**Design and Preparation of New Multifunctional Hydrogels Based on Chitosan**/**Acrylic Polymers for Drug Delivery and Wound Dressing Applications**

**Ioana A. Duceac 1,2,\*, Liliana Verestiuc 2,\* , Cristina D. Dimitriu 3, Vasilica Maier <sup>4</sup> and Sergiu Coseri <sup>1</sup>**


Received: 18 May 2020; Accepted: 26 June 2020; Published: 30 June 2020

**Abstract:** The dynamic evolution of materials with medical applications, particularly for drug delivery and wound dressing applications, gives impetus to design new proposed materials, among which, hydrogels represent a promising, powerful tool. In this context, multifunctional hydrogels have been obtained from chemically modified chitosan and acrylic polymers as cross-linkers, followed by subsequent conjugation with arginine. The hydrogels were finely tuned considering the variation of the synthetic monomer and the preparation conditions. The advantage of using both natural and synthetic polymers allowed porous networks with superabsorbent behavior, associated with a non-Fickian swelling mechanism. The in vitro release profiles for ibuprofen and the corresponding kinetics were studied, and the results revealed a swelling-controlled release. The biodegradability studies in the presence of lysozyme, along with the hemostatic evaluation and the induced fibroblast and stem cell proliferation, have shown that the prepared hydrogels exhibit characteristics that make them suitable for local drug delivery and wound dressing.

**Keywords:** superabsorbent hydrogel; *N*-citraconyl-chitosan; poly(acrylic acid)/poly(methacrylic acid)

#### **1. Introduction**

Hydrogels are a momentous collection of materials with highly diverse applications in engineering, biomedical and pharmaceutical sciences [1]. Among these, drug delivery and wound dressing are specifically of interest for scientists due to the increasing number of patients having various types of acute or chronic wounds (surgical, ulcers or burns that need emergency or constant, long-term medical assistance) combined with a considerable economic burden. The significance of this problem is easily illustrated by the growing wound dressing market share and size, from USD 7.1 bln. of the global market in 2019 to an estimated USD 12.5 bln. by 2022 [2].

An effective treatment is a constant challenge which leads to a careful selection of materials in medical practice and requests an imperative development of new advanced wound dressings with combined properties. These smart materials are able to absorb blood and wound fluids, protect the injury, accelerate the healing process by one or more mechanisms: promoting fibroblast proliferation and keratinocyte migration, which are both necessary for complete epithelialization; prevention of the wound contamination with opportunistic pathogen species; efficient transport of biologically active molecules (e.g., antimicrobial agents and other pharmaceuticals) [2].

Aiming at products that could be marketed, one of the research strategies focuses on multifunctional advanced dressings. These smart, high performance materials in the form of superabsorbent hydrogels, are expected to show various characteristics and fulfil several demands for application as wound dressing, such as to provide a porous structure required to absorb exudates, while maintaining a moist environment; a high swelling ratio and fluid retention, which entail the ability to rehydrate dry wounds (e.g., eschars); transparency to allow wound monitoring; offer the possibility to endow bioactive molecules in their matrix; good biodegradability, biocompatibility, and hemocompatibility [3,4].

Furthermore, the wound healing process is improved by using the same hydrogel system for the additional delivery of a therapeutic payload, such as antibiotics, anti-inflammatory drugs or growth factors [5]. Given the premises that hydrogels enable large amounts of bioactive molecules to be loaded into the polymeric network, they permit controlled release at the desired site as the hydrogels are placed directly at the targeted location (like a dressing or injectable/in situ gelling material), and the drug release occurs through a specific mechanism or strategy (swelling, network relaxation, temperature-induced transition, etc.), it can be hypothesized that hydrogels can be used as dressings able to perform controlled drug delivery [6].

The challenge of multifunctional hydrogel design and preparation consists of finding the optimal formulation that yields in the material with the best performance as both a drug delivery system and wound dressing. Such a macromolecular structure can be obtained by using a wise selection from the large variety of natural and synthetic polymers able to provide good biological interactions and modulated mechanical resistance and biochemical interactions [7].

Chitosan, a copolymer which consists of β-(1→4)-linked 2-acetamido-2-deoxy-D-glucopyranose and 2-amino-2-deoxy-D-glucopyranose units, is one of the most abundant polysaccharides in nature and possesses specific properties highly required for materials suitable in drug delivery, wound healing, and, ultimately, in regenerative medicine. This natural macromolecule is widely used for hydrogel fabrication due to its intrinsic properties such as excellent biocompatibility, low toxicity and immune-stimulatory activity [8], leading to positive effects on wound healing [9]. However, chitosan is poorly soluble in water and in the common organic solvents, except in aqueous acidic medium. To overcome this drawback, chitosan is often subjected to chemical modifications [10]. Chitosan is already the major component of various commercial wound dressings with hemostatic activity, such as ChitoSAM™ (SAM Medical, Wilsonville, OR, USA, 2018), ChitoGauze XR pro (North American Rescue, Greer, SC, USA, 2018), ChitoFlex (H.M.T. Inc., The Woodlands, TX, USA, 2007), and Axiostat®(AXIO, New York, NY, USA, 2018), which recommends this polymer to be further investigated for this type of application [11].

On the other hand, poly(acrylic acid) (PAA) and poly(methacrylic acid) (PAM) are synthetic polymers with biocompatible and antibacterial properties, and, therefore, are widely used as (bio)adhesives or superabsorbent materials [12–14]. Polymers grafted with acrylic acid yield in highly hydrophilic materials and tunable scaffolds for drug delivery systems due to the presence of pendant carboxylic groups [15–18]. In other words, acrylic acid-based hydrogels make excellent wound dressings due to their ability to retain large volumes of water and water-soluble drugs loaded into their matrix [19].

This study focused on the design, synthesis and advanced characterization of novel multifunctional hydrogels, having a specific composition optimized according to the targeted applications. Thus, a series of hydrogels based on *N*-citraconyl-chitosan and acrylates with various structures has been obtained, and their level of performance was comparatively assessed. The variation of two parameters was considered for the optimization: (i) the nature of the cross-linking agent, either PAA or PAM, and (ii) the initiator ratio used for free radical polymerization. In addition, L-arginine, an amino acid which is a key factor in accelerated wound healing and in cellular recognition and adhesion, was conjugated with the hydrogels network in order to modulate their interaction with various fluids and contact with cells. The influence of different components on the ability of hydrogels to respond to relevant biological media was also investigated. Moreover, the hydrogels capacity to act as a drug carrier and delivery system was evaluated by studying the release profiles and kinetics of ibuprofen, a nonsteroidal anti-inflammatory drug.

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

#### *2.1. Materials and Hydrogel Preparation*

Chitosan (MW 80 kDa, Sigma-Aldrich, Darmstadt, Germany) was purified by the solubilization–precipitation method. Briefly, chitosan was solubilized in 5% aqueous acetic acid solution, and then, precipitated with 20% NaOH solution, centrifuged and dialyzed against distilled water. Citraconic anhydride, acrylic acid (AA), methacrylic acid (AM), ammonium persulfate (APS), *N,N,N ,N -*tetramethylethylenediamine (TEMED), l-arginine, 1-ethyl-3-(3-dimethyl-aminopropyl)carbodiimide (EDC), *N*-hydroxysuccinimide (NHS) and ibuprofen (Ib) were purchased from Sigma-Aldrich, Germany. AA and AM were purified by passing through an inhibitor removal column (HQ, Sigma-Aldrich, Germany) and APS was recrystallized from methanol. For the cytotoxicity assays, the following were used: Hank's Balanced Salt Solution (HBSS), Dulbecco's Modified Eagle's Medium (DMEM), thiazolyl blue tetrazolium bromide (MTT) and calcein AM (Sigma-Aldrich, Germany). All other solvents and chemicals were obtained from Sigma-Aldrich, Germany, and were used as received.

#### 2.1.1. Chitosan Chemical Modification

Chitosan was chemically modified with citraconic anhydride by the *N*-acylation reaction using a previously described method. [20] Briefly, 27.5 mL citraconic anhydride were solubilized in 40 mL acetone and the solution was added dropwise under stirring to a chitosan solution (1% w/wt. in 5% acetic acid aqueous solution) in the presence of 80 mL methanol, to prevent chitosan precipitation in the presence of acetone, and has been allowed to react at room temperature (25 ◦C), for 18 h. The chitosan/anhydride molar ratio was 1:15, chosen based on our previous studies [20]. The obtained product was further purified through dialysis against distilled water for 7 days and then freeze-dried.

#### 2.1.2. Hydrogels Preparation

The chemical method of hydrogel preparation is the copolymerization of citraconyl-chitosan with AA or AM by free radical polymerization with the formation of cross-linked networks. Thereby, a stock solution of 1% citraconyl-chitosan in distilled water was prepared and the necessary volume for each hydrogel was transferred. AA or AM, followed by APS and TEMED, were added to the chitosan solution under vigorous stirring at 800 rpm and allowed to homogenize for 10 min. In Table 1, the hydrogels composition is presented in terms of synthetic monomer/initiator ratio, which is the variable for the hydrogels. The AA and AM volumes are not given in the table, as it is a constant ratio between the natural and synthetic polymers of 1:7 *w*/*w* citraconyl-chitosan/synthetic monomer. The APS/TEMED thermolabile initiator system was chosen due to its low toxicity and it was added to the polymerization mixture in different ratios, as presented in Table 1. The mixtures were transferred into molds (10 mL glass cylinders) and the radical polymerization reaction was carried out for 2 h at 70 ◦C. The obtained hydrogels have been thoroughly washed in distilled water until constant pH (pH ≈ 6.5) to ensure the removal of all unreacted monomers, APS, TEMED, oligomers and other residues. The purified materials were freeze-dried for further characterization, thus, ensuring porous network conservation.


**Table 1.** Hydrogels composition, the cross-linking reaction yield, and final chitosan and synthetic polymer content present in 100% hydrogel, as determined by elemental analysis.

\* an additional A was added to hydrogels code after the reaction with arginine.

#### 2.1.3. l-Arginine Coupling with the Hydrogels Network

The amino acid coupling reaction evolved in two stages. Firstly, the weighted samples were immersed in specific volumes of EDC/NHS solution buffered at pH = 5.4 for 6 h, for the activation of carboxylic groups. The pH value was chosen to ensure the availability of COOH groups from the synthetic polymer chains for the coupling reaction. The EDC/NHS solution volumes were determined for each sample based on the molar ratios COOH:EDC = 1:2, EDC:NHS = 1:1, and on the concentration of COOH groups for each hydrogel calculated from the elemental analysis data. Then, all hydrogels were washed three times with PBS, and then, put in contact with the arginine solution, at pH = 10 for 24 h, to enable the formation of the amide bond between the activated carboxylic groups from hydrogels and the α-amino moieties in the structure of arginine. Finally, the functionalized scaffolds were washed thoroughly with PBS and distilled water, then freeze-dried.

#### *2.2. Hydrogels Characterization*

#### 2.2.1. Structural and Morphological Analysis

The chemical composition was evaluated by elemental analysis, quantifying the nitrogen content in hydrogels by the Kjeldahl method. All FTIR spectra were recorded on dried samples in KBr pellets using a Bruker Vertex 70 spectrophotometer (Berlin, Germany) and scanned within the range 400–4000 cm−<sup>1</sup> in transmittance mode. Gold coated cross-sections of hydrogels were examined with a SEM Tescan-Vega microscope (Brno, Czech Republic), at ambient temperature, with an operating voltage of 30 kV, and the morphology images were analyzed with ImageJ software.

#### 2.2.2. Fluid Retention Study

The swelling behavior of all hydrogels was tested in phosphate-buffered solution (PBS) and simulated wound fluid (SWF), using the gravimetric method (50 mg aliquots immersed in 10 mL solution, incubated at 37 ◦C). The PBS solution had precise parameters: pH = 7.4, density 1.02 g/mL and concentration 0.01 M. The SWF solution was prepared using albumin (2%), NaCl (0.4 M), CaCl2 (0.02 M) and was finally buffered at pH = 7.4 with Trisbase (0.08 M); all components were solubilized in deionized water. Measurements were made for all aliquots at regular intervals and the swelling degree was calculated using the following equation:

$$SD = \frac{W\_t - W\_0}{W\_0} \times 100\tag{1}$$

where *SD*—swelling degree (%), *W*t—hydrogel weight at different times and *W*0—initial hydrogel weight.

Before each gravimetric measurement, the swollen aliquots were gently tapped on absorbing paper to eliminate the surface liquid excess. The tests were performed in triplicate.

#### 2.2.3. Drug Release Study

Ibuprofen was used as a model for drug release study. Initially, hydrogels were loaded by the swelling-diffusion method [21,22]. Weighted dry aliquots of 80 mg were immersed in 10 mL of 8 mg/mL ibuprofen solution in water/ethanol (1:1 *v*/*v*). After 24 h, the excess was removed and the samples were freeze-dried (using a freeze-dryer with pump for solvents, Labconco, USA). The in vitro release study was performed at 37 ◦C and neutral pH, chosen to mimic the extracellular environment. The aliquots were introduced in dialysis membrane and immersed in 500 mL PBS (pH = 7.2, 0.035 M), under dynamic conditions, and the samples were analyzed. The release was studied for 600 min and readings were performed automatically at preset time intervals. The absorbance was recorded at 220 nm, using an Erweka Dissolution tester DT 700 (Germany) coupled with a PharmaSpec UV-1700 spectrophotometer (Shimadzu, Japan). Blank experiments using drug-free hydrogels confirmed that there was no interfering background absorption. The released amount of ibuprofen was calculated using a standard calibration curve measured for the drug at 220 nm on a dilution series. Finally, the cumulative amount of released drug was plotted against time. Data analysis was done by fitting various kinetic models to the curves.

#### 2.2.4. In Vitro Degradation Study

Aliquots of the materials (30 mg) were immersed in 10 mL PBS, pH = 7.4, 0.01 M, containing 1200 μg/mL 1ysozyme, in a dialysis membrane (MWCO = 14,000 Da) that was further immersed in 30 mL PBS and incubated at 37 ◦C. For measurements, 1 mL of solution was sampled and evaluated. The concentration in saccharide reducing units was measured by the potassium ferricyanide method [23]: the extracted solution was mixed with 4 mL of 0.05% potassium ferricyanide solution in 0.5M Na2CO3 solution, and boiled for 15 min, allowed to cool and then, diluted with 5 mL PBS. The sample absorbance was measured at 420 nm, employing a PharmaSpec UV-1700 spectrophotometer (Shimadzu, Japan) and the amount of chitosan reducing ends was calculated using a calibration curve for *N*-acetyl-d-glucosamine.

#### 2.2.5. In Vitro Cytocompatibility Study

Selected hydrogels (synthesized with 0.8%, 1.2%, and 1.4% APS) were sterilized by immersion in 5 mL of 70% ethanol for 15 min and then, washed several times with HBSS and DMEM. The circular samples with 5 mm diameter cut with a sterile biopsy circular scalpel were incubated with normal human dermal fibroblasts (NHDF, Lonza, Basel, Switzerland), in 48-well plates and, as a reference, cells were seeded under the same conditions. Cell viability was assessed at 24, 48 and 72 h using a direct contact MTT assay. The absorbance of the obtained solutions was measured at 570 nm using a Tescan Sunrise Plate Reader. Cell viability was calculated by Equation (2), where RMA is the relative metabolic activity, As is absorbance of the sample and Ac is absorbance of the negative control. Each result represents the mean viability ± standard deviation of four independent experiments.

$$RMA = \frac{A\_s}{A\_\odot} \times 100\tag{2}$$

In addition to the MTT test, a live/dead staining assay was performed. Sterile hydrogel samples were put in direct contact with stem cells. At pre-set time intervals, the calcein solution was used to color the live cells by incubation for 20 min at 37 ◦C, followed by a microscopy analysis. Images were taken in phase contrast and fluorescence using a Leica DM IL LED Inverted Microscope with a Phase Contrast System and Fluorescence (Leica Microsystems GmbH, Wetzlar, Germany).

#### 2.2.6. Hemostatic Properties

The prothrombin time (PT) of blood and fibrinogen concentration in blood after contact with the prepared hydrogels were measured. Integral blood from healthy, non-smoker volunteers (venous puncture) was incubated with anticoagulant (aqueous sodium citrate 3.8% *w*/*v*; ratio 1/9 *v*/*v*). The hydrogels (5 × 10 × 1 mm) were added to 5 mL blood, and the control sample was considered the free integral blood, and both the sample and control were incubated at room temperature for 30 min. After that, the hydrogels were separated from blood by centrifugation (2500 rpm, 10 min). PT in blood plasma was determined (as a mean of three values) using a semi-automatic coagulometer Helena with 2 channels and photo-optical technique coagulation systems and PT-Fibrinogen kit (International Sensitivity Index (ISI) = 1.07). The International Normalized Ratio (INR) was calculated as the ratio of prothrombin times recorded for the hydrogels (PTH) and control (PTC) samples, raised to the power of the ISI value:

$$INR = \left(\frac{PTH}{PTC}\right)^{ISI} \tag{3}$$

#### **3. Results and Discussions**

#### *3.1. Hydrogels Structure*

The design of chitosan-based hydrogels considered the parameters that influence the formation of the cross-linked polymeric networks. A schematic illustration of the hydrogel preparation is illustrated in Figure 1. In the first stage, chitosan was chemically modified with citraconic anhydride. The anhydride reacted with the amino groups available on the chitosan backbone and yielded in amide bonds, which induced novel useful properties. The C=C bonds from the anhydride moieties are available for further copolymerization.

**Figure 1.** Schematic representation of the hydrogel network.

In order to observe and compare the hydrogels structure and properties depending on the acrylate nature, a constant 1:7 ratio (*w*/*w*) between citraconyl-chitosan and the synthetic monomer was chosen for both AA- and AM-containing materials.

The free radical polymerization in the second stage was initiated by the thermal decomposition of APS molecules, in the presence of TEMED, at 70 ◦C, when the free radicals reacted with either monomer molecules, AA or AM, or with citraconyl sequences in the chitosan structure. Consequently, it can be assumed that the network was finally formed of (1) PAA or PAM cross-links between *N*-citraconyl-chitosan chains, as designed, (2) *N*-citraconyl-chitosan/PAA or PAM graft copolymer, and (3) a semi-IPN of citraconyl-chitosan and PAA or PAM. All polymeric chains have been stabilized through hydrogen bonds, ionic interactions between –COOH and –NH2 groups, or other physical interactions.

From the material science point of view, the aim of this study was to assess the influence of APS concentration on the hydrogel cross-linking process. The APS percentage is reflected in the cross-linking density of the polymeric network, due to the occurrence of more or less reaction centers. The acrylic chain length further affects the hydrogel morphology, pore dimension, fluid absorption and retention etc. The cross-linking reaction yield was between 58 and 79% for hydrogel with PAA and in the range 41–59% for those with PAM (data in Table 1). A correlation can be observed between the polymerization yield and the initiator ratio: the yield decreased along with the increase in APS concentration in the reaction mixture. Moreover, the values were lower for PAM polymerization, but the value range was similar for both synthetic polymers.

Finally, in order to achieve improved biological interactions, the hydrogels were optimized by the bioconjugation of l-arginine, an amino acid that could be coupled with the carboxylic groups existing on the synthetic polymer chains in both PAA and PAM. l-arginine is an amino acid precursor of proline which is converted to hydroxyproline and then, to collagen; it has a positive influence on the insulin-like growth factor (IGF-1), a hormone which promotes wound healing. Given its properties, l-arginine is expected to impart valuable characteristics to the conjugated materials. Consequently, the properties were assessed prior and after this reaction in order to compare the hydrogels performance and arginine influence.

The elemental analysis was performed on chitosan, citraconyl-chitosan and all hydrogel samples, before the arginine conjugation reaction. The composition was calculated and the data are shown in Figure 2.

**Figure 2.** Nitrogen concentration of chitosan, *N*-citraconyl-chitosan and prepared hydrogels.

Compared to raw chitosan, the modified polysaccharide contained less nitrogen, thus, confirming the presence of citraconic anhydride in its structure. The decrease in nitrogen percentage is more significant in hydrogel samples due to the acrylic cross-links. The amount of nitrogen increased along with APS ratio in the hydrogels prepared with AM, indicating a higher reaction yield when less initiator was used. These data allowed the determination of the final composition in each hydrogel and the content of chitosan and synthetic polymer in 100% hydrogel is presented in Table 1. Based on these results, the carboxylic groups available for the final arginine coupling reaction could be determined.

The comparative FTIR spectra recorded for chitosan and citraconyl-chitosan, and for the selected scaffolds CA10, CA10A and CM10, respectively, are presented in Figure 3.

**Figure 3.** FTIR spectra for chitosan, *N*-citraconyl-chitosan (left), and selected hydrogels (right).

Chitosan exhibits specific absorption bands in the FTIR spectrum, as follows: one intense absorption peak at 3439 cm−<sup>1</sup> assigned to the OH and NH2 functional groups present in the polysaccharides structure; the amide I band was present at 1651 cm−1, shifted at higher wavenumber and combined with OH. The specific peak for C-O-C glycoside bond appeared at 1087 cm<sup>−</sup>1.

By comparison, the spectra of citraconyl-chitosan displayed new peaks correlated to the new amide bonds and to the structure of citraconyl residue. Hence, a large intense band was recorded between 1709 and 1568 cm−<sup>1</sup> as a result of the fusion of the peaks specific to amide I and the carboxylic group, due to C=O stretching. Furthermore, C-O-H in plane bending from the COOH group led to the peak at 1461 cm−1, while the peaks at 2924 and 2856 cm−<sup>1</sup> were correlated to the methyl groups symmetric and asymmetric stretching vibrations.

Scaffolds with 1% APS, cross-linked with AA or AM, before and after arginine conjugation, are presented in Figure 3 (right). The chemical cross-linking was confirmed by the presence of intense peaks at 1713 cm−<sup>1</sup> (CA10) and 1705 cm−<sup>1</sup> (CM10), correlated with the C=O bond from the carboxylic group and the intense frequency bands in the interval 2500–3500 cm−<sup>1</sup> specific to PAA–O–H bond stretching vibrations (from the carboxylic group); in the range 2600–3600 cm<sup>−</sup>1, for the hydrogels with PAM, the bands were more intense due to, and shifted by, the symmetric and asymmetric stretch specific to the methyl group. In addition, the presence of the amide bond previously formed between chitosan and the anhydride was confirmed by the signals at 1551 cm−<sup>1</sup> (CA10) and 1544 cm−<sup>1</sup> (CM10), bands assigned to N–H bending vibrations and C–N stretching specific to amide II, hence, confirming polymer modification.

Arginine coupling with hydrogels yielded in new peaks that appeared in the FTIR spectrum of sample CA10A, corresponding to the amide bond vibrations, as follows: at 1656 cm−<sup>1</sup> a new peak related to the stretching vibrations of C=O bond (amide I); at 1549 and 1323 cm<sup>−</sup>1, respectively, novel bands assigned to C–N stretching and N–H bending vibrations (amide II, III).

#### *3.2. Hydrogels Morphology*

SEM images from Figure 4 were recorded for samples with AA and AM, synthesized in the presence of 0.6% or 1% APS, before and after arginine coupling, in order to compare and determine the influence of the considered preparation parameters. All samples displayed a porous morphology with interconnected pores of variable dimensions. At a microscopic level, there are areas that differ in aspect, which may be explained by the fact that the radical polymerization did not occur uniformly in the whole reaction volume, thus, leading to regions where more synthetic polymer was formed.

**Figure 4.** SEM images of hydrogels with acrylic and methacrylic acid, obtained with 0.6% or 1% APS, before and after arginine immobilization.

Particularly, for the AA-based hydrogels, well defined pores were obtained and the arginine coupling caused the formation of larger pores and high pore dispersity (see Table 2). However, in the case of hydrogels prepared with AM, the pore size increased with the increasing amount of APS and the presence of the amino acid favored the formation of pores with an even higher size (the average pore dimension varied between 40 and 550 μm). In the case of the arginine-containing hydrogels (sample CM10A in Figure 4), a significant decrease in pore size was noticed. The explanation resides in the direct bonding of arginine to the synthetic polymer moieties and, thus, arginine molecules are pendant toward the interior of the pores, which finally entailed the reduction in pore diameter.


**Table 2.** Pore size variations with hydrogel composition.

#### *3.3. Swelling Behavior in Simulated Physiological Conditions*

The hydrogels' ability to absorb fluids is essential in skin treatment applications, as it concerns, for example, the wound exudates or a hemorrhage. The swelling behavior of the hydrogels can be affected by multiple factors: the hydrophilic or hydrophobic functional groups (–COOH, –OH, –CO–NH–) existing on the polymer chains, the internal morphology, the network parameters and its elasticity, temperature, pH and swelling medium [7]. The experimental data obtained during swelling assays are shown in Figure 5. In order to test the hydrogels, two types of fluids were chosen according to various methods described in the literature: phosphate-buffered solution (PBS) at pH 7.4 and simulated extracellular or wound fluid (SWF) [24].

Hydrogels with AA showed SDs of up to 30,000% due to the hydrophilicity of the synthetic polymer and the maximum swelling degree, associated to the plateau, was reached 20–60 min after contact, depending on the polymer chain length and network morphology, but mainly due to the carboxylic groups that provide an important hydrophilic behavior. By comparison, the hydrogels with AM hydrated slower and absorbed up to 16,000%. APS influence can also be noted: a higher initiator concentration is associated with a higher swelling capacity and a slower absorption rate. Such a result is considered to be determined by a higher cross-linking density and shorter polymeric chains, which led to pores with smaller size. Hence, the network's elasticity was altered, and the hydrogels became less flexible. [7] The presence of arginine in the polymeric networks determined a fast fluid absorption evident in all hydrogel samples, but with a smaller retained volume, up to 19,000% PBS. The influence of the swelling media was also investigated. The obtained experimental data suggested that SWF is absorbed in lower quantities as compared to PBS, presumably due to the high viscosity of the wound fluid and to existing albumin large molecules which do not access the domains with small pores, thus, leading to swelling degrees of a maximum of 12,000%.

Upon contact with any of the fluids, the polymer network started to swell progressively and the solution accessed the pores. The swelling mechanism may be either Fickian and the swelling is diffusion-controlled, or non-Fickian, when the swelling kinetics are based mostly on network relaxation, according to Ritger–Peppas model (Equation (4), data in Table 3) [25–28].

$$\frac{M\_{\rm f}}{M\_{\infty}} = kt^n\tag{4}$$

where *M*<sup>t</sup> is the material weight at time *t*, *M*<sup>∞</sup> is the initial material weight, *k* is the process rate constant, *t* is time and n is the diffusional exponent. The model was fit to experimental data for all hydrogels, before arginine conjugation, and tested in PBS.


**Table 3.** Kinetic parameters for hydrogels without arginine, immersed in PBS.

**Figure 5.** *Cont*.

**Figure 5.** Swelling behavior of the hydrogels: (**a**,**b**) Kinetic swelling degree data, in PBS for hydrogels with acrylic and methacrylic acid; (**c**,**d**) Maximum swelling degrees in PBS or SWF, before and after arginine coupling.

In the present case, the diffusional exponent n had values between 0.75 and 0.93 for hydrogels with AA and between 0.61 and 0.73 for those with AM; hence, for all hydrogel samples the value of the diffusional exponent n fell within the range 0.45–1, a fact that indicates a non-Fickian swelling mechanism and substantiated that the water retention was driven by the relaxation of the network.

The fast swelling, favored by the presence of arginine, and the large volumes retained in the hydrogel networks, due to the chitosan/acrylate cross-linking, indicated a superabsorbent behavior [12,29–34] for the prepared hydrogels, which is a highly advantageous characteristic for applications such as drug delivery and wound dressing, as intended.

#### *3.4. Ibuprofen Release Profiles and Kinetics*

Ibuprofen (Ib), a nonsteroidal anti-inflammatory drug, is used for the treatment of pain, fever, and inflammation. It is insoluble in water, but has a high solubility in most organic solvents, including ethanol. The mechanism that leads to the analgesic, antipyretic, and anti-inflammatory effects evolves through a non-selective inhibition of cyclooxygenase (COX) enzymes. Normally, COX enzymes convert arachidonic acid to a prostaglandin, which further mediates pain, inflammation, and fever [35]. It is well known that Ib systemic administration causes severe issues, such as stomach and intestine injuries, circulatory problems (heart failure, AVC), and kidney lesions. To limit these drawbacks, a local delivery of Ib is preferred, such as a dressing. Moreover, in this case, the therapeutic effects are enhanced.

The in vitro release of Ib from the hydrogels was analyzed in PBS, pH = 7.2, at 37 ◦C, mimicking the physiological environment. It is important to emphasize that the pH was maintained at the level recorded for the extracellular environment in physiological conditions, in order to be fit for the medical applications taken into consideration in the present article. The readings expressed as cumulative release were plotted versus time in order to obtain the drug release profiles and the results are shown in Figure 6.

**Figure 6.** The in vitro ibuprofen release kinetic data (**a**,**b**).

The release profiles for the hydrogels with AA indicated a steady, controlled release of a maximum of 4 mg ibuprofen over 10 h. However, in the case of hydrogels with AM, Ib was released in double the amount and in a shorter time, similar to a burst effect. In addition, AA-based hydrogels did not reach a plateau after 600 min, while hydrogels with AM reached the plateau after 300–360 min, meaning that the equilibrium in drug concentration was reached between the hydrogel and the environment. Moreover, as the APS ratio grows from 0.8% to 1.4 or 1.6%, respectively, the amount of released Ib is greater. This can be explained by the fact that more APS led to larger pores, according to SEM micrographs, which favored drug diffusion in both directions—at loading and at release.

The drug loading and release capacity depend on the polymeric matrices' structure and morphology, their ability to absorb and retain the drug solution, and the drug–hydrogel interaction [36]. The Ib molecule is amphiphilic and able to participate in various molecular/supramolecular associations. Ionic interactions between the COOH group from Ib and the NH2 moiety from chitosan were generally present in all hydrogels, since the same amount of chitosan was present in all matrices. It should

be emphasized that there was no arginine in the hydrogels tested for drug delivery. The significant difference in hydrogel–drug interactions was correlated with the type of synthetic polymer.

Interestingly, the hydrogels with PAM released higher amounts of drug and in shorter time intervals, results that were in contradiction with the swelling degrees, since the PAA favored a faster absorption and the retention of larger fluid volumes. It is worth mentioning that the drug loading took place in a water/ethanol solution, which consequently may have been favorable for the swelling in chitosan/PAM networks, rather than in matrices with PAA. Moreover, it can be assumed that hydrophobic interactions appeared between Ib and the methyl groups in PAM structure.

It is known that the drug release from hydrogels is governed by several phenomena: diffusion, erosion, network relaxation, all in various proportions, depending on the polymer nature, network stability and parameters, morphology, hydrophilicity, the nature of the drug and its interaction with the matrix, the release medium etc., [37–40]. In order to study the drug release kinetics and mechanism, the drug release data were fitted into four models, using the following equations [21,41–43]:

$$\text{Zero order}: \ \frac{Q\_t}{Q\_0} = k\_0 t \tag{5}$$

$$\text{First order}: \ln \frac{Q\_t}{Q\_0} = k\_1 t \tag{6}$$

$$\text{Higuchi model}: \frac{Q}{Q\_0} = k\_H t^{1/2} \tag{7}$$

$$\text{Korsmeyer-Pepps model} : \frac{Q\_t}{Q\_0} = kt^n \tag{8}$$

where *Qt* is the amount of drug released at time t, *Q*<sup>0</sup> is the original drug concentration in the material (40 mg), n is the release exponent and K is the release rate constant.

The models were fitted to the curves with the highest cumulative drug release in each series, namely CA16 (which had the highest APS ratio in the PAA-based hydrogels series) and CM12. The latter had a cumulative amount close to but higher than CA14, which had the highest APS ratio of the materials cross-linked with PAM. The results obtained for the correlation coefficient are shown in Table 4 and suggest that for the hydrogels with PAA, the release mechanism followed the Korsmeyer–Peppas kinetics [41], while the hydrogels with PAM fitted the Korsmeyer–Peppas and the Higuchi models well [42,43]. In addition to these data, it was important to determine the diffusional exponent n, which is a parameter that indicates the mechanism of drug release and varies depending on the geometry of the release device. Fickian diffusion is confirmed for *n* < 0.45; when 0.45 < *n* < 0.89, the drug transport is anomalous (non-Fickian)—in other words, the diffusion and relaxation rates are similar and the physical phenomena are diffusion and erosion of the matrix; if *n* > 0.89, then the key mechanism of drug release is Case II transport, thus, the drug release is determined by polymer relaxation [26,44]. The values obtained for the exponent n suggested anomalous (non-Fickian) transport, which indicated that the drug is released from the hydrogels by diffusion associated with the materials erosion. The drug release can be controlled by selecting the acrylate nature for the cross-linking and the initiator ratio. The drug is incorporated in a hydrophilic, initially glassy material, and the release is basically swelling-controlled [26].

**Table 4.** Drug release correlation coefficient values from different kinetic models.


#### *3.5. The In Vitro Hydrogels Degradation*

The biodegradation assays have proven that the chitosan-based scaffolds were susceptible to lysozyme attack under simulated physiological conditions of pH and temperature, as shown in Figure 7.

**Figure 7.** Degradation behavior of hydrogels in medium with enzyme.

The presence of the arginine molecules in the polymeric network caused the rapid absorption of the enzyme solution into the hydrogel networks, followed by their expansion, which entailed a fast degradation rate in all investigated samples, similar to a burst effect after 24–48 h. This phenomenon can be explained by the enhanced accessibility of lysozyme to the glycoside bonds within chitosan chains.

Lysozyme is present in any type of wound due to neutrophils secretion, in both infected and non-infected wounds. This phenomenon is associated with the inflammation phase of wound evolution and has the crucial role of cleaning the wound bed before any regeneration processes are triggered. The enzyme levels determined in the wound fluids were of 0.4 mg/mL for a wound with inflammation and at least 0.5 mg/mL lysozyme in an ulcer wound fluid. [45] Compared to the literature data, the 1.2 mg/mL lysozyme concentration used in this experiment was indicating that the hydrogels had a good stability in a more aggressive simulated environment, being able to maintain the cross-linked network for up to 48 h.

#### *3.6. Hydrogels Cytocompatibility*

All materials intended to be used for medical applications must be tested by means of biocompatibility. One of the standard assays is theMTT study forin vitro evaluation of cytocompatibility by the direct contact method. In addition to the obtained MTT data, a live/dead staining was performed and the results are shown in Figures 8 and 9.

**Figure 8.** Relative metabolic activity data from the MTT assays for hydrogels with AA or AM, obtained with 0.8%, 1.2% and 1.4% APS, respectively, tested for different time intervals.

**Figure 9.** Live/dead staining assay images of negative control and cells after incubation with different hydrogels (magnification 10×).

The relative metabolic activity induced by the hydrogels was between 76 and 100%, depending on the materials' composition. All tested materials were conjugated with arginine, an amino acid well-known for its key role in accelerated wound healing and in cellular recognition and adhesion. Therefore, arginine is a valuable element in multicomponent systems used for drug delivery and

wound dressing. [46] After the first 24 h, the best results were for materials with 1.2% APS, followed by a decrease in time, while the hydrogels with 1.4% APS showed a better cell response after a lag period of 72 h.

The fluorescence microscopy images were analyzed and several observations can be made: all hydrogels are transparent in the hydrated state, which can be an advantage for wound monitoring; the hydrogels induced an intense proliferation compared to the control with the formation of a confluent monolayer after 48 h; based on cell morphology, the stem cells appeared to have undergone differentiation. The cell response is the result of various cell–hydrogel interactions determined by the materials' structure and properties on the one hand, and, on the other hand, by cell biology and biochemistry. Thus, it is possible that the inclusion of L-arginine in the matrices induced this cell behavior—the intense proliferation and the differentiation. Due to the fact that the differentiation is associated with a lower metabolic effect and that it may have begun after a fast proliferation—a process that lasted less than 24 h—it may explain the low RMA values.

#### *3.7. Hemostatic Properties*

The thrombogenic character is an important property for wound dressing materials. The protein adhesion initiates the coagulation cascade, and thus, the material can accelerate the thrombus formation, stop the hemorrhage, and stimulate the healing process. The hydrogels porosity and their rough surface are known to be beneficial to blood coagulation [47]. Moreover, the colloid formed on the pore's surface upon the blood fluid sorption favors the adhesion of blood cells, and the dressing material presses the wound area and efficiently limits the bleeding as a physical barrier. The prothrombin time (PT) test is an efficient method to evidence the thrombotic or antithrombotic activity of biomaterials. Generally, the lower the value of PT, the faster the clotting rate is and the better the antithrombotic activity of these materials. [48] The experimental data recorded for the selected samples in terms of hemostatic activity are given in Table 5.



The excellent hemostatic properties of the prepared hydrogels were due to their strong swelling capacity and porous structure. For all tested hydrogels, the PT and INR values decreased in comparison with the control sample, while higher values for fibrinogen concentration were recorded. The PAA-based hydrogels exhibited the most intense hemostatic activity, compared with those with PAM, and this cumulative effect can be attributed to the synergic action of their specific characteristics: porosity, high hydrophilicity, and the positively charged moieties from chitosan and arginine immobilized onto networks, as was separately reported for chitosan-based materials and arginine [49–54]. Furthermore, at a lower APS ratio, the hemostatic effect was more intense for both hydrogel series. As superabsorbent hydrogels, these materials absorbed the plasma in the blood and, hence, promoted the local accumulation of the coagulation factors, erythrocytes and platelets, and accelerate the adhesion of the blood components to their surface, and thus, the blood coagulation was sped up. Therefore, taking into consideration these results and the literature data in the field, it can be stated that these new hydrogels successfully respond to the performance requirements for wound dressing materials.

#### **4. Conclusions**

In the present study, two series of hydrogels based on chemically modified chitosan grafted with acrylic polymers as cross-linkers were compared. The advantages of both natural and synthetic polymers led to hydrogels with combined desirable characteristics, and the networks were further conjugated with arginine in order to achieve enhanced properties. The results obtained from FTIR spectra, elemental analysis and SEM images confirmed the formation of networks with interconnected pores and evidenced a great influence on the hydrogel properties of the monomer and of the initiator ratio used during the synthesis. It has been confirmed that hydrogels have a high fluid absorption behavior after their interaction with buffer solution and simulated wound fluid, under physiological conditions. The ibuprofen release profiles were studied in vitro and the kinetics fitted the Korsmeyer–Peppas model best. The hydrogels have proved to be biodegradable in the presence of lysozyme with hemostatic properties and the cytocompatibility tests indicated the hydrogels' ability to induce cell proliferation and differentiation. In conclusion, these superabsorbent hydrogels with tunable properties may be considered suitable materials for both drug delivery and wound dressing applications.

**Author Contributions:** Conceptualization, I.A.D., L.V. and S.C.; Data curation, I.A.D.; Formal analysis, I.A.D.; Investigation, I.A.D., C.D.D. and V.M.; Methodology, I.A.D., L.V. and S.C.; Visualization, I.A.D.; Writing—original draft, I.A.D.; Writing—review and editing, I.A.D., L.V. and S.C. 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* **Developed Chitosan**/**Oregano Essential Oil Biocomposite Packaging Film Enhanced by Cellulose Nanofibril**

#### **Shunli Chen 1, Min Wu 1,2,\*, Caixia Wang 1, Shun Yan 1, Peng Lu 1,2 and Shuangfei Wang 1,2,\***


Received: 8 July 2020; Accepted: 8 August 2020; Published: 9 August 2020

**Abstract:** The use of advanced and eco-friendly materials has become a trend in the field of food packaging. Cellulose nanofibrils (CNFs) were prepared from bleached bagasse pulp board by a mechanical grinding method and were used to enhance the properties of a chitosan/oregano essential oil (OEO) biocomposite packaging film. The growth inhibition rate of the developed films with 2% (*w*/*w*) OEO against *E. coli* and *L. monocytogenes* reached 99%. With the increased levels of added CNFs, the fibrous network structure of the films became more obvious, as was determined by SEM and the formation of strong hydrogen bonds between CNFs and chitosan was observed in FTIR spectra, while the XRD pattern suggested that the strength of diffraction peaks and crystallinity of the films slightly increased. The addition of 20% CNFs contributed to an oxygen-transmission rate reduction of 5.96 cc/m2·day and water vapor transmission rate reduction of 741.49 g/m2·day. However, the increase in CNFs contents did not significantly improve the barrier properties of the film. The addition of 60% CNFs significantly improved the barrier properties of the film to light and exhibited the lowest light transmittance (28.53%) at 600 nm. Addition of CNFs to the chitosan/OEO film significantly improved tensile strength and the addition of 60% CNFs contributed to an increase of 16.80 MPa in tensile strength. The developed chitosan/oregano essential oil/CNFs biocomposite film with favorable properties and antibacterial activity can be used as a green, functional material in the food-packaging field. It has the potential to improve food quality and extend food shelf life.

**Keywords:** cellulose nanofibrils; chitosan; oregano essential oil; antimicrobial; oxygen barrier properties

#### **1. Introduction**

Petroleum-based plastic films have been widely used in recent decades in the packaging field due to their low cost, good chemical stability and excellent barrier performance [1]. However, the use of non-biodegradable materials in packaging applications has raised concerns about environmental pollution. The demand for advanced and eco-friendly packaging materials owing to their excellent physical, mechanical, and barrier properties and antimicrobial activity is significantly increasing, especially with increased consumer awareness of environmental protection and with increased attention being paid to food quality and safety. Recently, biodegradable materials, such as chitosan [2–4], starch [5], pectin [6], gelatine [7] and cellulose [8], as matrices to develop biologic composite packaging materials with good oxygen, water vapor barrier properties, antibacterial properties and mechanical properties, have become a focus of scholars. However, biodegradable packaging films prepared by a single biomass material often cannot simultaneously possess a variety of favorable properties which limits

the application of biomass packaging materials in the field of food packaging. Therefore, two or more kinds of material are blended together, and the functional components are added to prepare biocomposite packaging materials with good mechanical properties, oxygen and water vapor barrier properties and antibacterial or antioxidant properties.

Chitosan, which consists of (1,4)-linked-2-amino-deoxy-b-d-glucan, is a kind of cationic polysaccharide and is the deacetylated form of chitin [9]. As a renewable natural biopolymer, chitosan is derived from various sources and exhibits non-toxicity, biocompatibility, biodegradability and excellent film-forming properties. Meanwhile, chitosan has broad-spectrum antimicrobial activity against both Gram-positive and Gram-negative bacteria as well as fungi. Films prepared from chitosan were extensively used in food packaging as a degradable packaging material and inhibited bacterial reproduction, prolonged the shelf life of food and improved food quality and safety [10,11]. However, the poor mechanical and barrier properties of chitosan-based packaging materials compared to those of non-biodegradable materials have limited its widespread usage [12]. In order to develop the application of chitosan-based films in the field of food packaging, physical or chemical modification strategies have been tried. Poverenov et al. [13] prepared the alginate–chitosan coating by layer-by-layer electrostatic deposition, the coating showed excellent gas-exchange and water vapor permeability properties that protected the appearance of the fresh-cut melon. Caseinate and chitosan by ion interaction was carried out to obtain the composite films, which showed improved water vapor barrier properties [14]. The chitosan-grafted salicylic acid films have been found to maintain better quality in cucumber, making it a promising material for food packaging applications [15]. Cellulose nanofibrils (CNFs) from natural resources are recognized as the most abundant, renewable and biodegradable polymeric materials [16] and have been widely used due to their good biocompatibility, chemical stability, mechanical properties and oxygen barrier properties [17–20]. The addition of 3% (*w*/*w*) of CNFs increased the tensile strength, elongation at breaking and Young's modulus of corn starch film [21]. The water vapor permeability of bio-nanocomposite films was reduced by incorporating 4% (*w*/*w*) CNFs [22].

The antibacterial properties of chitosan are mainly related to its degree of deacetylation, molecular weight, types of microorganisms and other factors. Sanchez-Gonzalez et al. [23] showed that pure chitosan films had obvious antibacterial effects against *E. coli* and *L. monocytogenes*, but could not inhibit *S. aureus* growth. Song et al. [24] reported that chitosan film exhibited only slight inhibitory activity against *E. coli* and *S. aureus*. Some research showed that the antibacterial properties of chitosan were still controversial and unstable when chitosan was used as a single antibacterial agent when preparing film. Essential oil as a natural antibacterial agent was often added to the film to improve the antibacterial properties of the films. Oregano essential oil (OEO) mainly consists of phenolic compounds [25] that can effectively prevent spoilage and prolong the shelf life of food and is widely used in food packaging due to its favorable antioxidant and antibacterial properties. OEO can be directly incorporated into the matrix of packaging film and is released during transportation and/or storage of food and thereby contributes to reducing food spoilage [26]. However, to our knowledge, no information has been reported regarding the enhancement of cellulose nanofibers on the mechanical, thermal and barrier properties of chitosan/OEO packaging film.

The main objective of this study was to develop chitosan/OEO biocomposite packaging film enhanced by CNFs to obtain favorable physical characteristics and antimicrobial properties. The morphology, chemical structure, physical characteristics and antimicrobial properties of the developed biocomposite packaging films were measured and analyzed. The film will have potential applicability for food packaging and high-value goods.

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

#### *2.1. Chemicals and Materials*

Chitosan powder (80%–95%, deacetylated) was purchased from Sinopharm Chemical Reagent Co., Ltd., Shanghai, China. Pure OEO and Tween-80 were obtained from Shanghai Aladdin Bio-Chem Technology Co., Ltd., Shanghai, China. *E. coli* (ATCC 25,922) and *L. monocytogenes* (ATCC 19,115) strains were obtained from the China Center of Industrial Culture Collection, Beijing, China. Bleached bagasse pulp board was purchased from Guangxi GuiTang (Group) Co., Ltd, Guigang, Guangxi, China. All other chemicals were of analytical grade.

#### *2.2. Preparation of CNFs*

CNFs were prepared using a mechanical grinding method according to the process described by Nie et al. [27]. Initially, the bleached bagasse pulp board was soaked in deionized water overnight at room temperature and was then disintegrated by a fiber disintegrator (AG 04, Estanit GmbH, Muhlheim, Germany) for 30 min to obtain disintegrated pulp with 1% (*w*/*w*) concentration. The water in the disintegrated pulp was dehydrated and the disintegrated pulp was placed in a refrigerator at 4 ◦C overnight to balance the moisture levels. Then, the disintegrated pulp with a solid content of 2% (*w*/*w*) was ground using an ultrafine grinder (MKZA10-15J, Masuko Sangyo, kawaguchi, Japan) with -100-μm disc spacing at a speed of 1500 rpm to obtain MFC suspensions. After 10 grinding cycles, a CNFs suspension with a solid content of 2.56% (*w*/*w*) was obtained and stored in a 4 ◦C refrigerator for further use.

#### *2.3. Characterization of CNFs*

The morphology of CNFs was observed using a transmission electron microscope (TEM). The CNFs suspensions were diluted to concentrations of 0.008% (*w*/*w*) with deionized water and were ultrasonically dispersed for 30 min. A drop of the dispersed CNFs suspension was deposited on a carbon-coated grid and was then stained with 1.5% (*w*/*w*) phosphotungstic acid water for 15 min in a dark place. The dried grid was observed by a TEM (HT7700, Hitachi, Tokyo, Japan) with an acceleration voltage of 100 kV.

#### *2.4. Preparation Chitosan*/*OEO Films Enhanced by CNFs*

Chitosan (2%, *w*/*v*) was dispersed in a glacial acetic acid solution (1%, *v*/*v*) and magnetically stirred at 250 rpm for 8 h at room temperature to completely dissolve the chitosan. The pure OEO with addition amounts of 0%, 1%, 2%, 3% (*w*/*w*, chitosan-based) and Tween-80 (40% *w*/*w*, OEO-based) were added to the chitosan solution and stirred by a high-shear homogenizer (Unidrive-Model×1000D, CAT M.Zipperer GmbH, Ballrechten-Dottingen, Germany) at 8000 rpm for 10 min to obtain film-forming solutions. Tween-80 was used as surfactant to facilitate emulsion formation and stability. After ultrasonic deaeration for 1 h, 30-g film-forming solutions were cast onto Teflon plates (150 mm × 150 mm) and dried in an oven at 35 ◦C for 2 days. All dried film samples were removed from the molds and were stored at 25 ◦C and 50% RH until the antibacterial activity was evaluated.

To evaluate the enhancement of the CNFs on the properties of the chitosan/OEO biocomposite film with the best antimicrobial activity, the optimal added OEO amount was first determined. Two percent (*w*/*w*, chitosan-based) OEO was added in the follow-up experiments, based on the results (Section 3.2). The CNFs with a concentration of 2.56 wt % were dispersed in distilled water by a high-shear homogenizer at 13,000 rpm for 6 min to obtain a CNFs suspension with a concentration of 1.0 wt %. The 0%, 20%, 40%, 60% CNFs (*w*/*w*, chitosan-based) were added to the chitosan solution and stirred by a high-shear homogenizer at high speed (13,000 rpm) for 6 min to obtain mixtures of chitosan and CNFs. In addition, the 2% OEO (*w*/*w*, chitosan- and CNFs-based) and Tween-80 (40% *w*/*w*, OEO-based) were added to the mixture and homogenized using a high-shear homogenizer at 8000 rpm for 10-min to obtain film-forming solutions which contained CNFs. The film-forming solutions were deaerated, cast and dried as described above to obtain the chitosan/OEO/0% CNFs (COC0), chitosan/OEO/20% CNFs (COC20), chitosan/OEO/40% CNFs (COC40) and chitosan/OEO/60% CNFs (COC60) films.

#### *2.5. Antimicrobial Properties*

Antimicrobial activities of the sample films were evaluated using growth inhibition rates and disk inhibition zone assays. The *Escherichia coli* and *Listeria monocytogenes* cultures were regenerated through the exponential growth phase (24 h) in nutrient broth and brain heart infusion (BHI) broth in incubators at 37 ◦C and 75% RH to obtain a bacterial suspension with a concentration of 108 CFU/mL.

Before the inhibition rate test, the bacterial suspension was diluted to 10<sup>5</sup> CFU/mL by phosphate buffer saline (PBS) and the sample film was chopped into fragments and placed under ultraviolet sterilization for 1 h. Then, 100 mg of sample film fragments were placed into 5 mL of an *E. coli* and *L. monocytogenes* suspension of 10<sup>5</sup> CFU/mL and then shaken at 250 rpm in a water bath shaker at 37 ◦C for 2 h. After 2 h of contact time, 0.1 mL of *E. coli* and *L. monocytogenes* suspension diluted to 10<sup>2</sup> CFU/mL (with the sample film) was uniformly coil-coated on the nutrient surfaces and BHI agar plates, respectively. The plates were incubated for 24 h at 37 ◦C and 75% RH. The inhibition rates of *E. coli* and *L. monocytogenes* growth were calculated by the following equation:

$$\text{Growth inhibition rate (\%)} = (\text{A} - \text{B}) / \text{A} \times 100\% \tag{1}$$

where A and B are the bacterial counts from the control and the sample films, respectively. All values were averaged from three parallel experiments.

The bacterial suspension was diluted by 100 times with PBS to obtain an inoculum which contained approximately 10<sup>6</sup> CFU/mL for disk inhibition zone assays. All sample films were cut into circular discs of 10-mm diameter. All culture media were double-layered in which the concentration of the upper agar was 0.5 times that of the underlying agar (BHI agar was used as the medium for *L. monocytogenes* and nutrient agar as the medium for *E. coli*). One hundred microliters bacterial cultures with 105 CFU/mL were uniformly coil-coated on the surface of the BHI and nutrient agar plates and the film discs were placed on plates. The plates were incubated for 24 h at 37 ◦C and 75% RH. The diameters of the inhibition zones were measured with a vernier caliper.

#### *2.6. SEM Analysis*

The cross-sectional morphologies of the sample films were observed with a scanning electron microscope (F16502, Phenom, Eindhoven, Netherlands) at 5 kV. The cross-sections of the sample films were exposed by fracturing the films in liquid nitrogen and sprayed with a thin layer of gold under vacuum.

#### *2.7. FTIR Spectrum*

The chemical structures of sample films were characterized using a Fourier-transform infrared spectrometer (TENSEOR 27, Bruker, Ettlingen, Germany) over a range of 400–4000 cm−<sup>1</sup> that was operating in attenuated total reflection (ATR) mode and with a resolution of 4 cm<sup>−</sup>1.

#### *2.8. X-ray Di*ff*raction (XRD)*

X-ray diffraction (XRD) spectra of the sample films were measured by an X-ray diffractometer (MiniFlex600, Rigaku Corporation, Tokyo, Japan) with Cu Kα radiation (λ = 0.15418 nm) that was generated at 40 kV and 30 mA. The diffraction patterns of the films were recorded over an angular range of 2θ = 5◦–50◦ at a constant rate of 5◦/min.

#### *2.9. Thermal Stability*

The thermal stability of sample films was measured using a thermal gravimetric (TG) analyzer (STA449F5, NETZSCH, Bayern, Germany) in a nitrogen atmosphere and the films were heated from 30 to 600 ◦C at a heating rate of 10 ◦C/min and nitrogen flow rate of 20 mL/min.

#### *2.10. Mechanical Property*

Sample film thicknesses were measured using a digital micrometer (model 11,248–001, TMI, New Castle, DE, USA). Elongations at the breaking and tensile strengths of the sample films were determined by an electronic universal material testing machine (MODEL 3367, Instron, MA, USA). The films were cut into strips (100 mm × 15 mm). Stretching rates and initial grip separations were set to 10 mm/min and 50 mm, respectively.

#### *2.11. Light Transmittance*

The light transmittance of the developed biocomposite films was measured using a UV-visible spectrophotometer (Specord 50 Plus, Analytik Jena, Jena, Germany) in a wavelength range from 380 to 800 nm. An empty quartz cuvette was used as the blank. Each film sample was cut to 9 mm × 40 mm and was attached to the wall of the cuvette before measurement.

#### *2.12. Barrier Properties*

The oxygen transmission rate (OTR) values of the developed biocomposite films were measured by an automated oxygen permeability testing instrument (OX-TRAN 2/21, MOCON, Inc., Minneapolis, MN, USA) with a coulometric oxygen sensor method and followed the ASTM D3985 standard [28]. OTR is the volume of permeant oxygen passing through a film per unit surface area and time under equilibrium with testing conditions, and the unit of OTR was expressed as cc/m2·day [29]. The test area of the samples was 5 cm<sup>2</sup> and the tests were performed at 23 ◦C and 50% RH. The test gas was oxygen with a flow rate of 20 mL/min while a mixture of nitrogen (98%) and hydrogen (2%) was used as the carrier gas with a flow rate of 10 mL/min. The test mode was convergence by cycles.

The water vapor transmission rate (WVTR) of the developed biocomposite films was measured using a water vapor permeability testing instrument (TSY-T1, Labthink, Jinan, China). WVTR is the weight of permeant moisture passing through a film per unit surface area and time under equilibrium with testing conditions, and the unit of WVTR was expressed as g/m2·day [29].The film was cut into <sup>∅</sup>100 mm circular pieces and was placed onto a permeability cup with a 63.58 cm<sup>2</sup> testing area. The cup was previously filled with 10 mL of distilled water (RH 100%). The cup was sealed and placed into the dry chamber of the instrument at 38 ± 0.6 ◦C and 10% RH. The sealed cup was weighed periodically (0.001 g) until testing was complete. Water vapor amounts transported into the dry chamber were determined by the weight loss of the cup.

#### *2.13. Statistical Analysis*

The data were reported as mean ± standard deviation and analyzed by one-way analysis of variance (ANOVA) and Duncan's multiple range tests using the SPSS 22.0 statistical package for Windows (IBM SPSS Statistical software, Inc., Chicago, IL, USA). The significance level was always set to *p* < 0.05.

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

#### *3.1. Characteristic of CNFs*

The morphology of the CNFs is shown in Figure 1. The TEM image shows that the lengths of the prepared CNFs ranged from 200 nm to several microns, diameters ranged from 20 to 50 nm, length–diameter ratios were greater than 50, and there were intertwinements between the long fibrils [30].

#### *3.2. Antimicrobial Properties*

To determine the optimal addition amount of OEO in the following experiments, 1%, 2% and 3% OEO were added to the chitosan/OEO films. The antimicrobial activities of the chitosan/OEO films with different amounts of OEO against *E. coli* (Gram-negative) and *L. monocytogenes* (Gram-negative) were evaluated and the results are shown in Figures 2 and 3. As is shown in Figure 2, there was no clear inhibition zone either on *E. coli* or on *L. monocytogenes* around the pure chitosan film (0% OEO). With increased OEO addition amounts, the areas of the inhibition zones against both *E. coli* and *L. monocytogenes* gradually increased. The antibacterial properties of the chitosan/OEO films

significantly improved by adding OEO and demonstrated that OEO had good inhibition properties against *E. coli* and *L. monocytogenes*. As shown in Figure 3, it is worth noting that the inhibition rates against *E. coli* and *L. monocytogenes* of pure chitosan film (0% OEO) reached 40% and 43%, respectively. This suggested that pure chitosan film exhibited certain antibacterial properties toward *E. coli* and *L. monocytogenes,* but that the antibacterial properties were not obvious. However, the poor antibacterial properties of pure chitosan film cannot meet the requirements for packaging materials for some perishable foods. When the addition amounts of OEO were 2% and 3%, the growth inhibition rates of the developed films against *E. coli* and *L. monocytogenes* reached 99%. However, essential oils usually have a strong pungent, are volatile and excessive levels of essential oils in food packaging as antibacterial agents may affect the original food flavor [31]. Therefore, the optimal addition amount of OEO was chosen to be 2% to ensure a high antibacterial rate. The excellent antibacterial properties indicated that chitosan/OEO biocomposite films have the potential to be used as antimicrobial packaging materials to extend food shelf life.

**Figure 1.** Transmission electron microscopy (TEM) image of cellulose nanofibrils (CNFs).

**Figure 2.** Area of inhibition zone of the films with the different addition amount of oregano essential oil (OEO) against *E. coli* and *L. monocytogenes*.

**Figure 3.** Growth inhibition rate of the films with the different addition amount of OEO against *E. coli* and *L. monocytogenes*.

#### *3.3. SEM Analysis*

SEM was used to observe the microstructures of the developed chitosan/OEO/CNFs films to analyze the influence of adding CNFs on the film morphologies. Figure 4 shows the cross-sectional morphologies of the COC0, COC20, COC40 and COC60 films. The COC0 film showed a tight and homogenous structure and few pores may be caused by the volatilization of the OEO [32]. The fibrous-network structure of the films became more obvious with increased addition amounts of CNFs. This behavior was due to the hydroxyl group on the CNFs chains through hydrogen bonding interactions with chitosan. Moreover, the fibers overlapped each other and formed a dense three-dimensional network structure.

**Figure 4.** SEM images of the cross-section of the (**a**) chitosan/OEO/0% CNFs (COC0) film; (**b**) chitosan/ OEO/20% CNFs (COC20) film; (**c**) chitosan/OEO/40% CNFs (COC40) film and (**d**) chitosan/OEO/60% CNFs (COC60) film.

#### *3.4. FTIR Spectrum*

FTIR spectra are widely used to analyze changes in chemical structure and components of co-composites. The FTIR spectra of all sample films with different addition amounts of CNFs are shown in Figure 5. The characteristic peaks at 1629, 1543 and 1411 cm−<sup>1</sup> were assigned to C=O stretching (amide I), N–H bending (amide II) and C–N stretching (amide III), respectively [33]. These are the characteristic peaks of chitosan which appeared in all spectra of all films and confirmed that chitosan was the matrix material for all films. In the spectra of the COC0 films, the broad peak at 3276 cm−<sup>1</sup> was attributed to O–H and N–H stretching of chitosan. After the CNFs were incorporated with the chitosan, the position of the broad peak of the COC20, COC40 and COC60 films was shifted to around 3341 cm−<sup>1</sup> which was due to overlapping of the O–H bonds in both CNFs and chitosan. The peak at 1070 cm−<sup>1</sup> was related to the C–O–C stretching vibration of chitosan in the COC0 film spectra [34]. However, the C–O–C stretching vibration of the films containing CNFs was shifted to around 1059 cm−<sup>1</sup> which was due to the overlap of the C–O–C bonds in both CNFs and chitosan [35]. These results demonstrated that there are strong hydrogen bonds between CNFs and chitosan in the molecular chain. The peaks in the region of 2921 and 2863 cm−<sup>1</sup> were attributed to symmetric and asymmetric methylene stretching vibrations, respectively. Wu et al. [36] also found that the peaks at 2928 and 2864 cm−<sup>1</sup> became stronger in a gelatine–chitosan film with 4% OEO. These results indicated that OEO was successfully introduced into the films.

**Figure 5.** FTIR spectrum of the COC0 film, COC20 film, COC40 film and COC60 film; (COC0: chitosan/OEO/0% CNFs; COC20: chitosan/OEO/20% CNFs; COC40: chitosan/OEO/40% CNFs; COC60: chitosan/OEO/60% CNFs).

#### *3.5. X-ray Di*ff*raction (XRD)*

The XRD patterns of the chitosan/OEO/CNFs films with different added amounts of CNFs are shown in Figure 6. Soni et al. [37] reported that the characteristic peaks for pure chitosan films were near 2θ = 9.77◦ and 19.88◦. However, the positions of these characteristic peaks of chitosan-based films with OEO were slightly shifted (e.g., 2θ = 8.48◦ and 18.32◦) which indicated that the original crystalline structure of the chitosan was destroyed [38]. After the addition of CNFs, the characteristic peaks of the CNFs appeared in the region of 2θ = 16.5◦ and 22◦. The strengths of these peaks and the crystallinity of the films slightly increased with increased CNFs content and could be due to the ordered accumulation of chitosan chains on the surface of the crystalline domains of the CNFs [39]. Fernandes et al. [39] also reported that increased bacterial cellulose contents promoted crystallization of chitosan chains as observed in the diffractograms of water soluble chitosan/bacterial cellulose nanocomposite films. The strength of the crystalline peak increased with increased CNFs content which was due to the high biocompatibility between chitosan and cellulose [40].

**Figure 6.** X-ray diffraction patterns of the COC0 film, COC20 film, COC40 film and COC60 film; (COC0: chitosan/OEO/0% CNFs; COC20: chitosan/OEO/20% CNFs; COC40: chitosan/OEO/40% CNFs; COC60: chitosan/OEO/60% CNFs).

#### *3.6. Thermal Stability*

The thermal stabilities of the sample films were examined by TG to evaluate the effect of CNFs addition on the thermal degradation behavior of the films. Figure 7 shows the TG curves of the COC0, COC20, COC40 and COC60 films. The TG curves of all sample films showed the first stage of weight loss occurred between 90 and 250 ◦C which was associated with water evaporation. Similarly, the weight loss of cassara starch/chitosan/gallic acid films reinforced by CNFs occurred in the range of 90–225 ◦C and was also due to the weight loss of the absorbed moisture in the films [41]. The second degradation stage consisted of the disaggregation of chitosan molecules or/and disaggregation of cellulose chains which occurred between 250–340 ◦C. The third stage, between 340 and 500 ◦C, was due to oxidation of the char or/and breakdown of glucose units in CNFs. All sample films showed similar thermal behavior between 250 and 500 ◦C and therefore, the effect of CNFs on the thermal stability of chitosan/OEO/CNFs biocomposite films was insignificant [3]. However, the total residues of the COC60 films at 600 ◦C were the highest which indicated that 60% addition of CNFs reduced the rate of char oxidation and shifted the char oxidation to higher temperatures [42]. Therefore, the COC60 films are more suitable for food packaging applications even when used at a relatively high temperature.

**Figure 7.** Thermal gravimetric (TG) curve of the COC0 film, COC20 film, COC40 film and COC60 film; (COC0: chitosan/OEO/0% CNFs; COC20: chitosan/OEO/20% CNFs; COC40: chitosan/OEO/40% CNFs; COC60: chitosan/OEO/60% CNFs).

#### *3.7. Mechanical Property*

Tensile strength (TS) and elongation at break (EB) are fundamental properties for food-packaging films to resist the stresses and strains that the material may endure during food storage and transportation. Xu et al. [43] shown that the carboxylated CNF significantly enhanced the tensile strength of plasticized hemicelluloses/chitosan-based edible films. The mechanical properties of the COC0, COC20, COC40 and COC60 films are shown in Table 1. The TS of the COC0 film was determined to be 7.71 MPa and the TS of the COC20, COC40 and COC60 films increased to 10.24, 13.79 and 16.80 MPa, respectively (*p* < 0.05). The high TS of the films containing CNFs may result from the large aspect ratio of CNFs [43] and the stronger interfacial interaction between the chitosan and chains of CNFs [22]. Moreover, the result was also related to the mentioned in Section 3.5, the strength of crystalline peak increased in the chitosan amorphous matrix after addition of CNFs [39]. This may be because the intermolecular hydrogen bonding of chitosan was replaced by the new, strong hydrogen bonding between the hydroxyl groups in the CNFs and the hydroxyl groups in chitosan. Therefore, CNFs can be used as a good filler to enhance the mechanical strength of chitosan films. Khan et al. [3] observed a decrease in the EB values of chitosan films from 8.58% to 6.28% due to the addition of nanocrystal cellulose (NCC). In this work, the EB value was determined to be 31.31% for the COC0 film. Compared to the COC0 film, the EB was significantly reduced by 5.14%, 5.63% and 4.48% for the COC20, COC40 and COC60 films, respectively (*p* < 0.05). These results are attributed to the strong hydrogen bonding and electrostatic interactions between CNFs and the chitosan matrix [37].

**Table 1.** Mechanical property of the COC0 film, COC20 film, COC40 film and COC60 film; (COC0: chitosan/OEO/0% CNFs; COC20: chitosan/OEO/20% CNFs; COC40: chitosan/OEO/40% CNFs; COC60: chitosan/OEO/60% CNFs).


#### *3.8. Optical Properties*

The transparency of packaging films is important because light can lead to oxidation of nutrients including vitamins, fats and oils and can affect food quality. At the same time, packaging materials also need to have a certain amount of light transmittance to enable consumers to view the packaged products. The light transmittance spectra of the sample films are shown in Figure 8. The light transmittance of the COC0 film was the highest of all films which suggested that the light barrier effect of the COC0 film was poor. In addition, the light transmittance of the COC0 film was 39.73% at 600 nm (center of visible light spectrum) which was lower than most pure chitosan films mentioned in other research [34]; this indicated that the presence of OEO reduced the transmittance of the films [43]. With increasing CNFs contents, the light transmittance of the films decreased, and the opacity increased. This indicated that the addition of CNFs decreased the transparency of the films. The addition of 60% CNFs improved the light barrier effect the COC60 film which had the lowest light transmittance (28.53%) at 600 nm. These results suggested that CNFs were densely packed in the chitosan matrix and with compact lap between the fibers, light scattering was prevented by the small interstices between the fibers [44]. All of the results implied that the COC60 film has good prospects for food packaging because it has excellent shading properties.

**Figure 8.** Light transmittance of the COC0 film, COC20 film, COC40 film and COC60 film; (COC0: chitosan/OEO/0% CNFs; COC20: chitosan/OEO/20% CNFs; COC40: chitosan/OEO/40% CNFs; COC60: chitosan/OEO/60% CNFs).

#### *3.9. Barrier Properties*

Oxygen and water vapor are the important environmental factors that cause spoilage and deterioration of food during storage. Hence, there is concern about the barrier properties of oxygen and water vapor in food packaging materials. Figure 9 shows the barrier properties of the COC0, COC20, COC40 and COC60 films. As shown in Figure 9, the WVTR of the COC20, COC40 and COC60 films significant (*p* < 0.05) decreased when compared to the COC0 films (861.26 g/m2·day). The reduction in WVTR was due to the physicochemical interactions between CNFs and chitosan which led to reduced numbers of hydrophilic groups (–OH) [8]. As mentioned in Section 3.3, there was good biocompatibility between CNFs and the chitosan matrix and a three-dimensional network structure formed between the fibers by producing winding paths for the water vapor molecules and thus led to reduction of WVTR [22].

**Figure 9.** Oxygen transmission rate and water vapor transmission rate of the COC0 film, COC20 film, COC40 film and COC60 film; (COC0: chitosan/OEO/0% CNFs; COC20: chitosan/OEO/20% CNFs; COC40: chitosan/OEO/40% CNFs; COC60: chitosan/OEO/60% CNFs).

If the OTR value is in the region of 1–10 cc/m2·day, the packaging material is considered to have good oxygen barrier performance [45]. All sample films in this study had OTR < 10 cc/m2·day. However, the OTR of the COC20 films was 5.97 cc/m2·day which were lower than that of the COC0 film (8.64 cc/m2·day). Their decreased oxygen permeability could be due to the presence of a more tortuous path between the fibers for penetration by oxygen molecules [46]. Compared with the OTRs of the COC40 films (5.94 cc/m2·day) and COC60 films (6.03 cc/m2·day) there were comparable. Oxygen permeability was related to the addition of CNFs and was not affected by the CNFs content. Overall, the addition of CNFs to chitosan films shows good oxygen barrier performance and thus indicates that these composite films can be used as barrier packaging for food.

#### **4. Conclusions**

In this study, a novel biocomposite packaging film with good antibacterial activities in addition to good mechanical and barrier properties was successfully developed based on chitosan as the film matrix, CNFs as a reinforcing filler and OEO as an antibacterial agent. The chitosan film, which contained 2% OEO, exhibited significant antimicrobial activity and its growth inhibition rates against *L. monocytogenes* and *E. coli* reached 99%. The fibers overlapped with each other in the chitosan matrix to form a dense three-dimensional network structure that was observed by SEM. The FTIR spectrum showed that there were strong hydrogen bonds between CNFs and chitosan in the molecular chain. The TS of the chitosan/OEO film increased with the addition of CNFs. CNFs improved the barrier performance of the chitosan/OEO film to light, oxygen and water vapor by reducing light transmittance, oxygen permeability and water vapor permeability. However, the effect of CNFs on the thermal stability of the chitosan/OEO film was insignificant. The developed chitosan/oregano essential oil/CNFs biocomposite film can be used as an antibacterial and barrier materials in the field of food packaging. It has the potential to improve food quality and extend food shelf life.

**Author Contributions:** Conceptualization, M.W., P.L. and S.W.; methodology, S.C. and C.W.; formal analysis, S.C.; investigation, S.C., C.W. and S.Y.; data curation, M.W. and P.L.; writing—original draft preparation, S.C.; writing—review and editing, M.W. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Natural Science Foundation of Guangxi (No. 2019GXNSFAA185002) and by the Dean Project of Guangxi Key Laboratory of Clean Pulp & Papermaking and Pollution Control (ZR201708).

**Conflicts of Interest:** The authors declare no conflicts 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*
