Improved flame retardancy and mechanical properties of bacterial cellulose fabrics via solvent exchange and entrapment of zein and gluten

This study aimed to improve the flame retardancy and mechanical properties of bacterial cellulose (BC) by introducing cereal proteins, namely zein and gluten. The production conditions were determined by observing residual masses of samples at 1000 ℃ using thermogravimetric analysis (TGA). According to the TGA results, the optimized production conditions for the BCs with zein and gluten were combined solvent exchange and entrapment of 20 weight% (wt.%) of zein, and entrapment of 40 wt.% of gluten, respectively. Surface characterization of BC prepared with zein and gluten under the optimal conditions confirmed that the cereal proteins were incorporated into the BC nanostructures via solvent exchange and/or entrapment and the original chemical and crystal structures of BC were not significantly changed. Limiting oxygen index (LOI) analysis confirmed that cereal proteins improved the flame retardancy of BC. In particular, the LOI of BC entrapped with gluten was 50%, which was better than that of cowhide leather. Char morphology analysis confirmed that the as-produced BCs with cereal proteins exhibited condensed-phase flame-retardant mechanism by forming intumescent chars. Analysis of the mechanical properties confirmed that compared with cowhide leather, as-produced BCs with cereal proteins possessed high tensile strength and dimensional stability, making them suitable leather substitutes.


Introduction
Bacterial cellulose (BC) is an environmentally-friendly cellulosic material produced by Gluconacetobacter xylinus and Acetobacter xylinus (Kim et al., 2020).BC has received attention as a sustainable fabric because the appearance of dry BC is similar to that of natural leather, and it also has excellent moldability and a controllable shape (Kim et al., 2021a).However, BC is highly combustible due to its cellulosic structure (Kim & Kim, 2023); thus, the flammability of BC may restrict its application as a natural leather substitute.Thus, to expand the application of BC, flame-retardant finishing is needed.
In this study, cereal proteins (zein and gluten) obtained from the agricultural and biofuel industries are selected as additives to impart flame retardancy to BC.Because these proteins are biomass materials, their utilization would reduce waste generation (Peydayesh et al., 2023).Zein is a corn storage protein with good biodegradability and barrier properties (Gonçalves et al., 2020;Yang et al., 2020).Gluten is a wheat protein with good thermal stability because of its high sulfur content (Yuan et al., 2010).
Therefore, the first aim of this study was to produce BC with enhanced flame retardancy using zein and gluten, which are inexpensive biomass raw materials.However, cereal proteins have poor solubility in water (Kasaai, 2018), making them difficult to incorporate into BC.Thus, the second aim of this study was to introduce methods of finishing BC using cereal proteins; this was achieved via solvent exchange and entrapment methods.The solvent exchange involves replacing the solvent in a compound while maintaining the dissolved molecules in the solution (Fazlollahi & Wankat, 2018).Entrapment involves physically enclosing molecules inside BC without changing its nanostructure (Kim & Kim, 2023).The solvent exchange and entrapment methods are both very simple and easily implemented, which shortens the fabrication process.
The novelty of this study lay in improving the flame retardancy and mechanical properties of BC using inexpensive and eco-friendly biomass materials through simple methods.After producing BCs with cereal proteins under optimized conditions, the surface characterization, flame retardancy, and mechanical properties are analyzed.

Materials
Glucose, hydrogen peroxide (H 2 O 2 ; 34.5%), sodium hydroxide (NaOH; pellets), and 1 N NaOH solution were obtained from Duksan Pure Chemicals (Ansan, South Korea), and a kombucha symbiotic culture of bacteria and yeast (SCOBY) was purchased from Getkombucha (Broomfield, CO, USA).Yeast extract and peptone were obtained from BD Biosciences (San Jose, CA, USA).Zein from corn, wheat gluten, and sodium dodecylbenzenesulfonate (C 18 N 29 NaO 3 S; SDBS) were supplied by Sigma-Aldrich (St. Louis, MO, USA).All the reagents were used as received without further purification.Untanned raw cowhide leather scrap was obtained from a local leather shop in South Korea.The thickness of the scraps was 0.7 ± 0.4 mm, which was similar to that of the BC samples.

Production and pretreatments of original BC
BC was produced by acetic acid bacteria within the kombucha SCOBY and was subjected to pretreatments including washing, bleaching, and swelling (Han et al., 2019;Song et al., 2017).The pre-treated BC is denoted as 'original BC. '

Production of BCs with cereal proteins
Herein, the BC sample produced via combined solvent exchange and entrapment using zein is denoted as 'BC-zein-SE, ' and the BC sample produced by gluten entrapment is named as 'BC-gluten-E' .BC-zein-SE was produced based on the method described by Wan et al. (2017) with some modifications.Briefly, original BC (wet state) was immersed in an 80% (v/v) ethanol solution at 25 ℃ for 24 h for solvent exchange of water to ethanol inside original BC.Subsequently, a zein solution was prepared by dissolving zein powder in a range of 0-60 wt.% relative to original BC in an 80% (v/v) ethanol solution at a BC:liquor ratio of 1:10 (w/v), and was ultrasonicated at 25 ℃ for 30 min.Thereafter, solvent-exchanged BC was immersed in the zein solution, followed by ultrasonication at 25 ℃ for 30 min, shaking at 80 rpm for 1 h in a shaking water bath at 30 ℃, and drying in a drying oven (OF-22G, JEIO TECH Co., Daejeon, South Korea) at 25 ℃ for 24 h to obtain BC-zein-SE.BC-gluten-E was produced based on the method of Kim et al. (2021b) with some modifications; an alkaline NaOH solution was used to dissolve gluten powders (Mathew et al., 2019).Briefly, a gluten solution was prepared by dissolving gluten powder in the range of 0-60 wt.% relative to original BC in NaOH solution (0.25 M) at a BC:liquor ratio of 1:10 (w/v), and was ultrasonicated at 25 ℃ for 30 min.Thereafter, the gluten solution was denatured at 80 ℃ for 20 min with shaking at 80 rpm in a shaking water bath and cooled to 25 ± 2 ℃.Original BC was then immersed in the solution, followed by ultrasonication at 25 ℃ for 30 min, shaking at 80 rpm for 1 h in a shaking water bath at 30 ℃, and drying in a drying oven at 25 ℃ for 24 h to obtain BC-gluten-E.

Characterization of BCs with cereal proteins
The FT-IR spectra of the samples were recorded using a Nicolet iS50 FT-IR spectrometer (Thermo Fisher Scientific, Waltham, MA, USA).Each spectrum was baseline-normalized using the OMNIC program (Thermo Fisher Scientific, Waltham, MA, USA).XRD patterns of the samples were acquired using a D8 ADVANCE X-ray diffractometer (Bruker, Billerica, MA, USA) with a Cu-Kα radiation source (λ = 0.15418 nm).The DIF-FRAC.SUITE ™ program (Bruker, Billerica, MA, USA) was used to perform background subtraction and smoothing.The degree of crystallinity is calculated as follows (Eq.1): where A cryst is the sum of the crystalline band areas and A total is the total area under the diffractogram (Adepu & Khandelwal, 2021).The surface and cross-sectional morphologies of samples were evaluated using a FE-SEM instrument (JSM-7800F, Prime, JEOL, Tokyo, Japan).All samples were sputter-coated with Pt using a magnetron sputter-coater (108auto, Cressington Scientific Instruments, Watford, UK).Along with FE-SEM, EDX was used to detect C, O, N, P, and S on the surface of the samples.

Flame-retardant properties of BCs with cereal proteins
To determine the residual masses of samples at 1000 ℃, thermogravimetric analysis (TGA) was performed at a heating rate of 10 ℃/min under N 2 atmosphere (flow rate of 60 mL/min) using a thermal analyzer (Discovery SDT 650, TGA Instruments, New Castle, DE, USA).
The LOI of each sample was measured using a LOI instrument (Fire Testing Technology, West Sussex, UK) according to the ISO 4589-2:2017 standard (Plastics-Determination of burning behavior of oxygen index-Part 2: Ambient temperature test).The dimensions of each sample were 80 mm × 10 mm × 1 mm.

Mechanical property analyses of BCs with cereal proteins
The tensile strength and elongation of the samples were examined according to the ASTM D882-18 standard (Standard Test Method for Tensile Properties of Thin Plastic Sheeting) using a universal testing machine (HZ-1007A, Dongguan Lixian Instrument Scientific, Dongguan, China).
The softness of each sample was evaluated according to the ISO 17235:2015 standard (Leather-Physical and mechanical tests-Determination of softness) using a digital leather softness tester (XB-OTS-TF115, Dongguan Xinbao Instrument, Dongguan, China).The crease recovery angles of the samples were measured according to the ISO 2313:1972 standard (Textiles-Determination of the Recovery from Creasing a Horizontally Folded Specimen of Fabric by Measuring the Angle of Recovery).
The dimensional stability of the samples was evaluated according to the ISO 7771:1985 standard (Textiles-Determination of Dimensional Changes of Fabrics Induced by Cold-Water Immersion) with slight modifications.The size of each sample was 50 × 100 mm, and the duration of the study was 60-180 min.

Optimization of conditions for producing BCs with cereal proteins
Generally, a higher residual mass of a sample after thermogravimetric analysis (TGA) is attributed to greater char formation (Zhu et al., 2004).Hence, the optimal conditions for producing BC using cereal proteins were determined based on the residual mass data collected from TGA at 1000 ℃.
As shown in Fig. 1, the residual masses of BC-zein-SE and BC-gluten-E (see Experimental) increased as the amounts of zein and gluten increased, and were highest with the incorporation of 20 and 40 wt.% of zein and gluten relative to BC, respectively.However, when zein and gluten contents exceeded the highest value, the residual masses of BC-zein-SE and BC-gluten-E decreased.As the amounts of cereal proteins increased, fewer hydroxyl groups in BC would be available for reaction (Liu et al., 2019).Hence, the excess protein molecules were aggregated on the surface of BC, thereby decreasing the thermal stability (Wang et al., 2016).Therefore, the optimal conditions for producing BCs with cereal proteins were as follows: (1) Production conditions for BC-zein-SE: combined solvent exchange and entrapment, 80% (v/v) ethanol-water solution, immersion for 24 h, and 20 wt.% zein relative to BC; (2) Production conditions of BC-gluten-E: entrapment, 0.25 M NaOH solution, 40 wt.% gluten relative to BC.

FT-IR analysis
FT-IR spectroscopy was used to examine the chemical interactions between BC and cereal proteins.The FT-IR spectra of all samples (Fig. 2) show characteristic peaks of BC at 3300-3500, 2920-2960, 1054, and 1030 cm −1 , confirming that the chemical structure of BC remained unchanged after entrapping cereal proteins (Kim & Kim, 2023).Moreover, the O-H stretching peak in the FT-IR spectrum of the original BC was shifted to lower wavenumbers after protein entrapment.The blue shifts may originate from hydrogen bond formation, indicating a chemical interaction between BC and cereal proteins (Lee et al., 2015).Conversely, a shift to a higher wavenumber was observed for the C-H stretching peak in the FT-IR spectrum of BC-zein-SE.The red shift is attributed to the loose hydrogen bonds between BC and zein (Kudo & Nakashima, 2020), suggesting that compared with zein, gluten is more strongly bound to BC.The FT-IR spectra of BC-zein-SE and BC-gluten-E show typical absorption bands of proteins in the amide I (1600-1700 cm −1 ), amide II (1500-1600 cm −1 ), and amide III (1240-1450 cm −1 ) regions, suggesting that the cereal proteins were entrapped within BC (Mejia et al., 2007;Peng et al., 2022;Xie et al., 2020;Zhang et al., 2022).Furthermore, the new peaks at 2330-2360 cm −1 and 1150-1160 cm −1 in the FT-IR spectra of BC-zein-SE and BC-gluten-E also indicate the presence of zein and gluten within BC, respectively (Avila et al., 2020;Hosseini et al., 2022;Mohammadia et al., 2018).Thus, it was confirmed that the cereal proteins formed hydrogen bonds with BC and the chemical structure of BC was retained after entrapping cereal proteins.

XRD analysis
XRD was used to analyze the crystalline structures of BCs with cereal proteins.The diffraction pattern of the original BC (Fig. 3) shows peaks at 2θ = 14.8°, 17.2°, 23.0°,  (French, 2014).These peaks were also observed in the diffraction patterns of BC-zein-SE and BC-gluten-E, confirming that the entrapment of cereal proteins did not change the crystal structure of BC.In the diffraction pattern of BC-zein-SE, new peaks appeared at 2θ = 9.0° and 18.5°, corresponding to the α-helix and β-sheet structures of zein, respectively (Martelli-Tosi et al., 2018).Moreover, a new peak also appeared at 2θ = 27.3°,which might be attributed to the ionic interaction between zein and BC (Xu et al., 2022).In the diffraction pattern of BC-gluten-E, new peaks appeared at 2θ = 27.3°,30.9°, 32.2°, and 39.4°, which might reflect the intermolecular interaction between gluten and BC (Abugoch et al., 2011).The diffraction pattern of BC-gluten-E also showed a peak at 2θ = 37.8°, corresponding to the crystalline domain of gliadin (Rani et al., 2021).Hence, it was confirmed that zein and gluten were entrapped within BC.Moreover, the crystallinity of BC increased from 85.1% to 92.7% and 87.7% after entrapping zein and gluten, respectively.The self-assembly nature of cereal proteins can increase the strength of hydrogen bonding between BC and proteins, thereby forming better-packed structures, leading to increased crystallinity (Keshk & Sameshima, 2006;Poletto et al., 2012).

SEM analysis
The surface morphology of BCs with cereal proteins was examined (Fig. 4).The SEM images of the original BC show interlaced cellulose nanofibers.The SEM images of BC-zein-SE show dense structures with spherical zein particles aggregated on the surface.The aggregation of zein may result from the noncovalent interactions with BC (Ding et al., 2022).In the SEM images of BC-gluten-E show thickened BC fibers, where the spaces between the fibers were filled, which might be attributed to the entrapped gluten (Lin et al., 2009).The cross-sectional SEM image of the original BC shows a multilayered structure formed through the layer-by-layer assembly of BC sheets during cultivation (Jiamsawat et al., 2022).The cross-sectional SEM images of BC-zein-SE and BC-gluten-E show densely packed structures, which might influence the tensile strength of the composites (Jebel & Almasi, 2016).

EDX analysis
EDX was used to analyze the surface elemental compositions of samples.C and O, which are the major elements constituting BC fibers, were found in the EDX spectra of all samples (Fig. 5 and Table 1).This observation is consistent with the unchanged chemical structure of BC (Villarreal-Soto et al., 2020).N and S, which are typical elements of proteins, were observed in the EDX spectra of BC-zein-SE and BC-gluten-E.The presence of these elements confirms the entrapment of cereal proteins (Ahmed & Rehman, 2020;Chen et al., 2022).In addition, P was detected only in the EDX spectrum of BC-zein-SE because of the different elemental compositions of zein and gluten; zein is mainly composed of N, P, and S, whereas gluten is mainly composed of N and S (Ahmed & Rehman, 2020;Chen et al., 2022).

TGA analysis
The TGA curve of the original BC shows four mass-loss stages (Fig. 6).During the first degradation stage (≤ 200 ℃), the mass loss of the original BC is owing to the evaporation of water (Sheykhnazari et al., 2018).During the second (200-350 ℃) and third (350-570 ℃) stages, the mass loss of the original BC is due to the volatilization of BC (Kiziltas et al., 2015).In the fourth stage (570-1000 ℃), the original BC was finally decomposed to ash (Abral et al., 2020).The TGA curve of BC-zein-SE comprised two degradation stages.During the first degradation stage (≤ 400 ℃), the mass loss of BCzein-SE is due to the major degradation of zein (Altan et al., 2018).The mass loss of BC-zein-SE in the second stage (400-1000 ℃) corresponds to carbonization (Ghorbani et al., 2020).By contrast, the TGA curve of BC-gluten-E showed four degradation stages.
The first stage (≤ 190 ℃) corresponds to moisture loss from both BC and gluten (Khatkar et al., 2013).The second stage of mass loss (190-330 ℃) for BC-gluten-E is attributed to the thermal decomposition of the cellulose structure (Chong et al., 2019).During the third stage (330-550 ℃), gluten underwent chain breakage and volatilization (Peng et al., 2021).During the final degradation stage (550-1000 ℃), the mass loss of BC-gluten-E corresponds to the thermal decomposition of the char residue (Liu et al., 2020).The char residues of BC-zein-SE and BC-gluten-E at 1000 ℃ accounted for 29.890 and 17.048% of their initial masses, respectively.Because char residues can protect BC from combustion (Zhang et al., 2018), the results confirmed that the thermal stability of BC was improved after the entrapment of cereal proteins.

LOI analysis
The LOI experiment was held to determine the flame retardancy of the samples.The LOI of the original BC was 19%, which is similar to that of cotton (approximately 18%), indicating that original BC is a flammable material (Kim & Kim, 2023).The LOIs of BC-zein-SE and BC-gluten-E were 25% and 50%, respectively, which are similar to the literature data (Baishya et al., 2017;Kambli et al., 2018).The increase in the LOIs of BC-zein-SE and BC-gluten-E may be attributed to the synergistic effects of P-N and S-N, respectively, in the flame-retardant system of cereal proteins (Chang et al., 2017).In particular, the LOI of BC-gluten-E was considerably higher than that of cowhide leather (30%).The high LOI of BC-gluten-E may be attributed to the different oxygen barrier properties of zein and gluten.The oxygen permeability of gluten is relatively low due to the intramolecular self-crosslinking of gluten, whereas that of zein is relatively high owing to its helical conformation (Tang et al., 2012).Therefore, the LOI results confirmed that the entrapment of cereal proteins improved the flame retardancy of BC and that gluten has relatively better flame retardancy than zein.

Char morphology analysis
The char morphologies were observed to analyze the mechanism underlying the flame retardancy of the samples.The char residues of the original BC have a fractured morphology (Fig. 7), indicating insufficient char formation (Ao et al., 2020).This char cannot inhibit the heat and mass transfer during combustion, resulting in poor flame resistance (Xu et al., 2020).The char residues of BC-zein-SE exhibited sponge-like structures, whereas those of BC-gluten-E comprised swollen bubbles.These two structures are the common structures of intumescent char, which can act as an insulating barrier that hinders flame propagation and heat and mass transfer during combustion (Das et al., 2020).Moreover, the bubbles observed in the char residues of BC-gluten-E are carbon bubbles created by nonflammable gases, such as oxygen (Li et al., 2011).Thus, it was confirmed that the entrapment of cereal proteins involved a condensed-phase flame-retardant mechanism, forming intumescent chars on the surface of BC.

Tensile strength and elongation at break
As shown in Fig. 8a, the tensile strengths of BC-zein-SE and BC-gluten-E were significantly higher than that of the original BC.This can be explained by the effective stress transfer between BC and cereal proteins via interfacial adhesion (Li et al., 2020).Moreover, the densely packed structures of BC-zein-SE and BC-gluten-E (Fig. 4) might improve their tensile strengths (Jebel & Almasi, 2016).The tensile strengths of BC-zein-SE and BC-gluten-E were higher than those of cowhide leather.However, the elongation at break of BC-zein-SE and BC-gluten-E were similar to that of the original BC and were lower than that of cowhide leather (Fig. 8b).The strong hydrogen bonds between BC and cereal proteins may increase the rigidity of BC by restricting the movement of protein chains (Wang et al., 2017), thereby making BC less flexible and leading to early fracture (Wan et al., 2017).

Fabric softness and crease recovery
The fabric softness and crease recovery of samples were analyzed (Fig. 9), showing that the softness and crease resistance of BC-zein-SE and BC-gluten-E increased compared to those of the original BC, but were not better than those of cowhide leather.Typically, cellulosic fabrics exhibit poor crease resistance owing to the numerous free hydroxyl groups in their structure (Nallathambi et al., 2011).During creasing, cellulose chain slippage in the amorphous region causes hydrogen bonds to break and shift to new regions, increasing the rigidity and wrinkles (Chen et al., 2004;Raza et al., 2018;Tariq et al., 2022).However, if the hydroxyl groups of BC interact with protein molecules by forming hydrogen bonds, the stiffness of BC would be reduced by changing the BC fibril network through intermolecular interactions, thereby enhancing the crease resistance and softness (Barbi et al., 2021;Chen et al., 2018;Nallathambi et al., 2011).

Dimensional stability
The dimensional stability of the original BC was approximately 73-74% after immersion in the wetting solution for 180 min (Fig. 10).This is plausibly attributed to the aggregation of the BC fibers after the evaporation of water, which caused wrinkles on the surface (Domskiene et al., 2019).The dimensional stabilities of BC-zein-SE and BC-gluten-E were markedly higher than that of the original BC because the polymer chains within BC-zein-SE and BC-gluten-E were less mobile owing to the intermolecular interactions of the cereal proteins (Wang et al., 2017).Moreover, the dimensional stabilities of BCzein-SE and BC-gluten-E were comparable to that of cowhide leather.Therefore, the results confirmed that the entrapment of cereal proteins improved the dimensional stability of BC.

Fig. 1
Fig. 1 Residual masses at 1000 ℃ of BC-zein-SE and BC-gluten-E produced with different amounts of zein and gluten

Fig. 2
Fig. 2 FT-IR spectra of a original BC and BC-zein-SE, and b original BC and BC-gluten-E

Fig. 3
Fig. 3 XRD patterns of a original BC and BC-zein-SE, and b original BC and BC-gluten-E

Fig. 4
Fig. 4 SEM images of a the surface morphologies of original BC, BC-zein-SE, and BC-gluten-E at 5000× and 15000× magnifications, respectively, and b the cross-sections of original BC, BC-zein-SE, and BC-gluten-E at 150× and 1000× magnifications, respectively

Fig. 9 a
Fig. 9 a Fabric softness and b crease recovery of original BC, cowhide leather, BC-zein-SE, and BC-gluten-E