**1. Introduction**

Food packaging materials based on cellulosic fibers must keep their functionality under permanently changing conditions in the surrounding environment, such as temperature, storage time, or moisture, which are major influences on the shelf-life and quality of the packed food [1–3]. Paper-based packaging assures the strength and stability of the packaging, but due to its porous structure, paper lacks most of the important barrier functions needed nowadays. Furthermore, paperboard produced from so-called recycled or secondary cellulosic fibers can contain residues of mineral oils, better known as mineral oil saturated hydrocarbons (MOSH) and mineral oil aromatic hydrocarbons (MOAH), which represent a serious source of contaminants for packed food products. The sources of mineral oil residues in secondary fibers are ubiquitous, including mineral-oil-based printing inks in particular [4–8]. In order to control and prevent the permeation or migration of water vapor, mineral oils, grease, or liquids, paper must be further upgraded by a

suitable barrier coating, in order to ensure the required packaging function. Therefore, the surface of paper-based packaging materials is treated either by extrusion, using thermoplastic petroleum-based polymers, or by dispersion coating, using synthetic water-based polymer dispersions [9–11]. In recent years, significant research efforts in academia, as well as in industry, focused on the replacement of oil-based polymer materials in the surface treatment of paperboard. Driving forces for these developments are not necessarily only coming from the producers, but also from the consumers [2]. Bio-based materials applied on paper could provide interesting barrier functionalities while still maintaining the environmentally friendly characteristics of the packaging material. The challenge with paper-based packaging materials is that, for different products, different barrier properties are needed. Multiple layers of barrier materials are sometimes the solution chosen in practice [3,12,13].

Our investigation focused on sodium alginate and chitosan, and their application as barrier materials for paper intended to be used as primary or secondary food-packaging materials. Alginate is a polysaccharide naturally present in brown algae, and is usually available as salts of sodium and calcium. Alginates and its derivatives are already used in large amounts in the food industry as additives, and therefore, are also considered to be safe for their use as functional barriers for food-contact materials. Various water-soluble alginate formulations are available on the market, which can be applied with conventional coating equipment used in the paper and packaging industries [14–16]. Chitosan is an abundant, natural polysaccharide derived from chitin, a substance in the exoskeletons of crustaceans and insects. Economically interesting quantities are already produced from fishing industry waste, mainly during the processing of crabs and shrimps [17–21].

Our work investigated the coatability and barrier properties of these two water-soluble biopolymers, chitosan and alginate, from renewable resources. Both were applied under the same conditions onto two different paper grades, with the aim of evaluating their potential for reducing the migration and permeation of mineral oil components (MOSH/MOAH) [11], aromatic components, and water vapor. Furthermore, the coating layer's quality was analyzed via scanning electron microscopy, and its resistance toward grease, water absorption, and air permeability was determined. Similar studies with chitosan and alginate were reported in the literature for some of above-stated barrier functions [22–30]. The novel aspect of this work involves the systematic comparison and quantitative study of the barrier properties of alginate and chitosan, and their interaction with two different industrially produced paper substrates. In particular, the effect of these two bio-based barrier materials against the migration of mineral oil fractions (MOSH and MOAH) contained in paper, measured and quantified with HPLC–GC coupled with a flame ionization detector (FID), is yet to be reported.

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

#### *2.1. Coating Materials*

Powdered chitosan used in the preparation of the coating solution was kindly supplied by BioLog Heppe GmbH (Landsberg, Germany). This industrially produced chitosan with a degree of deacetylation 88%–95% was made from the carapace skin of crustaceans. According to the product specifications, the chitosan powder consisted of particles with a diameter ≤200 μm, and an ash content <1% (*w*/*w*). The dynamic viscosity of a 1% (*w*/*w*) chitosan aqueous solution dissolved in 1% (*w*/*w*) acetic acid at pH 4 and 20 ◦C was 20 mPa·s. Acetic acid (100%, Rotipuran) used for adjustment of the pH of dissolved chitosan in water was purchased from Carl Roth GmbH+ Co. KG (Karlsruhe, Germany). Sodium alginate (viscosity 15–25 mPa·s 1% (*w*/*w*) in water at 25 ◦C) was purchased from Sigma-Aldrich (Saint Louis, MO, USA).

#### *2.2. Preparation of Aqueous Coating Solutions*

A chitosan coating solution with a solid content of 4% (*w*/*w*) was prepared by dissolving it in heated deionized water (70 ◦C), adding the chitosan powder in small amounts and stirring for 6 h at 400 rpm. Subsequently, acetic acid was added in small portions in order to achieve a pH of 4, measured constantly by a portable pH meter (inoLab pH 7110, WTW, Weilheim, Germany). This chitosan aqueous solution was stirred and heated for 4 h at 70 ◦C until a yellow solution was obtained and no visible particles were observed.

A sodium-alginate coating solution with a solid content of 4% (*w*/*w*) was prepared by dissolving sodium-alginate powder in deionized water at a neutral pH. The sodium-alginate powder was added to water in portions, stirred at 400 rpm, and the aqueous solution was heated for 6 h at 75 ◦C. After this time, sodium alginate was completely dissolved, resulting in a homogenous coating solution. Due to the heating and evaporation of water, the solid contents of both coating solutions slightly increased. Therefore, the solid content was remeasured using a moisture analyzer (HR73, Mettler Toledo, Columbus, OH, USA), and adjusted to 4% (*w*/*w*) with deionized water. Finally, the coating solutions with the adjusted and desired solid contents were cooled to room temperature. The viscosities (Brookfield II+, at 50 rpm, *n* = 3) of 4% (*w*/*w*) chitosan and 4% (*w*/*w*) alginate coating solutions, measured at room temperature, were 2911 mPa·s ± 57 and 1448 mPa·s ± 20, respectively.

*2.3. Paper-Substrate Characterization*

Two different commercial paper grades were used in the coating trials. The first was a paper (PF) made from 100% primary-fiber furnish (mixture of hardwood and softwood), mass-sized using 100% active liquid alkenyl succinic anhydride (ASA), and surface-sized using starch and a calender machine. The second substrate was a paperboard (SF) made from 100% secondary or recycled fibers with no surface treatment. Prior to coating, the basic properties of the substrates were measured, and are summarized in Table 1 (*n* = 15).


**Table 1.** Basic characterization of substrates used for barrier coating (*n* = 15).

#### *2.4. Standardized Physical Paper Properties and Barrier Measurements*

Prior to the measurements, the raw (uncoated) paper substrates and paper samples coated with the alginate and chitosan formulations were conditioned for 48 hours at 23 ± 1 ◦C and 50 ± 3% relative humidity (RH) [31]. Measurements for grammage, thickness, density, roughness, air permeability, water-vapor transmission rate, water absorptiveness, contact angle, and grease resistance were performed according to the standardized methods listed in Table 2.

#### *2.5. Coating Trial with Laboratory Draw-Down Coater*

A coating trial was performed using a laboratory draw-down coater from RK Printcoat Instruments Ltd. (Litlington, UK). A target coat weight of 6 g·m−<sup>2</sup> (single-sided application), with a standard deviation of less than 10%, was achieved by applying two layers of barrier-coating solution. The coater speed was 4 m·min−1, and the wet-film thickness for the first and second coating layers, defined by the wire-wounded rod used, was 40 μm for both coating solutions. Drying of the coated paper samples was performed with hot air at 150 ◦C for 60 s.


**Table 2.** Paper properties and standard methods used for the testing of uncoated and coated samples.

#### *2.6. Surface Evaluation of Uncoated and Coated Paper Substrates*

The surface topography of uncoated raw paper, and chitosan- and alginate-coated paper was investigated using low-voltage scanning electron microscopy (LVSEM, Everhart-Thornley detector for the detection of secondary electrons; Zeiss Sigma 300, Oberkochen, Germany) [40]. The samples were cut (1 cm × 1 cm), then attached to SEM stubs using a double-sided conductive carbon tape, and imaging (magnification 500×) was performed at an acceleration voltage of 0.65 kV.

#### *2.7. Migration Experiments*

In this study, migration experiments were performed according to EU-Regulation No. 10/2011 [41]. As a food simulant, Tenax®, a poly (2,6-diphenyl-p-phenylene oxide) (Tenax® TA (refined), 60–80 mesh; SUPELCO, Bellefonte, PA, USA), was used for the simulation of dry foods such as rice, cereals, cocoa, coffee, and spices. The standard test conditions for long-term storage of these products for above and below six months at room temperature should be 60 ◦C for 10 days, but can be adapted using the Arrhenius equation. This was done, with conditions tested and set as 80 ◦C for two days. The Tenax® was applied in an amount of 4 g·dm−2. The experiments were performed in triplicate in migration cells (MigraCell®; FABES Forschungs-GmbH, Munich, Germany) with a tested surface area of 0.32 dm2. The cell was assembled according to the manufacturer's instructions, with the coated side facing the Tenax®, and placed in an oven for two days at 80 ◦C. Afterward, the Tenax® was drained into a glass vial with a screw cap, and 25 μL of an internal standard mix was added. The internal standard for migration experiments consisted of dodecane-d26 (C12D26; EURISO-TOP SAS, Saint-Aubin, France), nonadecane-d40 (C19D40; 98%; Cambridge Isotope Laboratories, Inc.; Tewksbury, MA, USA), benzophenone-d10 (C13D10O; 99 at.%; Sigma-Aldrich Co., St. Louis, MO, USA), and bis(2-ethylhexyl)phthalate-d4 and di-n-butyl phthalate-d4 (both "analytical standard", purchased from Sigma-Aldrich Co., St. Louis, MO, USA). All were used at a concentration of 200 mg·L−<sup>1</sup> in acetone (ROTISOLV® ≥99.9%, UV/IR-Grade; Carl Roth GmbH + Co. KG, Karlsruhe, Germany). The Tenax® was extracted three times with 10 mL of n-hexane (Picograde® for residue analysis; LGC Promochem GmbH; Wesel, Germany) and three minutes of vortexing. The extracts were combined through a folded filter in a 50-mL evaporation vial, and the solvent was evaporated to 0.5 mL in an automatic solvent evaporator (TurboVap® II; Biotage, Uppsala, Sweden). The extracts were then transferred into 1.5-mL glass vials with screw caps; the evaporation vials were rinsed with 0.5 mL of hexane, and this solvent added to the 1.5-mL vials. The extracts were stored in a refrigerator, and only a small amount was filled into a 1.5-mL glass vial with a micro insert and screw cap for measurements. The extracts were measured on a gas chromatograph with a flame ionization detector (GC–FID) to determine the overall migration. The separation was done using a Hewlett Packard 6890 Series GC System equipped with an Optima delta-6 capillary column (7.5 m × 100 μm × 0.10 μm, Macherey-Nagel, Germany). The oven was programmed to 60 ◦C (hold 1 min), then raised at <sup>15</sup> ◦C·min−<sup>1</sup> to 300 ◦C (3 min). The carrier gas used was hydrogen with a linear velocity of 48 cm·s<sup>−</sup>1. Aliquots of one microliter were injected with a split of 1:20. The injection-port temperature and

detector temperature were set to 280 and 320 ◦C, respectively. Data evaluation was done using the "GC ChemStation" software, version B.04.03 (Agilent Technologies, Santa Clara, CA, USA).

For the analysis of MOSH and MOAH, online-coupled HPLC–GC–FID was used as described in [7,8]. Prior to analysis, a MOSH/MOAH internal standard mix was added in a concentration of 1.5–6 <sup>μ</sup>g·mL<sup>−</sup>1. The standard purchased by Restek Corporation (Bellefonte, PA, USA) contained the following substances in 1-mL ampoules in toluene: n-undecane (300 <sup>μ</sup>g·mL−1), n-tridecane (150 <sup>μ</sup>g·mL−1), bicyclohexyl (300 <sup>μ</sup>g·mL−1), cholestane (5-α-cholestane; 600 <sup>μ</sup>g·mL−1), 1-methylnaphthalene (300 <sup>μ</sup>g·mL−1), 2-methylnaphthalene (300 <sup>μ</sup>g·mL−1), n-pentylbenzene (300 <sup>μ</sup>g·mL−1), perylene (600 <sup>μ</sup>g·mL−1), and 1,3,5-tri-tert-butylbenzene (300 <sup>μ</sup>g·mL−1). For the calculation of retention indices and the determination of cutting fractions in HPLC–GC–FID, a "C7–C40 saturated alkane standard" from SUPELCO (Bellefonte, PA, USA) was used. The concentration of the alkanes was 1000 <sup>μ</sup>g·mL−<sup>1</sup> in hexane, and was diluted to 1 <sup>μ</sup>g·mL−<sup>1</sup> with hexane prior to analysis.

The HPLC used was a Shimadzu LC-20AD (Shimadzu Corporation, Kyoto, Japan) equipped with an Allure Silica 5 μm column (250 mm × 2.1 mm). A gradient elution was used, starting with 100% n-hexane (flow 0.3 mL·min−1), before being raised to 35% CH2Cl2 within 2 min (hold for 4.20 min). The column was then backflushed at 6.30 min with 100% CH2Cl2 (flow 0.5 mL·min−1; hold for 9 min), and reconditioned to 100% n-hexane (flow 0.5 mL·min−1; hold for 10 min). The flow was subsequently decreased to 0.3 mL·min−<sup>1</sup> until the next injection. The UV-detector was equipped with a D2-lamp set at 230 nm and a cell temperature of 40 ◦C. The GC was a Shimadzu GC 2010 dual-FID (Shimadzu Corporation, Kyoto, Japan), equipped with two guard columns, Restek MXT Siltek (10 m × 0.53 mm inner diameter (id)), and two analytical columns Restek MTX®-1 (15 m × 0.25 mm id × 0.1 μm df). The carrier gas used was hydrogen with an analysis pressure of 150 kPa, and an evaporation pressure of 87 kPa for MOSH and 85 kPa for MOAH. The oven was programmed to 60 ◦C (hold 6 min), and raised at 20 ◦C·min−<sup>1</sup> to 100 ◦C (0 min) followed by 35 ◦C·min−<sup>1</sup> to 370 ◦C (9.29 min). The LC–GC interface was controlled by a Chronect-LC–GC by Axel-Semrau (Sprockhövel, Germany); data evaluation was done using the LabSolutions software version 5.92. According to a proposed method published by the German Bundesinstitut für Risikobewertung (BfR), quantification was done by integration of the hump for various molecular weight regions. They propose the ranges of C16–C25 and C25–C35 for food-contact materials for dry non-fatty food and storage at room temperature [5].

#### *2.8. Permeation Experiments*

The used migration cell allowed a one-sided migration experiment (as described in Section 2.7), and a two-sided application for simultaneously testing the migration and permeation. When performing a two-sided test in the migration cell, three changes of the experimental set-up were made. Firstly, a piece of cellulose was placed at the bottom of the cell where the modeling substances for the permeation were spiked. Secondly, the metal plate in the middle of the cell was removed. Therefore, thirdly, the colorless silicone ring had to be replaced by a FEP (Fluorinated ethylene propylene)-coated red ring to prevent the contamination of the Tenax® with siloxanes. As modeling substances, deuterated n-alkanes of various chain lengths were chosen because they best simulated a possible migration of mineral oil hydrocarbons through the sample (C14D30, C20D42, and C24D50, 98%-at.%D, purchased from Cambridge Isotope Laboratories, Inc. (Tewksbury, MA, USA); C16D34, 99%-at.%D, purchased from abcr GmbH (Karlsruhe, Germany); and C28D58 98%-at.%D, purchased from C/D/N/ Isotopes, Inc. (Pointe-Clair, QC, Canada)). The deuterated substances were used to prevent the interference of permeation and migration tests, because these n-alkanes were also present in the tested paper samples. To simulate aromatic permeability, a set of four aromatic compounds were selected (DL-Menthol, ≥95%; Eugenol, ReagentPlus®, 99%; Vanillin, ≥97%; and Acetovanillone, ≥98%; purchased from Sigma-Aldrich Co. (St. Louis, MO, USA)). One hundred microliters of a stock solution in acetone containing each of the mentioned substances in a concentration of 100 mg·L−<sup>1</sup> were spiked into the bottom of the cell. The test conditions, extraction, and analysis stayed the same as described above.

The methods described were used to test the barrier efficiency of the two uncoated papers, and the papers coated with alginate and chitosan.

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

#### *3.1. Physical Characterization of Coated Samples*

The substrates made out of primary fibers (PF) and coated with chitosan or alginate were labeled as PF chitosan or PF alginate, respectively. For the secondary-fiber (SF) substrates, this principle resulted in the sample descriptions, SF chitosan and SF alginate. The average values of thickness, density, grammage, and coat weight, with their corresponding standard deviations are shown in Table 3.

**Table 3.** Thickness, density, grammage, and pick-up values of alginate- and chitosan-coated primary-fiber (PF) and secondary-fiber (SF) samples (*n* = 15).


#### *3.2. Surface Evaluation, Film Formation, and Coating Quality*

The topography of the coated and uncoated samples was assessed based on SEM images. The conventional technologies for the barrier coating of paper, such as extrusion or lamination, are based on the application of a specific polymer, and the formation of a distinct film is indispensable. Depending on the substrate and its specific physical properties, the amount of barrier coating sometimes exceeded 20 g·m−<sup>2</sup> or 30 g·m−2, in order to ensure good barrier efficiency [9]. In this respect, it was of interest whether bio-based coating materials must form a film on top of the paper surface in order to perform well as a barrier.

The SEM images of the uncoated raw substrates, SF and PF, showed the expected clear difference between the two different paper grades (Figure 1). Voids, and different sizes and alignments of fibers and pores between the fibers were visible, and the measured values for roughness and density (see Table 1) confirmed the difference in the structure of substrates produced from primary and secondary furnishes (Figure 1a,b).

Due to these differences between the paper substrates, it is obvious that chitosan and alginate interacted differently with the substrates, as illustrated by the SEM images. The PF substrate coated with chitosan had a completely covered surface, with no voids or pores visible, and a significant amount of the chitosan was also visible on the single fibers. In contrast, the SF chitosan sample showed that the fibers were not completely covered, and the coating appeared to have impregnated the material so that no clearly visible film was formed. A similar behavior was observed for the alginate coatings. Alginate also formed a film on the PF paper's surface, and covered the paper's surface completely, while it impregnated the SF paper. This can be explained by the higher roughness of the SF paper, resulting in impregnation rather than a full coverage of the paper's surface.

**Figure 1.** Scanning electron microscopy (SEM) images of uncoated and coated paper substrates at 500× magnification: (**a**) primary-fiber (PF) uncoated paper; (**b**) secondary-fiber (SF) uncoated paper; (**c**) chitosan-coated PF paper; (**d**) chitosan-coated SF paper; (**e**) alginate-coated PF paper; (**f**) alginate-coated SF paper.

#### *3.3. Air Permeability and Grease Resistance (KIT Test)*

Air permeability is a purely physical measurement, and gives the volume of ambient air able to pass through voids or pinholes through a paper substrate during a defined time period (one minute). The air permeability of the uncoated samples, SF and PF, was 809 mL·min−<sup>1</sup> and 437 mL·min−1, respectively (Figure 2).

After coating with alginate and chitosan, the measured air-permeability values for all samples were 0 mL·min−1. Low air permeability also indicates that no pinholes or voids are present in the coated paper.

*Coatings* **2018**, *8*, 235

The KIT test is a common method for the evaluation of fat and grease resistance of paper. The method is primarily designed to evaluate fluorochemical-based coatings for grease barriers, but was successfully applied to bio-based barrier coatings as well [42]. KIT solutions are numbered from 1 to 12, with higher numbers indicating higher grease resistance, and vice versa. The grease resistance of a coated packaging material depends on its surface chemistry (hydrophilic or hydrophobic character), the barrier quality, density, present pores and voids, as well as thickness of the substrate and barrier. KIT solutions are organic, non-polar compounds (castor oil, toluene, and n-heptane) with low density (<1 g·cm<sup>−</sup>3), able to penetrate easily through the porous structure of uncoated paperboard. In order to build a good barrier against grease, assessed with the KIT method, the barrier should, therefore, be hydrophilic rather than hydrophobic [43]. Alginate and chitosan manifest hydrophilic characteristics, and are able to close the voids and pores of the paper surface, thus meeting the initial criteria for a good grease barrier. Alginate applied on SF paper improved grease resistance to a medium level (KIT Number: 7.0 ± 1), which could already be of interest for some applications in the packaging industry. The PF substrate coated with alginate reached the maximum KIT number of 12.0 ± 0.5, and thus, is classified as an excellent barrier material against grease. Contrary to alginate, where the performance on SF and PF paper was significantly different, the chitosan barrier gave rather similar KIT values on both papers (6.0 ± 0.5 on PF, and 5.0 ± 0.5 on SF substrate). Although chitosan fully covered the fibers and closed the surface of the PF substrate, it did not reach such high KIT numbers when compared with alginate. The reason for this could be the distinctive hydrophilicity of the alginate, which is of course higher than that of chitosan, which may also have hydrophobic characteristics [44].

**Figure 2.** Air-permeability (*n* = 15) and grease-resistance values (KIT test, *n* = 9) for uncoated and coated primary-fiber (PF) and secondary-fiber (SF) paper.
