**Biopolymers from Natural Resources**

Editors

**Rafael Antonio Balart Gimeno Marina Patricia Arrieta Dillon Daniel Garc´ıa Garc´ıa Lu´ıs Jes ´us Quiles Carrillo Vicent Fombuena Borr´as**

MDPI • Basel • Beijing • Wuhan • Barcelona • Belgrade • Manchester • Tokyo • Cluj • Tianjin


Daniel Garc´ıa Garc´ıa Universitat Politecnica de ` Valencia (UPV) `

*Editorial Office* MDPI St. Alban-Anlage 66 4052 Basel, Switzerland

This is a reprint of articles from the Special Issue published online in the open access journal *Polymers* (ISSN 2073-4360) (available at: https://www.mdpi.com/journal/polymers/special issues/ Biopolymers Natural Resources).

For citation purposes, cite each article independently as indicated on the article page online and as indicated below:

LastName, A.A.; LastName, B.B.; LastName, C.C. Article Title. *Journal Name* **Year**, *Volume Number*, Page Range.

**ISBN 978-3-0365-3967-6 (Hbk) ISBN 978-3-0365-3968-3 (PDF)**

© 2022 by the authors. Articles in this book are Open Access and distributed under the Creative Commons Attribution (CC BY) license, which allows users to download, copy and build upon published articles, as long as the author and publisher are properly credited, which ensures maximum dissemination and a wider impact of our publications.

The book as a whole is distributed by MDPI under the terms and conditions of the Creative Commons license CC BY-NC-ND.

## **Contents**


Reprinted from: *Polymers* **2021**, *13*, 384, doi:10.3390/polym13030384 ................ **107**


Thermomechanical Behavior Reprinted from: *Polymers* **2020**, *12*, 2248, doi:10.3390/polym12102248 ................ **259**


Reprinted from: *Polymers* **2020**, *12*, 1097, doi:10.3390/polym12051097 ................ **385**

## **About the Editors**

**Rafael Antonio Balart Gimeno** (h-index 41, Scopus) is a faculty staff member in the Department of Mechanical and Materials Engineering, in the area of Materials Science and Metallurgical Engineering. He has been a Full Professor at Universitat Politecnica de Val ` encia since 2015. ` His research activity has been developed in the Institute of Materials Technology at UPV in the "RESEARCH GROUP ON POLYMERS AND GREEN COMPOSITES - GiPCEco", a group that he has led since its formation. This group focuses its activity on the following research lines, all of them with a marked environmental concern: (1) the potential of polymers of renewable origin and their transfer to industrial sectors; (2) minimizing the impact of additives and active ingredients of natural origin; (3) the development of materials under the "bio-refinery" concept; (4) the use of waste from the agro-food industry and agro-forestry sector through their incorporation into polymer and composite formulations; and (5) circular economy.

**Marina Patricia Arrieta Dillon** (h-index 30, Scopus) works at the Department of Industrial and Environmental Chemical Engineering at the School of Technical Industrial Engineering of the Technical University of Madrid, (ETSII-UPM) Madrid, Spain. She holds a Biochemistry BS from the National University of Cordoba (Argentina), a MS in Food Technology from the Catholic ´ University of Cordoba (Argentina), a MS in Polymer Science and Technology from UNED (Spain), and an international PhD in Science, Technology and Food Management from the Polytechnic University of Valencia (Spain), awarded with the Extraordinary PhD Thesis Award. She has been an AECID and Santiago Grisol´ıa predoctoral fellow, as well as a Juan de la Cierva (formacion and ´ Incorporacion) and UCM postdoctoral fellow. Dr. Arrieta has experience on the synthesis, processing ´ (melt-blending, extrusion, injection moulding, electrospinning, etc.), recycling, and characterization of multifunctional sustainable polymers and their nanocomposites with active and multifunctional properties for sustainable food packaging or agricultural applications.

**Daniel Garc´ıa Garc´ıa** (h-index 19, Scopus) is an Associate Professor of Materials Science and Engineering at the Department of Mechanical and Materials Engineering (UPV Campus d'Alcoi). He is a Technical Engineer in Industrial Design and Materials Engineering. He studied a Master's degree in Engineering, Processing and Characterisation of Materials and obtained an international Ph.D. in Engineering and Industrial Production with a national FPU grant at the UPV. He was awarded with the National Award for End of Degree in University Education by the Spanish Government and with the Extraordinary Doctoral Thesis Award from the UPV. He joined the Technological Institute of Materials (ITM) in 2009 and specializes on the development and formulation of sustainable polymers. His areas of interest include the extraction of active compounds from agroforestry residues for use as additives in biopolymers and the development and characterization of bionanocomposites.

**Lu´ıs Jes ´us Quiles Carrillo** (h-index 18, Scopus) works at the Department of Mechanical and Materials Engineering of the Escuela Politecnica Superior de Alcoy (EPSA) of the Universitat ´ Politecnica de Val ` encia, (UPV) Alcoy, Spain. He holds a degree in Mechanical Engineering, a Master's ` degree in Engineering, Processing and Characterisation of Materials, and obtained an international PhD in engineering and industrial production with a national FPU grant at the UPV. He was awarded the Best Doctoral Thesis in Polymers Award by the Royal Spanish Society of Chemistry and Physics (RSEQ and RSEF), granted by the group specialised in polymers (GEP).

His research work has focused on the search for new formulations of polymers with high environmental performance from waste and natural compounds. He has extensive experience in the characterisation of polymeric and composite materials, the reuse of waste and by-products for the creation of highly eco-efficient packaging, and works extensively with techniques such as electrospinning and polymer injection moulding.

**Vicent Fombuena Borr´as** (h-index 17, Scopus) works at the Department of Nuclear and Chemical Engineering of the Escuela Politecnica Superior de Alcoy (EPSA) of the Universitat ´ Politecnica de Val ` encia, (UPV) Alcoy, Spain. He holds degrees in Technical Engineering in Industrial ` Chemical Engineering and Materials Engineering. He joined the Technological Institute of Materials (ITM) in 2008 as a research technician and obtained an international PhD in Engineering and Industrial Production in 2012. His research work has focused on the implementation of a circular economy model in the polymer industry. He has focused his efforts on obtaining multiple active compounds from seeds and agroforestry residues to be applied in thermoplastic and thermosetting polymers. He has extensive experience in techniques of the chemical, thermal and mechanical characterization of materials.

## **Preface to "Biopolymers from Natural Resources"**

The increase in the production and application of biobased polymers has been posed as an extremely promising method of sustainable development goals, offering the option of replacing traditional petroleum polymers in several industrial sectors. In this context, there is great interest in biobased building blocks, including: the revalorization of agri-food wastes; biopolymers extracted directly from biomass, such as polysaccharides and proteins, as well as those produced by yeast biomass or by bacterial fermentation; medical devices; food packaging; agricultural films; membrane process applications, and so on. Despite the successful development of biobased polymers to an industrial scale, huge interest in their optimization still exists in order to extend their industrial exploitation. Therefore, significant attention has been given to the improvement of these sustainable plastics' overall performance, with the hope of further expanding their advantageous properties. As a versatile product, biobased polymers can be modified through chemical synthesis, copolymerization, or surface modification, as well as by the use of different additives, i.e., micro and nanoparticles, plasticizers, and active agents, among others.

This work covers all aspects related to recent, original, and cutting-edge research works focused on enhancing the performance of biopolymers from natural resources, and emphasising their potential use in industrial sectors, in line with the worldwide trend of building a circular economy. Each chapter is concerned with a given polymer's sustainable origin, in addition to their prospective obtainment, production, design, and processing at an industrial level, as well as possible improvements for specific industrial applications.

We acknowledge each author for their substantial contribution to the research field of biopolymers from natural resources.

### **Rafael Antonio Balart Gimeno, Marina Patricia Arrieta Dillon, Daniel Garc´ıa Garc´ıa, Lu´ıs Jes ´us Quiles Carrillo, and Vicent Fombuena Borr´as** *Editors*

## *Editorial* **Biopolymers from Natural Resources**

**Rafael Balart 1,\*, Daniel Garcia-Garcia 1,\*, Vicent Fombuena 1,\*, Luis Quiles-Carrillo 1,\* and Marina P. Arrieta 2,3,\***


During the last decades, the increasing ecology in the reduction of environmental impact caused by traditional plastics is contributing to the growth of more sustainable plastics with the aim to reduce the consumption of non-renewable resources for their production. Thus, in the last few years, the increased bio-based polymer production and application have positioned biopolymers as one of the most promising ways to meet the sustainable development goal of replacing traditional petroleum polymers with more sustainable materials in several industrial sectors.

In this context, the revalorization of agro-food wastes, biopolymers extracted directly from biomass such as polysaccharides, proteins and lipids, as well as those produced by yeast biomass, algae or by bacterial fermentation, have attracted a significant amount of interest, especially for medical devices, food packaging, agricultural films, membrane process applications, sustainable clothes and so on—interest which, despite successful recent developments in bio-based polymers up to the industrial scale, extends to their optimization for industrial exploitation. Therefore, significant attention has been paid to the improvement of those sustainable plastics directly derived from biomass overall performance to provide more than one advantageous property attributed to their versatility to be modified through chemical synthesis, copolymerization, and surface modification, as well as through the use of different additives (i.e., micro- and nanoparticles, plasticizers, and active agents, among others). Moreover, nowadays, the starting monomers or building blocks to obtain traditional plastics can be obtained from biomass instead of petrochemical resources, known as drop-in bioplastics. Drop-in bioplastics present the same advantages of traditional plastics in terms of performance and are more sustainable, while at the same time they can be directly transformed into the desired products by means of the same technology already available at the plastic industrial sector. Figure 1 summarizes several biobased polymers classified on the basis of their origin and obtainment process.

The present Special Issue gathers a series of twenty-four articles focused on the manufacturing and characterization of biopolymers and building blocks extracted from natural resources, as well as their potential scalability to the industrial level for several industrial applications. The Special Issue paid special focus on the improvement of the overall performance of biobased polymers by chemical modification process, the incorporation of sustainable additives such as biobased oligomers, by blending with another biobased or biodegrable polymeric matrix, as well as by the development of composites and nanocomposites through the incorporation of naturally occurring particles and nanoparticles.

**Citation:** Balart, R.; Garcia-Garcia, D.; Fombuena, V.; Quiles-Carrillo, L.; Arrieta, M.P. Biopolymers from Natural Resources. *Polymers* **2021**, *13*, 2532. https://doi.org/10.3390/ polym13152532

Received: 30 June 2021 Accepted: 20 July 2021 Published: 30 July 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

1

**Figure 1.** Classification of biobased polymers.

Among other naturally derived biopolymers, polysaccharides are nontoxic and biodegradable, which increases their potential application in several fields. Polysaccharides, with cellulose as the most important representative since it is the most abundant polymer in nature and is found all around the world, are widely investigated. Although the most-used biopolysaccharides are obtained from plant origin (e.g., cellulose), they can be obtained from animal origin (e.g., chitin/chitosan) and also from microbial origin (e.g., bacterial cellulose). In this sense, the production of biobased polymers by using microorganisms has gained considerably attention during the last decade as a sustainable production method. In this regard, Irbe et al. [1] developed novel blends based on biopolysaccharides derived from biomass, softwood cellulose fibers such as softwood Kraft (KF) or hemp fibers (HF) and fungal fibers (hyphae). Regarding the fungal fibers, they concluded that among several fungal fibers, the fibers from screened basidiomycota fungi *Ganoderma applanatum* (Ga), *Fomes fomentarius* (Ff), *Agaricus bisporus* (Ab), and *Trametes versicolor* (Tv) were good for blending with softwood cellulose fibers. Thus, the fungal fibers were blended with KF and HF fiber pulp in different mass ratios (i.e., 50:50 and 33:33:33). The materials were characterized in terms of mechanical properties, air permeability and virus filtration efficiency. The overall performance of the hyphal polysaccharide/cellulose blends was highly affected by the microstructural features of each cellulosic material, while the ability of functional groups of hyphal polysaccharides to establish hydrogen bonding interactions with cellulose fibers was one of the key elements of network formation which determined the final blend properties and the potential applications. In this context, highly fibrillated hemp fibers showed the highest mechanical strength, while the blends containing Kraft fibers increased air permeability and showed a high virus reduction capacity. Thus, depending on the intended use the different microstructure of Hemp fibers or Kraft fibers will allow obtaining materials with interesting properties. For instance, materials with a low air permeability are interesting for food packaging applications where high gas barrier properties are of fundamental importance. Meanwhile, materials with a higher air permeability and virus filtration efficiency have potential for being used as gas permeable membranes, for example, as biobased filter layers in face masks. Despite its outstanding characteristics, cellulose in its native form is hard and provides some limitations for the plastic industry in terms of cellulose processing. Nevertheless, cellulose derivatives can overcome such problems. Thus, cellulose chemical modifications have been widely

studied and nowadays several cellulose derivatives can be found commercially such as carboxymethyl cellulose (CMC), methyl cellulose (MC), hydroxyethyl cellulose (HEC), hydroxypropyl methylcellulose (HPMC), and cellulose acetate (CA).

Among others, CMC possesses good film-forming abilities because of the presence of a hydrophilic carboxyl group (–CH2COONa), which allows its dissolution in water. In this context, Rachtanapun et al. [2] revalorized coconut juice, a waste product from the coconut milk industry, into bacterial cellulose using *Acetobacter xylinum* to further obtain carboxymethyl cellulose (CMC) through chemical synthesis with different degrees of substitution. With this propose in mind, CMC was obtained from bacterial cellulose in the presence of monochloroacetic acid (MCA) via a carboxymethylation reaction and by simply varying the sodium hydroxide content (NaOH from 20% to 60%), the degree of substitution was controlled. The degree of substitution significantly increased with NaOH concentration up to 30% and then progressively decreased with further NaOH concentrations. As CMC is a water soluble cellulose derivative, the higher the degree of substitution, the higher the water vapor permeability. The flexibility of the material also increased with the degree of substitution. Thus, the developed method to modify cellulose from nata de coco allows tuning of the water permeability and the mechanical properties of the final material according to the need of the intended application by simple varying the NaOH content. CMC with different degrees of substitutions has been also obtained by Kluklin et al. [3] from the *Asparagus officinalis* stalk end, a typical herbal medicine used in Asia, by simply varying the NaOH content (from 20% to 60%) too, and it was also found that there is a higher degree of substitution for the NaOH 30% (*w*/*v*). CMC has been also used in the development of solid polymer electrolyte (SPE) film for sustainable energy storage systems due to its solubility in water and its naturally high degree of an amorphous phase that allows easier transport of conducting ions like lithium (Li+) and proton (H+). In this context, Sohaimy et al. [4] tuck this advantageous properties of CMC to develop CMC-based films for electrochemical applications. However, the inherent solid properties of the polymeric film impede ionic mobility leading to a low ionic conductivity, especially at room temperature. Thus, ammonium carbonate (AC), with two groups of ammonium ions, was used as a dopant in amounts from 1 wt.% to 11 wt.% with the main objective of injecting more H<sup>+</sup> into the system and the CMC-AC thin electrolyte film was obtained by a solvent casting method. Both CMC and AC salts dissolved completely and had a low degree of crystallinity (Xc) as it was demonstrated by X-ray diffraction (XRD). Although, the ionic conductivity for a biopolymer electrolyte intended for commercial energy storage technology needs to be at least a minimum value of ~10−<sup>4</sup> Scm<sup>−</sup>1, they obtained promising results since the highest ionic conductivity obtained for CMC doped with 7 wt.% was 7.71 × <sup>10</sup>−<sup>6</sup> Scm−1, resulting in higher results than other diammonium salts tested (i.e., ammonium adipate and ammonium sulfate), showing that further enhancement is still needed to increase the ionic conductivity for suitable electrochemical applications (~10−<sup>4</sup> Scm−1) and the CMC-AC films show promising prospects for electrochemical applications.

Panraksa et al. [5] developed orodispersible films (ODF) based on hydroxypropyl methylcellulose (HPMC: E5 and E15) by extrusion-3D printer technology to enhance the solubility and dissolution of poorly water-soluble drugs (i.e., phenytoin). ODFs are required to rapidly disintegrate in the buccal cavity (i.e., within a minute), without requiring water, thus HPMC, which is a well-known hydrophilic and film-forming polymer, is an ideal candidate for this propose. HPMC E5 and E15 were used as the film-forming polymers, while propylene glycol and glycerin were used as plasticizers. They optimized the processing conditions and the results showed that the phenytoin-loaded HPMC E15 (at a concentration of 10% *w/v* of HPMC E15) was the most suitable formulation of HPMC, since it exhibited a good physical appearance, good mechanical strength, low Young's modulus and high elongation at break, rapid in vitro disintegration time (within 5 s) as well as a rapid drug release (up to 80% within 10 min), showing the improvement of

solubility and dissolution rate of phenytoin at the same time as allowing ease of handling and application.

Another widely used cellulose derivative is cellulose acetate (CA). For instance, it has been used by Arrieta et al. [6] as a carrier of *Cucumis metuliferus* fruit extract to provide to corona-treated low-density polyethylene (LDPE) films with antioxidant activity by coating a CA-based polymeric solution, containing different amounts of *Cucumis metuliferus* fruit extract (1, 3 and 5 wt.%), onto the LDPE film. The materials showed antioxidant activity (between 0.3 and 0.6 mg Trolox/dm2) after performing migration studies in a fatty food simulant and also showed their ability to reduce the browning effect of fresh-cut apples in direct contact. Additionally, the coating not only provided an antioxidant active functionality, but also improved the oxygen barrier performance to the final bilayer material, showing their use in active food packaging applications.

Another interesting approach to obtain a solution of insoluble cellulose is regenerated cellulose (RC). RC refers to the chemical dissolution of insoluble natural cellulose followed by the recovery of the material from the solution by means of different methods, such as cellulose carbamate processes, lyocell processes, and viscose processes. In this context, Husseinsyah et al. [7] used an ionic liquid (1-butyl-3-methylimidazolium chloride) as a novel method to obtain RC from empty fruit bunches (EFB) to develop a self-reinforced composite for biodegradable packaging applications. They studied the possibility of improving the interphase bonding between the cellulosic matrix and the filler in the selfreinforced composite by means of EFB chemical treatment with a methyl methacrylate (MMA) chemical treatment. They observed that the substitution of the –OH group of the EFB cellulose with the ester group of the MMA allowed the dissolution process of the EFB during the regeneration process owing to weakening of the hydrogen bonding of the cellulose. This allowed us to obtain a more crystalline and homogeneous EFB RC biocomposite film, with an improved thermal stability and mechanical performance. Another very important cellulosic material is cotton, which is widely distributed worldwide. It has a higher yield compared to other commercial crops, and thus several products are obtained from cotton seed, such as oil. In this context, Chen et al. [8] used bio-based cottonseed oil (CO) and two of its derivatives, fatty acid of cottonseed oil (COF) and sodium soap of cottonseed oil (COS), to tune the density and mechanical strength of polysulfide polymers by means of a free radical addition mechanism. COF reacts with sulfur to generate serials of polysulfide-derived polymers with a lesser viscosity and tensile strength. Whereas COS was not involved in the reaction with sulfur, as a consequence of the high melting point of sodium linoleate and sodium oleate. Nevertheless, COS it was able to increase the density and tensile strength of polysulfide-derived polymers when it was used as a filler. Moreover, the developed polysulfide-based polymers showed good reprocessability and recyclability, showing their potential as bio-based functional supplementary additives.

Ammar et al. [9] developed a novel adsorbent material for cationic dyes based on a polyelectrolyte multi-layered (PEM) system cross-linked to a cellulosic-based material. The PEM was formed by an alternation of layers of two polysaccharides, sodium alginate polyanion and reticulated citric acid with k-carrageenan, thus providing many carboxylate and sulfonate groups grafted on the cellulosic surface. The developed system provides outstanding adsorption capacities for cationic dyes (e.g., with a capacity above 522.4 mg/g for methylene blue).

The processing and revalorization of lignocellulosic materials obtained from agro-food wastes, as well as the extraction through sustainable approaches, have gained considerable interest. Gómez-Patiño et al. [10] extracted cellulosic cutins from of *S. aculeatissimum* and *S. myriacanthum* fruit peels, using enzymatic (i.e., *A. niger pectinase*, *A. niger cellulose* and *A. niger hemicellulose*) and chemical (trifluoroacetic acid hydrolysis) methods. They selected these fruits since they are not for human or animal consumption and are thus interesting as a potential raw material to produce sustainable materials. The hydrolyzed cutins showed mainly 10,16-dihydroxyhexadecanoic acid (10,16-DHPA) monomers composition and the obtained films showed a good homogeneity.

Lignocellulosic derivatives have also gain considerable interest as micro and nanofillers. In this context, Adamcyk et al. [11] studied the extraction of lignin from wheat straw, spruce, and beech using ethanol organosolv pretreatment at temperatures from 160–220 ◦C, and further precipitated by solvent-shifting obtaining lignin micro- and nanoparticles (with mean hydrodynamic diameters from 67.8 nm to 1156.4 nm). They observed that higher pretreatment temperatures increased the delignification of the raw materials but also favor depolymerization and structural alteration of the extracted lignin, increasing the particle sizes as well as agglomeration at pretreatment temperatures of over 200 ◦C. The obtained particles were then purified by dialysis and the showed interesting antioxidant activity (i.e., from 19.1 and 50.4 mg Lignin/mg ascorbic acid equivalents).

On the other side, lignocellulosic nanoparticles have been obtained from yerba mate (*Ilex paraguariensis*) waste, a typical infusion widely consumed in Latin America (i.e., Argentina, Brazil, Uruguay and Paraguay), by means of an aqueous extraction procedure (with mean average size of 495 nm). The obtained nanoparticles were used by Beltrán et al. [12] to reinforce mechanically recycled bioplastic (i.e., polylactic acid). Poly(lactic acid) (PLA) is chemically synthesized starting from simple sugars obtained from biomass fermented to lactic acid. It is one of the most promising biopolyesters for massive industrial applications due to its availability in the market at a competitive cost, and it can be recycled following the mechanical recycling process used for other traditional plastics (e.g., polyethylene terephthalate, PET). Low amounts of yerba mate nanoparticles (i.e., 1 wt.%) were able to increase the intrinsic viscosity and to improve the oxygen barrier properties of mechanically recycled PLA, showing the interest of such nanocomposites in the food packaging field. Ramos et al. [13] also studied PLA-based nanocomposites for food packaging applications. In this case, PLA was loaded with commercial montmorillonite (D43B) at two different concentrations (2.5 and 5 wt.%) and was further incorporated with thymol as an antioxidant and antimicrobial agent for active food packaging applications. The results showed that the addition of 2.5 wt.% D43B and 8 wt.% thymol leads to a material with a good balance of properties, with antibacterial activity against *Escherichia coli* and *Staphylococcus aureus.*

Another interesting source for sustainable additives are seeds fruits, peel and/or nut's shells, from which can be obtained flours and natural particles and/or nanoparticles, as well as vegetable oils. In this regard, Jorda-Reolid et al. [14] obtained micronized Argan shell (MAS) particles from residues of the *Argania spinosa* plant, a tropical plant whose fruit is used in Morocco to prepare oil. The authors used MAS particles (with mean average long of 70 μm and mean average wide of 45 μm) as reinforcing fillers of high-density polyethylene obtained from sugarcane (Bio-HDPE). Meanwhile, to improve the low compatibility between MAS and Bio-HDPE, they studied two compatibilizing agents, polyethylene-grafted maleic anhydride (PE-g-MA) and maleinized linseed oil (MLO); as well as halloysite nanotubes (HNTs), a second reinforcing filler. The obtained materials showed improved stiffness, high ductility, and good thermal stability, as well as a visual appearance very similar to that of reddish-color woods, showing their interest as wood plastic composites with the additional advantage of being chipper rather than neat Bio-HDPE. González Martínez et al. [15] optimized the carbonation reaction of epoxidized linseed oil (ELO) with carbon dioxide (CO2) in the presence of tetrabutylammonium bromide (TBAB) as catalysts, to obtain carbonated vegetable oils (CVOs), which have potential applications as interesting starting materials for the formation of polymers (e.g., monomers, additives, lubricants and plasticizers), particularly in the synthesis of nonisocyanate polyurethanes (NIPUs). They were able to obtain large conversion (96%), carbonation (95%), and selectivity (99%) at low temperature (90 ◦C), and thus the high carbonate content obtained for carbonated ELO showed their promising performance for the synthesis of NIPU with the required properties and more sustainable characteristics. On the other hand, Dominici et al. [16] studied the possibility of obtaining bioplastics

from wheat flour particles as a novel source and an energetically and economically cheap alternative to other thermoplastics. The refined flours, with different contents of bran, were first plasticized with glycerol and then the authors attempted to find how different contents of grinded bran could affect the deformability of the flour by blending with low melting polymeric fractions (such as poly(ε-caprolactone) (PCL), and polybutyleneadipate-*co*-terephthalate (PBAT)). They studied different processing parameters and they were able to obtain a novel thermoplastic wheat flour (TPWF), which showed an improved performance when glycerol was partially replaced by water and when it was blended with PCL, as well as adding citric acid as a compatibilizer.

Ferrándiz et al. [17] studied the potential use of soy protein (SP) fibers functionalized with an undisclosed antimicrobial agent as well chitin (with inherent antimicrobial properties) fibers as sustainable active textiles to replace synthetic polymers widely used in this sector. The thermal resistance of both weft-knitted fabrics was similar to that of cotton, whereas their air permeability was higher, particularly in the case of chitin due to its higher fineness, which makes these natural fibers very promising for summer clothes.

After cellulose, starch is the second most abundant polymer in nature and, thus, has been widely studied for several sustainable industrial applications. Zhang et al. [18] developed potato starch-based foams by means of a microwave treatment. They blended potato starch with chitosan as a reinforcing phase and it was found that starch-based foams with a larger proportion of starch showed a small pore size and a low density with higher compressive strength ascribed to the good compatibility between both polymeric matrices due to the formation of hydrogen bond interactions between amino groups of chitosan and hydroxyl groups of starch. Those interactions were also responsible to maintained intact the morphological structure of the foam in water for 10 days, while the starch-based foams completely degraded in water after 30 days. Thus, the developed starch/chitosan foams resulted in interesting results for their use as active materials in biomedical applications as well as in drug and food packaging.

For several industrial applications, starch is particularly used in its thermoplastic form, known as thermoplastic starch (TPS). To obtain TPS, native starch requires disruption of the granule organization by means of a combination of high water content and heat, leading to the starch granule swelling and consequently starch gelatinization. In this regard, Crucean et al. [19] studied the gelatinization process of three native starches: (i) wheat, (ii) potato, and (iii) waxy corn starch, as well as starch from a wheat flour in mixtures of water and choline chloride (an ionic compound, which has a "structure making" effect). They demonstrated that choline chloride/water system exhibits an allotropic change at low water concentrations and solubilization for water contents greater than 30%. They observed a stabilization, or a better organization, of the ordered regions of the starch in the presence of choline chloride as it was demonstrated by an X-ray diffraction analysis in a heating cell that the crystalline rearrangement of the structure of the starch grain takes place simultaneously with the solubilization phenomenon of amylose.

Osman et al. [20] studied thermoplastic starch (TPS)-based biocomposites using dolomite (DOL) as a filler in its pristine form (DOL(P)), as well as dolomites, after a simple and scalable sonication process (DOL(U)). TPS-DOL biocomposites intended for packaging applications were prepared at different loadings (i.e., 1, 2, 3, 4 and 5 wt.%). The TPS-based biocomposites with a high dolomite loading (i.e., 4 and 5 wt.%) showed better mechanical performance, showing greater tensile and tear properties, particularly in the case of biocomposites loaded with a sonicated process assisted by dolomite due to a reduction in particle size allowing their better dispersion within the TPS matrix. Thus, the high abundancy and low cost of dolomite, in combination with the simple, scalable and environmental friendly method of sonication, showed their interest to be of use as reinforcing fillers for TPS to produce a sustainable biocomposite for packaging applications. Although biomass-derived biopolymers are mainly extracted from plants, they can also be obtained from animal resources. For instance, gelatin has been traditionally used in the development of soft capsules (softgels) due to its biodegradable nature, as well as its ability to form

thermo-responsive hydrogels, inspiring interest in the biomedical sector and food industry. In recent years, new materials have been explored in the food industry to partially or completely replace gelatin with other non-animal natural hydrocolloids. In this context, Otálora et al. [21] developed gelatin/cactus mucilage softgel capsules to encapsulate oil extracted from sacha inchi (*Plukenetia volubilis L.*) seeds, a rich source of polyunsaturated fatty acids (PUFAs) that are beneficial to human health. They submitted the softgel capsules to an in vitro digestion process to simulate gastric conditions and the protective capacity of the gelatin/cactus mucilage-based softgel against digestive processes was evaluated. Although the study revealed a reduction in the content of polyunsaturated fatty acids (PUFAs) after the digestion process and, thus, a reduction of the nutritional value, they concluded that gelatin/cactus mucilage microcapsules can act as interesting bioactive delivery systems for acidic food (e.g.,: fruit juices or dairy drinks) before being subjected to digestive processes.

As has already been mentioned, the biobased polymer obtainment, by means of microorganisms, has gained a lot of interest. In this context, the family of polyhydroxyalkanoates (PHAs) are polyesters biologically synthesized by controlled bacterial fermentation in response to nutrient limitation as an intracellular storage of food and energy. The homopolymer, poly(hydroxybutyrate) (PHB), is the simplest and most common representative of the PHA family, but PHAs also comprise many copolyesters, polyhydroxybutyrate-*co*hydroxyalkanoates, that have gained high industrial interest in the bioplastic sector, such as polyhydroxybutyrate-*co*-hydroxyvalerates (PHBV), or polyhydroxybutyrate-*co*-hydroxyhexanoate (PHBH), polyhydroxybutyrate-*co*-hydroxyoctonoate (PHBO), and polyhydroxybutyrate-*co*-hydroxyoctadecanoate (PHBOd). For instance, Ivorra-Martinez et al. [22] developed PHBH/PCL blends with different compositions ranging from 0 to 40 of PCL wt.% by means of extrusion process followed by injection molding, with the main objective of reducing the inherent embrittlement of PHBH produced due to the aging process (i.e., secondary crystallization). Although they mainly found a lack of miscibility between both polymeric matrices, as revealed by the thermogravimetric analysis, they concluded that the addition of high amounts of PCL (i.e., 40 wt.%) contributed to an increase in ductility and to a decrease in the typical brittleness of PHA. PCL considerably improved the toughness, as well as the impact resistance, of neat PHBH. Therefore, PHBH/PCL blends showed their suitability for the packaging industry. Then, Ivorra-Martinez et al. [23] developed Wood Plastic Composites (WPCs) based on PHBH loaded with lignocellulosic particles of almond shell flour (ASF), a by-product from the agro-food industry, by means of extrusion process followed by injection molding. Since the addition of ASF (with a maximum particle size of 150 μm) leads to an embrittlement and reduced toughness of the WPC, they used oligomeric lactic acid (OLA) to provide improved properties to the final PHBH-ASF/OLA material. Interestingly, a remarkable increase in impact strength with 20 phr OLA addition was achieved. In fact, OLA provides PHBH polymer chains of a high mobility due to a plasticizing effect, leading to an improvement in toughness, even on composites with 30 wt.% ASF. Additionally, OLA decreased the water absorption capacity of WPC, thus broadening potential industrial applications of PHBH-ASF/OLA in high humidity environments.

As it was previously mentioned, nowadays many traditional plastics, which were obtained for decades through the classic petrochemical production routes, have recently found "green" routes for their production (e.g., Bio-PE, Bio-PET, etc.). In this regard, polyamide 6 (PA6) can be replaced with a biobased variant that exhibits a high renewable content and with very similar properties known as PA610. In this context, Marset et al. [24] used PA610 partially biobased polyamide (Bio-PA) to develop nanocomposites reinforced with 10%, 20%, and 30% halloysite nanotubes (HNTs) as a natural flame retardancy filler by means of an extrusion process, followed by injection molding. The results showed that HNTs promote a significant reduction in the optical density and in the number of toxic gases (i.e., CO2) emitted during combustion, especially when PA610 was loaded with 30% HNTs. Thus, the nanocomposites showed good flame retardancy properties.

In summary, the Special Issue Biopolymers from Natural resources published in *Polymers*, compiles the recent research works in biopolymers obtained from natural resources and the strategies to improve their overall performance for sustainable industrial applications. From the above, improvements of biobased polymers was successfully achieved by means of copolymerization, blending, the use of compatibilizers, or plasticizers (vegetable oils, oligomers, etc.), as well as with naturally occurring or naturally derived particles on the micro- and nano-scale. Special interest has been shown to the revalorization of agro-food wastes for the extraction/production of additives or polymers with an interest in the industrial sector. The resulting, more sustainable materials can find potential applications in several industrial sectors, such as in the controlled release of active compounds for active food packaging, bioactive delivery systems, or biomedical devices (e.g., soft capsules, etc.), textiles (e.g., sustainable clothes) or as wood plastic composites (e.g., furniture and outdoor applications).

Future research efforts on biobased polymer extraction and production methods, as well as the optimization of their final production into co-polymers, blends, composites, and nanocomposites, are still needed to properly transfer the proposed developments, as well as future ones from the laboratory scale level to the industrial sector to reach more environmentally friendly materials within a circular economy approach.

**Author Contributions:** All the guest editors read the twenty-four articles of the Special Issue and wrote this editorial letter. All authors have then reviewed the final version of the manuscript and agreed its publication. All authors have read and agreed to the published version of the manuscript.

**Acknowledgments:** The guest editors thank all the authors for submitting their work to this Special Issue and for its successful completion. We also acknowledge all the reviewers participating in the peer-review process of the submitted manuscripts for enhancing their quality and impact. We are also grateful to Chris Chen and the editorial assistants of *Polymers* who made the entire Special Issue creation a smooth and efficient process.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


## *Article* **Influence of the Presence of Choline Chloride on the Classical Mechanism of "Gelatinization" of Starch**

**Doina Crucean 1,2,3, Bruno Pontoire 2,3, Gervaise Debucquet 4, Alain Le-Bail 1,3 and Patricia Le-Bail 2,3,\***


**Abstract:** The aim of this research is to contribute to a better understanding the destructuration of three native starches and a wheat flour in mixtures of water and choline chloride. Model systems have thus been defined to allow a better approach to hydrothermic transformations related to the interactions between choline chloride and starch. We have observed that choline chloride has an impact on the gelatinization of starch which corresponds to the stabilizing salts phenomenon. The depolymerization and dissolution of the starch have also been demonstrated and can there dominate the gelatinization. However, the results obtained in X-ray diffraction by heating cell have shown that the exotherm which appeared was not only related to the depolymerization of the starch, but that a stage of crystalline rearrangement of the starch coexisted with this phenomenon.

**Keywords:** allotropic transition; choline chloride; plasticizer; starch dissolution

### **1. Introduction**

Choline chloride, because of its unique nutrient functionality and its ability to enhance flavor, provides a new option as a partial substitute for sodium chloride in reformulations of processed foods to reduce their sodium content. To date, very few studies have explored the potential of choline chloride to act as a substitute for salt. Locke and Fielding were the first to identify the properties of choline chloride to improve the salty taste of food products [1–3].

Starch is a natural polymer with particular properties unlike those of traditional polymers. As a heterogeneous material, it has macromolecular structures bound in a granular superstructure.

When native starch granules are heated in water (excess of water conditions, around 60% water wb), their semicrystalline nature and three-dimensional architecture are gradually disrupted, resulting in a phase transition from the ordered granular structure to a disordered state in water, which is known as "gelatinization" [4,5]. Gelatinization is an irreversible process that includes several steps, such as granule swelling, native crystalline melting (loss of birefringence), and leaching of fractions of polymers or of polymers in particular amylose chains resulting in a molecular solubilization [6]. The gelatinization process is essential in the processing of foods. For improving the properties of starch and adding new functionalities, it is common to carry out the hydrothermal transformation of starch in environments containing substances other than water. Various plasticizers and additives for starch processing have been used, including polyols (glycerol, glycol, sorbitol, etc.) and nitrogen-containing compounds (urea, ammonium derived, and amines) [7,8]. An alternative class of materials known as ionic liquids, now commonly defined as salts, that

**Citation:** Crucean, D.; Pontoire, B.; Debucquet, G.; Le-Bail, A.; Le-Bail, P. Influence of the Presence of Choline Chloride on the Classical Mechanism of "Gelatinization" of Starch. *Polymers* **2021**, *13*, 1509. https:// doi.org/10.3390/polym13091509

Academic Editors: Rafael Antonio Balart Gimeno, Daniel García García, Vicent Fombuena Borrás, Luís Jesús Quiles Carrillo and Marina Patricia Arrieta Dillon

Received: 21 March 2021 Accepted: 28 April 2021 Published: 7 May 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

melt below 100 ◦C has recently attracted much interest for the processing of biopolymers such as starch.

Choline chloride solubilized in water gives an ionic liquid. Several studies have shown the effects of ionic liquids (solvents and plasticizers) based on the effects of choline chloride on starch [9–14]. A study conducted by Decaen et al. highlights the plasticization of starch by ionic liquids based on an examination of choline [15]. The destructuration of native maize starch in mixtures of water and ionic liquids containing choline cations was studied in dynamic heating conditions, combining calorimetry, rheology, microscopy, and chromatographic techniques by Sciarini et al. [16].

This work presents the mechanism of choline chloride penetration into the starch grain and its impact on gelatinization and the structural evolution of starch during heating kinetics.

#### **2. Materials and Methods**

#### *2.1. Materials*

Wheat flour (type 65) was supplied by Evelia, La Varenne, France. Wheat starch was purchased from Loryma GmbH (Zwingenberg, Germany), and waxy corn starch was purchased from Roquette Italia SpA (Cassano Spinola, Italy). Potato starch and choline chloride (C5H14ClNO) were supplied by Sigma-Aldrich (France).

### *2.2. Sample Preparations*

The measurements were carried out on wheat flour (Fw) and wheat, potato, and waxy corn starch (Sw, Sp, and Swc, respectively) solubilized in an aqueous solution containing different concentrations of choline chloride.

The model systems studied being composed of three ingredients (flour, chlorine choline, and water) and were presented under the form of Equation (1) with Fx, Ccy, and Wz being the flour (or starch), choline chloride, and water content, respectively.

$$\text{F}\_{\text{x}}\text{C}\text{c}\_{\text{Y}}\text{W}\_{\text{x}}\tag{1}$$

A mixing design has been used (1). The factors are the concentrations of each component of the mixture. Responses are expressed as a function of these concentrations with the sum of the mass fraction of each component being equal to 100 (x+y+z= 100). For each analysis, 5 g of mixture was prepared.

The mode of preparation (quoted ILS = "Ionic Liquid + Starch" for the rest of the paper) consisted in adding the choline chloride under the form of ionic liquid [15] (previous dissolution of Cc and water for one hour before starting the experiments); in that case, the script that was used was "(Cc + W)", and the ionic liquid was then added to the flour. These systems were written with the following script: Fx[CcyWz].

All analyses were repeated in triplicate. A statistical analysis was performed with variance analysis (*p* < 0.005) to detect significant differences.

#### *2.3. Methods*

#### 2.3.1. Differential Scanning Calorimetry Measurements

Samples weighing 500 mg were placed in stainless steel pans, and a reference cell was prepared by adding an amount of water equal to the volume of water in the sample cells. The cells were then sealed with a heat-resistant seal. The cell containing the sample was agitated (50 rpm at room temperature) for one hour and then placed in the oven of the appliance. The pans were heated at a rate of 1 ◦C/min from 10 to 120 ◦C by using the SETARAM microcalorimeter (μDSC) III (France). Scans were run on the following starch suspensions: SP20[CcxWy], SW20[CcxWy], SWC20[CcxWy], and FW20[CcxWy]. All measurements were made at least in triplicate.

Thermal transitions were defined as TGe (peak temperature gelatinization), and ΔHGe denotes the transition enthalpies for different endotherms. The enthalpies were determined by endotherm integration, and all traces were normalized to 1 mg of dried sample. The partial enthalpies were calculated from the onset of the endotherm to the end (every 1 ◦C).

#### 2.3.2. X-ray Diffraction

The samples containing flour and starch suspended in [Cc56W24] were prepared as follows: SWC20[Cc56W24], SP20[Cc56W24], SW20[Cc56W24], and FW20[Cc56W24]. The appropriate amounts of choline chloride, water (when required), and starch were weighed. The same formulations without choline chloride were studied: SWC45W55, SP45W55, SW45W55, and FW45W55, corresponding to the same F/W ratio as the previous mixtures.

The samples were examined by wide-angle (WAX) X-ray diffraction. The measurements were performed using a D8 Discover spectrometer from Bruker-AXS (Karlsruhe, Germany). Cu Kα<sup>1</sup> radiation, produced in a sealed tube at 40 kV and 40 mA, was selected and parallelized using a double Gobël mirror parallel optics system and collimated to produce a 500 μm beam diameter with sample alignment by a microscopic video and laser. The data were monitored by a VANTEC 500 2D detector (Bruker, Karlsruhe, Germany) for 10 min and normalized.

The sample was placed in a capillary with a diameter of 1.5 mm, and a second capillary was introduced into the first to avoid loss of water during the rise in temperature; the two capillaries were sealed. The as-prepared capillary was placed in a heating stage HFS91 (Linkam, Tadworth, UK). The detector was positioned at a focusing distance of 8.6 cm from the sample surface. It was in direct beam position [17]. The heating kinetic applied to the sample was 1 ◦C/min from 20 to 120 ◦C. Every 10 ◦C, a plateau of 10 min was introduced to enable the acquisition of the diffraction spectrum.

### **3. Results and Discussion**

### *3.1. Gelatinization of Different Starches*

The thermograms of wheat flour (FW), wheat starch (SW), potato starch (SP), and waxy-corn starch (SWC) in excess water (W = 80%) should serve as a reference and allow comparison with the F20[CcxWy] systems, i.e., when choline chloride is added to the aqueous phase. All the samples exhibited an endotherm on their thermograms, which corresponded to the gelatinization of the starch at the following temperatures: FW 61.7 ± 0.2 ◦C, SW 58.1 ± 0.1 ◦C, SWC 70.8 ± 0.1 ◦C, and SP 65.7 ± 0.1 ◦C (Table 1). The gelatinization temperatures of the different samples studied increased in the following order: TGe (SW) < TGe (FW)<TGe (SP)<TGe (SWC), whereas the gelatinization enthalpies increased in the following order: ΔHGe (FW) < ΔHGe (SW) < ΔHGe (SWC) < ΔHGe (SP).


**Table 1.** Enthalpies (ΔHGe, J/g) and gelatinization temperatures (TGe, ◦C) of different flour/starch and plasticizer model systems. Enthalpies were calculated on flour or starch total dry basis.

The thermograms of the FW and SW samples showed a second endotherm characterized by the fusion of the amylose-endogenous lipid complexes formed during gelatinization [18,19]. The melting temperatures of the amylose-lipid complexes for SW and FW were 97 ± 1 ◦C and 92.3 ± 0.2 ◦C, respectively.

#### *3.2. Influence of Choline Chloride on Starch Destructuration*

Since choline chloride behaves as an ionic liquid when it is in the presence of water, its behavior during the heating of a starch granule suspension may strongly influence gelatinization. A study comparing the previously discussed aqueous-phase suspensions containing solubilized choline chloride was carried out.

Micro-DSC was used to study the physicochemical characteristics of the following suspensions: SWC20[CcxWy], SP20[CcxWy], SW20[CcxWy], and FW20[CcxWy].

The thermograms of the suspensions are shown in Figure 1, and their enthalpies ΔHGe and gelatinization temperatures TGe are shown in Table 1.

**Figure 1.** μDSC heating curves, 10–120 ◦C at 1 ◦C/min, of: (**a**) waxy-corn starch, (**b**) potato starch, (**c**) wheat starch, and (**d**) flour at different water and chlorine chloride contents.

At low concentrations of Cc, the starch undergoes a typical gelatinization, represented by an endothermic transition. At the same concentration, the endothermic transition attributed to the fusion of endogenous amylose-lipid complexes is also observed for FW and SW. When the concentration of Cc increases from 0 to 40% for SWC, SW, and FW and from 0 to 48% for SP, the peak of gelatinization moves to a higher temperature and then decreases for all three starches and flour. The observed gelatinization temperature is always higher than that observed in pure water. However, beyond 40% for SWC, SP, and FW and beyond 48% for SW, the gelatinization of the starch decreases with an increase in the B4 concentration. The same trend is observed for the enthalpies of gelatinization.

The increase in gelatinization temperature caused by Cc and the subsequent decrease with an increasing Cc level agree with the results of Chiotelli et al. for wheat and potato starches, Chungcharoen and Lund for rice starch, Jane for corn starch, and those of Ahmad and William and Ghani et al. for sargo starch [20–24].

The shift of the gelatinization temperature toward higher temperatures as well as the increase in the total enthalpy of gelatinization in the presence of Cc may be due to the reduction of the water activity in the starch/plasticizer solution, which causes an increase in the energy required for chemical and physical reactions involving water [25] Jane [22] claimed that the effect of salts on starch gelatinization follows the "Hofmeister" series, the increase in the gelatinization temperature follows the charge density order of the ions (LiCl > NaCl > KCl > RbCl), that is, the order of the "structure making" effect. However, at concentrations greater than 4 M, these salts decrease the onset temperature of gelatinization. Choline reacts as an "ion structure maker".

For high Cc concentrations, exothermic transitions were observed. The exotherms started at higher temperatures as the water content decreased, and the corresponding heat released (ΔH) increased as well. There was a critical concentration, which depends on the type of starch used, for which both exothermic and endothermic transitions took place. For this concentration, both phenomena seem to happen at very close temperatures, at which they overlapped, resulting in a neutralization with no visible thermal effect. The same behavior was found by Sciarini et al. and Mateyawa et al. working with choline acetate and EMIMAc and by Koganti et al. using N-methyl morpholine N-oxide (NMMO) [12,16,26]. These authors attributed the exothermic transition to starch dissolution in these solvents. For low water contents (16%), the thermograms of SP and SWC show wide endotherms at low temperatures (19.7 ◦C). This phenomenon was not visible for SW and FW. A DSC study was then carried out on the SWC20[CcxWy] systems (with x included between 56 and 70 and y between 24 and 10) to understand this phenomenon (Figure 2).

The thermogram of pure Cc (data not shown) shows an endotherm near 73.6 ± 0.1 ◦C. With the addition of water, this endotherm moves to lower temperatures and undergoes a strong shift from the baseline until disappearing. In the presence of waxy maize starch (20%), the concentration of Cc was increased in steps of 1% (from 56 to 70%), and water was decreased likewise (from 24 to 10%), which produced an endotherm at low temperatures for the sample with 17% water.

During these investigations, the Cc was solubilized in water to form the plasticizer, and then the starch was added to the plasticizer. When the amount of water was sufficient, complete solubilization of the Cc was achieved. When the water concentration was less than 17%, the solubilization of Cc was partial, and we observed a solubilization endotherm that broadened with a decrease in the water content. For the 14% water sample, there was enough unsolubilized choline chloride for the allotropic change in choline chloride at 68 ◦C to be observed.

It was assumed that the solution (water + Cc) became more viscous with increasing Cc concentrations, resulting in more difficulty for the Cc solution to penetrate inside the starch granule; in turn, a shift to higher gelatinization temperatures was observed.

**Figure 2.** Microcalorimeter heating curves, 10–120 ◦C at 1 ◦C/min, of waxy-corn starch at different water, and choline chloride contents.

The effect of chlorine choline (Cc) on the gelatinization of starch is very different from that of NaCl observed in the literature. Indeed, if the gelatinization temperature follows the same trend when adding Cc or NaCl, the enthalpy of gelatinization shows the opposite behavior. The decrease in the total enthalpy of starch gelatinization at high NaCl concentrations suggests a destabilization of the ordered regions of the starch in the presence of NaCl. As explained by Chiotelli et al., sodium chloride may be hindered by polymer–polymer interactions in favor of water–polymer interactions, resulting in a lower enthalpy for fusion of organized regions. The addition of Cc causes a reorganization of the internal structure of the starch grain [20].

#### *3.3. Penetration of the Plasticizer into the Starch Grain*

To understand the penetration kinetics of the ionic liquid (Cc + water) in the starch granule, the cumulative enthalpy curves were calculated and reported as a function of temperature.

The curves of the partial enthalpy of gelatinization for all the plasticizer contents studied are shown in Figure 3. This study presents the kinetics of the loss of ordered structure in starch for temperatures between 40 and 110 ◦C. An offset of the melting toward higher temperatures was observed when the Cc content increased; this was observed for the three starches and the flour.

The shapes of the curves shown in Figure 3 are rather different for SP and SW: the melting of the potato starch ordered zones was more abrupt than it was for the wheat starch and the wheat flour. These observations are in good agreement with the results of Waigh et al. on the two-stage destructuration of SW (crystallinity loss and helices destruction) [27]. Fannon, Hauber, and Bemiller showed that the SP granules appear smooth on the surface by scanning electron microscopy, whereas cereal grains such as SW have pores of about 100 nm at their surfaces, which are generated at the time of granule biosynthesis and cause a greater hydration sensitivity [28]. These observations may partly explain the differences in the plasticizer penetration into the granules for wheat starch and potato starch.

**Figure 3.** Curves of the partial enthalpy of gelatinization for all the plasticizer contents studied. (**a**) [SWCxCcy]Wz systems; (**b**) [SPxCcy]Wz systems; (**c**) [SWxCcy]Wz systems; (**d**) [FWxCcy]Wz systems.

For potato starch, the total gelatinization enthalpy (ΔH) in the presence of Cc remained nearly constant at low concentrations of Cc but decreased slightly for higher concentrations. For wheat starch, the ΔH increased significantly as the concentration of Cc increased, to a greater extent than that for the potato starch. This considerable increase in the total enthalpy of gelatinization of wheat starch at high Cc concentrations suggests a better organization of the ordered regions of starch in the presence of Cc; on the other hand, when the concentration reached 40% for the Cc, the total enthalpy of the starch due to the fusion decreased perceptibly. The effect of high Cc concentrations on the gelatinization process

was different for potato starch compared to waxy maize and wheat starch and wheat flour; it is assumed that this was due to the differences in the polymorphic structures.

The effect of Cc on the gelatinization of SP is more pronounced in the first step of the gelatinization process (retardation of the loss of crystalline order), whereas for SW, FW, and SWC, Cc also affects the second stage (delay in the loss of molecular order) and the overall enthalpy of the transition. These results are again opposite to those found in the literature for potato starch and wheat starch in the presence of high salt concentrations (NaCl) [20].

#### *3.4. Impact of the Formulation on Starch Gelatinization*

As previously described, choline chloride in the absence of water is not considered as an ionic liquid. Therefore, it seemed interesting to observe the action of Cc on starch when Cc is no longer acting as a plasticizer, but when it enters into the formulation as dry matter, Cc is placed directly in competition with water in front of starch. The relevant suspensions were formulated as follows: the starch (or flour) was mixed with the choline chloride (dry powder), and the water was added to the mixture. The obtained mixes were quoted using the following script: [FxCcy]Wz. These mixtures were studied by μDSC.

The temperatures and enthalpies of starch gelatinization did not vary with the formulation. However, there was a marked increase in the solubilization/rearrangement exotherm when the choline chloride was used as an ionic liquid (data not shown). This observation is explained by the fact that the solubilization of Cc before deposition on the starch allowed better penetration into the grain.

Another comparative study was carried out on the FW20Cc62W18 system, with three different formulations (Figure 4):


**Figure 4.** Comparative study of the thermograms of the systems [FW20Cc64]W16, FW20[Cc64W16], and [FW20W16]Cc64.

The results clearly showed that when the water (in small quantities) is premixed with the flour, it becomes inaccessible to the choline chloride, causing gelatinization at 90 ◦C corresponding to a hydration of approximately 50%, which agrees with the starting hydration (following formula).

There is also a broad endotherm of allotropic change in Cc, suggesting that choline chloride has access to only a very small amount of water.

For the other two formulations, very similar results were obtained: namely, no change in the gelatinization temperature, a slightly more prominent exotherm for formulation 2, and a small endotherm associated with the allotropic change in Cc.

In this part, we have shown that the order of incorporation of water into the system can influence the gelatinization of the studied starches. Indeed, if the water is added to the starch first, it binds enough to the starch to become unavailable for Cc. On the other hand, if the water is first added to the choline chloride, the latter is solubilized before entering as an ionic liquid in the starch.

### *3.5. Evolution of the Structure of Different Starches in the Presence of Choline Chloride during Heating Kinetics*

To understand the phenomena underlying the exothermic transitions observed with the microcalorimeter, measurements by an XRD heating cell were made.

Two types of systems were compared with and without Cc. The starch/flour and water concentrations were the same in both systems. The X-ray diffraction spectra of each of the suspensions of waxy corn starch and potato starch are collated in Figure 5.

SWC, SW, and FW possess A-type polymorphic structures, and SP possesses a B-type polymorphic structure. In the diffraction spectra of the SW20W24 (not shown), FW20W24 (not shown), and SWC20W24 systems, characteristic peaks of the A-type structure were identified. Spectra of the SP20W24 system showed characteristic peaks of the B-type structure.

A notable difference is seen in the evolution of the spectra during heating between the samples without Cc (Figure 5a,c) and those containing Cc (Figure 5b,d). Indeed, the peaks are more well defined in the systems without Cc, as with the addition of Cc at 20 ◦C. This confirms the previous conclusion: Cc causes a loss of crystalline order for all types of starch at room temperature. For the SP45W55 system, the type-B structure disappears at a temperature of 90 ◦C. However, for type-A systems, the structure collapses near 90 ◦C; this is consistent with Waigh's observations [29]. When Cc was added to the systems, a collapse of structure A or B was observed at much higher temperatures, which agrees with the results obtained by DSC.

For the SP20[Cc56W24] system, an increase in the peak intensity occurred at 5.6◦ in 2Θ at a temperature of 60 ◦C, followed by the onset of collapse at 100 ◦C and the total disappearance of the structure at 110 ◦C. This phenomenon corresponds to the appearance of the exotherm on the thermogram from the DSC study.

For systems SWC20[Cc56W24] and SW20[Cc56W24], the transition to an increase in crystallinity before melting is much less visible, and it is even absent for the FW20[Cc56W24] system. This explains the absence of the exotherm for the FW20[Cc56W24] system. In the literature, the exotherm is related to the solubilization of the starch in the plasticizer. However, during the X-ray diffraction study, we showed that the exotherm also corresponded to the increase in the intensity of the diffraction peaks and therefore to the increase in the crystallinity of the starch before its solubilization and then its melting.

**Figure 5.** Evolution of the X-ray diffraction spectra during heating associated with the thermogram: waxy maize starch without Cc: SWC45W55 (**a**) and with Cc: SWC20[Cc56W24] (**b**); potato starch without Cc: SP45W55 (**c**) and with Cc: SP20[Cc56W24] (**d**).

To complete this work, a μDSC study was conducted on the SP20[Cc56W24] system at low rates (0.1 ◦C/min). Two peaks were clearly observed (Figure 6), indicating that more than one phenomenon is responsible for the exothermic transition observed.

**Figure 6.** μDSC heating curves, 20–120 ◦C at 0.1 ◦C/min, for SP20[Cc56W24].

#### **4. Conclusions**

The starch/water system shows a gelatinization accompanied or unaccompanied by a fusion, according to the water content of the matrix. The choline chloride/water system possesses the characteristics of an ionic liquid and exhibits an allotropic change at low water concentrations and solubilization for water contents greater than 30%.

Choline chloride is an ionic compound, which, like NaCl, has a "structure making" effect: that is, it increases the viscosity of the aqueous solution while decreasing the water fraction with high mobility. As a result, the gelatinization temperatures are displaced to higher temperatures. In contrast to the addition of NaCl, the addition of choline chloride entails a significant increase in the gelatinization enthalpy, which suggests stabilization or better organization of the ordered regions of the starch in the presence of choline chloride.

An X-ray diffraction study in a heating cell demonstrated the crystalline rearrangement of the structure of the starch grain, which takes place simultaneously with the solubilization phenomenon of amylose.

**Author Contributions:** Conceptualization, A.L.-B. and P.L.-B.; methodology, D.C. and B.P.; formal analysis, D.C.; data curation, D.C., P.L.-B.; writing—original draft preparation, D.C., A.L.-B. et P.L.-B.; writing—review and editing, D.C., G.D. and P.L.-B.; supervision, P.L.-B.; project administration and funding acquisition, A.L.-B. All authors have read and agreed to the published version of the manuscript.

**Funding:** Co fundings by French Ministry of Agriculture (ONIRIS-ID for FOODS programme), by ONIRIS-GEPEA and by INRA-BIA.

**Institutional Review Board Statement:** Nothing to report.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Nothing to report.

**Acknowledgments:** This project was supported by the French Ministry of Agriculture (ONIRIS-ID for FOODS programme), by ONIRIS-GEPEA and by INRA-BIA co fundings.

**Conflicts of Interest:** I certify that there is no conflict of interest.

### **References**


## *Article* **Characterization of Novel Biopolymer Blend Mycocel from Plant Cellulose and Fungal Fibers**

**Ilze Irbe 1,\*, Inese Filipova 1, Marite Skute 1, Anna Zajakina 2,\*, Karina Spunde <sup>2</sup> and Talis Juhna <sup>3</sup>**


**Abstract:** In this study unique blended biopolymer mycocel from naturally derived biomass was developed. Softwood Kraft (KF) or hemp (HF) cellulose fibers were mixed with fungal fibers (FF) in different ratios and the obtained materials were characterized regarding microstructure, air permeability, mechanical properties, and virus filtration efficiency. The fibers from screened Basidiomycota fungi *Ganoderma applanatum* (Ga), *Fomes fomentarius* (Ff), *Agaricus bisporus* (Ab), and *Trametes versicolor* (Tv) were applicable for blending with cellulose fibers. Fungi with trimitic hyphal system (Ga, Ff) in combinations with KF formed a microporous membrane with increased air permeability (>8820 mL/min) and limited mechanical strength (tensile index 9–14 Nm/g). HF combination with trimitic fungal hyphae formed a dense fibrillary net with low air permeability (77–115 mL/min) and higher strength 31–36 Nm/g. The hyphal bundles of monomitic fibers of Tv mycelium and Ab stipes made a tight structure with KF with increased strength (26–43 Nm/g) and limited air permeability (14–1630 mL/min). The blends KF FF (Ga) and KF FF (Tv) revealed relatively high virus filtration capacity: the log10 virus titer reduction values (LRV) corresponded to 4.54 LRV and 2.12 LRV, respectively. Mycocel biopolymers are biodegradable and have potential to be used in water microfiltration, food packaging, and virus filtration membranes.

**Keywords:** air permeability; fungal fibers; hemp fibers; microstructure; mechanical properties; mycocel; softwood fibers; virus membrane filtration

### **1. Introduction**

The importance of biobased polymers is well known, and much research and development activities concern the use of biobased polymers in science, engineering, and industry. Biopolymers from renewable resources are used in multiple fields, namely health, food, energy, and the environment, due to their intrinsic features, versatility, biocompatibility, and degradability. Besides, the widespread use of biopolymers also addresses concerns about environmental sustainability [1]. As long as biopolymers stand as a sustainable, biodegradable compound, new biopolymer products and materials filled, blended, or reinforced with natural fibers, will predominate with the promise of sustainability benefits [2].

Generally, biobased polymers are classified into three classes: (1) naturally derived biomass polymers such as cellulose, cellulose acetate, starches, chitin, modified starch, etc.; (2) bio-engineered polymers bio-synthesized by using microorganisms and plants such as poly(hydroxy alkanoates (PHAs), poly(glutamic acid), etc.; (3) synthetic polymers produced from naturally derived molecules or by the breakdown of naturally derived macromolecules through the combination of chemical and biochemical processes such as polylactide (PLA), poly(butylene succinate) (PBS), bio-polyolefins, bio-poly(ethylene terephtalic acid) (bio-PET) [3]. The first and second class polymers are biodegradable, they

**Citation:** Irbe, I.; Filipova, I.; Skute, M.; Zajakina, A.; Spunde, K.; Juhna, T. Characterization of Novel Biopolymer Blend Mycocel from Plant Cellulose and Fungal Fibers. *Polymers* **2021**, *13*, 1086. https:// doi.org/10.3390/polym13071086

Academic Editor: Rafael Antonio Balart Gimeno

Received: 3 March 2021 Accepted: 25 March 2021 Published: 30 March 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

allow for more efficient production, which can produce desired functionalities and physical properties, but chemical structure designs have limited flexibility. The third class polymers such as bio-polyolefins and bio-PET are not biodegradable and the only contribution for reducing environmental impact comes from reducing the carbon footprint. The origin of a polymer does not determine its biodegradability; this condition depends on the chemical structure of the polymer [3].

The naturally derived biopolymers, among others, include polysaccharides of plant and fungal origin. Polysaccharides are nontoxic and biodegradable, which increases their potential application in biopolymers. The most used biopolysaccharides are obtained from plant origin (e.g., cellulose), microbial origin (e.g., bacterial cellulose), and animal origin (e.g., chitin/chitosan). Moreover, these natural derivatives present a considerable number of reactive functional groups (e.g., hydroxyl, carboxyl, and amino groups), which significantly increase their applicability through chemical modification or physical blend [1]. Cellulose is the most abundant polysaccharide of natural origin in the world, and is mostly produced by plants. The cellulose chains structure leads to areas of high crystallinity within the polymer and to high stability structures, which as a consequence promote considerable strength, remarkable inertness, and insolubility in water and common organic solvents [4]. In its turn, fungal cell walls share a common chemical structure composed of homo- and heteropolysaccharides, protein, protein–polysaccharide complexes, lipids, melanin, and polysaccharide chains of chitin. Chitin is a biopolymer of N-acetylglucosamine with some glucosamine, which is the main component of the cell walls of fungi and is considered the second most abundant natural polymer after cellulose [5].

Current research in membrane science is now focusing more on biopolymers from natural raw materials with a well-defined structure to develop new membrane materials. In fact, the combination of polysaccharides and proteins is a method frequently used to design blended materials with improved performance regarding swelling, mechanical resistance, and biocompatibility, among other features. The attention has been focused also on membranes based on chitosan blended with other biomacromolecules such as alginate, cellulose, collagen, gelatin, keratin, sericin, and soy protein [6]. Chitin-related materials from fungal sources with focus on nanocomposites and nanopapers have been suggested as greener alternative to synthetic polymers [7]. Fungal chitin–glucan nanopapers have manufactured from chitin nanofibrils, which is a native composite material (chitin–glucan) combining the strength of chitin and the toughness of glucan. These nanopapers showed distinct physico-chemical surface properties, being more hydrophobic than crustacean chitin [8]. Nanopapers from chitin nanofibrils have exhibited tunable mechanical and surface properties with potential use in coatings, membranes, packaging [9] and ultrafiltration of organic solvents and water [10]. Fungal mycelium–nanocellulose has been produced by the agitated liquid culture of a white-rot fungus with nanocellulose as part of the culture media. The obtained biomaterials are suggested for diverse applications, including packaging, filtration, and hygiene products [11].

The literature survey demonstrates that fungal chitin nanofibrils have been used to manufacture nanopapers. Additionally, fungal mycelium has been combined with nanocellulose to obtain a blended biomaterial. However, there are no studies on blended biopolymers from plant cellulose and fungal hyphae. The aim of this study was to develop a novel blended biopolymer from naturally derived biomass that is made of softwood cellulose fibers and fungal fibers (hyphae) and to test it for air permeability, mechanical properties, and virus filtration efficiency. More detailed investigation of specific properties of biomaterial blend containing fungal fibers from basidiomycete *Ganoderma applanatum* is published in our previous paper [12]. The cellulose fibers from softwood and hemp shives are easily available natural resource. The basidiomycetes selected for this study have shown a high importance in medical and nutritional applications. For example, *Ganoderma applanatum* has been reported for its phytochemical properties for potential application in nanotechological engineering for clinical use [13]. *Fomes fomentarius* has been used in a traditional medicine for centuries and is still economically important as

a source of medicinal and neutraceutical products [14]. *Trametes versicolor* is known for its general health-promoting effects and is widely employed in traditional medicine [15]. *Agaricus bisporus* represents the leading position among edible cultivated mushrooms [16].

### **2. Materials and Methods**

#### *2.1. Synthesis of Biopolymers*

The screening of several basidiomycetes was carried out to select the fungal candidates forming homogenous hyphal biomass convenient for biopolymer material development. The fungal biomass was obtained from (1) fruiting bodies of the forest growing species *Ganoderma applanatum* (Pers.) Pat., *Fomes fomentarius* (L.) Fr., *Lentinus lepideus* (Fr.) Fr., *Polyporus squamosus* (Huds.) Fr., *Fomitopsis betulina* (Bull.) B.K. Cui, (2) commercially cultivated mushroom stipes of *Agaricus bisporus* (J.E. Lange) Imbach and (3) mycelium of pure culture strain *Trametes versicolor* CTB 863 A.

Fungal biomass was kept in 4% NaOH solution for 24 h at a room temperature in order to extract proteins and alkali soluble polysaccharides. Then the samples were washed in tap water and mechanically disintegrated using Blendtec 725 (Orem, UT, USA) at 360 W for 30 s. Obtained fungal fibers (FF) were dried at room temperature and kept in a dry state until used.

Bleached softwood Kraft fibers (KF) were provided by Metsä Fibre (Äänekoski, Finland) as pressed sheets and used without specific pre-treatment.

Hemp fibers (HF) were obtained from industrial hemp *Cannabis sativa* (USO-31). After the decortication fibers were treated in 4% NaOH solution at 165 ◦C for 75 min, then washed with tap water to neutral and refined using Blendtec 725 (Orem, UT, USA) at 179 W for 7 min at 1.5% consistency, dried at room temperature and kept in dry state until used.

For biopolymer blend development, a certain amount (g) of FF in combinations with KF and HF fiber pulp in different mass ratios (50:50 and 33:33:33) was placed in a glass baker and soaked in 1–2 L of distilled water for 8 h, then disintegrated using 75,000 revolutions in the disintegrator (Frank PTI, Laakirchen, Austria). Material sheets were produced according to ISO 5269-2:2004 with a Rapid Köthen paper machine (Frank PTI, Laakirchen, Austria). Grammage (weight per unit area in g/m2) of samples was calculated by dividing the mass with area according to ISO 536:2019. At least five parallel samples at grammage <sup>50</sup> ± 15 g/m<sup>2</sup> of each composition were prepared.

#### *2.2. Micromorphology*

The micromorphology was examined by a light microscope (LM) Leica DMLB (Leica Microsystems GmbH, Wetzlar, Germany) at a magnification of 200×. Lactophenol blue solution (Fluka) was used to observe cyanophilic reaction of fungal hyphae. The images were captured by a video camera Leica DFC490 using calibrated image analysis software Image-Pro plus 6.3 (Media Cybernetics, Inc., Rockville, MD, USA).

For scanning electron microscopy (SEM), the surface of samples was coated with gold plasma using a K550X sputter coater (Emitech, Ashford, UK) and examined with Vega TC (Tescan, Brno-Kohoutovice, Czech Republic) with accelerating voltage of 15 kV, software 2.9.9.21.

#### *2.3. Air Permeability*

Air permeability was tested by Bendsen method according to ISO 5636-3:2013 using an air permeability tester 266 (Lorentzen and Wettre, Stockholm, Sweden). Samples were clamped between a metal ring and a rubber gasket and the air flow rate was measured through sample area of 10 cm<sup>2</sup> under 1.47 kPa air pressure for 5 s. Commercially available disposable face masks consisting of three layers made of polypropylene spunbondmeltblown-spunbond nonwoven fabrics were tested for air permeability as comparison for developed biopolymer samples.

### *2.4. Mechanical Properties*

For evaluation of tensile properties samples with a width of 1 cm were prepared with a strip cutter (Frank-PTI, Laakirchen, Austria) and tested according to ISO 1924-1:1992 using a tensile tester vertical F81838 (Frank-PTI, Laakirchen, Austria). Tensile index (Nm/g) was calculated by dividing measured maximum tensile strength (N/m) by grammage (g/m2) of test sample. Breaking length (km) and stretch (%) was used as calculated by software of equipment. Burst index (kPa m2/g) was calculated dividing the measured burst strength (kPa) by grammage (g/m2), where burst strength was measured as hydrostatic pressure necessary to cause rupture in a circular area of a 4 cm diameter of sample according to ISO 2758:2014 using burst tester (Frank-PTI, Laakirchen, Austria).

#### *2.5. Virus Preparation and Membrane Filtration Procedure*

The recombinant Semliki forest virus (SFV) pSFVenh/Luc, encoding firefly luciferase gene, was produced as previously described [17,18]. The virus-containing cell medium was harvested and concentrated by ultracentrifugation through two sucrose cushions, as previously described [19]. The virus was rapidly frozen and subsequently used as a virus stock. The virus titre expressed in infectious units per ml (i.u./mL) was quantified by immunostaining with rabbit polyclonal antibodies specific to the nsp1 subunit of SFV replicase, as previously described [20].

The ability of developed biopolymer materials to retain/remove virus particles was tested using centrifugal filtration test. EMA/CPMP/BWP/268 guidelines were considered for development of virus filtration procedures and calculation of virus reduction values. Two-layer cellulose hygienic paper (AB Grigeo, Vilnius, Lithuania), three-layer cotton mask with antibacterial *Silverplus* coating (SIA P.E.M.T., Riga, Latvia) and a standard surgical mask type II EN 14,683 (Matopat, Toru ´n, Poland) were used as reference materials.

A 1.5 × 1.5 cm sample was cut out from each material, folded in a form of a conical funnel, placed into truncated 1 mL plastic pipette tip (for better fixation), and subsequently placed into an Eppendorf 1.5 mL tube as presented in Figure 1.

**Figure 1.** Schematic representation of the virus filtration test (technical details are provided in text). (1) 1.5 cm × 1.5 cm filter sample is cut out from each material, folded into the form of a conical funnel, and placed into a 1.5 mL tube; (2) 50 μL of recombinant Semliki forest virus (SFV)-enh/Luc virus solution (107 i.u./mL) is added into the cone; (3) the tube is centrifuged to allow the virus to pass through the material; (4) the filtrate (indicated by arrow) is collected; (5) the filtrated is diluted and used for cell infection in a 24-well cell culture plate; (6) after overnight incubation of the plate the cell lysates are prepared and the virus infection is measured by detection of the luciferase activity in infected cells (luminometry). The cell infection with the standard dilutions of the virus is used to generate a standard curve and to calculate the amount of virus in the filtrate.

50 μL of recombinant SFVenh/Luc virus solution with a titer of 10<sup>7</sup> infectious units per milliliter (i.u./mL) were poured into each cone containing the test material. The tubes were then centrifuged for 20 s at 500× *g* (3500 rpm) by FVL-2400N Combi-Spin, Mini-Centrifuge/Vortex (Biosan, Riga, Latvia) to allow the liquid to pass (filtrate) through the material under low pressure conditions. The filtrated virus samples were collected (at least 20 μL each) and used for BHK-21 cell infection in 24-well plate as previously described [18]. Briefly, 10 μL of the virus sample in duplicate was mixed with 190 μL PBS (containing M2+ and Ca2+) and incubated with BHK-21 cells for 1 h at 37 ◦C, 5% CO2, then 800 μL of BHK medium (1% fetal bovine serum) was added. In parallel, standard dilutions of the SFVenh/Luc virus in a range 1 × 103–5 × 105 i.u./mL per well were generated and used for BHK-21 cell infection in duplicate in the 24-well plate. The cells were incubated overnight at 37 ◦C, 5% CO2 to allow complete cell infection to complete and expression of firefly luciferase gene.

To quantify the virus titer in filtered samples the relative luminescence units (RLU) were measured in cell lysates and the respective values were plotted on the standard curve with serial SFVenh/Luc virus dilutions. The RLUs were measured by the Luciferase assay (Promega, Madison, WI, USA), as recommended by manufacturer. Briefly, the cell medium was removed, and the cells (24-well) were lysed in 100 μL of the Cell Culture Lysis buffer (Promega, Madison, WI, USA), centrifuged at 600 cfr for 5 min, and 1 μL of the cell lysate was used immediately to measure the luciferase enzymatic activity by luminometer Luminoskan Ascent (Thermo Scientific, Loughborough, UK). The cell infection was done in duplicate in each independent experiment. Two independent experiments were performed with each sample (repeats). The virus titer standard curve was generated in each experiment and the negative control signal (RLU of uninfected cells) was subtracted from all values. The log10 reduction value (LRV) represents the difference between loaded and eluted virus infectious units per ml, and respectively was calculated according to the following equation:

LRV = [log10 (virus titer before filtration) − log10 (virus titer after filtration)]. (1)

Statistical analysis of obtained results was performed using an Excel 2016 MSO data statistical analysis tool.

### **3. Results and Discussion**

The fruiting bodies of *L. lepideus*, *P. squamosus*, and *F. betulina* formed gelatinous biomass in NaOH solution which was unusable for further experiments. FF were successfully separated from the biomass of fungi *G. applanatum*, *F. fomentarius*, *A. bisporus*, and *T. versicolor* and integrated in biopolymer compositions with softwood and hemp cellulose fibers.

#### *3.1. Morphological Characterization*

The microscopic structure of mycocel biopolymer blends and macrostructure of developed materials is shown in Figure 2.

Kraft fibers, hemp fibers, and fungal fibers were most likely physically bound together making a net of the biopolymer material. The fungal fibers separately or in bundles were randomly distributed within the net of cellulose fibers. The structure of materials was affected by individual properties of fibers. The size of Kraft fibers was average 2 mm in length and 30 μm in width. Hemp fibers were 1 mm long and 10–20 μm wide with a net of microfibers (2–5 μm). Fungal fibers from fruiting bodies of poroid species (*G. applanatum*, *F. fomentarius*) (Figure 2a,b) were 2–7 μm across. The fibers of *T. versicolor* mycelium were 2–3 μm wide (Figure 2c). The stipes of mushroom *A. bisporus* were composed of swollen hyphae, ascending 7–20 μm across (Figure 2d).

**Figure 2.** *Cont*.

(**e**)

**Figure 2.** Microstructure (left column) and macrostructure (right column) of mycocel biopolymer compositions made of Kraft fibers (KF), hemp fibers (HF), and fungal fibers (FF): (**a**) KF FF (Ff); (**b**) KF FF (Ga); (**c**) KF FF (Tv); (**d**) KF FF (Ab); (**e**) KF HF FF (Ga). Microimages show flat hemp and softwood fibers (10–30 μm) (**a**–**e**) and narrow fungal fibers (2–7 μm) of polypores (**a**–**c**,**e**) and swollen hyphae (7–20 μm) of agaric (**d**). LM, 200×. Bar = 100 μm.

The hyphae of basidiomycetes (primarily poroid species) are divided in three main types: generative, binding, and skeletal hyphae with key differences in cell wall thickness, internal structure, and branching characteristics. Monomitic species comprise only generative hyphae, dimitic species comprise two hyphal types (usually generative and skeletal) and trimitic species contain all three hyphal types [21].

The fungal fruiting bodies under this study, namely, *G. applanatum* and *F. fomentarius*, had trimitic hyphal system. Depending on the species, the generative hyphae were 2–5 μm, skeletal hyphae 3–7 μm and binding hyphae 2–4 μm across (Figure 2a,b). In the case of poroid species *T. versicolor*, the mycelium of pure culture strain consisted of thin-walled generative hyphae, while hyphae of agaric *A. bisporus* were flat and swollen. The fibers of two later species formed dense hyphal bundles among the cellulose fibers. These can be observed in Figure 2c,d where cyanophilic reaction turned hyaline hyphae of Tv and Ab blue after staining. The trimitic species Ga and Ff presented brown pigmented hyphal network (Figure 2a,b,e). The pigmented layer within the hyphal cell wall is likely crosslinked to polysaccharides and has a mesh-like structure with pores, through which small and large molecules can penetrate into cells. Pigment melanin is believed to enhance the strength of the cell wall and has antioxidant properties and propensity to bind to a variety of substances [22].

SEM micrographs (Figure 3) support the findings of LM and display a detailed ultrastructure of the raw materials and mycocel biopolymer blends. Fiber orientation appeared randomly distributed, entangled fibers were physically bound together with H bonds, forming non-oriented, multi-layered net. Kraft fibers (Figure 3a) were flattened and made a microporous network. Similarly, the fungal fibers Ga (Figure 3c) formed a network of microporous structures. Hemp material (Figure 3b) displayed a dense structure composed of flat fibers and thin microfibers.

The microstructure of mycocel biopolymers was affected by individual properties of raw materials. The structure of KF FF (Ga) (Figure 3d) was determined by individual properties of Kraft fibers and fungal fibers. Distribution FF among the flattened KF favoured formation a loose microporous structure of the material. The structure of biopolymer HF FF (Ga) was affected by a dense network of variable diameter hemp fibers (Figure 3e). The FF were incorporated in a tight HF network of microfibers forming a compact structure with decreased porosity. The surface of KF HF FF (Ga) material (Figure 3f) displayed dense areas of hemp fibers and Kraft fibers with few areas of loose fungal fibers. The porosity and the tortuosity fractal dimension are two critical parameters to determine the permeability. It is reported [23] that increase in the tortuosity fractal dimension leads to decrease in the dimensionless permeability and absolute permeability; an increase in the porosity increases

the dimensionless permeability; increase in the fiber diameter yields an increase in the absolute permeability.

**Figure 3.** Ultrastructure of raw materials (left column) from (**a**) Kraft fibers (KF), (**b**) hemp fibers (HF), and (**c**) fungal fibers (FF Ga), and mycocel biopolymer compositions (right column): (**d**) KF FF (Ga); (**e**) HF FF (Ga); (**f**) KF HF FF (Ga). SEM, 1000×. Bar = 100 μm.

### *3.2. Air Permeability*

The air permeability results were supported by microscopy findings. Mycocel biopolymer compositions had significant variations in air permeability properties starting from almost non-air permeable ones to materials with air permeability above measuring limit of device, which was 8820 mL/min (Table 1). The fungal type of hyphae had a significant effect on results. Biopolymer blends having *F. fomentarius* (Ff) or *G. applanatum* (Ga) trimitic hyphal system in combination with Kraft fibers (KF FF (Ff) and KF FF (Ga)) had the highest

air permeability >8820 mL/min. Biopolymers consisting of cellulose fibers (KF and HF) and *A. bisporus* (Ab) hyphae had the lowest air permeability 14.1 mL/min, <1 mL/min and 7.5 mL/min for KF FF (Ab), HF FF (Ab), and KF HF FF (Ab), respectively.

**Table 1.** Mechanical and air permeability properties of mycocel biopolymer materials (KF = Kraft fibers; HF = hemp fibers; FF = fungal fibers; Ga = *G. applanatum*; Ab = *A. bisporus*; Tv = *T. versicolor*; Ff = *F. fomentarius*). The marked (\*) samples have been described previously [12].


It should be noted that materials with HF in their composition had the lowest air permeability numbers because of high Shopper Riegler freeness of hemp cellulose fiber (91.5 ◦SR) [12], which led to high bonding and tight networks of fibers (Figure 3b). The bonding of cellulose fibers is based primarily on hydrogen bonding between hydroxyl functional groups during close contacting of fibers [24]. The same bonding theory can be attributed to investigated mycocel biopolymer materials, since cellulose and chitin the main constituent of hyphae—are biopolymers and have similar polysaccharide chain structure with the main difference being the replacement of one of the three hydroxyl groups with acetyl amine group in chitin monomeric unit [25]. The ability of functional groups of hyphal polysaccharides to make hydrogen bonding with cellulose fibers is one of the key elements of network formation in investigated mycocel materials. Furthermore, better bonding leads to a more compact packing of fibers and lower free volumes associated with lower air permeability [26], which can be seen in the cases of biopolymer blends with Tv mycelial and Ab stipe fibers. The higher numbers of air permeability showed lower ability of Ga and Ff hyphae to get involved in the network of cellulose fibers through hydrogen bonding. It can be explained by presence of non-polysaccharidic substances, such as pigments, which were indicated by the dark color of fruiting bodies and isolated hyphae from Ga and Ff.

Mycocel blends with higher air permeability have potential for using as gas permeable membranes, for example, as biobased filter layer in face masks. Disposal medical face masks, used as reference material, correspondingly showed high air permeability above 8820 mL/min, therefore only KF FF (Ga) and KF FF (Ff) are appropriate for this application. Blends with very low air permeability, such as those containing Ab hyphae can be used in applications, where high air or gas barrier properties are important, for example, food packaging materials [27].

### *3.3. Mechanical Properties*

Comparison of cellulose and fungal fiber biopolymers showed significant differences among different fungal species and hyphal types regarding their effects on mechanical properties of investigated materials (Table 1). It was possible to produce a pure fiber material from *G. applanatum* (Ga) and *A. bisporus* (Ab) biomass, however FF (Ga) material had a rather low tensile index 8.2 Nm/g, but FF (Ab) material was too fragile for handling and it was not possible to measure mechanical properties. When compare two-component blends of Kraft fibers and fungal fibers, the best results were obtained in the case of KF FF

(Ab) and KF FF (Tv), 32.5 and 26.5 Nm/g, respectively, which is by 92% and 57% more than tensile index measured for material containing only KF. However, addition of *G. applanatum* (Ga) or *F. fomentarius* (Ff) fibers to cellulose fibers decreased mechanical strength by 18% for KF FF (Ga) and by 48% for KF FF (Ff) if compare with pure KF. This can be explained by specific character of Ab and Tv hyphae which formed a dense microstructure with cellulose fibers increasing the mechanical strength of material. It is reported [28] that generative hyphae alone (monomitic hyphal system), which are hollow and contain cytoplasm, are suggested to provide limited mechanical performance, with binding hyphae (dimitic and trimitic hyphal systems) responsible for material strength. Contrary, our results showed that the trimitic hyphal system of fungi Ga and Ff did not favor an improvement of mechanical properties of mycocel compositions. Lower mechanical performance of Ga and Ff hyphae containing materials can be explained also by the aspects mentioned in description of air permeability properties and are related to lower ability for bonding with cellulose caused by presence of non-polysaccharidic substances and lower amount of hydrogen bonds formed. It is known that fiber to fiber bonding or bonding strength is directly related to mechanical strength of cellulose fiber-based materials and can be evaluated using results of mechanical tests [29].

Tensile index was almost threefold higher in wet stage when compare KF and KF FF (Ab) showing the ability of Ab hyphae to significantly improve the strength of fiber material in wet conditions. Two-component blends consisting of hemp fibers and fungal fibers HF FF were produced using hemp cellulose and Ga or Ab fibers. It must be noted that highly fibrillated hemp fibers with freeness 91.5 ◦SR used in the research, had the most significant effect on the mechanical strength of all investigated materials. Furthermore, pure HF material had the highest tensile index among all the tested materials by reaching 60.4 Nm/g. Addition of hyphae significantly decreased tensile index reaching 30.8 Nm/g in the case of HF FF (Ga) and 32.5 Nm/g in the case of HF FF (Ab). When compare threecomponent blends KF HF FF, here also Ab hyphae showed the highest effect on mechanical properties of composites by KF HF FF (Ab) reaching tensile index 46 Nm/g, while KF HF FF (Ga) reached 35.9 Nm/g. Overall, *A. bisporus* (Ab) and *T. versicolor* (Tv) fibers showed higher potential as component of biopolymer for improving the mechanical properties of cellulose fiber-based materials, whether they consist of wood or hemp cellulose fibers. Biopolymers containing Ga hyphae demonstrated the highest flexibility measured as percentage elongation or stretch. HF FF (Ga) and KF HF FF (Ga) samples reached 3.4% and 3.9% stretch, while other composites had numbers equal or below 2.0%.

Mechanical performance of materials was evaluated also by measuring burst index, which is the hydrostatic pressure necessary to cause rupture in a circular area of a given diameter and tells how much pressure material can tolerate before rupture. Three component blends showed the highest burst index 2.8 kPa m2/g and 2.4 kPa m2/g for KF FF FF (Ga) and KF HF FF (Ab) respectively, indicating the prevalence of fiber diversity presented in mechanical blend. If two-component blends are compared, Tv and Ab containing KF FF materials have higher results than Ga and Ff containing ones. Breaking length is used to characterize inherent strength of material and is defined as the length of imaginary material strip, if suspended vertically from one end, would break by its own weight. Numeric values of breaking length of biopolymer blends were in direct correlation with tensile strength values showing the predominance of Ab and Tv containing materials and lower results in Ga and Ff containing materials. Mechanical properties of disposal medical face masks layers were not possible to measure according to the standard and device used for biopolymer blends testing, however apparently they are less strong than materials of biopolymer blends.

### *3.4. Virus Filtration Properties*

In order to evaluate the filtration properties of the experimental biomaterials, the recombinant Semliki forest virus (SFV) was applied, which belongs to the Togaviridae family of enveloped RNA viruses (60–70 nm in diameter) structurally similar to human pathogenic viruses such as influenza and coronaviruses. The test system was based on a safe replication deficient SFV vector (pSFVenh/Luc), allowing to perform one round cell infection to be performed with precise quantification of the virus filtration rates using rapid measurement of luciferase activity in infected cells.

The efficiency of the SFV/enhLuc virus filtration through the tested materials are summarized in Table 2. The efficient virus retention properties were observed both for raw materials (Kraft fibers, hemp fibers, and fungal fibers Ga) and mycocel blends KF FF (Ga) and KF FF (Tv), which revealed very low virus permeability rates (<2%). Cellulose hygiene paper also showed the ability to retain the virus, albeit with lower efficiency (the permeability <30%) comparing to cellulose containing biopolymers. The properties of cellulose materials are determined by production technology and additives which can result in lower mechanical strength, density, and virus filtering efficiency as observed in the case of hygiene paper. Regarding the reference materials, i.e., face masks, the highest virus permeability was observed for the tested surgical mask (>92%) and hydrophobic outer layer of the cotton mask (>70%), whereas the *Silverplus* layer retained the virus significantly (the permeability <1%).

**Table 2.** SFV1enh/Luc recombinant virus titer change after pressure filtration through experimental materials (KF = Kraft fibers; HF = hemp fibers; FF = fungal fibers; Ga = *G. applanatum*; Tv = *T. versicolor*).


According to EMA guidelines (CPMP/BWP/268/95, European Medicines Agency) the log10 reduction value (LRV) is an important parameter to quantify the virus reduction capacity. Only the LRV > 4 is considered as a "very high" virus reduction potential. In this study, the biopolymer KF FF (Ga) demonstrated sufficiently high virus reduction capacity (4.54 LRV). Remarkably, the silver containing cotton layer of the commercial mask also showed efficient virus removal properties (3.6 LRV), which can be related to the viricidal capacity of the immobilized silver nanoparticles [30].

Due to safety aspects, handling of human viruses is a labor-intensive and timeconsuming process. In this study, we have established a protocol for a relatively simple virus filtration method, which is based on replication deficient SFV vector encoding firefly luciferase used for rapid virus quantification. The proposed method can be efficiently applied for primary screening of virus filtration properties of the membranes.

The mechanisms underlying the membrane filtration of viruses include size exclusion and/or adsorptive interactions (e.g., hydrophobic/hydrophilic and electrostatic interactions) between virus envelope and membrane compounds [31]. The diameter of the virus particles is much smaller than the pore size of the tested membranes. Therefore, the adsorptive interactions can be considered as the main mechanism of filtration in the tested system. A surgical mask, which is made of hydrophobic polypropylene layers, did not adsorb aqueous virus containing solution, resulting in low virus retention values (Table 2). Practically, the surgical masks serve as a barrier to aqueous aerosol and are designed to block direct fluid entry into the wearer's respiratory tract and mostly act as a repellent of the water-based liquids [32]. In contrast to surgical mask and tested cotton mask, the row fiber materials and mycocel blends exhibited hygroscopic properties and showed the ability to binding, or adsorption of SFV particles. A high virus filtration capacity of mycocel biopolymer blends might be attributed to the structural properties of the fungal cell wall as an insoluble polysaccharide-based sorbent. The fungal polysaccharides possess direct virus inactivation properties as shown for several highly pathogenic human viruses, including human immunodeficiency virus, herpes simplex virus, etc. [33]. The application of mycocel-based filters alone or in combination with polypropylene-based layers can represent an advanced individual respiratory protective device against airborne pathogens. Furthermore, chitin, one of the main polymers of fungal cell walls, is widely used for controlled drug delivery systems, protein and enzyme carriers, and packaging materials, based on its natural antimicrobial activity. Chitin and its derivatives (e.g., chitosan) have many useful properties that make them suitable for a wide variety of biomedical applications. Their products are known to be antibacterial, antifungal, antiviral, nontoxic, and nonallergic [34]. Therefore, the proposed mycocel-based biopolymers are promising multifunctional materials for biomedical and bioengineering applications.

### **4. Conclusions**

A novel mycocel biopolymer from naturally derived biomass of plant cellulose fibers and fungal fibers (hyphae) was developed and characterized regarding its air permeability, mechanical properties, and virus filtration efficiency.

Air permeability and mechanical properties of mycocel biopolymer blends were affected by microstructural features of raw materials. Highly fibrillated hemp fibers had the most significant effect on the mechanical strength while Kraft fibers revealed increased air permeability. The ability of functional groups of hyphal polysaccharides to make hydrogen bonding with cellulose fibers was one of the key elements of network formation which determined biopolymer properties. The loose fiber net of trimitic fungal species *G. applanatum* and *F. fomentarius* with incorporated Kraft fibers formed a microporous structure of mycocel blends with improved air permeability (>8820 m/min) and limited mechanical properties in comparison with individual raw fibers. *A. bisporus* and *T. versicolor* fibers showed higher potential as components of biopolymer for improving the mechanical properties of cellulose fiber-based materials, whether they consist of wood or hemp cellulose fibers.

Virus testing provided promising results regarding virus filtration efficiency of biopolymer blends. Mycocel biopolymer KF FF (Ga) demonstrated sufficiently high virus reduction capacity (4.54 LRV) than surgical mask and outer and inner layers of commercial face mask. Feasibly, the adsorptive interactions can be considered as the main mechanism of filtration in tested system.

The natural, biodegradable mycocel blends have a potential for use in biomaterial membranes depending on the target application. Blends with higher air permeability and virus filtration efficiency have potential for being used as gas permeable membranes, for example, as biobased filter layer in face masks. Blends with low air permeability can be used in areas where high air or gas barrier properties are important, for example, food packaging materials. The water microfiltration and ultrafiltration also are considered for future application.

**Author Contributions:** Conceptualization, I.I. and I.F.; methodology, M.S., K.S., A.Z.; validation, I.I., I.F. and A.Z.; formal analysis, K.S., A.Z.; writing—original draft preparation, I.I.; writing—review and editing, I.F., A.Z. and T.J.; project administration, I.F. and T.J.; funding acquisition, T.J. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research is funded by the Ministry of Education and Science, Republic of Latvia, Project "Integration of reliable technologies for protection against Covid-19 in healthcare and high-risk areas", Project No. VPP-COVID-2020/1-0004. The Article Processing Charge was kindly provided by the Latvian State Institute of Wood Chemistry Base financing source.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


## *Article* **Upgrading Argan Shell Wastes in Wood Plastic Composites with Biobased Polyethylene Matrix and Different Compatibilizers**

**Maria Jorda-Reolid 1, Jaume Gomez-Caturla 2, Juan Ivorra-Martinez 2,\*, Pablo Marcelo Stefani 3, Sandra Rojas-Lema <sup>1</sup> and Luis Quiles-Carrillo 1,\***


**Abstract:** The present study reports on the development of wood plastic composites (WPC) based on micronized argan shell (MAS) as a filler and high-density polyethylene obtained from sugarcane (Bio-HDPE), following the principles proposed by the circular economy in which the aim is to achieve zero waste by the introduction of residues of argan as a filler. The blends were prepared by extrusion and injection molding processes. In order to improve compatibility between the argan particles and the green polyolefin, different compatibilizers and additional filler were used, namely polyethylene grafted maleic anhydride (PE-g-MA 3 wt.-%), maleinized linseed oil (MLO 7.5 phr), halloysite nanotubes (HNTs 7.5 phr), and a combination of MLO and HNTs (3.75 phr each). The mechanical, morphological, thermal, thermomechanical, colorimetric, and wettability properties of each blend were analyzed. The results show that MAS acts as a reinforcing filler, increasing the stiffness of the Bio-HDPE, and that HNTs further increases this reinforcing effect. MLO and PE-g-MA, altogether with HNTs, improve the compatibility between MAS and Bio-HDPE, particularly due to bonds formed between oxygen-based groups present in each compound. Thermal stability was also improved provided by the addition of MAS and HNTs. All in all, reddish-like brown wood plastic composites with improved stiffness, good thermal stability, enhanced compatibility, and good wettability properties were obtained.

**Keywords:** argan shell particles; wood plastic composite; polyethylene; mechanical properties; compatibilization

### **1. Introduction**

In the last decades, an increasing concern about environmental issues related to the great use of petrochemical-derived plastics has been rising, as well as the necessity to reduce the carbon footprint associated with the production processes of those polymers. This trend is driving the industry toward the use of polymeric materials derived from natural resources, with the objective of moving away from petroleum. Moreover, this fact has propitiated the use of biodegradable polymers, some of which are capable of decomposing in composting conditions. At the same time, they exhibit very similar properties to those of their petrochemical counterparts [1]. One of the main drawbacks of plastics derived from natural sources is that they are more expensive than those from petrochemical origin. This leads to a direct rejection by industry. For this reason, incorporation of fillers and creation of new more economic materials is gaining increasing relevance.

**Citation:** Jorda-Reolid, M.; Gomez-Caturla, J.; Ivorra-Martinez, J.; Stefani, P.M.; Rojas-Lema, S.; Quiles-Carrillo, L. Upgrading Argan Shell Wastes in Wood Plastic Composites with Biobased Polyethylene Matrix and Different Compatibilizers. *Polymers* **2021**, *13*, 922. https://doi.org/10.3390/ polym13060922

Academic Editor: Gianluca Tondi

Received: 27 February 2021 Accepted: 15 March 2021 Published: 17 March 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Among those materials, "Wood plastic composites" (WPC) are becoming increasingly popular. This technology implies the addition of natural organic fillers obtained from wastes into polymer matrices, with the objective of decreasing their processing cost. At the same time, those fillers enhance their environmental value and vary their mechanical, physical and chemical properties [2]. Initially, only wood flour or sawdust were used as bio-fillers [3–6], but over time fillers from agroforestry waste or food industry have started to be used, such as orange peel [7,8], almond shell [9], pomegranate peel [10], or argan seed shell [11]. Those fillers are introduced in the structure of polymers like poly (lactic acid) (PLA), high density polyethylene (HDPE) or polyester [12–14]. Nowadays, WPC can be employed as a substitute of wood due to its appearance for the fabrication of indoor and outdoor furniture, decks pergolas, and so on [15,16].

In this sense, high density bio-polyethylene (Bio-HDPE) is one of the most interesting polymer matrices, as it combines its natural origin with the ease of processability of polyolefins. This polymer is produced by conventional polymerization of ethylene obtained from catalytic dehydration of bio-ethanol [17], which is extracted from natural sources, such as sugarcane [18]. Bio-HDPE possesses the same physical properties than its fossilderived counterpart; particularly, good mechanical resistance and high ductility [19]. Injection-molded Bio-HDPE pieces have great application in packaging [20]. However, one of the main problems that limits the use of bio-fillers as reinforcing agents is their hydrophilic nature, which causes a poor interfacial adhesion with a hydrophobic (non-polar) matrix, such as Bio-HDPE. This provokes a deterioration in the mechanical properties of the material and a bad compatibility between filler and matrix [21,22].

Several studies have been carried out to improve the compatibilization between those fillers and polymeric matrices. On the one hand, some of the methods that have been applied to improve this compatibility include the modification of the surface of the filler particles from the waste, such as mercerization, benzoylation, acetylation, silanation, or graft copolymerization [23,24]. On the other hand, compatibilizing agents have also been used in order to improve the adhesion and affinity between filler and matrix. Additionally, they have also been utilized to increase the dispersion of the reinforcing particles within the matrix and give additional properties to the blend [25–27]. In this context, polyethylene-grafted maleic anhydride (PE-g-MA) is one of the most efficient compatibilizers. It acts as a chemical connection between ligno-cellulosic particles and polymeric chains, owing to its double functionality. First, the polyethylene fraction of PE-g-MA interacts with polymeric chains due to its chemical affinity, while the anhydride groups can react with the hydroxyl groups from the ligno-cellulosic structure in the organic particles by esterification. Thus, leading to an increase in the matrix–filler interaction and a positive effect in the dispersion of the particles and the stiffness of the blend [28–30]. However, ductile properties tend not to get better, because of the fragility given by those fillers when they form stress-concentrator aggregates in the matrix.

In recent years, vegetable oils have generated a great deal of interest as compatibilizing additives. Modified vegetable oils have been used as plasticizers, stabilizers, crosslinkers, compatibilizers, and so on. The use of epoxidized vegetable oils (EVOS) should be remarked, such as epoxidized soybean oil (ESBO) [31,32], epoxidized linseed oil (ELO) [33,34], or epoxidized palm oil (EPO) [35]. All of them have been tested in polymeric blends and composites, achieving positive compatibilization/plasticization effects, owing to the reactivity of the oxirane group. Maleinization is an interesting alternative chemical modification for vegetable oils. Maleinized linseed oil (MLO) has been widely used in WPC, for example, in a PBS matrix with almond shell flour [36]. Furthermore, the excellent ductile properties that this oil is capable of giving to composites have also been reported [13]. In addition, inorganic particles can also be introduced as reinforcing agents in polymeric matrices, such as carbon nanotubes (CNTs), nanoclays, or metallic oxide nanoparticles [37–39]. In recent years, halloysite nanotubes (HNTs) have gained great importance in this sense. They are cheaper and more abundant than, for example, carbon nanotubes (CNTs). In this sense, halloysite is a natural mineral clay with nanotube

morphology, similar to kaolinite (Al2[Si2O5(OH)4].nH2O), where n is 2 for hydrated HNTs, and 0 for dehydrated HNTs [40]. Frost et al. [41] reported two types of hydroxyl groups, inner and outer hydroxyl groups, placed between the nanotubes layers and in the surface of HNTs [41]. Because of their multilayer structure, most of the hydroxyl groups are in the inner side, while a few of them are in the surface of HNTs. As a result of this, HNTs dispersion in polymer matrices is simpler than other inorganic nanoparticles [40]. Numerous studies have reported the ability of HNTs to improve thermal stability, fire resistance, mechanical properties, and crystalline behavior of polymers such as polypropylene [42– 46], low density polyethylene (LLDPE) [47], polyamide 6 (PA6) [48], epoxy resin [49], polylactide (PLA) [50,51], and so on. These are the reasons for utilizing HNTs in this work. As it is the case with organic particles, the challenge here is to achieve a good compatibility and adhesion between HNTs and the polymer matrix, especially in non-polar matrices, such as polyolefins.

The objective of this work is the reuse of micronized argan shell (MAS) as a reinforcing organic filler in a Bio-HDPE matrix derived from sugarcane, with the objective of developing a wood plastic composite with enhanced properties compared to the original matrix following the trends proposed by the circular economy, which seeks to reuse or recycling the waste and to reintroduce them into the industry [52]. Argan (*Argania spinosa*) is a tropical plant that belongs to the *Sapotaceae* family, whose fruit is used in morocco to prepare oil [53]. This oil has great application in cosmetics owing to its high E vitamin content [53], while fruit residues and seeds are generally used to feed cattle [54]. The main drawback of micronized argan shell (MAS) is its low compatibility with Bio-HDPE. In particular, the challenge of this investigation is to improve the affinity between both elements through the use of compatibilizing agents. Several formulations have been developed to meet this end, using the compatibilizers PE-g-MA and MLO; and halloysite nanotubes (HNTs) as additional reinforcing fillers; altogether with Bio-HDPE as the polymer matrix and argan shell particles as the main reinforcing filler; these elements have been used both individually and in combination. Additionally, MLO and HNTs concentrations have been varied (7.5 parts per hundred resin (phr) when used individually and 3.75 phr when used in combination). In order to evaluate the effects of these compatibilizers and fillers over each blend, mechanical, morphological, thermal, thermomechanical, colorimetric, and wettability properties are presented.

### **2. Materials and Methods**

### *2.1. Materials*

Bio-HDPE, SHA7260 grade, was supplied in pellets form by FKuR Kunststoff GmbH (Willich, Germany) and manufactured by Braskem (São Paulo, Brazil). This green polyethylene has a density of 0.955 g cm−<sup>3</sup> and a melt flow index (MFI) of 20 g/10 min, measured with a load of 2.16 kg and a temperature of 190 ◦C. Micronized argan shell (MAS) was supplied by MICRONIZADOS VEGETALES S.L. (Benamejí (Córdoba), Spain) company. Halloysite nanotubes were supplied by Sigma Aldrich (Madrid, Spain) with CAS number 1332-58-7.

Interlocking is one of the most important mechanisms in polymer composites. This mechanism depends on the filler shape. Figure 1a shows the morphology of the argan particles observed by SEM. Most of the particles exhibit a rough surface, which can be ascribed to the milling process due to the high hardness of this filler. A closer observation of the particles shows some porosity on their surface, which can produce a good interaction with the polymer matrix acting as mechanical anchoring points. A similar structure was observed by Laaziz et al. [55] when studying argan nut shell particles with different treatments. Additionally, Figure 1b and c show histograms for the length and diameter of the particles, respectively, determined from the SEM image. Particles were average 70 μm long and 45 μm wide. These parameters are also important to consider when talking about mechanical properties. Too large particles could lead to a great mechanical impairment, increasing the heterogeneity of the blends. Crespo et al. [56] observed this effect for almond shell particles superior to 150 μm diameter.

**Figure 1.** (**a**) Scanning electron microscope image (SEM) of micronized argan shell (MAS). Image was taken with a magnification of 100× and a scale marker of 100 μm; (**b**) histogram of the argan particles length; (**c**) histogram of the argan particles diameter.

In relation to the additives, polyethylene-grafted maleic anhydride (PE-g-MA) with CAS Number 9006-26-2 and MFI values of 5 g/10 min (190 ◦C/2.16 kg), was obtained from Sigma-Aldrich S.A. (Madrid, Spain). This PE-based copolymer was selected due to its functionality. The maleinized linseed oil—MLO, VEOMER LIN was supplied by Vandeputte (Mouscron, Belgium). This modified vegetable oil is characterized by a viscosity of 10 dPa s at 20 ◦C and an acid value comprised in the 105–130 mg KOH g−<sup>1</sup> range.

### *2.2. Preparation of Bio-HDPE Blends*

Bio-HDPE, PE-g-MA, MAS y HNTs were initially dried at 40 ◦C for 48 h in a dehumidifying dryer MDEO (Barcelona, Spain) to remove any residual moisture prior to processing to avoid the possibility of hydrolysis due to the moisture. Then, the corresponding wt.% of each component (see Table 1) were mixed and pre-homogenized in a zipper bag. The corresponding formulations were compounded in a twin-screw co-rotating extruder from Construcciones Mecánicas Dupra, S.L. (Alicante, Spain). This extruder has a 25 mm diameter with a length-to-diameter ratio (L/D) of 24 to allow a correct blending process. The extrusion process was carried out at a rate of 22 rpm, using the following temperature profile (from the hopper to the die): 140–145–150–155 ◦C. The compounded materials were pelletized using an air-knife unit. In all cases, residence time was approximately 1 min. Table 1 shows the compositions of the formulations developed in this work.

To transform the pellets into standard samples, a Meteor 270/75 injector from Mateu & Solé (Barcelona, Spain) was used. The temperature profile in the injection molding unit was 135 ◦C (hopper), 140 ◦C, 150 ◦C, and 160 ◦C (injection nozzle). A clamping force of 75 tons was applied while the cavity filling and cooling times were set to 1 and 10 s, respectively. Standard samples for mechanical and thermal characterization with an average thickness of 4 mm were obtained.


**Table 1.** Summary of compositions according to the weight content (wt.%) of Bio-HDPE/MAS and different compatibilizers and additives.

### *2.3. Characterization of Bio-HDPE Blends*

2.3.1. Mechanical Characterization

Mechanical properties were obtained with different tests like tensile test, shore hardness, and Charpy impact test. Tensile properties of Bio-HDPE/PE-g-MA/MAS blends were obtained in a universal testing machine ELIB 50 from S.A.E. Ibertest (Madrid, Spain) as recommended by ISO 527-1:2012 with dog bone samples 1B specification. A 5-kN load cell was used and the cross-head speed was set to 5 mm/min. Shore hardness was measured in a 676-D durometer from J. Bot Instruments (Barcelona, Spain), using the D-scale, on rectangular samples with dimensions 80 × <sup>10</sup> × 4 mm3, according to ISO 868:2003. The impact strength was also studied on injection-molded rectangular samples with dimensions of <sup>80</sup> × <sup>10</sup> × 4 mm3 in a Charpy pendulum (1-J) from Metrotec S.A. (San Sebastián, Spain) on notched samples (0.25 mm radius V-notch), following the specifications of ISO 179-1:2010. All mechanical tests were performed at room temperature, and at least six samples of each material were tested and the corresponding values were averaged.

### 2.3.2. Morphology Characterization

The morphology of fractured samples from Charpy tests, obtained from the impact tests, were studied by field emission scanning electron microscopy (FESEM) in a ZEISS ULTRA 55 microscope from Oxford Instruments (Abingdon, UK). Before placing the samples in the vacuum chamber, they were sputtered with a gold-palladium alloy in an EMITECH sputter coating SC7620 model from Quorum Technologies, Ltd. (East Sussex, UK). The FESEM was operated at an acceleration voltage of 2 kV.

#### 2.3.3. Thermal Analysis

The most relevant thermal transitions of Bio-HDPE/PE-g-MA/MAS blends were obtained by differential scanning calorimetry (DSC) in a Mettler-Toledo 821 calorimeter (Schwerzenbach, Switzerland). Samples with an average weight of 6–7 mg, were subjected to a thermal program divided into three stages: a first heating from 25 ◦C to 160 ◦C followed by a cooling to 0 ◦C, and a second heating to 250 ◦C. Both heating and cooling rates were set to 10 ◦C/min. All tests were run in nitrogen atmosphere with a flow rate of 66 mL/min using standard sealed aluminum crucibles with a capacity of 40 μL.

The thermal degradation of the Bio-HDPE/PE-g-MA/MAS blends was assessed by thermogravimetric analysis (TGA). TGA tests were performed in a LINSEIS TGA 1000 (Selb, Germany). Samples with a weight of 15–17 mg were placed in 70 μL alumina crucibles and subjected to a dynamic heating program from 40 ◦C to 700 ◦C at a heating rate of 10 ◦C/min in air atmosphere. The first derivative thermogravimetric (DTG) curves were also determined. All tests were carried out at least three times in order to obtain reliable results.

### 2.3.4. Thermomechanical Properties

In order to obtain the thermomechanical properties of the Bio-HDPE composites a dynamical mechanical thermal analysis (DMTA) was carried out in a DMA1 dynamic analyzer from Mettler-Toledo (Schwerzenbach, Switzerland), working in single cantilever flexural conditions. Rectangular samples with dimensions 20 × <sup>6</sup> × 2.7 mm3 were subjected to a dynamic temperature sweep from −150 ◦C to 120 ◦C at a constant heating rate of 2 ◦C/min. The selected frequency was 1 Hz and the maximum flexural deformation or cantilever deflection was set to 10 μm.

#### 2.3.5. Color and Wetting Characterization

The colorimetric modifications of the samples were measured with a Konica CM-3600d Colorflex-DIFF2, from Hunter Associates Laboratory, Inc. (Reston, VA, USA) was used. Color coordinates (L\*a\*b\*) were measured according to the following criteria: L\* = 0, darkness; L\* = 100, lightness; a\* represents the green (a\* < 0) to red (a\* > 0); b\* stands for the blue (b\* < 0) to yellow (b\* > 0) coordinate.

Hydrophilicity/ hydrophobicity of each blend was assessed by contact angle measurements through time were carried out with an EasyDrop Standard goniometer model FM140 (KRÜSS GmbH, Hamburg, Deutschland) which is equipped with a video capture kit and analysis software (Drop Shape Analysis SW21; DSA1). Double distilled water was used as test liquid, a drop of it was put in each sample and contact angle measurements were taken at 0, 5, 10, 15, 20, and 30 min after the administration of the water. For color and wetting measurement, traction samples were used to facilitate the measures.

### 2.3.6. Water Absorption Test

The water absorption capacity of the Bio-HDPE/PE-g-MA/MAS blends was evaluated by the water uptake method. Impact specimens (80 × <sup>10</sup> × 4 mm3) from each blend were first weighted in a balance and then put inside a beaker filled with distilled water, all of them wrapped with tiny pieces of a metal grid so they could sink. After that, the weight of all samples was measured in intervals of several hours the first day, and then measured each week for 14 weeks in order to evaluate the amount of taken up water. In every measurement, the moisture in the surface of the samples was removed with tissue paper. During the immersion time, the temperature was controlled to 23 ◦C.

### 2.3.7. Infrared Spectroscopy

In order to obtain the interactions between the different elements, a chemical analysis of the Bio-HDPE/PE-g-MA/MAS blends was carried out by attenuated total reflection-Fourier transform infrared (ATR-FTIR) spectroscopy. Spectra were recorded using a Bruker S.A Vector 22 (Madrid, Spain) coupled to a PIKE MIRacleTM single reflection diamond ATR accessory (Madison, WI, USA). Data were collected as the average of 10 scans between 4000 and 500 cm−<sup>1</sup> with a spectral resolution of 2 cm−1. Samples from impact specimens (80 × <sup>10</sup> × 4 mm3) were employed in this test and the measurement was performed at room temperature.

### 2.3.8. Statistical Analysis

The significant differences among the samples were evaluated at 95% confidence level (*p* ≤ 0.05) by one-way analysis of variance (ANOVA) following Tukey's test. Software employed for this propose was the open source R software V4.0.3 (http://www.r-project. org (accessed on 12 March 2021))

#### **3. Results**

### *3.1. Mechanical Properties*

The results concerning the mechanical properties of Bio-HDPE/MAS blends with different compatibilizers are shown in Table 2. Those results are of great interest to evaluate

**Code E (MPa)** σ**max (MPa)** ε**<sup>b</sup> (%) Shore D Hardness Impact Strength (kJ/m2)** Bio-HDPE 750 <sup>±</sup> <sup>47</sup> <sup>a</sup> 14.48 <sup>±</sup> 0.78 <sup>a</sup> Nb <sup>a</sup> 56.2 <sup>±</sup> 1.3 <sup>a</sup> 2.7 <sup>±</sup> 0.2 <sup>a</sup> Bio-HDPE/PE-g-MA/MAS <sup>846</sup> <sup>±</sup> <sup>36</sup> <sup>b</sup> 7.57 <sup>±</sup> 0.81 <sup>b</sup> 20.7 <sup>±</sup> 2.0 <sup>b</sup> 59.2 <sup>±</sup> 0.8 <sup>a</sup> 1.4 <sup>±</sup> 0.1 <sup>b</sup> Bio-HDPE/PE-g-MA/MAS/HNTs <sup>1126</sup> <sup>±</sup> <sup>65</sup> <sup>c</sup> 7.57 <sup>±</sup> 0.90 <sup>b</sup> 6.0 <sup>±</sup> 0.9 <sup>c</sup> 60.6 <sup>±</sup> 1.3 <sup>a</sup> 1.7 <sup>±</sup> 0.1 <sup>c</sup> Bio-HDPE/PE-g-MA/MAS/MLO <sup>442</sup> <sup>±</sup> <sup>33</sup> <sup>d</sup> 6.66 <sup>±</sup> 0.39 <sup>c</sup> 41.5 <sup>±</sup> 1.7 <sup>d</sup> 53.2 <sup>±</sup> 0.8 <sup>b</sup> 2.2 <sup>±</sup> 0.3 <sup>d</sup> Bio-HDPE/PE-g-MA/MAS/HNTs/MLO <sup>523</sup> <sup>±</sup> <sup>26</sup> <sup>e</sup> 6.98 <sup>±</sup> 0.59 <sup>c</sup> 33.9 <sup>±</sup> 3.5 <sup>e</sup> 54.6 <sup>±</sup> 0.5 <sup>b</sup> 2.1 <sup>±</sup> 0.3 <sup>d</sup>

and compatibility between Bio-HDPE and MAS.

**Table 2.** Summary of mechanical properties of the injection-molded samples of Bio-HDPE blends. Tensile modulus (E), maximum tensile strength (σmax) elongation at break (εb), Shore D hardness, and impact (Charpy) strength.

the effectivity of the different agents used in terms of resistance improvement of Bio-HDPE

a–e Different letters in the same column indicate a significant difference among the samples (*p* < 0.05).

As it can be observed in Figure 2, neat Bio-HDPE presents a Young modulus (E) and tensile strength of 750 and 14.48 MPa, respectively. Elongation at break (εb) could not be determined because the tensile test machine reached its maximum elongation without breaking the sample. Those values are indicative of certain stiffness and high ductility for Bio-HDPE, as it has been also reported by other authors [57]. Incorporation of MAS (30 wt.%) with PE-g-MA (3 wt.%) into the Bio-HDPE matrix increases the Young modulus up to 846 MPa, in this case the statical analysis showed a significant difference by the incorporation of the filler. This can be directly related to a proper distribution of MAS particles in the polymer matrix, leading to a good adhesion of the filler with the polymer and obtaining a more rigid material [26,27]. However, there is a clear reduction in tensile strength and elongation at break (Figure 2), this can be related to an excess of MAS utilized, increasing the stiffness of the material. In this context Essabir et al. [11] observed a similar behavior when treating HDPE with argan particles, in a way that Young modulus increased with MAS concentration, but tensile strength started to decrease with a MAS particle content superior to 5%. Addition of HNTs in the blend showed an additional enhancement of stiffness; the incorporation of only 7.5 phr HNTs produced a significant difference against the previous sample and a 33% of improvement. In this context, Bio-HDPE/PE-g-MA/MAS/HNTs sample presented the highest Young modulus of all blends (1126 MPa). This can be attributed to the reinforcing effect of HNTs, which seem to be well-dispersed within the matrix owing to the presence of the compatibilizer PE-g-MA. Similar behavior was reported by Pratap et. al. [58]. On the other hand, HNTs do not produce changes in tensile strength (no significant differences could be appreciated between them) and reduce elongation at break. Therefore, the material becomes less ductile, as a result of the combination of MAS and HNTs reinforcing effects. MLO-compatibilized blends saw their Young modulus and tensile strength reduced in comparison with the rest of the blends. Nevertheless, elongation at break increased 100% referring to the Bio-HDPE/PE-g-MA/MAS blend and the Tukey test showed a significant difference between the samples. In this sense, the plasticizing effect of MLO can be verified, highly increasing the ductile properties and compatibility between components in the blend. Similar results were observed by Quiles-Carrillo et. al. [59] when treating a PA1010/Bio-HDPE blend with MLO. As to Bio-HDPE/PE-g-MA/MAS/HNTs/MLO blend, it shows intermediate values between the sample with HNTs and the sample with MLO individually. Although the plasticizing effect of MLO seems to prevail over the reinforcing effect of HNTs, due to the very high elongation at break and the relatively low Young modulus and tensile strength.

**Figure 2.** Mechanical properties of the injection-molded samples of Bio-HDPE blends. Tensile modulus (E), maximum tensile strength (σmax) elongation at break (εb). \* Elongation at break for the Bio-HDPE could not be assessed because breakage is not achieved.

Regarding the impact strength results, neat Bio-HDPE presents a value of 2.7 kJ/m2, which is a ductile behavior indicator. Bio-HDPE/PE-g-MA/MAS blend shows a reduction of impact strength down to 1.4 kJ/m2, confirming the great stiffness of the blend, probably due to the quantity of MAS utilized [11]. The presence of HNTs in the polymer matrix increases the impact strength to 0.3 kJ/m2, which means a significant difference between the mentioned samples. This effect is probably associated with an interaction between argan particles and HNTs. Moreover, it is possible that HNTs presence leads to a good dispersion of the argan particles, as well as HNTs themselves, boosting the reinforcing effect of the fillers [58]. As it was expected, MLO-compatibilized system shows a great impact strength due to the plasticizing effect of MLO. Several studies have reported the effectivity of MLO as a compatibilizing agent to increase the impact strength of fragile materials, such as PLA or PLA/almond shell flour blends [33]. Impact strength of HNTs/MLO combinated system is very similar to that of MLO blend, remarking the plasticizing effect of this oil. In this sense, the addition of HNTs to the composites with MLO did not provide a significant difference.

With regard to Shore D hardness, a similar trend to that observed in Young modulus (E) can be observed. HNTs sample shows the highest hardness (60.6), giving evidence of the reinforcing effect of nanotubes and MAS. A very similar value is displayed by the ternary blend Bio-HDPE/PE-g-MA/MAS (59.2), reflecting the hardening effect of MAS, composites without MLO did not show significant differences between them. As expected, MLO samples exhibit the lowest hardness (53.2), due to the great ductility this component bestows upon the blend. Ferri et. al. reported this same effect in PLA/TPS blends with different proportions of MLO [60]. The combination of HNTs and MLO leads to a slightly superior hardness to that of the MLO sample, this provoked by the hardening effect of HNTs and the reduction in MLO concentration (going from 7.5 phr to 3.75 phr). The ANOVA analysis showed that MLO had a clear effect on the hardness.

#### *3.2. Morphology of Bio-HDPE/MAS Blends*

Internal morphology of these materials is directly related to their mechanical properties. Figure 3 shows the field emission scanning electron microscopy (FESEM) images at 1000× of argan particles and the surface of fractured impact samples of each one of the blends. First, Figure 3a shows the morphology of an argan particle, which presents a rough surface, emphasized by the presence of holes on its structure (arrow). Those holes play an important role in the adhesion of the particles into the matrix according to Crespo et al. [56,61], who reported a very similar morphology in almond shell flour particles. Figure 3b corresponds to neat Bio-HDPE, which exhibits the typical irregular, rough, and cavernous surface of a ductile polymer, as it was also reported by Quiles-Carrillo et. al. [57]. Regarding the incorporation of argan particles, they show great cohesion within the matrix, as it is shown in Figure 3c–f. The gap between the perimeter of the particles and the polymer matrix is really narrow (arrows), which implies good interaction between both elements. This behavior can be well-related to the action of PE-g-MA as a compatibilizer and justifies the increase in stiffness, in terms of elastic modulus, of the samples that do not have MLO in their structure. Samples in Figure 3e,f show wire-shaped long structures in the polymer matrix (circles). They indicate a more ductile fracture and they confirm the plasticizing effect of MLO. This is directly related to the high elongation at break displayed in mechanical properties. Moreover, argan particles seem to be more immersed in the matrix in samples with MLO, revealing a contribution of the oil to compatibilization. This was also observed by Quiles-Carrillo et. al. [13], when adding MLO to PLA and almond shell flour blends. Additionally, with the presence of HNTs and MLO in the structure, a positive decrease in the number of gaps inside the matrix can be observed. Those gaps are originated by the fall of argan particles after fracture, thus being indicative of a worse interaction between MAS and Bio-HDPE.

**Figure 3.** Field emission scanning electron microscopy (FESEM) images at 1000× of the fractured surfaces of: (**a**) MAS; (**b**) neat Bio-HDPE; (**c**) Bio-HDPE/PE-g-MA/MAS; (**d**) Bio-HDPE/PE-g-MA/MAS/HNTs; (**e**) Bio-HDPE/PE-g-MA/MAS/MLO; (**f**) Bio-HDPE/PEg-MA/MAS/HNTs/MLO.

Presence of HNTs in the blends can be observed in argan particles in Figure 3d,f. Figure 4a,b better illustrates them, where 5000× FESEM images of Bio-HDPE/PE-g-MA/MAS/HNTs and Bio-HDPE/PE-g-MA/MAS/HNTs/MLO samples are shown, respectively. In this context, a correct distribution of nanotubes can be observed in the surface of argan particles, which are indicated in the images by circles. This behavior could be attributed to a synergy between hydroxyl groups of HNTs and hydroxyl and anhydride groups present in PE-g-MA and lignocellulosic compounds of MAS [62]. As expected, nanotubes density is superior in the 7.5 phr HNTs sample (Figure 4a), where nanotubes cover almost all the argan particle surface. While in the 3.75 phr HNTs sample (Figure 4b), nanotubes are disposed in several agglomerates in the surface of MAS. This fact could be responsible for the low increase in resistance in 3.75 phr HNTs sample, altogether with the plasticizing effect of MLO.

**Figure 4.** Field emission scanning electron microscopy (FESEM) images at 5000× of the fractured surfaces of samples compatibilized with HNTs: (**a**) Bio-HDPE/PE-g-MA/MAS/HNTs (7.5 phr HNTs); (**b**) Bio-HDPE/PE-g-MA/MAS/HNTs/MLO (3.75 phr HNTs).

### *3.3. Thermal Properties of Bio-HDPE/MAS Blends*

Figure 5 gathers the DSC thermograms corresponding to the second heating cycle of the studied samples; Table 3 gathers melting temperature (Tm) and crystallinity (Xc) values of every Bio-HDPE blend. Regarding thermograms, in this case only melting temperature can be observed, because glass transition temperature is really low for Bio-HDPE (−100 ◦C) [63]. In the case of neat Bio-HDPE, it shows a melting temperature of 131 ◦C and a crystallinity XC of 66.3%. Similar values were reported by Quiles-Carrillo et al. [64]. After the incorporation of argan particles alongside PE-g-MA into the Bio-HDPE, melting temperature undergoes a slight decrease down to 130.4 ◦C. On the other hand, crystallinity highly decreases to 49.7%. This could be due to the fact that, although argan particles can induce heterogeneous nucleation over the crystallization process of the polymer, high MAS concentrations (over 10% MAS) can compromise nucleation because of the particle–particle contact. Therefore, a limitation in space for crystal formation and growth is produced. This phenomenon was observed by Essabir et al. [65]. HNTs increase Tm up to 133.1 ◦C. This effect was also observed by Erding et al. [66] after studying the melting temperature of polyethylene and halloysite nanotubes composites (PE/HNTs). This can be ascribed to the fact that HNTs have some insulating properties against heat transfer, which is common in clay-based materials. Crystallinity suffers a great decrease in this case (37.6%), probably due to an excess of nucleating agents in the blend. This excess is originated by the combined presence of HNTs and argan particles. In this context, several studies have demonstrated the ability of HNTs to favor crystallization of polymer matrices, as long as there is no overload (+20 phr), to avoid the formation of aggregates that hinder the crystal nucleation and growth process [67]. Regarding MLO, it does not vary Tm, but reduces, along with argan, crystallinity down to 39.8%. MLO improves compatibilization and, therefore, argan particle dispersion in the polymer matrix, thus contributing to reducing polymer–polymer interactions and favoring crystal formation [68]. However, in this case there are excess of argan particles, as it was aforementioned, that diminishes the crystallinity of the matrix. HNTs, MAS, and MLO sample hardly presents any change referring to the melting temperature and it reduces crystallinity in a very similar manner to that of MLO sample, due to obstruction of heterogeneous nucleation, as it has been previously explained. Regarding the statistical analysis of the variance, the melting temperatures did not showed modifications. A different trend could be observed when talking about the melting enthalpy. The first significant difference could be observed when the filler was introduced and as a result a dilution effect was produced on the blends. Depending on the compatibilization strategy of each material, the degree of crystallinity showed minor modifications as commented above.

**Figure 5.** Differential scanning calorimetry (DSC) thermograms of Bio-HDPE/PE-g-MA/MAS blends with different compatibilizers.

**Table 3.** Melting temperature (Tm), melting enthalpy (ΔHm) and crystallinity (XC) of Bio-HDPE/PE-g-MA/MAS blends with different compatibilizers, obtained by differential scanning calorimetry (DSC).


a–d Different letters in the same column indicate a significant difference among the samples (*p* < 0.05).

Thermal stability has also been studied. Figure 6 shows both TGA curves and their first derivative (DTG), while Table 4 gathers the main quantitative parameters related to thermal degradation. Neat Bio-HDPE presented a temperature of approximately 345 ◦C for a mass loss of 5 wt.-% (T5%), while its maximum degradation rate temperature was (*Tdeg*) 447 ◦C. Additionally, after a single-phase degradation, its residual mass at 700 ◦C was 0.3%. Montanes et. al. [69] observed a similar thermal profile for neat Bio-HDPE. Addition of argan fillers and compatibilizers seem to have decreased the initial degradation temperature to values between 275 and 290 ◦C, the Tukey analysis showed a significant difference in these cases. This is mainly attributed to the presence of MAS in the polymer structure, which reduces in certain manner its thermal stability. This was also reported by Essabir et. al. [26] when they studied the properties of polypropylene and almond shell flour blends. Interestingly, it can be seen as all other samples undergo several degradation stages. Particularly, Bio-HDPE/PE-g-MA/MAS sample shows a threephase degradation, attributed to the hemicellulose and pectin degradations (280–340 ◦C); cellulose degradation (340–448 ◦C) and lignin degradation (448–477 ◦C) which are present in argan particles [26,70]. These different phases appear in all samples except for neat Bio-HDPE. MLO samples seem to provide some more thermal stability to the blend, due to its capacity to link polymer chains (crosslinking), which is a positive effect toward stability. With regard to the maximum degradation rate temperature, it has increased in all cases except the MLO sample. This is highly related to the compatibilizing effect of PE-g-MA altogether with argan particles and HNTs [58], leading to a good particle dispersion and retarding the maximum degradation peak to 481.3 and 458.3 ◦C for Bio-HDPE/PE-gMA/MAS and Bio-HDPE/PE-g-MA/MAS/HNTs samples, respectively. Referring to HNTs, their tubular hollow structure, which is capable of holding degradation products inside, is responsible for slowing the mass transport mechanism, as it was reported by Pratap et al. [58].

**Figure 6.** (**a**) Thermogravimetric analysis (TGA) curves and (**b**) first derivative (DTG) of Bio-HDPE/PE-g-MA/MAS blends with different compatibilizers.

**Table 4.** Main thermal degradation parameters of the Bio-HDPE/MAS blends with different compatibilizers in terms of the onset degradation temperature at a mass loss of 5 wt.% (*T5%*), maximum degradation rate (peak) temperature (*Tdeg*), and residual mass at 700 ◦C.


a–e Different letters in the same column indicate a significant difference among the samples (*p* < 0.05).

According to residual mass results, all samples have more residual mass than neat Bio-HDPE. This is especially logic with HNTs, as they are mostly inorganic and possess a very high degradation temperature [58]. In this case 7.5 phr HNTs sample retains a residual mass of 7.8 wt.-% while the sample with 3.75 phr HNTs maintains a 4.3 wt.-% residual mass. The introduction of the lignocellulosic fillers also provided an increasement of the residual mass due to the introduction of lignocellulosic compounds that cannot be degraded at 700 ◦C as Liminana et al. reported for the Almon Shell Flour [36]. Regarding the residual mass, all the samples show significant differences between the previous blend, this is manly by the incorporation of the HNT and the MAS that provide different residual mass as mentioned before.

### *3.4. Dynamic-Mechanical Behavior of Bio-HDPE/PE-g-MA/MAS Blends*

Thermomechanical properties of blends were studied by DMTA technique. Figure 7a shows the evolution of the storage modulus (G') from −150 to 120 ◦C. In relation to neat Bio-HDPE, an initial decrease in G' can be observed until −100 ◦C approximately. This is directly connected with the glass transition of the material. Then, the storage modulus further decreased due to softening of the polymer matrix. Quiles-Carrillo et. al. [57] observed a very similar profile for this polyolefin. Addition of MAS and PE-g-MA into the matrix slightly increases G' modulus along all the temperature range. Specifically, a rise is produced at −140 ◦C (see Table 5) from 2513 MPa (Bio-HDPE) to 2523 MPa (Bio-HDPE/PE-g-MA/MAS). This effect emphasizes as temperature decreases, going

from 1309 to 1413 MPa at −25 ◦C. This is caused by a stiffening of the polymer due to the reinforcing effect of argan particles [65]. As expected, HNTs presence highly augments the storage modulus, providing values of 3111 and 1898 MPa at −140 and −25 ◦C (see Table 5), respectively, in comparison to the values obtained previously for neat Bio-HDPE. This is indicative of an increase in the stiffness of the material, confirming the results commented in the mechanical properties section, associated with a high dispersion of HNTs in the matrix. Thus provoking a decrease in polymer chain mobility due to physical interactions between nanotubes and adjacent chains [71]. In the case of Bio-HDPE/PE-g-MA/MAS/MLO sample, superior values of storage modulus were observed until 0 ◦C in comparison to Bio-HDPE and Bio-HDPE/PE-g-MA/MAS samples, but the modulus was inferior in the last section, till 120 ◦C. This reduction can be associated with the plasticizing effect of MLO. Although the difference is not significant and, given the modulus increase in the initial section of the diagram, it can be asserted that MLO acts as a dispersing element of the argan particles, in combination with PE-g-MA. Quiles-Carrillo et. al. [57] observed a similar behavior by studying the effect of MLO to compatibilize Bio-HDPE/PLA blends. Finally, the combined effect of HNTs (3.75 phr) with MLO (3.75 phr) enhances the stiffness of Bio-HDPE in all the temperature range, but to a lesser extent than that of 7.5 phr HNTs sample, as the concentration is lower in this case. Nonetheless, a good compatibilization and dispersion effect of MLO toward HNTs and MAS can be seen. The statistical analysis shows a similar trend of the storage modulus with respect to the elastic modulus.

**Figure 7.** Plot evolution of (**a**) the storage modulus (*G* ) and (**b**) the dynamic damping factor (tan δ) of the injection-molded samples of Bio-HDPE/PE-g-MA/MAS blends with different compatibilizers.



**\*** The Tg has been measured using the tan δ peak maximum criterion. a–e Different letters in the same column indicate a significant difference among the samples (*p* < 0.05).

> Figure 7b shows the evolution of the damping coefficient (tan δ) with temperature for every studied blend. The peak observed at −115 ◦C for Bio-HDPE corresponds to γ-relaxation of the green polyolefin, which is directly related to its glass transition temperature Tg [72], which is considered as the maximum peak value. This value is very similar to those observed in other studies [57]. A second relaxation, which is called α-relaxation,

can be located between 50 and 120 ◦C and is associated with an interlaminar shear process. α-relaxation is often separated into two processes (α y α') due to an inhomogeneity in crystalline regions [73]. The Tg change given by the addition of MAS with different compatibilizers is not so significant. In the case of Bio-HDPE/PE-g-MA/MAS and Bio-HDPE/PE-g-MA/MAS/HNTs samples, a reduction of about 3–4 ◦C is produced, while MLO reduces it even further (5 ◦C). It is logic that the 7.5 phr MLO sample possesses the lowest Tg due to the plasticizing effect previously mentioned. Regarding the combination of HNTs (3.75 phr) and MLO (3.75 phr), they seem to exert a synergetic effect that moves the glass transition peak +1 ◦C approximately. This can be associated with a good dispersion of HNTs into the polymer matrix, which immobilize Bio-HDPE chains to some extent. Only with the superposition of all the effects commented before, a significant difference could be observed, HNT/MLO blend reduced the Tg to the lowest temperature (−114.3 ◦C).

### *3.5. Color Measurement of BIO-HDPE/PE-g-MA/MAS Blends*

Color, luminance, and transparency are essential issues to be considered in materials that try to imitate wood or WPC, in this case reddish-like, as those parameters determine how similar in appearance are these materials to wood. Table 6 gathers color coordinates values of Bio-HDPE/PE-g-MA/MAS blends, while Figure 8 shows the visual appearance of tensile test samples. All samples are opaque, mainly due to the semicrystalline nature of Bio-HDPE, which does not allow light to pass through [74].

**Table 6.** Luminance and color coordinates (L\*a\*b\*) of the Bio-HDPE/PE-g-MA/MAS blends with different compatibilizers.


a–c Different letters in the same column indicate a significant difference among the samples (*p* < 0.05).

L\*a\*b\* color coordinates were measured on injection-molded samples in order to analyze the color variations and luminance. Luminance (L\*) is indicative of lightness. Neat Bio-HDPE showed high brightness as a consequence of its characteristic white color,

in contrast with the brown-like color of the other samples, which present low luminance as a result of the addition of fillers and compatibilizers. Referring to color coordinate a\*, it is indicative of the color change between green (negative) and red (positive). Particularly, neat Bio-HDPE presents a negative value but close to 0, −2.29 in this case, because of the proximity of the sample to white color. Similar value was reported by Rojas-Lema et al. [75] in neat Bio-HDPE films. The rest of materials have positive and quite similar values each due to the characteristic brown color of micronized argan shell [76]. However, a slight change in brown tonality can be observed between those samples with and without MLO. MLO samples display a darker brown color, as a result they present a lower a\* value than samples lacking MLO, which have a clearer brown color as well as closer to pure red. This can be due to the trend of MLO to take a\* coordinate to more negative values, as it was observed by Quiles-Carrillo et. al. [77]. Regarding b\* coordinate color, it defines blue (negative) and yellow (positive) colors. Neat Bio-HDPE has a negative value of −5.35, which is similar to that reported by Rojas-Lema et. al. [75]. While the other materials present similar values between 4 and 5, due to approaching the intrinsic yellow-like color of the argan particles and MLO. It should be noted that yellow tonality of MLO is less intense than that of argan particles, therefore reaching lower b\* values. On the other hand, HNTs do not affect color significantly but seem to reduce color coordinate a\* and slightly increase coordinate b\*. The statistical results show that the biggest color difference was assessed by the introduction of the filler, the compatibilization strategies followed slightly provided modifications between them.

Colors showed by the materials studied here make them perfect candidates for wood plastic composites [78], presenting reddish-like brown colors that allow them to be used in wood-based products fabrication, where quality and aesthetic are vital and determine the success or failure of the product [79].

#### *3.6. Wetting Properties and Water Absorption of Bio-HDPE/PE-g-MA/MAS Blends*

One of the main disadvantages of green composites or any material based on hydrophilic fillers is their tendency to absorb water. With the objective of evaluating the behavior of Bio-HDPE/PE-g-MA/MAS blends against water, contact angle was measured at different times after applying a water drop over each sample surface. A great contact angle is indicative of a poor affinity toward water. Table 7 shows different contact angles for the blends at 0, 5, 10, 15, 20, and 30 min after the administration of the water drop. It can be seen as initially, all samples are hydrophobic, as their contact angles are far superior to 65◦, which is the hydrophilic threshold according to Vogler [80]. This can be ascribed to the fact that the polymer matrix (Bio-HDPE) is a completely nonpolar compound, formed only by C-H bonds. This bond is formed by atoms with practically the same electronegativity. It is this reason that makes neat Bio-HDPE barely reduce its contact angle over time. When PE-g-MA and MAS are introduced into the polymer, the contact angle suffers a quick reduction (56.7◦ at 30 min). This can be attributed to the oxygen-based groups in PE-g-MA (anhydride groups) and to cellulose, hemicellulose, and lignin present in argan particles (hydroxyl and carbonyl groups). These groups give polarity to the material and can form hydrogen bonds with water (polar solvent), providing hydrophilicity over time [81]. Addition of HNTs into the blend provokes a faster decrease in contact angle than the previously mentioned samples. However, the contact angle at 30 min is very similar and slightly superior to that of Bio-HDPE/PE-g-MA/MAS. This could be attributed, on the one hand, to polar functional groups (hydroxyl) present in HNTs structure and, on the other hand, to their ability to retain different molecules within their tubular hollow structure, water being one of those molecules [82]. As expected, Bio-HDPE/PE-g-MA/MAS/MLO sample shows the lowest contact angle over time, reaching a considerably lower value than the rest of the samples (22.1◦). This is ascribed to the high polarity of the MLO molecule, which has a great amount of anhydride and carbonyl groups, through which water can form hydrogen bonds as long as water and sample remain in contact [83]. Figure 9 perfectly reflects this behavior on the MLO sample, where the drop of water flattens over

time. Finally, MLO and HNTs sample greatly reduces the hydrophobicity of the material due to the aforementioned reasons, verifying the effect of MLO over wettability.

**Table 7.** Contact angle (*θw*) of different Bio-HDPE/PE-g-MA/MAS blends with different compatibilizers at several times of exposure to water: 0, 5, 10, 15, 20 and 30 min.


a–e Different letters in the same column indicate a significant difference among the samples (*p* < 0.05).

**Figure 9.** Water contact angle of the sample Bio-HDPE/PE-g-MA/MAS/MLO over time: 0, 5, 10, 15, 20, and 30 min.

From all these results it can be concluded that all fillers and compatibilizers used do not significantly vary hydrophilicity of Bio-HDPE at first, but they do increase water absorption in case of exposure for long periods of time. This could be corroborated by the statistical study where only some significant differences appeared at the initial time, but these were clearly accentuated over the time due to the different behaviors of the blends.

Additionally, water absorption capabilities under long time exposure were studied for each material by means of water absorption test. Figure 10 shows the water absorption of each sample after 14 weeks of immersion in distilled water. It can be observed as neat Bio-HDPE barely absorbed water, presenting an asymptotic value of 0.05 wt.-%. This remarks the hard hydrophobic nature of Bio-HDPE, as it has been previously said during the contact angle analysis. With the addition of argan particles and PE-g-MA, water absorption increases up to 0.98 wt.-%, after a period of 14 weeks. This means an increase of 94% in relation to neat Bio-HDPE. This is closely related to hydroxyl groups in lignocellulosic compounds of argan particles, which increase the facility of the material toward capturing humidity [84]. The incorporation of HNTs (7.5 phr) in the blend causes an increment in water absorption up to 1.27 wt.-%, which can be ascribed to the hollow tubular morphology of nanotubes. Their structure helps them to trap water molecules with ease, along with the effect of polar hydroxyl groups. Lastly, MLO sample achieves the highest water absorption (1.55 wt.-%) due to the plasticizing effect that it exerts over Bio-HDPE matrix, increasing its free volume and making the diffusion of water within its structure easier, as it was observed by Quiles-Carrillo et. al. [85]. As expected, 3.75 phr HNTs and 3.75 phr MLO sample presents intermediate results between those of individual HNTs and MLO samples (7.5 phr each). It should be also noted that addition of HNTs and MLO increases the water absorption speed during the first week, denoted by a more pronounced slope in the initial section showed by all three samples, in comparison with Bio-HDPE/MAS/PE-g-MA sample.

According to these results, although hydrophilicity provided by argan particles (MAS) could prove a disadvantage in some ambits, it could certainly give applicability to these materials in some other fields. One of the applications could be flowerpot fabrication, so that when the plant is irrigated, water excess is absorbed by the pot material over time.

#### *3.7. Infrared Spectroscopy*

Chemical composition of the injection-molded samples was analyzed by means of Fourier transformed infrared spectroscopy (FTIR). Figure 11a gathers the individual spectrums of the compatibilizers and fillers used (PE-g-MA, MAS, HNTs y MLO) from 4000 to 600 cm−1. First, PE-g-MA spectrum is very similar to that of Bio-HDPE, as it is a polyethylene-based (PE) compound. Absorption bands at 2840 and 2910 cm−<sup>1</sup> are associated with the stretching vibration for -CH-, -CH2-, or -CH3 [86]. The peak at 1468 cm−<sup>1</sup> corresponds to the deformation vibration of -CH2- or -CH3 [86]. On the other hand, the peak at 720 cm−<sup>1</sup> is due to (CH2)n rock when n ≥ 4 [86]. In the case of MAS, two bands stand out at 1030 and 900 cm−1, which are related to the stretching vibration of C-O and C-OH in polysaccharide rings in cellulose [87,88]. A peak at 1114 cm−<sup>1</sup> can also be appreciated due to the symmetric glycosidic stretching of C-O-C bond in polysaccharide compounds of cellulose. Moreover, little absorption bands in the 1500 and 1130 cm−<sup>1</sup> region can be observed, which indicate the presence of characteristic groups of cellulose, hemicellulose, and lignin. Some examples of those groups are the peak at 1227 cm<sup>−</sup>1, related to the stretching vibration C=O of the acetyl group in lignins [89]; and the band at 1420 cm−1, which corresponds to the vibration of aromatic rings in lignin [87]. The low-intensity band at 1720 cm−1, is ascribed to non-cellulosic compounds (pectin, lignin, and hemicellulose) [89]. Another weak peak can be found at 2900 cm−1, which is related to stretching vibration of C-H bonds in CH and CH2 groups, present in cellulose and hemicellulose [65]. A final little band is found at 3300 cm<sup>−</sup>1, associated to the stretching of O-H bonds in carbohydrates (cellulose and hemicellulose) [55]. Regarding the halloysite nanotubes (HNTs) spectrum, peaks at 3695 and 3619 cm−<sup>1</sup> are due to the stretching vibration of Al2OH- (each OH is linked to two Al atoms) [90]. Bands at 3670 and 3650 cm−<sup>1</sup> are ascribed to the O-H stretching of inner-surface hydroxyl groups and the out-of-phase vibration of the inner surface hydroxyl groups, respectively [90–93]. The low intensity peak at 1649 cm−<sup>1</sup> is attributed to bending vibrations due to absorbed water [94], the band at 1117 cm−<sup>1</sup> is due to apical Si-O bonds, while 1022 and 682 cm−<sup>1</sup> bands are associated with the perpendicular stretching of Si-O-Si bonds [93,94]. Peaks at 937 and 909 cm−<sup>1</sup> are consequence of the O-H deformation of inner-surface hydroxyl groups and

O-H deformation of inner hydroxyl groups, respectively [93,94]. Two last bands analyzed are located at 790 and 749 cm−1, and can be ascribed to the O-H translation vibrations of halloysite O-H units [93]. With regard to MLO spectrum, a first peak is located at 3006 cm<sup>−</sup>1, corresponding to =C-H stretching of double carbon-carbon bonds, and those at 2925 and 2850 cm−<sup>1</sup> refer to the antisymmetric and symmetric stretching vibration C-H of saturated carbon-carbon bonds (C-C), respectively [95]. Other bonds can be found at 1740 and 1709 cm−1, indicative of the C=O carbonyl stretching vibration of the ester and maleic anhydride, respectively; another band at 1158 cm−<sup>1</sup> is due to the stretching vibration C-O-C, C-O, and C-C in ester groups; the peak at 720 cm−<sup>1</sup> is normally related to the out-of-plane C-H stretching vibration of saturated C-C bonds [95]. Finally, the band at 1462 cm−<sup>1</sup> is due to C-H bending vibration, while the absorption peaks at 1862 and 1784 cm−<sup>1</sup> are attributed to anhydride groups.

**Figure 11.** (**a**) Fourier transform infrared (FTIR) spectra, from bottom to top, of maleinized linseed oil (MLO), halloysite nanotubes (HNTs), micronized argan shell (MAS), Polyethylene-graft-maleic anhydride (PE-g-MA) (**b**) FTIR spectra of the different blends, from top to bottom Bio-HDPE, Bio-HDPE(67 wt.%)/PEg-MA(3 wt.%)/MAS(30 wt.%),Bio-HDPE(67 wt.%)/PE-g-MA(3 wt.%)/MAS(30 wt.%)/HNTs(7.5 phr), Bio-HDPE(67 wt.%)/PE-g-MA(3 wt.%)/MAS(30 wt.%)/MLO(7.5 phr), Bio-HDPE(67 wt.%)/PE-g-MA (3 wt.%)/MAS(30 wt.%)/HNTs(3.75 phr)/MLO(3.75 phr).

Figure 11b gathers infrared spectrums for neat Bio-HDPE as well as all the developed blends. The main bands of neat Bio-HDPE are located at 2914, 2846, 1466, and 718 cm−1, and they are related to the stretching vibration, bending deformations, and rocking of methylene groups (CH2) [96]. These peaks are found in all samples, but with higher

intensity, due to the presence of PE-g-MA, which is a PE-based material and as such presents the same nature as Bio-HDPE. Thus, their spectrums are very similar between each other. The weak absorption bands between 1370 and 1350 cm−<sup>1</sup> are associated with wagging and symmetric deformations of CH2 and CH3 groups, respectively. The addition of MAS compatibilized with PE-g-MA makes a band in the range 1025–1091 cm−<sup>1</sup> to appear, which is mainly attributed to the formation of PE-g-MA dimers or oligomers, identified with the stretching of C-C and C-O bonds [59]. Moreover, this band could be combined with another one at 1030 cm<sup>−</sup>1, relative to the presence of C-O and C-OH bonds in MAS particles cellulose. The presence of HNTs in the blend deforms and increases the intensity of the peak at 1030 cm−1, due to the perpendicular Si-O-Si stretching, as it has been previously mentioned in the individual spectra of HNTs. Additionally, in HNTs samples a weaker band at 900 cm−<sup>1</sup> can be observed, ascribed to the deformation of O-H groups [90]. Moreover, little peaks can be located in the 3650 cm−<sup>1</sup> region, which have been previously identified in the individual analysis of HNTs, associated with the stretching vibration of inner-surface O-H groups in the nanotubes. Regarding the MLO presence in the blend, it produces changes found at 1690 cm−<sup>1</sup> approximately, related to the C=O stretching in hydrolyzed anhydride groups in MLO [13]. This peak has already been observed in MLO individual spectrum, as it has been aforementioned, and it does not appear in those samples without MLO on their structure. The appearance of low-intensity bands at 3700 and 3600 cm−<sup>1</sup> can be ascribed to the absorption of water, catalyzed by MLO and argan particles, which are polar compounds with great amount of functional oxygen-based groups and great affinity toward water. HNTs and MLO blend shows the same characteristic peaks described for each individual sample, but with less absorption intensity, as their concentration reduces from 7.5 phr to 3.75 phr.

These results seem to reveal certain compatibilization between elements. This fact is denoted by a general increase in the intensity of peaks related to oxygen-based groups, indicating a possible bonding between anhydride groups in PE-g-MA and MLO with hydroxyl groups in MAS and HNTs, increasing the affinity reciprocally within the polymer matrix.

#### **4. Discussion**

The present work shows the incorporation of argan shell wastes in wood plastic composites with biobased polyethylene matrix and different compatibilizers. The incorporation of this type of loads can be effectively used as new reinforcement elements in order to create parts prepared by conventional industrial processes for thermoplastic materials, in special injection molding. In relation to the mechanical properties, the incorporation of MAS (30 wt.%) with PE-g-MA (3 wt.%) into the Bio-HDPE matrix increases the Young modulus up to 846 MPa. This can be directly related to a proper distribution of MAS particles in the polymer matrix, leading to a good adhesion of the filler with the polymer and obtaining a more rigid material. Addition of HNTs in the blend showed an additional enhancement of stiffness. In this context, Bio-HDPE/PE-g-MA/MAS/HNTs sample presented the highest Young modulus of all blends (1126 MPa). The incorporation of MLO in the green composites increased 100% of the elongation at break of the Bio-HDPE/PE-g-MA/MAS blend. From a morphological point of view, the incorporation of argan particles in the BioHDPE matrix, showed great cohesion, which implies good interaction between both elements. Furthermore, the addition of MLO verified the results obtained in the mechanical properties, showing a more ductile fracture and confirming the plasticizing effect of MLO. However, the incorporation of the lignocellulosic filler leads to a reduction in thermal properties and a decrease in the crystallinity of the compound. In relation to the thermomechanical properties, the incorporation of HNTs highly augmented the storage modulus, providing values of 3111 and 1898 MPa. This is indicative of an increase in the stiffness of the material, confirming the results commented in the mechanical properties section, associated with a high dispersion of HNTs in the matrix. In general, the incorporation of argan particles and additives provided the materials with reddish-brown colors that allow

their use in the manufacturing of wood-based products. In relation to one of the main drawbacks of green composites, it can be seen how the incorporation of lignocellulosic particles greatly increases the water absorption of the composites, generating certain disadvantages. In addition, the incorporation of the MLO increases the water absorption capacity of the composites, due to the plasticization effect it exerts over the Bio-HDPE matrix.

#### **5. Conclusions**

The results obtained in this work indicate that it is possible to obtain WPC with high renewable content with Bio-HDPE, lignocellulosic fillers and natural additives. This type of green composites can greatly favor the generation of circular economies focused on giving added value to agri-food waste from the Mediterranean basin, also favoring the creation of highly efficient polymers at very competitive costs. The materials obtained in this work have proved to possess excellent properties at a reduced cost in comparison with the neat Bio-HDPE. High ductile properties, relatively high stiffness and good thermal stability were reported, as well as a visual appearance very similar to that of reddish-color woods, which is essential in a wood plastic composite. The samples showed certain hydrophilicity over time, which can prove to be a disadvantage, although it is unusual for these materials to be immersed in water for long periods of time. However, if it were the case, they have some applications too, as it is the fabrication of plant pots. All in all, the affinity between argan particles and Bio-HDPE has been successfully increased in this study with the use of compatibilizers, and their properties vastly displayed and demonstrated. This work opens a new line of research regarding the development of new materials formed by polar fillers and non-polar polymeric matrices.

**Author Contributions:** Conceptualization, L.Q.-C. and P.M.S.; methodology, S.R.-L.; validation, J.I.- M., J.G.-C., and L.Q.-C.; formal analysis, M.J.-R.; investigation, M.J.-R.; data curation, J.I.-M.; writing original draft preparation, M.J.-R.; writing—review and editing, J.G.-C. and J.I.-M.; supervision, P.M.S.; project administration, L.Q.-C.; All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Ministry of Science, Innovation, and Universities (MICIU) project number MAT2017-84909-C2-2-R.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** L. Q.-C. wants to thank Universitat Politècnica de València for his post-doctoral PAID-10-20 grant from (SP20200073). J. I.-M. wants to thank the Spanish Ministry of Science, Innovation and Universities for his FPU grant (FPU19/01759). J. G.-C. wants to thank Universitat Politècnica de València for his FPI grant from (SP20200080). S.R.-L. for her Santiago Grisolia grant from Generalitat Valenciana (GVA) (GRISOLIAP/2019/132).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Article* **Yield and Selectivity Improvement in the Synthesis of Carbonated Linseed Oil by Catalytic Conversion of Carbon Dioxide**

**David Alejandro González Martínez, Enrique Vigueras Santiago and Susana Hernández López \***

Laboratorio de Investigación y Desarrollo de Materiales Avanzados, Facultad de Química, Universidad Autónoma del Estado de México, Campus Rosedal, Toluca 50200, Mexico; dgonzalezmartinez31@gmail.com (D.A.G.M.); eviguerass@uaemex.mx (E.V.S.) **\*** Correspondence: shernandezl@uaemex.mx

**Abstract:** Carbonation of epoxidized linseed oil (CELO) containing five-membered cyclic carbonate (CC5) groups has been optimized to 95% by reacting epoxidized linseed oil (ELO) with carbon dioxide (CO2) and tetrabutylammonium bromide (TBAB) as catalysts. The effect of reaction variables (temperature, CO2 pressure, and catalyst concentration) on the reaction parameters (conversion, carbonation and selectivity) in an autoclave system was investigated. The reactions were monitored, and the products were characterized by Fourier Transform Infrared Spectroscopy (FT-IR), carbon-13 nuclear magnetic resonance (13C-NMR) and proton nuclear magnetic resonance (1H-NMR) spectroscopies. The results showed that when carrying out the reaction at high temperature (from 90 ◦C to 120 ◦C) and CO2 pressure (60–120 psi), the reaction's conversion improves; however, the selectivity of the reaction decreases due to the promotion of side reactions. Regarding the catalyst, increasing the TBAB concentration from 2.0 to 5.0 *w*/*w*% favors selectivity. The presence of a secondary mechanism is based on the formation of a carboxylate ion, which was formed due to the interaction of CO2 with the catalyst and was demonstrated through 13C-NMR and FT-IR. The combination of these factors makes it possible to obtain the largest conversion (96%), carbonation (95%), and selectivity (99%) values reported until now, which are obtained at low temperature (90 ◦C), low pressure (60 psi) and high catalyst concentration (5.0% TBAB).

**Keywords:** carbonation reaction; selectivity optimization; carbonated epoxidized linseed oil; non-isocyanate polyurethane

### **1. Introduction**

Obtaining cyclic carbonates (CCs) has received significant attention because CCs have attractive properties, such as low toxicity, high solubility, and boiling points, can be used as solvents, and have high reactivity with amines [1,2]. Although 5-membered cyclic carbonates (CC5) are less reactive than 6-, 7-, and 8-element cyclic carbonates (CC6, CC7, and CC8), respectively, the synthesis of CC5 has been more studied because CC5 synthesis does not involve the use of toxic precursors such as CS2, phosgene, and ethyl chloroformate, among other environmentally harmful solvents [3,4]. The most common synthesis method of CC5 remains the insertion of carbon dioxide (CO2) in cyclic ethers because it is considered a safe process, and the atom economy is close to 100% [4]. This synthesis route has been extensively studied with small molecules such as ethylene and propylene oxide [2,5], while long-chain molecules, such as vegetable oils (VOs), have been less studied. Carbonated vegetable oils (CVOs) have various applications as solvents, lubricants, additives, plasticizers, and monomers for the formation of polymers [6–9]. Currently, the main application is in the synthesis of non-isocyanate polyurethanes (NIPUs) through the aminolysis reaction of CVO because NIPU does not require highly toxic materials such as isocyanates [10–16].

Vigueras Santiago, E.; Hernández López, S. Yield and Selectivity Improvement in the Synthesis of Carbonated Linseed Oil by Catalytic Conversion of Carbon Dioxide. *Polymers* **2021**, *13*, 852. https:// doi.org/10.3390/polym13060852

**Citation:** González Martínez, D.A.;

Academic Editor: Rafael Antonio Balart Gimeno

Received: 14 January 2021 Accepted: 5 March 2021 Published: 10 March 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Although CC has been synthesized since 1933 by Carothers et al. [17], it was not until 2004 that Tamami et al. obtained CC5 from epoxidized soybean oil (ESO) and CO2 for the first time in the presence of TBAB as a catalyst (Scheme 1) [18]. A significant amount of research has focused on developing catalysts or co-catalysts to enhance carbonation kinetics. Generally, catalysts are composed of a Lewis acid for the oxirane's electrophilic activation and a Lewis base that acts as a nucleophile. Some of these catalysts are quaternary phosphate compounds, phosphine complexes, metal complexes, alkali metal halides, quaternary ammonium salts, and ionic liquids, among others [19–25]. Additionally, it has been observed that the catalytic activity increases when the acidity of the cation and the nucleophilicity of the halide increase [19]. Therefore, co-catalyst systems have been proposed, including CaCl2, SiO2-I, palladium doped with H3PW12O40/ZrO2, TBAB+SnCl4, and TBAB+H2O [10,18,19,26]. Among the catalyst systems studied, TBAB is the most commonly used catalyst in the carbonation reaction due to bromine's effectiveness as a leaving group [27]. However, one disadvantage of TBAB is that at high temperatures (150–190 ◦C), it shows decomposition to volatile components (Hofmann reaction) (Scheme 2) [6].

**Scheme 1.** Cycloaddition reaction of carbon dioxide (CO2) into oxirane rings [18].

**Scheme 2.** Products obtained from the Hofmann elimination reaction [6].

Furthermore, several investigations have focused on studying operational and equipment parameters using TBAB as a catalyst. Tamami et al. (2004) carried out the carbonation reaction under mild conditions (atmospheric pressure; 110 ◦C). Despite obtaining relatively high conversion values (94% by Fourier transform infrared spectroscopy (FT-IR) and 78% by titration) [26], the reaction time was considered long (70 h) and involved an obstacle to commercialization [20]. After that, the studies' main objective was to optimize the reaction time (Table 1). Javni et al. studied the carbonation of ESBO at higher CO2 pressure (56.5 bar), demonstrating that reaction time and conversion are a function of temperature, pressure, and catalyst concentration [2,18]. Doll et al. intensified the carbonation conditions

using CO2 in the supercritical state (103 MPa, 100 ◦C) and reduced the reaction time to 40 h and 100% conversion. Similar results were obtained by Mann et al.; however, using supercritical conditions is a disadvantage due to high energy consumption [6,28]. Regarding the efforts to develop equipment to improve carbonation, Mazo et al. used microwave heating (1 atm, 120 ◦C) and a water ratio (1:3; H2O/poxy), reaching 87% conversion in 40 h [29]. Zheng et al. proposed using a continuous-flow microwave reactor, obtaining 73% conversion in 7 h (6 bar, 120 ◦C). Nevertheless, homogeneity and microwave penetration are still problems to scale to the industry level [30].

Despite the progress made in the investigations of carbonation reactions in vegetable oils, there are still some drawbacks. As shown in Table 1, the main objective of these studies is to achieve high conversion degrees, that is, ensuring that the epoxide groups react mostly. However, the carbonation levels that have been reached are not high, reaching no more than 78% [26,31]. Furthermore, this variable is not measured in most of the reported works. The carbonation level allows us to know if effectively the epoxides are converted into cyclic carbonates or if there is the generation of reaction byproducts that may impact the selectivity of the carbonation reaction.

The conversion kinetics of the reaction improves with increasing temperature and pressure. Nevertheless, the impact of these variables on the degree of carbonation and selectivity has not been studied in detail. Therefore, there are differences in the published results, so it has not been possible to establish the reaction conditions that render a higher percent of carbonation and selectivity, not just conversion (Table 1). This is because we consider it relevant to systematically study the influence of temperature, CO2 pressure and TBAB concentration on the conversion, carbonation and selectivity degree to find a balance of those variables for obtaining the highest reaction parameters.


**Table 1.** Literature Reports of Carbonation Reaction in Epoxidized Vegetable Oils (EVOs) Using Tetrabutylammonium Bromide (TBAB).

ESO = Epoxidized soybean oil; ELO = Epoxidized linseed oil; ECSO = Epoxidized cottonseed oil; ECO = Epoxidized castor oil; EVNO = Epoxidized vernonia oil.

### **2. Materials and Methods**

### *2.1. Materials*

Linseed oil (LO) was purchased commercially through a local distributor (Jalisco, Mexico). LO has a clear yellow oil appearance. The molecular weight (MW), the number of double bonds (DB), and iodine value (IV) were determined by proton nuclear magnetic resonance (1H-NMR) as described in [9,35,36], rendering 920.5 g/mol, 6.86 (DB), and 189.16 (IV), respectively. IR120 (AIR-120H) Amberlite catalyst (1.8 meq/mL by wetted bed volume), TBAB (≥98.0% assay), solvents such as ethyl acetate (≥99.5% assay) and toluene (≥99.5% assay) were purchased from Sigma-Aldrich Química, S.L. (Toluca, EdoMex, México) Acetic acid (≥99.7%), and hydrogen peroxide (50% concentration) were obtained from Fermont (Monterrey, N.L, México). Industrial grade CO2 gas (≥99.5%) was purchased from Praxair (Toluca, EdoMex, México). All reagents were used as received except the LO, which was passed through a chromatographic column filled with α-alumina.

### *2.2. Characterization*

The structural analysis of products and the monitoring of epoxidation and carbonation reactions were studied using FT-IR, Carbon-13 nuclear magnetic resonance (13C-NMR), and proton magnetic resonance (1H-NMR) spectroscopies. FT-IR spectra were obtained on an FT-IR Prestige 21 spectrometer, Shimadzu Scientific Instruments, Inc. in México (Tultitlán, EdoMex, México) equipped with a diamond crystal and a horizontal attenuated total reflectance (HART) module. The infrared spectra were obtained in absorbance mode with 64 scans and a resolution of 4 cm−<sup>1</sup> in the range of 560–4000 cm<sup>−</sup>1. All FT-IR spectra were normalized to the signal at 1736 cm−<sup>1</sup> [37], corresponding to the triglyceride ester group's carbonyl vibration. An Avance III spectrometer (Bruker Mexicana, S.A. de C.V, Cd. de México, México) was used for 13C-NMR and 1H-NMR analysis. The analysis was performed at 300 MHz, with a spectrum width of 3689.22 Hz, a pulse width of 4.75 μs, 32 scans at 293 K, 90 pulse width of 9.5 μs. CDCl3 was used as the solvent, and tetramethylsilane was used as the internal standard. Through 1H-NMR, the number of epoxide and carbonate groups was quantified, as well as conversion, epoxidation, carbonation, and selectivity values of the reactions [8,13,35].

#### *2.3. Synthesis of Epoxidized Linseed Oil (ELO)*

ELO was obtained according to the Prileschajew reaction (industrial method), where DB of LO reacts in situ with a percarboxylic acid (peracetic acid) in the presence of a heterogeneous catalyst (Scheme 3), such as an ion-exchange resin (Amberlite IR120) [38–40]. The conditions of the epoxidation reaction were obtained from a recently optimized methodology [35]. The general procedure consists of placing LO (0.054 mol), toluene (25 mL), acetic acid (0.53 mol/DB), and Amberlite IR120 (12.5 g) inside a three-necked flask equipped with a thermometer, magnetic stirring, and condenser with reflux. The initial mixture was heated to 50 ◦C, and H2O2 (50 wt%) (1.54 mol/DB) was added. Subsequently, the reaction temperature was adjusted to 80◦C and maintained under these conditions for 90 min. Immediately, the reaction mixture was cooled to room temperature, and the product was purified. Then, 150 mL of ethyl acetate was added to the mixture and filtered under vacuum to remove the catalyst. The organic phase was washed with 500 mL of a 10% sodium bicarbonate solution to neutralize acetic acid. Subsequently, the organic phase was dried with anhydrous magnesium sulfate and filtered. The solvent is removed using a rotary evaporator and then placed in a vacuum desiccator [37]. Finally, the product obtained was characterized by both FT-IR and 1H-NMR.

**Scheme 3.** Epoxidation reaction of vegetable oils (VOs).

### *2.4. Synthesis of Carbonated Epoxidized Linseed Oil (CELO)*

As shown in Table 1, various authors have systematically investigated the effect of different variables, both in the process and in operation, on the epoxidation degree of vegetable oils [8,26,32]. From the results obtained, it has been shown that high epoxidation values can be obtained and are spectroscopically similar to each other (e.g., soybean oil, cottonseed oil, linseed oil). However, in most of the studies, a low carbonation degree was obtained. Therefore, it was necessary to study different reaction conditions to determine the best possible parameters to obtain carbonated oils with the highest carbonation degree and reaction selectivity. The carbonation reaction was monitored as a function of time, temperature, CO2 pressure, and catalyst concentration, quantifying by 1H-NMR the content of oxirane and carbonate groups to determine the change in the epoxidation, carbonation, and selectivity percentage. The synthesis was carried out in 500 mL Teflon vessel, inserted into a stainless steel reactor equipped with a temperature controller from room temperature to 400 ◦C and inlets necessary to flow gas and pressurize the reaction in a range from 15 to 200 psi. All the settings of the experimental conditions and the reaction parameters are shown in Table 2. The first variable that was studied was temperature. The reaction temperature was modified from 90 to 120 ◦C, keeping the pressure (90 psi) and the catalyst concentration (2.5%) constant and setting the temperature that presented the best carbonation reaction results. Subsequently, CO2 pressure was modified in an interval from 60 to 120 psi, keeping the temperature (90 ◦C) and catalyst concentration (3.5%) constant. Finally, once the impact of both the temperature and the CO2 pressure on the reaction performance was defined, the catalyst concentration (from 2.5% to 5.0%) was varied at temperature (90 ◦C) and constant pressure (60 psi). Finally, once the effect of both temperature and CO2 pressure on the reaction yield was defined, the catalyst concentration (from 2.5% to 5.0%) was varied at constant temperature (90 ◦C) and pressure (60 psi). The general methodology consisted of placing 5 g of ELO (MW = 1007.3 g/mol; 6.26 epoxide content, EC, values obtained by 1H-NMR) and the amount of TBAB (2.5–5.0% mol with respect to EC) inside the reactor. The mixture was stirred manually for 10 min until complete homogeneity was achieved and the reactor was closed. Prior to initiating the reaction, an oxygen-free atmosphere is generated by performing three CO2 purges. Afterward, the system is adjusted to reaction conditions by holding both temperature (90–120 ◦C) and CO2 pressure (60–120 psi) constant for a defined time (12–92 h). After the reaction time was increased, the system was cooled to room temperature and depressurized. ELO purification is initiated by solving the reaction mixture with ethyl acetate. Subsequently, at least 3 or 5 washes with hot water (50 ◦C) are carried out to remove the catalyst. The organic phase is dried with anhydrous magnesium sulfate. The highest amount of ethyl acetate is removed by distillation at reduced pressure, and the residual solvent is eliminated in a vacuum desiccator for 24 h. Finally, the dry samples were characterized by FT-IR and 13C-NMR. The amount of carbonate groups formed was quantified by 1H-NMR. The runs were performed in triplicate and the results plotted are the average value of each test.


**Table 2.** Experimental Settings and Average Values of the Reaction Parameters: Conversion, Carbonation and Selectivity Percentages.

#### **3. Results**

### *3.1. Characterization of Epoxidated Linseed Oil (ELO)*

FT-IR spectra were obtained for both the raw material (LO) and the corresponding epoxide (ELO), consistent with those reported by different authors [35,37,41–43]. The most representative LO vibration signals (Figure 1a) correspond to the ester carbonyl group at 1736 cm−<sup>1</sup> (C=O) and double-bound signals at 3021 cm−<sup>1</sup> (=C–H), 1652 cm−<sup>1</sup> (C=C) and 720 cm−<sup>1</sup> (HC=CHcis). The other vibration bands correspond to ester carbonyl (1736 and 1159 cm−1), methyl (2922, 1456, and 1377 cm−1), and methylene (2852 and 719 cm−1) groups. In the ELO spectrum (Figure 1b), the epoxy ring's vibration signals are identified at 1250 and 823 cm−<sup>1</sup> (C–O–C), the last being the most representative. The FT-IR technique allowed us to qualitatively verify the formation of ELO from LO through the disappearance of DB signals and the appearance of bands corresponding to the epoxy rings. Moreover, it is important to mention that there is no evidence of hydrolyzed chains or interruption of ester bonds, which generally occur at 3200–3550 cm−<sup>1</sup> (hydroxyl zones) and at 1650 cm−<sup>1</sup> (carboxylic acids), respectively [37,44].

**Figure 1.** Fourier transform infrared spectroscopy (FT-IR) spectrum of (**a**) linseed oil (LO), (**b**) epoxidized linseed oil (ELO), and (**c**) carbonated epoxidized linseed oil (CELO).

The signals of representative 1H-NMR spectra for LO and ELO are shown in Figure 2, which are consistent with those reported in the literature [35,37,45]. In the 1H-NMR spectrum of LO (Figure 2a), the most characteristic signal belongs to DB (vinyl hydrogens) at 5.25–5.45 ppm (L), which overlaps with the central hydrogen of glycerol (K). The rest of the hydrogen signals are found at 4.10–4.34 ppm (J, glycerol methylene), 2.74–2.80 ppm

(G, internal allylic), 2.27–2.35 ppm (F, α-carboxylic), 1.97–2.13 ppm (E, external allylic), 1.55–1.68 ppm (D, β-carboxylic), 1.23–1.40 ppm (C, aliphatic methyl), 0.94–1.01 ppm (B, fatty acid methylene) and 0.84–0.92 ppm (A, methyl). The 1H-NMR spectrum of ELO (Figure 2b) is similar to that of LO; hence, the presence of an oxirane ring was corroborated mainly with the appearance of the signal in the region of 2.86–3.23 ppm (I, –CHOCH–). Furthermore, a significant decrease and signal shift corresponding to the unreacted vinyl hydrogens (5.6 ppm region) is observed. The epoxide group content present in ELO, as well as the molecular weight of oil, was determined by 1H-NMR, giving values of 1007.3 g/mol and 6.26 epoxide groups, respectively [35].

**Figure 2.** Proton nuclear magnetic resonance (1H-NMR) spectra of (**a**) LO and (**b**) ELO.

*3.2. Characterization of Carbonated Epoxidized Linseed Oil (CELO)*

CELO was obtained by reacting the oxirane groups of ELO obtained above with CO2 in the presence of TBAB. The reaction mechanism that has been adopted involves a nucleophilic attack by the bromide ion on the epoxy ring. Alkoxide generation is promoted, which in turn carries out a nucleophilic attack on CO2. Finally, the oxyanion displaces the bromine, generating the corresponding CC5 (Scheme 1) [8,18].

The structural characterization of CELO was carried out using FT-IR, 13C-NMR, and 1H-NMR spectroscopic techniques that are consistent with the results reported in the literature [6,17,20,31,46–48]. To corroborate the formation of the carbonate group, the infrared spectra of ELSO and CELO were compared. In the CELO spectrum (Figure 1c), the carbonate group's vibration signals appear at 1798 and 1045 cm−1. The most notable signal is at 1798 cm−1, which corresponds to the carbonyl of the cyclic carbonate (C=O), while the signal at 1045 cm−<sup>1</sup> is assigned to the C-O bond. Also, the disappearance of epoxide group signals (823 cm<sup>−</sup>1) is observed.

As additional qualitative evidence, 13C-NMR spectra were useful to identify the carbonate group's presence in the CELO structure, as reported by Doley and Dolui, Zheng et al. and Liu and Lu [10,47,48]. The 13C-NMR spectrum of CELO (Figure 3b) showed the following carbon signals: end methyl (9.7–13.3 ppm), triglyceride chain methylene (20.2–33.4 ppm), remaining epoxy rings (53.2–57.4 ppm), central methine (61.5 ppm), methylene of glycerol (68.2 ppm), and carbonyls of fatty acids (172.5–173.0 ppm). Carbon signals at 79.2 ppm (C–O), 81.4 ppm (C–O), and 154.1 ppm (C=O) confirm the formation of CELO.

**Figure 3.** Carbon-13 nuclear magnetic resonance (13C-NMR) spectra of (**a**) ELO and (**b**) CELO.

Figure 4 shows a representative 1H-NMR spectrum of CELO, which was useful for qualitative characterization of CELO and quantifying the carbonation reaction progress. In the 1H-NMR spectrum of CELO (Figure 4), the signal corresponding to the hydrogen of the central carbon of glycerol at 5.25 ppm (K) is observed. In the 2.85–3.28 ppm (I) region, the hydrogen is associated with the unreacted epoxide groups. To corroborate CELO formation, signals associated with cyclic carbonate hydrogens appear in the interval of 4.45–5.10 ppm (M).

The 1H-NMR technique is widely used to characterize materials' chemical structure, both qualitatively and quantitatively [49]. 1H-NMR has been shown to be an effective technique, compared to traditional methods, to determine the chemical composition of triglycerides and their derivatives and reaction parameters [36,50]. The performance of the carbonation reaction was evaluated through conversion (% C), carbonation, or yield (% Y), and selectivity (% S) parameters. The numbers of epoxide and carbonate groups present in CELO were calculated from the integrals of the 1H-NMR spectrum (Figure 4). The central carbon (K) signal was taken as the spectrum normalization factor. The number of epoxide groups (Em) was calculated with the signals corresponding to epoxy rings (I) (Equation (1)). Similarly, through Equation (2), the number of carbonate groups (Cm) was calculated from the associated signals (M). Regarding the reaction parameters, the conversion, carbonation, and selectivity values were obtained using Equations (3)–(5) [8,29]. The runs were performed in triplicate and the results plotted are the average value of each test:

$$\mathbf{E\_m} = \frac{\mathbf{I}}{2\mathbf{K}}\tag{1}$$

$$\mathbf{C\_m} = \frac{\mathbf{M}}{2\mathbf{K}}\tag{2}$$

$$\% \text{C} = \left(\frac{\text{E}\_{\text{mi}} - \text{E}\_{\text{mf}}}{\text{E}\_{\text{mi}}}\right) \times 100 \tag{3}$$

$$\text{V}\,\%\text{Y} = \left(\frac{\text{C}\_{\text{m}}}{\text{E}\_{\text{mi}}}\right) \times 100\tag{4}$$

**Figure 4.** Values obtained by 1H-NMR in the carbonation reaction at 90 ◦C, 60 psi, 3.5% TBAB, and 68 h.

#### 3.2.1. Effect of Temperature

To determine the optimal conditions for obtaining CELO that contains the highest degree of carbonation (high content of carbonate groups) and selectivity (least amount of byproducts), the carbonation reaction was monitored based on temperature, CO2 pressure, and the amount of catalyst. In general, in the CELO structure, the epoxide groups show steric hindrance because the epoxides' positions are located in the middle of the molecular chain (25). To obtain high levels of conversion, it is necessary to use a slightly elevated temperature (Scheme 2), avoiding reaching the decomposition temperature of the catalyst (determined as 150–190 ◦C by Doll et al. [6]). Therefore, temperature is a critical parameter in the carbonation reaction.

Figure 5 shows the effect of reaction temperatures (90, 100, 110, and 120 ◦C) on the conversion, carbonation, and selectivity degree while keeping the pressure (90 psi), catalyst concentration (2.5% TBAB), and time (24 h) constant. At 90 ◦C, conversion reaches 55%. When the temperature increases from 90 to 120 ◦C, the conversion increases from 55 to 85.3%. Similar behavior is observed in the carbonation percentage. By increasing the temperature to 120 ◦C, carbonate formation increases slightly from 51 to 57.1%. It is observed that the growth rate in carbonation is slower than that in conversion. The opposite behavior is observed in selectivity, whereby increasing the temperature to 120 ◦C, the selectivity decreases from 92.7 to 66.9%. Through FT-IR, it was detected that as the reaction temperature increased from 90 to 120 ◦C, the signal corresponding to the hydroxyl zone also increased (3200–3600 cm−1). Usually, in chemical reactions, it is desirable to have high selectivity values. Low values indicate that some of the epoxide groups are not being converted to the desired product, i.e., they are not being formed to cyclic carbonates, which indicates that some alternative reactions.

**Figure 5.** Effect of temperature on the conversion, carbonation, and selectivity degree.

Although the carbonation reaction was carried out below the decomposition temperature of TBAB, evidence of the formation of byproducts at higher temperatures is presented in Figure 6. To verify this statement, the interaction between TBAB and CO2 was studied in the absence of ELO at 120 ◦C and 90 psi for 48 h. TBAB decomposition is similar to the Hoffman reaction (Scheme 3). However, due to the presence of CO2, the first step was the addition of the bromide to the CO2 molecule to form the carboxylate ion, which extracts beta hydrogen from a butyl substitute of tetrabutylammonium, decomposing into CO2 and hydrogen bromide, and generating butene and tributylamine as final products (Scheme 4). Figure 7 shows the 13C-NMR spectra of the catalyst before (pure TBAB) and after being subjected to heat treatment in the presence of CO2 (Figure 7b). The carbon signals of pure TBAB (Figure 7a) are found at 12.2 ppm (*d*, terminal methyl), 18.5 ppm (*e*, methylene), 22.8 ppm (*f*, methylene), and 57.8 ppm (*g*, quaternary amine). In Figure 7b, the formation of tributylamine is confirmed by the appearance of signals at 11.5 ppm (*h*, –CH3), 18.2 ppm (*i*, –CH2–), 23.2 ppm (*j*, –CH2–), and 50.4 ppm (*k*, tertiary amine). Because the spectrum was only obtained from the solid sample, the presence of butene and carboxylate was determined indirectly. However, it is necessary to continue with the byproducts characterization.

**Figure 6.** FT-IR spectrum of CELO at different temperatures.

**Scheme 4.** Byproducts resulting from the interaction between TBAB and CO2 at 120 ◦C.

**Figure 7.** 13C-NMR spectrum of (**a**) TBAB, and (**b**) trimethylamine as a byproduct or the TBAB+CO2 reaction in absence of ELO at 120 ◦C, 90 psi, and 48 h.

The formation of carboxylate ions from metal catalysts and CO2 at elevated pressures is well known [51]. The presence of another additional reaction mechanism to the one reported (Scheme 1) is possible at elevated temperatures and in the presence of the carboxylate ion. Hence, the interaction between epoxide and CO2 may also depend on the reaction temperature. Similar results were reported by Kiara et al., 1993, who proposed the influence of CO2 pressure on the reaction mechanisms [52]. At a lower temperature, the bromide ion generates the alkoxide that is subsequently transformed into the corresponding cycle (species 5a). However, when the carbonation reaction is carried out at a higher temperature, there may also be the formation of other species (oligomers) that negatively impact selectivity. A possible mechanism is shown in Scheme 5 (pathway 1). HBr promotes the formation of carboxylate ions, which by nucleophilic addition, interact with epoxide cations to form carboxylates. However, this species is not stable, so it reacts with another carboxylate molecule to form the final macrocyclic dimers (species 5b) or oligomer. However, analyzing closely the spectra in Figure 6 for runs 1 through 4 (Table 2), new bands corresponding to O–H bonds in 2200–3600 cm−<sup>1</sup> region were observed for experiments at 100, 110 and 120 ◦C, as well as a broadening of the band from 990–1070 cm−<sup>1</sup> due to the overlap of new C–O bonds. This broadening was determined from the area under the curve in the 990–1122 cm−<sup>1</sup> region, from the normalized spectrum. These facts suggest that formation of hydroxyl groups. One explanation is that HBr generated at temperatures above 100 ◦ C carried out an epoxy ring opening reaction by the mechanism shown in Scheme 5 (pathway 2) forming a bromohydrin, instead of a CC5 (a) or a dicarbonate (c). Table 3 summarizes signals observed in the 13C– and 1H-NMR spectra for run 4 (120 ◦C), which would correspond to a bromohydrin (c). These assignments

are supported by the work of Eren et al. [53]. These signals confirm that an increase in temperature enables the epoxy rings to react in alternative ways, decreasing the selectivity towards the carbonation reaction.

**Scheme 5.** Possible reaction pathways between CO2 and epoxides at low and high temperatures.

**Table 3.** Signals that Corroborate the Formation of a Bromohydrin as a Side Product of the Carbonation Reaction of ELO (Run 4, 120 ◦C).


### 3.2.2. Effect of Pressure and Catalyst Concentration

Like temperature, pressure is a key parameter in the carbonation reaction, not only because it increases the solubility of CO2 in oils but it also favors the carbonation reaction since it is the volume reduction reaction and it promotes the interaction between the oil and the catalyst as observed and studied by other authors [28,32,54,55]. Their studies show that the solubility of a gas (CO2) in the liquid phase (ELO) is a function of the pressure of the system and is related by the Henry coefficient. In turn, this coefficient

depends on both the properties of the liquid and the temperature [55]. Results published by Zhang et al. [32] have shown that the solubility of CO2 increases with pressure, which can favor the conversion of epoxidized vegetable oil (opening of the oxirane ring) due to the higher concentration of CO2 in the system. Regarding the glyceride series, they showed that the difference in polarity and molecular weight of the tested molecules affects the solubility in CO2. Monoglyceride is logically more soluble in CO2 than diglyceride and triglyceride, due to its lower number of carbons, evidence that the solubility of these derivates in carbon dioxide is governed mainly by their molecular weight. A terminal epoxy fatty acid diester was found to be more soluble and more reactive in CO2 than an internal epoxy fatty acid diester [56,57]. The reactivity centers, i.e., epoxide group of epoxidized fatty esters is more accessible than the ones of epoxidized vegetable oils. Considering temperature and pressure, the higher they are, usually the greater the solubility of the epoxy ester in CO2. They explain this effect by the fact that the increase in temperature leads to a decrease in the cohesiveness of the oil and, therefore, to an increase in its solubility in CO2. Furthermore, due to the high molecular weight, the steric hindrance of ELO structure with limitedly accessible internal epoxy groups; it is believed that pressure and temperature can play an important role in increasing yield and selectivity in this particular epoxidized triglyceride. Therefore, in order to increase the degree of carbonation of ELO, it was established to evaluate the carbonation reaction at 90 ◦C, 3.5% TBAB, 68 h, and under different moderate pressures of CO2 (60, 90 and 120 psi). The results obtained from the reaction parameters (% V, % C and% S) are presented in Figure 8a. A considerable increase in conversion from 60 to 120 psi is observed (74% and 88.3%, respectively). Cyclic carbonate formation is also enhanced with increasing CO2 pressure, reaching a maximum carbonation degree of 77.2% at 120 psi. However, the growth rate of carbonation is less than the conversion rate. Hence, the selectivity of the reaction decreases from 90.1 to 87.2% with increasing CO2 pressure. According to Ochiai and Endo, the mechanism in reactions between oxirane rings and CO2 depends on the system pressure [46]. Therefore, at high pressures, the generation of carboxylate ions is favored, reducing the reaction's selectivity.

**Figure 8.** (**a**) Effect of pressure at 90 ◦C, 3.5% TBAB, and 68 h. (**b**) Effect of catalyst concentration at 120 psi, 90 ◦C, and 86 h.

Finally, the concentration of catalyst in the reaction parameters was studied. The experiment was performed at 90 ◦C, 120 psi, and 86 h. The results are shown in Figure 8b. The conversion degrees achieved were 80.8, 94.1, and 96.1% at 2.5, 3.5, and 5.0% TBAB, respectively. The increase in conversion using 2.5 to 3.5% TBAB is greater than 5.0% because as the reaction progresses, the material's viscosity is higher, decreasing the absorption of CO2. The carbonation degree improves from 70.9 to 95.2% using 2.5–5.0% TBAB, indicating a better interaction between the catalyst, CO2, and ELO at 5.0%. Similar behavior occurs with selectivity because as the concentration of TBAB increases, selectivity increases, obtaining maximum values of 99%.

The systematic study of the reaction variables made it possible to find the optimal conditions to obtain CELO with a high content of cyclic carbonates and a low content of

residual epoxide groups. The temperature was found to be the most critical variable in the reaction. The selectivity value was lower with increasing reaction temperature. In the literature (Table 1), the carbonation reaction was carried out at 120–140 ◦C to accelerate it. However, they fail to achieve high carbonation percentages. Therefore, in our work, the temperature of 90 ◦C was determined as the optimum temperature value because at this temperature, the highest selectivity percentage was obtained. Subsequently, the CO2 pressure was modified, and when the pressure increased, both the conversion and carbonation percentage increased. Although the selectivity decreases slightly as the pressure increases, the pressure at 120 psi has a higher reaction rate than that at 60 psi. Hence, 120 psi was considered acceptable and was selected as the optimum pressure value. Finally, at 90 ◦C and 120 psi, the catalyst concentration was changed from 2.5 to 5.0%. It was observed that as the catalyst concentration increased, both the conversion, carbonation, and selectivity percentage increased considerably. Therefore, 5% was set as the optimal catalyst concentration value because values of 96.1, 95.2, and 99% conversion, carbonation, and selectivity, respectively, were obtained. Therefore, 5% was set as the optimal catalyst concentration value because, at this concentration, the highest conversion (96.1%), carbonation (95.2%), and selectivity (99%) values were obtained. Although conversion values equal to 100% have been obtained in the literature, carbonation values above 77% and selectivity values above 89% have not been reported until now. Obtaining vegetable oils with a high carbonate content is essential to improve the quality of subsequent polymers such as NIPU with the required properties.

### **4. Conclusions**

The carbonation reaction between epoxidized linseed oil and CO2 in the presence of TBAB as a catalyst was studied by modifying the parameters of temperature (90–120 ◦C), CO2 pressure (60–120 psi), and amount of catalyst (2.5–5.0% TBAB). The presence of CELO was corroborated using FT-IR, 13C-NMR, and 1H-NMR techniques by the appearance of signals corresponding to cyclic carbonates and a decrease in signals belonging to epoxide groups. The different reaction conditions have a significant impact on the conversion, carbonation, and selectivity degree. It was observed that by increasing the reaction temperature from 90 to 120 ◦C, both the conversion rate and the carbonation percentage are favored; however, selectivity is negatively impacted. An additional reaction mechanism is promoted, based on the generation of carboxyl ions as the result of the interaction between CO2 and TBAB. Pressure exhibits similar behavior, and high CO2 pressures (120 psi) also favor side reactions. Contrary to the catalyst concentration, selectivity is favored by increasing the percentage of TBAB in the system. From quantification of the 1H-NMR spectra, it was established that at 90 ◦C, 60 psi, and 5.0% catalyst, it is possible to obtain the highest values reported up to the moment of conversion equal to 96%, 95% carbonation, and selectivity values of 99%.

**Author Contributions:** Conceptualization, S.H.L., D.A.G.M., and E.V.S.; validation, S.H.L. and E.V.S.; writing—original draft preparation, D.A.G.M.; writing—review and editing, S.H.L., D.A.G.M., and E.V.S.; visualization, S.H.L., and E.V.S.; supervision, E.V.S. and S.H.L.; funding acquisition, S.H.L. and E.V.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Project SIEA-UAEM No. 4965/2020CIB.

**Data Availability Statement:** The data is not yet publicly available as it is necessary to first present the dissertation paper. Once submitted, the data is made public in the Institutional Repository of the Universidad Autónoma del Estado de México (http://ri.uaemex.mx/ accessed on 1 March 2021).

**Acknowledgments:** Authors thank CONACYT for scholarship provided to the student (701376). To M.C. Nieves Zavala for the technical support in 1H-NMR measurements (Project SHL2018).

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


## *Article* **Development and Characterization of Weft-Knitted Fabrics of Naturally Occurring Polymer Fibers for Sustainable and Functional Textiles**

**Marcela Ferrándiz 1, Eduardo Fages 2, Sandra Rojas-Lema 2,\*, Juan Ivorra-Martinez 2, Jaume Gomez-Caturla <sup>2</sup> and Sergio Torres-Giner 3,\***


**Abstract:** This study focuses on the potential uses in textiles of fibers of soy protein (SP) and chitin, which are naturally occurring polymers that can be obtained from agricultural and food processing byproducts and wastes. The as-received natural fibers were first subjected to a three-step manufacturing process to develop yarns that were, thereafter, converted into fabrics by weft knitting. Different characterizations in terms of physical properties and comfort parameters were carried out on the natural fibers and compared to waste derived fibers of coir and also conventional cotton and cottonbased fibers, which are widely used in the textile industry. The evaluation of the geometry and mechanical properties revealed that both SP and chitin fibers showed similar fineness and tenacity values than cotton, whereas coir did not achieve the expected properties to develop fabrics. In relation to the moisture content, it was found that the SP fibers outperformed the other natural fibers, which could successfully avoid variations in the mechanical performance of their fabrics as well as impair the growth of microorganisms. In addition, the antimicrobial activity of the natural fibers was assessed against different bacteria and fungi that are typically found on the skin. The obtained results indicated that the fibers of chitin and also SP, being the latter functionalized with biocides during the fiber-formation process, showed a high antimicrobial activity. In particular, reductions of up to 100% and 60% were attained for the bacteria and fungi strains, respectively. Finally, textile comfort was evaluated on the weft-knitted fabrics of the chitin and SP fibers by means of thermal and tactile tests. The comfort analysis indicated that the thermal resistance of both fabrics was similar to that of cotton, whereas their air permeability was higher, particularly for chitin due to its higher fineness, which makes these natural fibers very promising for summer clothes. Both the SP and chitin fabrics also presented relatively similar values of fullness and softness than the pure cotton fabric in terms of body feeling and richness. However, the cotton/polyester fabric was the only one that achieved a good range for uses in winter-autumn cloths. Therefore, the results of this work demonstrate that non-conventional chitin and SP fibers can be considered as potential candidates to replace cotton fibers in fabrics for the textile industry due to their high comfort and improved sustainability. Furthermore, these natural fibers can also serve to develop novel functional textiles with antimicrobial properties.

**Keywords:** natural fibers; soy protein; chitin; coir; comfort; functional textiles; Circular Bioeconomy

### **1. Introduction**

Over the last several years, valorization of natural resources has become an important topic. Therefore, the idea of using agricultural and food processing by-products [1], or

**Citation:** Ferrándiz, M.; Fages, E.; Rojas-Lema, S.; Ivorra-Martinez, J.; Gomez-Caturla, J.; Torres-Giner, S. Development and Characterization of Weft-Knitted Fabrics of Naturally Occurring Polymer Fibers for Sustainable and Functional Textiles. *Polymers* **2021**, *13*, 665. https:// doi.org/10.3390/polym13040665

Academic Editor: Hong Hu

Received: 3 February 2021 Accepted: 18 February 2021 Published: 23 February 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

even wastes of the agricultural and food industries can be very interesting due to it may contribute to the development of the Circular Bioeconomy [2]. In addition, it provides the possibility of develop fiber-based products with new characteristics and applications. More recently, fibrous porous media have also been elucidated with fractal models [3,4].

In this context, an interesting natural and renewable source for textiles fibers is represented by the antimicrobial polysaccharides. Among these, chitin offers, together with cellulose, show the highest worldwide availability. Chitin is a β-(1,4)-N-acetyl-D-glucosamine found in the shells of crabs, lobsters, and shrimps, which are marine fishery by-products. It is biodegradable and can be additionally blended with cellulose or silk [5]. Furthermore, chitin and chitosan, which is obtained from chitin deacetylation, are carbohydrates that exhibit several active and bioactive properties. For instance, some novel functional applications of chitin fibers are focused on the medical sector such as bandages, sutures, etc. since it shows non-toxicity to humans [6]. Chitin also offers antimicrobial protection due to its structure prevents the growth of microorganisms such as bacteria, fungi, and viruses [7–9]. However, despite some chitin fibers have been reported in literature, chitin is not widely used at industrial scale in textile applications. In particular, chitin has a semicrystalline structure with highly extended hydrogen bonds, which cause some difficulty in its solubilization. In addition, its degree of acetylation (DA), which is normally 0.9, indicates the presence of some amino groups and it also affects its solubility. However, these groups are susceptible to modifications in order to improve this characteristic [10]. Therefore, artificially converting it to fibrous architectures has emerged as a novel approach to modulate their surface activity, structure, and molecular flexibility/rigidity for expanded textile applications. Therefore, various strategies have been adopted to fabricate fibers from chitin, including electrospinning as well as dry and wet spinning [11]. Fibers characteristics vary according to the solvent and carbohydrate contents, showing values of tenacity from ~1.59 to 3.2 g/d and elongation from ~3 to 20% [12].

Proteins represent another novel source for natural fibers derived from agricultural and industrial by-products and wastes [13]. Their properties and functionalities are greatly influenced by their amino acid sequences that initiate the self-assembly into higher order structures with unique physicochemical properties [14]. Among protein-based plants, soy proteins (SPs) are widely processed and consumed all over the world. In particular, different types of SP resins can obtained from soybean harvesting and processing as a by-product [15]. In this process, soy flour is purified to obtain soy protein powder. For that, soybean meal is first centrifuged followed by acidification and neutralization processes [16]. SPs can be classified into four categories, namely 2S, 7S, 11S, and 15S, according to their sedimentation coefficients, in which 7S and 11S surpass 80% of the total amount. β-conglycinin is the most prevalent 7S globulin and composed of three major subunits, namely α (~67 kDa), α' (~71 kDa), and β (~50 kDa), jointed through noncovalent interactions. Native glycinin (11S globulin, hexamer) is made up of six subunits, owing a molecular weight (MW) of 300–380 kDa. These subunits are either acidic or basic and form a hollow cylinder by electrostatic interaction, hydrogen, and disulfide bonds [17]. The quaternary structure of SPs is susceptible to changes of external conditions such as pH, temperature, and ionic strength. Moreover, its high β-sheet content (≥40%) provides the peptides necessary for self-assembled fibrillation [18]. During fabrication of amyloid-like fibrils, proteins dissociate and are hydrolyzed into peptides, which then assemble into protofilaments that intertwine in the final stage. This sequential process can be achieved by heating the proteins in an acid environment [19–21], where conditions of 80–95 ◦C and pH 1.6–2 are commonly used. In general, SP fibers exhibit a worm-like and periodic structure, typically with diameter of a few nm and length of several μm. SP can be molded into a wide variety of shapes, which include fibers, films, hydrogels, nanofibers, and solid parts [22,23]. Moreover, SP can be mixed with other components in order to improve their properties or achieve new ones, such as antioxidant capacity or antimicrobial activity [24]. SP fibers also present a luxurious appearance, good comfort, good chromaticity, which

makes them brilliant, and improves dyeing fastness compared to other protein-based fibers, such as silk [25].

Nevertheless, one of the main drawbacks of natural fibers is that they are prone to the biological attack of microorganisms and insects [26]. Microorganism growth depends on their chemical structure, specific surface area, thread thickness, and capacity to retain moisture, oxygen, and nutrients [27–29]. Some of the problems caused by microorganisms are odor emissions, discoloring, degradation, and risk of causing infections. In addition, microorganisms can release enzymes and other organic compounds, which can trigger degradation. Additionally, it is necessary to take into account that the microbial growth can result in a deterioration of the mechanical properties with a loss in tensile strength, elongation, and visual changes [29]. As a result, the use of antimicrobial materials has recently gained importance in a wide variety of applications related to medical clothes and implements, housing, decorative, construction as well as textiles [30]. In this regard, *Escherichia coli* (*E. coli*, gram-negative) and *Staphylococcus aureus* (*S. aureus*, gram-positive) are commonly used to test the antimicrobial properties since both are human pathogens that show high resistance to common antimicrobial agents [31,32]. Furthermore, *Candida albicans* (*C. albicans*), *Trichophyton mentagrophytes* (*T. mentagrophytes*), and *Epidermophyton stockdaleae* (*E. stockdaleae*) are among the fungi tested since they can be located on the skin as well as in mucous membranes and are responsible for systemic infections with high mortality rate [33,34]. All of these microorganism have been associated with several infections due to their growth on the surface of textile fabrics, especially in hospitals or health care institutions [35].

This study originally assesses the potential of natural fibers of chitin, SP, and coir, which can be obtained from agricultural and food wastes, to produce functional fabrics produced by weft knitting for uses in textile applications. To this end, it involves the physical characterization and antimicrobial properties of the natural fibers against the most commonly tested bacteria and fungi. Finally, the thermal and tactile comforts of the natural fiber fabrics were also determined and compared to those of cotton and cotton-based fibers. The tactile comfort was related to mechanical properties and the interaction of the clothes with the human body, while thermal comfort characterization involved thermal resistance, water vapor resistance, and air permeability properties.

### **2. Materials and Methods**

#### *2.1. Materials*

Chitin fibers were supplied by Tec Service S.r.L. (Villorba, Italy), whereas the SP fibers were provided by Swicofil (Emmen, Switzerland). According to the supplier, SP fibers are functionalized with undisclosed antimicrobial agents during the spinning process, which can restrain the growth of colon bacillus, impetico bacteria, and sporothric. Coir fibers, obtained from Swarna Trades (Kochi, India), were also tested for comparison purposes. Table 1 summarizes the main physical properties of these natural fibers indicated by their suppliers.

**Table 1.** Main physical properties of soy protein (SP), chitin, and coir fibers.


\* dry properties.

### *2.2. Yarn Processing*

The three types of natural fibers were selected for yarn processing and subsequent weaving. This process started with the opening of the fiber bales, which consists of the

extraction of the pressed fiber from its packaging by an automatic machine UNIfloc A 12 from Rieter China Textile Instruments Co., Ltd. (Changzhou, China). The second step was the separation of the flock into homogenous small flakes of fibers to facilitate their disaggregation and individualization in subsequent processes. Finally, the third step was the carded process within the opening and cleaning of the fibers. This last operation completed the individualization of the fibers to obtain a regular tape, which was folded on a coil in a cylinder card. Thereafter, the resultant folded tape was stretched, in order to obtain a wick, which was subjected to a certain twist using a ring twisting machine form MEERA Industries limited (Gujarat, India) to obtain the yarn.

### *2.3. Fabric Manufacturing*

In order to perform the comfort tests, it was necessary to obtain the fabrics from the fibers. Knitting is, after weaving, the second most frequently used method to produce fabrics and its popularity has recently grown due to the adaptability of the new fibers, high versatility, and the consumer demand for stretchable, wrinkle resistant, and snug-fitting fabrics. Knitting to shape can be realized either by weft knitting or by warp knitting. In general, the weft-knitting process is more flexible due to the needle selection capability available and the variety of structural designs possible. In this technique, the same thread feeds an entire mesh pass, progressively feeding all the needles. Therefore, during this process, fabrics of several consecutive rows of intermeshed loops were made in a horizontal way from a single yarn and each consecutive rows of loop built upon the prior loops consecutively. To this end, the small-diameter circular machine Galan Ratera (Manresa, Spain) was used with a speed of 350 rpm. The resultant fabrics are shown in Figure 1.

**Figure 1.** Weft-knitted fabrics of (**a**) soy protein (SP) (**b**) chitin fibers.

#### *2.4. Evaluation of Fiber Geometry and Mechanical Tests*

The evaluation of fiber geometry, which is related to the fiber fineness, was carried out according to the classification of Sekhri [36], as shown in Table 2.

**Table 2.** Fineness classification of fibers.


The mechanical properties of the fibers were studied according to the UNE EN ISO 5079-96 and UNE EN ISO 1973:1996 analyzing tensile strength, elongation at break, tenacity, and fineness. The UNE 40152:1984 was followed for determining the length of the fibers. All the mechanical properties were obtained by averaging five experimental values, and all of the samples were stored and tested under room conditions.

### *2.5. Moisture Content Determination*

The ASTM D2495-07 standard test method, applied for moisture determination in cotton, was used to the study moisture content by oven drying. To this end, fibers were previously conditioned at 20 ± 2 ◦C and 65 ± 2% RH for 48 h, then, they were weighed and placed for drying at 105 ± 2 ◦C in an oven for 24 h. After this time, samples were cooled in a desiccator and weighed again. The moisture content was calculated following Equation (1),

$$\text{MC} \left( \% \right) = \left[ \left( M - D \right) / M \right] \times 100 \tag{1}$$

where *MC* represents the moisture content (%), *M* is the mass of the specimen, and *D* is the oven-dry mass of the specimen. Samples were tested in triplicate.

### *2.6. Antimicrobial Tests*

To ascertain the antibacterial and antifungal activities of the fibers, these were subjected to the presence of *E. coli* (ATCC 10536) and *S. aureus* (ATCC 6538) bacteria and *C. albicans* (CECT 1394), *T. mentagrophytes* (CECT 2958), and *E. stockdaleae* (CECT 2988) fungi. All bacteria and fungi were obtained from the Leibniz Institute DSMZ-German Collection of Microorganisms and Cell Cultures (Braunschweig, Germany). The evaluation of the antimicrobial efficacy of the fibers was carried out by following the principles of the AATCC Test Method 100-2004 standard. To this end, fiber samples of approximately 1 ± 0.1 g were introduced in the broth with the bacterial or fungal inoculum. Then, it was plated out by triplicate in agar plates after serial dilutions. Finally, samples were incubated for 18 h at 37 ◦C. After this time, the bacteria colony-forming units (CFU/mL) were determined for each plate. The microbial reduction or growth inhibition was quantified based on the number of CFU/mL identified after the incubation time for each type of fiber in relation to the number of bacteria initially counted. The same procedure was performed to analyze the antifungal activity.

### *2.7. Comfort Assessment*

The comfort parameter is complex and subjective feature. It depends on both psychological and physiological perceptions of the body and the own fabric properties. It can be divided into tactile comfort and thermal comfort [37]. The comfort generated by a fabric depends, among other factors, on sensory parameters such as perception through the sense of touch, which is much more demanding and subjective than the rest. The aim of this measurements was to ascertain whether the comfort characteristics of the natural fiber fabrics could serve to replace textiles based on cotton. To this end, the KES-FB or Kawabata Evaluation System (Kato Tekko, Kyoto, Japan) was used to analyze the fabric low-stress mechanical and surface properties. It is based on different experimentation modulus: KES-FB1—automated module designed to measure the tension and shear properties of fabrics (tension energy, tension force, and stiffness and shear hysteresis), which applies opposing and parallel forces to the fabric until reach a maximum offset angle; KES-FB2 automated module to measure the flexion of a fabric or non-woven, where measures the force that a given fabric requires to be bended; KES-FB3—automated module to determine the compression properties of fabrics, in which a fabric of ~2 cm2 is subdued to a compress between two plates and increasing the pressure at a maximum value of 50 gf/cm2; and KES-FB4—automated module for measuring the surface of fabrics by determining the friction coefficient and geometric roughness with a displacement speed of the sample 1 mm/s [38]. Therefore, through the Kawabata Evaluation System, the three primary hand parameters shown in Table 3 were determined, which are habitually considered for analyzing the applications of fabrics in winter and autumn suiting [39].


**Table 3.** Primary hand parameters for fabrics following the Kawabata Evaluation System.

The results obtained of tactile comfort in terms of primary hand values (PHVs) and total hand values (THVs) of the fabrics were qualified according to the evaluations described in Tables 4 and 5, respectively [40].

**Table 4.** Primary Hand Values (PHVs) for fabrics according to the feeling grade.


**Table 5.** Total Hand Values (THVs) for fabrics.


The thermal comfort is defined as the condition in which the user feels satisfied with the thermal environment. Therefore, this concept is related to the fabric heat transmission behavior. In this analysis, three different tests were performed: Water vapor resistance, thermal resistance, and air permeability [41]. All these measurements were performed through Skin Model of Weiss Umwelttechnik (Lindenstruth, Germany), whose device allows quantifying the breathability of a given fabric and, therefore, the heat flow by varying the environmental conditions (%RH and temperature).

To compare the comfort results of the natural fabrics with conventional ones, the same tests were carried out on pure cotton and blends of cotton/polyester (35/65) and cotton/acrylic (40/60) fabrics obtained from Tela Prat S.A (Sabadell, Barcelona), which have the same stitch per weft. All the measurements were carried out at 65 ± 3% of RH at 20 ± 1 ◦C, averaging five experimental values.

### *2.8. Statistical Analysis*

Mean values and standard deviations were calculated from the experimental data obtained by analysis of variance (ANOVA) with 95% confidence interval level (*p* ≤ 0.05). For this purpose, a multiple comparison test (Tukey) was followed using the software OriginPro8 (OriginLab Corporation, Northampton, MA, USA).

### **3. Results**

#### *3.1. Geometry and Mechanical Properties of the Natural Fibers*

In the clothing industry, it is important to develop textiles with flexibility and elasticity that also offer functional properties that enjoy the growing demand for workwear and sportswear, thereby, improving the wearer's comfort and protection [42,43]. Therefore, mechanical parameters of the fibers have a great importance both in terms of performance and also during the spinning process since they influence on the machine preparation. In Table 6, one can observe the different mechanical values of the natural fibers tested herein and obtained from the test. In this regard, it is important to consider that the characteristics of the natural fibers regarding mechanical properties and chemical composition can vary considerably according to the cultivation place, environmental conditions, extraction and processing methods, the part of the plant from which the fibers are obtained (roots, stem, leaves, etc.), maturity of the plant, among others [44,45]. Fineness, also called linear density, determines the quality and price of the fibers. In general terms, fibers are not constant or regular. Therefore, if a given fiber is thick, it will be stiff, firmer, and with wrinkle resistance. As opposite, if the fiber is fine, it will present softness, flexibility, and good fall [46]. In this sense, Hari [47] reported that fineness for cotton is between 1 and 3 dtex. According to the present results, the SP fiber value was within this range. On the contrary, the chitin fiber showed a higher value, being statistically different, but it was close to the upper limit. According to the evaluation reported in previous Table 2, both natural fibers can be classified as "semi-fine" fibers. These results in fineness are promising because SP and chitin are within the same range of cotton.

**Table 6.** Mechanical properties of soy protein (SP), chitin, and coir fibers.


\* Obtained from the technical datasheet of the product; a,b Different letters in the same column indicate a significant difference among the samples (*p* < 0.05).

Fiber length is another very important feature due to it affects the processability that, in turn, also influences the physical properties. In the case of cotton fibers, their length ranges from 10 to 40 mm [48], being the most used worldwide those with a length between 20 and 35 mm [46]. Besides, it is known that the diameter of the cotton fiber, in most cases, is not more than 1/100 of the fiber length. The results indicate that the here-tested natural fibers are approximately 2 to 4 times longer than those of cotton, which is an indispensable factor to consider when choosing the spinning process and their application in textiles. In addition, another positive aspect of longer fibers is that they will require less twist than shorter fibers for yarn strengthen [47]. In this case, the SP fibers presented the highest value followed by coir and chitin, respectively.

The mechanical properties of the fibers also depend on their constitution, the amount of cellulose, and the crystallinity [49]. In this context, fibers obtained from fruits or seeds are habitually weaker than those obtained from steams of hemp or ramie. In particular, the higher tensile strength and modulus derive from their low microfibrillar angle and high cellulose content. However, it is also important to consider that the fibers strength can be more affected by their own defects than their structure [49]. In this context, SP and chitin, both natural fibers, showed similar values of elongation at break, that is, 16.21, and 17.87%, respectively. The ductility of these natural fibers is higher than that of cotton fibers, which has been reported to show an elongation-at-break value of approximately 7% [50]. For chitin, Fan et al. [7] indicated that blend fibers of chitin at 70 wt% with alginate yielded an elongation at break of 16.4%, which is close to the value attained in the present study. In addition, as it was mentioned above, the mechanical properties of the coir fibers were not analyzes due to their high brittleness, although other studies have reported values of elongation at break between 15 and 30% [51]. For instance, in the study performed by Sumi et al. [52], the elongation at break of coir fibers achieved a value of 16.67%, while Tomczak et al. [45] reported for 25-mm long fibers a value of approximately 25%.

In terms of tenacity, Jackman [53] described that the minimum value for most fiber fabrics is 22.12 cN/tex, though some fibers with lower values can be used in textiles due to they also show high elasticity and resilience. For instance, McKenna [50] reported that tenacity of cotton fiber is around 30 cN/tex, whereas Singh [54] indicated that the cotton fiber tenacity ranges between 25 to 40 cN/tex. Therefore, the previously reported values are close to the tenacity value of 32.12 cN/tex obtained herein for the SP fibers. The chitin fibers achieved a tenacity value of 24.73 cN/tex, which is similar to that reported previously by Pang et al. [55] for chitin/cellulose blended fibers, that is, 23 cN/tex. Therefore, one can consider that both SP and chitin fibers are above the minimum tenacity values required for being used in textiles. Considering that this property is related to their strength, which is the force per unit linear density necessary to break the fiber [36,56], these results make these natural fibers very promising to produce sustainable textiles.

### *3.2. Moisture Content of the Natural Fibers*

Moisture content is one of the most important characteristics to be considered as it may affect the physical properties of fabrics [57]. For instance, it has been reported that moisture absorption for cotton can produce a significant increase in flexibility, toughness, elongation, length, tensile strength, and color, besides affecting the electrical properties [58,59]. Therefore, in textiles, controlling moisture content is necessary as an inadequate level of moisture can lead to fiber breakage, which causes a decrease in the quality of the fiber during harvesting and ginning. Additionally, moisture content of the fiber may also influence the fiber mass, which is an important factor in the textile market [60]. For instance, the increase in fiber strength in cotton is related to an increase in moisture content due to the formation of hydrogen bonding, while other fibers, such as rayon, become weaker when wet [36,56]. Another consequence of water absorption is swelling, which causes the increase in fiber diameter, and consequently, the contraction in the length of twisted ropes. In addition, a low moisture content will require excessive energy at the bale press, whereas a high moisture content can produce a deterioration of the fiber quality during storage due to microorganisms' attack. Furthermore, the control of moisture is also important to facilitate cleaning. Therefore, a good control of this parameter can guarantee the adequate quality of the fibers all over the production and supplying chain [59].

In textiles, this characteristic is normally expressed as the wet percentage material. One can observe in Figure 2 that the value obtained for SP fiber was 5.8%, whereas in the case of coir and chitin fibers, these values were 9.5%, and 9.2%, respectively. These moisture contents were also compared to those of cotton because this is the most popular natural fiber, particularly in clothing [48]. According to the study performed by Higgins et al. [61], the cotton fabric moisture content is around 8%, measured at 20 ± 2 ◦C with 65 ± 2% of RH. Similar results were attained for cotton fibers and cotton fabrics, that is, 8.5% in both cases [62]. According to Netravali [49], Tomczak et al. [45], and Mohanty et al. [63], the moisture content of coir fibers can be as high as 8%, while Cunha [64] indicated that chitin fiber is relatively hydrophobic. It is important to mention that, in general, plant fibers tend to absorb more quantity of water due to the presence of hydroxyl (-OH) groups in cellulose [48]. Therefore, the values reported, herein, agree with previous results and indicate that SP fiber absorbs less water amounts than the other natural fibers, including cotton. This is a positive factor since it favors its processability, cleaning, and mechanical performance stability. Besides, it is also important to note that the water content of SP fibers is higher than that of synthetic fibers based on polyesters, such as polyethylene terephthalate (PET), which shows a value of approximately 0.4% at 65% RH. However, excessively low moisture may cause static electricity due to friction [58] and it also produces dust and dirt attraction. For these reasons, synthetic fibers are habitually mixed with other hydrophilic material fibers [47].

**Figure 2.** Percentage of moisture content (%) of soy protein (SP), chitin, and coir fibers. a,b Different letters indicate a significant difference among the samples (*p* < 0.05).

### *3.3. Antimicrobial Activity of the Natural Fibers*

In the next step, the antimicrobial activity of the fibers was assessed against the *S. aureus* and *E. coli* bacteria and *T. mentagrophytes*, *E. stockdaleae*, and *C. albicans* fungi. The antimicrobial results of the natural fibers are gathered in Figure 3. One can observe in Figure 3a that chitin fibers showed a high antibacterial activity, especially against *E. coli*. In particular, after 18 h of contact, a reduction of 100% and 89.3% in the CFUs/mL of *S. aureus* and *E. coli* was respectively achieved. In this regard, Cheng et al. [28], reported that chitosan, a derivative from chitin, yielded a bacterial reduction for *S. aureus* of 98.57% and 98.59% with contact times of 30, and 50 min, respectively. In addition, for the same contact times, a reduction of 60.87% and 54.35% was observed for *E. coli*. Similarly, Qin et al. [65] developed chitosan fibers with silver sodium hydrogen zirconium phosphate by spinning solution, yielding a reduction of 98.60% in the growth of *S. aureus*. In another study, solution-spun blend fibers of chitin and cellulose were developed by Pang et al. [55]. In the fiber samples containing 6.46% of chitin, a reduction of 65.12% and 55.76% was obtained for *S. aureus* and *E. coli* growth. A high antibacterial activity was also observed by Cheah et al. [66], who developed a chitosan modified membrane with a quaternary amine (P-CS-GTMAC-A), having a reduction value of 76.7% for *E. coli*. The antibacterial activity of chitin and its derivatives involves electrostatic attraction. In the case of chitosan, for example, it has a positive charge that attracts bacterial cells charged negatively that produces the rupture of the bacterial cells and also causes the leak of the intracellular components [66,67]. In relation to the antifungal activity of chitin, shown in Figure 3b, one can observe that a reduction in the *T. mentagrophytes*, *E. stockdaleae*, and *C. albicans* strains of 88.8%, 79.1%, and 88.5% was respectively achieved after 18 h of contact. In the case of *C. albicans*, Qin et al. [65] observed a similar reduction trend by using chitosan fibers, reporting a reduction of 78.62% for this microorganism. Moreover, Surdu et al. [68] also analyzed the effect on the growth of *C. albicans* of cotton fabrics with different contents of chitosan, showing that the antimicrobial effect increased with the quantity of chitosan. In particular, for samples with 5 g/L of chitosan, the antimicrobial efficiency was nearly 83%, yielding an increase of 45% regarding the samples with the lowest chitosan value, that is, 1 g/L. Authors also analyzed the growth of other fungi of the Trichophyton family, that is, *T. interdigitale*, observing a reduction in the microbial growth of only 27% for cotton samples with 5 g/L of chitosan. This value is relatively lower compared with the one obtained in the present study, which can be related to the fact that chitosan was applied in the form of coating in the former study.

**Figure 3.** Antibacterial and antifungal activities in terms of colony-forming units (CFU/mL) of chitin fibers (**a**,**b**), soy protein (SP) fibers (**c**,**d**), and coir fibers (**e**,**f**). a–c Different letters in the same period for different samples indicate a significant difference (*p* < 0.05). A,B Different letter of the same sample in different periods indicate a significant difference (*p* < 0.05).

In relation to the SP fiber, its antibacterial activity against *S. aureus* and *E. coli* is shown in Figure 3c. Similar to the chitin fibers, at the beginning of the process, the bacterial growth was high, especially for *E. coli*. However, a reduction of 100% of the bacterial growth was observed after 18 h of contact. It is worth noting that SP alone has no antimicrobial effect and the present results differ from other studies regarding the use of SP fibers. For instance, soy protein isolated (SPI) film did not presented a decrease of the antimicrobial activity against Gram-positive *S. aureus* after 24 h of incubation [69]. In the same way, Xiao et al. [70] indicated that SPI did not presented antibacterial activity against both *S. aureus* and *E. coli*. On the other hand, in the study performed by Raeisi, et al. [71] it was observed for SPI/gelatin blends a slight decrease in the percentages of colonies (CFU) of 13% against *S. aureus*, while no colonies decrease was observed against *E. coli* after 24 h of incubation. Therefore, the antimicrobial performance of the SP fibers, reported herein, can be ascribed to the antimicrobial compounds added during the spinning process. These biocidal agents are mainly based on phenolics or acidic and bacteriocin since the functional groups present in the side chains of SP are known to alter their retention [72]. Some other examples of biocides added to the textiles include triclosan, silver, and quaternary ammonium compounds [73–76]. The mechanism used by antimicrobial textiles varies

depending on their type, but in most cases, the main acting-mechanisms include preventing cell reproduction, damage on cell walls, protein denaturation, and enzyme blocking [77,78]. In relation to the antifungal activity of the SP fibers, which can be seen in Figure 3d, *T. mentagrophytes* and *E. stockdaleae* showed more than 1200 CFU/mL at the beginning of the test but both fungi were fully eliminated after 18 h of contact. In case of *C. albicans*, this reduction was of 63.3%, which still represents a good antifungal activity because it is above 50%. In comparison to other materials, Mishra et al. [79] reported the antibacterial activity of bamboo fibers, which presented lower antibacterial activities against *S. aureus* and *E. coli* with reduction values of 63.08%, and 68.44%, respectively.

As opposed to the chitin and SP fibers, in the case of coir, one can observe that the number of microorganisms increased after 18 h of contact with the fibers. In particular, in Figure 3e one can observe that the number of CFU/mL at the beginning of the test was under 10,000 CFU/mL for both bacteria. However, it increased by approximately 3 and 7 times for *S. aureus* and *E. coli*, respectively. This behavior agrees with the results of the study performed by Chandy et al. [80], in which coir fibers did not present any inhibition zone against *E. coli*. These results mean that the fibers do not show any antibacterial activity. Similar results can be observed in Figure 3f for fungi, showing no evidence of antifungal activity after 18 h of contact. On the contrary, an increase in the number of CFU/mL was observed for all the fungi tested. In particular, the growth of *T. mentagrophytes*, *E. stockdaleae*, and *C. albicans* increased by 16.6%, 21.35%, and 66.6%, respectively. In this regard, Sumi et al. [52] reported that coir fibers do not present good antimicrobial properties against the *Aspergillus niger* fungus.

Therefore, one can conclude that both chitin and SP fibers present a high antibacterial activity against these two types of microorganisms. Additionally, the high increase in the number of bacteria and fungi observed for the coir fibers can be related to its high moisture content, which represents a more favorable medium for the development of microorganisms. However, it is also worth mentioning that the test conditions applied for all of the three types of fibers were the same, taking into account that microorganism growth also depends on temperature, pH, nutrients, oxygen and moisture content, etc. [81].

#### *3.4. Comfort of the Natural Fiber Fabrics*

Thermal comfort responds to the thermal environment that, in turns, depends on the temperature of the air and the room, the speed of the air and its humidity as well as on the type of clothing and the activity that is carried out. Therefore, water vapor resistance in fabrics is an important parameter due to the necessity that water vapor on the skin can pass through their pores, while an uncomfortable sensation can be produced if water vapor remains on the skin. In terms of thermal resistance, this parameter will depend on the type of fabric, its thickness, and bulk density, among others, whereas heat transmission involves the three different ways of transfer, that is, convection, conduction, and radiation. Furthermore, air permeability is related to the initial warm or cool feeling when the user wears clothes. If the airflow value is high, it means that the feeling of warm or cool will be more intense [82].

Table 7 shows the results of the thermal comfort analysis obtained through the skin model performed on the chitin and SP fabrics. These values were compared with fabrics of commercial cotton and blends of cotton with polyester and acrylic polymer to ascertain their application in textiles. Cotton is also a natural fiber, but it currently faces some environmental issues derived from the use of pesticides and insecticides to guarantee the good growth of the plants [83]. Furthermore, the coir fabric was discarded because as shown above, it did not achieve the expected properties.


**Table 7.** The results obtained through skin model for fabrics of soy protein (SP), chitin, cotton, cotton-based fibers.

a–c Different letters in the same column indicate a significant difference among the samples (*p* < 0.05).

One can observe that the water vapor resistance of the SP fabric was 5.01 m2·Pa/W, which was slightly but still significantly higher than that of cotton, that is, 4.82 m2·Pa/W, and also than the combinations of cotton with polyester and acrylic polymer, that is, 4.85, and 4.44 m2·Pa/W, respectively. In the case of the chitin fabric, the water vapor resistance value was 4.38 m2·Pa/W. These results are related to the moisture absorption capacity of each natural fiber, which agree with the moisture content shown in previous Figure 2, and also to its lower air permeability [82]. The SP fibers presented the lowest water absorption and showed the highest water vapor resistance. Moreover, according to the results obtained in this study, it can be concluded that the chitin fabric behaves similarly to that of cotton/acrylic, while the SP fabric would be more similar to that of neat cotton. These results indicate that the fabrics studied herein, being in a range of 4–5 m2·Pa/W, would have good breathability to water vapor. In comparison with previous works, Mishra et al. [79] reported that the water vapor resistance for cotton fabrics was 3.82 m2·Pa/W, being this value lower than that of the chitin and SP fabrics obtained in this study. This result can be explained by the fact that the previous cotton fabrics showed a fineness of 1.23 dtex, which is also lower than the present natural fibers.

Regarding the thermal resistance of the fabrics, it can be observed that the SP one also presented the highest value, that is, 0.0463 m2·K/W. In the case of the chitin fabric, it showed a value of 0.0426 m2·K/W. However, it is important to mention that the thermal resistance of both fabrics was similar to those of cotton and cotton/polyester, which did not present significant differences, whereas the cotton-based fabric containing the acrylic polymer presented the lowest value. Thermal resistance is a very important parameter for fabrics intended for underwear applications, for example, in which cotton is mostly used. Therefore, in view of these results, it can be indicated that both the SP and chitin fabrics can be suitable for this type of application. For other types of textiles where the human body protection becomes more important due to climatic fluctuation, the SP fabric is the only one expected to provide the required properties.

Finally, the air permeability is the measure of air flow passed through a given area of a fabric, which is widely used in the textile industry for outdoor garment manufacturers to describe their functional performance. One can observe that the chitin fabric presented higher permeability than the pure cotton, showing a value of 2716 mm/s. In particular, this value was nearly 2 times higher than that of the cotton fabric, which was 1638 mm/s. Meanwhile, the SP fiber air permeability did not present a significant difference versus the cotton and cotton/polyester fabrics. The higher values attained herein can be related to the lower hairiness of these natural fibers due to they presented longer fiber lengths [82]. In the case of chitin, the high value of air permeability can be additionally related to high fineness of this natural fiber, which was 3.25 dtex, by which more air could penetrate through the fabric [41]. Therefore, considering the values obtained, herein, one can conclude that fibers of chitin and SP are more suitable to produce fabrics for summer clothing because in these fabrics the pass of the air is more favored.

Other important characteristic of fabrics is tactile sensation, which is crucial when determining the degree of comfort. The fabric hand properties allow knowing the performance of the material in contact with the skin, which depends on different factors such as fiber type, fabric structure, and, in some cases, the fabric finishing. For fabrics, there

are two types of performance analysis. The first one is called utility performance and it is related to strength, color, and shrinkage, among others. The second one is quality performance, which is related to the appearance and comfort. It is important to consider that the second analysis is more difficult to conduct than the first one [84]. The assessment used to carry out the quality performance analysis is hand evaluation, being considered as a subjective method and known as Kawabata. Through this method the three PHVs, namely Koshi (stiffness), Numeri (smoothness), and Fukurami (fullness and softness), and also THVs were calculated in order to characterize the different fabrics. PHV is normally rated on a scale from 0 (weak) to 10 (strong), whereas THV is rated from 0 (not useful) to 5 (excellent) [85]. In Table 8 the results of the tactile comfort studied using Kawabata Evaluation System are gathered.

**Table 8.** Results obtained through Kawabata of the primary and total hand values (PHVs and THVs) of men's winter suit fabrics of soy protein (SP), chitin, cotton, and cotton-based fibers.


a–e Different letters in the same column indicate a significant difference among the samples (*p* < 0.05).

The fabric PHV concept covers the stiffness, fullness and softness of the fabric, and smoothness, which allows to quantify comfort based on the determination of the mechanical properties [86]. As seen in the Table 8, chitin presented values without significative differences, regarding cotton, and cotton/acrylic in Koshi parameter, which are related to the stiffness/elasticity when folded. In general, all samples values were ranged between 7.5 and 8.5, which indicates that they were high values. According to the reference used by Kawabata et al. [39], it can be considered that all of the samples presented a high-quality suiting for winter-autumn because their values are within the range of 4.5 ≤ HV ≤ 6.5. In the case of Fukurami, which indicates the fullness and softness in terms of body feeling and richness, it can be observed that both the SP and chitin fabrics showed relatively similar values than pure cotton. Nevertheless, it is important to consider that, within the current samples, the only value that was in the range of 5 ≤ HV ≤ 8 of high-quality suiting for winter-autumn was the cotton/polyester blend. Regarding Numeri, which determines the smoothness and it is ascribed to a smooth, flat, soft touch fabric, both chitin and SP fabrics showed lower values, that is, 4.06 and 6.15, respectively, than the pure cotton, which presented the highest value, that is, 8.56. However, the values attained for chitin fibers is in the range of 6 ≤ HV ≤ 8, which is considered as high-quality suiting for winter-autumn. Furthermore, the THV was also determined since this parameter is used to indicate the quality of the fabric for a specified end use and it gives a general idea of a fabric hand [87,88]. Therefore, the higher this value, the better the fabric touch and the more comfortable is the textile. The neat cotton fabric showed the highest value, that is, a THV of 3.72. As indicated in the table, it is considered as good, being relative similar to that reported by Behera [82], that is, 3.01. While, chitin presented no significant difference compared to cotton.

The results also indicated that the cotton/polyester fabric presents a good value, whereas in the case of SP and chitin their values are in the range that corresponds to average. The fabric with the lowest THV was attained for the cotton/acrylic combination, with a value of 2.72, although it could be still considered as average for this type of application. However, it must be necessary a THV ≥ 4 for winter-autumn and a THV ≥ 3.5 for midsummer, according to the criteria for an ideal fabric [39]. Based on this evaluation,

only pure cotton and the blend of cotton/polyester fulfill this requirement for midsummer with values of 3.72, and 3.60, respectively. As reported by Sun et al. [89], the PHVs for Koshi, Numeri, and Fukurami and THV were 4.78, 0.05, 3.04, and 2.03, respectively. Therefore, it can be said that the stiffness and softness values were within the good range, however, the smoothness value was low. Thus, THV was also low and the fabric can be considered weak for winter-autumn clothes. In the study carried out by Verdu et al. [90], the behavior of the cotton/polyester blend in proportions of 35/65 showed values of PHV for Koshi, Numeri, and Fukurami and THV of 5.8, 4.6, 4.4, and 3.1, respectively. All comfort values were within the range of average, which indicates that the fabric can be potentially applied for winter clothes, while the THV parameter indicated that the fabric has good quality. From the above, one can consider that the SP fabric and, more importantly, the chitin fabric presented Koshi (stiffness), Fukurami (fullness and softness), and total hand values (THV) close to those of cotton fabrics, which opens up their potential applications in sustainable textiles.

### **4. Conclusions**

Over the last years, new chemical and synthetic fibers have been developed to fulfill different technical requirements of the textile industry. These fibers result from chemical processes applied to natural fibers or, more habitually, derive from synthetic polymers. However, synthetic fibers show several sustainability issues due to the polymers used for their manufacturing derives from petroleum, which is not a renewable source, and the resultant textiles are not biodegradable or compostable, that is, they do not disintegrate in natural environments or in controlled compost conditions, respectively. In this regard, the valorization of wastes or processing by-products of the agricultural and food industries can potentially increase the economic value of natural fibers and also provide enhanced environmental benefits that improve the overall sustainability of the textiles. The present study reported on the use of chitin and SP fibers as an environmentally friendly alternative to chemical and synthetic fibers, whereas fabrics of both fibers were also considered as potential candidates to replace cotton. In terms of mechanical properties, both natural fibers presented values similar to cotton for fineness and tenacity. Furthermore, the chitin and SP fibers were longer than common cotton fibers, which is something desirable in textiles since their yarns will be less prone to twist when knitted. Moreover, in the case of SP, these fibers showed the lowest water absorption value, which is a positive aspect to improve the mechanical stability and reduce microbial growth. The antimicrobial tests also demonstrated that the chitin and SP fibers are very promising to avoid the growth of bacteria and fungi. In the case of chitin, its antimicrobial properties are ascribed to the interaction between positively charged biopolymer molecules and negatively charged microbial cell membranes. For SP, these fibers were functionalized by the manufacturer with microbial resistant elements integrated in the fiber molecule chains during the spinning process. The thermal comfort analysis also revealed that thermal resistance of the chitin and SP fibers had similar values than cotton whereas, regarding air permeability, both fabrics showed higher values than the cotton fabric. The lowest value was observed for chitin due to its higher fineness, indicating that these novel fabrics allow the air pass more easily through the fabric and, thus, can be potentially considered for summer clothes. For water vapor, the SP fabric presented higher resistance than chitin, which is related to the low moisture content of these fibers. In terms of PHV, the stiffness for the SP and chitin fabrics turned out to be high, being more than the average, however softness and fullness had similar values to that of cotton. Alternatively, smoothness presented more variability among the tested fabrics, showing cotton the highest value. In addition, the THV parameter showed adequate values for SP and chitin, being in the range of average, while the cotton/polyester fabric showed the only value that corresponded to the good range to meet the requirements for winter-autumn cloths.

**Author Contributions:** Conceptualization, S.T.-G.; methodology, S.R.-L. and J.I.-M.; validation, J.I.-M. and S.T.-G.; formal analysis, S.R.-L., J.G.-C. and J.I.-M.; investigation, S.R.-L., M.F. and J.I.-M.; data curation, E.F., J.G.-C., J.I.-M. and S.R.-L.; writing—original draft preparation S.R.-L.; writing—review

and editing, S.T.-G.; visualization, S.R.-L.; funding acquisition, S.T.-G. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research work was funded by the Spanish Ministry of Science and Innovation (MICI) project number MAT2017-84909-C2-2-R.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data is contained within the article.

**Acknowledgments:** S.R-L is a recipient of a Santiago Grisolia grant from Generalitat Valenciana (GVA) (GRISOLIAP/2019/132). S.T.-G. acknowledges MICI for his Ramón y Cajal contract (RYC2019-027784-I).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Article* **Preparation and Characterization of a New Polymeric Multi-Layered Material Based K-Carrageenan and Alginate for Efficient Bio-Sorption of Methylene Blue Dye**

**Chiraz Ammar 1,2, Fahad M. Alminderej 3, Yassine EL-Ghoul 2,3,\*, Mahjoub Jabli 4,5 and Md. Shafiquzzaman <sup>6</sup>**


**Abstract:** The current study highlights a novel bio-sorbent design based on polyelectrolyte multilayers (PEM) biopolymeric material. First layer was composed of sodium alginate and the second was constituted of citric acid and k-carrageenan. The PEM system was crosslinked to non-woven cellulosic textile material. Resulting materials were characterized using FT-IR, SEM, and thermal analysis (TGA and DTA). FT-IR analysis confirmed chemical interconnection of PEM bio-sorbent system. SEM features indicated that the microspaces between fibers were filled with layers of functionalizing polymers. PEM exhibited higher surface roughness compared to virgin sample. This modification of the surface morphology confirmed the stability and the effectiveness of the grafting method. Virgin cellulosic sample decomposed at 370 ◦C. However, PEM samples decomposed at 250 ◦C and 370 ◦C, which were attributed to the thermal decomposition of crosslinked sodium alginate and k-carrageenan and cellulose, respectively. The bio-sorbent performances were evaluated under different experimental conditions including pH, time, temperature, and initial dye concentration. The maximum adsorbed amounts of methylene blue are 124.4 mg/g and 522.4 mg/g for the untreated and grafted materials, respectively. The improvement in dye sorption evidenced the grafting of carboxylate and sulfonate groups onto cellulose surface. Adsorption process complied well with pseudo-first-order and Langmuir equations.

**Keywords:** polyelectrolyte multi-layers; sodium alginate; k-carrageenan; cellulosic nonwoven textile; surface functionalization; characterization; bio-sorption; isotherms

### **1. Introduction**

The discharge of dyes from industries and their elimination has received much attention as they can damage human health and animals [1–3]. Indeed, dye molecules are resistant to natural degradation, allergenic, carcinogenic, and stable in the presence of oxidizing agents [4–6]. This interesting topic requires the development of various technologies to treat colored waters. Biological treatment and coagulation/flocculation processes are viewed as ineffective to treat soluble dyes [7–9]. Adsorption appeared more effective as it is simple and economic and it is especially used to remove pollutants, which are not easily biodegradable. Thus, a specific attention is devoted to explore new adsorbents, which could be cheaper, proficient, and easy regenerated.

**Citation:** Ammar, C.; Alminderej, F.M.; EL-Ghoul, Y.; Jabli, M.; Shafiquzzaman, M. Preparation and Characterization of a New Polymeric Multi-Layered Material Based K-Carrageenan and Alginate for Efficient Bio-Sorption of Methylene Blue Dye. *Polymers* **2021**, *13*, 411. https://doi.org/10.3390/ polym13030411

Academic Editors: Rafael Antonio Balart Gimeno, Daniel García García, Vicent Fombuena Borrás, Luís Jesús Quiles Carrillo and Marina Patricia Arrieta Dillon Received: 22 December 2020 Accepted: 25 January 2021 Published: 28 January 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

In this sense, several polymeric adsorbents were used for the removal of dyes from contaminated matters [10–12]. Concerning textile adsorbent materials, they are rarely investigated in the literature. Some sorbent material based synthetic functionalized textiles were studied for the removal of cationic dyes. Despite their pretreatment and/or functionalization, adsorption capacities were limited due to their hydrophobicity [13–15]. Cellulose, alginates, and k-carrageenan biopolymers were naturally abundant polysaccharides. Cellulose is recognized for its good hydrophilicity [16–21]. Alginate, a polysaccharide biopolymer, is an essential component of the cell wall of brown algae. It could be an excellent bio-sorbent of organic dyes due to its high contents of carboxylic and hydroxyl functional groups [22–25]. K-carrageenan is a natural sulfated polysaccharide extracted from red edible seaweeds. It is widely used in the food industry, owing to gelling, thickening, and stabilizing properties [26,27]. As examples of works conducted in previous literatures, Rahman et al., in a comprehensive review, discussed synthesis, characteristics, and methylene blue adsorption of various cellulose nanocrystal-based hydrogels [28]. They indicated that the addition of other polysaccharides including chitosan and alginate into cellulose displayed remarkably improved adsorption capacities for methylene blue. Guesmi et al. demonstrated that the addition of sodium alginate (5–20%) to hydroxyapatite improved well methylene blue adsorption [29] and the adsorption capacity increased from 77.51 to 142.85 mg/g, using hydroxyapatite and hydroxyapatite-alginate, respectively. In our previous work, we have demonstrated that the adsorption of methylene blue onto extracted cellulose, from Aegagropila L., reached 109 mg/g and it was only about 47mg/g for the raw marine macroalgae [30]. Yang et al. synthesized gel beads based on k-carrageenan and graphene oxide [31] and observed that the adsorption capacity of the gel beads for methylene blue attained 658.4 mg/g at 25 ◦C. As globally observed, the developed materials-based polysaccharides are outstanding adsorbents with excellent adsorption capacities of cationic dyes.

In line with this emergent topic, the current study proposes a novel adsorbent design based on polyelectrolyte multi-layered (PEM) biopolymeric material as a potent bio-sorbent. This is formed by an alternation of layers of two polyelectrolyte biopolymers. The first layer is composed of alginate polyanion and the second is obtained after the reticulation of the citric acid with k-carrageenan polyanion polymer. The PEM system is crosslinked to cellulosic natural material and applied as adsorbent of cationic dyes. Indeed, the proposed technique offers the adsorption characteristics of not only cellulose as the main support but also k-carrageenan and sodium alginate as immobilized biopolymers. The system provides massive hydroxyl, sulfonate, and carboxylate groups suggesting therefore the use of the resulting material as a bio-sorbent of cationic species. Therefore, the importance of our proposal lies in the preparation of a new PEM biomaterial based on alginate and k-carrageenan crosslinked to cellulosic material, which is very rich in anionic functional groups. In addition, the investigation of the layer-by-layer functionalization method in the adsorption of textile rejects is a novel study and was not reported in the literature. The prepared materials were characterized using FT-IR, SEM, TGA, and DTA. The bio-sorbent performances were evaluated under different experimental conditions including pH, time, temperature, and the initial dye concentration. The theoretical kinetic and isotherms equations were used to analyze the experimental data.

#### **2. Experimental**

### *2.1. Materials*

A non-woven cellulosic textile material was used in this study. Its surface weight was 240 g/m<sup>2</sup> and the thickness had an average of 0.6 mm. The textile is a 3-dimensional calendared fiber network. Sodium alginate (AG) was supplied from Sigma-Aldrich with a medium molecular weight (30,000 g/mol) having a degree of deacetylation of 70%. Citric acid (CTR, 226.2 g/mol) and Kappa carrageenan (a fine white powder) were purchased from Sigma-Aldrich. All chemicals were used without further purification. Methylene blue (MB) was supplied by Central Drug House (India).

### *2.2. Preparation of PEM Bio-Sorbent*

PEM bio-sorbent was prepared via an alternating grafting process of sodium alginate and k-carrageenan using citric acid as a crosslinking agent. The process of grafting of the cellulosic material is based on a pad-dry technique. PEM bio-sorbent was carried out by functionalizing the cellulosic material with alternating baths in sodium alginate/citric acid solution and then in a K-carrageenan solution. The PEM is deposited by alternating successive baths according to the "layer-by-layer deposition" method. Different pairs of layers (from 1 to 8) are prepared for bio-sorbent cellulosic material. Each impregnation was performed in a total volume of 50 mL. All polymer solutions are completely renewed after the deposition of 2 pairs of layers. The samples (4 cm × 4 cm) were treated in a solution of soluble sodium alginate polymer (0.4% *w/v*) and citric acid (0.4% *w/v*) under constant stirring at 170 rpm for 20 min at room temperature. After each impregnation, the samples were dried 20 min at 90 ◦C. After drying, the samples were impregnated in a solution of k-carrageenan (at 0.8% *w/v*) with stirring (170 rpm) for 20 min at room temperature. The cycle was repeated as many times as necessary. At each end of the cycle, a pair of layers was thus deposited and crosslinked on the cellulose material.

During the preparation process of PEM materials, the weight gain was calculated after depositing each pair of layers of sodium alginate and k-carrageenan. The following formula was applied illustrating the weight gain as a function of the number of pairs of grafted copolymer layers.

$$\%-\text{Weight gain} \left(n\right) = \frac{\left(m\_n - m\_i\right)}{m\_i} \times 100\tag{1}$$

where *n* is the number of pairs of layers and *mi* and *mn* are the sample weights before and after functionalization, respectively.

### *2.3. Characterizations*

For the analysis of the chemical structure of prepared multilayered adsorbent, infrared spectroscopy analysis was conducted using a FT-IR spectrometer (Agilent Technologies, Cary 600 Series FTIR Spectrometer) via the attenuated total reflection mode (ATR). Spectra of cellulose material and multilayered bio-adsorbent polymeric system were recorded at a range of 4000 to 400 cm<sup>−</sup>1, and with a resolution of 2 cm−1.

Thermal stability and different thermal properties of natural cellulosic material and PEM bio-sorbent were determined by thermogravimetric measurements using TA Instruments apparatus. The fixed parameters for different analysis were a heating rate of 10 ◦C min−<sup>1</sup> and a temperature from 25 to 600 ◦C.

Surface morphology was assessed using a scanning electron microscopy (FEI Quanta SEM). An accelerating voltage of 5 KV with various magnification essays was applied for the surface analyzing of different samples. The SEM analysis was preceded by a coating procedure of samples with a carbon layer to enhance their conductivity.

#### *2.4. Adsorption Experiments*

The adsorption tests were carried out in a batch reactor by stirring the colored synthetic solutions in the presence of each of the adsorbents at a constant agitation speed (150 rpm). We studied the effect of the main parameters influencing the adsorption capacity such as pH (ranged from 3 to 9), contact time (in a range of 0 to 120 min.), initial dye concentration (varied from 50 to 1000 mg/L), and temperature (22, 40, and 60 ◦C). The residual concentration of each of the dyes was determined using an UV/visible spectrophotometer (Shimadzu UV-2600).

The adsorbed methylene blue dye amount (q(mg/g)) onto the surface of PEM crosslinked to non-woven cellulosic textile material was calculated using the following formula:

$$\mathbf{q(mg/g)} = \frac{(\mathbf{C}\_0 - \mathbf{C}\_t) \cdot \mathbf{V}}{m} \tag{2}$$

where *C*<sup>0</sup> and *Ce* are the initial and residual concentration (mg/L), respectively, *V* is the volume of methylene blue dye (L) used for the adsorption, and m is the mass of the biosorbent (g) used for the adsorption.

#### **3. Results and Discussion**

### *3.1. Preparation of PEM Biopolymer Adsorbent*

The design of PEM was carried out by grafting the cellulosic material with alternating layers of sodium alginate and K-carrageenan biopolymers crosslinked by citric acid (Figure 1). The layer-by-layer deposition technique was used to elaborate different multilayered bio-sorbent (from 1 to 8-layer pairs). The variation of the weight gain according to the number of grafted layers of sodium alginate and K-carrageenan was presented in Figure 2. We noticed the progressive increase of total weight gain with the number of polymeric layers grafted to the cellulosic material. Results revealed a more significant increase of the weight gain within 3-pairs-layers finishing the cellulosic sample. For the characterization and adsorption investigations, the samples treated with three polymeric layers of reticulated alginate and K-carrageenan will be investigated.

**Figure 1.** Illustration of the functionalization of cellulosic material upon polyesterification reactions via CTR (Citric acid) crosslinking, leading to the polyanionic PEM bio-sorbent system after curing at 140 ◦C for 15 min.

### *3.2. FT-IR Spectroscopy Analysis*

FT-IR-ATR analysis was performed in order to characterize the chemical grafting and conception of the bio-sorbent material via the identification of the different functional groups appearing after functionalization. This characterization was applied to analyze both the untreated cellulosic sample and the PEM grafted adsorbent material. Spectra of untreated cellulosic sample and grafted PEM material (3-layers PEM) were showed in Figure 3. Different new peaks appeared in the PEM grafted cellulosic material proving the chemical grafting. A peak centered at 1705 cm–1 appeared referring to the ester groups confirming the polyesterification reaction established between carboxylic groups of citric acid and hydroxyl functions of both alginate and carrageenan biopolymers. This is in line with previous research studies, in which FT-IR were used to confirm the evidence of a polyesterification reaction between cellulosic material and different polysaccharides via polycarboxylic acids as crosslinking agents [32–34]. A peak closed to 1310 cm–1 appeared clearly, which corresponded to the symmetric stretching vibration of the carboxylic acid groups COO\_ of galacturonic acid of alginate grafted polymer [35]. In addition, a more significant wide peak around to 3290 cm–1 appeared with grafted PEM material was attributed to hydroxyl groups of cellulose, alginate, and carrageenan grafted polymers. Briefly, FT-IR analysis of the two spectra, permitted us to confirm the chemical interconnection of the PEM bio-sorbent system and indicated the efficiency of the applied grafting chemical process.

**Figure 2.** Weight gain variation as function of the number of layers grafted onto the cellulosic bio-sorbent material.

**Figure 3.** FT-IR-ATR spectra of virgin cellulosic sample (**a**) and PEM grafted bio-sorbent material (**b**).

### *3.3. TGA and DTA Analysis*

Figure 4 showed thermograms of untreated and grafted PEM samples. The untreated cellulosic sample showed only one zone showing one temperature of degradation referring to the cellulosic material closed to 370 ◦C. However, two distinct parts appeared in the thermogram of the designed PEM sample. The first with a temperature of degradation around 250 ◦C was attributed to the crosslinked polymer of alginate and K-carrageenan. The second closed to 370 ◦C, with a high loss of weight referring to the degradation of the

cellulosic sample. The observed weight loss for the two samples, around 100 ◦C, was due the evaporation of water absorbed inside their structures. At this temperature we noticed a higher weight loss in the case of the grafted PEM material. This was due to the increase of the hydrophilicity after biopolymers grafting onto the cellulosic samples. Furthermore, the prepared PEM samples presented a higher residual weight after degradation (20%) compared to the untreated one (0%). This proved well the improvement thermal stability of the PEM bio-sorbent material after grafting.

**Figure 4.** TGA (**a**) and DTA (**b**) thermograms of virgin cellulosic sample and grafted PEM bio-sorbent material.

DTA results confirmed the TGA analysis and revealed one significant peak in the derivative signal ascribed to the cellulose degradation in the case of untreated sample. For the untreated PEM sample two principal peaks appeared, which were attributed to the degradation of the grafted polymers (at 250 ◦C) and the degradation of the cellulosic material (at 370 ◦C).

### *3.4. Morphological Analysis*

Figure 5 showed the micrographs of untreated cellulosic sample and grafted PEM material with 3-layer pairs. We noticed a significant surface modification after grafting. The microspaces between fibers were filled with the layers of the two functionalizing polymers. The grafted PEM revealed a higher surface roughness compared to the virgin sample. This modification of the surface morphology confirmed the stability and the effectiveness of the applied method of grafting.

**Figure 5.** SEM micrographs of (**a**) untreated cellulosic sample and (**b**) PEM grafted bio-sorbent material.

### *3.5. Application to the Adsorption of Methylene Blue*

In this investigation, the prepared untreated and grafted materials were applied as adsorbents of methylene blue in synthetic medium by varying pH, time, initial dye concentration, and temperature. Figure 6a represents the evolution of the adsorbed methylene blue amount as a function of pH. The maximum adsorbed amount was reached at pH = 6. In fact, under high acidic conditions, the positively charged methylene blue ions opposed the positively charged adsorbent surface leading to low adsorbed rates. At higher pH values (pH < 6), the adsorbent surface became negative, favoring an electrostatic interaction with methylene blue. At basic conditions, the adsorbed dye amount decreases, which could be explained by the repulsion forces between dye molecules and biosorbent surface.

**Figure 6.** (**a**) Effect of pH on methylene blue adsorption (C0 = 600 mg/L, t = 120 min), (**b**) effect of time, (**c**) effect of initial dye concentration, and (**d**) effect of temperature.

The time required to achieve equilibrium was observed at 120 min (Figure 6b). It is seen that the adsorption was fast during the first period of times (from 0 to 50 min) where more than 80% of target was accomplished. This trend could be explained by the fact that many adsorption sites are available during this first stage at the surface of the adsorbents. After this period of times, the adsorption attained a steady state, which could be interpreted by the saturation of the adsorption sites. Results showed also that the maximum adsorbed amount of methylene blue are 124.4 mg/g and 522.4 mg/g for the untreated materials and grafted ones, respectively. This difference in the sorption capacities could be explained by the addition of new functional groups (carboxylate and sulfonate groups) grafted on the surface of cellulose.

Scheme 1 provides a schematic representation of hydrogen bonding and ionic interaction between methylene blue molecule and PEM crosslinked to non-woven cellulosic textile material surface. Indeed, the free hydroxyl groups of cellulose non-woven could interact with nitrogen atom of methylene blue via hydrogen bonding. However, the sulfonate groups (SO3 −) of K-carrageenan and/or carboxylate groups of both sodium alginate (COO−) and CTR crosslinking agent could react with positive nitrogen atom (N+) through ionic interaction.

**Scheme 1.** A schematic representation of hydrogen bonding and ionic interaction between methylene blue molecule and the surface of PEM crosslinked to non-woven cellulosic textile material.

The adsorbed dye amount was found to greatly increase with the increase of initial methylene blue concentration. At equilibrium, it reached 171.4 mg/g for the untreated samples and 590.5 mg/g for the grafted samples (Figure 6c). This adsorbed amount depends on the temperature value and it decreased for example in the case of grafted samples from 590.5 mg/g at 22 ◦C to 453.5 mg/g at 60 ◦C (Figure 6d). This indicated that the adsorption of methylene blue was exothermic in this case. Indeed, at higher temperatures values, the dye could be desorbed from the samples.

### *3.6. Kinetic Modeling*

Kinetics data were significant to recognize the attraction between adsorbates and adsorbents at equilibrium state. It could suggest if the studied mechanism is physical and/or chemical, and mass transport. Herein, modeling kinetic data was assessed using pseudo first order, pseudo second order, Elovich, and intradiffusion kinetic models. Results are depicted in Figure 7. The computed kinetic parameters for the different equations were summarized in Table 1. The obtained curves and computed kinetic parameters indicated that pseudo-first-order equation could describe well the kinetic data (0.96 < R2). The correlation coefficients obtained within the pseudo-second-order were also high (0.95 < R2). These results indicated that the adsorption process is so complex and could be considered as physical and chemical [36]. The divergence of the plots for the intraparticle diffusion model from the origin suggests that this model was not the sole rate-controlling step [37].

**Figure 7.** Kinetic data modeled using: (**a**) pseudo first order, (**b**) pseudo second order, (**c**) Elovich, and (**d**) intraparticular diffusion.


**Table 1.** Summarized kinetic parameters for the grafted and ungrafted samples.

Where: α (mg/g/ min) and β (mg/g/min) are Elovich constants; R<sup>2</sup> is the regression coefficient.

### *3.7. Isotherms Investigation*

The relationship between the grafted materials and the studied adsorbate was evaluated using the equations of Langmuir, Freundlich, and Temkin (Figure 8). The calculated parameters were shown in Table 2. We observed that the Langmuir equation fitted well with the experimental data (0.99 ≤ <sup>R</sup>2) compared to the other studied equations. This trend suggested that the adsorption process is a monolayer, and the sorption sites are homogeneous having similar adsorption capacities [38]. The values of (1/*n*) revealed adsorption intensity or surface heterogeneity. Indeed, when the value of 1/n ranges from 0.1 to 1.0, it therefore reflects a good adsorption [39]. In our study 2.3 ≤ *n* ≤ 3.8, which indicates that the adsorption of the molecules of methylene blue onto the surface of the grafted samples is so favorable. The decrease of the values of the adsorption energy constant (bT), determined from the Temkin model, with the increase of temperature indicated the exothermic nature of the adsorption mechanism. This is consistent with the trends observed within the effect of temperature discussed above.

**Figure 8.** Isotherms data modeled using: (**a**) Langmuir, (**b**) Freundlich, and (**c**) Temkin.



### **4. Conclusions**

In this study, a new multilayered polymeric bio-sorbent was elaborated using the layer-by-layer grafting method. The prepared cellulosic material based k-carrageenan and alginate crosslinked biopolymers were then chemically and morphologically characterized. Chemical grafting via polyesterification reactions and the stability of the multilayer functionalization were confirmed by FT-IR, TGA-DTA, and SEM analysis. Methylene blue dye was used as an adsorbate for the bio-sorption evaluation on the PEM grafting material in batch experiments. The influence of the different process parameters was studied with respect to the sorption equilibrium. The designed bio-sorbent showed excellent sorption capacities of MB with a capacity above 522.4 mg/g. The improvement in dye sorption was evidently due to the presence of many carboxylate and sulfonate groups of the crosslinked k-carrageenan and alginates grafted on the cellulosic surface material. The correlation of the experimental data with the theoretical equations showed that the kinetic data could be described with both pseudo first order and pseudo second order equations suggesting

that the adsorption process involved physical and chemical interactions. The adsorption phenomenon occurred in heterogeneous adsorption sites and it was exothermic and spontaneous. Langmuir isotherm fitted better to adsorption data, indicating that the adsorption process was localized on a monolayer, and all adsorption sites on the adsorbent were homogeneous and had the same adsorption capacity. In brief, we developed a simple and new polymeric material for removing one of the most dangerous contaminants from textile industrial aqueous discharges. Further investigation could be extended for the exploration of this material in the removal of other pollutants such as metals and pesticides.

**Author Contributions:** Data curation, C.A., Y.E.-G. and M.S.; Formal analysis, C.A., F.M.A., Y.E.-G. and M.J.; Investigation, C.A., F.M.A. and M.S.; Methodology, C.A., Y.E.-G. and M.S.; Project administration, F.M.A. and Y.E.-G.; Software, F.M.A. and M.S.; Supervision, F.M.A.; Validation, F.M.A., Y.E.-G. and M.J.; Writing—original draft, C.A., Y.E.-G. and M.J.; Writing—review and editing, Y.E.-G. and M.J. All authors have read and agreed to the published version of the manuscript.

**Funding:** The authors gratefully acknowledge Qassim University, represented by the Deanship of Scientific Research, on the financial support for this research under the number 5656-cos-2019-2-2-I, during the academic year 1440 AH/2019 AD.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Article* **Production and Properties of Lignin Nanoparticles from Ethanol Organosolv Liquors—Influence of Origin and Pretreatment Conditions**

**Johannes Adamcyk 1,\*, Stefan Beisl 1,2, Samaneh Amini 1, Thomas Jung 1, Florian Zikeli 1, Jalel Labidi <sup>2</sup> and Anton Friedl <sup>1</sup>**


**Abstract:** Despite major efforts in recent years, lignin as an abundant biopolymer is still underutilized in material applications. The production of lignin nanoparticles with improved properties through a high specific surface area enables easier applicability and higher value applications. Current precipitation processes often show poor yields, as a portion of the lignin stays in solution. In the present work, lignin was extracted from wheat straw, spruce, and beech using ethanol organosolv pretreatment at temperatures from 160–220 ◦C. The resulting extracts were standardized to the lowest lignin content and precipitated by solvent-shifting to produce lignin micro- and nanoparticles with mean hydrodynamic diameters from 67.8 to 1156.4 nm. Extracts, particles and supernatant were analyzed on molecular weight, revealing that large lignin molecules are precipitated while small lignin molecules stay in solution. The particles were purified by dialysis and characterized on their color and antioxidant activity, reaching ASC equivalents between 19.1 and 50.4 mg/mg. This work gives detailed insight into the precipitation process with respect to different raw materials and pretreatment severities, enabling better understanding and optimization of lignin nanoparticle precipitation.

**Keywords:** lignin; nanoparticles; biorefinery; organosolv pretreatment

### **1. Introduction**

Lignin as the second most abundant biopolymer after cellulose has gained massive interest in recent years. It is estimated that approximately 2 × 1011 tons of lignocellulosic biomass residues are produced worldwide every year [1], making lignin a renewable phenolic compound available in large quantities. Since it is already produced in side-streams of pulp-productions and biorefineries, it has the potential to enable the transition from a fossilbased to a biobased economy by supplanting synthetic compounds currently produced from fossil resources. Since only around 40% of the lignin produced in pulping processes are needed to cover the internal energy demand of the processes [2,3], the possibility exists to vastly increase the amount of lignin used for material purposes providing efficient use of all biomass components and improving the sustainability of the process. Thus, the use of lignin as a material should be heavily expanded.

One challenge in the utilization of lignin is its diversity, as it is known to differ for different plant species, but also depending on the extraction process [4]. Lignin from the most common pulping processes, the Kraft and sulfite processes, contains a relatively-high amount of sulfur and may therefore not be suitable for many practical applications. In contrast, ethanol organosolv pretreatment yields lignin of high purity, with a structure closely resembling that of native lignin, and without sulfur contamination through the

**Citation:** Adamcyk, J.; Beisl, S.; Amini, S.; Jung, T.; Zikeli, F.; Labidi, J.; Friedl, A. Production and Properties of Lignin Nanoparticles from Ethanol Organosolv Liquors—Influence of Origin and Pretreatment Conditions. *Polymers* **2021**, *13*, 384. https:// doi.org/10.3390/polym13030384

Academic Editor: Rafael Antonio Balart Gimeno Received: 29 December 2020 Accepted: 22 January 2021 Published: 26 January 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

process [5,6], opening up possibilities for new applications [6]. A disadvantage of this process is its lower efficiency and high energy demand for solvent recovery [7]. Improvement of the process economy while still maintaining the high quality of the lignin produced will be necessary to compete with other pretreatment technologies on the one hand and fossil-based products that are aimed to be replaced by lignin on the other.

Although increased process severity tends to increase delignification and lignin concentration in the liquor [8], which would be desirable for increased process efficiency, simultaneously more sugar degradation products are formed [9], and the lignin changes structurally, which might be problematic for the final product. Thus, process efficiency needs to be balanced out with the final product quality.

Another possibility to improve the economy of a biorefinery based on organosolv pretreatment is finding high value applications for the produced lignin. Recent works have shown that colloidal lignin particles have many desirable and improved properties which lead to a wide range of possible applications [10], due to the high ratio of surface to volume. Especially the application of lignin for UV-protection in cosmetics [11–13] but also wood conservation [14] has been reported and commercial use in these areas seems within reach. Qian et al. [15] reported that the ability to block UV-radiation is improved for smaller particle sizes, which underlines the connection between improved properties and particle size. However, previous works have shown that a major portion of the lignin remained in solution after the precipitation of colloidal particles, resulting in a lower overall yield [16,17]. This raises the suspicion that lignin gets fractionated by molecular weight in the precipitation step, which was confirmed for commercial soda lignin by Sipponen et al. [18]. Further understanding of the precipitation of lignin should make it possible to improve the process towards higher yields while still maintaining small particle sizes.

Based on this, the present work investigates the extraction with ethanol organosolv pretreatment and the subsequent precipitation of lignin into colloidal particles. Three different raw materials (wheat straw, spruce wood, and beech wood), representing three combinations of guaiacyl (G), *p*-hydroxy phenol (H) and syringyl (S) lignins common in nature (GSH, G and GS lignins, respectively), and four temperatures (160–220 ◦C) were applied in the pretreatment process. Through this experimental plan, the influence of raw material and process severity on the lignin precipitation was studied. The extracts' compositions were characterized, and the prepared lignin particles were analyzed regarding their size and physico-chemical properties. The different process fractions were analyzed regarding their molecular weight to further investigate the precipitation process.

#### **2. Materials and Methods**

An overview of the experimental procedure is shown in Figure 1, indicating process steps, fractions, and analytics. Generally, lignin was extracted from the different raw materials and precipitated into colloidal particle suspensions by addition of an antisolvent. A part of these suspensions was centrifuged for analytics, the remainder was dialyzed to remove dissolved impurities. The last step was membrane filtration to produce a thin particle layer for color measurements.

**Figure 1.** Schematic of the experimental plan conducted for each pretreatment condition and raw material. The analyzed process fractions are depicted in green.

The particle fraction separated via dialysis is termed "particles after dialysis" and the particle fraction separated via centrifugation is termed "particles" throughout the whole manuscript. The particles after the centrifugation are considered the same as the ones in the suspension directly after precipitation.

### *2.1. Materials*

The wheat straw used was harvested in 2015 in Lower Austria, the spruce wood was supplied by Laakirchen Papier (Laakirchen, Austria), the beech wood by Lenzing AG (Lenzing, Austria). The particle size of all raw materials was reduced in a cutting mill with a 2 mm mesh, the resulting material was stored under dry conditions until the pretreatment. Ultra-pure water (18 MΩ/cm) and Ethanol (Merck, 96 vol %, undenaturated, Darmstadt, Germany) was used in the organosolv treatment.

#### *2.2. Pretreatment/Extraction*

The organosolv pretreatment was conducted ina1L stirred autoclave (Zirbus, HAD 9/16, Bad Grund, Germany) using a 60 wt% aqueous ethanol mixture as solvent under consideration of the water content in the raw material. The biomass content in the reactor based on dry matter was 8.3 wt%. The reactor was heated to the pretreatment temperature and then held at this temperature. After 60 min of total pretreatment time, the reactor was cooled to room temperature. The solid and liquid fractions were separated using a hydraulic press (Hapa, HPH 2.5, Achern, Germany) at 200 bar, the extract was then centrifuged (Thermo Scientific, Sorvall, RC 6+, Waltham, MA, USA) at 24,104× *g* for 20 min to remove residual solids. The particle-free extracts of the single batches were unified, the composition was analyzed, and it was stored at 5 ◦C until the precipitation experiments were performed.

### *2.3. Precipitation*

To eliminate the influence of the lignin concentration on the precipitation, the concentration of all extracts was set to the lowest total lignin concentration (3.52 g/L) by dilution with 60 wt% aqueous ethanol. The precipitation experiments were conducted at 25 ◦C, the volume ratio of extract to antisolvent was kept constant at 1:5, where the antisolvent consisted of Ultra-pure water. Setup (b) as described in [16] was used for the precipitations. The resulting suspensions were stored at 5 ◦C.

#### *2.4. Downstream Processing*

Parts of the suspensions were centrifuged in a Thermo WX Ultra 80 ultracentrifuge (Thermo Scientific, Waltham, MA, USA) at 288,000× *g* for 60 min. The supernatants were decanted and stored at 5 ◦C, the liquid free particles were freeze dried and stored in a desiccator.

The rest of the suspensions were dialyzed for 7 days in excess of water (daily replaced) using Nadir® dialysis hoses with 10,000–20,000 Dalton cut-off. The particles after dialysis were stored at 5 ◦C.

### *2.5. Analytics*

#### 2.5.1. Extract Characterization

The organosolv extract was analyzed for carbohydrates, lignin, and degradation products. The carbohydrate content was determined by the sample preparation following the National Renewable Energy Laboratory (NREL) laboratory analytical procedure (LAP) "Determination of sugars, byproducts, and degradation products in liquid fraction process samples" [19] but with no neutralization of the samples after hydrolysis. A Thermo Scientific ICS-5000 HPAEC system equipped with a photo array detector (PAD) with deionized water as the eluent was used for the determination of arabinose, glucose, mannose, xylose, and galactose. The concentrations of the degradation products acetic acid, hydroxymethylfurfural (HMF) and furfural were determined with a Shimadzu LC-20A

"prominence" HPLC system and a Shodex SH1011 (Showa Denko, Tokyo, Japan) column at 40 ◦C, with 0.005 M H2SO4 as eluent. The acid insoluble and acid soluble lignin contents were determined following the NREL LAP "Determination of Structural Carbohydrates and Lignin in Biomass" [20] using the dry matter of the extract obtained at 105 ◦C.

### 2.5.2. HP-SEC Analysis

Molar mass distributions of the standardized organosolv extracts, suspensions, centrifuged lignin particles, supernatants, and lignin particles after dialysis were determined by alkaline High Performance Size Exclusion Chromatography (HP-SEC) analysis (eluent: 10 mM NaOH) using three TSK-Gel columns in series at 40 ◦C (PW5000, PW4000, PW3000; TOSOH Bioscience, Darmstadt, Germany) with an Agilent 1200 HPLC system (flow rate: 1 mL/min, DAD detection at 280 nm, Santa Clara, CA, USA). Solid lignin fractions were dissolved in the eluent, and liquid lignin fractions were diluted with the eluent to adjust the pH for analysis. Calibration of the columns set was done using polystyrene sulfonate reference standards (PSS GmbH, Mainz, Germany). The molar masses at peak Maximum (Mp) were 78,400 Da, 33,500 Da, 15,800 Da, 6430 Da, 1670 Da, 891 Da, and 208 Da.

### 2.5.3. Hydrodynamic Diameter

Particle size measurements were conducted in a ZetaPALS (Brookhaven Instruments, Holtsville, NY, USA) using dynamic light scattering (DLS). The refractive index of the particles was set to 1.53 and the imaginary refractive index to 0.1. All reported mean values for hydrodynamic diameters are based on intensity based distributions.

### 2.5.4. Color

The colorimeter PCE-CMS 7, PCE Instruments (Albacete, Spain) was used to measure the Hunter color values (L, a, and b) of the lignin. The lignin particles after dialysis were filtered against a HFK™-131 ultrafiltration membrane (Koch Membrane Systems, Rimsting, Germany) resulting in layer densities of lignin ranging from around 1 g/m2 to 14 g/m2. All membranes were purged with 10 wt% aqueous ethanol and the color background (L = 99.08, a = 0.001, b = −4.911) was measured after purging. The color of each lignin sample was measured at three different lignin layer densities and was standardized to 3 g/m2 via linear regression of the L, a, and b values.

### 2.5.5. Antioxidant Activity

All particle suspensions were filtrated using a PES-Filter (Merck, Darmstadt, Germany) with a pore size of 0.22 μm to remove agglomerates and yield comparable particle size distributions.

*ABTS:* The determination of 2,2 -azinobis-(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) radical inhibition was based on the method of Re et al. [21]. Twenty-five milliliters of ABTS stock solution was prepared in dark bottle by dissolving 69.7 mg of ABTS salt and 11.8 mg K2S2O8 in water. The stock solution was left in a refrigerator and diluted with deionized water to obtain an initial absorbance of 0.7 at 734 nm in a Jasco V-630 spectrophotometer (JASCO, Tokyo, Japan) before use. To determine the radical scavenging ability, the sample was added to the diluted ABTS solution. The absorbance was measured after 6 min incubation at 734 nm. The percentage of inhibition was calculated from below equation:

$$\mathbf{AA}[\%] = (\mathbf{Ac} - \mathbf{At}) / \mathbf{Ac} \times 100\% \tag{1}$$

where: AA-antioxidant activity, Ac—absorbance of control sample, At—absorbance of tested sample.

*FRAP:* The ferric reducing antioxidant power (FRAP) test is based on the reduction of Fe(III) to Fe(II) by the antioxidant compound which forms a colored complex (593 nm) with 2,4,6-tripyridyl-s-triazine (Fe(II)-TPTZ) in acetate buffer at pH 3.6 [22]. Briefly, the reactive solution was freshly prepared with 25 mL of 300 mM acetate buffer (pH 3.6), 2.5 mL of 10 mM 2,4,6-tripyridyl-s-triazine in 40 mM HCl and 2.5 mL 20 mM FeCl3·6H2O in distilled water. Samples (0.1 mL) were mixed with 3 mL of the FRAP reactive solution. The reference standard used was ascorbic acid (ASC) and results are given as ASC equivalents.

#### **3. Results**

#### *3.1. Extract Composition*

The composition of the extracts is shown in Figure 2. The amount of each compound in the extract increases with temperature, the largest increase of acid insoluble lignin (AIL) and the degradation products happens from 200 to 220 ◦C for all raw materials, indicating a considerable increase in process severity. Lignin is the dominant compound in the dry matter of the extracts, making up between 50 wt% for extracts from straw and 70 wt% for extracts from the woods. From 160 to 200 ◦C, the total lignin content of the wheat straw extracts is higher than that of the woods due to the consistently higher amount of acid soluble lignin (ASL) in wheat straw extracts, while at 220 ◦C the lignin content in the extract from spruce is highest, followed by beech (Figure 2a). This could suggest that although the initial lignin content of wheat straw is lower than that of the woods, the extraction is more efficient already at lower temperatures, and a higher severity is required for efficient delignification of the woods when using ethanol organosolv pretreatment. Siika-aho et al. [23] applied an oxygen/carbonate process to three different raw materials at varying process severities and also found delignification of straw to be more efficient at low process severities compared to spruce and beech.

**Figure 2.** Concentration of lignin (**a**), acetic acid and carbohydrates (**b**), and degradation products (**c**) in organosolv extracts. Abbreviations: AIL—acid insoluble lignin, ASL—acid soluble lignin, HMF—hydroxymethylfurural.

The trends of the carbohydrates in the extracts (Figure 2b) are similar for all raw materials, though beech wood has consistently the highest concentrations. For acetic acid, extraction from wheat straw results in a higher concentration than from the woods in most cases, though there is a strong increase for beech from 200 ◦C to 220 ◦C. For the degradation products (Figure 2c), the amounts are of the same order of magnitude from 160 ◦C to 200 ◦C, at 220 ◦C hydroxymethylfurfural (HMF) of spruce and furfural of beech stand out.

To remove the influence of the lignin concentration in the precipitation step, the extracts were diluted to the lowest lignin concentration (3.52 g/L) with aqueous ethanol (60 wt %). As demonstrated in Figure 3a, the molecular weight distribution of the lignin in the extracts is influenced by the pretreatment temperature and the raw material. The changes between the raw materials become more visible in the molecular weight distributions (Figure 3b). There are four distinct peaks of different molecular weight, with maxima at approximately 900, 610, 410, and 255 Da. The two lower molecular weight fractions are more dominant in the wheat straw extracts than in those of the woods, which could be linked to the higher concentration of ASL found in wheat straw extracts. In earlier works the third peak was found to be strongly influenced by p-hydroxycinnamic acids like ferulic acid and p-coumaric acid, in monomeric form or connected to lignin fragments present in wheat straw [24]. Apart from the differences in the raw materials, the pretreatment temperature influences the molecular weight distribution of the dissolved lignin; large, dissolved lignin molecules are depolymerized at higher temperatures, and lignin fragments can again re-condensate at severe conditions [25]. The combination of raw material and temperature decides which effect is dominant. The Mw of all three raw materials show a trend to lower values with increasing temperature (Figure 3b), which is more pronounced for wheat straw than for the woods. In Figure 3c, the absorbance is multiplied with the molecular weight to highlight the strong influence of high molecular weight fractions on the Mw. These molecular weight distributions also indicate (stronger than the weight averaged values) that the temperature increase from 200 to 220 ◦C significantly influences the molecular weight.

### *3.2. Precipitation*

The lignin dissolved in the extracts was precipitated into micro- and nanoparticles by addition of water as an antisolvent. The resulting suspensions were characterized on the particles' hydrodynamic diameter by dynamic light scattering. The particles and supernatant were separated by centrifugation for analysis; parts of the suspensions were dialyzed to remove impurities before further analytics. Visible agglomerates formed during dialysis were removed by filtration before determination of the antioxidant activity in order to analyze exclusively colloidal particles. Molecular weight of lignin particles after precipitation, dissolved lignin in the supernatants, and lignin particles after dialysis was measured. The yields of the precipitations could not be determined due to too low sample amounts.

The hydrodynamic diameter of the particles stays relatively similar for each raw material between 160 ◦C and 200 ◦C, with only a slight trend upwards and diameters between 67.8 nm and 105.9 nm. However, the extracts from 220 ◦C yielded significantly bigger particles with diameters ranging from 152.8 nm for spruce to 1156 nm for beech (Figure 4), suggesting that the precipitation is strongly influenced by changes in the lignin occurring above 200 ◦C of pretreatment temperature.

**Figure 3.** Mw and PDI of the extracts from the different raw materials (**a**), molecular weight distribution of extracts from different raw materials at 180 ◦C (**b**), and molecular weight distribution of wheat straw at different pretreatment temperatures (**c**). Abbreviations: W—wheat straw, S—spruce, B—beech; the number next to the letter is the temperature of the corresponding pretreatment temperature.

**Figure 4.** Hydrodynamic diameters of lignin particles.

This leads back to the differences in molecular weight distribution found in the extracts for different pretreatment temperatures and raw materials. In Figure 5, the Mws of all

raw materials, pretreatment temperatures and fractions are presented. The results show that Mws of the lignin in the extracts are lower than those of the particles as well as of the particles after dialysis, but higher than those of the supernatants for all experiments. This supports the assumption that during the solvent shifting precipitation mainly lignin of high molecular weight is precipitated into particles while lignin of lower molecular weight remains in solution. Figure 6 showcases this fractionated precipitation for beech wood pretreated at 180 ◦C. Sipponen et al. [18] also found this molar-mass-fractionation when precipitating wheat straw soda lignin from aqueous ethanol by water addition and attributed it to electronic interactions of the aromatic lignin structures, which lead to lower solubility of larger lignin molecules.

**Figure 5.** Mw in the different process fractions for wheat straw, spruce, and beech.

**Figure 6.** Molecular weight distributions of extract, particles, particles after dialysis, and supernatant from beech wood experiments at 180 ◦C pretreatment temperature.

Interestingly, the Mws of the particles separated from the suspension directly after precipitation are consistently higher than those of the particles separated after dialysis (light and dark green columns in Figure 5). This indicates that some lignin with lower molecular weight is still in solution after precipitation but not removed by dialysis, despite a much higher membrane molecular weight cut-off. A possible explanation for this is that the gradual ethanol removal during dialysis lowers the lignin solubility. Smaller lignin molecules that are dissolved in the supernatant would precipitate onto the existing particles due to this, increasing their diameter and facilitating agglomeration. The results show that a portion of the lignin dissolved in the supernatant is removed by dialysis, since the Mw of the particles after dialysis is always higher than that of the extract.

The trend to lower Mws at higher pretreatment temperatures found in the extracts (Figure 3a) is also present and more pronounced for the particles (Figure 5), which coincides with the increase in particle size at higher temperatures (Figure 4). A possible explanation for this is that at higher pretreatment temperatures more lignin is depolymerized into smaller molecules which have a higher solubility than the larger molecules in the same extract. Because of the higher solubility of smaller lignin molecules, the degree of local supersaturation after the mixing would be lower for extracts produced at 220 ◦C, causing slower precipitation and larger particles [26]. Figure 7c shows this decrease of particle size with increasing Mw. Zwilling et al. [27] also found this increase of particle size with decreasing molecular weight, and connected it with more hydrophilic character of low molecular weight lignin. However, in this work lignin structure was not investigated.

**Figure 7.** Mw and PDI of lignin particles (**a**), molecular weight distribution of particles from different raw materials at 180 ◦C pretreatment temperature (**b**), Mw of particles over diameter of particles (**c**).

Comparison of the particles' molecular weight distributions for the different raw materials (Figure 7) shows that spruce wood extracts yield particles with lower Mws and PDIs than wheat straw or beech wood extracts. Figure 7b exemplifies this and additionally shows that the lignin particles from the woods have a similar distribution, while wheat straw results in a much broader distribution with more dominant peaks at 410 Da and 210 Da. At 220 ◦C, Mws and PDIs of all raw materials are within the same range and generally lower, suggesting that the lignin of all raw materials is depolymerized to a similar degree (Figure 7a). Comparing Figures 3a and 7a shows that the differences between the Mws or PDIs found for the particles are not present to the same extent in the respective extracts, which means that the precipitation amplifies differences between lignin from different raw materials. This might in turn have an influence on the particles' properties.

#### *3.3. Lignin Properties*

Since the particle size generally increased strongly for the precipitations from 220 ◦C extracts, and the precipitation did not give reasonable results for beech at that temperature, the color and antioxidant potential were only determined for lignin particles after dialysis from pretreatment temperatures 160 ◦C to 200 ◦C.

### 3.3.1. Color

For many of the possible commercial material applications of colloidal lignin particles, like in sunscreens or food packaging [10], it is important to consider the color of the produced lignin particles. For example, if lignin particles are being used in a product for their antioxidant activities to avoid color change in a product, a dark, brown particle color is obstructive to the application and a limiting factor to the maximum particle concentration applicable. Figure 8 shows the colors of equally thin layers of lignin particles on a membrane and the respective values in the CIELAB-color space. Higher pretreatment temperatures lead to darker lignin and stronger colors, demonstrated by the lowered brightness (L \*) and increased reddening (+a \*) and yellowing (+b \*) values in the spider graph, which can be considered undesirable. The darkening of the lignin particles can be attributed to increase of condensation reactions at higher process severities. The color changes most for wheat straw and least for beech wood with increasing pretreatment temperature. The ratio of red to yellow color component is similar for all lignins, except for wheat straw at 160 ◦C, where the yellow component is more dominant. Generally, through variation of raw material and pretreatment temperature a wide array of colors in the brown spectrum can be achieved for lignin particles. The influence of the particle size on the color found by other groups [28] could not be investigated with the given experimental plan, since too many other contributing factors were varied, and the investigated particles where of similar diameters.

**Figure 8.** Color of a 3 g/m2 layer of lignin particles (**a**) and affiliated components in CIELAB-color space (**b**).

#### 3.3.2. Antioxidant Activities

Before determining the antioxidant activities of the particles after dialysis, agglomerates and large particles were removed from the particle suspensions after dialysis by filtration, which resulted in suspensions with hydrodynamic diameters between 65.7 and 90.7 nm. Both methods used to determine the antioxidant activity (FRAP and ABTS) show similar trends (Figure 9): Higher pretreatment temperatures lead to lower antioxidant activity for beech, for spruce this effect is only witnessed when the temperature is increased

from 180 to 200 ◦C. For straw, the antioxidant activity does not significantly change with the temperature for ABTS, for FRAP it reaches a maximum at 200 ◦C. The antioxidant potential of the beech lignin particles is highest at 160 and 180 ◦C according to both methods, at 200 ◦C wheat straw reaches similar values. The divergence in the results between ABTS and FRAPS makes it difficult to compare the antioxidant potential of particles from spruce and wheat straw at 160 and 180 ◦C.

**Figure 9.** Antioxidant activities of particles after dialysis.

It is important to note that the antioxidant potential was measured for the particles (not solubilized lignin), and therefore is also affected by the surface area. Zhang et al. [29] investigated the antioxidant activities of lignin nanoparticles from corncob residue and found IC50 values between 101.2 and 296.6 mg/L. Due to the variation in raw materials and pretreatment temperatures combined with comparatively small differences in the particle diameters, the influence of the particle size on the antioxidant activity found in other works could not be investigated in this work.

#### **4. Conclusions**

The production and properties of colloidal lignin particles from organosolv extracts from different raw materials and the influence of the pretreatment temperature was investigated in the present work. All investigated raw materials (wheat straw, spruce wood, beech wood) were successfully applied for the production of colloidal particles with diameters below 110 nm whereas the origin of the lignin showed only minor influence in the size of the resulting particles.

Higher pretreatment temperatures increase the delignification of the raw materials but also favor depolymerization and structural alteration of the extracted lignin. This leads to increased particle sizes of over 110 nm and agglomeration at pretreatment temperatures of over 200 ◦C.

The investigation of the precipitation via solvent-shifting showed lignin fractionation by molecular weight, meaning that large lignin molecules precipitate while smaller lignin molecules stay in solution.

The resulting lignin nanoparticles were characterized on their application relevant properties of color and antioxidant potential. Higher pretreatment temperatures generally results in particles with darker colors and lower antioxidant activity especially for particles produced from beech and spruce. The antioxidant activity of the particles ranged from 19.1 and 50.4 mg Lignin/mg ASC equivalents.

From a process perspective, it is desirable to reach a high concentration of lignin in the extract, which can be achieved by higher pretreatment temperatures. However, this increase in process efficiency comes at the cost of lignin quality, which is demonstrated by the sharp increase in particle diameter, darker color of the resulting particles, and the decrease in antioxidant activity of the lignin particles. Therefore, optimal conditions and choice of raw material in a commercial lignin nanoparticle production process heavily depend on the final application.

**Author Contributions:** Conceived and designed the experiments, S.B. and J.A.; performed the experiments and analyzed the data, S.A., T.J., J.A., S.B., and F.Z.; wrote and edited the manuscript, J.A. with significant input and editing from A.F., J.L., S.B., and F.Z. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Acknowledgments:** The authors acknowledge TU Wien Bibliothek for financial support through its Open Access Funding Program.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


## *Article* **Carboxymethyl Bacterial Cellulose from Nata de Coco: Effects of NaOH**

**Pornchai Rachtanapun 1,2,3,\*, Pensak Jantrawut 2,4, Warinporn Klunklin 1, Kittisak Jantanasakulwong 1,2,3, Yuthana Phimolsiripol 1,2,3, Noppol Leksawasdi 1,2,3, Phisit Seesuriyachan 1,2, Thanongsak Chaiyaso 1,2, Chayatip Insomphun 1,2, Suphat Phongthai 1,2, Sarana Rose Sommano 3,5, Winita Punyodom 3,6, Alissara Reungsang 7,8,9 and Thi Minh Phuong Ngo <sup>10</sup>**


**Abstract:** Bacterial cellulose from nata de coco was prepared from the fermentation of coconut juice with *Acetobacter xylinum* for 10 days at room temperature under sterile conditions. Carboxymethyl cellulose (CMC) was transformed from the bacterial cellulose from the nata de coco by carboxymethylation using different concentrations of sodium hydroxide (NaOH) and monochloroacetic acid (MCA) in an isopropyl (IPA) medium. The effects of various NaOH concentrations on the degree of substitution (DS), chemical structure, viscosity, color, crystallinity, morphology and the thermal properties of carboxymethyl bacterial cellulose powder from nata de coco (CMCn) were evaluated. In the carboxymethylation process, the optimal condition resulted from NaOH amount of 30 g/100 mL, as this provided the highest DS value (0.92). The crystallinity of CMCn declined after synthesis but seemed to be the same in each condition. The mechanical properties (tensile strength and percentage of elongation at break), water vapor permeability (WVP) and morphology of CMCn films obtained from CMCn synthesis using different NaOH concentrations were investigated. The tensile strength of CMCn film synthesized with a NaOH concentration of 30 g/100 mL increased, however it declined when the amount of NaOH concentration was too high. This result correlated with the DS value. The highest percent elongation at break was obtained from CMCn films synthesized with 50 g/100 mL NaOH, whereas the elongation at break decreased when NaOH concentration increased to 60 g/100 mL.

**Keywords:** bacterial cellulose; biopolymer; carboxymethyl cellulose; CMC; nata de coco; sodium hydroxide

### **1. Introduction**

Primary cell walls of eukaryotic plants, algae and the oomycetes consist of cellulose as the major component. The cellulose consists of D-glucose units linked as a linear chain,

**Citation:** Rachtanapun, P.; Jantrawut, P.; Klunklin, W.; Jantanasakulwong, K.; Phimolsiripol, Y.; Leksawasdi, N.; Seesuriyachan, P.; Chaiyaso, T.; Insomphun, C.; Phongthai, S.; et al. Carboxymethyl Bacterial Cellulose from Nata de Coco: Effects of NaOH. *Polymers* **2021**, *13*, 348. https:// doi.org/10.3390/polym13030348

Academic Editor: Rafael Antonio Balart Gimeno Received: 25 December 2020 Accepted: 15 January 2021 Published: 22 January 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

ranging from several hundred to over ten thousand β (1→4) units [1]. Cellulose is insoluble in water. However, sodium carboxymethyl cellulose (CMC), which is one of cellulose's derivatives, can be dissolved in water [2]. CMC is an anionic, linear polymer, watersoluble cellulose reacted with monochloroacetic acid (MCA) or monochloroacetate acid (NaMCA). CMC is an important cellulose derivative applied in several industrial fields, such as the food industry, cosmetics, pharmaceuticals, detergents, textiles, [3] ceramics [4], etc. However, intensive utilization of wood has caused environmental problems and is expensive to manufacture. Therefore, there have been many studies about utilizing agricultural waste to be sources of CMC, such as cellulose from papaya peel [4], sugar beet pulp [5], sago waste [6], mulberry paper [7], *Mimosa pigra* peel [8] and durian husks [2,9], palm bunch and bagasse [10] and asparagus stalk ends [11]. The uses of CMC in food manufacturing require high purity of CMC grades ranging between 0.4 and 1.5 [12].

Cellulose has been obtained from bacterial cellulose (BC), including the genera *Agrobacterium*, *Rhizobium*, *Pseudomonas*, *Sarcina* and *Acetobacter* [13]. Acetobacter produced a pure bacterial cellulose aggregate containing impurities, such as hemicelluloses, pectin and lignin [14,15]. There are many desirable properties of bacterial cellulose, such as high purity, high degree of polymerization, high crystallinity, high wet tensile strength, high water-holding capacity and good biocompatibility [16,17]. *Acetobacter xylinum* is an acetic acid bacterium which can ferment and digest the carbon source in the coconut juice before converting it to the extracellular polysaccharide or cellulose [12]. The coconut juice is a waste product from the coconut milk industry known for its use in manufacturing cellulose [15,18]. Bacterial cellulose becomes a cellulosic white-to-creamy-yellow product called nata de coco. Therefore, the high intrinsic purity of bacterial cellulose from nata de coco together with the low environmental impact of the bacterium isolation means it is used in several applications, such as hydrogels [19], tissue engineering [20], cell culture [21], wound dressing and cancer therapy [22], CMC production [12], etc. However, the study of production and characterization of CMC films from nata de coco (CMCn) is limited.

This study aims to determine the effect of NaOH concentrations (20–60 g/100 mL) on the thermal properties, degree of substitution (DS), chemical structure, viscosity, crystallinity, and morphology of CMCn powder. With this aim, the effects of NaOH concentrations on solubility, mechanical properties (tensile strength, percentage elongation at break), percentage of transmittance, water vapor permeability (WVP) and the morphology of CMCn films were evaluated.

### **2. Materials and Methods**

#### *2.1. Materials*

Coconut juice was collected from Muang District, Chiang Mai Province, Thailand. *Acetobacter xylinum* came from the Division of Biotechnology, Faculty of Agro-Industry, Chiang Mai University, Thailand. Di-ammonium hydrogen orthophosphate was obtained from Univar, Williamson Rd Ingleburn, Australia; magnesium sulphate from Prolabo, England; and commercial grade citric acid and sugar powder from the Thai Roong Ruang Sugar Group, Thailand. Glacial acetic acid and sodium hydroxide (artificial grade) were purchased from Lab-scan; hydrochloric acid and sodium chloride were purchased from Merck, Darmstadt, Germany; m-cresol purple, indicator grade, from Himedia, Marg, Mumbai, India; chloroacetic acid from Sigma-Aldrich, Burlington, MA, USA; isopropyl alcohol (IPA), ethanol and absolute methanol were purchased from Union Science, Muang District, Chiang Mai, Thailand.

### *2.2. Preparation of Bacterial Cellulose*

Twenty liters of coconut juice were boiled in a pot and then 20 g of di-ammonium hydrogen orthophosphate, 10 g of magnesium sulphate, 30 g of citric acid and 100 g of sugar powder were added and mixed in. One and a half liters of the mixture were poured into each sterile plastic tray and then 55 mL of 95% *v*/*v* of ethanol was added. The mixture stood at room temperature for 30 min. One hundred milliliters of *Acetobacter xylinum* were

then added to each tray with an aseptic technique. The tray was covered with fabric and paper to protect it from any contaminants. Nata de coco was obtained after the mixture had stood at room temperature for 10 days without being disturbed. The nata de coco was sliced, boiled 5 times in a pot with 15 L of water and dried in a hot air oven (Model: 1370FX-2E, Sheldon Manufacturing Inc., Cornelius, OR, USA) at 55 ◦C for 12 h. It was then ground using a grinder (Philips-HR1701, Simatupang, Jakarta, Indonesia), and screened through a 0.5 mm mesh sieve (35 mesh size). The nata de coco bacterial cellulose powders were contained in propylene bags until used.

### *2.3. Synthesis of Carboxymethyl Cellulose from Bacterial Cellulose*

Fifteen grams of bacterial cellulose powder from nata de coco, NaOH solution (100 mL) in different concentrations (30, 40, 50 and 60 g/100 mL) and 450 mL of isopropanol (IPA) were mixed for 30 min. Eighteen grams of monochloroacetic acid (MCA) was added to initiate the carboxymethylation process and the mixture was stirred continuously at 55 ◦C for 30 min. The mixture beaker was enclosed by aluminum foil and stood in a hot air oven at 55 ◦C for 3.5 h. After the heating process, the mixture separated into solid and liquid phases. The solid phase was collected and suspended in 300 mL of absolute methanol, and then 80 mL/100 mL of acetic acid was added to neutralize it. The suspended mixture was filtered by a Buchner funnel. Undesirable by-products were removed from the final product by soaking 5 times in 300 mL of 70% *v*/*v* ethanol for 10 min. Then, the final product was washed again with 300 mL of absolute methanol. The carboxymethylated bacterial cellulose (CMCn) obtained from nata de coco was dried in an oven at 55 ◦C for around 12 h [8].

### *2.4. Degree of Substitution (DS)*

The degree of substitution (*DS*) of CMCn is defined as the average number of hydroxyl groups in the cellulose structure which have been replaced with carboxymethyl and sodium carboxymethyl groups at C2, 3 and 6. The DS determination was described in a crosscarmellose sodium monograph in USP XXIII. The method included two steps: titration and residue on ignition [8]. The DS is calculated using the following Equation (1):

$$DS = A + S \tag{1}$$

where *A* is the degree of substitution with carboxymethyl acid and *S* is the degree of substitution with sodium carboxymethyl. *A* and *S* were calculated using information from the titration and ignition steps using the following Equations (2) and (3):

$$A = \frac{1150M}{(7120 - 412M - 80C)}\tag{2}$$

$$S = \frac{(162 + 58A)C}{(7102 - 80C)}\tag{3}$$

where *M* is the *mEq* of the base required for endpoint titration. *C* is the percentage of ash remaining after ignition. The reported *DS* values are means of three repetitions.

### *2.5. Titration*

Precisely one gram of CMCn was weighed and added to a 500 mL Erlenmeyer flask. Three hundred milliliters of sodium chloride (NaCl) solution (10 g/100 mL) was added next. The Erlenmeyer flask was covered by a stopper and intermittently shaken for 5 min. The mixture was mixed with 5 drops of m-cresol purple, and then 15 mL of 0.1 N hydrochloric acid (HCl) was added. If the solution color did not change, 0.1 N HCl was gradually added until the solution changed to yellow. The solution was gradually titrated with 0.1 N NaOH. The solution changed from yellow to violet at the endpoint. The net amount of milliequivalent base required for neutralization of 1 g of CMCn (M) was calculated using Equation (4):

$$M(mEq) = mmol \times Valence\tag{4}$$

where *m* is 10<sup>−</sup>3, mole is mass in grams per molecular weight of NaOH and the valence of NaOH is 1.

### *2.6. Residue on Ignition*

A crucible was placed in an oven at 100 ◦C for 1 h and kept in a desiccator until the weight achieved a constant value (weighing apparatus, AR3130, Ohaus Corp. Pine, Brook, NJ, USA). Next, 1.000 g of CMCn was added to a crucible. To obtain black residue, a crucible containing CMCn was ignited in a 400 ◦C kiln (Carbolite, CWF1100, Scientific Promotion, Parsons Ln, Hope S33 6RB, England) for 1–1.5 h and then placed into a desiccator. The whole residue was wetted by adequate sulfuric acid and then heated until white fumes completely volatilized. White residue was obtained from the ignition crucible containing residue at 800 ± 25 ◦C and the residue was placed in desiccator to reach an accurate weight. All experiments were performed three times. The percentage of residue on ignition was calculated by following Equation (5).

$$\mathcal{C} = \frac{Weight\ of\ residue}{Weight\ of\ CMC} \times 100\tag{5}$$

### *2.7. Fourier Transform Infrared Spectroscopy (FTIR)*

An infrared spectrophotometer (Bruker, Tensor 27, Fahrenheitstr 4, D-28359, Bremen, Germany) was used to evaluate the functional groups of bacterial cellulose from nata de coco and CMCn. The bacterial cellulose from nata de coco and CMCn samples (~2 mg) with KBr were used to make pellets. Transmission was measured at a wave number range of 4000–400 cm<sup>−</sup>1.

### *2.8. X-Ray Diffraction (XRD)*

X-ray diffraction patterns of bacterial cellulose from nata de coco and CMCn were recorded in a reflection mode on a JEOL JDX-80-30 X-ray diffractometer, Brucker, Milton, ON, Canada. The scattering angle (2Ө) was from 5◦ to 60◦ at a scan rate of 2◦/min.

### *2.9. Viscosity*

The bacterial cellulose from nata de coco and CMCn were analyzed by a Rapid Visco Analyzer (Model: RVA-4, Newport Scientific, Warriewood, Australia) to measure viscosity. Three grams of cellulose and CMCn was dissolved in 25 mL of distilled water, heated to 80 ◦C and continuously stirred for 10 min. The sample solutions were tested in 2 steps. In the first step, the samples were tested at 960 rpm for 10 s. Next, the temperature was set at 30, 40 and 50◦ C at 5 min-intervals with a speed of 160 rpm. All tests were repeated three times [8].

### *2.10. Scanning Electron Microscopy (SEM)*

The morphology of bacterial cellulose from nata de coco, CMCn powder and CMCn films were analyzed using a scanning electron microscope (SEM) (Phillip XL 30 ESEM, FEI Company, Hillsboro, OR, USA) equipped with a large field detector. The acceleration voltage was 15 kV under low settings with 5000×.

### *2.11. Thermal Properties*

The thermal properties of the melting point of bacterial cellulose from nata de coco (CMCn) were determined using a differential scanning calorimeter (DSC) (Perkin Elmer precisely, Inst. Model Pyris Diamond DSC, Hodogaya-Ku, Godocho, Kanagawa, Japan). Aluminum pans containing 10 mg of samples were heated from 40 to 240 ◦C. A heat rate was set as 10 ◦C min−1. The tests were performed under N2 gas with a flow rate of 50 mL min<sup>−</sup>1. The data was represented in terms of a thermogram, which indicated the melting point. All samples were tested three times [8].

### *2.12. Film Preparation*

To obtain a film-forming solution, 3 g (Weighing apparatus, AR3130, Ohaus Corp. Pine, Brook, NJ, USA) of CMCn was dissolved in 300 mL of 80 ◦C distilled water with constant stirring for 15 min. Glycerol (15 g/100 g) was added and the film-forming solution was casted onto plates (20 cm × 15 cm). The plates containing the film-forming solution was placed at room temperature for 48 h to obtain dry CMCn films. Then, the CMCn films were removed from the plates. The CMCn films were kept at 27 ± 2 ◦C with 65 ± 2% relative humidity (RH) for 24 h (Thai industrial standard, TIS 949-1990).

### *2.13. Film Solubility*

The method to determine the percentage of solubility of CMCn films was modified from Phan et al. [23]. CMCn films were dried at 105 ◦C in a hot air oven for 24 h and kept in desiccators for 24 h. Then, about 0.2000 g of the CMCn films was weighed. This weight was recorded as an initial dry weight (*Wi*). Each weighed CMCn film was suspended in 50 mL of distilled water with shaking at 500 rpm for 15 min and poured onto filter paper (Whatman, No.93). The films then were dried in an oven at 105 ◦C for 24 h and weighed to obtain the final dry weight (*Wf*). The percentage of soluble matter (%SM) of the films was calculated by using the following Equation (6):

$$\% \text{SM} = \frac{\left(\mathcal{W}\_i - \mathcal{W}\_f\right)}{\mathcal{W}\_i} \tag{6}$$

#### *2.14. Percent Transmittance*

The percent transmittance of CMCn film was measured using a spectrophotometer (LaboMed, inc., Los Angeles, CA, USA) with a wavelength of 660 nm.

### *2.15. Mechanical Properties*

The CMCn films were cut into rectangular strips (1.5 cm × 14 cm) for tensile strength and percentage elongation measurements. To precondition them, CMCn films were kept for 24 h at 27 ± 2 ◦C and relative humidity (RH) was controlled in the range of 65 ± 2% (Thai Industrial Standard; TIS 949-1990). Film thickness was measured at five different location in each sample by a micrometer, model GT-313-A (Gotech testing machine Inc., Taichung Industry Park, Taichung City, Taiwan). Tensile strength and water vapor permeability (WVP) was calculated using average film thickness. The tensile strength (TS) and percentage elongation at break (EB) were measured using a Universal Testing Machine Model 1000 (HIKS, Selfords, Redhill, England) as per the ASTM Method (ASTM, D882-80a, 1995a). The upper grips were separated. The machine was operated at 100 mm and 20 mm min−<sup>1</sup> for the initial grip separation distance and crosshead speed, respectively. To calculate the TS, the maximum load at break was divided by the cross-sectional area of the sample. The EB was defined by the increasing of sample length at break point as the percentage of the initial length (100 mm). All mechanical tests were repeated 10 times.

### *2.16. Water Vapor Permeability (WVP)*

The method for measuring the water vapor permeability (WVP) of the CMCn films was described in the ASTM method (ASTM, E96-93, 1993). Circular aluminum cups (8 cm diameter and 2 m depth) containing 10 g of silica gel were covered by the circle shape of CMCn films (7 cm diameter). Then, the aluminum cups were closed by paraffin wax, weighed and kept in a desiccator. NaCl saturated solution was used to maintain the conditions (25 ◦C, 75% RH) in the desiccator. The closed cups were weighed daily for 10 days. A slope was obtained from plotting a graph between weight gain (Y axis) and time (X axis). The water vapor transmission rate (WVTR) was calculated using the following Equation (7):

$$\text{WVTR} \left( \frac{\text{g}}{\text{m}^2}.day \right) = \frac{slope}{\text{Fillm Area}} \tag{7}$$

where the film area was 28.27 cm<sup>2</sup> and the WVP (g.m/m2.mmHg.day) was calculated using the following Equation (8):

$$\left(\text{WVP}\left(\text{g.}\frac{m}{m^2}.day.mmHg\right) = \frac{\text{WVP}R}{\Delta P} \times L\right) \tag{8}$$

where *L* is the average film thickness (mm) and Δ*P* is the partial water vapor pressure difference (mmHg) across two sides of the film specimen (the vapor pressure of pure water at 23.6 ◦C = 21.845 mmHg). These samples were tested 3 times [8].

### *2.17. Statistical Analysis*

All the data were presented as the mean ± SD. One-way ANOVA was used to evaluate the significance of differences at the significance level of *p*-value < 0.05. Statistical analysis was performed using SPSS software version 16.0 (SPSS Inc., Chicago, IL, USA).

#### **3. Results and Discussion**

#### *3.1. Degree of Substitution (DS)*

The effects of NaOH concentrations on the DS of CMCn is shown in Figure 1. As NaOH concentrations increased from 20 to 30 g/100 mL NaOH, the DS of CMCn increased as well, while the DS dropped at 40 to 60 g/100 mL NaOH concentration. The values of the obtained CMCn were from 0.30–0.92. This phenomenon can be explained by the carboxymethylation procedure, where two competitive reactions occurred concurrently. A cellulose hydroxyl reacts with sodium monochloroacetate (NaMCA) to obtain CMCn in the first reaction, as shown in Equations (9) and (10).

$$\text{CLL-OH} + \text{NaOH} \rightarrow \text{CLL-ONa} + \text{H}\_2\text{O} \tag{9}$$

(Cellulose) (Reactive alkaline form)

$$\text{CLL-ONa} + \text{Cl-CH}\_2\text{-COONa} \rightarrow \text{CLL-O-CH}\_2\text{-COONa} + \text{NaCl} \tag{10}$$

$$\text{(CMC)}$$

NaMCA alternated to sodium glycolate as a byproduct by reacting with NaOH in the second reaction, as shown in Equation (11).

$$\text{NaOH} + \text{Cl-CH}\_2\text{COONa} \rightarrow \text{HO-CH}\_2\text{COONa} + \text{NaCl} \tag{11}$$

The second reaction overcame the first at stronger alkaline concentrations. In conditions of excessive alkalinity, DS was low because a side reaction dominates, causing the formation of sodium glycolate as a byproduct. This result agrees with Rachtanapun and Rattanapanone [8], who reported that degradation of CMC from *Mimosa pigra* occurred because of high concentrations of NaOH. Similar results were reported for the maximum DS value (0.87) of carboxymethyl cellulose from durian rind [2] and the maximum DS value (0.98) of carboxymethyl cellulose from asparagus stalk ends with a NaOH concentration of 30 g/100 mL [11].

**Figure 1.** Effect of the amount of NaOH on the DS of CMCn. CMCn: carboxymethyl bacterial cellulose powder from nata de coco. DS: degree of substitution.

#### *3.2. Percent Yield of Carboxymethyl Cellulose from Nata de Coco*

The effect of NaOH concentrations on percent yield of carboxymethyl bacterial cellulose from nata de coco (CMCn) powder synthesized is shown in Figure 2. This study investigated the effect of NaOH in concentrations of 20–60 g/100 mL on the percent yield of bacterial cellulose powder synthesis. At 20–40 g/100 mL NaOH concentrations, the percentage yield increased. However, the percentage decreased at 50–60 g/100 mL NaOH concentrations. This resulted from the alkali-catalyzed degradation of bacterial cellulose. The trend of percentage yield of CMCn was similar to the DS results (Figure 2). This result was in agreement with carboxymethyl cellulose from durian husks [2] and asparagus stalk ends [11].

**Figure 2.** Percent yield of carboxymethyl cellulose from nata de coco synthesized with various NaOH concentrations (20, 30, 40, 50 and 60 g/100 mL). The different letter, e.g., 'a', 'b', 'c' or 'd' are statistically different (*p* < 0.05).

#### *3.3. Fourier Transform Infrared Spectroscopy (FTIR)*

The substitution reaction in carboxymethylation was identified using FTIR. The FTIR spectrums of bacterial cellulose from nata de coco and CMCn synthesized with a 30 g/100 mL NaOH concentration are shown in Figure 3. The same functional groups appeared in both cellulose and CMCn. A broad peak at 3441 cm−<sup>1</sup> referred to the hydroxyl group (–OH stretching). A peak at 1420 cm−<sup>1</sup> was related to –CH2 scissoring. Strong peaks at 1604 cm−<sup>1</sup> and 1060 cm−<sup>1</sup> indicated carbonyl group (C=O stretching) and ether groups (–O– stretching), respectively [8]. In the CMCn sample, the obvious increase of the carbonyl group (C=O), methyl group (–CH2) and ether group (–O–) peaks was observed, but the intensity of the hydroxyl

group (–OH) was lower when compared to bacterial cellulose (Figure 3). This result indicated that cellulose molecules changed due to carboxymethylation occurrence [2,6,10,11,17].

**Figure 3.** Fourier transform infrared spectroscopy of (**a**) bacterial cellulose from nata de coco and (**b**) CMCn synthesized with 30 g/100 mL NaOH.

### *3.4. Effect of Various NaOH Concentrations on Viscosity of CMCn*

The effect of using various NaOH concentrations on the viscosity of CMC is shown in Figure 4. The peak viscosity of CMCn increased when the concentration of NaOH increased from 20 to 30 g/100 mL because hydroxyl groups on the cellulose molecule were substituted by more carboxymethyl groups, being hydrophilic groups. Cohesive forces reduced because CMCn temperature increased, whereas the rate of molecular interchange concurrently increased [8,24]. The viscosity of CMCn solution was in accordance with the DS value at the same temperature. A previous study found that NaOH concentrations influenced the viscosity of CMC [25]. They stated that more carboxymethyl groups (a higher DS value) increased CMC viscosity [8]. In addition, they reported that the degradation of CMC polymers causes decreasing viscosity with too much NaOH, leading to a lower DS [26].

**Figure 4.** Effect of NaOH concentrations on the viscosity of bacterial cellulose and CMCn.

### *3.5. Effects of Various NaOH Concentrations on Thermal Properties of CMCn Powder*

The effects of various NaOH concentrations on the thermal properties of bacterial cellulose from nata de coco and CMCn powder are shown in Figure 5. The melting temperature ™ of bacterial cellulose was 167.4 ◦C, and those of CMCn synthesized with 20, 30, 40, 50 and 60 g/100 mL NaOH were 175.3, 188.2, 179.4, 164.7 and 151.0 ◦C, respectively. The melting temperature of CMCn was increased at 20–30 g/100 mL because the substituent of carboxymethyl groups caused increases in the ionic character and intermolecular forces between the polymer chains. At 40–60 g/100 mL NaOH, the melting temperature decreased because of the side reaction predominating with the formation of sodium glycolate as a byproduct and chain breaking of the CMCn polymer. This result was in agreement with carboxymethyl cellulose from asparagus stalk ends [11].

**Figure 5.** Differential scanning calorimetry of cellulose from nata de coco and CMCn synthesized with various amounts of NaOH.

### *3.6. X-Ray Diffraction (XRD)*

The effect of NaOH concentrations on the amount of crystallinity in bacterial cellulose from nata de coco and CMCn powder is shown in Figure 6. The treatment of cellulose with NaOH caused a decrease in the amount of cellulose crystallinity because NaOH leaded to the dividing of hydrogen bond [18]. By comparison with cellulose without alkalization with NaOH, monochloroacetic acid molecules substituted into bacterial cellulose molecules more easily because the distance between each polymer molecule increased. Thus, before the carboxymethylation reaction, NaOH had an effect on cellulose structured to decrease the crystallinity of the CMCn. These results were also found in the reduction of crystallinity of CMC from asparagus stalk ends [11] and cavendish banana cellulose [26] due to alkalizing by 20 g/100 mL and 15 g/100 mL, respectively.

**Figure 6.** X-ray diffractograms of cellulose from nata de coco and CMCn synthesized with various amounts of NaOH.

### *3.7. Scanning Electron Microscopy of Cellulose from Nata de Coco and CMCn Powder*

The scanning electron micrographs of bacterial cellulose from nata de coco and CMCn powder with 20, 30, 40, 50 and 60 g/100 mL NaOH are shown in Figure 7a–f. The micrograph of bacterial cellulose from nata de coco (Figure 7a) exhibited a compact appearance with a smooth surface without pores or cracks, and a lot of small fibers appeared. However, the microstructure of samples dramatically changed when the NaOH concentration increased. The morphology of CMCn powder with 20 g/100 mL NaOH (Figure 7b) showed small fibers with minimal damage. As shown in Figure 7c, the morphology of CMCn powder with 30 g/100 mL NaOH was compact and dense, with no signs of cracks and pits. The morphology of CMCn powder with 40 g/100 mL NaOH (Figure 7d) was deformed and had some cracks and pits. Moreover, the morphology of CMCn powder with 40 g/100 mL NaOH also correlated with the DS value, due to a chain degradation of the CMC polymer. This result is similar to those of other studies [2,8,11]. CMCn powder with 50 g/100 mL NaOH (Figure 7e) showed increased surface irregularities with a more indented and collapsed surface. Moreover, when increasing NaOH concentration to 60 g/100 mL (Figure 7d), the surface became rougher and totally deformed. Thus, it can be concluded that synthesis with increasing NaOH concentrations damages the surface area of bacterial cellulose powder. This result was similar to those of carboxymethyl rice starch [27] and carboxymethyl cassava starch [28] and carboxymethyl cellulose from asparagus stalk ends [11]. The principal parameter for weakening the structure of cellulose and causing a loss of crystallinity that allowed etherifying agents to reach the cellulose molecules for carboxymethylation processes was the alkalization [27]. Consequently, this result was consistent with the DS value. To indicate carboxymethyl reactions, the color formation was investigated by color measurement. The main effects on CMC color value were increasing a\* (redness), b\* (yellowness) and yellowness index (YI) values as NaOH concentrations increased (20–30 g/100 mL NaOH). The L\* value of CMCn synthesized with various NaOH concentrations decreased as the concentrations increased from 20 to 30 g/100 mL NaOH. The a\* and b\* values decreased, but the L\* value of CMCn increased as NaOH was increased to 40 g/100 mL NaOH. Moreover, the trend of YI of CMCn was decreased when increasing the NaOH concentration up to 40%. This phenomenon was probably due to the first step in the carboxymethylation of cellulose (Equation (9)), which delivered CMC or sodium glycolate [10]. At a high NaOH concentration (50 and 60 g/100 mL NaOH), all color values were decreased. A carboxymethylation reaction might have been a reason for the color change in this study [2,11,27]. The ΔE of cellulose and CMCn had same trends as the a\* and b\* values. The WI of cellulose and CMCn had the same trends as the L\* value (Table 1).


**Table 1.** Color values of bacterial cellulose from nata de coco and CMCn synthesized with various amounts of NaOH.

Note: Obtained by Duncan's test (*p* < 0.05). L\* = lightness, a\* = redness, b\* = yellowness, ΔE = total color difference, YI = yellowness index, WI = whiteness index, hab = hue. The different letter, e.g., 'a', 'b', 'c' or 'd' are statistically different (*p* < 0.05).

**Figure 7.** Scanning electron micrographs of (**a**) cellulose from nata de coco and CMCn powder: with (**b**) 20 g/100 mL NaOH, (**c**) 30 g/100 mL NaOH, (**d**) 40 g/100 mL NaOH, (**e**) 50 g/100 mL NaOH and (**f**) 60 g/100 mL NaOH.

### *3.8. Solubility and Transmittance*

The effects of NaOH concentrations on the solubility of CMCn films are shown in Figure 8. At 20–30 g/100 mL NaOH, the solubility of CMCn film increased; however, at 40–60 g/100 mL NaOH, the film solubility decreased. The percentage of soluble matter could be indicated by the DS. The percentage of soluble matter rose as the DS value increased [6]. The effects of NaOH concentrations on the percent transmittance of the CMCn films are shown in Figure 9. The percent transmittance increased with NaOH concentrations of 20–30 g/100 mL and decreased at 40–60 g/100 mL. This result correlated with the percentage of solubility of the CMCn films.

**Figure 9.** Effect of NaOH concentrations on the percentage of transmittance of the CMCn films. The different letter, e.g., 'a', 'b', 'c' or 'd' are statistically different (*p* < 0.05).

#### *3.9. Scanning Electron Microscopy of Cellulose from Nata de Coco and CMCn Films*

Cross-section views of scanning electron micrographs of CMCn films synthesized with various NaOH concentrations are shown in Figure 10a–e. As shown in Figure 10a, the morphology of CMCn films with 20 g/100 mL NaOH was a rather smooth surface without pits and cracks. CMCn films with 30 g/100 mL NaOH (Figure 10b) were compact and dense. The morphology of CMCn films synthesized with 40 g/100 mL NaOH (Figure 10c) was still compact but rougher and some scraps protruded on the surface. CMCn films with 50 g/100 mL NaOH (Figure 10d) showed increasing surface irregularities and defects such as cracks and pits. The morphology of CMCn films with 60 g/100 mL NaOH (Figure 10e) was deformed and revealed lots of cracks and some scraps.

**Figure 10.** Scanning electron micrographs of CMCn films: with (**a**) 20 g/100 mL NaOH, (**b**) 30 g/100 mL NaOH, (**c**) 40 g/100 mL NaOH, (**d**) 50 g/100 mL NaOH and (**e**) 60 g/100 mL NaOH.

### *3.10. Water Vapor Permeability (WVP)*

The effects of NaOH concentrations on the water vapor permeability (WVP) of the CMCn films are shown in Figure 11. When the concentration of NaOH increased, cellulose transformed to CMCn, causing a higher polarity. This result led to a rise of the WVP. Moreover, the morphology of CMCn films with 40–60 NaOH showed cracks and pits. Damage dramatically increased in the WVP of CMCn with 50–60 g/100 mL NaOH. This result can be explained through a SEM micrograph (Figure 10). SEM morphology demonstrated that CMCn films with 40–60 NaOH showed cracks and pits. The results were in agreement with studies of CMC from rice starch [24], which showed that the WVP of carboxymethyl rice starch films rose due to increases in NaOH concentrations. Furthermore, carboxylation influenced increases of polarity, reductions in crystallinity and changes in granule morphology [29].

**Figure 11.** Effect of NaOH concentration (20, 30, 40, 50 and 60 g/100 mL) on the water vapor permeability (WVP) of CMCn films. The different letter, e.g., 'a', 'b', 'c' or 'd' are statistically different (*p* < 0.05).

#### *3.11. Tensile Strength (TS) and Elongation at Break (%)*

The tensile strength values of CMC films with various NaOH concentrations are shown in Figure 12a. When NaOH increased (20–30 g/100 mL NaOH), the tensile strength also rose. This is because TS values are correlated with increases in DS values because of the carboxymethyl group substitution, causing a rise in the ionic character and intermolecular forces between the polymer chains [23]. Moreover, this is related to why the morphology of CMCn film synthesized with 30 g/100 mL NaOH was compact and dense. However, at higher NaOH concentration, the TS dropped because of sodium glycolate, which is a secondary product from the CMC synthesis reaction and polymer degradation. These results are related to those of CMC films from *Mimosa pigra* [8], which shows that the TS of CMC films from *Mimosa pigra* rose when concentrations of NaOH increased (20–30 g/100 mL). Studies of CMC films from mulberry paper waste showed that the TS increased with increasing NaOH concentrations, but too-high NaOH concentrations (60 g/100 mL) could cause a hydrolysis reaction in the cellulose chain [7]. The percentage of elongation at break (EB) of CMCn films with various NaOH concentrations used in CMCn synthesis are shown in Figure 12b. CMCn films exhibited higher EB when NaOH concentrations increased (20–50 g/100 mL NaOH). However, when NaOH concentration was increased to 60 g/100 mL, the EB of CMCn films decreased. This phenomenon can be explained as a cellulose structure with decreased crystallinity caused the CMCn films to have increased elasticity due to higher concentrations of NaOH. Nevertheless, at 60 g/100 mL NaOH, the occurrence of cellulose hydrolysis resulted in lower flexibility of CMCn films [8].

**Figure 12.** Effect of various NaOH concentrations (0, 20, 30, 40, 50 and 60 g/100 mL) on (**a**) tensile strength and (**b**) percent elongation at break of CMCn films. The different letter, e.g., 'a', 'b', 'c' or 'd' are statistically different (*p* < 0.05).

#### **4. Conclusions**

From the results, this study was successful in using nata de coco as a resource to obtain carboxymethyl bacterial cellulose (CMCn) using different NaOH concentrations. This demonstrates that the major parameter relating to the properties of CMCn is NaOH quantity. The DS of CMCn dramatically increased when NaOH concentrations increased from 20 to 30 g/100 mL in CMCn synthesis, and the DS value gradually dropped at NaOH concentrations of 40–60 g/100 mL. The mechanical properties and viscosity of CMCn films were correlated with the DS. CMCn consisting of high DS can have improved mechanical properties. Therefore, the CMCn films synthesized with 30 g/100 mL NaOH exhibited good mechanical properties, such as high TS and EB with no evidence of cracking and pitting when examined by SEM. CMC is widely applied in several fields, including in the food, nonfood and pharmaceutical industries, and thus the results from this study are useful. The investigation of methods to modify cellulose in this study, such as varying the NaOH content, supports new and potentially useful applications of cellulose from nata de coco.

**Author Contributions:** P.R. conceived and planned all experiments; conceptualization, investigation and methodology, P.R.; C.I., S.R.S. and T.M.P.N. performed the measurements; W.K., K.J., N.L., P.S., T.C., C.I., S.R.S., T.M.P.N. and P.R. wrote the manuscript with input from all authors; W.K., S.P., P.J., W.P. and A.R. analyzed the data; K.J. and N.L. interpreted the results; Y.P. and P.R. contributed reagents and materials; validation, P.R.; project supervision, P.R. All authors discussed the results, commented, and agreed to publish this manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** We wish to thank the Center of Excellence in Materials Science and Technology, Chiang Mai University, for financial support under the administration of the Materials Science Research Center, Faculty of Science, Chiang Mai University. This research was supported by the Program Management Unit for Human Resources & Institutional Development, Research and Invitation, NXPO [Frontier Global Partnership for Strengthening Cutting-edge Technology and Innovations in Materials Science, and Faculty of Science (CoE64-P001) as well as the Faculty of Agro-Industry (CMU-8392(10)/COE64), CMU. The present study was partially supported by The Thailand Research Fund (TRF) Senior Research Scholar (Grant No. RTA6280001). This research work was also partially supported by Chiang Mai University.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Acknowledgments:** We wish to thank the Center of Excellence in Materials Science and Technology, Chiang Mai University, for financial support under the administration of the Materials Science Research Center, Faculty of Science, Chiang Mai University. This research was supported by the Program Management Unit for Human Resources & Institutional Development, Research and Invitation, NXPO [Frontier Global Partnership for Strengthening Cutting-edge Technology and Innovations in Materials Science, and Faculty of Science (CoE64-P001) as well as the Faculty of Agro-Industry (CMU-8392(10)/COE64), CMU. The present study was partially supported by The Thailand Research Fund (TRF) Senior Research Scholar (Grant No. RTA6280001). This research work was also partially supported by Chiang Mai University.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


## *Article* **Improving the Tensile and Tear Properties of Thermoplastic Starch/Dolomite Biocomposite Film through Sonication Process**

**Azlin Fazlina Osman 1,2,\*, Lilian Siah 1, Awad A. Alrashdi 3, Anwar Ul-Hamid <sup>4</sup> and Ismail Ibrahim 1,2**


**Abstract:** In this work, dolomite filler was introduced into thermoplastic starch (TPS) matrix to form TPS-dolomite (TPS-DOL) biocomposites. TPS-DOL biocomposites were prepared at different dolomite loadings (1 wt%, 2 wt%, 3 wt%, 4 wt% and 5 wt%) and by using two different forms of dolomite (pristine (DOL(P) and sonicated dolomite (DOL(U)) via the solvent casting technique. The effects of dolomite loading and sonication process on the mechanical properties of the TPS-DOL biocomposites were analyzed using tensile and tear tests. The chemistry aspect of the TPS-DOL biocomposites was analyzed using Fourier transform infrared spectroscopy (FTIR) and X-Ray Diffraction (XRD) analysis. According to the mechanical data, biocomposites with a high loading of dolomite (4 and 5 wt%) possess greater tensile and tear properties as compared to the biocomposites with a low loading of dolomite (1 and 2 wt%). Furthermore, it is also proved that the TPS-DOL(U) biocomposites have better mechanical properties when compared to the TPS-DOL(P) biocomposites. Reduction in the dolomite particle size upon the sonication process assisted in its dispersion and distribution throughout the TPS matrix. Thus, this led to the improvement of the tensile and tear properties of the biocomposite. Based on the findings, it is proven that the sonication process is a simple yet beneficial technique in the production of the TPS-dolomite biocomposites with improved tensile and tear properties for use as packaging film.

**Keywords:** thermoplastic starch; dolomite; biocomposite; mechanical properties; sonication

## Published: 15 January 2021

**Citation:** Osman, A.F.; Siah, L.; Alrashdi, A.A.; Ul-Hamid, A.; Ibrahim, I. Improving the Tensile and Tear Properties of Thermoplastic Starch/Dolomite Biocomposite Film through Sonication Process. *Polymers* **2021**, *13*, 274. https://doi.org/ 10.3390/polym13020274 Received: 9 December 2020 Accepted: 8 January 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

**1. Introduction**

Nowadays, petroleum-based plastics are daily and widely used packaging materials. Unfortunately, this contributes negative effects to the environment as a result of the long degradation time of plastics leading to waste disposal problems [1]. Scientifically, the molecular structure and chemical bonding of the petroleum-based plastic caused them to be too durable and resistant to the natural process of degradation. For example, polyolefins such as polypropylene (PP) and polyethylene (PE) are very resistant to hydrolysis and totally non-biodegradable. Consequently, this causes a large accumulation of plastic wastes in the landfill, causing a serious waste management problem and soil contamination. Moreover, the toxic chemicals present in the plastic will also cause pollution to the environment. When plastics are burned, they emit dioxin, which is the most toxic substance that may harm the health, reproduction, development, immune systems, disrupt hormones, cause cancer, and create other issues [2]. To reduce those effects, recycling of the plastic wastes is performed. However, several problems may arise during the recycling process, especially

during the collection, separation, and cleaning, due to possible contamination on the plastics and difficulty of finding an economically viable outlet, where incineration may give off some toxic gas [3].

Due to the above-mentioned problems, effort to replace the synthetic plastics with the natural-based plastics is gaining more traction from year to year. The development of new plastic materials from natural and renewable resources has been the object of intensive academic and industrial research [4–6]. Starch-based biodegradable plastics are an ideal replacement to petroleum-based plastics. Even though these biopolymers cannot replace synthetic polymers in every application, at least they can be used to produce specific products, normally for application in which the recovery of plastic is not viable, economically feasible, and controllable such as in one-time use plastic [4–7].

Thermoplastic starch (TPS) can be obtained through plasticization of the native starch granules. TPS can be used to form biodegradable biocomposites films for packaging application. However, TPS-based products usually have several disadvantages. They may exhibit poor mechanical performance due to severe issues of water sensitivity, environmental stress cracking, and high fragility over time. Their low mechanical properties are a major concern, especially when exposed to hydrolytic and oxidative conditions and elevated temperatures [4,8]. These restrict the use of the TPS packaging films for long-term application as opposed to the petroleum-based plastic. In order to improve the properties, the TPS needs to be reinforced with fillers. In this research, a mineral filler, which is dolomite (DOL), was used to reinforce the TPS film. In this context, dolomite particles dispersion in the TPS matrix is very important, since it can affect the mechanical properties of the resultant TPS-DOL biocomposite films. Previous research indicated that a good dispersion of filler leads to improvement in the mechanical properties of the TPS composites, such as tensile and tear strength [4,8]. Generally, the pristine dolomite particles exist in agglomerated form. This may hinder their homogenous dispersion throughout the TPS matrix. The poor dispersion of dolomite can inhibit their reinforcing capability; thus, biocomposites films with poor properties may be produced [9]. Due to these reasons, dolomite needs to undergo a pre-dispersion process to diminish particle agglomeration, thus ensuring a good dispersion of dolomite in the TPS matrix. According to previous research, sonication is a practical method for the pre-dispersion process of the particles and nanoparticles [10]. Thus, this method was employed in this study to obtain a good dispersion of dolomite in the TPS matrix.

To the best of our knowledge, no research related to TPS/dolomite biocomposite film has been conducted and reported. Our research group is the first to study and investigate the TPS/dolomite biocomposite system for potential use as packaging film. Both starch and dolomite are abundant sources of materials to be transformed into biocomposite or other products with several beneficial applications. Thus, research in this area is considered impactful toward sustainability. Of particular interest, an optimization study on the TPS/dolomite biocomposite system can be the first step to formulate the biocomposite with the best property profiles. Therefore, the objective of this study was to investigate the effects of dolomite loading (0, 1, 2, 3, 4, and 5 wt%) on the tensile and tear properties of the TPS/dolomite biocomposite film. The novelty of this work is the use of a simple technique for pre-dispersing the dolomite, which is tip sonication process to improve dispersion of dolomite particles in the TPS matrix. This technique involved no harmful chemicals, as the medium used is only water. Furthermore, the application of this predispersion technique allows the enhancement in both tensile and tear properties of the TPS biocomposite incorporating dolomite. This article highlights all of these aspects.

### **2. Materials and Chemicals**

Corn starch powder, dolomite filler, glycerol, and sodium bicarbonate have been used to produce the virgin TPS film and TPS-DOL biocomposites film. Corn starch was manufactured by Sigma-Aldrich (St. Louise, MO, USA) with code number S4126 in 2 kg poly bottle packing. It was supplied by Euroscience Sdn. Bhd (Kuala Lumpur, Malaysia)

in white powder form. It consists of about 27% amylose and 73% amylopectin. Dolomite was kindly supplied by Perlis Dolomites Industries Sdn. Bhd (Padang Besar, Malaysia) and is known by the locals as 'batu reput'. It came in powder form with a particle size of 150 μ. Dolomite was used as filler in the TPS biocomposite film in pristine and sonicated form. Glycerol with a minimum concentration of 99.5% was manufactured by HmbG Chemicals (Hamburg, Germany) with code number C0347-91552409. This chemical was purchased through A.R. Alatan Sains Sdn Bhd (Alor Setar, Malaysia). Glycerol served as a plasticizer in the formulation of TPS film. Sodium bicarbonate was manufactured by HmbG Chemicals (Hamburg, Germany) with code number C0725-2134330 and supplied by A.R. Alatan Sains Sdn Bhd (Alor Setar, Malaysia). It acted as a thickener in the production of TPS film.

### *2.1. Preparation of Dolomite (DOL(P)) and Ultrasonicated Dolomite (DOL(U)*

Dolomite powder having a particle size of 150 μ was ground into finer and smaller particle size by using a mortar and pestle. After that, it was sieved by using a 50 μ sieve with 100 mesh. The resultant powder is called pristine dolomite (DOL(P)).

The sonication of dolomite was done as a pre-dispersion process prior to biocomposite production to assist dolomite dispersion and distribution in the TPS matrix. The suspension of dolomite was prepared by mixing the powder form of dolomite with distilled water in a ratio of 1:100. The dolomite was sonicated by using a Branson Digital Ultrasonic Disrupter/Homogenizer, Model 450 D with 20% amplitude, 20 s pulse on, and 10 s pulse off for 120 min. The resultant sonicated dolomite is referred to as DOL(U).

### *2.2. Plasticization of Starch to Form Thermoplastic Starch (TPS)*

Thermoplastic starch (TPS) was obtained through the combination of corn starch with plasticizers under a stirring and heating process. Firstly, distilled water, corn starch, and glycerol were combined in a mass ratio of 100:5:2 in one beaker. The amount of water, corn starch, and glycerol used to produce one film of TPS in a Teflon mold were 100 mL, 5 g, and 2 g, respectively. Next, the mixture was stirred continuously in a water bath under a temperature range of 75–85 ◦C by using a magnetic stirrer (500 rpm). This procedure took about 45 min or until the mixture turned into paste.

#### *2.3. Preparation of TPS-DOL Biocomposites*

The preparation of the TPS-DOL biocomposites films was done through the solution casting and drying process method. The suspension of dolomite (sonicated) was added into the TPS suspension and keep stirred in the range between 75 and 85 ◦C. Next, 0.5 g of sodium bicarbonate was dissolved in 30 mL of distilled water and then added into the TPS suspension, which was continuously stirred for another 5 min. After that, the biocomposite suspension was poured into Teflon-coated mold and put into the oven at 45 ◦C for 24 h for drying. The dried film of the TPS-DOL biocomposite is shown in Figure 1. It was removed from the Teflon-coated mold and stored in a desiccator to avoid moisture while conditioning it in standard environment prior to testing. The process was repeated for pristine dolomite (without sonication) to form control biocomposites films. All the film samples were prepared in the same volume and using the same-size molds. Therefore, all samples had almost similar thickness.

**Figure 1.** Dried film of the thermoplastic starch-dolomite (TPS-DOL) biocomposite in the Tefloncoated mould.

The formulations of TPS-DOL biocomposite films are shown in Table 1. In preliminary works, we have prepared the TPS biocomposite films with dolomite loading greater than 5 wt% (7 and 9 wt%). However, no good films were observed, which was perhaps due to agglomeration and the overcrowding of dolomite particles in the matrix. The films were brittle and cracked after being cut into dumbbell shape. Furthermore, the samples took a longer time for drying. Based on these observations, we did not proceed for testing.


**Table 1.** The formulation of thermoplastic starch-dolomite (TPS-DOL) biocomposite films with their respective acronyms.

The overall experimental procedures to produce the TPS-DOL biocomposite films are summarized in Figure 2.

**Figure 2.** Flow chart of experimental procedures.

### **3. Characterization and Mechanical Testing**

The tensile test, tear test, and FTIR analysis were conducted to test the properties and chemical structure of the TPS film and TPS-DOL biocomposite films.

### *3.1. Fourier Transform Infrared Spectroscopy (FTIR)*

By using a Perkin Elmer RXI FTIR spectrophotometer (Waltham, MA, USA), the FTIR spectra of all the samples were analyzed. The chemical functional groups of TPS-DOL biocomposites (incorporating the DOL(P) and DOL(U)) were characterized by FTIR. The samples were recorded for 32 scans in the frequency range of 650–4000 cm−<sup>1</sup> wavenumber with a resolution of 32 cm<sup>−</sup>1.

#### *3.2. X-ray Diffraction (XRD)*

X-ray diffraction (XRD) was used to analyze the structure of dolomite before and after the sonication process. The analysis was also performed on the TPS film and TPS-DOL biocomposite films incorporating the DOL(P) and DOL(U) to determine their crystallinity. Each sample was placed in sample holder and scanned over a range of 2θ = 10◦–35◦ and with a scan speed of 2◦/min.

The degree crystallinity of corn starch, TPS, DOL(P), DOL(U), and TPS-DOL biocomposite films was evaluated by using the formula below [11].

$$\text{Degree of Crystallinity} = \frac{(\text{total area of cristalline peaks})}{(\text{total area of all peaks})} \times 100\%... $$

### *3.3. Tensile Test*

The tensile properties such as the tensile strength, elongation at break, and Young's modulus of the TPS-DOL biocomposite films were determined through a tensile test using the Instron Machine 5569 (Norwood, MA, USA), according to angle test specimen (ASTM D 882) [12]. The test specimens were cut into dumbbell shape according to an ASTM D 638 Type V die cutter [13]. Details are shown in Figure 3 and Table 2.

**Figure 3.** Schematic drawings of dumbell specimen according to ASTM D638 Type 5 [13].


**Table 2.** Dimension of a dumbbell specimen (ASTM D638 Type V).

The samples were tested with a crosshead speed of 10 mm/min. Five replicates of each sample were tested, and the average values of tensile strength, elongation at break, and Young's modulus were recorded.

### *3.4. Tear Test*

The tear strength of the TPS-DOL biocomposite films was determined by using an Instron Machine 5569. The samples were cut according to angle test specimen (ASTM D624 Type C) using a wood-base laser cutter [14]. Dimensions of the tear test specimen are shown in Figure 4 The test was run at a crosshead speed of 500 mm/min. Three replicates of each biocomposite film were tested and average values of tear strength were taken.

**Figure 4.** Schematic drawings of tear test specimen according to ASTM D624 Type C [14].

Setup for tensile and tear tests is shown in Figure 5. Both testings were carried out using the same Instron machine.

**Figure 5.** Setup for tensile and tear tests: (**a**) dumbbell-shaped samples for tensile test, (**b**) sample loading for tensile test, (**c**) Instron machine model 5569 to perform tensile and tear tests, (**d**) samples for tear test, (**e**) sample loading for tear test.

### **4. Results and Discussion**

### *4.1. Functional Group Analysis of the Dolomite Filler (DOL(P) and DOL(U))*

Fourier-transform infrared spectroscopy (FTIR) analysis was conducted for DOL(P), DOL(U), TPS, and TPS-DOL biocomposite films. Figure 6 illustrates the FTIR spectra of DOL(P) and DOL(U).

**Figure 6.** Fourier transform infrared spectroscopy (FTIR) spectra of pristine dolomite (DOL(P)) and sonicated dolomite (DOL(U)).

Generally, the FTIR spectra of DOL(P) indicate that the most prevalent component present in dolomite is carbonate [15]. The carbonate mineral group (magnesium and calcium carbonate) illustrates the strong peak at 1420 cm−1, which is associated with lattice CO3 <sup>2</sup><sup>−</sup> [16]. On the other hand, the other main components of dolomite can be seen through the appearance of spectral peaks around 3000 and 719 cm−<sup>1</sup> [15]. The FTIR absorption at around 3000 cm−<sup>1</sup> relates to organic residue, while the band at around 2870 cm−<sup>1</sup> represents the calcite combination band, and the peak at 719 cm−<sup>1</sup> is attributed to magnesian calcite or calcite [16,17]. In agreement with our result, Chen et al. also stated that there are few FTIR absorptions at 1420, 873, and 719 cm−1, which are verified as the presence of dolomite [17]. DOL(U) refers to the dolomite, which has underwent a 2 h sonication process. It is observable that the features of the FTIR spectra of DOL(P) and DOL(U) are quite similar. There are only slight changes in the peak position of the overall spectra, which suggest that there is no structural and chemical modification of the dolomite when subjected to the sonication process. However, there are more intense peaks around 1418, 872, and 719 cm−<sup>1</sup> for DOL(U) when compared with DOL(P). This could be due to the smaller particle size but larger surface area of dolomite after it was subjected to the sonication process. It is widely understood that the infrared (IR) spectroscopy involves the analysis on how the infrared light interacts with matter. Therefore, the concentration of molecules in the tested sample affects the peak intensity in the IR spectra [18]. Particles with higher surface area may allow penetration of the IR light into a greater number of molecules. As a result, more intense peaks can be observed. Note that the aim of conducting the sonication process in this research was to break the dolomite particles into smaller and finer size so that the dolomite filler can be more efficiently dispersed throughout the TPS matrix. This can enhance the tensile and tear properties due to better interaction between dolomite filler and the TPS matrix.

### *4.2. Functional Groups Analysis on TPS Film and TPS-DOL Biocomposite Films*

Figure 7 indicates the FTIR spectra of the TPS film, TPS-5wt%DOL (P) and TPS-5wt%DOL(U) biocomposite films. It can be observed that the FTIR signal of the TPS film indicates a strong spectral band in the range of 600 to 1500 cm−1, which confirms the fingerprint of the starch material [5]. TPS film displays the characteristics FTIR absorptions at 3265 and 2925 cm−1, which correspond to the O-H and C-H stretching, respectively. The appearance of the peak at 1645 cm−<sup>1</sup> was due to the hydrophilic characteristics of the

TPS itself, as it can be tightly bound with water [19]. The spectral peak at 1140 cm−<sup>1</sup> is contributed to the C-O-H bond. Within the region below 860 cm<sup>−</sup>1, the spectrum of the TPS film reveals the complex vibrational mode, which was likely due to the skeletal vibrations of the pyranose ring in the glucose unit [19].

**Figure 7.** FTIR spectra of thermoplastic starch (TPS) film and TPS-5wt%DOL biocomposites films.

Overall, the FTIR spectra of the TPS film were almost similar to the TPS-5wt%DOL biocomposite films. Only slight changes can be observed in the range of 1800 to 2500 cm−<sup>1</sup> and 750 to 1200 cm−1. When benchmarked with the TPS and TPS-5wt%DOL(P) biocomposite film, we can see slight broadening of the peak that relates with molecules of starch in the TPS-5wt% DOL(U) film (refer to the bands around 750 to 1200 cm−1). Interestingly, this was also accompanied by the broadening and weakening of peaks, which was related to dolomite (1800–2500 cm<sup>−</sup>1). These could be a signal that some matrix–filler interactions have occurred (perhaps hydrogen bonding) between the starch and DOL(U) filler. As mentioned earlier, the size of dolomite became much smaller upon the sonication process. Consequently, these smaller particles could be better dispersed and diffused into the TPS chains, which interact well through the formation of hydrogen bonding. Better interaction between the DOL(U) filler and TPS molecules resulted in the TPS-5wt%DOL(U) biocomposite film that possesses better mechanical properties than the TPS film, as discussed in the following section.

### *4.3. Crystallinity Analysis of the Dolomite Filler (DOL(P) and DOL(U))*

The structure of dolomite is ordered; thus, its crystalline structure can be analyzed through XRD. The structure of the DOL(P) and DOL(U) was characterized and compared by using XRD analysis. An XRD pattern was used to observe the changes in the dolomite structure before and after undergoing the sonication process. Figure 8 illustrates the XRD pattern of the DOL(P) and DOL(U).

**Figure 8.** X-ray diffraction (XRD) pattern of DOL(P) and DOL(U).

The XRD pattern of dolomite (DOL(P) and DOL(U)) reveals diffraction peaks at around 2*θ* = 31.2◦ and 33.8◦, which are similar to that observed by Mohammed et al. [15]. However, when benchmarked with the DOL(P), the DOL(U) shows more intense diffraction peaks at both angles. This signifies that the crystallinity of dolomite was enhanced by the sonication process. To prove this, the degree of crystallinity of both types of dolomite was measured and tabulated in Table 3. The degree of crystallinity for DOL(P) and DOL(U) was found to be 76.57% and 77.10%, respectively. The higher degree of crystallinity of the DOL(U) could be as a result of the 2-h sonication process. The reduction in particle size of dolomite has increased its surface area. As a result, signals from X-ray diffraction that went through the crystalline lattices of dolomite particles became stronger [5].


**Table 3.** Crystallinity of the pristine dolomite (DOL(P)), sonicated dolomite (DOL(U)), corn starch, TPS film, TPS-5wt%DOL (P), and TPS-5wt%DOL (U) biocomposite films based on XRD analysis.

### *4.4. Crystallinity Analysis of the Corn Starch, TPS Film, and TPS-DOL Biocomposite Films*

XRD can be used to determine the degree of crystallinity based on the ratio of crystalline to amorphous phase in the starch. Starch is a semi-crystalline material, which consists of two main components; amylose (forming crystalline region) and amylopectin (forming amorphous phase). The existence of these two phases can be proved through the XRD spectrum.

The degree of crystallinity values for the corn starch, TPS film, TPS-5wt%DOL(P), and TPS-5wt%DOL(U) films were also further studied by using XRD analysis. Table 2 indicates the percent crystallinity of all the samples, while their XRD pattern in the range of 10◦ to 35◦(2*θ*) is illustrated in Figure 9.

XRD pattern of the corn starch powder indicates strong and sharp diffraction peaks at 2*θ* = 15.1◦, 20◦, and 22.9◦ and an unresolved doublet at 17.1◦ and 18◦, which signify a typical crystalline structure of cereal starch (A-type crystallinity). Due to the double-helical arrangement of amylopectin chains, these A-type crystallites are usually denser and less hydrated. Similar results of the pure corn starch were proved by other researchers [5,20]. The peaks at 2*θ* = 15.1◦ and 18◦ that exist in the corn starch disappeared in the XRD signal of the cast TPS film, which means the original crystalline structure was destroyed during the solution mixing process. However, there is a minimum residue of A-type crystalline that could still be observed in the TPS film produced by the solvent casting technique [21]. In the preparation of the solvent cast TPS film, shear and thermal energies applied during the stirring process were adequate to melt granular the crystallites; therefore, most of the crystalline phase of the starch has been diminished.

The XRD pattern of the TPS film has shown the combination of crystallinities Type B and Type Vh as well as the low value of the degree of crystallinity, which is about 25.38%. There are diffraction peaks at 2*θ* of 17.2◦, 19.5◦, 22.6◦, and 34.1◦. Peaks at 17.2◦ are assigned to B-type crystals, while peaks at 19.5◦ and 22.6◦ belong to Vh-type crystals, which are related to amylose crystallization into a single helical structure [5,22].

Upon the addition of dolomite into the starch matrix, the peak at 2*θ* = 31◦ (for TPS-5wt% DOL(P)) and doublet peak at 2*θ* = 31.2◦ and 31.6◦ (for TPS-5wt%DOL(U)) appear. Those peaks are correlated with dolomite, further proving the existence of dolomite particles in the structure of the biocomposite films. As referred to Table 3, it can be seen that the addition of dolomite filler significantly increased the degree of crystallinity of the TPS film. Although the peak location of starch does not change obviously in the XRD pattern, the intensities increased significantly due the addition of dolomite into the TPS film (refer to the diffraction peaks 2*θ* at around 22◦ for TPS film and TPS-5wt%DOL(P)) film. The degree of crystallinity of the TPS film was increased from 25.38% to 51.79% when incorporated with DOL(P).

**Figure 9.** XRD pattern of corn starch, TPS film, and TPS-5wt%DOL (P) and TPS-5wt%DOL (U) biocomposite films.

According to Table 3, TPS-5wt%DOL(U) (52.43%) has a higher degree of crystallinity as compared with TPS-5wt%DOL(P) (51.79%). Based on Figure 9, the peak intensity for TPS-5wt%DOL(U) is higher than that of TPS-5wt%DOL(P), (refering to diffraction peaks at around 2*θ* = 31◦). In addition, TPS-5wt%DOL(U) also showed a strong doublet peak at around 2*θ* at 31◦ due to the presence of well-dispersed dolomite particles. The deagglomeration of dolomite upon the sonication process led to the increased surface area of the dolomite particles, thus producing a stronger diffraction peak when dispersed in the TPS matrix. Another reason could be related to the sonication process that facilitated the dispersion of dolomite particles throughout the TPS matrix and improved the matrix–filler interactions. Subsequently, this has encouraged the nucleation process for the TPS matrix crystallization and thus increased the degree of crystallinity of the biocomposites film [23]. The increase of degree of crystallinity is commonly associated with its improvement in tensile strength [5,21]. This is also the reason why the TPS-DOL(U) biocomposite film has better tensile strength when compared with TPS-DOL(P) biocomposite film, which will be shown and discussed in the next section.

### *4.5. Mechanical Properties Evaluation through Tensile Test*

Figures 10–12 illustrate the effect of dolomite loading (0, 1, 2, 3, 4 and 5 wt%) and sonication process on the tensile properties of TPS biocomposite films. According to Figure 10, low loadings of dolomite cause a deterioration of the tensile strength of the TPS-DOL biocomposite film. The TPS film possesses the tensile strength of 2.64 MPa. Meanwhile, the tensile strength of the TPS-DOL (P) biocomposites with 1, 2 and 3 wt% filler was found to be 1.76 MPa, 1.73 MPa, and 1.98 MPa, respectively. On the other hand, TPS-DOL(P) biocomposite films with high loading of filler (4 and 5 wt%) show almost the same tensile strength value with the TPS film. This assessment was made based on the tensile strength and standard deviation values of the TPS film and TPS-DOL(P) biocomposite films containing 4 and 5 wt% DOL(P), which show statistically no significant difference. The results suggest that dolomite in pristine form was not a good reinforcing filler for the TPS film and thus could not improve the tensile strength of the host biopolymer when added in 1 to 5 wt%. Conversely, for TPS-DOL(U) biocomposite film cases, the tensile strength was enhanced when DOL(U) was added in 4 wt% and 5 wt%. Based on the results, the tensile strength of TPS-4wt%DOL(U) biocomposites was 3.06 MPa, which showed an improvement of about 15.9% when compared with the TPS film. However, when less DOL(U) was added into the TPS (1 wt%, and 2 wt%), the tensile strength was 1.88 MPa and 1.87 MPa, respectively, which is somewhat lower than the TPS film. The TPS-5wt% DOL(U) biocomposite has the highest tensile strength among all the samples (3.61 MPa) by demonstrating 36.74% higher in tensile strength when benchmarked with the TPS film.

**Figure 10.** Tensile strength of TPS film and TPS-DOL biocomposite films.

Overall, the results indicate that the low loading of dolomites, either in pristine or sonicated form, could not improve the tensile strength of the TPS biocomposite film. This could be due to the insufficient content of dolomite particles to allow a reinforcing effect to the TPS chains. This happened due to the lack of dispersed particles to act as a stress-transferring medium when subjected to the tensile load. In fact, the lowering of tensile strength can be seen due to the presence of filler that induces inhomogeneity in the biocomposite system, rather than providing reinforcing effect. Syed Adam et al. have obtained and reported the same findings through their research on TPS biocomposite [7]. However, when a higher content of dolomite was added, there was a greater number of particles being diffused into the TPS chain and acting as reinforcing filler. Upon the sonication process, the dolomite filler turned into smaller size particles, which can be

more easily dispersed throughout the TPS matrix. There were higher numbers of welldispersed dolomite particles to interact with TPS molecular chains. As observed through the FTIR analysis, such interactions can be observed through the formation of hydrogen bonding between the dolomite filler and TPS molecular chains. As a result, a better stress-transforming mechanism between the matrix and filler occurred. Apparently, the tensile strength of the TPS-5wt%DOL(U) biocomposite film (3.61 MPa) was better than the tensile strength of the TPS-5wt%DOL(P) biocomposite film (2.68 MPa). This shows that the sonication process can assist the distribution and dispersion of dolomite filler throughout the TPS, which eventually improved the tensile strength of the biocomposite film. A better dispersion of dolomite could increase the efficiency of the filler to transfer the stress effectively within the composite structure. This is in agreement with the findings of Nik Adik et al. [24].

**Figure 11.** Elongation at break of TPS film and TPS-DOL biocomposite films.

**Figure 12.** Young's modulus of TPS film and TPS-DOL biocomposite films.

Elongation at break values of the TPS film, TPS-DOL(P) biocomposite, and TPS-DOL(U) biocomposite films are compared in Figure 11. It can be seen that as the DOL(P) and DOL(U) loading increased in the TPS matrix, reduction in the elongation at break occurred. Elongation at break is a reflection of the ductility of the material: the direct opposite for brittleness. When dolomite filler is incorporated into a TPS matrix, it will increase its stiffness. The increase in stiffness resulted in a decrease in the ductility of the material. Thus, as the dolomite loading increases, the ductility decreases. Such a trend was also reported by other researchers [24,25]. The elongation at break of the TPS film was found to be 95.6%. With the addition of 1 wt% DOL(P) in TPS film, the elongation at break was 126.13%. The elongation at break was reduced significantly (25.43%) when the DOL(P) loading was increased to 2 wt%. When the loading of DOL (P) increased to 3 wt% and 4 wt%, the elongation at break value continued to decrease to 94.7% and 85.67%, respectively. TPS-5wt%DOL(P) has the lowest elongation at break among all the samples (66.37%), showing 30.58% reduction when benchmarked with the TPS film. Apparently, TPS filled with DOL(U) has a similar trend with TPS filled with DOL(P). The elongation at break of TPS-1wt%DOL(U) biocomposites was 165.77%, but it reduced significantly by 23.59% to 134.13% when the loading of DOL(U) was increased to 2 wt%. A further increase of DOL(U) loading to 3 wt% and 4 wt% resulted in the decrease of elongation at break to 106.9% and 100.37%, respectively. The TPS-DOL(U) biocomposite film achieved the lowest elongation at break (96.17%) when the filler loading was 5 wt%. As expected, the elongation at break value of the TPS-DOL(P) biocomposite films is lower than the TPS-DOL(U) biocomposite films when the same amount of filler was added. For instance, the elongation at break of TPS-1wt%DOL(P) (126.13%) was lower compared with TPS-1wt%DOL(U) (165.77%). This was probably due to the greater aggregation of dolomite filler when no sonication process was involved, thereby resulting in a greater stiffening effect to the TPS molecular chains. Subsequently, the flexibility of the biocomposite film reduced more significantly.

Young's modulus values of the TPS film, TPS-DOL(P), and TPS-DOL(U) biocomposite films at different dolomite content (1 wt%, 2 wt%, 3 wt%, 4 wt%, and 5 wt%) are demonstrated in Figure 12. It can be seen that the addition of DOL(P) and DOL(U) into the TPS matrix resulted in the increase of the Young's modulus of the biocomposites. In general, the Young's modulus of these samples was enhanced by the incorporation of dolomite filler in the TPS matrix. This was due to the stiffness of dolomite, which can restrict the chain mobility of the TPS matrix. The results are in tandem with findings reported by a few researchers [24–26]. It can be concluded that the TPS film has the lowest Young's modulus (7.1 MPa) among all the samples. The Young's modulus of TPS-5wt%DOL(P) biocomposite film was 10.23 MPa, which possesses a 44.08% increment when benchmarked with the TPS film. The TPS-5wt%DOL(U) biocomposite achieved the highest Young's modulus (which represents the stiffest samples) among all the samples. It possesses an 87.32% increment in Young's modulus when benchmarked with the TPS film. Moreover, the value of Young's modulus obtained for TPS-DOL(U) biocomposite films was higher than that of TPS-DOL(P) biocomposite films. The most significant difference can be observed when incorporating 5 wt% of dolomite into the TPS matrix. The TPS-5wt%DOL(U) (13.3 MPa) achieved a 30% increment in Young's modulus when compared with TPS-5wt%DOL(P) (10.23 MPa). As explained previously, dolomite that underwent a sonication process can be more homogeneously dispersed throughout the TPS matrix. Better filler–matrix interactions cause more restriction in TPS chain mobility and thus increase the Young's modulus of the biocomposite film [5,7].

### *4.6. Mechanical Properties Evaluation through Tear Test*

Figure 13 compares the values of the tear strength of the TPS film and TPS-DOL biocomposite films. Generally, the tear strength of the TPS film (0.80 MPa) is lower than that of the TPS-DOL(P) and TPS-DOL(U) biocomposite films. This indicates that both pristine dolomite and sonicated dolomite have successfully improved the tear strength of the TPS. Both TPS-DOL(P) and TPS-DOL(U) biocomposite filmss achieved the highest tear strength when 5 wt% of filler was employed. TPS-5wt%DOL(U) biocomposite film shows the highest tear strength (2.30 MPa) among all the samples. It has achieved 187.50% higher tear strength when compared with the TPS film.

**Figure 13.** Tear strength of TPS film and TPS-DOL biocomposite films.

Obviously, when the same amount of filler was used, the TPS-DOL(U) biocomposite film possesses greater tear strength as compared to the TPS-DOL(P) biocomposite film. This happened due to the sonication process that helps to homogenously distribute dolomite particles into TPS molecule chains and improve the matrix–filler interactions [5]. The homogenous distribution of dolomite can increase the efficiency of filler to transfer the load within the TPS biocomposite structure. Obviously, the tear strength data of the biocomposite seem to follow the trend of the tensile strength data, in which the higher loading of dolomite filler causes a better reinforcing effect. The following statement explains this phenomenon: When a greater amount of dolomite particles penetrated through the polymer chain, they would develop more surface interactions with the TPS matrix; therefore, a better reinforcing effect can be seen. This is translated through the tensile and tear strength values. Osman, Edwards, and Martin have observed the same phenomenon in their composite system with mineral filler (clay) [27].

#### *4.7. Morphology of Tensile Fractured Surface of the TPS Film and TPS-DOL Biocomposites*

Figure 14 presents the SEM images of the tensile fractured surface of the TPS film and TPS-DOL biocomposite films. In its neat form (without filler), the TPS film shows a smooth and homogeneous surface. However, upon the addition of pristine dolomite (DOL(P)), the surface morphology of the matrix became rough and non-homogeneous with pores and pits. Apparently, at the same dolomite loading, the TPS-DOL biocomposites with sonicated filler (TPS-DOL(U)) exhibit smoother and more homogeneous surface morphology than the TPS-DOL biocomposites with pristine dolomite (TPS-DOL(P). These findings support the results of a test where the addition of DOL(U) in the TPS resulted in a more homogeneous biocomposite due to the well-dispersed filler distributed in the matrix. As a result, the tensile properties of the TPS film were enhanced. Li et al. have observed similar features in their poly (3-hydroxybutyrate-co-3-hydroxyvalerate) (PHBV)-based biocomposites with miscanthus biocarbon as the filler. A homogeneous dispersion of filler has been found to improve the mechanical and thermal stability of the host biopolymer. The biocomposite with a poor dispersion of filler indicated a rough fractured surface with the appearance of voids, suggesting the poor wetting of filler by the biopolymer matrix [28].

**Figure 14.** SEM images of tensile fractured surface of the (**a**) TPS film (**b**) TPS-3wt%DOL (P) (**c**) TPS-3wt%DOL (U) (**d**) TPS-5wt% DOL (P) and (**e**) TPS-5wt%DOL (U) biocomposite films at 500× magnification (left) and 1000× magnification (right).

### **5. Summary**

The effects of dolomite loading and the sonication process on the mechanical properties of the TPS-DOL biocomposite films have been explored through this research. Indicatively, the use of high loading of dolomite (4 to 5 wt%) can improve both the tensile and tear properties of the biocomposite film, provided that the dolomite is being sonicated prior to being added into the TPS matrix. The use of dolomite in pristine form (DOL(P)) could not

improve the tensile properties, even when added in 5 wt%. Conversely, upon the sonication process, the dolomite (DOL(U) brought an increment in both tensile and tear strength of the biocomposite film when added in 4 and 5 wt%. Based on the FTIR results, the spectral peaks of DOL(U) were more intense when compared with DOL(P). This means that when dolomite underwent the sonication process, a reduction in particle size occurred, causing a larger surface area of particles for the penetration of IR light into a greater number of molecules. In addition, the sonication process also reduced the particle size of the dolomite, assisting the dispersion and distribution of its particles throughout the TPS matrix. XRD analysis showed that the sonication process has increased the crystallinity of dolomite. This factor also contributed to the enhancement of the tensile and tear properties of the biocomposite film. The fractured surface morphology of the TPS-DOL (U) biocomposites is smoother and more homogeneous than the TPS-DOL (P) counterparts, further supporting the fact that the sonication process can improve the dispersion of the dolomite filler throughout the TPS matrix and produce more homogeneous biocomposites. In conclusion, sonicated dolomite has potential to be used as a reinforcing filler in the TPS film in order to improve its viability for packaging application. The low cost and abundancy of dolomite and the simple and environmental friendly method of sonication can be adopted to produce a sustainable biocomposite for future needs.

**Author Contributions:** Conceptualization, A.F.O.; data curation, L.S. and A.A.A.; formal analysis, L.S. and A.A.A.; investigation, A.F.O. and L.S.; methodology, A.F.O. and I.I.; project administration, A.F.O.; software, A.A.A.; writing—original draft, A.F.O.; writing—review & editing, A.U.-H. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Fundamental Research Grant Scheme (FRGS) under a grant no: FRGS/1/2019/TK05/UNIMAP/02/6 from the Ministry of Education Malaysia.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable for studies not involving humans.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **Abbreviations**


### **References**


## *Article* **Synthesis, Characterization, and Application of Carboxymethyl Cellulose from Asparagus Stalk End**

**Warinporn Klunklin 1, Kittisak Jantanasakulwong 1,2,3, Yuthana Phimolsiripol 1,2,3, Noppol Leksawasdi 1,2,3, Phisit Seesuriyachan 1,2, Thanongsak Chaiyaso 1,2, Chayatip Insomphun 1,2, Suphat Phongthai 1,2, Pensak Jantrawut 3,4, Sarana Rose Sommano 3,5, Winita Punyodom 3,6, Alissara Reungsang 7,8,9, Thi Minh Phuong Ngo <sup>10</sup> and Pornchai Rachtanapun 1,2,3,\***


**Abstract:** Cellulose from *Asparagus officinalis* stalk end was extracted and synthesized to carboxymethyl cellulose (CMCas) using monochloroacetic acid (MCA) via carboxymethylation reaction with various sodium hydroxide (NaOH) concentrations starting from 20% to 60%. The cellulose and CMCas were characterized by the physical properties, Fourier Transform Infrared spectroscopy (FTIR), Differential scanning calorimetry (DSC), Scanning electron microscopy (SEM) and X-ray diffraction (XRD). In addition, mechanical properties of CMCas films were also investigated. The optimum condition for producing CMCas was found to be 30% of NaOH concentration for the carboxymethylation reaction, which provided the highest percent yield of CMCas at 44.04% with the highest degree of substitution (DS) at 0.98. The melting point of CMCas decreased with increasing NaOH concentrations. Crystallinity of CMCas was significantly deformed (*p* < 0.05) after synthesis at a high concentration. The *L\** value of the CMCas was significantly lower at a high NaOH concentration compared to the cellulose. The highest tensile strength (44.59 MPa) was found in CMCas film synthesized with 40% of NaOH concentration and the highest percent elongation at break (24.99%) was obtained in CMCas film treated with 30% of NaOH concentration. The applications of asparagus stalk end are as biomaterials in drug delivery system, tissue engineering, coating, and food packaging.

**Keywords:** agricultural waste; asparagus; biopolymer; carboxymethyl cellulose; CMC; degree of substitution; DS; cellulose extraction

**Citation:** Klunklin, W.;

Jantanasakulwong, K.; Phimolsiripol, Y.; Leksawasdi, N.; Seesuriyachan, P.; Chaiyaso, T.; Insomphun, C.; Phongthai, S.; Jantrawut, P.; Sommano, S.R.; et al. Synthesis, Characterization, and Application of Carboxymethyl Cellulose from Asparagus Stalk End. *Polymers* **2021**, *13*, 81. https://dx.doi.org/ polym13010081

Received: 8 November 2020 Accepted: 21 December 2020 Published: 28 December 2020

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ( https://creativecommons.org/ licenses/by/4.0/).

### **1. Introduction**

Asparagus (*Asparagus officinalis* L.) is a nutritious and perennial vegetable continually used as antifungal activities, anticancer and anti-inflammatory herbal medicine in Asia. In the processing of asparagus, the spears which are 2–3% of total weight of the asparagus are typically processed into three types of products: canned, fresh, and frozen [1]. The residues from the processing were used for animal feeding and produced low-value products due to the high content of cellulose [2]. Therefore, tons of asparagus stalk end are the agriculture by-product which can cause environmental pollution [2]. With abundant cellulose and photochemical properties of cellulose in asparagus by-products, it acts as a good source of new value-added products [1], including biological compounds and functions [3,4], a dietary fiber [5] and nanocellulose [6]. Agricultural by-products from asparagus are rich source of celluloses which can be isolated from asparagus stalk end. A little work has been carried out to extract the cellulose from the asparagus; however, the previous works have not comprehensively considered to synthesis and apply CMC from asparagus stalk end as a film or coating to a food industry.

Cellulose is the most abundant polymer present in the primary cell wall of plants, algae and the oomycetes [6,7]. The abundant cellulose consists of repeated links of β–Dglucopyranose which can convert into high value cellulose esters and ethers [7]. The cellulose is insoluble in water which limits many applications. To increase the utility of cellulose, the carboxymethylation process of cellulose is able to syntheses water-soluble cellulose derivatives called carboxymethylcellulose (CMC) [8]. The forecast for global carboxymethyl cellulose market is estimated to reach approximately USD 1.86 billion by 2025. Food market generated the highest revenue of the global CMC market in 2016 with a market share of 33.7% [9]. The productions of CMC synthesized from agricultural waste have been studied such as sago waste [10], *Mimosa pigra* peel [11], durian rind [8], cotton waste [12], corn husk [13], corncob [14], grapefruit peel [15], rice stubble [16], papaya peel [17], *Juncus* plant stems [18], and mulberry wastepaper [19].

Thus, CMC can be versatile applied in various field such as drug delivery carrier [20], wound healing application [21], active corrosion protection [22], binder in ceramics [23], composited in hydrogel film [24] and hydrogel food packaging [25], fresh fruit and vegetable coatings [26,27] and hydrocolloid in noodles, bakery products and other foods [28,29] and it can produce composite films. CMC films exhibit many worthwhile attributes than many other biopolymers such as biodegradability, ease of production, forming a stable cross-linked matrix for packaging, etc. However, no research has presented the production of CMC powder from asparagus stalk end (CMCas) and its application such as packaging film so far.

Therefore, this study aimed to determine the effect of NaOH concentration at 20–60% on percent yield, the degree of substitution (DS), thermal properties, chemical structure, viscosity, crystallinity, morphology of CMCas. The effect of NaOH concentration mechanical properties (tensile strength and percentage elongation at break) and morphology of CMCas films was also carried out.

#### **2. Materials and Methods**

### *2.1. Materials*

Asparagus stalk end was purchased from Nong Ngulueam Sub-District (Nakhon Prathom, Thailand). All chemicals used in the preparation and analysis of synthesized CMC were analytical reagent (AR) grade or the equivalent. Absolute methanol, ethanol, and sodium silicate from Union science Co., Ltd. (Chiang Mai, Thailand), hydrogen peroxide from QReCTM (Auckland, New Zealand), sodium hydroxide and glacial acetic acid from Lab-scan, isopropyl alcohol (IPA), Monochloroacetic acid from Sigma-Aldrich (Steinheim, Germany).

### *2.2. Materials Preparation*

Two grams of asparagus stalk end was cut into small pieces and dried in an oven 105 ◦C for 6 h. Then dried asparagus was weighed to calculate the percent dryness as described in Equation (1). Dried asparagus stalk end was grounded by using a hammer mill (Armfield, Ringwood, Hampshire, UK) to less than 1 mm. The dried powder was then stored in polypropylene (PE) bags at ambient temperature until used. The percent dryness was calculated by the following Equation (1)

$$\%Dryness = \frac{Weight\ of\ cellulose\ without\ moisture\ content}{Weight\ of\ cellulose\ content} \times 100\tag{1}$$

### *2.3. Extraction of Cellulose from Asparagus Stalk End*

Cellulose from dried asparagus stalk end powder was extracted according to the method of [11,16]. Briefly, the dried powder of asparagus stalk end was extracted with a ratio of cellulose to 10% (*w*/*v*) NaOH solution at 1:20 (*w*/*v*) treated at 100 ◦C for 3 h. The black slurry was filtered and rinsed with cold water until a neutral pH of rinsed water was obtained. To obtain the cellulose fiber, the washed fiber residue was dried in an oven at 55 ◦C for 24 h. Hydrogen peroxide and sodium silicate were used to bleach the cellulose fiber. The bleached cellulose was then grounded by using the hammer mill size 70 mash (Armfield, Ringwood, Hampshire, UK). The bleached cellulose powder was kept in PE bags at ambient temperature until used.

### *2.4. Carboxymethyl Cellulose (CMC) Synthesized from Asparagus Stalk End*

50 mL of various concentrations of NaOH at 20, 30, 40, 50, and 60% (*w*/*v*) and 350 mL of isopropanol (IPA) were blended with 15 g of cellulose powder extracted from asparagus stalk end and this was mixed for 30 min. Subsequently, the carboxymethylation reaction was synthesized according to the method of [11]. The final CMCas product was in powder form. The percent yield was calculated by the following Equation (2) [11]:

$$\text{Yield of } \text{CMC } \left( \% \right) = \frac{\text{Weight of } \text{CMC } \left( \text{g} \right)}{\text{Weight of } \text{cellulose } \left( \text{g} \right)} \times 100 \tag{2}$$

### *2.5. Determination of the Degree of Substitution (DS) of CMCas*

The degree of substitution (*DS*) of CMCas presents the average number of hydroxyl group replaced by carboxymethyl and sodium carboxymethyl groups at C2, 3, and 6 in the cellulose structure. The *DS* of CMCas was evaluated by the USP XXIII method for Crosscarmellose sodium which are titration and residue on combustion [11,23]. Calculation of the *DS* was showed in the following Equation (3);

$$DS = A + S \tag{3}$$

where *A* is the *DS* of carboxymethyl acid and *S* is the *DS* of sodium carboxymethyl which *A* and *S* were calculated using Equations (4) and (5):

$$A = \frac{1150M \text{ content}}{(7120 - 412M - 80C) \text{Content}} \tag{4}$$

$$S = \frac{(162 + 58A)\mathcal{C}\,\text{content}}{(7120 - 80\mathcal{C})\,\text{content}}\tag{5}$$

where *M* is consumption of the titration to end point (mEq) and *C* is the amount of ash remained after ignition (%).

#### *2.6. The Determination of Percentage of Residue on Ignition*

A crucible was dried in an oven at 100 ◦C for 1 h and transferred to a desiccator until the weight reached at an accurate value (Scale, AR3130, Ohaus Corp. Pine, NJ, USA) [23]. CMCas was added into a crucible. The crucible containing CMCas were ignited at 400 ◦C using a kiln (Carbolite, CWF1100, Scientific Promotion, Sheffield, UK) for approximately l to l.5 h and placed into the desiccator to obtain black residue. Sufficient sulfuric acid was added to moisten the entire residues and heated up until the gas fumes (white smoke) was utterly disappeared. The crucible with the black residue was ignited at 800 ± 25 ◦C to produce white residue and placed in the desiccators to get a constant weight. The residue on ignition was presented in the percentage and calculated using Equation (6).

$$\text{The percentage of residue} = \frac{\text{Weight of residue content}}{\text{Weight of CMC content}} \times 100\tag{6}$$

### *2.7. Fourier Transform Infrared Spectroscopy (FTIR)*

The aldehyde groups of the cellulose from stalk asparagus and CMCas were determined by using FTIR spectra (Bruker, Tensor 27, Billerica, MA, USA). 2 mg of dry sample was pressed into a pellet with KBr. Transmission level measured at the wavenumber range of 4000–400 cm−<sup>1</sup> [28].

### *2.8. Viscosity of Cellulose and CMCas*

A Rapid Visco Analyzer (RVA-4, Newport Scientific, Warriewood, Australia) was used to determine the viscosity of samples. 1 g of cellulose or CMCas sample was weight and dissolved in 25 mL of distilled water with stirring at 80 ◦C for 10 min—automatic stirring action set at 960 rpm for 10 s. The temperature of the sample was varied from 30, 40, 50 and 60 ◦C at 5 min intervals and held speed at 160 rpm until end of the test [11].

### *2.9. Thermal Conductivity Measurements*

The thermal characterization of cellulose from asparagus stalk end and CMCas were determined using Differential Scanning Calorimeter (DSC) (Perkin Elmer, Kitakyushu, Japan) according to the melting temperature determined from the thermogram. The determination was performed under a nitrogen atmosphere at a flow rate of 50 mL min−1. Samples (10 mg) contained in aluminum pans were heated from 40 to 450 ◦C with a heating rate of 10 ◦C min−<sup>1</sup> [11].

### *2.10. X-ray Diffraction (XRD)*

The crystallinity of the samples was determined using X-ray diffraction patterns carried out by X-ray powder diffractometer (JEOL, JDX-80-30, Shimadzu, Kyoto, Japan) in the reflection mode. Prior to the test, the samples were dried in a hot air oven (Memmert, Büchenbach (Bavaria, Germany) at 105 ◦C for 3 h to produce a powder form. Scans were carried out in the range of scattering angle (2θ) from 10 to 60◦ with at a scan rate of 5◦/min [13].

### *2.11. Scanning Electron Microscopy (SEM)*

Scanning Electron Microscopy (SEM) (JEOL JSM-5910LV SEM; Tokyo, Japan) was used to analyze the surface morphology of cellulose from asparagus stalk end and CMCas. The samples were analyzed through a large field detector. The acceleration voltage was used at 15 kV with 1500× original magnification [12].

### *2.12. Color Characteristics*

The color characteristic of cellulose form asparagus stalk end and CMCas was evaluated by using a Color Quest XE Spectrocolorimeter (Hunter Lab, Shen Zhen Wave Optoelectronics Technology Co., Ltd., Shenzhen, China) in order to express the CIELAB color as three values: *L\** for the lightness from blackness (0) to whiteness (100), *a\** from greenness (−) to redness (+), and *b\** from blueness (−) to yellowness (+). The total color differences

(Δ*E*) take into account the comparisons between the *L\**, *a\** and *b\** value of the sample and standard and calculated by the following Equation (7) [8]

$$
\Delta E = \sqrt{\left(L\_{Standard}^\* - L\_{Sample}^\*\right)^2 + \left(a\_{Standard}^\* - a\_{Sample}^\*\right)^2 + \left(b\_{Standard}^\* - b\_{Sample}^\*\right)^2} \tag{7}
$$

The whiteness index (*WI*) was also calculated by the following Equation (8) to represent the degree of whiteness of samples [30].

$$WI = \sqrt{\left(100 - L^{\*^2}\right) + a^{\*^2} + b^{\*^2}}\tag{8}$$

### *2.13. Preparation of CMCas Film*

The film-forming solutions were prepared by dissolving 4 g of CMCas in 180 mL of distilled water with a constantly stirred at 80 ◦C under magnetic stirring for 15 min. 15% (*w*/*w*) glycerol was added to the solution and stirred for additional 5 min. The solution was poured on acrylic plates (20 cm × 15 cm) and the dried at room temperature for 72 h. After that, the film was peeled from the plates and stored in polyethylene bags at ambient temperature [11].

### *2.14. Mechanical Properties of CMCas Film*

The thickness of CMCas films was determined using a micrometer model GT-313-A (Gotech Testing Machine Inc., Taichung, Taiwan, China). The examined mechanical properties which were tensile strength (*TS*) and percent elongation at break (*EB*) of the CMCas film were tested (10 measurement each) using a universal texturometer (Model 1000 HIK-S, UK) according to the standard procedure of ASTM D882-80a [31] with preconditioning for 24 h and determined at 27 ± 2 ◦C with a relative humidity (RH) of 65 ± 2% according to Thai industrial standard for oriented polypropylene film (TIS 949-2533). The rectangular CMCas films were cut into 15 × 140 mm as test specimens. The specimens were carried out by using an initial grip separation distance of 100 mm and crosshead speed at 20 mm/min. The *TS* and *EB* were calculated by following Equations (9) and (10), respectively.

$$TS\ (MPa) = \frac{The\ maximum\ load\ (N)}{Weight\ of\ each\ film\ \times\ Thicksness\ of\ each\ film} \tag{9}$$

$$EB\left(\%\right) = \frac{\text{The length of the film rupture} - \text{The initial length of the film}}{\text{The initial length of the film}} \tag{10}$$

### *2.15. Statistical Analysis*

Statistic data were analyzed by a one-way analysis of variance (ANOVA) using SPSS software version 16.0 0 (SPSS, an IBM company, Chicago, IL, USA). Duncan's multiple range test was employed to evaluate significant differences among the treatments (*p* < 0.05). All measurements were analyzed in triplicated. The results were represented the mean values ± standard deviation. The figures present the standard deviation as the appropriate values. The error bars for some data points overlap the mean values.

### **3. Results and Discussion**

### *3.1. Percent Yield of Carboxymethyl Cellulose from Asparagus Stalk End (CMCas)*

The percent yield of CMCas synthesized with various NaOH concentrations at 20, 30, 40, 50 and 60% (*w*/*v*) is shown in Figure 1. The percent yield of CMCas increased when increasing of NaOH concentration from 20% to 30% (*w*/*v*) and then decreased with further increasing NaOH concentration at over 30% (*w*/*v*). The resulted phenomenon might be occurred due to a limitation of sodium monochloroacetate (NaMCA) as etherifying agents for substituting cellulose. Moreover, the NaOH concentration was not high enough to complete for converting the cellulose into alkali cellulose [32]. This result was similar to the work of synthesis carboxymethyl cellulose from durian rind [8].

**Figure 1.** Percent yield of carboxymethyl cellulose from asparagus stalk end (CMCas).

### *3.2. Degree of Substitution (DS) of CMCas*

The effect of various concentrations of NaOH on DS of CMCas was carried out shown in Figure 2. The DS of CMCas was in between 0.49 and 0.98 which reached the highest value at the concentration of NaOH at 30% (*w*/*v*). However, the DS of CMCas was reduced at higher concentration of NaOH at 40–60% (*w*/*v*) ranged from 0.83–0.49. This phenomenon can be explained by investigating the carboxymethylation process, where two reaction occur concurrently. The first reaction involved a cellulose hydroxyl reacting with sodium monochloroacetate (NaMCA) in the presence of NaOH to obtain CMCn shown in Equations (11) and (12).

**Figure 2.** Effect of NaOH on degree of substitution of CMCas.

$$\text{R-OH} + \text{NaOH} \rightarrow \text{R-ONa} + \text{H}\_2\text{O} \tag{11}$$

$$\text{R-ONa} + \text{Cl-CH}\_2\text{-COONa} \rightarrow \text{R-O-CH}\_2\text{-COONa} + \text{NaCl} \tag{12}$$

The second reaction involves NaOH reacting with NaMCA to form sodium glycolate as by-product shown in Equation (13) [33].

$$\text{NaOH} + \text{Cl-CH}\_2\text{COONa} \rightarrow \text{HO-CH}\_2\text{COONa} + \text{NaCl} \tag{13}$$

The second reaction overwhelms the first reaction together with a strong alkaline concentration. If the alkaline level in the second reaction is too high, a side reaction will form a high level of sodium glycolate as a by-product, thus lowering the DS. This result was agreed with the result found by [11] who studied CMC from *Mimosa pigra* peel. The degradation was taken place due to a high concentration of NaOH [11]. Moreover, similar results have been reported in [8]. The maximum DS value (0.98) of CMC from durian rind was also got from the NaOH concentration at 30% and this was also related to the trend in percent yield presented in Figure 1.

### *3.3. Fourier Transform Infrared Spectroscopy (FTIR) of Cellulose from Asparagus Stalk End and CMCas*

FTIR was used to evaluate functional groups changes in the cellulose from asparagus stalk end and CMCas structures. The substitution reaction of CMCas during the carboxymethylation was confirmed by FTIR [28]. Cellulose and each CMCas have similar functional groups with same absorption bands such as hydroxyl group (–OH stretching) at 3200–3600 cm−1, C–H stretching vibration at 3000 cm−1, carbonyl group (C=O stretching) at 1600 cm−1, hydrocarbon groups (–CH2 scissoring) at 1450 cm−1, and ether groups (–O– stretching) at 1000–1200 cm−<sup>1</sup> [34]. The FTIR spectra of the cellulose and CMCas synthesized with various NaOH concentrations are shown in Figure 3. The carbonyl group (C=O), methyl group (–CH2) and ether group (–O–) notably increased in the CMCas sample; however, the absorption band of hydroxyl group (–OH) reduced when compared to those cellulose samples (Figure 3). This result confirmed that the carboxymethylation had been substituted on cellulose molecules [32,34].

Wavenumber cmƺ<sup>1</sup>

**Figure 3.** FTIR spectra of (**a**) cellulose, (**b**) CMCas synthesized with 20% (*w*/*v*) NaOH concentration, (**c**) CMCas synthesized with 30% (*w*/*v*) NaOH concentration, (**d**) CMCas synthesized with 40% (*w*/*v*) NaOH concentration, (**e**) CMCas synthesized with 50% (*w*/*v*) NaOH concentration and (**f**) CMCas synthesized with 60% (*w*/*v*) NaOH concentration.

### *3.4. Effect of Various NaOH Concentrations on Viscosity of CMCas*

The relationship between different NaOH concentrations and viscosity of CMCas solution is shown in Figure 4. The viscosity of the CMCas solution (1 g/100 mL) reduced with increase in the temperature. Increasing temperature plays a role in reducing the cohesive forces while simultaneously increasing the rate of molecular interchange, causing the lower viscosity [11]. According to the literature, the viscosity of CMCas solution affected by NaOH concentrations was changed in accordance with DS value at the same temperature [33,35]. The CMCas synthesized with 20% (*w*/*v*) NaOH concentration was the highest in viscosity due to the swelling and gelatinization of cellulose comparable to previous

studies [11]. As far as NaOH concentration rose above 20%, the viscosity fell down. The decreasing of viscosity at a higher NaOH concentration was because of the degradation of CMC polymer led to a lower DS value that provides a small number of hydrophilic groups reducing the ability of the polymer to bond among water molecules [32,36]. An increase of the DS value also increased the CMC's ability to immobilize water in a system [36]. Nevertheless, the viscosity of CMC depends on many influencing factors such as solution concentration [32], pH value, NaOH concentrations [36], and temperatures [11].

**Figure 4.** Effect of various NaOH concentrations on viscosity of CMCas.

### *3.5. Effect of Various NaOH Concentrations on Thermal Properties of CMCas*

The effect of various NaOH concentrations on the thermal property of cellulose from asparagus stalk end and CMCas are shown in Figure 5. The melting temperature (Tm) of cellulose and CMCas synthesized with 20, 30, 40, 50, and 60 % (*w*/*v*) NaOH concentrations were 178.79, 206.61, 193.13, 182.34, 178.34 and 161.21 ◦C, respectively. The highest Tm of CMCas increased when 20% (*w*/*v*) NaOH concentration was employed. The Tm at 20–30% (*w*/*v*) of NaOH concentrations were higher than Tm of cellulose due to the substituent of carboxymethyl groups influenced on increasing ionic character and intermolecular bonds between the polymer chains [37]. The Tm at 40–50% (*w*/*v*) of NaOH concentrations decreased due to the melting temperature of CMCas slightly decreased as increasing the concentration of NaOH because of the increase in alkalization reaction together with a high substitution of carboxymethyl group. Moreover, the decrease in Tm was caused by an increase in number of the carbon atoms producing carbon skeleton, after a rapid quenching cooling with rate of 80 ◦C/min from the isotropic state at 200 ◦C [38]. As the Tm of 60% (*w*/*v*) of NaOH concentration decreased, side reaction was in the majority with sodium glycolate forming as a by-product and chain breaking of CMCas polymer similar to the results presented in *Mimosa pigra* peel [11].

**Figure 5.** Thermal property of cellulose from asparagus stalk end and CMCas synthesized with various NaOH concentrations.

### *3.6. X-ray Diffraction (XRD) of Cellulose from Asparagus Stalk End and CMCas*

The changes on the structure of cellulose from asparagus stalk end and CMCas were evaluated by using XRD shown in Figure 6. The strength of hydrogen bonding and crystallinity contribute to the microstructure of CMCas material. All CMCas samples were less value in a peak of intensity (au) compared to cellulose from asparagus stalk end. Decrease in crystallinity index of cellulose may be due to the reorganization or cleavage of molecular according to alkalization of NaOH solution [32]. The increased aperture among cellulose polymer molecules was also caused by the substitution of monochloroacetic acid (MCA) molecules into the hydroxyl group of cellulose macromolecules easier than the cellulose without treating with an alkali solution [23].

These results are accord with results of various studies which were tested in cavendish banana cellulose [32] and durian rind [8]. The crystallinity index was also decreased when alkalizing by 15% (*w*/*v*) NaOH [32]. Thus, the crystallinity of CMCas decreased by the effect of alkali solution prior to the carboxymethylation reaction on cellulose structure.

**Figure 6.** X-ray diffractograms of (**a**) cellulose from asparagus stalk end and CMCas powder: with (**b**) 20% (*w*/*v*) NaOH, (**c**) 30% (*w*/*v*) NaOH, (**d**) 40% (*w*/*v*) NaOH, (**e**) 50% (*w*/*v*) NaOH and (**f**) 60% (*w*/*v*) NaOH.

### *3.7. Morphology of Cellulose from Asparagus Stalk End and CMCas*

Morphology of cellulose from asparagus stalk end and CMCas powder with different NaOH concentrations were characterized by SEM with an acceleration voltage of 15 kV and 1500× original magnification (Figure 7a–f). The surface of cellulose from asparagus stalk end (Figure 7a) showed a smooth without pores or cracks and with a dense structure; however, the appearance of small fibers has been found in the cellulose. The increase of NaOH concentration affected changes in microstructure of the samples. The morphology of CMCas powder treated with 20% of NaOH concentration emerged small fiber with minimal damage (Figure 7b). By increasing the content of NaOH from 30% to 60% (*w*/*v*), the surface uniformity decreased. The surface of CMCas powder mixed with NaOH concentration at 30% started to peel without cracking as shown in Figure 7c. The surface of CMCas powder has been begun to crack and deform when the powder treated with 40% of NaOH concentration (Figure 7d), due to the degradation of CMC polymer chain. Increasing the NaOH concentration at 50% (Figure 7e), the morphology of CMCas powder increased irregularity surface with more collapsed and indented. The surface was more roughly and totally deformed when CMCas treated with a higher NaOH concentration. Thus, this phenomenon could be inferred that CMCas was synthesized with a high NaOH concentration causing a damage in surface area of cellulose powder. This result was comparable with a synthesized carboxymethyl from rice starch [23] and cassava starch [39]. Alkaline solution probably reduced the strength of structure, loss the crystallinity and changed the stability of the molecular organization of the cellulose causing the etherifying agents to have more access to the molecules for the carboxymethylation processes [36]. Consequently, this result was according to DS value.

2Ό (degree)

**Figure 7.** Scanning electron micrographs of (**a**) cellulose from asparagus stalk end and cmcas powder: with (**b**) 20% (*w*/*v*) naoh, (**c**) 30% (*w*/*v*) naoh, (**d**) 40% (*w*/*v*) naoh, (**e**) 50% (*w*/*v*) naoh and (**f**) 60% (*w*/*v*) naoh. The acceleration voltage was 15 kv under low with 1500×.

### *3.8. Color of Cellulose from Asparagus Stalk End and CMCas*

Color measurement was carried out to evaluate the color formation caused by carboxymethylation reaction. The color values of CMCas were decreasing in *L\** value as NaOH concentrations increased from 20 to 60% (*w*/*v*) of NaOH concentration. The a\* and b\* values showed the highest increased in CMCas treated with 40% (*w*/*v*) of NaOH concentration. Moreover, an increase in NaOH concentration up to 40% decreased the yellowness of CMCas probably due to the first reaction of carboxymethylation (Equation (12)) to produce CMC or sodium glycolate [36]. At higher NaOH concentration (50 and 60%, *w*/*v*), the decreasing of all color values was occurred. Therefore, the carboxymethylation reaction might be the reason the color values of cellulose and CMCas has been changed [35]. The ΔE of the cellulose and CMCas had similar trend compared to a\* and b\* values. These results are akin to the results from the previous studies [23]. The WI of the cellulose and CMCas had same trend with *L\** value (Table 1).


**Table1.**ColorvaluesofcelluloseandCMCassynthesizedwithvariousNaOHconcentration.

### *3.9. Mechanical Properties of CMCas Films*

The tensile strength (TS) and percentage of elongation at break (EB) of CMCas films synthesized with various NaOH concentrations were determined by using the ASTM Standard Method D882-80a (Table 2). The TS value increased gradually with increasing NaOH concentration up to 40% (*w*/*v*) of NaOH. The TS value was correlated with an increasing DS value due to the substituted carboxymethyl groups in anhydroglucopyranose unit affecting an increase in intermolecular force among the polymer chains and the ionic character [32]. Nevertheless, the TS value of CMCas films started to decline at a high NaOH concentration due to the polymer degradation and formation of by-product generated from sodium glycolate increased the reduction of CMC content and thus decreasing the intermolecular forces.

The EB of CMCas films also increased with increasing NaOH concentrations up to 30% (*w*/*v*) of NaOH and drop down after that. At high NaOH concentrations, the crystallinity of the cellulose structure decreased with increasing a flexibility. In addition, CMCas films has been obtained the lower flexibility at 60% (*w*/*v*) of NaOH concentrations because of an occurrence of hydrolysis reaction in cellulose chain [36].

**Table 2.** Mechanical properties for CMCas films synthesized with various NaOH concentrations.


Mean ± Standard deviation values within a column in the same group, followed by the different letters (a–d) are significantly different (*p* < 0.05).

### **4. Conclusions**

This study revealed that *Asparagus officinalis* stalk end could be favorably used as a raw material to produce carboxymethyl cellulose. The alkali solution is the main parameter that affected all characteristics of CMCas synthesized with various NaOH concentrations. The percent yield of CMCas initially increased with NaOH concentrations increase up to 30% (*w*/*v*) and then gradually decreased with further NaOH concentrations increase related with the DS value of CMCas. The lightness (*L\**) of CMCas has been changed after treated with increasing NaOH concentration. The crystallinity of CMCas was found to decrease after synthesized with alkali solution. The highest tensile strength and elongation at break values of CMCas films were found in 40% (*w*/*v*) NaOH-synthesized CMCas films and 30% (*w*/*v*) NaOH-synthesized CMCas films, respectively.

**Author Contributions:** Conceived and planned the whole experiments, P.R.; Conceptualization, Investigation and Methodology, P.R.; wrote the manuscript with input from all authors, W.K., K.J., N.L., P.S., T.C., C.I., S.R.S., T.M.P.N. and P.R.; analyzed the data, W.K., S.P., P.J., W.P. and A.R.; contributed reagents and materials, Y.P. and P.R.; Validation, P.R.; supervised the project, P.R. All authors have read and agreed to the published version of the manuscript.

**Funding:** The present study was partially supported by TRF Senior Research Scholar (Grant No. RTA6280001). This research work was partially supported by Chiang Mai University.

**Acknowledgments:** We wish to thank the Center of Excellence in Materials Science and Technology, Chiang Mai University for financial support under the administration of the Materials Science Research Center, Faculty of Science, and Chiang Mai University. We also acknowledge partial financial supports and/or in-kind assistance from Program Management Unit—Brain Power (PMUB), the Office of National Higher Education Science Research and Innovation Policy Council (NXPO) in Global Partnership Project, Basic research fund, Thailand Science Research and Innovation (TSRI), Chiang Mai University (CMU), Faculty of Science (CoE64-P001) as well as Faculty of Agro-Industry (CMU-8392(10)/COE64), CMU.

**Conflicts of Interest:** The authors declare that there are no conflicts of interest regarding the publication of this paper.

### **References**


## *Article* **The E**ff**ect of Halloysite Nanotubes on the Fire Retardancy Properties of Partially Biobased Polyamide 610**

**David Marset 1, Celia Dolza 1, Eduardo Fages 1, Eloi Gonga 1, Oscar Gutiérrez 1, Jaume Gomez-Caturla 2, Juan Ivorra-Martinez 2,\*, Lourdes Sanchez-Nacher <sup>2</sup> and Luis Quiles-Carrillo 2,\***


Received: 15 December 2020; Accepted: 17 December 2020; Published: 19 December 2020

**Abstract:** The main objective of the work reported here was the analysis and evaluation of halloysite nanotubes (HNTs) as natural flame retardancy filler in partially biobased polyamide 610 (PA610), with 63% of carbon from natural sources. HNTs are naturally occurring clays with a nanotube-like shape. PA610 compounds containing 10%, 20%, and 30% HNT were obtained in a twin-screw co-rotating extruder. The resulting blends were injection molded to create standard samples for fire testing. The incorporation of the HNTs in the PA610 matrix leads to a reduction both in the optical density and a significant reduction in the number of toxic gases emitted during combustion. This improvement in fire properties is relevant in applications where fire safety is required. With regard to calorimetric cone results, the incorporation of 30% HNTs achieved a significant reduction in terms of the peak values obtained of the heat released rate (HRR), changing from 743 kW/m2 to about 580 kW/m<sup>2</sup> and directly modifying the shape of the characteristic curve. This improvement in the heat released has produced a delay in the mass transfer of the volatile decomposition products, which are entrapped inside the HNTs' lumen, making it difficult for the sample to burn. However, in relation to the ignition time of the samples (TTI), the incorporation of HNTs reduces the ignition start time about 20 s. The results indicate that it is possible to obtain polymer formulations with a high renewable content such as PA610, and a natural occurring inorganic filler in the form of a nanotube, i.e., HNTs, with good flame retardancy properties in terms of toxicity, optical density and UL94 test.

**Keywords:** PA610; halloysite nanotubes (HNTs); nanocomposites; flame retardant; cone calorimeter

### **1. Introduction**

The current social awareness about the environmental problems derived from the use of non-renewable polymers and additives is generating a great change in the industry. Moreover, governments are beginning to become concerned about this problem and are starting to develop legislation that favors environmental protection and the use of materials that reduce the harmful impact on nature [1,2]. In particular, a great effort has been made in recent decades to develop and use new materials and additives that are biodegradable and possess sustainable properties as well as a reduced carbon footprint by reducing greenhouse gases during their production [3–6].

In this context, until relatively recently, high performance polymers such as polyamides (PAs) were materials derived entirely from oil; however, new technologies and research have succeeded in obtaining them from monomers, both fully and partially renewable [7,8]. Monomers of biological origin include, for example, brazilic acid, sebacic acid, 1,4-diaminobutane (putrescine), and 1,5-diaminopentane (cadaverine) [9–12]. From these types of monomers, different kinds of polyamides can be obtained. It is always desirable to generate polymers with properties similar to those of petrochemical counterparts. This fact is vital for industry, because obtaining biobased polyamides (bio-PA) that behave similarly to PA6 and PA66 regarding stiffness, and similarly to PA12 regarding flexibility, is increasingly important for economic and ecological issues [13,14].

Currently, more than 6 million tons of polyamides are required annually, with growing demand [15]. In particular, polyamide 6 (PA6) and polyamide 66 (PA66) make up approximately 90% of the total polyamide use in the plastic industry [16]. However, as noted, the development of a "green" route for the production of biobased polyamides (bio-PA) has generated increasing interest due to the inevitable stoichiometric waste associated with the classic petrochemical production routes, which are commonly thought to cause global warming and other environmental problems [8,17]. For example, in some applications, PA6 can be replaced with a biobased variant called PA610, which possesses very similar properties, but it is more ductile as well as it exhibits high renewable content. Because of the fact that the dicarboxylic acid can readily be condensed with the petroleum-based 1.6-hexamethylenediamine (HMDA) obtained from butadiene, the material PA610, containing 60–63 natural content, may be obtained [18].

Within the polymer industry, polyamide 6 (PA6) is a highly relevant engineering polymer, which finds applicability in certain areas where high flame retardant and fire retardant properties are required [19]. The development of flame retardant additives that are attractive from an ecological point of view has become the focus of much effort. In particular, halogen-free flame retardants based on renewable sources have attracted great interest [20]. Many of these are phosphorus-based additives [21,22], such as aluminum hypophosphite (AlHP), with very good results for blends considering sustainable polymers such as polylactic acid [23] or polyvinyl alcohol [24], and others such as PA6 [25]. Other elements that are having large application as halogen-free retardants are ammonium polyphosphate (APP) or antimony trioxide, with great application at present in polyolefins [26,27]. Other materials like polyurethane foams, ammonium polyphosphate [28], or triphenyl phosphate (TPhP) [29,30] are often used as intumescent due to their low toxicity, those being halogen-free materials, and highly efficient. In other cases, flame-retardant additives containing elements other than phosphorus, such as expanded graphite [31], nanoclays, and nanosilicates [32,33] or graphite oxides [34], have been utilized. In this context, the search for this type of more natural additives has found elements such as halloysite nanotubes (HNTs) [35].

Blends containing halloysite nanotubes (HNTs) present great potential in the generation of natural fire retardancy polymeric materials distinguished by their high sustainability and low emission of toxic gases and fumes during combustion [36,37]. Normally, one of the main disadvantages of nanofillers is the low dispersion it presents in a polymer matrix, directly affecting both mechanical and flame-retardant properties [38,39]. However, because of their polar structure, HNTs can be efficiently dispersed in different polyamides [40,41]. Over the last two decades, nanocomposites based on inorganic clay minerals have attracted a lot of attention. In addition, due to their nanoscale structure, nanocomposites may exhibit significant improvements in aspects such as mechanical properties, reduced gas permeability, increased thermal stability, and improved flame retardancy compared to the properties of the polymer without these nanocomposites [23–26]. For this reason, the use of halloysite nanotubes is becoming more attractive. In this context, some authors have incorporated HNTs in polyamides of petrochemical origin such as PA 6 or PA66, providing promising results for blends between 5% and 40% of HNTs regarding mechanical properties and fire protection [42,43]. On the other hand, trying to get away from oil products, authors such as Sahnaoune et al. [44] are introducing these types of natural fillers into biobased polyamides such as Polyamide 11, taking into account as

a disadvantage the fact that this type of polyamide is more expensive and has less applicability in today's industry. For this reason, the aim of this project is to find a balance between sustainability and direct application in industry for Polyamide 610.

In previous works, the impact of the incorporation of HNTs into biobased PA610 at different levels on mechanical, thermal, and morphological properties has been deeply studied [45]. However, the main objective of the work reported here was the use of different levels of halloysite nanotubes (HNTs) to improve the flame-retardant properties of PA610. The objective was to determine how the incorporation of nanotubes into PA610 affects its fire retardancy and flame-retardant properties in high-performance injected parts. In order to determine the properties of the blends, several samples have been characterized using cone calorimetry, limiting oxygen index (LOI), and a calorimetric pump.

### **2. Materials and Methods**

### *2.1. Materials*

Partially BioBased Polyamide 610 (PA610) was supplied by NaturePlast (Ifs, France), in the form of pellets. According to the manufacturer, this is a biobased medium-viscosity injection-grade homopolyamide with a density of 1.06 g/cm<sup>3</sup> and a viscosity number (VN) of 160 cm3/g. This polyamide has 63% of biological content. As flame retardant, the halloysite nanotubes (HNTs) were supplied by Sigma Aldrich (Madrid, Spain) with CAS number 1332-58-7. This material had an average tube diameter of 50 nm and inner lumen diameter of 15 nm. Typical specific surface area of this halloysite was 65 m2/g.

### *2.2. Sample Preparation*

Prior to processing, the biobased PA610 and Halloysite nanotubes were dried at 60 ◦C for 48 h in the dehumidifying dryer MDEO. Both components were mechanically pre-homogenized in a zipper bag. The materials were then fed into the main hopper of a co-rotating twin-screw extruder (Construcciones Mecánicas Dupra, S.L., Alicante, Spain). The screws featured 25 mm diameter with a length-to-diameter ratio (L/D) of 24. The extrusion process was carried out at 20 rpm, setting the temperature profile, from the hopper to the die, as follows: 215–225–235–245 ◦C. The different PA610/HNTs composites were extruded through a round die to produce strands and, subsequently, pelletized using an air-knife unit. In all cases, residence time was approximately 1 min. The four prepared compositions are shown in Table 1.


**Table 1.** Summary of compositions according to the weight content (wt.%) of polyamide 610 (PA610) and halloysite nanotubes (HNTs).

In the final step, the compounded pellets were shaped into square plates of 150 <sup>×</sup> 150 <sup>×</sup> 5 mm3 by injection molding in a Meteor 270/75 from Mateu and Solé (Barcelona, Spain). The temperature profile in the injection molding unit was 220 ◦C (hopper), 225 ◦C, 230 ◦C, and 235 ◦C (injection nozzle). A clamping force of 75 tons was applied while the cavity filling and cooling times were set to 2 and 20 s, respectively.

### *2.3. Material Characterization*

### 2.3.1. Cone Calorimeter Test (CCT)

The cone calorimeter model was 82121 (FIRE Ltd., Surrey, UK) and the tests were performed according to ISO 5660 standard procedures. The dimensions of the samples were 100 <sup>×</sup> 100 <sup>×</sup> 5 mm3. Each sample was wrapped in aluminum foil (0.0025 to 0.04 mm thick) and horizontally exposed to an external heat flux of 50 kW/m2, 25 mm conical distance, and 20 min test time.

### 2.3.2. Limiting Oxygen Index (LOI) and UL94

LOI was carried out in an 82121 model (FIRE Ltd., UK) according to the standard oxygen index test stated in the UNE-EN ISO 4589-2 norm. Type I test pieces and the ignition procedure (A) related only to the upper surface were used. Prior to the test, the specimens were conditioned at 23 ◦C and 50% relative humidity for 24 h. The size of the samples used was 150 <sup>×</sup> 10 <sup>×</sup> 4 mm3. Three samples were studied using the LOI test.

The UL-94 horizontal burn tests were carried out following the testing procedure UL 94:2006; EN 60695-11-10:1999/A1:2003 with a test specimen bar that was 150 mm long, 10 mm wide, and about 5 mm thick.

### 2.3.3. Toxicity and Opacity Test

The smoke density chamber tests the opacity of the emitted fumes according to the EN ISO 5659-2 norm. This test allows to obtain the optical density using a simple chamber; simultaneously, the toxicity of the fumes was determined according to the UNE-EN 17084 norm. The test was carried out in a chamber model NBS Smoke Chamber (Concept Equipment Ltd., Arundel, UK), and a FTIR model MG2030 MKS Instruments, Inc. (San Diego, USA) to study the toxicity.

The dimensions of the samples were 75 <sup>×</sup> 75 <sup>×</sup> 5 mm3. The samples must be conditioned to a constant mass at a temperature of 23 ◦C and a relative humidity of 50% in accordance with ISO 291. Each sample was wrapped in aluminum foil (0.04 mm thick) and exposed to a radiation of 50 kW/m2, 25 mm conical distance, and 600 s test time. Regarding toxicity, a flow rate of 4 L/min was extracted for 30 s at minutes 4 and 8 of the test. Three replicates of each material were performed.

Regarding the calculation of the specific optical density (Ds), Equation (1) was used:

$$D\_s(t) = \frac{V}{AL} \log\_{10} \frac{100}{T(t)} [Adimensional] \tag{1}$$

where *Ds*(*t*) is the specific optical density; *<sup>V</sup> AL* is the ratio between the volume of the camera (*V*), the exposed area of the specimen (*A*), and the length of the light path (*L*). This ratio is equivalent to 132 and, finally, *T*(*t*) is the value of transmittance measured in %.

In relation to the toxicity, the concentration of toxic gas (CO2, CO, HF, HCl, HCN, NO2, SO2, HBr) and optical density of smoke were recorded following the previous norm (UNE-EN 17084). During the smoke toxicity test, the concentrations of eight toxic gases were used as quantifying terms for the conventional index of toxicity (CIT). CIT was calculated using Equation (2):

$$CTT\_G = 0.0805 \sum\_{i=1}^{i=8} \frac{c\_i}{C\_i} \tag{2}$$

where, *CITG* was the conventional toxicity index for general products, *ci* was the concentration of the gas in the chamber, and *Ci* was the reference concentration of the gas.

### 2.3.4. Calorific Value

The equipment used for this test was a PARR 6200 calorimeter (Parr Instrument Company, Moline, IL, USA). The samples in pellet form were turned into fine powder by a grinder and liquid nitrogen was used in order to avoid thermal decomposition in the samples. The temperature of the distilled water was set at 26 ◦C and the closing pressure was set between 3.0 and 3.5 MPa with no air in the inside. The reagents used were distilled water, pressurized oxygen with purity greater than 99.5%, a standardized benzoic acid tablet, and a pure iron wire of 0.1 mm.

### **3. Results**

### *3.1. Cone Calorimeter Test (CCT)*

The calorimetric cone test (CCT) provides a great deal of information on the fire behaviour of the material under study when exposed to a variable radiation source [46]. The CCT is based on the principle of oxygen consumption, simulating the combustion of polymers in real fire situations, demonstrating great utility in research studies, and allowing the development of new materials with excellent fire-retardant properties [47,48]. Table 2 summarizes the main results obtained from this test in relation to thermal parameters and Table 3 summarizes the results of smoke parameters.

**Table 2.** Summary of thermal parameters obtained with the calorimetric cone test (CCT) on the PA610 and HNTs samples.



**Table 3.** Summary of smoke parameters obtained with the CCT on the PA610 and HNT samples.

In relation to the ignition time of the samples (TTI), the incorporation of halloysite reduces the ignition start time by about 25 s in all the cases studied. Furthermore, the total duration of the inflammation does not present great variations, except for PA610/HTN10, which reaches values of 850s. Authors like Marney et al. [49] showed similar behaviour in PA6 mixtures with HNTs, where a reduction in the ignition time of the halloysite mixtures was observed. This factor may be due to the early release of the internal elements of the HNTs, which leads to the formation of small combustible molecules and results in an accelerated decomposition.

The flame retardant effect of aluminum phosphinate is in combination with zinc borate, borophosphate, and nanoclay in polyamide-6Mehmet. It is sometimes difficult to evaluate the performance of incorporating a flame-retardant additive because the results obtained are expressed in terms of time or released energy, thus generating problems of comparison. To overcome this issue, Vahabi et al. [50] have defined a dimensionless concept called the "Flame Retardancy Index" (*FRI*), which allows a very simple comparison between the pure polymer and its fire retardancy composite. This dimensionless concept is born from the following equation:

$$FRI = \frac{\left[THR \cdot \frac{pHRR}{TTT}\right]\_{\text{Next\text{ }Polomer}}}{\left[THR \cdot \frac{pHRR}{TTT}\right]\_{\text{Composite}}} \tag{3}$$

Thanks to the use of the *FRI*, a better comparison can be made between the different values obtained from the calorimetric cone test, obtaining a direct comparison between the pure polymer and its compounds.

If the FRI values in Table 2 are analyzed, it can be seen directly whether the HNTs improve the fire-retardant characteristics. Initially, it is expected that by introducing the flame-retardant additive and dividing the term calculated for the neat polymer by that of its composite with HNTs, a dimensionless quantity greater than 1 is obtained. However, it can be seen how the incorporation of 10% of HNTs into the mixture reduces the value to 0.46, obtaining a very poor value. The incorporation of a higher amount of HNTs progressively improved the previous results. In particular, 30% of HNTs improve by more than 47% of the previous value, but both are below the unit. Following the guidelines of Vahabi's work, any compound that has an FRI value below 1 is taken as the "Poor" level of performance in terms of fire retardancy. Following these general premises, HNTs seem to not provide a fire retardancy improvement when compounded with PA610. This factor may be related to the large fire-retardant capacity of PA, an engineering plastic with intrinsic fire retardancy properties, due to its base structure.

Some characteristic results of the calorimetric cone test are presented above.

### 3.1.1. Heat Release Rate (HRR)

The HRR measured by cone calorimeter is a very important parameter as it expresses the intensity of fire [51]. In this section, Table 2 and Figure 1 show the results related to the heat released by the samples through the CCT. Referring to the maximum peak of heat released (pHRR), most of the samples exhibit a decrease in the heat released thanks to the presence of halloysite, with the exception of PA610/HTN10, which presents a higher peak than the base compound without additive, possibly due to a low concentration of the additive. This reduction of the peak indicates that little energy has been released to the system, verifying the possible application of HNTs as fire retardancy additives in certain applications. In relation to the time when the maximum pHRR value is produced, there seems to be no apparent relationship between this value and the amount of halloysite.

On the other hand, Figure 1 shows the evolution of the heat release rate as a function of time. A heat reduction can be observed for the first 10 min of the test, reaching maximum values around 300 s. The results show a slight increase in heat release in the PA610/HTN10 sample and a very slight reduction in the heat released in the PA610/HTN20 sample, both compared to the polymer without load. However, in the case of 30% halloysite, a significant reduction is achieved in terms of the maximum values obtained, going from 743 to about 580 kW/m2 and directly modifying the shape of the curve. This improvement in the heat released is produced by the carbonization of the halloysite, forming a layer that acts as a barrier that slows down the combustion of the polyamide, making it difficult for the sample to burn. In this context, authors like Marney et al. [42] showed similar results with the incorporation of HNTs in petrochemical polyamides. This happens due to the formation of a thin layer of carbon (or skin) on the surface, which breaks during the first stages of combustion, as shown by the small plateau at about 100, verifying how the modification in the shape of the curve is directly related to the formation of carbon in the external layers [52]. In addition, the HRR increases until it reaches a peak because of the increasing amount of combustible volatile compounds caused by the rise in total temperature of the substrate. The apparent temperature increases because the material properties are changing so that the unexposed surface reaches temperatures close to that of the exposed surface at

the time of ignition. When the release of combustible volatiles has been exhausted, the combustion reaction ceases, and the peak is usually followed by a sharp linear decrease to zero.

**Figure 1.** Heat release rate as a function of time.

The integration of the HRR curves allows to obtain the THR (total heat realised) values. The PA610 produced 128.1 MJ/m2; the introduction of HNTs slightly increased the heat to 164.3 for the PA610/20HNTs. The range of values obtained matches with the proposed by Do ˘gan et al. [53] for a polyamide-6 and different flame-retardant additives.

### 3.1.2. Effective Heat Combustion (EHC)

In the CCT, the effective heat of combustion (EHC) represents the heat released during combustion per unit mass. Figure 2 shows the graph of the results obtained as a function of time. A slight increase in the effective heat of combustion in the samples with HNTs can be seen, which is determined by the values of HRR and mass loss of the different materials during the test. Particularly, it can be seen how the PA610/10HNTs sample has provoked an increase of 20 MJ/kg compared to the sample of PA without any load. Qin et al. [54] showed similar results with the incorporation of montmorillonite to polyamide 66, where the average EHC value of PA66 increased after the addition of nano-loads, which were obtained from decomposition processes. This increase in EHC may be related to a low load concentration in the mixture, which generates an increase in the amount of energy released. This can be seen directly in samples where the amount of HNTs is higher, as it can be observed in 20% and 30% HNT samples, where the EHC is close to the PA610.

**Figure 2.** Effective heat of combustion as a function of time.

### 3.1.3. CO and CO2 Production

Another important aspect to consider during a combustion produced by fire is the amount of CO and CO2 that is released from the burned products, mainly because a great release could cause anoxia conditions, making it difficult to evacuate the site. Table 3 shows the maximum production values for these gases. The analyzer of the equipment used is able to determine the concentration of CO and CO2 during the course of the test, making the analysis of the emission of these gases deeper.

Figure 3 shows the values of CO2 produced during the test in relation to the quantity released in kg per kg of material analysed and according to the release rate. Regarding the emitted quantity, a slight increase can be observed in the 20% and 30% HNT samples compared to the base material, while in the case of the 10% HNT sample, the emission of CO2 doubles the quantity emitted by the base polyamide. On the other hand, referring to the release rate, although it seems that the amount of total CO2 emitted is greater as the additive load is increased, as observed with the 10% and 20% HNT samples, a significant reduction in the maximum peak for the CO2 rate in the case of the 30% HNT sample can be observed. This fact could mean that the multi-layered porous nature of the HNT structure may prevent those evolved gases from entering the combustion zone of the burned sample. In addition, the heat feedback from the flame zone to ensure a faster polymer decomposition may also be restricted by the insulating nature of the carbon structure in the HNTs [55].

**Figure 3.** CO2 production during the CCT test: (**a**) CO2 rate and (**b**) CO2 yield.

In relation to the values obtained for CO, Figure 4 shows how the values obtained both in total production and ratio are relatively low for all samples. However, from a global point of view, it can be seen how the incorporation of HNTs into the PA610 matrix means a slight decrease in the amount of CO generated during combustion. However, these values are not as representative as those obtained for the CO2 produced. This factor indicates that the compounds are burned reasonably efficiently (since carbon monoxide can be measured due to an incomplete combustion) [42]. In this sense, Gilman et al. [56] studied silicate nanocomposites in PA6 and suggested that if EHC, smoke extinguishing area (SEA), and CO yields did not change, flame inhibition of the condensed phase was involved, and this was accompanied by a decrease in PHRR and mass loss ratio (MLR) as well as a change in carbon yield.

**Figure 4.** CO production during the CCT test: (**a**) CO rate and (**b**) CO yield.

### 3.1.4. Rate of Smoke Production (RSP)

The smoke performance of a fire retardancy material is a vital parameter in terms of fire safety. In the case studied, the significant increase in the amount of smoke produced is related to HNTs incorporation into the blends. This increase can be linked to the formation of a carbonized layer of material during combustion produced by halloysite itself. It is this outer layer on the surface of the test piece that could cause the combustion to release a greater amount of smoke. It should be noted that an increase in the amount of carbonized residue is observed at the end of the test in materials with halloysite. The appearance of carbonized combustion products generates an emission of darker fumes, which causes an increase in the RSP as shown in Figure 5. Similar results were reported by Levchik et al. [57] where they showed that the amount of smoke produced by the incorporation of HNTs was almost double than that of the original polymer because HNTs appear to aid the process of carbon formation in the composite material by acting as a "glue" between the HNTs, thus ensuring the formation of a consistent and strong porous carbon layer over the polymer.

The increase in the RSP value during the test is strongly related to the smoke formation as determined by the smoke extinguishing area (SEA) (Figure 6) during the combustion of HNT-containing compounds; however, it does not seem to depend on the amount of HNT. The presence of HNTs seems to accelerate the rate of PA610 smoke production during the first stages of the combustion process, being that this effect is especially marked in the 10% HNT sample. Similar results were reported by Marney et al. [42].

**Figure 5.** Smoke production rate of samples with HNTs.

**Figure 6.** Specific extinction area for PA610 samples loaded with HNTs.

#### 3.1.5. MASS

Regarding the amount of final mass obtained after the test, Figure 7 shows the results obtained in terms of mass loss ratio and percentage of residual mass. In particular, Figure 7a shows an increase in residual mass with the incorporation of HNTs. This increase is closely related to the amount of carbonized mass remaining as waste at the end of the test, which is largely made up of carbonized halloysite. The creation of this layer of carbonized material has been reported by various authors even with the use of other flame retardants such as APP or TiO2 [58]. An intumescent carbon layer may appear on the surface of materials during combustion, creating a physical protective barrier able to

contain heat and mass transfers. The carbon layer limits the diffusion of oxygen to the underlying part of the material or insulates it from heat and combustible gases; it also further delays the pyrolysis of the material. The mass loss decrease was attributed to the formation of carbon and the morphological structure at the surface of the materials [59,60]. Similarly, the result of each test specimen before and after exposure to cone radiation can be seen in Figure 8. In the case of PA610, total disappearance of the material can be seen, while in the rest of the tests, a layer of carbonized material appears, formed mainly by halloysite remains. These remains may be evidence of the increase in smoke generation previously observed.

**Figure 7.** (**a**) Percentage of residual mass during the testing of PA610 samples with HNTs; (**b**) Loss of mass ratio of PA610 samples with HNTs.

**Figure 8.** Visual difference between samples before and after the CCT test: (**a**) PA1010 before, (**b**) PA1010 /10HNTs before, (**c**) PA1010/20HNTs before, (**d**) PA1010/30HNTs before and (**e**) PA1010 after, (**f**) PA1010/10HNTs after, (**g**) PA1010/20HNTs after, (**h**) PA1010/30HNTs after.

In this sense, the incorporation of HNTs in relation to the residual mass is closely related to the values obtained in TGA degradation behavior in the previous work [45], focused on mechanical, thermal, morphological, and thermomechanical properties. The previous TGA results on the PA610/HNT system indicated a slight increase in the onset degradation temperature from 417.4 ◦C up to 419.6 ◦C, as well as the maximum degradation rate temperature (from 461.5 ◦C up to 466.6 ◦C). The residual weight is, as expected, close to 10, 20, and 30 wt.%, as HNTs do not decompose below 700 ◦C. In the previous work, scanning electron microscopy images (FESEM) also revealed good particle dispersion

(especially for 10 wt.% and 20 wt.% HNTs). It can be seen how the residual mass of the compositions are almost coincident with the nominal HNTs content.

On the other hand, Figure 7b shows the mass loss ratio (MLR), which is a relevant factor in terms of selecting a fire retardancy additive, as it provides valuable information about the behavior against fire of the analyzed materials. The results obtained in the MLR are closely related to those observed in the smoke extinction area (SEA) in Figure 6, where the samples with halloysite present lower mass loss than the PA610 samples. Some authors have shown how the HNTs reduce the fire hazard of the butadiene-acrylonitrile rubber (NBR) vulcanizates. They clearly extend the time to ignition (TTI), and positively influence the average mass loss rate (MLR) parameter. Because of the synergetic relation between flame retardants and halloysite nanotubes, the parameters connected with the amount of heat released during NBR composites' combustion are reduced [36].

### 3.1.6. The Maximum Average Heat Rate of Emission (MARHE)

The maximum average heat emission index (MARHE) is a parameter used in the EN 45545-2:2013+A1:2015 standard to classify the materials to be tested for railway applications. This value is obtained by dividing the maximum heat emission value recorded (in kW) by the area of the test specimen (0.01 m2). The maximum heat emission value per unit area (kW/m2) alongside other results such as opacity and smoke toxicity are used to classify the tested material according to the applicable risk stated in the standard norm stated above. In this context, Table 4 shows MARHE's results obtained in the tests carried out.

**Table 4.** Summary of results of maximum average heat emission index (MARHE).


The values obtained in all cases were very high. The PA610 shows values above 335 kW/m2 and the incorporation of HNTs increases this value up to 409 kW/m2 for a concentration of 10% HNTs. In the European context, the MARHE value must be inferior to 90 kW/m<sup>2</sup> for these materials to be used in railway applications; the green composites studied here exhibit higher MARHE values (330–400). Additional flame-retardant additives or protective coatings for these biocomposites should be able to reduce their MARHE value, but for the time being, these green composites are not suitable for European rail applications due to the limitation exposed [61].

### *3.2. Limiting Oxygen Index (LOI) and UL94 Results*

This test is defined as the minimum percentage of oxygen needed in a mixture in order to maintain the combustion of the sample after ignition. LOI tests are widely used to evaluate fire-retardant properties of materials, especially for the screening of fire-retardant polymer formulations. In this context, Figure 9 shows the LOI values obtained for PA610 with different concentrations of HNTs in their structure.

The mean of the results obtained for the PA610 without HNTs stands at 27.2%. This value is relatively high for this type of polymer. Other authors have reported values close to 20–25% for PA6 [42,53], verifying the good application that this type of biopolyamide can have from the point of view of flammability. The inclusion of HNTs means a slight decrease in the LOI values, standing at 26% for a 30% load of HNTs. Authors like Li et al. [62] showed how the incorporation of 2% HNTs into PA6 slightly improved the LOI values of this polymer, but always remained in values below 25%. In general, from the experimental results obtained for LOI, a decrease in the oxygen limit value could be observed. This means that the increase of halloysite in the polyamide makes the combustion of the

material easier in low oxygen concentration conditions. In spite of this fact, variations in numerical results do not suppose a considerable change leaving aside the facility for ignition of the compound (differences smaller than a 2% of concentration in volume of O2). Similar results have been reported by Sol et al. [63], where the increase of HNTs in the mixtures supposed a slight decrease in the LOI values, but always stayed at values superior to 24% of LOI.

The incorporation of HNTs in large quantities does not improve flammability for PA610 due to its good initial results. Authors like Vahabi et al. [64] have reported that the best results of HNTs as flame retardants are found in percentages close to 10%, verifying the results obtained in this experiment relating to LOI values.

**Figure 9.** Graphic representation of the limiting oxygen index (LOI) values of each sample.

The UL-94 test shows that all the samples comply remarkably with the test, providing in all cases a V-0 classification. This factor verifies that the incorporation of HNT does not impair the original properties of PA610, showing promising results.

### *3.3. Smoke Density and Toxicity Analysis*

Among the parameters that can be obtained by means of the CCT, the specific optical density (Ds) is a measurement of the degree of opacity of smoke, taken as the negative decimal logarithm of the relative light transmission. Figure 10 shows the optical density evolution along time. Except for the 20% HNTs mixture, the incorporation of nanotubes generates a slight decrease in the maximum values obtained for optical density as the halloysite load in the polyamide matrix increases. In addition, an increase in the time needed to generate smoke can be seen in the samples with HNTs. This type of behavior is very important in applications where properties against fire are needed, verifying the applicability of these types of natural loads in different fields.

**Figure 10.** Evolution of optical density as a function of time.

In addition, the representation of density evolution as a function of time is shown according to the UNE-EN ISO 5659 standard. It is also important to point out the specific optical density values after 10 min of testing, as well as the maximum value (Dsmax). In this context, Table 5 shows values obtained by Equation (1). These values are closely related to those that can be seen in Figure 10.

**Table 5.** Summary of results of the maximum specific optical density values (Dsmax) and at 10 min of testing (Ds10).


The results obtained here are very similar to those obtained by Zhang et al. [65]. These authors reported that the addition of 5% in weight of HNT in a thermosetting resin managed to reduce the density of the smoke emitted by the mixtures. This result means that the incorporation of this additive generates a synergetic effect on the suppression of smoke from the resin, PA610 being the case studied. The reduction in optical density may be related to good dispersion of HNTs in the PA610 matrix. As already verified by mechanical properties, reducing agglomerates makes the HNTs act more adequately.

The smoke toxicity test procedure for railway industry applications follows EN 45545-2/ISO 5659-2 standards [66] performed at 50 kW/m2. Related to toxicity, one of the main causes of death in cases where fire is involved is toxic gases generation [48]. Table 6 shows the results obtained in terms of toxicity of the analysed fumes, such as CO2, CO, HCl, HF, HCN, NO2, SO2, and NO. It can be seen how the incorporation of HNTs produces, in most cases, a clear reduction in the emission of certain gases, such as CO2 and NO2. In particular, it can be seen how the 30% HNTs mixture reduces this value from 437.6 kg/kg to 151.2 kg/kg in the case of CO2. Those are quite promising values referring to this type of material. Attia et al. [67] showed similar gas reduction results with the incorporation of different HNT loads in an ABS matrix. In addition, certain charges present in silicates, such as iron oxides, can participate in the flame-retardant mechanism by trapping radicals during polymer degradation, thus improving thermal stability and flammability properties for these nanocomposites [68,69].


**Table 6.** Volumetric fraction of the combustion gases.

In order to make a deeper analysis of the final toxicity of the samples, the CITG was obtained. This value is a dimensionless index that provides information about the overall toxicity of all the combustion gases analyzed.

Table 7 shows the results obtained from the CITG index. It can be seen that most of the samples with halloysite present a notable reduction in emitted gases concentration compared to pure PA610. However, the 10% HNTs sample shows a slight increase in the final concentration of emitted gases, which results in a higher index in comparison to the base material (PA).



### *3.4. Calorific Value*

The calorific value test is obtained by means of a calorimetric pump. It is used to establish the combustion power in MJ/kg of the material to be characterized, and to analyze the existing difference made by the incorporation of fire retardancy additives. To carry out this test, the material is first introduced in a crucible together with an ignition wire, inside an airtight container with oxygen under pressure. It is then filled with pressurized oxygen gas. Next, the determination of combustion energy is carried out, so it is introduced inside a container with water and two electrodes are connected to each side of the previously introduced ignition wire. Finally, by means of agitation made by the equipment, the temperature of water is maintained homogeneous so that the temperature probe determines the increase in centigrade degrees of water produced by the material combustion.

Table 8 shows combustion energy and temperature difference values for the characterized samples. It can be seen how the incorporation of HNTs supposes a clear reduction of the combustion heat, verifying the application of these types of natural loads as fire retardancy additives. The combustion energy value for PA610 is 33.2 MJ/kg, while the samples with HNTs as additives, particularly the mixture with 30% HNTs, reduce the combustion value to 23.9 MJ/kg. This reduction is a clear advantage of this type of nanocomposites, since they have direct impact on flammability. During combustion, a halloysite-rich barrier is formed, delaying mass loss/transfer (i.e., less fuel available in the flame zone). Additionally, the refractory nature of the HNT limits the heat conduction and results in the reduction

of available energy, as proposed by Marney or Smith [42,70]. There have been proposed several fire retardancy mechanisms that HNTs can exert. One of the proposed mechanisms is the typical formation of a char layer that prevents direct contact of the polymer with oxygen and slows down the escape of volatile products. Another mechanism considers that iron oxides contained in HNTs can trap free radicals during the decomposition, thus allowing to delay the burning process as observed in other iron-containing nanoparticle systems for improved flammability [71]. Nevertheless, the amount of iron oxide in HNTs is relatively low (0.29 wt.%), and there would be not enough iron oxide to play main role in fire retardancy with HNTs [72]. Du et al. [68] reported that the particular nanotube structure of HNTs allows entrapment of the decomposition products in the lumen, with a positive effect on delaying the mass transport and thus, leading to increased thermal stability. This phenomenon that allows to improve the fire retardancy properties has been represented in Figure 11. Authors like Hajibeygi et al. [73] obtained similar energy values for polyamide, besides corroborating the reduction of heat release thanks to the incorporation of additives such as zinc oxide (ZnO) nanoparticles. On the other hand, regarding the incorporation of natural nanocomposites as fire retardancy additives, Majka et al. [74] showed very similar heat release results for polyamide 6 with montmorillonite (MMT), justifying the incorporation of natural additives as fire retardancy elements.

**Figure 11.** Scheme of the fire retardancy enhancement by HNTs by the entrapment of volatile decomposition products and delaying the mass transfer in (**a**) bioPA610 and (**b**) bioPA610+HNTs.

This improvement in heat release may be closely related to the good dispersion obtained by HNTs in PA610. This factor can be seen directly in the mechanical properties section, and in particular, in the morphology of the samples from the previous work [45].

**Table 8.** Summary of calorific values results obtained in the PA610/HNT samples.


### **4. Conclusions**

This work shows that the incorporation of nano-loads with fire-retardant properties, such as HNTs, can be effectively used as new reinforcement elements in partially biobased PA610 parts prepared by conventional industrial processes for thermoplastic materials, such as injection molding.

In relation to cone calorimetric values, the incorporation of 30% HNTs achieved a significant reduction in terms of the maximum values obtained of HRR, changing from 743 kW/m2 to about 580 kW/m<sup>2</sup> and directly modifying the shape of the curve. This improvement in the heat released is produced by the entrapment of the volatile decomposition products into the HNTs' lumen, which slows down the combustion of the polyamide, making it difficult for the sample to burn. However, in relation to ignition time of the samples (TTI), the incorporation of HNTs reduces the ignition start time about 20 s. This factor causes the FIR value of the compounds with HNTs to be below 1, showing poor behavior in terms of fire-retardant characteristics. In addition, the incorporation of these natural nano-loads favors the delay of the time needed to provoke an increase in generated smoke. From the point of view of gas emission and average toxicity, the CITg index, the incorporation of 20 and 30% HNTs, generates a clear reduction in emitted gases concentration, close to 50%, compared to neat PA610.

On the other hand, considering released energy during combustion, a clear reduction can be seen in all the cases where HNTs have been added as nano-loads. The incorporation of 30% of HNTs generates a reduction of 9 MJ/kg of energy released in relation to pure PA610. The good dispersion analysed in the previous work between the HNTs and the PA610 matrix could be the cause of the good results obtained in energy release. Because it prevents HNTs agglomerates from being generated and avoids the loss of fire properties.

The presence of HNTs in the PA610 matrix means a notable reduction in the number of toxic gases emitted. Especially in the case of the 30% HNTs mixture, which reduces the quantity of CO2 emitted from 437.6 kg/kg to 151.2 kg/kg. The results obtained indicate that it is possible to obtain compounds with high renewable content such as PA610, and natural inorganic filler with nanotube structure, that is, HNTs, which exhibit acceptable fire-retardant properties, at the same time that it is possible to obtain a highly balanced material from the mechanical, thermal, and flame-retardant point of view. This improvement in properties against fire is very relevant in many applications where fire safety is crucial.

**Author Contributions:** Conceptualization, L.S.-N., L.Q.-C., and E.F.; methodology, L.Q.-C., C.D., and D.M.; validation, O.G., E.F., and E.G.; formal analysis, L.Q.-C., J.G.-C., and D.M.; investigation, E.M., L.S.-N., and D.M.; data curation, J.I.-M., D.M., and C.D.; writing—original draft preparation, L.Q.-C. and D.M.; writing—review and editing, J.G.-C. and L.Q.-C.; supervision, E.G., L.Q.-C., and L.S.-N.; project administration, E.F., L.S.-N., and L.Q.-C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Ministry of Science, Innovation, and Universities (MICIU) project numbers MAT2017-84909-C2-2-R and AGL2015-63855-C2-1-R.

**Acknowledgments:** J.I.-M. wants to thank Universitat Politècnica de València for his FPI grant from (SP20190011) and Spanish Ministry of Science, Innovation and Universities for his FPU grant (FPU19/01759). AITEX wants to thank the European Regional Development Fund (ERDF) from the European Union, for co-funding the project "NABITEX—Innovative technical textiles based on SUDOE natural fibers to be applied in Habitat Sector" through the Interreg SUDOE Program. This publication (communication) is the sole responsibility of its author. The Management Authority is not responsible for the use that may be made of the information herein disclosed

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*

## **Hydroxypropyl Methylcellulose E15: A Hydrophilic Polymer for Fabrication of Orodispersible Film Using Syringe Extrusion 3D Printer**

**Pattaraporn Panraksa 1, Suruk Udomsom 2, Pornchai Rachtanapun 3, Chuda Chittasupho 1,4, Warintorn Ruksiriwanich 1,4 and Pensak Jantrawut 1,4,\***


Received: 29 October 2020; Accepted: 11 November 2020; Published: 12 November 2020

**Abstract:** Extrusion-based 3D printing technology is a relatively new technique that has a potential for fabricating pharmaceutical products in various dosage forms. It offers many advantages over conventional manufacturing methods, including more accurate drug dosing, which is especially important for the drugs that require exact tailoring (e.g., narrow therapeutic index drugs). In this work, we have successfully fabricated phenytoin-loaded orodispersible films (ODFs) through a syringe extrusion 3D printing technique. Two different grades of hydroxypropyl methylcellulose (HPMC E5 and HPMC E15) were used as the film-forming polymers, and glycerin and propylene glycol were used as plasticizers. The 3D-printed ODFs were physicochemically characterized and evaluated for their mechanical properties and in vitro disintegration time. Then, the optimum printed ODFs showing good mechanical properties and the fastest disintegration time were selected to evaluate their drug content and dissolution profiles. The results showed that phenytoin-loaded E15 ODFs demonstrated superior properties when compared to E5 films. It demonstrated a fast disintegration time in less than 5 s and rapidly dissolved and reached up to 80% of drug release within 10 min. In addition, it also exhibited drug content uniformity within United States Pharmacopeia (USP) acceptable range and exhibited good mechanical properties and flexibility with low puncture strength, low Young's modulus and high elongation, which allows ease of handling and application. Furthermore, the HPMC E15 printing dispersions with suitable concentrations at 10% *w*/*v* exhibited a non-Newtonian (shear-thinning) pseudoplastic behavior along with good extrudability characteristics through the extrusion nozzle. Thus, HPMC E15 can be applied as a 3D printing polymer for a syringe extrusion 3D printer.

**Keywords:** 3D printing; syringe extrusion 3D printing; hydroxypropyl methylcellulose; orodispersible film; phenytoin

### **1. Introduction**

Over the last few decades, there has been a growing interest in the use of three-dimensional (3D) printing technology within the medical and pharmaceutical fields to fabricate the customizable solid dosage forms that suit different needs, preferences and individual characteristics of each patient [1]. Three-dimensional printing is a manufacturing method that can fabricate 3D-printed

products of any shape and size on-demand from digital design software through depositing materials layer-by-layer [2]. This technology involves three commonly used techniques: printing-based inkjet (IJ) systems, nozzle-based deposition systems or extrusion (solid or semi-solid)-based printing technique and laser-based writing systems. Among these, the extrusion-based printing technique has been recognized as the most popular technique for the fabrication of solid oral dosage forms owing to their excellent capability to print with a wider selection of polymers and drugs at room temperature and the capability to incorporate high amounts of drugs with low-cost [1,3]. Numerous studies have been performed regarding the benefits of extrusion-based 3D printing to design various novel dosage forms such as polypills, gastro-floating tablets and orodispersible films (ODFs) [4]. ODFs are the relatively novel dosage form prepared by using hydrophilic polymers and designed to rapidly disintegrate within a minute in the buccal cavity, without requiring water [5]. This dosage form exhibits several advantages over other oral dosage forms, including ease of administration to pediatric and geriatric patients experiencing dysphagia (swallowing difficulty), dose flexibility and improving the bioavailability of drugs due to high vascularity and high permeability in the buccal cavity [6]. The major advantages of preparing an ODF by 3D printing over the standard film solvent casting are the ability to print objects with different filling (hollow, matrix or full) and the dose of drugs can be controlled by calculating the material consumption during the resizing of the printed object at the design stage, which is proper for personalized therapy. Moreover, 3D-printed films can be formulated with less amount of time.

A variety of hydrophilic polymers such as polyvinyl alcohol (PVA), polyvinylpyrrolidone (PVP), polyethylene glycol (PEG), hydroxypropyl cellulose (HPC) and hydroxypropyl methylcellulose (HPMC) are used as film-forming polymers for the preparation of ODFs, and most of them can also be used as printing materials for extrusion-based 3D printers [7,8]. Hydroxypropyl methylcellulose (HPMC), also known as hypromellose, is widely implemented in pharmaceutical manufacturing as a binder, thickening agent, hydrophilic matrix material and film-forming material. It is classified into several grades based on viscosity, degree of hydroxypropyl substitution and degree of methoxy substitution. The low viscosity HPMC grades (e.g., HPMC E3, HPMC E5 and HPMC E15) are often used for ODFs preparation and suitable for extrusion-based 3D printing of oral dosage forms [8–10]. Moreover, the use of HPMC, which is a hydrophilic polymer, can further be advantageous in terms of enhancing the solubility and dissolution of poorly water-soluble drugs in the manufacture of solid dispersion. However, there are still limited studies available on the preparation of 3D-printed ODFs while using low viscosity HPMC as a film-forming polymer. In a previous study, levocetirizine dihydrochloride ODFs consisting of HPMC E15 and pregelatinized starch were prepared using semi-solid extrusion (SSE) 3D printer. The 3D-printed ODFs exhibited good flexibility and rapid drug release in vitro by dissolving completely in two minutes [11]. In addition, previous work on the extrusion-based 3D-printing of HPMC in the pharmaceutical field can be found [12]. The extrusion-based (fused-deposition modeling) 3D printer was used to fabricate 3D-printed tablets by using their developed HPMC filament.

Phenytoin, which was selected as a model drug in this study, is an antiepileptic drug widely used in the treatment of partial seizures, generalized seizures and status epilepticus [13]. It belongs to the Biopharmaceutical Classification System (BCS) class II drug in which its bioavailability is limited due to poor water solubility (32 μg/mL). Various approaches were employed to overcome the solubility problem. Solid dispersion of the drug in a hydrophilic polymer is also one of the promising techniques for enhancing its solubility. There are many available dosage forms of phenytoin in the market, such as oral suspension, chewable tablets, capsules, and intravenous injections. However, the commercial production of orodispersible phenytoin dosage form has not yet been available. The development of phenytoin ODF has numerous advantages over conventional dosage forms such as convenience for patient administration, accurate drug dosing, rapid onset of action with increased bioavailability due to bypassing hepatic first-pass effect and noninvasiveness. Moreover, phenytoin ODF can be used for dysphasic and schizophrenic patients and can be taken without water due to their ability to disintegrate within a few minutes to release medication in the mouth.

Consequently, the present study aimed to assess the possibility of phenytoin ODFs fabrication by syringe extrusion 3D printer using two different low viscosity grades of HPMC (HPMC E5 and HPMC E15) as film-forming polymers. The extrudability and printability of different grades of HPMC were investigated and discussed. Then, the developed 3D-printed ODFs were evaluated for their physicochemical properties, mechanical properties, in vitro disintegration time and in vitro release profiles. The most important impact of our work is the use of the syringe extrusion 3D printer developed by Biomedical Engineering Institute, Chiang Mai University, Thailand and the use of a polymer (HPMC E15) with optimum viscosity as a single printing material to fabricate phenytoin-loaded ODFs for leading to personalized medicine. Until now, there has been no previous study/experiment that has used this syringe extrusion 3D printer to fabricate 3D-printed products. Our syringe extrusion 3D printer was developed to print varieties of fluid gels like materials, such as hydrogels and pastes. Moreover, it can control the printing material temperature by using a temperature control system on the syringe socket. This temperature control system will help control the viscosity of printing material and keep the printing material in a semi-solid state, which makes the material printable through the 3D printer.

### **2. Materials and Methods**

### *2.1. Materials*

The model drug, 5,5-diphenylhydantoin (phenytoin: PT), with purity >98%, was purchased from Merck (Darmstadt, Germany). Hydroxypropyl methylcellulose E5 (HPMC E5, AnyCoat®-C AN5, substitution type 2910, viscosity 5 mPa·s) and Hydroxypropyl methylcellulose E15 (HPMC E15, AnyCoat®-C AN15, substitution type 2910, viscosity 15 mPa·s) were purchased from Lotte Fine Chemical Co., Ltd. (Seoul, South Korea). Refined glycerin was purchased from Srichand United Dispensary Co., Ltd. (Bangkok, Thailand). Propylene glycol (PG) was purchased from Dow Chemical Company (Midland, MI, USA). Ethanol ≥95% (Bangkok Alcohol Industrial Co., Ltd., Bangkok, Thailand) and distilled water were used as the solvents for preparing the printing dispersions. All of the other reagents were analytical grade.

### *2.2. Determination of Extrudability and Printability*

### 2.2.1. Rheological Characterization

To investigate the printability of the different concentrations of two different grades of HPMC that were proposed for use in syringe extrusion 3D printing, the rheological and dimensional accuracy tests of 10, 12.5, 15 and 20 % *w*/*v* of HPMC E5 and HPMC E15 were conducted. The samples were prepared by dispersing HPMC E5 and HPMC E15 in ethanol–water mixtures (9:1) and stirred for 4 h with a magnetic stirrer.

The rheological behaviors of HPMC E5 and HPMC E15 were investigated using a plate-and-plate Brookfield Rheometer (Brookfield Rheometer R/S, P25 DIN plate, Brookfield engineering laboratories, Middleboro, MA, USA). The measurements were carried out by measuring the shear stress and viscosity with varying the shear rate ranging from 0 to 100 s−<sup>1</sup> at 25 ◦C. The gap between plate and base was set at 1 mm. All the experiments were carried out in triplicate. The flow behavior index (*n*) and consistency coefficient (*K*) were calculated using the power-law model equation:

$$
\mathfrak{r} = \mathbb{K}\dot{\mathcal{V}}^{\mathbb{w}}
$$

where <sup>τ</sup> is the shear stress (Pa), . <sup>γ</sup> is the shear rate (s<sup>−</sup>1), and *<sup>K</sup>* is a consistency coefficient (Pa·s*n*)

2.2.2. 3D Printer, Design and Printing Parameters

In this study, the syringe extrusion 3D printer (Figure 1) developed by the Biomedical Engineering Institute of Chiang Mai University was used to fabricate the 3D-printed ODFs. This customized syringe extrusion 3D printer is based on a core-XY 3D printer in which an extrusion nozzle can move in an X and Y axis and a separate building plate on a *Z*-axis while printing layer-by-layer for generating 3D structure to prevent vibration on the sample. A custom syringe-based extruder was designed and built in-house for precision deposition. In Figure 1, the syringe-based extruder used a stepper motor to move a plunger of a 10 mL syringe via a direct lead screw drive. The syringe extrusion 3D printer was controlled by a computer and a user interface on the printer. Before 3D printing, an object was designed using an open-source program and was divided into numerous two-dimensional (2D) layers with a defined thickness, infill and speed of printing. These 2D layers can be piled up by selectively adding the desired materials in a highly reproductive layer-by-layer manner under the instruction of computer-aided design (CAD) models.

**Figure 1.** Schematic diagram of syringe extrusion 3D printer.

The 3D printing process was performed using a syringe extrusion 3D printer. Initially, as shown in Figure 2, the 3D designs and models of the printed ODF with the dimension of 32.5 mm width <sup>×</sup> 32.5 mm length <sup>×</sup> 0.19 mm height were obtained using Tinkercad® software (2020, Autodesk Inc., San Rafael, CA, USA). The 3D models were constructed by a simplified constructive solid geometry method and exported as a 3D printer readable stereolithography (.stl) file format to Repetier-Host software version 2.1.6 (Hot-World GmbH & Co. KG., Willich, Germany). Then, the .stl file was sliced and converted to a 3D printable code (G-code) by the open-source Slic3r software version 1.3.1 (GNU Affero general public license, version 3). Thereafter, the samples were transferred into a 10 mL disposable syringe (14.5 mm diameter), and the 3D models were printed with a syringe extrusion 3D printer equipped with a single head printing extruder nozzle with a diameter of 0.51 mm (21 G). The printing process was conducted at 25 ◦C, 10 mm/s printing speed and 120 mm/s nozzle traveling speed, and the printing parameters were preset as the following: the layer height was 0.19 mm, the fill angle was 45◦, the perimeters were 2, and the infill was defined as rectilinear with 100% ratio (For the measurement of the diameter of printed filaments, the infill was set as 0% of the volume, and the perimeter was 1).

**Figure 2.** Computer-aided design (CAD) of the 3D-printed orodispersible films (ODF) (length 32.50 mm, width 32.50 mm and height 0.19 mm).

### 2.2.3. Dimensional Accuracy Tests

The diameters of printed filaments (0% infill) and area of printed ODFs (100% infill) were determined using photos taken by a digital camera (Canon EOS 750D with an 18–55 mm lens, Canon, Inc., Tokyo, Japan) and measured with ImageJ software version 1.52 (National Institutes of Health, Bethesda, MD, USA; available at https://imagej.nih.gov/ij/index.html). Then, the shape fidelity factor (SFF), which is the ratio between the 3D-printed area and CAD model area, was calculated using the following equation:

$$\text{Shape fidelity factor} = \frac{\text{Prrinted area} \left(cm^2\right)}{\text{CAD model area} \left(cm^2\right)}$$

### *2.3. Fabrication of 3D-Printed Orodispersible Film*

In order to prepare the printing dispersions for fabrication, the following steps were taken. First, HPMC and phenytoin were dispersed in specific proportions as listed in Table 1 in the mixed solvent of ethanol and distilled water (9:1, *v*/*v*) and stirred for at least 3 h with a magnetic stirrer at room temperature (25 ± 2 ◦C). Subsequently, the two different plasticizers (glycerin and propylene glycol) were added into the dispersions at 10% *w*/*w* of the polymer. The dispersions were gently stirred for 10 min and then left until all of the air bubbles disappeared. Ten milliliters of the dispersion were loaded into a 10 mL disposable syringe attached with 21 G needle tips and then printed with the parameters described in Section 2.2.2. The total printing time was approximately 3.5 min for each film. After printing, the printed ODFs were placed at room temperature for 15–30 min to complete the drying process.


**Table 1.** Composition of different formulations of 3D-printed ODFs.

### *2.4. Characterization of 3D-Printed ODFs*

#### 2.4.1. Morphological Characterization

The morphology of the printed ODFs was investigated by scanning electron microscopy (SEM) with a JEOL scanning electron microscope (JSM-5410LV, JEOL, Ltd., Peabody, MA, USA) at 10 kV under low vacuum mode. The film characterizations were performed without any coating solution at magnifications of ×500. The thickness and surface of the films were evaluated.

### 2.4.2. Film Thickness and Weight Variation

The thickness and weight variation were determined by selecting ten printed ODFs randomly. The thickness of each 3.25 cm × 3.25 cm sized film was measured at three points by an outside micrometer (3203-25A, Insize Co, Ltd., Suzhou New District, Jiangsu, China). Weight variation was evaluated by an analytical balance (PA214, Ohaus Corporation, Parsippany, NJ, USA). The average thickness (in mm) and weight (in g) of printed ODFs with the standard deviation were calculated. The measurements were conducted in triplicate.

### 2.4.3. Mechanical Properties of 3D-Printed ODFs

Mechanical strength tests were carried out on square ODFs pieces (3.25 cm × 3.25 cm) with a texture analyzer TX.TA plus (Stable Micro Systems, Surrey, UK) using a cylindrical stainless probe (2 mm diameter) with a plane flat-faced surface (probe contact area =3.14 mm2). The film was fixed on the plate with a cylindrical hole of 9.0 mm-diameter (area of the sample holder hole =63.56 mm2), and a cylindrical probe was moved down at a velocity of 1.0 mm/s (Figure 3). Measurement started when the probe had contacted the sample surface (triggering force). The probe moved on at constant speed until the film was torn. The maximum force (N), distance (mm) and slope of the linear region of the force–time curve (N/s) were recorded. All of the experiments were conducted in triplicate for each sample at room condition (25 ◦C, 70% relative humidity). The mechanical strength of the film was characterized by puncture strength (MPa), elongation to break (%) and Young's modulus (MPa), which were calculated using the following equations [14,15]:

$$Pumclure\,strength = \frac{F\_{max}}{A r\_p}$$

where *Fmax* is the maximum force required to tear the film (N) and *Arp* is the probe contact area (mm2).

$$Elongation\text{ to }break = \left(\frac{\sqrt{a'^2 + b^2} + r}{a} - 1\right) \times 100$$

where *a* is the radius of the film in the radius of sample holder cylindrical hole, *a* is the length of the film sample is not torn by the probe (*a* − *r*), *b* is the probe displacement, and *r* is the probe radius.

$$\text{Young's modulus} = \frac{\text{Slope of } force - time \text{ curve (N/s)}}{\text{Film thickness} \times \text{Probe speed}}$$

$$\text{Figure 3. } Count.$$

**Figure 3.** Rheological behaviors of hydroxypropyl methylcellulose (HPMC) E5 and HPMC E15 at different concentrations: (**a**) stress versus shear rate; (**b**) viscosity versus shear rate and fitting power-law model.

### 2.4.4. In Vitro Disintegration Time Study

The disintegration test method used in this study was modified from Preis et al. (2014) [16]. The printed ODF was clamped between the sample holder and the magnetic clip (attached to the bottom side of the film). The magnetic clip had a weight of 3 g (0.03 N) that represented the approximate minimal force applied by the human tongue. Then, the attached film was half-immersed (50%) in 70 mL of simulated salivary fluid pH 6.8, which was prepared according to Marques et al. (2011) [17], at 37 ± 0.5 ◦C. The time required for the film to break and the magnetic clip to drop down was recorded visually and noted as in vitro disintegration time. All studies were performed in triplicate for each formulation. The optimum formulations in terms of mechanical properties and disintegration time were selected for further experiments.

### *2.5. Phenytoin Content*

Three printed ODFs of size 3.25 × 3.25 cm dimension of each selected formulation (E5-PT-G and E15-PT-G) were taken in the vials containing 5 mL of distilled water and stirred at 600 rpm until the film was completely dissolved, after which 15 mL of ethanol was added and then set aside to ensure complete solubility of phenytoin. The solution was properly diluted with Tris buffer pH 6.8 with 1% *w*/*v* sodium lauryl sulfate (SLS) [18] and filtered through a 0.45 μm membrane filter. The phenytoin content was then determined by using HPLC. An HPLC system (HP 1100 Series HPLC, Agilent Technologies, Inc., Santa Clara, CA, USA) equipped with a Capcell Pak AQ 250 mm × 4.6 mm C18 column having a particle size of 5 μm (Shiseido Co. Ltd., Tokyo, Japan) was utilized. The mobile phase consisted of 23% *v*/*v* acetonitrile, 27% *v*/*v* methanol and 50% *v*/*v* of pH 3.0 phosphate buffer solution. The pump flow rate was maintained at 1.0 mL/min, and the injection volume was 10 μL. The UV detection wavelength was set as 240 nm. The phenytoin contents were calculated from the standard calibration curve of phenytoin in Tris buffer pH 6.8 with 1% *w*/*v* SLS, which demonstrated linearity with a high correlation coefficient (r<sup>2</sup> = 0.9999). The following regression equation was obtained: *y* = 5.7615*x* − 1.8235, where *y* was the peak area and *x* was the concentration of phenytoin (μg/mL). All the measurements were performed in triplicate, and the average percentages of phenytoin content with the standard deviation were calculated.

### *2.6. In Vitro Phenytoin Release Study*

The in vitro release test of printed ODFs was carried out in 300 mL of Tris with 1% *w*/*v* SLS buffer solution (pH 6.8) maintained at 37 ± 0.5 ◦C at a stirring rate of 50 rpm. At predetermined time intervals (1, 3, 5, 10, 15, 30 and 60 min), the samples were withdrawn, and then an equivalent volume of fresh dissolution medium was replaced in order to maintain sink conditions throughout the experiment. Withdrawn samples were filtered and analyzed by HPLC, as described in the drug content studies. The cumulative percentage of drug release was calculated using the standard calibration curve of phenytoin in Tris buffer pH 6.8 with 1% *w*/*v* SLS. All experiments were performed in triplicate.

### *2.7. Statistical Analysis*

All the data were presented as mean ± SD. One-way ANOVA was used to evaluate the significance of differences at the significance level of *p*-value < 0.05. Statistical analysis was performed using SPSS software version 16.0 (SPSS, Inc., Chicago, IL, USA).

### **3. Results and Discussion**

### *3.1. Rheological Characterization and Printability Study*

The flowability and extrudability of printing dispersions are the important factors to ensure that the ODFs containing the required amount of phenytoin are printed precisely. The rheological behaviors of 10, 12.5, 15 and 20 % *w*/*v* of HPMC E5 and 10%, 12.5% HPMC E15 are shown in Figure 3. As observed in the figure, all the HPMC dispersions exhibited a non-Newtonian (shear-thinning) pseudoplastic behavior since their viscosity values decreased as well as shear stress increased when the shear rate increased from 0 to 100 s−1. Furthermore, the power-law model was fitted to the obtained results (shear stress–shear rate flow curves), and the power-law model parameters (flow behavior index (*n*) and consistency coefficient) are shown in Table 2. The results showed that *n* values were less than 1.0 (0.66 to 0.78) for all HPMC E5 and HPMC E15 dispersions indicating the shear-thinning pseudo-plasticity nature, and the *n* values decreased with increasing of polymer concentration, indicating more intense shear-thinning. For the consistency coefficient, the results exhibited that the consistency coefficient values of HPMC E5 were lower than those of HPMC E15 with the same concentration, indicating that HPMC E5 was less pseudoplastic and less viscous than HPMC E15. The viscosities of 10, 12.5, 15 and 20 % *w*/*v* of HPMC E5 were 0.80 ± 0.04, 1.70 ± 0.03, 3.02 ± 0.18 and 9.84 ± 0.08 Pas and the viscosities of 10 and 12.5 % *w*/*v* of HPMC E15 were 8.10 ± 0.51 and 16.85 ± 0.76 Pas, respectively. As expected, the viscosities were found to be increased with increasing concentration of each HPMC. Moreover, a higher viscosity value was also observed when phenytoin was incorporated into the formulations. In the case of E5 20% and E15 10% formulations, after phenytoin was incorporated, the viscosities were found to be 13.39 ± 0.74 and 12.17 ± 0.47 Pas, respectively. Additionally, the results indicated that, for lower molecular weight polymer (HPMC E5), the higher polymer mass was required to achieve the same viscosity as that of the high molecular weight polymer (HPMC E15).

**Table 2.** Power-law model parameters and shape fidelity factor of different concentrations of HPMC E5 and HPMC E15.


Note: NA (not applicable) means the printing dispersions cannot be extruded through the nozzle.

According to Table 2, it was found that the diameters of the printed filaments were decreased when the concentration of HPMC E5 and HPMC E15 increased. The increase in the viscosity values could lead to the resolution improvement of printed ODFs in which the ODFs could be printed more accurately and more precisely. Shape fidelity factors of all printed formulations were close to one, suggesting that a stable shape of gel filament was deposited in the desired place during extrusion. Nonetheless, the results found that the addition of the drug could affect the viscosities and printability of the printing dispersion. The E15(12.5%)-PT formulation could not be extruded through the extruder nozzle smoothly due to its high viscosity and fast drying of the printing dispersion at the tip of the nozzle, ultimately causing a blockage of the extruder nozzle, whereas the E5(20%)-PT and E15(10%)-PT formulations were well extruded through the nozzle. Their diameters of printed filaments were found to be mostly close to the actual extruder nozzle diameters (0.51 ± 0.02 and 0.52 ± 0.01, respectively), thereby making them more suitable for this customized syringe extrusion 3D printer.

Thus, the results indicated that 20 and 10 % *w*/*v* were the suitable final polymer concentrations in printing dispersions for HPMC E5 and HPMC E15, respectively. Then, the E5(20%)-PT and E15(10%)-PT formulations, which exhibited similar rheological behavior and viscosity (approximately 12–13 Pas) and enabled extrusion through nozzles, were subsequently selected for plasticizers-loading and evaluated for their mechanical properties and disintegration time.

### *3.2. Film Preparation and Characterization*

The control E5 and E15 ODFs were slightly yellowish to colorless and transparent, whereas the phenytoin-loaded ODFs (E5-PT, E5-PT-PG, E5-PT-G, E15-PT, E15-PT-PG and E15-PT-G ODFs) were white in color and opaque due to phenytoin content. It was also observed that E5-PT-G, E15-PT, E15-PT-PG and E15-PT-G ODFs were smooth, uniform and flexible while the E5-PT and E5-PT-PG were more brittle and easily cracked during the drying process due to lack of plasticizers or an inadequate amount of plasticizer. Moreover, it was observed that the printed surface of E5-PT-G, E15-PT, E15-PT-PG and E15-PT-G ODFs showed the rectilinear pattern printed diagonally without the interruption and did not exhibit any noticeable difference upon visual inspection.

Additionally, the thickness and weight variation of all printed films were carried out in order to ensure the consistency of the printing process and printed ODFs. The average thickness of all printed ODFs varied in the range of 0.013–0.061 mm while the average thickness of phenytoin-loaded ODFs varied in the range of 0.056–0.061 mm, which falls within a typical range for fast dissolving oral films (0.05–0.15 mm) and would be adequate for handling to avoid any discomfort when applied on the buccal mucosa [19]. Therefore, as can be observed in Table 3, the average thickness and weight of printed ODFs was increased when the drug and/or plasticizers were incorporated into the film formulations; however, there were no significant differences (*p* > 0.05) observed between the thickness and weight of phenytoin-loaded E5 ODFs and phenytoin-loaded E15 ODFs.


**Table 3.** Weight and thickness of 3D-printed ODFs.

Note: 20% *w*/*v* of HPMC E5 and 10% *w*/*v* of HPMC E15 were used. For each test, means with the same letter are not significantly different. Thus, means with different letters, e.g., 'a' or 'b' are statistically different (*p* < 0.05).

### *3.3. Mechanical Properties of 3D-Printed ODFs*

The mechanical properties of all printed ODFs are displayed in Table 4, whereas the mechanical properties of E5-PT and E5-PT-G ODFs were not performed since they cracked during the drying process. The mechanical properties of the ODFs were taken to ensure that the ODFs possess suitable mechanical properties for easy handling and ease of removal from the packaging during use. In this study, the puncture strength, elongation to break and Young's modulus were used as parameters to characterize the mechanical properties of printed ODFs. The puncture strength is a measure of the toughness of film or the required force to apply on the film surface and puncture the film until breaking point [20]. The elongation to break is a parameter that reflects the ductility and stretchability of film (the ability of a film to be stretched without being torn). The Young's modulus is a parameter that reflects the elasticity and flexibility of film [21].

**Table 4.** Mechanical parameters and disintegration time of 3D-printed ODFs.


Note: ND (not determined) means films cannot be performed for E5-PT and E5-PT-PG ODFs, and the disintegration time of control E5 andE15 ODFs were not evaluated. For each test, means with the same letter are not significantly different. Thus, means with different letters, e.g., 'a' or 'b' are statistically different (*p* < 0.05).

Mechanically, the films in which phenytoin was incorporated (E5-PT-G, E15-PT, E15-PT-PG and E15-PT-G ODFs) were found to have significantly lower puncture strength, elongation to break and Young's modulus than the control E5 and E15 ODFs (*p* < 0.05), indicating that the E5-PT-G, E15-PT, E15-PT-PG and E15-PT-G ODFs were becoming less ductile and more flexible. This may be due to the disruption of the polymeric matrix continuity when drug particles were incorporated. This result was consistent with the previous studies showing that the addition of the drug to the formulations significantly influenced the mechanical properties of the film by decreasing the tensile strength and increasing the film's brittleness [22–24].

Furthermore, this study also showed that, without any plasticizers, the phenytoin-loaded E5 films (E5-PT) could not be formed, whereas the E5 films with a plasticizer (E5-PT-G) showed no significant difference in the puncture strength, elongations to break and Young's modulus value (*p* > 0.05) when compared with the phenytoin-loaded E15 films (E15-PT, E15-PT-G and E15-PT-PG). Additionally, it was observed that the E15-PT-PG and E15-PT-G ODFs showed slightly lower puncture strength, lower Young's modulus and higher elongation to break than the films without any plasticizers (E15-PT ODFs), but the difference was not significant (*p* > 0.05). Thus, the results indicated that the addition of glycerin at a concentration of 10% (*w*/*w* of polymer) as a plasticizer in the ODFs formulation had the effect on improving the mechanical properties of both E5 and E15 formulations, whereas it may require a higher concentration of propylene glycol for E5 formulation in order to avoid cracking of the films, to induce sufficient flexibility and to obtain the similar mechanical properties as E5-PT-G ODFs. Due to the hydrogen-bonding capabilities of plasticizers (e.g., glycerin and propylene glycol), the addition of suitable amounts of plasticizers can improve the flexibility of films by allowing the interaction of hydrogen bonding between polymer chain functional groups and the plasticizing agent, resulting in the reduction of polymer–polymer interaction [25].

#### *3.4. Disintegration Time*

The results of in vitro disintegration time of E5-PT-G, E15-PT, E15-PT-PG and E15-PT-G ODFs are presented in Table 4. For the film disintegration tests, there is still no official guidance available for determining the disintegration time of ODFs in the pharmacopeia. Nevertheless, according to the United States Pharmacopeia (USP) and CDER guidance regarding the orally disintegrating tablets (ODTs), disintegration time (which may be applied to ODFs), a time limit of 30 s or less for disintegration is specified [26]. In this study, the results showed that E5-PT-G, E15-PT, E15-PT-PG and E15-PT-G ODFs disintegrated within 6 s, which passes the time limit for disintegration. Moreover, all of the phenytoin-loaded E15 ODFs disintegrated significantly faster than the E5-PT-G ODFs (*p* < 0.05). This finding was consistent with the results of previous studies reporting that the formulation with the higher concentration of HPMC had a significantly longer disintegration time than that of formulations containing less concentration of HPMC [27].

According to the mechanical properties and disintegration time results, the optimal formulations, E15-PT, E15-PT-PG and E15-PT-G, which exhibited good mechanical properties and faster disintegration time, were selected for further experiments.

### *3.5. Morphological Characteristics of 3D-Printed ODFs*

Scanning electron microscopy images of E15, E15-PT, E15-PT-PG and E15-PT-G ODFs are shown in Figure 4. The E15-PT, E15-PT-PG and E15-PT-G ODFs demonstrated uniform and well-distributed extruded dispersions with a highly porous structure in the films. Generally, 3D printing of polymer inks with in situ evaporation of solvents has allowed fabrication of 3D porous structures with stringent requirements of rheological properties of the printing ink, e.g., high viscosity and high vapor pressure [28]. In our study, spontaneous solidification via in situ evaporation of mixed solvent of ethanol and distilled water (9:1, *v*/*v*) generated porosity at micro-to-nano scales. The addition of different plasticizers (PG and glycerin) did not affect the morphology of printed ODFs. In addition, there was no significant difference in the thickness observed from SEM images compared to the thickness measured by using a micrometer.

**Figure 4.** Scanning electron microscopy images of E15 (**a**), E15-PT (**b**), E15-PT-PG (**c**) and E15-PT-G (**d**).

### *3.6. Phenytoin-Loading Content*

In order to assess the uniformity of the printed ODFs and the robustness and precision of the film manufacturing process (3D printing process), phenytoin-loading content was evaluated. In this study, the theoretical phenytoin content in the printed ODFs size of 3.25 cm × 3.25 cm was 30 mg/film. Considering the theoretical content as 100%, the phenytoin-loading content was found to be 99.4 ± 5.2, 101.3 ± 3.8 and 99.7 ± 2.4% for E15-PT, E15-PT-PG and E15-PT-G ODFs, respectively. The results indicated that the phenytoin-loading content values of all phenytoin-loaded E15 ODFs were in agreement with the theoretical drug-loading (30 mg) and met acceptable criteria (95.0–105.0%) endorsed by the USP [18], confirming the uniformity of printed ODFs and the precision of 3D printing process of our customized syringe extrusion 3D printer.

### *3.7. In Vitro Phenytoin Release Studies*

In vitro phenytoin release experiments were carried out in Tris with 1% *w*/*v* SLS buffer solution pH 6.8 at 37 ± 0.5 ◦C. The dissolution profiles of selected 3D-printed ODFs (E15-PT, E15-PT-PG and E15-PT-G ODFs) are shown in Figure 5. All of the selected ODFs demonstrated a similar dissolution profile pattern showing the rapid phenytoin release up to 80% within 10 min, followed by a slow constant release rate to complete drug release (100% level) in 60 min. However, the dissolution rate of all selected films was not significantly different at all time points (*p* > 0.05). The rapid release behavior of phenytoin from all selected printed ODFs could be attributed to the highly porous structure of printed ODFs, fast disintegration time within 5 s, which creates more pores to allow the drug to diffuse out of the films and the hygroscopic nature of HPMC, which increases the water uptake capability of the films. According to the obtained results, this study indicated that the improvement of the dissolution rate of poorly water-soluble drugs, such as phenytoin, in the printed ODFs can be achieved. At present, there have been not many pharmaceutical products adopted 3D printing in manufacturing. In this study, we have successfully used 3D printing technology for fabricating phenytoin-loaded ODF. Although 3D printing has been used to prepare the ODFs to enhance the solubility of poorly water-soluble drugs, there has not been any research works reporting the preparation of ODFs containing phenytoin. According to the literature, Najafi et al. formulated phenytoin sodium mucoadhesive film using Carbopol 934, sodium carboxymethyl cellulose and HPMC as film formers by solvent casting method. They reported that the best formulation released 80% of the drug within 120 min [29]. However, phenytoin-loaded ODF developed by using the 3D printing technique in this study significantly enhanced drug disintegration and dissolution to less than 5 s and up to 80% in 10 min drug release, respectively, which was shown to be a novel pharmaceutical preparation possessing optimal mechanical strength and drug release profile compared with the previous reports.

**Figure 5.** In vitro phenytoin release from E15-PT, E15-PT-G and E15-PT-PG 3D-printed ODFs (*n* = 3) in Tris with 1% *w*/*v* SLS buffer solution (pH 6.8).

### **4. Conclusions**

In this study, the orodispersible films containing the required amount of phenytoin (30 mg) were successfully fabricated by 3D printing technology. The rheological properties and dimensional accuracy tests for different concentrations of HPMC E5 and HPMC E15 dispersions were performed to determine the optimal concentration and viscosity that were appropriate for the customized syringe extrusion 3D printer. The results showed that 20% *w*/*v* of HPMC E5 and 10% *w*/*v* of HPMC E15 were the most suitable polymers for incorporating drug and printing. Among all developed 3D-printed ODFs, the phenytoin-loaded E15 ODFs exhibited good physical appearance, good mechanical strength, rapid disintegration time (within 5 s) and a rapid release (up to 80% within 10 min), showing the improvement of solubility and dissolution rate of a poorly water-soluble drug, phenytoin. Overall, this study indicated that HPMC E15 is feasible to be applied as one of the 3D printing polymers for extrusion-based 3D printer and for the fabrication of orodispersible films. Additionally, this approach could be explored further for the stability of the film and the individualization of the ODFs for each patient by adjusting the volume of printing formulations.

**Author Contributions:** Conceptualization, P.P. and P.J.; methodology, P.P. and P.J.; software, S.U.; supervision, P.J.; validation, P.P. and P.J.; writing—original draft, P.P., S.U., P.R., C.C., W.R. and P.J.; writing—review and editing, P.P. and P.J. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study received financial support from the National Research Council of Thailand (NRCT) and partial funding from Chiang Mai University.

**Acknowledgments:** This research project is supported by the National Research Council of Thailand (NRCT): NRCT5-RGJ63007-079.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## **Methyl Methacrylate (MMA) Treatment of Empty Fruit Bunch (EFB) to Improve the Properties of Regenerated Cellulose Biocomposite Films**

### **Salmah Husseinsyah 1, Nur Liyana Izyan Zailuddin 1, Azlin Fazlina Osman 1,2,\*, Chew Li Li 1, Awad A. Alrashdi <sup>3</sup> and Abdulkader Alakrach <sup>4</sup>**


Received: 23 October 2020; Accepted: 4 November 2020; Published: 6 November 2020

**Abstract:** The empty fruit bunch (EFB) regenerated cellulose (RC) biocomposite films for packaging application were prepared using ionic liquid. The effects of EFB content and methyl methacrylate (MMA) treatment of the EFB on the mechanical and thermal properties of the RC biocomposite were studied. The tensile strength and modulus of elasticity of the MMA treated RC biocomposite film achieved a maximum value when 2 wt% EFB was used for the regeneration process. The treated EFB RC biocomposite films also possess higher crystallinity index. The morphology analysis indicated that the RC biocomposite film containing MMA treated EFB exhibits a smoother and more homogeneous surface compared to the one containing the untreated EFB. The substitution of the –OH group of the EFB cellulose with the ester group of the MMA resulted in greater dissolution of the EFB in the ionic liquid solvent, thus improving the interphase bonding between the filler and matrix phase of the EF RC biocomposite. Due to this factor, thermal stability of the EFB RC biocomposite also successfully improved.

**Keywords:** empty fruit bunch; regenerated cellulose; ionic liquid; methyl methacrylate

### **1. Introduction**

Biopolymers that derive from renewable resources have sparked great interest to the world community due to their benefits to the environment and sustainability. Their properties and characteristics demonstrated capability as substitutes for the common traditional petrochemical plastics for various applications [1–4]. Biopolymers are plenty and can be derived from various sources. Generally, biopolymers can be divided into poly (amino acids) and proteins, poly-, di- and monosaccharides such as chitin [5], starch, glucose, fructose and cellulose [6,7]. Cellulose which have the formula of (C6H10O5)n are recognized as renewable materials that can be found naturally abundant on Earth. These natural resources exist vastly in plants, animals and some microorganisms and can cover a range or variety of properties for instance from biodegradable and biocompatibility to being economical and low cost [5–8]. They are now considered as one of the most promising polymeric materials with their derived products applicable to numerous parts and sectors, principally in fibers, polymers, paints, papers and films industry [5,6]. Despite such outstanding characteristics

shown, there are also some limitations or disadvantages that are raised specifically in terms of cellulose processing. In order to produce cellulosic based film, the cellulose needs to be dissolved prior to the casting process. However, the dissolution process of cellulose is challenging since cellulose are known to be non-soluble in common solvents (water and conventional organic solvents). The insolubility of cellulose is due to the presence or appearance of intra- and intermolecular hydrogen bonding and Van der Waals interaction which closely pack the chains together, and also the partially crystalline structure of cellulose [8,9]. These restraints have motivated further discoveries and investigations for suitable cellulose dissolution methods to facilitate or aid in the use of cellulose, for example, in applications of regenerated cellulose [8,10,11]. Regenerated cellulose (RC) can be referred to as the chemical dissolution of insoluble natural cellulose followed by the recovery of the material from the solution [10,11]. Regenerated celluloses (RCs) are produced or formed commercially using various methods, for example, cellulose carbamate process, lyocell process and viscose process. These different routes can result in fibers with a range of mechanical properties like fiber with low elongation but high tenacity and modulus, and fiber with great elongation but low or small modulus and tenacity [10,11]. Ionic liquids (ILs) are recognized as purely ionic, salt-like materials that are liquid at unusually low temperatures or in simple context they are ionic compounds which are liquid below 100 ◦C [12,13]. ILs demonstrate exceptional performance or behavior for a variety of applications in the chemical industry, for instance catalysts, chemical synthesis, separation and preparation of materials [14]. Xia et el. have recently reviewed recent advances in processing and volarization of lignocellulosic materials in ILs, including the use of ILs as pre-treatment solvents for lignocellulose to improve and enhance accessibility for lignocellulose-based biocomposite production. They summarized that current research trends have shown that the ILs have a huge potential in multifarious, efficient and environmentally friendly utilization of lignin, lignocellulose and cellulose in green technology advancement. By utilizing ILs as the solvent, additive and dispersant, the production of high performance materials generated from natural plants or biomass materials can be increased and applied in various industries and fields [15]. Other findings also agreed that ILs are considered as an adequate solvent to dissolve cellulose for the cellulose regeneration process and this approach be employed for advanced applications [10,12]. The recovery of ILs after cellulose regeneration could be achieved via methods like evaporation, reverse osmosis, ion exchange and salting out for reuse and recycling purposes [16]. Other solvents that can be used for dissolving cellulose to generate regenerated cellulose are N-methylmorpholine-N-oxide (NMMO) [17], N-dimethylacetamide (DMAC) / lithium chloride (LiCl) [10], 1-butyl-3-methylimidazolium chloride (BMIMCl) [18], 1-allyl-3-methylimidazolium chloride (AMIMCl) [19], 1-ethyl-3-methylimidazolium acetate (EMIMAc) [20] and 1-ethyl-3-methylimidazolium chloride (EMIMCl) [21]. The solvent 1-butyl-3-methylimidazolium chloride (BMIMCl) was first determined or established by Swatloski and co-workers to have good dissolving power for cellulose [22]. This finding generates a way towards a new class of cellulose solvent system, and accelerates or prompts the investigation into the potential of other ILs in dissolving cellulose.

Oil palm empty fruit bunches (OPEFBs) are known as one of the most abundant biomasses from palm oil plantations. It is considered as a lignocellulosic residue of palm oil plants, which is regarded as an important agricultural source in Southeast Asian countries. For instance, Malaysia has produced about 15.8 million tonnes of OPEFB annually due to the vigorous activity of oil palm cultivation [10,23]. Upon oil extraction from fresh fruit bunches (FFB), empty fruit bunch (EFB) is attained as biomass with advantageous properties such as being inexhaustible, renewable, biodegradable and environmental friendly [10]. EFB consists of cellulose content ranging from 42.7–65.0%, where as much as 60.6% of this cellulose can be turned into value added products [24]. Cellulose which is derived from EFB can be employed for several applications such as making bioplastic film, paper and polymer composite products. Chemical pre-treatments are usually implemented for cellulose extraction and also to enhance or improve the cellulose dissolution [10,25]. Cellulose pre-treatments are usually carried out on natural fibers to remove the lignin and partly eliminate the hemicelluloses. It also disrupts or breaks the crystalline structure of cellulose, allowing the cellulose to swell and dissolve in

suitable solvent [10,25]. Some examples of pre-treatment methods include alkaline treatment [10,25], acid hydrolysis [26], hot compressed water [27] and enzymes [28]. Since cellulosic fibers are innately hydrophilic, they can easily degrade and absorb moisture from the environment. Therefore, chemical modification is applied and utilized to increase hydrophobicity and moisture resistance. Examples of chemical modification methods are silane functionalization [29], acetylation [30] and polymer grafting [31]. However, in the production of regenerated cellulose biocomposite, chemical treatment can also facilitate the regeneration process by reducing the hydrogen bonds of the cellulose and improving its solubility in the ionic liquid [25].

To-date, the mechanical and thermal analyses data of the RC film derived from the EFB based cellulose and produced through the use of 1-butyl-3-methylimidazolium chloride (BMIMCl) as ionic liquid and methyl methacrylate (MMA) as chemical treatment not available in the literature, but is truly important to contribute to the progress of the RC biocomposite films intended for use in biodegradable packaging applications. This is because the combination of both chemicals is practical and efficient to produce RC biocomposite with improved performance. BMIMCl has excellent cellulose-dissolving ability while MMA is an industrially relevant chemical with good capability to treat cellulose and improve its dissolution process during the regeneration process. The MMA was previously used to treat chitosan and *Nypa frutican* based-cellulose for the production of biocomposite [32,33]. However, this is the first attempt to utilize MMA in the chemical treatment of the EFB. In order to produce the EFB RC biocomposite film with improved mechanical and thermal properties, the role of MMA as the EFB's modifier must be clarified while the best EFB loading needs to be determined. This research was aimed at conducting an investigation into all of these aspects, thus revealing the never reported findings for future development of biodegradable plastics.

### **2. Experimental**

### *2.1. Materials*

Empty fruit bunch (EFB) fibers produced as biomass from palm oil trees were collected and received from the Malaysian Palm Oil Board (MPOB) in Bangi, Selangor, Malaysia. The EFB fibers were ground using a ring mill into powder form. The average particle size of EFB was 45 microns which was determined using Malvern (Malvern, United Kingdom) particle size analyzer. Microcrystalline cellulose (MCC) was purchased from Aldrich (St. Louise, MO, USA). Sulfuric acid, ethanol and sodium hydroxide (NaOH) were acquired from HmbG Chemicals (Hamburg, Germany). Acetic acid was purchased from BASF Chemical Company (Ludwigshafen, Germany) while sodium chlorite (NaClO2) was supplied by Sigma Aldrich (St. Louise, MO, USA). Furthermore, 1-butyl-3-methyl-imidazolium chloride and methyl methacrylate (MMA) were purchased from Merck (Darmstadt, Germany).

### *2.2. Pre-Treatment of EFB*

Firstly, EFB powder was treated using 4 wt% of NaOH solution for 1 h at 70 ◦C. The treatment was performed under mechanical stirring at the speed of 750 rpm. This procedure was repeated 3 times, and the EFB was washed with distilled water and filtered several times after each treatment to remove the soluble alkaline. Then, the alkaline treated EFB was subjected to a bleaching process for further delignification. The bleaching process was carried out in solution containing acetate buffer (solution of 2.7 g NaOH and 7.5 mL glacial acid in 100 mL distilled water), 1.7% w/v aqueous sodium chlorite and distilled water. The EFB was mixed together with the bleaching solution which was stirred for 60 min at the temperature of 70 ◦C. The bleached fibers were then washed with distilled water and filtered. Acid hydrolysis was carried out using 65 wt% of H2SO4 at the temperature of 50 ◦C under mechanical stirring for 45 min. The EFB attained was washed with cold distilled water to stop the acid hydrolysis reaction. The EFB was washed and filtered repeatedly until pH 7 was reached. The filtered EFB was dried in an oven for a day at 70 ◦C.

### *2.3. EFB Modification*

Chemical treatment was carried out on pre-treated EFB using methyl methacrylate (MMA). Firstly, the MMA solution was prepared by dissolving 3% of MMA in ethanol (v/v). Then, the EFB was slowly added into the MMA solution and the mixture was stirred continuously for 1 h. The solution was left overnight. The next day, the EFB was filtered and dried in an oven at 70 ◦C for 24 h.

### *2.4. Preparation of EFB Regenerated Cellulose (RC) Biocomposite Films*

The concept of this RC biocomposite is the same as self-reinforced composite (SRC). Both EFB cellulose and MCC were used as the reinforcement and matrix when they were partially dissolved in the 1-butyl-3-methylimidazolium chloride (BMIMCl) solvent system. The dissolved cellulose was transformed into the matrix phase surrounding the remaining non-dissolved cellulose. EFB RC biocomposite films with 1, 2, 3 and 4 wt% EFB (untreated and MMA treated), plus a constant amount of microcrystalline cellulose (MCC) (3 wt%) were prepared for this study. BMIMCl was used as solvent in the regeneration process of the EFB cellulose. Firstly, the BMIMCl was heated in order to dissolve it into liquid form. Furthermore, 3 wt% of MCC and 1 wt% of EFB was dispersed in BMIMCl. The mixture was heated at the temperature of 100 ◦C and stirred for 30 min until a homogeneous cellulose solution was achieved. The cellulose solution was cast onto a glass plate mold. As the film sample developed, it was soaked in a water bath to allow the diffusion of ionic liquid out of the sample. Water was changed a few times to ensure complete removal of ionic liquid from the film sample. The film attained was dried in an oven for 24 h at 50 ◦C. Similar steps were repeated to produce the EFB RC biocomposite films with different contents of untreated and treated EFB. Note that the purpose of adding MCC into the mixture was to provide a strengthening effect to the RC biocomposite films. Preliminary work was done in which the RC films were produced without the addition of MCC. Unfortunately, the films produced were too brittle. In addition, we have tried to prepare the film with 0% EFB; however, no good film has been successfully produced. These factors were the reason for preparing the RC biocomposite films with both MCC and EFB. As MCC was added in constant amounts (3 wt%) in all the RC biocomposite films, the changes in the morphology and properties of the films reported in the discussion part were assumed due to the EFB contents.

The overall process to produce the EFB RC biocomposite films is illustrated in the schematic diagram in Figure 1.

**Figure 1.** Procedures to prepare EFB RC biocomposite films.

### **3. Texting and Characterization**

### *3.1. Tensile Testing*

Instron Universal Testing Machine (Norwood, MA, USA), model 5590 was used to carry out the tensile test which was based on ASTM D 882. The tensile test was operated at the speed of 10 mm/min. The EFB RC biocomposite films were cut into rectangular shapes with dimensions of 15 mm × 100 mm. The thickness of the sample was measured using a digital vernier caliper along its gauge length area three times (at center, upper and bottom sides). The mean value was taken and recorded. Five replicates of each material were used for the tensile test, and the average values of tensile strength, modulus of elasticity and elongation at break were attained and recorded. Statistical analysis was performed using two-tailed Student's t-test for unpaired data to compare among materials. A significance level of 0.05 or less was accepted as statistically significant.

### *3.2. Scanning Electron Microscopy (SEM)*

Scanning electron microscope (SEM) model JEOL JSM-6460LA (Akishima, Japan) was utilized to analyze and study the morphology of EFB RC biocomposite films. The fractured surface of the biocomposite film was subjected to morphological analysis to study the deformation behavior of the film samples. The films were coated with a thin layer of palladium for conductive purpose. Scanning was done at a voltage of 10 kV.

### *3.3. X-Ray Di*ff*raction (XRD)*

XRD analysis was conducted using a Bruker DS X-ray diffractometer (Billerica, MA, USA) utilizing X-ray sources of Cu-Kα radiation (λ = 1.5418 Å). The analysis was carried out under normal atmospheric conditions and at ambient temperature. Samples used for XRD testing were strips of films with dimensions of 10 mm × 15 mm [32]. The crystallinity index (CrI) for both untreated and treated EFB RC biocomposite films were calculated based on [10,34]:

$$\text{CrystalTriity Index (CrI)} = \left(\text{I/I'}\right) / \left(\text{I}\right) \times 100\tag{1}$$

where I = the height of peak assigned to (200) planes, measured in the range 2θ = 20–23◦, and I' = the height of peak assigned to (100) planes, located at 2θ = 12–16◦.

### *3.4. Fourier Transform Infrared (FTIR)*

A Perkin Elmer L1280044 FTIR spectrometer (Waltham, MA, USA) was used to study the functional groups of untreated and treated EFB RC biocomposite films. Four scans were conducted on the films in the wavenumber range from 4000–650 cm–1 with resolution of 4 cm–1 using attenuated total reflectance (ATR) method.

### *3.5. Thermogravimetric Analysis (TGA)*

TGA Pyris Diamond Perkin Elmer (Waltham, MA, USA) was used to analyze the thermal behavior of both untreated and treated EFB RC biocomposite films. The film samples used were weighed to about 5 ± 2 mg. The films were placed in aluminum pans and heated from 25 to 700 ◦C with a heating rate of 10 ◦C/min. The testing was performed in nitrogen atmosphere with nitrogen flow rate of 50 mL/min.

#### **4. Results and Discussion**

### *4.1. Tensile Properties*

Figure 2 shows the effect of EFB content on the tensile strength of the untreated and MMA treated EFB-based RC biocomposite films (EFB RC biocomposite). The tensile strength for both untreated and treated EFB RC biocomposite films demonstrated similar trends in which the value increased initially when the EFB content increased from 1 to 2 wt%. Further increase in the EFB content resulted in the decrease of the tensile strength of the EFB RC. Large accumulations of EFB particles led to non-homogeneous dispersion of the EFB during the formation of the EFB RC biocomposite film. This would develop more stress concentrated areas that hasten the failure of the biocomposite's tensile deformation. The highest tensile strength value (34 MPa) was obtained when 2 wt% MMA treated EFB was employed to form the RC biocomposite. Obviously, at similar EFB content, the treated EFB RC biocomposite films obtained higher tensile strength compared to the untreated biocomposite films. For instance, when 2 wt% MMA treated EFB was used to form the RC biocomposite; a 26% improvement in tensile strength was obtained vs. the untreated RC biocomposite film. It is proven that the treatment of EFB with MMA has enhanced and improved the interfacial adhesion between the non-dissolved (filler phase) and dissolved (matrix phase) cellulose, allowing greater stress transferring mechanisms throughout the biocomposite structure.

**Figure 2.** Effect of EFB content on tensile strength of untreated and MMA-treated EFB RC biocomposite films.

Figure 3 displays the effect of EFB content on the elongation at break of the untreated and MMA treated EFB RC biocomposite films. Generally, the elongation at break values of the biocomposites does not show any clear trend. The fluctuating of the elongation at break value could be related to environmental influences such as moisture, absorption rate of the film and different levels of void content due to random dispersion of the EFB particles in the biocomposite structure. These factors affect the flexibility and conformation of the biopolymer chains when subjected to tensile deformation. In addition to that, the use of single material (EFB cellulose which functioned as both matrix and reinforcement) to form the biocomposite can cause an uncontrollable portion of the dissolved cellulose (matrix) and undissolved cellulose (filler) in the RC biocomposite structure. In this case, the "real composition" of filler phase may not be proportional with the EFB content, thus the portion that contributes to elasticity may also vary. These reasons might explain the unexpected elongation at break results of these biocomposite films. However, regardless of the EFB content, the MMA treated films exhibit lower elongation at break compared to the untreated biocomposite films. Enhancement in the interfacial bonding between matrix and filler obtained through the chemical treatment has restrained the flexibility and mobility of the polymer chain.

**Figure 3.** Effect of EFB content on elongation at break of untreated and MMA-treated EFB RC biocomposite films.

The effect of EFB content on the modulus of elasticity of the untreated and MMA treated EFB RC biocomposite films is illustrated in Figure 4. The modulus of elasticity for the untreated and treated EFB RC biocomposite films increased as the EFB content increased from 1 to 2 wt%. Addition of more EFB content (3 and 4 wt%) resulted in the declination of modulus of elasticity for both types of film. At 1, 2 and 3 wt% EFB content, the treated RC films demonstrated higher modulus of elasticity than the untreated biocomposite films. As mentioned earlier, MMA treatment of EFB has improved the matrix-filler interactions, thereby increasing the stiffness of the biocomposite films. The higher stiffness required higher stress to prompt movement, thereby resulting in the increase in the modulus of elasticity of the biocomposite films. However, when 4 wt% EFB was employed, the modulus of elasticity value of the RC biocomposite film with the untreated EFB statistically shows no significant difference to the RC biocomposite film with the MMA-treated EFB. Even though the EFB has been treated with the MMA, the high content and overcrowding of the cellulose might hinder the aligning of the cellulosic molecular chains in an orderly manner during the regeneration process, hence more flexible chains are formed. This results in no significant increase in the modulus of elasticity of the treated EFB RC biocomposite when compared with the untreated one.

**Figure 4.** Effect of EFB content on modulus of elasticity of untreated and MMA-treated EFB RC biocomposite films.

### *4.2. X-Ray Di*ff*raction (XRD)*

Figure 5 shows the XRD pattern of the untreated and treated RC biocomposite films with MMA at 2 and 4 wt% of EFB content, while Table 1 summarizes their crystallinity index (CrI). The diffraction peaks of all the samples were observed in the 2θ range of 11.0–13.0◦ and 19.0–21.0◦. These peaks suggest the cellulose II structure, as stated by Zhang et al. [35]. The intensity of the peaks of the treated films is higher compared to the untreated biocomposite films, thus the CrI is also higher. This was due to the treatment of EFB with MMA which has improved and enhanced the dissolution and regeneration of cellulose, thus a higher degree of cellulose II was formed.

**Figure 5.** XRD pattern of untreated and MMA-treated RC biocomposite films containing 2 and 4 wt% EFB.


**Table 1.** The crystallinity index (CrI) of untreated and MMA-treated EFB RC biocomposite films containing 2 and 4 wt% EFB.

### *4.3. Morphology Study*

Figure 6a, and Figure 6b–e represent the SEM images of the EFB powder and fractured surface of the RC biocomposite films at 2 and 4 wt% of EFB content, respectively. The SEM image of the EFB powder indicates that it contains particles with irregular shape. Upon regeneration process, the particulate form of the EFB disappeared as observed through the images of the biocomposite films (Figure 6b–e). A smoother and more homogeneous surface can be observed in the EFB RC biocomposite films that contain 2 wt% EFB as compared to the ones containing 4 wt% EFB (untreated and MMA treated). This shows that, lower content of EFB (2 wt%) resulted in better dissolution and dispersion of the EFB in the structure of the RC biocomposite. The presence of higher EFB content (4 wt%) resulted in poorer dispersion of the non-dissolved EFB (filler phase) and lower rate of dissolution of the regenerated cellulose (matrix phase). Similar results were attained by Liu et al. [12], which concluded that the homogeneous surface was obtained in films regenerated from cellulose solution containing low cellulose content. Furthermore, it can also be witnessed that the chemical treatment of the EFB using MMA resulted in a more homogeneous and smoother surface morphology of the RC biocomposite. The micrographs also indicate that fewer aggregates were observed. This implied that the reinforcement was well-surrounded by the regenerated cellulose matrix, due to improved interactions between both filler and matrix phases. The untreated RC biocomposite film with 4 wt% of EFB content (Figure 6c) illustrates a rough and heterogeneous morphology, which could be the cause of its tensile property reduction as mentioned previously.

**Figure 6.** *Cont*.

**Figure 6.** (**a**) SEM of EFB powder. Tensile fractured surface of (**b**) untreated RC biocomposite film containing 2 wt% EFB, (**c**) untreated RC biocomposite film containing 4 wt% EFB, (**d**) treated RC biocomposite film containing 2 wt% EFB and (**e**) treated RC biocomposite film containing 4 wt% EFB as observed through SEM.

### *4.4. Thermal Gravimetric Analysis*

Thermal gravimetric analysis (TGA) curves of the untreated and MMA-treated EFB RC biocomposite film containing 2 wt% of EFB were illustrated in Figure 7, while Table 2 tabulated the TGA data for all those materials. At the early stages of the thermal decomposition process (which is below 100 ◦C), the weight loss was due to the volatilization and vaporization of water [34,35]. At temperature range of 95–191 ◦C, the weight loss for both types of biocomposite film remained almost unchanged. This could be due to the removal of the remaining hemicellulose and extractives upon the alkaline treatment [34]. At temperatures higher than 190 ◦C, the decomposition of the RC biocomposite films generally occurred in two stages and this trend was also observed by Yeng et al. [34] and Reddy et al. [35]. At the first stage, a sharp weight loss can be observed in the temperature range of 191–282 ◦C for the treated RC film and 182–285 ◦C for the untreated film. This is associated with the degradation process of the crystalline cellulose. The second degradation stage occurred at 289–427 ◦C for treated film and 288–425 ◦C for untreated film, indicating that the cellulose almost completely burnt due to the continuing thermal oxidative degradation [35,36]. Apparently, the treated biocomposite films exhibit higher maximum degradation temperature (Tdmax) as opposed to the untreated films. This is ascribed to the better compatibility between the matrix and filler phase. Most obvious at temperatures between 300 and 600 ◦C, the treated films demonstrate lower weight loss. This suggests that the MMA treatment of the EFB can improve the thermal stability of the EFB RC biocomposite

film. Other researchers also realized the benefit of chemical treatment of fiber in improving its thermal stability. For instance, Chen et al. [37] have analyzed and reported the thermal properties of the cellulose fibers from rice straw, and concluded that the treated cellulose fibers possess higher degradation temperatures and higher thermal stability than the untreated ones.

**Figure 7.** Thermal gravimetric analysis (TGA) curves of untreated and MMA-treated EFB RC biocomposite films containing 2 wt% EFB.


**Table 2.** TGA data of EFB RC biocomposite films containing 2 wt% EFB.

### *4.5. Fourier Transform Infrared Spectroscopy Analysis (FTIR)*

Figure 8a,b illustrates the FTIR spectra of the untreated and MMA-treated EFB RC biocomposite films. Both untreated and treated EFB RC biocomposite films demonstrate similar peaks, suggesting that both films have comparable chemical compositions. The broad peak at 3373 cm–1 relates to the appearance of –OH stretching vibration. The band which can be seen at 2961 and 2930 cm–1, is attributed to the C-H group's stretching vibration [10,38]. Vibration of CH2 groups is implied by the peak at 2857 cm–1. The peak at 1725 cm–1 corresponds to the carbonyl groups (C=O) appears due to the chemical treatment of the EFB using MMA. The reason for the appearance of this peak in the spectra of both untreated and treated EFB RC biocomposite films is that there might be some residual of the lignin which also exhibits these functional groups. The absorption band observed at 1463 cm–1 relates to the CH3 and CH2 deformation vibration, while the peak at 1270 cm–1 displays the presence of C-O bond stretching vibration [10,38]. The absorption band detected at 1188 cm–1 for the treated biocomposite films indicated the anti-symmetric C-O-C stretch from the ester [10]. The band which is seen at 1118 cm–1 suggests the C-O-C pyranose ring skeletal vibration. The appearance of the band at 1020 cm–1 in both untreated and treated films indicates the presence of C-O bond stretching vibration of C-O-C linkage in the anhydroglucose unit (AGU). The band at ~877 cm–1 is associated with β-glycosidic linkages between glucose in the EFB cellulose as a result of ring vibration and –OH

bending of the glycosidic C-H group. This type of response can only be seen in the regenerated cellulose. Figure 9 demonstrates the schematic reaction between EFB and MMA. Upon chemical treatment, the hydrogen bonds of the EFB cellulose reduced due to replacement of its –OH group with the ester group of the MMA. As a result, the EFB cellulose can be better dissolved in the solvent, leading to an improved regeneration process and greater interphase bonding between the filler phase and matrix phase of the EFB RC biocomposite. As observed through the SEM analysis, more homogeneous RC biocomposite was obtained that also led to the improved mechanical and thermal properties of the EFB RC biocomposite film.

**Figure 8.** FTIR spectra of (**a**) untreated and (**b**) treated RC EFB biocomposite films with MMA.

**Figure 9.** The schematic reaction between EFB and MMA.

### **5. Conclusions**

The effects of EFB content and chemical treatment of EFB using methyl methacrylate (MMA) on the mechanical and thermal properties of the EFB RC biocomposite films were studied. Results demonstrated that the treated EFB films exhibit higher and greater tensile strength and modulus of elasticity, but lower elongation at break compared to the untreated films. The treated EFB RC films also attained higher crystallinity index. Surface morphology of the treated biocomposite films indicates smoother surface due to the MMA treatment. Thermal analysis demonstrates that the treated biocomposite film possesses higher Tdmax and lower weight loss at 600 ◦C. This suggests that the treated biocomposite film has greater thermal stability. The presence of ester in the treated EFB RC biocomposite films with MMA was proven and affirmed by FTIR analysis. The substitution of the –OH unit of the EFB cellulose with the ester unit of the MMA has facilitated the dissolution process of the EFB during the regeneration process due to weakening of the hydrogen bonding of the cellulose. A more homogeneous biocomposite structure developed which led to the improvement of both tensile properties and thermal stability of the EFC RC biocomposite film.

**Author Contributions:** Conceptualization, S.H. and A.F.O.; methodology, C.L.L.; software, A.A.; validation, N.L.I.Z., A.F.O. and S.H.; formal analysis, A.F.O.; investigation, A.A.; resources, A.A.A.; data curation, A.A.A.; writing—original draft preparation, N.L.I.Z.; writing—review and editing, A.F.O.; visualization, C.L.L.; supervision, S.H.; project administration, A.F.O.; funding acquisition, A.A.A. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** The authors would like to thank the School of Materials Engineering for the facilities and Madam Marliza Mosthapa for her kind help in experimental works.

**Conflicts of Interest:** This manuscript, or its contents in some other form, has not been published previously by any of the authors and/or is not under consideration for publication in another journal at the time of submission. All authors declare that there is no conflict of interest regarding the publication of this paper.

### **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Biodegradable Starch**/**Chitosan Foam via Microwave Assisted Preparation: Morphology and Performance Properties**

### **Xian Zhang, Zhuangzhuang Teng and Runzhou Huang \***

Co-Innovation Center of Efficient Processing and Utilization of Forest Products, College of Materials Science and Engineering, Nanjing Forestry University, Nanjing 210037, China; shinezhang17@163.com (X.Z.); RowanZJ@163.com (Z.T.)

**\*** Correspondence: runzhouhuang@njfu.edu.cn

Received: 19 October 2020; Accepted: 2 November 2020; Published: 6 November 2020

**Abstract:** The effects of chitosan (CTS) as the reinforcing phase on the properties of potato starch (PS)-based foams were studied in this work. The formic acid solutions of CTS and PS were uniformly mixed in a particular ratio by blending and then placed in a mold made of polytetrafluoroethylene for microwave treatment to form starch foam. The results showed that the molecular weight and concentration of CTS could effectively improve the density and compressive properties of starch-based foams. Furthermore, orthogonal experiments were designed, and the results showed that when the molecular weight of CTS in foams is 4.4 <sup>×</sup> 105, the mass fraction is 4 wt%, and the mass ratio of CTS–PS is 3/4.2; the compressive strength of foams is the highest at approximately 1.077 mPa. Furthermore, Fourier transform infrared spectroscopy analysis demonstrated the interaction between starch and CTS, which confirmed that the compatibility between CTS and PS is excellent.

**Keywords:** chitosan; potato starch; microwave; foam; orthogonal experiments

### **1. Introduction**

Global pollution is increasingly becoming a serious concern, causing extreme climate changes and natural disasters. The huge pollution of traditional plastic products in the environment has extensively raised people's concern. Traditional plastics such as polystyrene (PS), soft (hard) polyurethane, polyvinyl chloride, and polyolefin are widely used as the matrix of foaming materials, but due to the poor operability and degradability of these plastics, most of the foaming materials are wasted, leading to increasingly serious white pollution. With people's increased awareness of environmental protection, considerable attention has been paid to the development of environmentally friendly materials. Therefore, bio-based and biodegradable polymers as replacement materials for traditional nonbiodegradable plastics have been extensively studied [1–4]. Biodegradable plastics in the natural environment are decomposed into carbon dioxide (CO2) and oxygen (O2) by the action of water (H2O), light, and microorganisms, and they hardly pollute the environment. Conventional plastics are nondegradable, which degrade slowly and require a long time, even hundreds of years, to degrade completely. They also impact the soil negatively during this time. Therefore, the exploration of degradable plastics that can replace traditional plastics holds great practical significance.

Starch, a rich polysaccharide substance found on earth, is an important raw material for the synthesis of biodegradable plastics. Starch has excellent biodegradability, and it can be decomposed by microorganisms in the natural environment and finally be metabolized into H2O and CO2. Starch-based plastics are beneficial in being nontoxic, harmless, widely available, and cheap and also possess relatively superior properties. Use of these plastics effectively can alleviate the problem of white pollution and the crisis of biochemical energy shortage [5,6]. Biodegradable porous materials prepared from

starch are being widely used in the fields of food packaging and medical tissue engineering [7]. To ensure complete degradability of the system, various biodegradable natural polymers such as cellulose, lignin, and chitosan (CTS) and polyesters such as polylactic acid (PLA), polycaprolactone, polybutylene succinate, and polyvinyl alcohol have been blended with starch in an attempt to synthesize biodegradable plastics [8–13]. Loercks et al. [14] successfully developed a degradable thermoplastic starch-based product by compounding starch with a hydrophobic biodegradable polymer, which greatly enhanced its aerobic degradation rate. Ferri et al. [15] reported that the addition of glycerin, polyethylene glycol, and other plasticizers to the blend of thermoplastic starch (TPS) and PLA could significantly reduce the cost, enhance mechanical performance, and broaden the application ranges of PLA-TPS blends. Bénézet et al. [16] prepared fiber-reinforced foams by the extrusion method and reported that the addition of fibers increases the expansion index and significantly reduces the water absorption of starch-based foams; the foams with 10% hemp fibers presented excellent mechanical properties.

CTS consists of β-(1,4)-linked 2-amino-deoxy-d-glucopyranose, and it can be synthesized by deacetylation of chitin [17,18]. It is soluble in acidic solutions because of the protonation of its –NH2 group at the C–2 position of the glucosamine unit. Owing to its biodegradability and unique physicochemical properties, it is widely used in the preparation of hydrogels, films, fibers, or sponges and in the field of biomedicine [19]. Although CTS is a novel substance that possesses excellent film-forming potential, permeability, and antibacterial properties, it has many deficiencies. Dang et al. [20] prepared TPS or CTS films by using different CTS concentrations and confirmed that CTS is distributed on the surface of films. Due to the high crystallinity and hydrophobicity of CTS and its ability to interact with starch-forming intermolecular hydrogen bonds, CTS on the membrane surface improves the water vapor and oxygen barrier properties of materials and decreases the hydrophilicity of the membrane surface.

Starch and chitosan are natural high molecular polymers with excellent biocompatibility and biodegradability, which can be completely degraded without causing harm to the environment [21]. Currently, most research has focused on the synthesis of the starch–CTS membrane [22–25], and the application of CTS-starch foams in the field has been rare. The foam prepared by starch and chitosan is a kind of biological material, which is easy to be processed and molded. The chitosan in foam can promote the formation of blood coagulation and thrombosis, inhibit the growth of various bacteria and fungi, and promote the repair of damaged tissues [26]. Therefore, the foam is expected to be a good hemostatic material and packaging material for food and medicine. In this paper, we aimed to manufacture a new foam material by using CTS and potato starch (PS) in a polytetrafluoroethylene mold for microwave treatment, analyze the effects of molecular weight and concentration of CTS on the morphological characteristics and functional properties, such as physicochemical, mechanical, and thermal properties, of starch-based foam, and determine the optimum proportion of CTS and PS in foams to achieve excellent performance. We hope to enable further development of starch-based foams for use as active materials in biomedical applications as well as food and drug packaging.

#### **2. Materials and Methods**

#### *2.1. Materials*

PS (with approximately 75% amylopectin and 25% amylose) and CTS (low viscosity ≤200 mPa·s; medium viscosity: 200–400 mPa·s; high viscosity ≥400 mPa·s) were purchased from Shanghai Meryer Company (Shanghai, China). Formic acid (96%) was purchased from Shanghai Macklin Company (Shanghai, China).

### *2.2. Sample Preparation*

(1) Determination of the molar mass of CTS

Acid hydrolysis was used to determine the molecular weight of three CTS samples used in this study. The dried CTS powder was dissolved in 0.2 M acetic acid (HAc)/0.1 M sodium acetate (NaAc) solvent (pH = 4) and different concentrations of the solution were prepared; CTS powder was dissolved completely in the solvent by moderately heating the solution at 40 ◦C. The 10 mL solvent and cooled CTS solution were injected into the Ubbelohde viscometer (capillary diameter = 0.8 mm) (Shanghai Liangjing Glass Instrument Factory, Shanghai, China), and the vertical state Ubbelohde viscometer was placed in a constant temperature water bath at 30 ◦C for 15 min. The outflow time of solvent and CTS solution (*t*, *t*0) was measured with a stopwatch, and the mean value of the three measurements was recorded (the difference was not more than 0.2 s).

(2) Preparation of CTS-starch based foams

The starch-based foaming materials were formulated using CTS and PS by heating in a microwave oven. AWD900ASL23-2 Galanz microwave oven (Galanz Shanghai Company Limited, Shanghai, China) with microwave emission frequency of 2450 MHz was used to heat CTS-starch compounds under microwaves at 100% power of 900 watts. The container carrying the CTS-starch mixtures was located at the center of the microwave chamber and irradiated by microwaves for the same time (40 s). Different CTS-starch mixtures were packed in special containers and the specific experimental process is shown in Figure 1. After microwave treatment, the samples were conditioned overnight at 25 ◦C before further characterization.

**Figure 1.** Schematic diagram of preparation of CTS-starch foams.

#### (3) Experiment design

The comprehensive experiment was designed with three factors at three levels respectively (33 = 27 combinations, without considering the reproducibility of experiments), namely the viscosity of CTS (A), the mass fraction of CTS (B), and the weight ratio of CTS solution/starch (C). Table 1 shows the experimental design under different factors and levels.


**Table 1.** Formula of CTS-starch foams.

On the basis of the comprehensive experiment, orthogonal test analysis and multi-factor analysis of variance were carried out to explore the interaction factors' influence on the mechanical properties, which would further help the experiment to determine the best process parameter of each level and to determine the relative importance of individual parameters and the combination of process parameters with high performance.

### *2.3. Characterization*

### (1) Calculation of CTS molecular weight

When the viscosity of polymer in a wide range of molecular weight is measured with the same viscometer, the relative value of solution viscosity to solvent viscosity or the relative viscosity, η*r*, is expressed as *t*/*t*<sup>0</sup> (*t*<sup>0</sup> and *t* are the outflow time of dilute solution and pure solvent, respectively, under the same temperature and measured using the same viscometer,). The specific viscosity, η*sp*, which reflects the internal friction effect between polymer and polymer and between pure solvent and polymer, is expressed as follows:

$$
\eta\_{\rm sp} = (\eta - \eta\_0)\eta\_0 = (\eta\_0 \phi \eta\_0) - 1 = \eta\_r - 1 \tag{1}
$$

The intrinsic viscosity [η] is defined as the reduced specific viscosity (η*sp*/*c*) or the relative viscosity of the inherent relative viscosity (*ln* η*r*/*c*) of a solution in the case of infinite dilution of solution mass concentration (*c*), which is expressed as follows:

$$\lceil \eta \rceil = \lim\_{\mathbf{c} \sim 0} \eta\_{\mathbf{sp}} / \mathbf{c} = \lim\_{\mathbf{c} \sim 0} \ln \eta\_{\mathbf{r}} / \mathbf{c} \tag{2}$$

The intrinsic viscosity [η] of polymer solution exhibits a specific relation with the molecular weight of polymer at a specific temperature. According to the Mark–Houwink–Sakurada (MHS) equation, the following empirical formula is used for calculating [η] and the molecular weight of polymer solution [27–29]:

$$
\mathbb{L}[\eta] = \mathbb{K}\mathbb{M}^{\mathfrak{a}},\tag{3}
$$

where *M* is the molecular weight, *K* and α are constants for given solute–solvent system and temperature, respectively. In this study, *<sup>K</sup>* and <sup>α</sup> were 6.59 <sup>×</sup> <sup>10</sup>−<sup>3</sup> and 0.88, respectively.

The relationship between *ln* η*r*/*c* (or η*sp*/*c*) and mass concentration *c* is linear. Thus, a straight line can be drawn by plotting *ln* η*r*/*c* (or η*sp*/*c*). When *c* is extrapolated from Equation (2), the line should cross a point on the y-axis; the intercept is the intrinsic viscosity [η]; the molecular mass of polymer can be calculated using Equation (3).

(2) Morphology of starch-based foams

The starch-based composite structure was investigated using scanning electron microscopy (SEM) (FEI 200 Quanta FEG, pressure 0.9 Torr) (FEI Company, Hillsboro, OR, USA). Foam samples were sputter-coated with platinum for 20 min after cutting with a razorblade into 3-mm thick slices, and the transversal cross-section area was examined using SEM.

(3) Pore size, porosity, and density of CTS–PS foams

Image analysis (Image-J) software (National Institutes of Health, Bethesda, MD, USA) was used to measure the average diameters of nearly 100 pores. Plot of pore size distribution enabled the comparison of cellular structure for different formulations of starch foams. The finished foam was cut into five samples of 24 <sup>×</sup> 24 <sup>×</sup> 32 mm3 size, and then the exact size of the sample was measured to ensure that the error was less than 0.02 mm. The sample mass was weighed with an electronic balance to make the result accurate to 0.001 g, and finally the density of the material was calculated using the following formula:

$$
\rho = \frac{M}{V} \times 10^3 \tag{4}
$$

where ρ = density (g/cm3), *M* = foam weight (g), *V* = foam volume (cm3)

(4) Mechanical property

According to the standard GB/T 8813-2008, the compressive strength of foam materials was tested using a CMT 4000 universal testing machine (Mechanical Testing & Simulation Company, Eden Prairie, MN, USA). The vertical direction of foam growth was consistent with the direction of compression. The compressive strength of the foam material is the ratio of the maximum compressive stress to the original cross-sectional area of the sample, when the relative deformation of the material is less than 10%, which is expressed in kPa.

The specific formula is as follows:

$$P = F/S = F \times 10^4 / (l \times w) \tag{5}$$

where *P* = compressive strength (kPa), *F* = compressive stress (kN), *S* = cross-sectional area (cm3), *l* = length (mm), *w* = width (mm).

(5) Fourier transform infrared analysis

The dried samples and potassium bromide were ground into a fine powder in an agate mortar and pressed into tablets. The Fourier transform infrared (FTIR) spectrum of material was measured using the VERTEX 80 V type FTIR spectrometer (German Bruker company, Karlsruhe, Germany) and the spectrum at a range of 4000–400 cm−<sup>1</sup> was acquired for every sample. The processing software OMNIC 8.0 (Thermo Nicolet Corporation, Fitchburg, WI, USA) was used to perform automatic baseline correction of the collected infrared spectrum to ensure its baseline level.

(6) Thermal characterization

The total mass loss on a TG209F3 analyzer (Netzsch Corporation, Bavarian, Germany) was assessed using the thermogravimetric (TG) method. For the 6 mg samples dried for 24 h at 70 ◦C, the experimental temperature was set as 30 ± 3 ◦C to 800 ◦C in nitrogen atmosphere. The heating rate was 10 ◦C/min, and the experimental flow rate was 30 mL/min.

(7) Differential scanning calorimetry (DSC) test

The glass transition temperature of sample was evaluated by a 200 F3 Differential Scanning Calorimeter (Netzsch Corporation, Bavarian, Germany). The dried sample of around 3 mg was placed in an aluminum crucible and sealed. An empty aluminum box was used as a control and nitrogen was used as purge gas. The temperature ranged from 20 to 180 ◦C and the heating rate was 10 ◦C/min.

(8) Solubility of foams in water

The samples were dried in a drying oven at 60 ◦C for 24 h and then soaked in deionized water at room temperature. The shape of the sample in water was recorded until it dissolved.

### **3. Results and Discussion**

### *3.1. Molecular Weight of CTS*

The viscosity of CTS solution at a particular concentration directly reflects its molecular weight. Assuming other factors to be fixed, the higher the relative molecular weight of CTS, the higher will be the viscosity of the CTS solution, and the smaller the relative molecular weight is, the smaller will be the viscosity of the CTS solution. Figure 2 illustrates the trend line of the CTS sample, with a low viscosity extrapolated to the intersection point of the straight line and y-axis to be 502.39, implying that the characteristic viscosity of the CTS, [η] was 502.39, and the molecular weight of the CTS sample calculated according to the MHS equation was 3.5 <sup>×</sup> 105. Likewise, the molecular weights of other two types of CTS were also estimated through the same calculation; Table 2 summarizes the values of the molecular weight. The correlation coefficient (R2) of CTS with different viscosities was greater than 0.95, which indicates the reliability of the calculated results. The molecular weight of CTS with viscosity between 200 and 400 mPa·s was 4.4 <sup>×</sup> 105 and that of CTS with a viscosity of 400 mPa·s was 5.2 <sup>×</sup> 10<sup>5</sup> (Table 2), and the difference was significant; the trend was completely consistent with the above-mentioned.

**Figure 2.** Intrinsic viscosity of CTS in 0.2 M HAc/0.1 M NaAc at 30 ◦C.

**Table 2.** Molecular weight of chitosan with different viscosities.


### *3.2. Morphology of Starch-Based Foams*

Figure 3 presents SEM observations of the starch-based foam prepared using the molecular weight of CTS as a variable. Figure 3a,b indicate that the size and distribution of pores on the foam surface produced by microwave radiation are not as apparent as those observed in the inner layer structure. The pores on the surface layer were small and unevenly distributed, whereas the bubbles in the inner structure were significantly larger and more uniform. Starch-based foams made of CTS with different molecular weights also demonstrated distinct characteristics. Comparison of the three types of foam surface structure (Figure 3a,c) revealed the largest number of pores in the foam with CTS of molecular weight 4.4 <sup>×</sup> 105 (CTS-4.4/PS) (Figure 3e); the quantity of pores in foam with the CTS of molecular weight 3.5 <sup>×</sup> 105 (CTS-3.5/PS) was relatively smaller (Figure 3a) than that in the CTS of molecular weight 5.2 <sup>×</sup> <sup>10</sup><sup>5</sup> (Figure 3c). In short, the increasing order of the number of pores according to the SEM observation is as follows: Figure 3e > Figure 3a > Figure 3c. The trend observed in the pore quantity on the inner structure was exactly the opposite of that observed in the surface morphology. From the SEM images in Figure 3b,d,f, we can clearly observe that the starch-based foam made from medium-viscosity CTS (Figure 3d) has the largest number of pores, relatively smaller pore size (a rough comparison by the yellow circles in Figure 3b,d,f), and relatively regular bubble morphology compared with other foam types.

**Figure 3.** The surface morphologies of foams made from low (**a**), medium (**c**), and high (**e**) viscosity CTS and the corresponding inner morphologies (**b**,**d**,**f**). Scale bar = 500 μm

### *3.3. Pore Size, Porosity, and Density of CTS–PS Foams*

The pore size and porosity were measured using Image-J software, and Table 3 presents the specific values. We observed that at a fixed amount of CTS solution and with an increase in the starch mass, the bubble porosity and porosity decrease in all the CTS–PS composites. The decrease in foam hole and porosity in the same volume mold with increase in starch content is quite reasonable. However, the pore size and porosity of foams with different levels of starch quality also changed with changes in the molecular weight of CTS. When the mass ratio of CTS solution and starch was 3:4, the bubble size and porosity of CTS-4.4/PS foam composites were the smallest (281.7 μm and 55.39%, respectively), whereas the difference in bubble size between CTS-3.5/PS and CTS-5.2/PS foams was not

significant, which indicates that the molecular weight of CTS affects the starch-based foams to a certain extent; however, the effect is limited.


**Table 3.** The pore size and porosity of foams with CTS solution for 4 wt%.

According to the density of the starch-based foam shown in Figure 2, when the mass fraction of CTS solution was 4%, the density of foams prepared from CTS solution with the same molecular weight gradually increased with the increase in starch content. However, when the mass ratio of CTS solution and starch was constant, the change was different from that observed earlier, and the density of CTS-4.4/PS foams was apparently higher (Figure 4). Figure 4 shows the influence of mass fraction of different CTS solutions on starch-based foams, when the mass ratio of CTS/PS was 3/4.2. Notably, the mass fraction of CTS solution had only a minor effect on the density of the entire foam material, whereas the molecular weight of CTS exerted a greater effect on the same. Table 4 summarizes the specific values of all composites. Combined with the aforementioned SEM observations on the pore size and porosity of foam materials, we can assume that by using a particular mass fraction of the CTS solution, the pore structure and density of materials can be improved by appropriately increasing the starch content. This is because when the starch content increases, the formic acid in the CTS solution and PS generates more esterification reactions in the microwave, and since the volume of the mold is constant, it appears as a prepared foam in a macroscopic view. Thus, the pore size and porosity gradually decrease, and correspondingly, the density of foams gradually increases.

**Figure 4.** The density of foams with CTS solution for 4.2 wt%.


**Table 4.** The density values of CTS–PS foams.

### *3.4. The Compressive Property of Foams*

Biodegradable foam must possess definite mechanical strength to maintain its integrity during transportation, and the density and compressive strength of foams are somewhat correlated. Figure 5 illustrates the curves for the compressive strength of PS-based foams with change in the molecular weight of CTS and mass ratio of CTS/PS to 3/4.2 at 10% compressive ratio. The variation trend of these curves was basically consistent, as shown in the figure. The CTS-4.4/PS foams exhibited greater density and superior compressive performance compared with CTS-3.5PS and CTS-5.2PS foams. The increase in the density of the foam increases the bearing capacity of a single bubble hole and the effective bearing area within the unit section, which is generally reflected in the increased overall compressive strength of the foams.

**Figure 5.** Compressive strength of foams with chitosan solution for 4.2 wt%.

The foams composed of CTS and starch exhibit such strength mainly because of the formation of a high intermolecular hydrogen bond between NH3 <sup>+</sup> in the main chain of CTS and OH<sup>+</sup> in the starch molecule. The amino group of CTS is easily protonated to form NH3 <sup>+</sup> in formic acid solution, and the ordered crystal structure in the starch molecule is destroyed during the microwave irradiation treatment, which leads to the formation of a hydrogen bond between the hydroxyl group of starch and amino group of CTS. In addition, electrostatic interactions between NH3 <sup>+</sup> in CTS and phosphate groups in PS may occur. Phosphorus, the most important element in PS, exists in the form of a covalent bond in PS. Glucose-6-phosphate (pKa1 = 0.94, pKa2 = 6.11) is the main structure of esterified phosphate in PS, and its acidity is stronger than that of orthophosphate [30,31]. Therefore, phosphate groups in PS possess a negative charge in aqueous solutions and do not combine with other negatively charged substances. The amino groups in the CTS solution possess a positive charge after being protonated and interact with the negatively charged phosphate groups in PS, which further improves the compressive performance of the entire CTS-PS foams. Under the condition of 4.2 wt% CTS, the compressive strength of CTS-PS foam increased with the increase in starch quality and the maximum compressive value was obtained at a mass ratio of 3/4.2. The molecular weight of CTS demonstrated a certain influence on the compressive strength of the starch-based foams. Among the foam composites prepared from CTS with three different molecular weights, the CTS-4.4/PS foams had the highest compressive strength value and those of CTS-3.5/PS and CTS-5.2/PS foams were extremely close (Table 5). The interaction between starch and CTS has been reported to enhance the functional properties [32]; however, when the number of NH3 <sup>+</sup> species increases beyond a critical value, synthesis of uniform CTS/PS foams becomes challenging and results in weak interaction at the boundary and poor compressive performance.


**Table 5.** Compressive strength of starch-based foams.

To determine the effect of each factor on the foam compression performance, we further designed an orthogonal experiment and variance analysis. Table 6 presents the experiment results. According to range (R) analysis, R (C) > R (A) > R (B), and combined with the variance analysis in Table 7, we found that factor A (CTS molecular weight) and factor C (the mass ratio of CTS to starch) were all significantly affected, whereas factor B (CTS mass fraction) displayed no significant difference in the experimental results (P > 0.05), which indicates that the mass fraction of CTS fluctuating slightly above 4 wt% does not considerably affect the compressive performance of the foams. According to the results of intuitive and variance analyses, the optimal combination of foam preparation was A2B2C3, implying that the CTS molecular weight of 4.4 <sup>×</sup> 105, CTS mass fraction of 4%, and the CTS/PS mass ratio of 3/4.2 are optimal for foam synthesis. We verified the compression performance of the CTS starch-based foam under these conditions, which was higher (1.077 mPa) than those observed in other conditions.


**Table 6.** Design and results of orthogonal test.

**Table 7.** Variance analysis of orthogonal test.


### *3.5. FTIR Analysis*

To study the interaction between CTS and PS, an infrared spectrum of each material was obtained. Figure 6 illustrates an extremely wide stretching vibration band at 3413 cm−<sup>1</sup> for the dried PS corresponding to the characteristic absorption peak of the hydroxyl group in the starch. A moderate intensity peak at 2930 cm−<sup>1</sup> was associated with –CH2 antisymmetric stretching vibration, and an

absorption band at 1647 cm−<sup>1</sup> was attributed to water absorption in the amorphous region of PS; the absorption peak near 1380 cm−<sup>1</sup> was ascribed to –CH3 bending vibration and –CH2 twisted vibration. C–C, C–O, and C–H stretching vibrations and bending vibration of C–OH formed obvious absorption peaks in the 1300–800 cm−<sup>1</sup> region of the spectrum. CTS molecules also displayed a strong and wide absorption peak at 3428 cm<sup>−</sup>1, corresponding to the stretching vibration of O–H and N–H; the absorption peaks at 1660, 1500, and 1425 cm−<sup>1</sup> corresponded to amide-I belt stretching vibration, –NH2 deformation vibration, and –CH2 and –CH3 bending vibrations, respectively. The characteristic absorption peak at 1157 cm−<sup>1</sup> has been ascribed to the CTS C–O–C asymmetric vibration [33,34].

**Figure 6.** Infrared spectrogram of materials.

In a study, the interaction between CTS and starch in starch-based biological foam was shown to cause clear changes in the surroundings of polymer groups, which led to variations in the frequency and intensity of relevant absorption bands, and a unique response was obtained with changes in the characteristic spectra peak wave numbers [35]. Relative to the infrared spectra of pure CTS and starch, the amide peaks of CTS at 1660 cm−<sup>1</sup> and the characteristic peaks of starch at 1646 cm−<sup>1</sup> moved to a low-frequency region, and the intensity of the absorption band increased. Correspondingly, the absorption band strength of foam composites at 3413 cm−<sup>1</sup> increased, which indicates that the hydrogen bond density in the CTS-starch system increased and inter- and intra- molecular hydrogen bonds formed between amino and hydroxyl groups in molecular chains of CTS and starch, thus improving the compatibility between CTS and starch [36,37].

### *3.6. Thermogravimetric Analysis (TGA)*

The thermogravimetric curves and corresponding thermogravimetric rate curves of raw materials and starch-based foams are shown in Figure 7. From Figure 7a,b, it was clear that all materials exhibited three stages of heat loss during the whole pyrolysis process. The weight loss in the first stage below 100 ◦C could be attributed to the evaporation of water. Polysaccharides usually have a strong affinity for water, become hydrated easily, and form large molecules with a messy structure. Hydration properties of these polysaccharides depend on their primary and supramolecular structures [38]. The peak temperatures (Tp, shown in Table 8) for the water evaporation rates of chitosan with different molecular weights were 72.11, 74.70, and 72.17 ◦C, respectively. Moreover, the foams prepared by chitosan with different molecular weights reached their peaks at these temperatures. The water evaporation

rate of chitosan increased with an increase in the molecular weight even though the increment was small. At the same time, the characteristic temperature points of the chitosan–starch foams moved to a higher temperature at this stage. This was due to the cross-linking between chitosan and starch to form new hydrogen bonds that inhibited the increase in water diffusion and thus amplified the water penetration resistance.

**Figure 7.** Thermal decomposition process of materials (**a**): raw materials, (**b**): CTS–PS foams.


**Table 8.** Thermal decomposition parameters of materials.

<sup>a</sup> T1 = Temperature at the peak below 100 ◦C on derivative thermogravimetry (DTG) curve, Tp = Temperature at the peak of the whole DTG curve, Residue = Mean value of residual mass.

The second thermogravimetric stage (between 200 to 400 ◦C) involved the depolymerization of chitosan molecular chains and the cleavage of starch molecular chains; this included the dehydration of sugar rings and the depolymerization and decomposition of glucose units [39]. Due to the similar depolymerization temperatures of starch and chitosan molecular chains, chitosan–starch foams presented the peak of the mixed weight loss rate of starch and chitosan molecules. The corresponding peak temperatures for the three kinds of foams were 302.88, 305.31, and 305.22 ◦C. The pyrolysis temperatures of the foams were higher than the temperatures observed for pure starch and chitosan. This indicated that the molecular interaction between the two biomolecules increased the molecular depolymerization temperature. Finally, the high-temperature "tail" between 340 and 500 ◦C was

the third stage of the pyrolysis of chitosan and the vaporization and elimination of starch volatile products [40]. TGA curves showed the high-temperature characteristics of the biological starch-based foams. They depicted that the hydrogen bonding between chitosan and starch resulted in better thermal stability than that observed for neat starch and chitosan. Moreover, the maximum weight loss rate was associated with the molecular weight of chitosan. The value Tp of CTS-4.4/PS was higher than that of the others. This result could be attributed to increased hydrogen bonding and charge interactions.

#### *3.7. DSC Analysis*

The DSC diagrams of chitosan and starch and CTS/PS foams are shown in Figure 8. The convex peak represents the heat absorption of materials during the glass transition, the intersection point between the peak baseline and the tangent of the rising stage is the initial temperature of glass transition (To), the apex of the peak is the peak temperature (Tp), the peak stage of tangent and baseline by the intersection of terminated for gelation temperature (Tc) and the area of the peak for absorption of heat enthalpy (ΔH); these values are listed in Table 9. As can be seen from Table 9, To, Tp, Tc for chitosan were 34.65, 69.57, 104.72 ◦C, respectively, and ΔH was 133.53 J/g. The To, Tp, ΔH of porous foams after microwave treatment showed a backward trend with the increase in molecular weight of chitosan. Tester et al. [41] believed that the vitrification transition temperatures (To, Tp, and Tc) were associated with the perfection and stability of crystal. The advancement of glass transition temperature indicated that the temperature required to dissolve the crystal was reduced, indicating the worse stability of the crystal, and the backward shift was the opposite. Therefore, it can be concluded that the molecular weight of chitosan affected the stability of crystals in starch-based foam: the higher the molecular weight of chitosan, the greater the stability of the crystal in foams. The content of crystals could be determined by comparison of enthalpy and the enthalpy of samples increased with the increase in the crystal content. In addition, the improvement of crystal integrity, the enhancement of interaction between amylose or between amylose and amylopectin, the formation of chitosan and starch complex, and the formation of orderly crystalline regions may also increase the enthalpy value [42]. As shown in Table 9, the enthalpy of pure starch is 131.33 J/g, which is higher than that of some chitosan–starch composites; this is due to the easy formation of a double helix structure between short-chain molecules in starch, making the internal structure of starch stronger and more close. CTS/PS foams increased with the increase in the molecular weight of chitosan, which may be caused by the formation of partially ordered crystal regions between starch molecules and chitosan complexes and the longer the molecular chain of chitosan is, the slower its crystal structure is destroyed in the glass transition.

**Figure 8.** DSC diagrams of raw materials and CTS/PS foams.


**Table 9.** The glass transition temperature and enthalpy of samples.

To = initial temperature of glass transition, Tp = the apex of the peak is the peak temperature, Tc = the peak stage of tangent and baseline by the intersection of terminated for glass transition temperature; ΔH = the area of the peak for absorption of heat enthalpy.

### *3.8. Water Solubility Analysis*

Figure 9 shows the solubility of CTS-starch-based foams in deionized water. The overall structure of starch-based foams was uniform and the edges were clearly visible (Figure 8a) when foams were submerged in water. After 3 days, the volume of foams increased and the edges appeared to soften. Furthermore, the turbidity of deionized water increased relative to that in the beginning of the experiment. After 10 days, the exterior foam color turned milky white, but the structure of foams remained intact. The foams began to dissolve in water after 18 days, the volume of foams gradually decreased, and water gradually mixed until foams were completely dissolved after 30 days. The whole dissolution process of CTS-starch-based foams demonstrated that the interaction between CTS and starch molecules can keep the morphological structure of foams intact for 10 days in water, but this interaction is effective for a limited time period.

**Figure 9.** *Cont.*

**Figure 9.** The morphology of CTS-starch-based foams in deionized water (**a**): start time, (**b**): 3 days later, (**c**): 10 days later, (**d**): 18 days later, (**e**): 24 days later, (**f**): 30 days later.

### **4. Conclusions**

Biological foams containing CTS and PS were prepared successfully in a new polytef mold in a microwave oven. The effects of different viscosities and concentrations of CTS and different mass ratios of CTS-starch on the morphological, physicochemical, mechanical, and thermal properties of CTS–PS foams were comprehensively studied. The research conclusions can be summarized as follows:

(1) The viscosity of CTS solution at a particular concentration directly reflects the molecular weight of CTS. The molecular weights of CTS with different viscosities were 3.5 <sup>×</sup> <sup>10</sup>5, 4.4 <sup>×</sup> <sup>10</sup>5, and 5.2 <sup>×</sup> <sup>10</sup>5. The results confirmed that high viscosity corresponds to high relative molecular weight of CTS.

(2) Analysis of morphology, pore size, and density of starch-based biological foams indicated that CTS foams comprising CTS of molecular weight 4.4 <sup>×</sup> 10<sup>5</sup> exhibit small pores and low density. Moreover, the foams with a larger proportion of starch demonstrated small pore size and low density. Interestingly, the reverse was found to be true in the case of compressive properties; foams comprising CTS of molecular weight 4.4 <sup>×</sup> 105 exhibited the highest compressive strength, and foams with a larger proportion of starch exhibited higher compressive strength. In addition, orthogonal experiments were set up to confirm that the compressive performance of foams is highest when the molecular weight of CTS is 4.4 <sup>×</sup> 105, the mass fraction is 4 wt%, and the mass ratio of CTS-PS is 3/4.2.

(3) The change in the characteristic peaks of the foams composed of CTS and PS in the FTIR diagram indicated that the amino group in CTS interacted with the hydroxyl group in the starch. In TGA and DSC analysis, it could be seen that the interaction between starch and chitosan improved the thermal stability of the foam, and in the water solubility test, the interaction maintained the integrity of the foam's morphological structure in water within a certain period of time, and the foams completely degraded after 30 days.

**Author Contributions:** R.H. designed the study and conducted parts of experiments and data analysis. R.H. and X.Z. wrote the initial manuscript. X.Z. and Z.T. performed the composite property testing/data analysis. All authors have read and agreed to the published version of the manuscript.

**Funding:** This collaborative study was supported by Major projects of Natural Science Foundation of Jiangsu (18KJA220002); China Postdoctoral Science Foundation: Special Program (2017T100313); China Postdoctoral Science Foundation: General Program (2016M601821). Natural Science Foundation of China (Grant No. 51606103). And the Postgraduate Research and Practice Innovation Program of Jiangsu Province (2019NFUSPITP0004).

**Conflicts of Interest:** There is no conflict of interest among authors.

### **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Natural Inspired Carboxymethyl Cellulose (CMC) Doped with Ammonium Carbonate (AC) as Biopolymer Electrolyte**

### **Mohd Ibnu Haikal Ahmad Sohaimy <sup>1</sup> and Mohd Ikmar Nizam Mohamad Isa 1,2,\***


Received: 30 July 2020; Accepted: 21 September 2020; Published: 26 October 2020

**Abstract:** Green and safer materials in energy storage technology are important right now due to increased consumption. In this study, a biopolymer electrolyte inspired from natural materials was developed by using carboxymethyl cellulose (CMC) as the core material and doped with varied ammonium carbonate (AC) composition. X-ray diffraction (XRD) shows the prepared CMC-AC electrolyte films exhibited low crystallinity content, *Xc* (~30%) for sample AC7. A specific wavenumber range between 900–1200 cm−<sup>1</sup> and 1500–1800 cm−<sup>1</sup> was emphasized in Fourier transform infrared (FTIR) testing, as this is the most probable interaction to occur. The highest ionic conductivity, σ of the electrolyte system achieved was 7.71 <sup>×</sup> <sup>10</sup>−<sup>6</sup> Scm−<sup>1</sup> and appeared greatly dependent on ionic mobility, μ and diffusion coefficient, *D*. The number of mobile ions, η*,* increased up to the highest conducting sample (AC7) but it became less prominent at higher AC composition. The transference measurement, *tion* showed that the electrolyte system was predominantly ionic with sample AC7 having the highest value (*tion* = 0.98). Further assessment also proved that the H<sup>+</sup> ion was the main conducting species in the CMC-AC electrolyte system, which presumably was due to protonation of ammonium salt onto the complexes site and contributed to the overall ionic conductivity enhancement.

**Keywords:** biopolymer; carboxymethyl cellulose; solid polymer electrolyte; ionic transport

### **1. Introduction**

Apart from increased energy and power density of energy storage, one of the remaining challenge in advancement of energy storage technologies such as portable electronic devices, smart grids and electric vehicles is to lower the production cost as well as to reduce the environmental effect [1–3]. Thus, one of the steps in order to achieve this is to transform the electrolyte used in energy storage. The electrolyte is a medium that facilitates the movement of ions across the electrode in order to complete the circuit and usually the electrolyte is in liquid form since ionic transport is easier in liquid phase. However, the development of solid polymer electrolyte (SPE) film has garnered much attention as it offers outstanding advantages compared to other form of electrolytes [3–5]. Several types of polymer were used to investigate the possibilities for energy storage material such as polyethylene oxide (PEO), polyvinyl chloride (PVC), poly(methyl methacrylate) (PMMA), polyvinylidene fluoride (PVDF) and polyacrylonitrile (PAN) [4,6–8]. Alternatively, polymers originated from natural sources (biopolymer) such as starch, chitosan, gelatin and agar [9–12] are also becoming an attractive substitute for synthetic polymers partly due to superior mechanical, electrical properties and for being cheaper [9]. Equally, cellulose in particular is another good choice since it is literally found everywhere but cellulose

in its original form is hard to utilize due to its inability to dissolved and shaped. However, derivatives of the base polymer can overcome this problem.

Currently several cellulose derivatives biopolymer can be found commercially such as methyl cellulose (MC), hydroxyethyl cellulose (HEC) and carboxymethyl cellulose (CMC) [13–15]. CMC in particular is a good SPE biopolymer candidate. This is attributable to its good film-forming abilities due to the presence of a hydrophilic carboxyl group (–CH2COONa) which allows the biopolymer to dissolve in cheap solvent (water) [16]. On top of that, CMC has a naturally high degree of amorphous phase [17]. These allow easier transport of conducting ions like lithium (Li+) and proton (H+). The major flaw with SPE is the low ionic conductivity due to inherent solid properties of the polymer itself, which impede ionic mobility especially at room temperature [18]. For a biopolymer electrolyte to be utilized in commercialized energy storage technology, the ionic conductivity needs to be enhanced to at least a minimum value of ~10−<sup>4</sup> Scm−1, and one of the most prominent methods for ionic conductivity enhancement is by introducing the ionic dopant into the polymer electrolyte. Ammonium salts are the dopant materials focused upon in this research. Ammonium nitrate (A–N) and ammonium fluoride (A–F) are some examples of ammonium salts [17,19] and are the most common proton donors. Deprotonated of H<sup>+</sup> from the ammonium ion group (NH4 <sup>+</sup>) of ammonium salts helps to improve ionic conductivity of the electrolyte system. Thus in theory, ammonium carbonate (AC) which have two group of ammonium ions could inject more H<sup>+</sup> into the system.

The ionic conductivity is the most vital parameters in order to determine the viability of the electrolyte system. Nevertheless, the value calculated only gives a general overview of the electrolyte performance. Fundamentally, the ionic conductivity of the electrolyte is the product of ionic density, η, ionic mobility, μ and the elementary charge, *e*. Thus, it can be beneficial to acquire a detailed transport number for the electrolyte system as it can help to evaluate and later suggest an appropriate technique to improve the ionic conductivity of the electrolyte system. Several methods have been employed previously by researchers to calculate the transport number [20,21]. The commonly used one is by using the Fourier transform infrared (FTIR) deconvolution approach, which is also the method chosen for this study [22–24]. Apart from that, a high cation transference number is also crucial in electrolyte system as it can reduce charge concentration gradient and subsequently produce higher power density when assembled in a battery [25].

This paper aims to develop SPE from a biopolymer, which is CMC doped with ammonium carbonate (AC) as the ionic source, then to find the ionic conductivity and the transport parameters contributing towards the ionic conductivity behavior in CMC-AC biopolymer films. Ionic conductivity was measured using electrical impedance spectroscopy (EIS). The X-ray diffraction (XRD) and Fourier transform infrared (FTIR) spectrum of the electrolyte was de-convoluted to quantitatively determine the crystallinity content and the transport number (number of mobile ions, η mobility of mobile ions, μ and diffusion coefficient, *D*), respectively. A direct current (dc) polarization technique was utilized to confirm the ionic transference number of CMC-AC biopolymer films. From these analyses, the correlation between the results towards the ionic conductivity behavior can be ascertained.

#### **2. Materials and Methods**

### *2.1. Preparation of Carboxymethyl Cellulose-Ammonium Carbonate (CMC-AC) Electrolytes Films*

Carboxymethyl cellulose (CMC) (Across, NJ, USA) were dissolved in 100 mL of distilled water until homogenous before adding (1–11 wt.%) of ammonium carbonate salts (AC) (Merck, Darmstadt, Germany) which is calculated using Equation (1), where the *x* represents the weight of added salt and *y* the weight of CMC. The compositions of the CMC-AC biopolymer films are shown in Table 1. The CMC-AC biopolymer electrolyte solution was casted into petri dishes and left for a period of time (1 month) at room temperature to dry into thin film. Three (3) sample replicates were produced for each

composition. The CMC-AC biopolymer films obtained are free standing with no phase separation observed (Figure S1).

$$\text{composition} = \frac{\text{x}}{\text{x} + \text{y}} \times 100\tag{1}$$



#### *2.2. Electrical Impedance Spectroscopy (EIS)*

A HIOKI 3532-50 LCR Hi-Tester (HIOKI, Nagano-ken, Japan) used to investigate the conductivity of CMC-AC biopolymer film. Each samples tested in the frequency range of 50 Hz to 1 MHz by utilizing stainless steel as blocking electrodes at a temperature range of 303–363 K.

### *2.3. X-ray Di*ff*ractometer (XRD)*

The degree of crystallinity of CMC-AC biopolymer films obtained using Rigaku Miniflex II X-ray diffractometer (Rigaku, Tokyo, Japan) with *CuK*α radiation sources. The radiation sources directed the X-ray beam onto the films and scanned at angle, 2θ between 5◦ and 80◦.

### *2.4. Fourier-Transform Infrared (FTIR)*

Complexation or interaction between CMC and AC salts were determined by using a Thermo Nicolet 380 FTIR spectroscopy (Thermo Fischer, Madison, WI, USA) equipped with attenuated total reflection (ATR). The spectrum recorded in a wavenumber ranged from 800–4000 cm−<sup>1</sup> and with resolution of 4 cm<sup>−</sup>1.

### *2.5. Transference Number Measurement (TNM)*

The dc polarization techniques with 1.5 V fixed voltage were applied across the samples held by a stainless steel holder. A digital multimeter and computer were also connected to the circuit to collect the current (*I*) value against time (*t*). The transference number, *tion* of CMC-AC biopolymer films was calculated from the testing.

#### **3. Results**

### *3.1. XRD Analysis*

Figure 1a shows the X-ray diffraction (XRD) result of pure AC and CMC-AC biopolymer films. From the figure, pure AC salt displays several sharp peaks which represent the crystalline nature of AC salt with three prominent peaks can be observed at angle 2θ = 22.54◦, 30.2◦ and 34.54◦. The XRD pattern for CMC-AC biopolymer films show a similar pattern after the addition of AC salts where the amorphous hump is centered at 2θ = 21.1◦ (Figure 1a). It can be noticed that none of crystalline peak observed from AC salt appeared for all films, which indicates the AC salts dissolved within CMC amorphous phase to form an electrolyte system with improved electrolyte properties [26]. The amorphous phase will lead to higher ionic conductivity due to greater ionic diffusion, since the

amorphous polymers have a flexible polymeric backbone [27]. Several reports have proven that the high amorphous phase has led to ionic conductivity improvement [28–30].

**Figure 1.** (**a**) The X-ray diffraction (XRD) diffractograms of all CMC-AC biopolymer films. (**b**) The de-convoluted plot of XRD for sample AC7 where green line represents the crystalline phase and blue line represents the amorphous phase.

De-convolution techniques on XRD spectra can reveal the explicit details of the amorphous or crystallinity phase of electrolyte films [31]. The de-convoluted XRD diffractograms show several peaks where a narrow and sharper peak (solid green line) represents the crystalline phase and a broader peak (dotted blue line) represents the amorphous phase (Figure 1b). From the de-convoluted peak, the degrees of crystallinity (*Xc*) of each CMC-AC biopolymer films were calculated using Equation (2) and the results are given in Table 2.

$$\text{degree of crystallization, } X\_{\varepsilon}(\%) = \frac{A\_{\varepsilon}}{A\_{\varepsilon} + A\_{a}} \times 100\% \tag{2}$$


**Table 2.** The respective area of crystalline (*Ac*) and amorphous (*Aa*) phase and degree of crystallinity (*Xc*) of each CMC-AC biopolymer films.

The overlapping peaks and area forms under the diffractograms were determined for either amorphous (*Aa*) or crystalline (*Ac*) phases before calculating the *Xc*, and the results are tabulated in Table 2. The degree of crystallinity, *Xc* for the CMC-AC biopolymer films is in between 46% to 30% as shown in the table. The irregularity (random arrangement) of the polymer structure due to the increase in amorphous phase can led to lower energy barrier between hopping sites. Thus, the increasing amorphous phase allows for easier ion conduction which in this case is the H<sup>+</sup> ion through the polymer matrix and subsequently increases the ionic conductivity value. The degree of crystallinity, *Xc* trend of current work is similar compared to other works which the highest conducting sample has the lowest *Xc* [32,33].

### *3.2. FTIR Analysis*

Figure 2 shows the FTIR spectrum of the CMC-AC biopolymer films in the range of (a) 1500–1800 cm−<sup>1</sup> and (b) 900–1200 cm−1. Overall the FTIR spectrum can be seen in Figure S2 in the Supplementary Materials. These ranges are responsible for the vibration of carboxyl group (COO) of CMC, where it is the probable site for interaction between CMC and AC [17,19,34]. The FTIR peak centered at 1056 cm−<sup>1</sup> and 1591 cm−<sup>1</sup> correspond to the vibration peak of C–O and C=O of the COO, respectively. As seen in Figure 2, there is no shifting of the peak indicating that the AC did not directly interact with that particular site. However, the C=O peak shows an observable reduction in intensity as the AC composition increases which implies that the AC salt weakly interacts with the COO. The presence of free lone pair electron at C=O attract free ions to interact. The small bump appeared around ~1650 cm−<sup>1</sup> is believed due to the AC salt. The AC salt believed to dissociate to NH4 <sup>+</sup> and CO3 <sup>2</sup><sup>−</sup> ions. This allows a protonation process between the loosely bonded H<sup>+</sup> from NH4 <sup>+</sup> structure to the complex site [19]. The protonated H<sup>+</sup> then able to transport to the conduction site (C=O) and subsequently improve the ionic conductivity of the electrolyte films. This can be further clarify from the impedance and transport analysis.

**Figure 2.** Fourier transform infrared (FTIR) spectra of CMC-AC biopolymer films range from (**a**) 1500–1800 cm−<sup>1</sup> and (**b**) 900–1200 cm<sup>−</sup>1.

#### *3.3. Ionic Conductivity and Transport Properties*

The real (*Zr*) and imaginary (*Zi*) impedance data obtained from EIS testing presented in the form of Cole-Cole plot (Figure 3a,b). This plot reveals semi-circle at high frequency (left side of the plot) and slanted line at low frequency (right side of the plot). This shape can be explained in terms of the electronics component where the semi-circle represents parallel combination of capacitor and resistor [35]. The resistor represents the ionic migration through the polymer matrix while the capacitor represents the polarized polymer chains [35]. The slanted line is due to the electrode polarization effect where electrical double layer capacitance occur due to the blocking stainless steel electrode used in this testing [36]. The bulk resistance, *Rb* value for each CMC-AC biopolymer film was obtained from the intersection of the semi-circle and slanted impedance line value at the *x*-axis [15]. The value of ionic conductivity is calculated using Equation (3). The *t* and *A* in the equation correspond to film thickness and the contact area, respectively. The ionic conductivity value at room temperature (303 K) against AC composition (wt.%) is plotted in Figure 3c. As seen in the figure, the ionic conductivity increases for sample AC1 (1 wt.%) to AC7 (7 wt.%) with the value of 7.71 <sup>×</sup> 10−<sup>6</sup> Scm−<sup>1</sup> and later dropped for sample AC9 (9 wt.%) and sample AC11 (11 wt.%). This was attributed to the diminishing effect of bulk resistance, which was apparent with the size changes to the impedance semi-circle as seen in Figure 3a,b, that indicates lower resistance for the mobile ions to move [17]. Table 3 shows the comparison of ionic conductivity of current work with other polymer-doped ammonium salt electrolyte systems. This work exhibits higher ionic conductivity when compared to other single ammonium and di-ammonium salt system. Figure 3d shows the ionic conductivity at elevated temperature for selected biopolymer films. In the investigated temperature range, the ionic conductivity increases as temperature increases. This is due to two reasons, (1) the AC absorbs the thermal energy and allows more ions dissociation, (2) the CMC backbone vibrate and create space for ionic conduction thus increase ionic conductivity at high temperature. This indicates the CMC-AC biopolymer films is thermally assisted which is proven from good fitting (R<sup>2</sup> = 0.93–0.98) of the ionic conductivity to

the Arrhenius relationship (Equation (4)). Transport properties analysis can further reveal the details behind the ionic conductivity improvement of the CMC-AC biopolymer system.

$$\text{Ionic conductivity, } \sigma = \frac{t}{R\_b \times A} \tag{3}$$

$$
\sigma = \sigma\_0 \varepsilon x p^{(\frac{-k\_0}{kT})} \tag{4}
$$

**Figure 3.** (**a**,**b**) Impedance plot for CMC-AC biopolymer films and (**c**) the ionic conductivity of CMC-AC biopolymer films (**d**) ionic conductivity at elevated temperature.

**Table 3.** Ionic conductivity comparison of polymer with ammonium salt system measured at room temperature from literature and this work.


Arof et al. successfully calculated the transport number of their electrolyte system through FTIR peak (de-convoluted technique) [42]. Thus, the same technique was applied to the current work. The de-convoluted FTIR spectra were assigned into free ions and contact/aggregates ions to determine the area of de-convoluted peak before calculating the transport number of the electrolyte system. The FTIR spectra range of 1500–1750 cm−<sup>1</sup> was selected for de-convoluted process to find the peak originated from ammonium (NH4 <sup>+</sup>) ion. Figure 4a shows the de-convoluted peak of each CMC-AC biopolymer film. The peak centered at ~1589 cm−<sup>1</sup> is the peak of the carboxylic group (COO–) of

CMC [17]. The free ions peak centered from 1594–1643 cm−<sup>1</sup> while the contact ion centered from 1650–1684 cm<sup>−</sup>1. The percentage of free ions were first calculated using Equation (5), where *Af* is the area of free ions peak and *Ac* is the area of contact ions peak. Afterward, the percentage of free ions obtained were used to determine the transport number (number of mobile ions, η ionic mobility, μ and diffusion coefficient, *D*) using Equations (6)–(8) [43].

$$\text{Percentage of free ions } \left( \% \right) = \frac{A\_f}{A\_f + A\_c} \times 100\% \tag{5}$$

$$\eta = \frac{M \times N\_A}{V\_{Total}} \times \% \text{ of free ions} \tag{6}$$

$$
\mu = \frac{\sigma}{\eta \epsilon} \tag{7}
$$

$$D = \frac{\mu k\_B T}{\varepsilon} \tag{8}$$

**Figure 4.** (**a**) The de-convoluted spectra of each sample and (**b**) the plot of respective transport parameters in CMC-AC biopolymer film.

In Equation (6), *M*, *NA*, and *VTotal* stands for the number of moles of salts used, Avogadro's number and the total volume of electrolytes films respectively, where *VTota*<sup>l</sup> is equal to mass, *m* divided by density, ρ of each material used. In Equation (7), σ is the ionic conductivity of the electrolyte films and *e* is the electric charge. In Equation (8), *kB* is the Boltzmann constant and *T* is the absolute temperature in kelvin.

The peak for free ions originated from NH4 <sup>+</sup> ions, which have deprotonated into NH3 <sup>+</sup> ions. The peak position found from the de-convoluted plot is in good agreement with other reports [44,45]. The percentage of free ions peak is calculated and tabulated in Table 4. From the table, the CMC-AC biopolymer films has a percentage of free ions is in the range between 59% and 85%. For sample AC1–AC7, the percentage of free ions decreases from 85.48% for sample AC1 to 59.53% for sample AC7 which corresponds to the highest conducting sample, before it starts to increase for sample AC9 and AC11. This is in inverse manner compared to other reports where the reports show that the percentage

of free ions is usually with the highest ionic conductivity sample [46–48]. However, the percentage value only gives a general view of the transport number. Further analysis of the transport number (Equations (6)–(8)) will further help to explain the transport behavior.

**Table 4.** The respective area and percentage of free ions and contact ions of each sample.


Figure 4b presents the calculated transport number. From the figure, the value of η increase with each AC composition added for sample AC1 to AC7 (low AC composition film). At low composition, AC salt is able to dissociate into its respective ions (NH4 <sup>+</sup> and CO3 −) and has a small probability to form an ion cluster (contact ions), thus allowing one of the protons (H+) which is loosely bound to the NH4 <sup>+</sup> structure to be protonated and transported across the medium. As a result, the μ and *D* value is increased. At higher AC composition films (AC9-AC11), the value of η continues to increase in contrast to the ionic conductivity, which dropped along with μ and *D* value. The ions packed closely in the electrolyte when a huge amount ionic dopant supplied into the system. Consequently, this makes the de-protonation and protonation process at the ammonium (NH4 <sup>+</sup>) ions occurs almost spontaneously and reduces the effective mobility of mobile ions, and consequently the diffusion rate thus prove μ and *D* dependency [47]. The dependency of μ and *D* towards ionic conductivity value is similar to the factors affecting ionic conductivity in a machine-learning framework proposed by Shi et al. [1], which mentioned diffusion coefficient and ionic radius as some of the factors. In this work, the diffusion coefficient improvement is believed due to small H<sup>+</sup> ionic radii. Scheme 1 illustrates the behavior of ionic transport in the CMC-AC biopolymer at (a) low composition and (b) at high composition.

**Scheme 1.** The best ionic transport illustration for (**a**) at low AC composition and (**b**) high AC composition.

### *3.4. Transference Number Analysis*

In this analysis, the biopolymer films were subjected to polarization currents (dc), which the current will become saturated indicating ionic transference, *tion* value [48]. See Figure S3 for testing setup. Figure 5 shows the plot of normalized current against time for sample AC7. From the figure, the initial total current decreases against time and eventually reached a steady state current. The value of *tion* for each sample was normalized before calculated using Equation (9) and the corresponding value is tabulated in Table 4.

$$t\_{i\text{inv}} = \frac{\left(I\_{initial} - I\_{standard}\right)}{I\_{initial}}\tag{9}$$

**Figure 5.** The plot of polarized current against time for AC7 electrolyte film.

The current become saturated due to ion accumulation at the blocking electrodes (increase resistance) while the remaining current flow is due to electron conduction. From the table, the sample with the highest ionic conductivity value (AC7) has the highest *tion* (0.98). This verified that the optimal CMC-AC biopolymer film system (AC7) is primarily ionic which is desirable for an intercalation process at the electrode [1]. However, the ionic conduction in the electrolyte system can be either due to cations or anions. The number of mobile ions, η calculated previously is actually the overall number for both ionic species and ionic mobility, μ and diffusion coefficient, *D* is the mean value for cations and anions [42]. Thus it can be differentiated to cationic and anionic mobility (μ<sup>+</sup> μ−) and cationic and anionic diffusion coefficient (*D*+, *D*−) and calculated using Equations (11) and (13), respectively [42,43]. Table 5 shows the results of the respective calculated cationic and anionic value obtained for CMC-AC biopolymer system where the value of μ<sup>+</sup> and *D*<sup>+</sup> is higher than μ<sup>−</sup> and *D*−. This indicates that in the current work the primary conducting species is cations, which further proved that the CMC-AC biopolymer film is also a proton (H+) conductor. The results obtained in this study is similar to the previous electrolyte system reported in the literature using another type of ammonium salt [49,50].

$$D = D\_{+} + D\_{-} = \frac{k\mathfrak{p}T\sigma}{\eta\mathfrak{e}^{2}}\tag{10}$$

$$t\_{\rm ion} = \frac{D\_+}{D\_+ + D\_-} \tag{11}$$

$$
\mu = \mu\_+ + \mu\_- = \frac{\sigma}{\eta \epsilon} \tag{12}
$$

$$t\_{\rm ion} = \frac{\mu\_{+}}{\mu\_{+} + \mu\_{-}} \tag{13}$$


**Table 5.** The transference value and respective percentage of cation/anion of CMC-AC biopolymer films.

### **4. Conclusions**

To conclude, the CMC-AC biopolymer film has been successfully prepared. Both CMC and AC salts dissolved completely and had low degree of crystallinity, *Xc* for the highest conducting sample (AC7) as proven by the XRD de-convolution result, while the AC salt appeared to have weak interaction with carboxyl group of CMC. The highest ionic conductivity value for the system was 7.71 <sup>×</sup> <sup>10</sup>−<sup>6</sup> Scm−<sup>1</sup> and thermally assisted (Arrhenius relations). The ionic conductivity of the current work was higher than other diammonium salts (ammonium adipate and ammonium sulfate) as presented in this work. It was also noted that the best composition (wt.%) of diammonium salt in a polymer electrolyte was lower compared to a single ammonium salt which was due to abundance of ammonium ions in a mol of ammonium salts. The transport number analysis showed that the CMC-AC biopolymer films were influenced by the ionic mobility, μ*,* and the diffusion coefficient, *D,* since both values achieved the maximum value for sample AC7, while the increased value of η for higher salt composition films led to lower μ and *D* values. The transference number also verified that the conducting species was mainly cation, which in this system was the proton (H+). The H<sup>+</sup> was able to transport across the polymer matrix through the high amorphous phase of the electrolyte films and subsequently increased the ionic conductivity. This shows a promising prospect for CMC-AC biopolymer film in electrochemical applications. However, further enhancement is still needed to increase the ionic conductivity for suitable electrochemical applications (~10−<sup>4</sup> Scm<sup>−</sup>1).

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2073-4360/12/11/2487/s1, Figure S1: The physical appearance of CMC-AC biopolymer films, Figure S2: The overall FTIR spectrum of CMC-AC electrolyte, Figure S3: Transference measurement testing setup.

**Author Contributions:** Conceptualization, M.I.N.M.I.; methodology, M.I.N.M.I. and M.I.H.A.S.; validation, M.I.N.M.I. and M.I.H.A.S.; formal analysis, M.I.N.M.I. and M.I.H.A.S.; investigation, M.I.H.A.S.; resources, M.I.H.A.S.; data curation, M.I.H.A.S.; writing—original draft preparation, M.I.N.M.I. and M.I.H.A.S.; writing—review and editing, M.I.N.M.I. and M.I.H.A.S.; visualization, M.I.H.A.S.; supervision, M.I.N.M.I.; project administration, M.I.N.M.I.; funding acquisition, M.I.N.M.I. All authors have read and agreed to the published version of the manuscript.

**Funding:** Fundamental Research Grant Scheme (FRGS) from Ministry of Education, VOT 59452 and Universiti Sains Islam Malaysia, Grant Scheme 166919 funded this research.

**Acknowledgments:** The authors would like to thank Universiti Sains Islam Malaysia for internal grant scheme (166919), Ministry of Higher Education for FRGS (59452) grants scheme. Special thanks to Advanced Materials team (AMt) members for consultations and assistances provided. The authors hope that everyone to stay strong, be safe and remain physical distance during this hard time dealing with COVID-19 pandemic. We will get through this very soon, Insha-Allah.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**

1. Shi, S.; Gao, J.; Liu, Y.; Zhao, Y.; Wu, Q.; Ju, W.; Ouyang, C.; Xiao, R. Multi-scale computation methods: Their applications in lithium-ion battery research and development. *Chin. Phys. B* **2016**, *25*, 018212. [CrossRef]


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Biocomposites Based on Plasticized Wheat Flours: E**ff**ect of Bran Content on Thermomechanical Behavior**

### **Franco Dominici 1, Francesca Luzi 1, Paolo Benincasa 2, Luigi Torre <sup>1</sup> and Debora Puglia 1,\***


Received: 31 August 2020; Accepted: 26 September 2020; Published: 29 September 2020

**Abstract:** In the present work, the effect of different bran content on the overall thermomechanical behavior of plasticized wheat flours (thermoplastic wheat flour; TPWF) was investigated. Refined flour (F0) with negligible bran fiber content, F1 flour (whole grain flour, 20% wt. bran), F3 (50% wt. bran) and F2 (F1:F3, 50:50) film samples were realized by extrusion process. The effect of TPWF blending with two different biopolymers (polycaprolactone and poly butyrate adipate terephthalate), combined with the presence of citric acid as compatibilizer was also considered. Results from FESEM analysis and tensile characterization demonstrated that PCL was able to reach improved compatibility with the plasticized flour fraction at intermediate bran content (F2 based formulation) when 25% wt. of biopolymeric phase was added. Additionally, it was proved that improvements can be achieved in both thermal and mechanical performance when higher shear rate (120 rpm) and low temperature profiles (Tset2 = 130–135–140 ◦C) are selected. Disintegrability of the TPWF basic formulations in compositing conditions within 21 days was also confirmed; at the same time, an absence of any phytotoxic event of compost itself was registered. The obtained results confirmed the suitability of these materials, realized by adding different bran contents, to mechanically compete with bioplastics obtained by using purified starches.

**Keywords:** bran content; plasticized wheat flour; citric acid; biobased blends

### **1. Introduction**

The increasing cost of petrol-based plastics and the public concern about their contribution to environmental pollution have raised the interest towards biobased and biodegradable materials, which help to dispose of by-products from agricultural production and food industries. In the case of bioplastics, purified starch from many agricultural sources (e.g., cereals, tubers, etc.) is often used as a basic ingredient. However, there is literature on the use of wheat flours to obtain bioplastics as an energetically and economically cheap alternative to purified starch [1,2]. Previous research from our group demonstrated that the tensile properties of thermoplastic films depended on wheat grain hardness and baking properties of refined flours [3,4]. However, the use of wholegrain flours has received limited attention, even if this approach could be of relevance for the reinforcement effect, due to bran, on the overall performance of plasticized starches [5]. Bran represents the outer portion of the grain, including the pericarp and seed teguments, containing a relevant amount of lignin and cellulose, the so-called fiber. It accounts for around 15–25% wt. of the total grain weight and generally comes out as a by-product of grain milling. It is normally used as animal feed [6]; however, the progressive decrease of the whole national livestock that occurred in the last decade has led to an increase of bran

stocks to get rid of, with possible valorization as biological chemicals and energy source [7]. The effect of bran particle size on functionality of the gluten network was already explored in wholegrain flour and its baking properties [8]: It was found that a deleterious effect on time of dough development, gluten strength, starch gelatinization and the retrogradation was intensified by the presence of all constituents of the grain in the wheat mass formulation when compared to refined flour. In particular, it was evidenced that the quality of the protein and the differences between the particle sizes with respect to the stability and development time are broadly correlated with the quality of the gluten network. Additionally, it was proven that the presence of fibers limited the availability of water to the starch in the wholegrain samples, and that this effect was especially strong for flour with finer particle size, which also had the highest rate of absorption.

In another paper, Liu et al. [9] studied the adverse effects of wheat bran on gluten network formation, which may lead to the reduction in gluten viscoelasticity and quality deterioration of fiber enriched flour products. With the properties of starch, such as degree of gelatinization, gel stability, and retrogradation, being strongly influenced by the availability of water in the formed mass system [10], it is considered extremely important to investigate the role of bran content and its grinding level even on the thermomechanical behavior of plasticized flours.

Since it is expected that the bran level could, in general, increase the strength of the matrix to the expense of the deformability of the plasticized flour [11] by creating macroscopic defects in the material, the possibility of using biobased polymers in combination with thermoplastic flour to recover the plasticity has been also considered. The literature reports results on the effect of fiber reinforcement on mechanical behavior of thermoplastic starch blended with different polyesters [12], but no examples are available on how the presence of bran additive could tune the mechanical behavior; according to this, we here attempted to verify, for the first time, how different contents of grinded bran could affect the deformability of the flour; furthermore, blending with low melting polymeric fractions (polycaprolactone and polybutylene adipate terephthalate) was considered to increase the limited deformability of the polymeric matrix.

#### **2. Materials and Methods**

### *2.1. Materials*

Soft wheat produced in the Umbria region (Italy) was chosen as the reference wheat variety to obtain, by grinding and selective extraction, flours with different bran contents. The milling products were kindly supplied by Molini Spigadoro (Bastia Umbra, Italy). The chemicals, glycerol, magnesium stearate, D-sorbitol, water, polyvinyl alcohol (≥99% hydrolyzed) (PVA) and citric acid (CA) were supplied by Sigma Aldrich. Polybutylene adipate terephthalate (PBAT) Ecoflex F Blend C1200 was supplied by BASF. Polycaprolactone (PCL) Capa 6500 was kindly provided by Perstorp.

#### *2.2. Preparation of Milled Products, Their Plasticization and Blending*

Four wheat flour milling products, namely F0, F1, F2, F3, were considered. The detailed compositions are reported in Table 1.

**Table 1.** Content of plasticizable fraction (PF) and fiber in% wt. of the samples.


\* FLOUR (PF/UF = 100/0); \*\* BRAN (PF/UF = 30/70).

F0 was a refined flour with negligible percent bran content; F1 was a wholegrain flour as it would come out from milling the whole grain (including the pericarp and seed coats); F3, with a bran content of 50% (wt.), represented the outcome generally obtained as the grinding tail of milling; F2 was obtained by mixing F1 and F3 in equal parts (50:50 wt.).

Plasticization of the samples was carried out with an Xplore Microcompounder 5 & 15 cc extruder (DSM, Sittard, The Netherlands) by considering suitable contents of glycerol, water and other process facilitating additives, as reported in our previous work [3]: Flour (68%, *w*/*w*), glycerol (23%, *w*/*w*), magnesium stearate (1.8%, *w*/*w*), sorbitol (5.2%, *w*/*w*), PVA in aqueous solution PVA/water 1:20 (2%, *w*/*w*). The initial process parameters were set as reported in our previous paper on refined flours: Temperature profile (Tset2) in the three heating zones of the extruder at 130–135–140 ◦C and mixing at 30 rounds per minute (rpm) for 6 min [4].

The doses of the reagents for the plasticization were adapted by calculating the plasticizable fraction of material (PF) (starch, proteins and other components), excluding the fiber and the other not plasticizable constituents (UF). Indeed, fiber content, which represents the fibrous portion of flour, does not participate to the plasticization process, while the bran content, which has plasticizable fraction, should be taken into account. The fiber, not participating directly in the plasticization reaction, was considered in the formulation only for the evaluation of absorbed water, estimated to be at 15% by weight of fiber [13]. In order to give an explanation of the adopted methodology for plasticization, a detailed recipe for the F1 sample is given as an example. F1 flour, as typical wholegrain flour, consists of 80% flour and of 20% bran. The plasticizable fraction (PF) of F1 is 86% wt. (given by 0.8 × 1.0 + 0.2 × 0.3, where 1.0 is the PF of the flour and 0.3 is the PF of the bran). In a similar way, PF values have been calculated for all the samples and summarized in Table 1.

In the case of blends based on TPWF, two types of biodegradable polymers, PBAT and PCL, were initially considered at a weight amount of 20% wt. In the case of polycaprolactone, the research was extended to blends containing 25, 30 and 40% wt. of the polymeric component. Further attempts to optimize the formulations were made by changing the quantities and types of plasticizers. The amount of glycerol was reduced from 23 to 17% wt. and, at the same time, an additional water fraction of 17% wt. was added to provide hydroxyl groups functional to plasticization, less available due to glycerol reduction. Moreover, the plasticizing and compatibilizing effect of citric acid added at 0.8% wt. was also evaluated. Then, an optimization of the parameters was tried by varying the temperature and mixing speed. The effects of increasing the temperature profile were tested by setting Tset3 to 135–140–145 ◦C. An attempt was also made increasing the rotation speed of the screws from 30 to 120 rpm. All the samples were used to produce specimens of film with a thickness of about 300 μm with the aid of a Film Device Machine (DSM, Sittard, The Netherlands) coupled to the extruder. The list of samples is summarized in Table 2.

### *2.3. Characterization of Flours and TPWF-Based Composites*

#### 2.3.1. Alveographic Properties

F0 flour was used for the measurement of alveographic parameters as it is a good approximation to the plasticizable fraction of the processed samples. The tests were carried out by using a Chopin alveograph (Alveolink NG, Villeneuve-la-Garenne, France) in constant hydration (HC) mode, following the recommendations of the ISO 27,971 standard. Average values of the main alveographic parameters, Tenacity (P), Extensibility (L), Baking strength (W), and Configuration ratio (P/L), were determined with five replicates.

#### 2.3.2. Thermogravimetric Analysis

Thermal degradation of the milling products F0, F1 and F3, having different content of bran, was evaluated carrying out thermal dynamic tests, from 30 ◦C to 600 ◦C at 10 ◦C min−<sup>1</sup> by thermogravimetric analysis (TGA, Seiko Exstar 6300, Tokyo, Japan). About 5 mg of each sample was used, and dynamic tests were performed under nitrogen flow (200 mL min<sup>−</sup>1). Mass loss (TG) and derivative mass loss (DTG) curves for each tested material were evaluated.


**Table 2.** Samples produced: Main constituents and processing parameters.

\* Tset 2 = 130–135–140 ◦C; 3 = 135–140–145 ◦C.

### 2.3.3. Tensile Tests

A universal electronic dynamometer LR30K Plus (LLOYD Instruments, Bognor Regis, UK) was used to carry out a mechanical characterization of the materials. Tensile tests were performed by setting a crosshead speed of 5 mm min−<sup>1</sup> on 20 <sup>×</sup> 150 mm rectangular specimens about 300 mm thick, in accordance with ISO 527 standards. Ultimate tensile strength (σ) and strain at break (εb) were calculated from the resulting stress–strain curves with the support of a software specific to the test machine: NEXYGEN Plus Materials Testing. The measurements were done, after conditioning the samples at room temperature for 24 h at 50% relative humidity (RH), testing at least five specimens for each formulation.

### 2.3.4. Morphological Evaluation

A first visual analysis was performed on TPWF/bran-based film samples. Moreover, a morphological characterization of composites was carried out using a field emission scanning electron microscope (FESEM) Supra 25 by Zeiss (Oberkochen, Germany). Micrographs of fractured surfaces obtained by cry fracturing the samples in liquid nitrogen were taken with an accelerating voltage of 5 kV at different magnifications. Previously, the samples were gold sputtered to provide electric conductivity.

#### 2.3.5. Disintegration in Compost

The compost mineralization of the films was evaluated on the basis of the ISO 20,200 standard. A certain amount of compost inoculum, supplied by Gesenu Spa, was mixed together with synthetic organic waste, prepared with an appropriate amount of sawdust, rabbit feed, starch, sugar, oil and urea to thus constitute the soil for composting. The soil moisture content was maintained at values of 50% RH by adding water and mixing at regular intervals of time, as indicated by the legislation, while aerobic and thermal conditions were guaranteed during the test. Based on ISO 20200, a sample can be considered disintegrated when it reaches 90% mass disintegration in at least 90 days in contact with the composting soil in the ripening phase. The disintegration percentage after a time *t* in compost is calculated as reported in Equation (1):

$$D\_t = \frac{m\_i - m\_r}{m\_i} \times 100\tag{1}$$

where *mi* is the initial mass of the sample and *mr* is the mass of the extracted sample, after drying, at a given time *t*.

#### 2.3.6. Evaluation of Phytotoxicity

The phytotoxicity of the compost obtained from the disintegration test of the films was assessed at 40 days from the start of composting and, following the obtained results, the evaluation was repeated at 60 days. A germination test was carried out on cress seeds (*Lepidum sativum* L.), a test plant normally used for this purpose, as required by the IPLA, DIVAPRA, ARPA methods "Compost Analysis Methods", 1998. This method involves the evaluation of the effect of an aqueous extract of compost, picked up from disintegration tests, on seed germination. It was decided to evaluate three composts, those obtained with plastic films derived from flours F0, F1 and F3, assuming that F2 would give an intermediate result between F1 and F3. For each compost, the following standard procedure was used. Each sample to be tested (200 g) was brought to a humidity of 85% and left for two hours in contact with the added water. It was then centrifuged at 6000 rpm for 15 min and the supernatant was filtered under pressure at 3.5 atm with a sterilizing membrane. The aqueous extract was diluted up to concentrations of 50% and 75%. Five aliquots, each of 1 mL, of each of the two dilutions of the obtained samples (plus the same number of controls with water) were placed in 9 cm diameter Petri dishes containing bibulous paper. 10 seeds of *Lepidium Sativum* were added to each capsule, soaked for one hour in distilled water. The capsules were placed to incubate at 27 ◦C for 24 h. After this period, the germinated seeds were counted, and the root length of the buds was measured. The germination index (*Ig*) was calculated as indicated in Equation (2):

$$I\_{\mathcal{S}}(\text{\textquotedblleft}\_{\text{\textquotedblleft}}) = \frac{(G\_{\text{c}} \times L\_{\text{c}})}{(G\_{\text{t}} \times L\_{\text{t}})} \times 100\tag{2}$$

where:

*Gc* = average number of germinated seeds in the sample *Gt* = average number of seeds germinated in the control *Lc* = average root length in the sample *Lt* = average root length in the control

The values of the germination indices for 50% and 75% dilutions after 40 and 60 days of maturation of the compost extracted from the disintegration soils of the samples TPF0, TPF1 and TPF3 were analyzed.

### 2.3.7. Statistical Analysis

Data were analysed by analysis of variance (ANOVA), using the Statgraphics Plus 5.1. Program (Manugistics Corp. Rockville, MD, USA). To differentiate samples, Fisher's least significant difference (LSD) was used at the 95% confidence level.

### **3. Results and Discussion**

### *Wheat Flour Characterization*

The refined flour F0, having a moisture content of 14.5% wt. and a protein amount of 11.8%, was tested with the Chopin's alveograph at constant hydration (HC), showing the following alveographic parameters: Tenacity (P = 64 mm H2O), extensibility (L = 99 mm), baking strength (W <sup>=</sup> <sup>182</sup> <sup>×</sup> <sup>10</sup>−<sup>4</sup> J), configuration ratio (P/L = 0.65), elasticity index (Ie = 47.4%). These alveographic parameters are characteristic of a standard flour with moderate strength and standard quality for basic baking uses.

The results of the thermogravimetric analysis (Figure 1) show that, with the exception of weight loss due to water evaporation at around 80 ◦C, there were no significant weight losses due to thermal degradation within the temperature range for plasticization up to 150–160 ◦C. As reported in [14], the shape of this low temperature peak can be varied due to bran addition: It was shown that flour rich mixtures exhibit distinct features with an initial peak attributed to starch and a secondary shoulder attributed to gluten. In general, they observed that a gradual shift occurred in the gluten shoulder, in conjunction with the addition of bran to the mixture. In our case, we observed that in bran rich mixtures, i.e., F3 formulation, the peak was not symmetric and a shift to a lower temperature range was observed due to a modified moisture release from the flour during heating.

**Figure 1.** Mass loss (TG) (**a**) and derivative mass loss (DTG) (**b**) curves of F0, F1 and F3 samples.

The thermogravimetric analysis evidenced also the typical degradative pattern for cereal flours. The three wheat flours showed similar TG curves, with small differences in terms of residual weight at the end of the test, in line with the typical mass loss values (17% *w*/*w* for wheat flour) and the additional residue due to the presence of the bran component (Figure 1a): After water evaporation, the second main step centered at 300 ◦C corresponds to the decomposition of starch, while the third step (T > 400 ◦C) corresponds to the formation of inert carbonaceous residues [1] (Figure 1b). In the case of increasing bran amount, the F1 and F3 samples showed the presence of a further peak, in the range 200–250 ◦C, identified as the starting point for decomposition of lignocellulosic components (mainly cellulose and hemicellulose) in the bran fraction [15].

Having established that the flours with variable bran content could be processed and plasticized in the selected temperature range without losing thermal stability, films based on F0, F1, F2 and F3 flours were realized by extrusion, as detailed in Section 2.2. The visual analysis of the films in the top row (Figure 2a), produced without the addition of biopolymers, showed a progressive browning as the fraction of bran in the flour increased. However, films with low fiber content showed good transparency, satisfying a fundamental characteristic for some packaging applications. The yellowing/browning of the TPWF films, which it is normally caused by the non-enzymatic reactions that occur during plasticization, is emphasized by the presence of the bran, which adds opacity and darkening [16–18].

**Figure 2.** Visual observation of (**a**) thermoplastic wheat flour (TPWF) based films made with flours with increasing bran content (from F0 to F3), (**b**) F2 based films blended with 20, 25, 30 and 40% wt. of PC and (**c**) films based on plasticized F0, F1, F2 and F3 blended with 25% wt. of polycaprolactone (PCL).

In Figure 2b, the effect of the addition of PCL in the formulation based on F2 flour produces an improvement in transparency, which was enhanced as the proportion of biopolymer in the blend increased. In the last line of the picture (Figure 2c), images of the films obtained after the optimization of compositions (25% wt. of PCL fraction) and processing parameters (120 rpm) are included. The film based on refined flour F0 shows good transparency, which remained acceptable even in the F1-based film, despite the yellowing due to the presence of bran fibers. The level of transparency of F2\_25CL120R is also acceptable, although the darkening caused by the abundant bran fraction produced a color change on browning tones and a sensible reduction in transparency [19]. In the case of F3-based film, the transparency was compromised to an extent that makes the film unsuitable for applications requiring visibility of the underlying objects, while its use for opaque packaging or other applications, such as mulching or shading sheets, where opacity is functional, can be envisaged for the F3\_25CL120R composition.

The morphologies of the fractured surfaces for the TPWF films were observed by FESEM (Figure 3) and differences were found for the four milling products with different bran contents. In detail, the F0 flour (Figure 3a) appeared well plasticized with a uniform surface, no separate starch granules were noted and the absence of bran particles was evident. Plasticization of the F1 flour (Figure 3b) was also well achieved, since a smooth and homogeneous plastic phase was found in the analyzed surface. Bran fibers with a particle/lamellar appearance were uniformly distributed and well bonded to the plasticized starch, suggesting the realization of a composite material with good characteristics. Similarly to the previous ones, the plasticization of the F2 flour was also performed with good results.

**Figure 3.** FESEM images of fractured cross-sections of (**a**) TPF0, (**b**) TPF1, (**c**) TPF2 and (**d**) TPF3 films, acquired at 1000× magnification.

In this case (Figure 3c), we noted the prevalent presence of lamellar particles of bran that resulted oriented, as alternating layers with the plastic phase, due to the orienting effect of the production process. F3, with the highest fiber fraction among the selected flours, highlighted the prevalent presence of bran particles (Figure 3d), with the plasticized starch having reduced adherence to the bran particles. Fibrous agglomerates and not well plasticized starch particles were noted: Due to the hindering effect of the large amount of bran fiber, wheat flour granules were less capable of forming hydrogen bonds with plasticizers through their hydroxyl groups, leaving some domains unreacted with unplasticized starch particles [3,20]. In general, while observed morphologies in TPF1 and TPF2 can confirm the ductility of the films, to the expense of low strength (that actually increased in TPF2 due to increased filler content), TPF3 appears saturated with the reinforcement phase and a matrix phase close to the wettability limit of the fibers. In this case, behavior that maximizes strength and stiffness but limiting the elastic-plastic characteristics can be expected.

The observations made by analyzing the sample morphologies were confirmed by checking the results of tensile characterization made on the same series of materials In Table 3, the results of the tensile tests carried out on plasticized flour samples and their bioblends are included.


**Table 3.** Results of tensile tests made on plasticized flours and their blends with PCL and polybutylene adipate terephthalate (PBAT), by varying the biopolymer amounts, processing temperatures and mixing rate.

\* Tset 2 = 130–135–140 ◦C; 3 = 135–140–145 ◦C.

The refined F0 flour, following the plasticization process, shows mechanical properties in line with other flours, with comparable alveographic characteristics, tested in previous works [3]. As evidenced in Figure 4a, good elongation values (54%) correspond to a moderate tensile strength (1.23 MPa), which is the main drawback of TPWFs. The selection of flours containing different bran fractions offers the advantage of having a fibrous filler, which can be effective as a reinforcement phase and, at the same time, has a plasticizable fraction, able to guarantee good compatibility and bonding at the interface with the starch matrix upon plasticization. The bran plays the role of reinforcement by preventing creep and deformation of the thermoplastic phase. As the bran fraction and consequently the fiber content increased, the samples showed decreasing strain values. TPF1 showed an ε<sup>b</sup> of 32.2%, which decreased to 23.8% with TPF2, further dropping to 19.6% in the case of the TPF3 sample. On the other hand, when the percentage of bran increased, the tensile strength increased as well, reaching an σ value more than tripled in the case of the F3-based sample (3.83 MPa) when compared to refined F0 flour.

The tensile strength of TPF1 was more than doubled (2.63 MPa) compared to the sample without fiber TPF0; TPF2 shows the same tensile strength value (2.62 MPa) as TPF1, albeit with a higher bran content. This result suggested the possibility of an improvement of the mechanical properties for reference TPF2, which could be achieved by improving the dispersibility of the bran fiber in the plasticized matrix. To pursue this goal, the characteristics of the matrix must be enhanced by improving the compatibility with the reinforcement phase. Comparing the values obtained for all formulations, it can be commented that values for the maximum tensile strength and elongation at break changed significantly (*p* < 0.05).

**Figure 4.** Results of tensile strength and strain at break for (**a**) TPWF/bran film samples, (**b**) TPWF/bran and biopolymeric blends, (**c**) samples processed with two different plasticization temperature profiles (T2, T3), (**d**) samples processed with two different mixing rates (30 rpm, 120 rpm). (a–d, stress values) (1–5, strain values) Different superscripts within the same column group (stress or strain values) indicate significant differences among formulations (*p* < 0.05).

In order to further improve the characteristics of the F2 based TPWF, two actions were considered: The improvement of the intrinsic characteristics of the composite by using plasticizers/compatibilizers and the addition of another matrix in blend that could enhance the characteristics of the entire composite system. The selection of the matrices was carried out taking into account some preliminary criteria, such as physical–chemical compatibility, conservation of the biodegradability of materials and process compatibility, possible plasticization and blending in one step to minimize energy waste and environmental impact, according to an eco-sustainable development perspective. Polybutylene adipate terephthalate (PBAT) and polycaprolactone (PCL) were selected for this specific purpose and initially used at a nominal percentage of 20% wt. Citric acid (CA) was indeed considered as a suitable compatibilizer for TPWF: CA, other than having a plasticizing effect, can be also effective in the compatibilization between plasticized starch, bran fiber and biopolymers [21,22]. A nominal percentage of 0.8% wt. was chosen, on the basis on the results of previous literature works [23]. The effects on the composites were firstly evaluated by adding individually CA, PBAT and PCL, and then the formulations with the concurrent use of biopolymer and compatibilizer were studied.

In Figure 4b, it is demonstrated that the citric acid, added alone to the TPF2 formulation, has a plasticizing effect, increasing the deformation to 34.9% but reducing the strength to 1.61 MPa. The TPF2\_20BAT blend showed an increase in tensile strength, beside a reduction in deformability, which highlighted the poor compatibility between TPWF and PBAT. A decrease of both values of strength and deformation at break was also found in the TPF2\_20CL formulation, with only PCL in blend. The use of citric acid was found to have no positive effects when added in the presence of PBAT,

while its role was effective when combined with PCL. TPF2\_20CL showed an increase in strain at break up to 60.4% compared to the value of 16.8% for the same sample without CA. The addition of citric acid induced reactions able to favor the adhesion between PCL and TPS (trans-esterification), improving the wettability (hydrolysis) and inhibiting the formation of cross-linking (sulfhydryl (SH)-SS exchange) during flour plasticization [21,22,24–26].

Considering that glycerol has a much higher plasticizing effect, due to the presence of three hydroxyl groups, compared to water, and even a "lubricating" effect that lowers the stiffness of the plasticized system, it was planned to replace 6% wt. of glycerol with 17% wt. of water [27]. Furthermore, an attempt was made to improve the mechanical performance of the produced films by varying the plasticization temperature from T2 to T3 [4,28]. The increase of the temperature had the purpose of improving the tensile stress resistance of the materials, by intensifying the formation of bonds and crosslinking, typical of the plasticizing process, conferring strength and rigidity to the TPWF. To better understand the effect of these variations on mechanical properties of TPWF-based samples, both the samples of refined flour F0 and those of F2 flour were tested. The three pairs of samples processed at T2 and T3 (Figure 4c) showed that the increase in temperature, in the presence of water and CA, generally causes a worsening of the mechanical properties, by lowering the stress and strain values. At higher temperatures, the hydrolytic phenomena induced by CA prevailed over the effects of transesterification and cross-linking, supported by the kinetics of the plasticization reaction of the flours at T3. It should be noted that the new dosage with the partial replacement of glycerol with water produced a notable increase in strength (+74%), that moved from 1.66 MPa of TPF2\_CA20CL to 2.89 MPa of F2\_CA20CL as expected.

A further attempt to optimize the formulations was made by increasing the fraction of PCL in the blend to evaluate the ideal TPWF/PCL ratio (Figure 4d). Furthermore, the effect of shear stresses during plasticization was evaluated by processing samples at 30 and 120 rpm. It is known that the shear stresses applied during the plasticization phase can produce effects on the destruction of starch granules and, consequently, on the mechanical characteristics of the materials. In order to take into account the effect of the rheological characteristics of the system, the tests were repeated for materials with different PCL fractions [29,30]. A higher speed of rotation of the screws during plasticization in the extruder produced a general improvement of the mechanical properties, both in terms of strength and, albeit to a lesser extent, of strain. In particular, the increase of shear stresses raised the strength (+23%) from 3.0 to 3.7 MPa in the sample F2\_25CL120R. In Figure 5, the SEM micrographs of the samples processed at different screw speeds show a completely different morphology; higher shear stresses, produced with 120 rpm of screw speed, improved plasticization (Figure 5b); resulting images showed smoother and more uniform fracture surfaces, free of granules and fibrous conglomerates, with uniform distribution of the separate phases of PCL and TPWF.

**Figure 5.** FESEM morphologies of F2\_25CL30R (**a**) and F2\_25CL120R (**b**) samples.

Finally, samples of all flours were produced using the optimized formulation and process parameters. In Figure 6, the progressive improvement in tensile strength is closely related to the increase in the fraction of bran fiber. The lowest σ value (2.68 MPa) was obtained for F0\_25CL120R, which was free of fiber, and rose to 3.91 MPa with F3\_25CL120R. The intermediate fiber contents also corresponded to intermediate strength values for F1\_25CL120R and F2\_25CL120R, equal to 3.58 and 3.70 MPa, respectively. On the other hand, the increase in the bran fraction produced a progressive decrease in the strain at break, which reached 57.1% with F0 flour and dropped to 10.0% for the F3 flour-based material. Samples on F1 and F2, flour-based, showed intermediate ε<sup>b</sup> values of 37.4% and 18.9%, respectively. A clear effect of the bran fiber fraction on the mechanical properties of the TPWF-based composites was noted: The addition of fiber produced a 45% increase in strength but caused a drop to 18% of maximum elongation reached without fiber. This result suggests that the composite F3\_25CL120R has reached the fiber saturation (35% wt.) and a further increase of the reinforcement phase would lead to brittle behavior of the material. Materials made with quantities of fibers between the tested extremes show intermediate values of σ and ε<sup>b</sup> indicating the possibility of designing the mechanical properties of the material to be produced according to the formulation.

**Figure 6.** Results of tensile strength and strain at break for TPWF/PCL blends based on F0, F1 and F3 flours, containing 25% wt. of PCL and processed at 120 rpm. (a–c, stress values) (1–4, strain values) Different superscripts within the same column group (stress or strain values) indicate significant differences among formulations (*p* < 0.05).

The thermal stability of the optimized formulations was determined by thermogravimetric analysis. Figure 7 presents the thermal degradation profile (TG/DTG curves) of the TPWF/PCL blends based on F0, F1 and F3 flours, containing 25% wt. of PCL and processed at 120 rpm. Thermal degradation of blends presented four mass loss stages (Figure 7). Up to approximately 130 ◦C, there is a mass loss due to the presence of water, while following weight loss, observed between 130 and 230 ◦C, can be related to the evaporation of glycerol and other volatile compounds present in TPWF [31]. Then, the starch chains began to degrade at about 230 ◦C [32]; after that, the fourth stage of thermal degradation of the blends occurred from 350 to 430 ◦C, due to the degradation of PCL chains. The main differences in these profiles was found for the signal of the plasticized TPWF; while the maximum degradation rate of the polymeric PCL phase was almost constant in intensity for all the three blends, the second main weight loss accounted for the reduced amount of plasticized fraction. It essentially followed the trend that TPWF with more bran content (F3 based blend) showed reduced degradation rates. The increased amount of bran was also responsible for increased value of remaining mass at the end of the test, as observed in Figure 7b, due to the charred fraction of fiber.

Figure 8a shows the visual images of the samples during the progress of the disintegration process under composting conditions, while Figure 8b shows the trend of the mass disintegration rate in compost for the three tested formulations. All materials reach 90% disintegration after 15 days under composting conditions. The TPF3 film showed different disintegration kinetics, presenting lower disintegration values, in comparison with TPF0 and TPF1 films, between the 2nd and 4th day under composting conditions. This behavior can be justified considering that higher fiber content was present in F3 film, which slowed down the decomposition process of the plasticized fraction. Starting from the 10th day of the test, both degradation kinetics and final disintegration degree of the three systems were aligned and samples completely disintegrated within 21 days, confirming the compostability, at lab conditions, of the studied materials.

**Figure 7.** TG (**a**) and DTG (**b**) curves for TPWF/PCL blends based on F0, F1 and F3 flours, containing 25% wt. of PCL and processed at 120 rpm.

**Figure 8.** Visual images of the TPWF samples based on F0, F1 and F3 flours (**a**) and evolution of disintegration degree (**b**) at different compositing times.

The results obtained 40 days after the start of composting (Table 4) indicate an effect of both the type of compost and the concentration of the extract. All the composts tested were found to have a depressing effect on the germination and growth of watercress sprouts, as the germination index, Ig (%), which was always less than 70%, considered the minimum acceptable value. For all compost, a lower germination index corresponded to a higher concentration of the extract. Among the compost, the sample derived from refined flour (F0) is the one that gave the greatest phytotoxic effect, while the compost obtained from both plasticized flours containing bran contents (TPF1 and TPF3) were found to have a less depressing effect on germination performance, but still more toxic than desired. According to this, it was assumed that the revealed phytotoxicity was due to the incomplete maturation of the compost. For this reason, the germination test was repeated with compost extract taken 60 days after composting. In these conditions, it can be seen that, at 60 days, all compost allowed an acceptable germination index (i.e., >70%). In particular, no compost inhibited germination, which was always close to 100% even at the highest extract concentration, while root growth on compost extract obtained from refined flour (F0) was slightly reduced, but always within acceptable limits. On the other hand, there was a kind of hormetic or stimulating effect of the compost extract obtained from F1 flour when used at the lowest concentration (50% dilution); in this case the Ig (%) was 108%. This result is not surprising because it is known in the literature that various substances, both synthetic and natural (e.g., NaCl and other salts, herbicides and allelopathic substances), can have a depressive effect at high concentrations or a stimulating effect at low concentrations.

**Table 4.** Germination test results on compost extract taken 40 and 60 days after the incubation.


### **4. Conclusions**

The objective of this work was the study of thermomechanical behavior of eco-sustainable and biodegradable materials obtained by plasticizing wheat-milling products containing fractions of bran fiber as filler/reinforcement. Four flours, with different contents of bran fraction, were obtained by sampling along the wheat milling line. The standard alveographic characteristics of reference refined flour allowed the production of film samples, plasticized in the extruder, both with the refined flour F0 and with the milling products F1, F2 and F3 with fiber content of about 15, 25 and 35% wt. The TPWF/bran fiber composites proved to have acceptable mechanical characteristics, which can be improved by the use of suitable quantities of PCL in blend, with citric acid as compatibilizer and with the partial replacement of glycerol with water. Process parameter optimization tests have shown that the lowest plasticization temperature profile (T2) and the highest mixing rate (R120) produced materials with better mechanical properties. In light of the obtained results, we concluded that it is possible to design formulations and manage the process parameters to obtain eco-sustainable and compostable materials from plasticization of raw wheat flour/bran fiber reinforced, at affordable costs, with characteristics designed for different application sectors requiring different mechanical performance.

**Author Contributions:** Conceptualization, P.B., F.D., D.P.; investigation, F.D., F.L., D.P.; writing—original draft preparation, P.B., F.D., D.P.; writing—review and editing, P.B., D.P., F.D., L.T.; supervision, L.T.; funding acquisition, D.P. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*

## **Encapsulation E**ff**ect on the In Vitro Bioaccessibility of Sacha Inchi Oil (***Plukenetia volubilis* **L.) by Soft Capsules Composed of Gelatin and Cactus Mucilage Biopolymers**

### **María Carolina Otálora 1,\*, Robinson Camelo 2, Andrea Wilches-Torres 1, Agobardo Cárdenas-Chaparro <sup>2</sup> and Jovanny A. Gómez Castaño 2,\***


Received: 30 July 2020; Accepted: 23 August 2020; Published: 2 September 2020

**Abstract:** Sacha inchi (*Plukenetia volubilis* L.) seed oil is a rich source of polyunsaturated fatty acids (PUFAs) that are beneficial for human health, whose nutritional efficacy is limited because of its low water solubility and labile bioaccessibility (compositional integrity). In this work, the encapsulation effect, using blended softgels of gelatin (G) and cactus mucilage (CM) biopolymers, on the PUFAs' bioaccessibility of *P. volubilis* seed oil was evaluated during in vitro simulated digestive processes (mouth, gastric, and intestinal). Gas chromatography–mass spectrometry (GC–MS) and gas chromatography with a flame ionization detector (GC–FID) were used for determining the chemical composition of *P. volubilis* seed oil both before and after in vitro digestion. The most abundant compounds in the undigested samples were α-linolenic, linoleic, and oleic acids with 59.23, 33.46, and 0.57 (g/100 g), respectively. The bioaccessibility of α-linolenic, linoleic, and oleic acid was found to be 1.70%, 1.46%, and 35.8%, respectively, along with the presence of some oxidation products. G/CM soft capsules are capable of limiting the in vitro bioaccessibility of PUFAs because of the low mucilage ratio in their matrix, which influences the enzymatic hydrolysis of gelatin, thus increasing the release of the polyunsaturated content during the simulated digestion.

**Keywords:** softgels; mucilage; biopolymers; in vitro digestion; bioaccessibility

### **1. Introduction**

Sacha inchi (*Plukenetia volubilis* L.) is a plant belonging to the Euphorbiaceae family that grows in the Amazon rainforest in Northeastern Peru and Northwestern Brazil. It has aroused interest in the food industry because of the high quantity and quality of the edible oils contained in its seeds. Its fruit consists of a star-shaped capsule that is approximately 3–5 cm in size, and normally contains between four and six dark brown edible oval seeds that are 1.5–2 cm in size [1]. The seeds of this plant are an excellent source of edible oil (41–54%). This oil contains lipids (35–60%), free fatty acids (1.2%), and phospholipids (0.8%) [2]. The high nutritional value of sacha inchi oil is due to its high polyunsaturated fatty acid (PUFA) and monounsaturated fatty acid (MUFA) content, which varies between 77.5% and 84.4%, and 8.4% and 13.2%, respectively [2–4]. α-Linolenic acid (ALA; C18:3, ω-3) is the major fatty acid (47–51%), followed by linoleic acid (LA; C18:2, ω-6, 34–37%) and oleic acid (ω-9, 9–10%) [2]. These fatty acids are considered beneficial because of their antioxidant, antithrombotic, antidyslipidemic, and anticancer effects [5–7]. Although PUFAs are beneficial, they are very sensitive

to oxidative damage when exposed to oxygen, and are affected by heat, light, and humidity [8]. Therefore, it is essential to establish adequate systems for the transport and encapsulation of fatty acids, while maintaining their nutritional properties until they are released within the body.

Encapsulation with hydrocolloid biopolymers is an effective and widely used technique in the food industry in order to protect dietary supplements against oxidation and loss of nutritional value. Its effectiveness is as a result of the hermeticity of its walls and the safe supply of bioactive compounds, due to its rapid disintegration in biological fluids at body temperature [9,10]. This characteristic makes soft capsules (or softgels) one of the most widely used encapsulation methods because of their safe and nutritionally-accepted means of delivering aqueous liquid or semi-solid dosage formulations. Although there are several studies on the encapsulation of omega-3 fatty acids using different techniques [11–14], the use of soft capsules for the encapsulation of polyunsaturated fatty acids in scientific studies is scarce [15,16].

Softgels are generally manufactured with animal-derived gelatin (G), water, non-volatile plasticizers, and minor additives such as opacifiers and dyes [17,18]. The choice of G as a traditional material in the walls of softgels is related to its biodegradable nature and its ability to form thermo-responsive hydrogels [19,20]. An emerging trend in the food industry has recently sought to partially or completely replace G with other non-animal natural hydrocolloids. This has allowed plant mucilage to emerge as an attractive structural biopolymer alternative to soft capsules [21]. The mucilage extracted from the cladodes of *Opuntia ficus-indica* is a heteropolysaccharide matrix that has proven to be a suitable natural structuring material in soft capsules because of its high fiber content and desirable functional properties [22–24].

Bioaccessibility is defined as the amount of a bioactive compound that can cross the intestinal barrier as a result of its release from the matrix by the action of digestive enzymes [25]. Therefore, the stability of the encapsulated polyunsaturated fatty acids that are released into the gastrointestinal tract and are available for intestinal absorption is directly related to the biocompatibility and biodegradability rates of the capsule matrix [26]. As a result, the behavior of the capsule structure in relation to changes in pH and the presence of digestive enzymes and bile salts is an important factor that should be properly evaluated. This makes in vitro methods that simulate gastrointestinal media particularly advantageous for the quantification of the biocompatibility and biodegradability of soft capsules containing such nutrients. The advantages of in vitro methods include speed, cost efficiency, and a lack of ethical restrictions when compared to in vivo methods [27]. The use of in vitro conditions provides the opportunity to evaluate the suitability of a matrix's ability to carry functional compounds of interest for the food and pharmaceutical industries. The bioaccessibility of polyunsaturated fatty acids has been evaluated using in vitro models in different studies [28,29].

The primary interest of this study was to determine the effect of gelatin/cactus mucilage as a softgel wall material in the in vitro digestion of encapsulated sacha inchi oil (SIO). The bioaccessibility analysis was performed using a comparative study that involved the quantification and identification of bioactive compounds using gas chromatography with a mass spectrometry (MS) detector and a flame ionization (FID) detector. The quantification was performed both before and after different laboratory-scale digestion processes. To the best of our knowledge, this study is the first evaluation of soft capsule wall matrices in the bioaccessibility of edible SIOs, which may be a topic of relevant interest to the food supplement industry.

### **2. Materials and Methods**

### *2.1. Materials and Reagent*

Sacha inchi fruits were supplied by local farmers in Miraflores (Boyacá, Colombia). The seeds were selected manually by discarding those that presented physical damage. They were packaged in polyethylene bags and stored at 18 ◦C until use. Cactus mucilage (CM) was extracted from the cladodes of *Opuntia ficus-indica* provided by local farmers in Duitama (Boyacá, Colombia), following the methodology reported by Quinzio et al. [30], and were used without further purification. Food grade gelatin (Type B, bloom strength 285, pI 4.2–6.5, Mw 40,000–50,000 Da, and ~99% purity) was provided by Gelco (Medellín, Antioquia, Colombia). Glycerol was provided by Merck (Darmstadt, Germany) for use as a plasticizer. Commercial samples of digestive enzymes and bile salts (α-amylase, pepsin from porcine gastric mucosa, and pancreatin from porcine pancreas) were obtained from Sigma–Aldrich (Auckland, New Zealand).

### *2.2. Sacha Inchi Oil (SIO) Extraction*

The SIO extraction was performed using the Soxhlet methodology with chloroform as a solvent. Soxhlet extraction was selected as the most conventional, economical, and easiest process to implement. To maximize the SIO extraction, approximately 15 g of seed material was ground in an analytical mill (IKA A11 basic S1). Then, 1 g of seed sample was placed in a cellulose cartridge (33 × 80 mm) for extraction using a Soxhlet extractor SER 148/3 Velp Scientifica (Usmate Velate, Italy) for 1.5 h with 60 mL of solvent.

### *2.3. Characterization Sacha Inchi Oil*

### 2.3.1. Fatty Acid Identification with Gas Chromatography Coupled to Mass Spectrometry (GC–MS)

The chemical composition of the SIO sample was determined by gas chromatography coupled to mass spectrometry (GC/MS) using a 6890 N GC–MS instrument (Agilent Technologies Inc., Palo Alto, CA, USA) coupled to an Agilent 5973 N inert mass selective (IMS) detector. A capillary column DB-1MS (30 m × 0.25 mm ID with 0.25 μm film thickness) was employed for the analysis. The elution program started at a temperature of 70 ◦C, held for 2 min, and then increased to 320 ◦C at a speed of 8 ◦C/min and was held for 29 min. Both the IMS detector and the injector port temperatures were 320 and 250 ◦C, respectively. The injection volume used was 3.0 μL, and helium was used as the carrier gas at a flow rate of 1 mL/min in a splitless mode. The components were identified using a commercial library higher than 85% (WileyW9N08, Mass Spectral Database of the National Institute of Standards and Technology (NIST)).

### 2.3.2. Fatty Acid Profile with Gas Chromatography with a Flame Ionization Detector (GC–FID)

The SIO sample was also analyzed using a 6890 N GC–FID instrument (Agilent Technologies Inc.). A capillary column DB-225 (60 m × 0.25 mm ID with 0.25 μm film thickness) was employed for the analysis. The elution program started at a temperature of 75 ◦C and increased to 220 ◦C at a speed of 5 ◦C/min for 50 min. Both the detector and injector port temperatures were 220 and 250 ◦C, respectively. The injection volume used was 0.2 μL, and helium was used as the carrier gas at a flow rate of 1 mL/min in a 100:1 split mode.

### *2.4. Soft Capsule Preparation with SIO Inclusion*

The soft capsule design was developed according to the method reported by Camelo et al. [21]. CM (1.0 g) and food grade G (21.0 g) were separately dissolved in 100 mL of distilled water at 18 ◦C and 40 ◦C for 2 h and 30 min, respectively. Both solutions were constantly stirred at 300 rpm using a magnetic stirrer (C-MAG HS 7S000, IKA, Staufen im Breisgau, Germany) in order to ensure complete solubilization. Afterwards, the gelatin solution was mixed separately with glycerol (Gly) at a concentration of 15% (*w*/*v*) at 60 ◦C for 2 h to remove residual air bubbles and to obtain a homogeneous solution. The ratio of G/CM (3:1 *w*/*w*) was homogenized at room temperature for 1 h under constant magnetic stirring. The biopolymer solution was poured into an elliptically-shaped mold (22 mm length and 11 mm diameter) and dried in a Memmert UM 400 drying oven (Schwabach, Germany) at 25 ◦C with a relative humidity of 40% for 1 h. Subsequently, 2 mL of sacha inchi oil was injected into the formed soft capsule. The syringe hole in the capsule was then sealed by carefully applying heat using a small hot spatula.

### *2.5. Fatty Acids' Polyunsaturated Bioaccessibility*

The encapsulated oil was submitted to an in vitro digestion process using mouth, gastric, and intestinal simulation. This method was implemented as described by Pacheco et al. [31]. To simulate mouth digestion, 5 g of soft capsules were weighed and added to 9 mL of simulated saliva (1.59 mM CaCl2, 21.1 mM KCl, 14.4 mM NaHCO3, 0.2 mM MgCl2, pH adjusted to 7.0 using 1.0 N HCl, and the α-amylase enzyme). This mixture was incubated in a Schutzart DIN 60529-IP 20 shaking water bath (Memmert, Germany) for 5 min at 37 ◦C, and agitated at 185 rpm. Gastric digestion was then initiated by adding 36 mL of pepsin solution (25 mg/mL in 0.02 N HCl) to the samples. The mixture's pH was adjusted to 2.0 using 1.0 N HCl, and was incubated while continuously shaking at 130 rpm at 37 ◦C for 60 min. To simulate intestinal digestion, the gastric-digested mixture's pH was adjusted to 6.0 using 1 M NaHCO3. Afterwards, 0.25 mL of pancreatin solution (2 g/L) and biliary salts (12 g/L) dissolved in aqueous 0.1 M NaHCO3 were added and then the mixture was incubated at 37 ◦C for 120 min while constantly stirring at 45 rpm.

After intestinal digestion, the samples were immediately centrifuged at 5000 rpm for 10 min. The centrifuged samples were separated into two layers: an opaque sediment layer at the bottom, and a thin oily layer at the top. The oily layer was centrifuged again at 4000 rpm for 3 min and filtered using a Millipore membrane (0.45 μm), and then analyzed for its polyunsaturated fatty acid content and composition using GC–FID and GC–MS. The percentage of bioaccessibility was calculated using this equation:

$$\text{Biocessibility} \left( \% \right) = \frac{\left( \text{Content of factory acids polymerated present in the lightest product} \right)}{\left( \text{Content of factory acids polymerated present in the encapsulated matrix} \right)} \times 100 \quad (1)$$

### **3. Results**

### *3.1. Characterization of Sacha Inchi Oil*

### 3.1.1. Fatty Acid Identification with Gas Chromatography Coupled to Mass Spectrometry (GC–MS)

As shown in the GC–MS chromatograms in Figure 1, the fatty acid profiles revealed 13 main chemical constituents in the SIO sample prior to encapsulation. The peaks numbered 1, 2, and 3 in the oil, eluted at retention times of 31.34, 33.15, and 33.19 min, were identified as hexadecanoic acid methyl ester, 9,12–octadecadienoic acid methyl ester, and (Z,Z,Z)-9,12,15-octadecatrienoic acid methyl ester, respectively. This demonstrates that the oil is a rich source of polyunsaturated fatty acid. Peak number 4 in the chromatograms, eluted at a retention time of 33.52 min, was identified as (Z)-11-octadecenoic acid methyl ester and octadecenoic acid methyl ester. The peaks numbered 5, 6, 7, 8, and 9 were observed in the oil eluted at times of 33.52, 36.91, 37.68, 38.41, and 39.11 min, respectively, and were identified as eicosane. In the sample, the compounds hexadecane (peak number 10), 1,4-phthalazinedione 2,3 dihydro-6-nitro (peak number 11), cyclotrisiloxane hexamethyl (peak number 12), and 5-methyl-2-phenylindolizine (peak number 13) were eluted at retention times of 39.79, 40.45, 41.08, and 41.74 min, respectively.

**Figure 1.** GC–MS chromatograms of sacha inchi oil before the simulated gastrointestinal digestion.

#### 3.1.2. Fatty Acid Profile with Gas Chromatography with a Flame Ionization Detector (GC–FID)

The fatty acid contents analyzed by GC–FID in the unencapsuslated SIO sample (i.e., prior to encapsulation) are shown in Table 1. The presence of α-linolenic (C18:3 ω-3), linoleic (C18:2 ω-6), palmitic (C16:0), stearic (C18:0), and oleic (C18:1 ω-9) acids, in decreasing order of abundance, indicated a rich source of polyunsaturated fatty acids. These values agreed with the results of Chirinos et al. [3] and Gutiérrez et al. [32]. The content of ω-3 (59.23 g/100 g) is important and desirable because of its contribution to preventing several diseases such as obesity, diabetes, allergies, Alzheimer's, and coronary and neurodegenerative diseases [33]. The ω-6/ω-3 ratio of 0.56 has many health and nutritional benefits, such as the reduction of chronic diseases, and in cardiovascular and hypertension disease prevention [34,35]. The low ratio of linoleic/α-linolenic is similar to the results obtained in both irradiated and non-irradiated sacha inchi oils [32].

**Table 1.** Fatty acid composition (g/100 g) of sacha inchi oil before the simulated gastrointestinal digestion.


#### *3.2. Chemical Content of Encapsulated SIO after Simulated In-Vitro Digestion*

In order to study the impact of soft capsules on the bioaccessibility of the fatty acids, the encapsulated sacha inchi oil was subjected to in vitro digestion (i.e., mouth–gastric–intestinal media simulation), and its chemical content was analyzed using GC–MS and GC–FID spectrometry.

3.2.1. Products Derived from the Degradation of Polyunsaturated Fatty Acids after In Vitro Digestion

The chromatogram profile of the main SIO compounds and their degradation by-products using GC–MS showed a maximum of 17 chemical constituents, as displayed in Figure 2. No peak attributable to 9,12–octadecadienoic acid methyl ester was detected in the chromatograms of the encapsulated oil after digestion, and therefore its bioaccessibility was considered negligible. Conversely, the presence of (Z,Z,Z)-9,12,15-octadecatrienoic acid methyl ester was detected before and after in vitro digestion in peaks number 3 and 13, which were eluted at a retention time of 33.20 and 32.06 min, respectively. The presence of this fatty acid was detected in the chromatograms of the encapsulated oil after

digestion in a 43.96% peak area, making its bioaccessibility considerably significant. The presence of peaks with different retention times was also evidenced. For instance, 10,13-octadecadienoic acid methyl ester was identified as a degradation by-product of 9,12–octadecadienoic acid methyl ester [36], along with derivatives such as propanal, alcoholic amines, and aromatic compounds. All were related as by-products of simulated gastrointestinal digestion, demonstrating a low bioaccessibility of the encapsulated oil.

**Figure 2.** GC–MS chromatograms of sacha inchi oil after the simulated gastrointestinal digestion. Abbreviations: (**1**) Ethanol 2-(2-methoxyethoxy); (**2**) 3-[2-Diethylaminoethyl]-2,4-pentanedione; (**3**) Diethyl carbamoyl *t*-butoxy sulfide; (**4**) Ethane 1,2-bis(methylthio); (**5**) Ethane 1,2-bis(methylthio); (**6**) Cyclotetrasiloxane octamethyl; (**7**) Cycloserine; (**8**) 1-Propanol 2-amino; (**9**) 2-Pentanamine *N*-(1-methylbutyl); (**10**) *n*-Hexylmethylamine; (**11**) Methylpent-4-enylamine; (**12**) 10,13-Octadecadienoic acid methyl ester; (**13**) (Z,Z,Z)-9,12,15-Octadecatrienoic acid, methyl ester; (**14**) Epinephrine; (**15**) 1-Octanamine *N*-methyl; (**16**) 2-Amino-1-(*o*-hydroxyphenyl)propane; (**17**) Tetrasiloxane, decamethyl.

### 3.2.2. Fatty Acid Content under the Simulated In Vitro Digestion

The relative concentration of the encapsulated fatty acids before and after the simulated gastrointestinal digestion, determined using GC–FID spectrometry, is shown in Figure 3. It was found that the bioaccessibility of the α-linolenic, linoleic, and oleic polyunsaturated fatty acids was 1.70%, 1.46%, and 35.8% respectively, while the saturated stearic and palmitic acids presented bioaccessibility values of 2.26% and 1.72%, respectively.

**Figure 3.** Relative concentration of fatty acids in sacha inchi oil (SIO) samples before and after the simulated gastrointestinal digestion.

### **4. Discussion**

As seen in Figure 3, during the digestion of the SIO encapsulated in G/CM softgels, there was a significant decrease in the content of the two most abundant PUFAs (i.e., α-linolenic and linoleic acid content), which entailed a reduction in the nutritional and functional value of this natural oil. As can be inferred by comparing the GC–MS chromatograms of the SIO before and after in vitro digestion (Figures 2 and 3), the amounts of unsaturated fatty acids originally present in the SIO stimulated a higher generation of oxidation by-products during digestion [28]. Similar results were reported by Nieva-Echevarría, Goicoechea, and Guillén [37] during the in vitro gastrointestinal digestion of flaxseed oil. The amount of non-encapsulated oil (i.e., the oil that remained on the surface of the capsule) may also have affected the stability of the bioactive compound. In other words, the oil that was present at the surface of the capsule might have undergone oxidation and thereby affected the oxidative stability of the sample [38]. Alpizar-Reyes et al. [39] observed a similar situation when sesame seed oil was microencapsulated with tamarind seed mucilage.

Compared with the free sacha inchi oil [40], during in vitro digestion, the soft capsules showed a low protection of the encapsulated PUFAs against gastric conditions because of the nature of the wall materials and the G/CM ratio in the matrix. This behavior was associated with the intermolecular interactions between the functional groups of CM combined with G, which gradually disappeared as a result of the repulsion forces of the biopolymers [21,41,42], as well as the rapid degradation of the gelatin by the proteolytic enzymes present in the stomach [20], which reduced the barrier properties of the matrix. This was insufficient at protecting the oil against oxidation, and in turn affected the porosity of the biopolymer matrix and the oil release rate during in vitro digestion. Furthermore, the low amount of CM hydrocolloid in the soft capsule allowed for the degradation and facilitation of the enzymatic hydrolysis of G under gastric conditions [40].

These results seem to contradict the results of other published studies. Cortés-Camargo et al. [43] found that there was a delayed release of lemon essential oil microencapsulated using mesquite gum–chia mucilage mixtures. Papillo et al. [44] reported a high in vitro bioaccessibility of curcuminoids that were microencapsulated using gum arabic and maltodextrins as encapsulating agents. Da Silva Stefani et al. [45] reported a good bioavailability of nanoencapsulated linseed oil using chia seed mucilage as a structuring material. Jannasari et al. [42] studied the microencapsulation of vitamin D using gelatin and cress seed mucilage, and found release rates in the gastric and intestinal media of 28% and 70%, respectively. In addition, Barrow et al. [46] reported a high bioavailability of omega-3 fish oil microencapsulated using the technique of complex coacervation (140–180 mg EPA/DHA/g powder) in contrast with softgel capsules (data not shown). These results suggest that the G/CM matrix is not an encapsulating biopolymer that is sufficiently resistant to gastric conditions, thus reducing the bioaccessibility of the bioactive compounds carried by G/CM softgels.

### **5. Conclusions**

In this work, oil samples extracted from sacha inchi seeds were encapsulated in softgels composed of gelatin (G) and cactus mucilage (CM) biopolymers, and then exposed to simulated gastric conditions. The nutritional composition of the oil samples was evaluated before and after in vitro digestion by means of GC–MS and GC–FID spectrometry. In this way, the protective capacity of the contents of sacha inchi oil offered by the G/CM biopolymeric wall of the softgel against digestive processes was evaluated.

α-Linolenic (C18:3 ω-3), linoleic (C18:2 ω-6), oleic (C18:1 ω-9), stearic (C18:0), and palmitic (C16:0) acids were the main fatty acids present in the non-encapsulated sacha inchi oil. It was found that the content of polyunsaturated fatty acids (PUFAs), especially α-linolenic (C18: 3 ω-3) and linoleic (C18: 2 ω-6), carried by the G/CM softgels, decreased significantly during in vitro digestion (bioaccessibility equal to 1.70% or 1.46%, respectively), which revealed a reduction in the nutritional value of the encapsulated oil after undergoing gastric processes. The low protective capacity of the G/CM wall material was attributed to the low concentration of the CM hydrocolloid, which left the gelatin biopolymer exposed to enzymatic hydrolysis.

Although the bioaccessibility of the PUFAs obtained was relatively low, we believe that the use of a mixture of proteins (gelatin) and heteropolysaccharides (cactus mucilage) for the manufacture of microcapsules can act as a suitable delivery system for the incorporation of other bioactive compounds within acidic food matrices before being subjected to digestive processes, for example, by encapsulating functional agents and subsequent release (by shacking) in media such as fruit juices or dairy drinks. This result will undoubtedly be interesting for certain applications in the food and pharmaceutical industries.

**Author Contributions:** Data curation, R.C. and A.C.-C.; formal analysis, J.A.G.C. and M.C.O.; investigation, A.W.-T. and M.C.O.; project administration, M.C.O.; writing (original draft preparation), M.C.O.; writing (review and editing), J.A.G.C. and M.C.O. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was funded by the Universidad de Boyacá and the Universidad Pedagógica y Tecnológica de Colombia through the interinstitutional Project SGI 2384 of the Vicerrectoría de Investigaciones of the Universidad Pedagógica y Tecnológica de Colombia.

**Conflicts of Interest:** The authors declare that there is no conflict of interest.

### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
