Graphene oxide enhanced ionic liquid plasticisation of chitosan/alginate bionanocomposites ‡

: This study reports that the effect of graphene oxide (GO) or reduced GO (rGO) on the structure and properties of polyelectrolyte-complexed chitosan/alginate bionanocomposites is highly dependent on plasticiser type (glycerol or 1-ethyl-3-methylimidazolium acetate ([C 2 mim][OAc])) due to the competing interactions between the components. For the glycerol-plasticised chitosan/alginate matrix, inclusion of GO/rGO enhanced the chitosan crystallinity and increased matrix ductility. While the chitosan/alginate matrix plasticised by [C 2 mim][OAc] showed dramatically weakened interactions between the two biopolymers, GO was highly effective at counteracting the effect of [C 2 mim][OAc] by interacting with the biopolymers and the ionic liquid ions, resulting in enhanced mechanical properties and decreased surface hydrophilicity. Compared with GO, rGO was much less effective at promoting chitosan–alginate interactions and even resulted in higher surface hydrophilicity. However, irrespective of the plasticiser type, inclusion of rGO resulted in reduced crystallinity by restricting the interactions between [C 2 mim][OAc] and the biopolymers, and higher ionic conductivity.

On the other hand, for combined or enhanced properties, hybridisation of different biopolymers or biopolymers with synthetic polymers has been commonly employed (Šimkovic, 2013; van den Broek, Knoop, Kappen, & Boeriu, 2015;Yu, Dean, & Li, 2006). For example, in food packaging, the antimicrobial efficacy of chitosan can be combined with the barrier properties of other synthetic polymers, increasing shelf life and product quality (van den Broek et al., 2015). Moreover, chitosan, as a polycation, can be complexed with negatively charged biopolymers such as alginate, carboxymethyl starch, pectin, and proteins (Mateescu, Ispas-Szabo, & Assaad, 2015), resulting in polyelectrolyte complexation (PEC). The advantages of PEC have recently been demonstrated by creating hybridised biopolymer materials with superior properties to those of either single biopolymer, such as mechanical properties (Li, Ramay, Hauch, Xiao, & Zhang, 2005;Meng, Xie, J o u r n a l P r e -p r o o f 6 Zhang, Wei et al., 2019), barrier properties (Basu, Plucinski, & Catchmark, 2017), hydrolytic stability (Chen, Xie, Tang, & McNally, 2020c), and cell adhesiveness (Iwasaki et al., 2004). Deposition of chitosan together with silk fibroin or collagen on nanofibres could result in materials with excellent antimicrobial activity and cytocompatibility, promising for biomedical applications (Wu et al., 2020;Xia et al., 2019). However, there have been limited studies that have focused on the effects of plasticisers on PEC and biopolymer structure and properties. Previously, we reported that PEC between chitosan and alginate could be strongly influenced by an IL plasticiser, resulting in inferior properties (e.g. largely increased surface hydrophilicity) (Chen, Xie, Tang, & McNally, 2020b). How to improve the plasticisation of such hybridised biopolymer materials while maintaining their complexation and properties is a question of both scientific and practical interest.
The aim of this study is to understand the effects of graphene oxide (GO) and reduced graphene oxide (rGO) as nanofillers on the structure and properties of chitosan and chitosan/alginate blends plasticised by glycerol or 1-ethyl-3-methylimidazolium acetate ([C2mim][OAc]). The advantages of inclusion of GO into chitosan have been demonstrated. For example, chitosan materials with GO added generally showed improved mechanical properties due to efficient load transfer between the nanofiller (GO) and chitosan matrix (Han, Yan, Chen, & Li, 2011;Yang, Tu, Li, Shang, & Tao, 2010). Furthermore, the excellent performance of chitosan/GO composites as adsorbents for metal ions (Fan, Luo, Sun, Li, & Qiu, 2013;Liu et al., 2012) and methylene blue (Fan et al., 2012) was indicated. Moreover, it was reported that biodegradable chitosan-graphene oxide (GO) composites possessed improved mechanical properties and drug delivery performance over chitosan alone (Justin & Chen, 2014). However, how the interplay between GO/rGO and plasticiser affect the J o u r n a l P r e -p r o o f structure and properties of biopolymer PEC systems such as chitosan/alginate has not been widely explored.
In this study, the bionanocomposites were prepared by thermomechanical processing, which imparts high shear stresses enabling excellent dispersion of the nanofillers in the biopolymer matrices. Our hypothesis is that GO and rGO, containing different concentrations of oxygencontaining groups and negative charges, can be used to tailor the plasticisation of hydrophilic biopolymers in different ways. In contrast, most previous studies (Han et al., 2011;Pan, Wu, Bao, & Li, 2011;Yang et al., 2010) have focused on the surface chemistry of the nanomaterials and their direct interactions with biopolymers. Our results highlight largely unexplored routes in which GO or rGO, even in rather small loadings, interfere with blend structure and determine properties (e.g. mechanical properties and hydrophilicity/hydrophobicity) of such multiphasic biopolymer composites, broadening our knowledge of the potential of such biopolymer composites.
Alginate sodium (viscosity: 200±20 mPa·s; M/G ratio: 1:1) was purchased from Shanghai Macklin Biochemical Co., Ltd (China). Graphene oxide (aqueous acid paste with 25% GO, 74% water, and 1-1.5% HCl) was acquired from Abalonyx AS (Norway). Glycerol (≥99% Analytical Grade) was supplied by Fisher Scientific UK Ltd; [C2mim][OAc] (≥95.0%) and triacetin (99%) by Sigma-J o u r n a l P r e -p r o o f Aldrich Company Ltd (UK); formic acid (98% w/w AR) and NaBr (pure) by Scientific Laboratory Supplies Ltd, (UK); hydrazine hydrate solution (78-82% iodometric, Honeywell Fluka) and ammonia solution (35%, AR, d = 0.88) from Fisher Scientific UK Ltd. Deionised water was used throughout the study. Reduced graphene oxide (rGO) was synthesised from GO following the method described previously (Chen, Xie, Tang, & McNally, 2020a). The characteristics of GO and rGO can be found in our previous publication (Chen et al., 2020a). Table 1 shows the formulations and codes of different samples prepared. The matrix was either chitosan alone (represented by "X") or chitosan/alginate (indicated by "Y"). The codes also signify the plasticiser used, with "G" for glycerol or "E" for [C2mim] [OAc]. The suffix "F" indicates the processed samples were films. The samples were prepared by pre-blending, thermomechanical kneading at 80 °C for 15 min, and hot-pressing at 110 °C and 160 bar for 10 min, followed by conditioning at 57% relative humidity for 3 weeks as described previously (Chen et al., 2020c).

Sample preparation
Additionally, one of the plasticisers (20 wt% based on biopolymer weight) and either GO or rGO (0.75 wt% based on biopolymer weight) were added during the pre-blending step. The samples without GO or rGO, namely XG-F, XE-F, YG-F, and YE-F, have been reported previously (Chen et al., 2020b) and are termed as controls throughout the discussion.
where θ is the angle of incidence, λ is the wavelength of the incident light, n is an integer.
Thermo-gravimetric analysis (TGA) was undertaken using a Mettler Toledo TGA apparatus over a temperature range of 30-700 °C at 10 K/min under nitrogen.
Dynamic mechanical thermal analysis (DMTA) was performed using a Tritec 2000 DMA (Triton Technology Ltd, UK) in dual cantilever mode with a sample length of 5 mm at a displacement of 0.01 mm. Temperature scans were performed from −100 °C to 180 °C at 2 °K/min and 1 Hz.
Tensile testing was performed using an Instron 3367 universal testing machine with a 1kN load cell at a crosshead speed of 3 mm/min. As the specimens were in the form of thin sheets, specimen extension was measured by grip separation as suggested by ASTM Standard D882. At least seven replicates were used for each sample.
Contact angle (θc) data was obtained from sessile tests at RT based on Young-Laplace using an Attension Theta Lite instrument (Biolin Scientific, UK). As θc kept changing after a drop of water was placed onto the sample surface, θc values at 0 s, 30 s, and 60 s (denoted as θc0s, θc30s, and θc60s respectively) were recorded.
Electrical impedance spectroscopy ( where, ω is the angular frequency (= 2πf), ε0 is the permittivity of free space (≈ 8.854 ×10 −12 F⋅m −1 ), A is the tested area of the sample (m 2 ), and t is the sample thickness (m).
The bulk resistance (Rb) was determined from the Nyquist plots of impedance (Z″ vs. Z′) from the points where the semicircle and the straight line meet. Then, the conductivity (σdc) can be calculated using equation (5) (Chen et al., 2020b)  The morphology of the different bionanocomposites was further studied using STEM, as shown in Figure 1. For the X-series of composites, minor traces of GO/rGO agglomerations can be seen.
Given this observation, it is considered that the GO/rGO nanosheets were largely exfoliated and dispersed in the matrices either as few-layer nanoplatelets, which lost contrast under STEM and difficult to image. GO nanosheets are generally negatively charged resulting from the ionisation of the oxygen-containing groups (e.g. ─COOH and ─OH). Therefore, dispersion could also be promoted by hydrogen-bonding and electrostatic interactions between the chitosan polycation and the negatively charged GO nanosheets (Yang et al., 2010). Excellent dispersion of GO in chitosan materials has also been noted in previous studies (Han et al., 2011;Pan et al., 2011;Yang et al., 2010). In the Y-series of composites, large GO/rGO agglomerations are more frequently observed although they are small in number. It is likely that PEC between chitosan and alginate competed with the interactions between GO/rGO and chitosan to some extent.
J o u r n a l P r e -p r o o f

J o u r n a l P r e -p r o o f
For the Y-series of samples, a "new structure" was also observed normally at the edges of the areas imaged (where the material has no or much less interaction with the electron beam) ( Figure   S2), which is highly interesting. Given this, the energy from the electron beam could possibly facilitate coordination between alginate and [C2mim] + and the packing of polysaccharide chains to form crystals, in an analogy to the formation of junction zones by alginate with Ca 2+ (Li, Fang, Vreeker, Appelqvist, & Mendes, 2007;Morris, Rees, Thom, & Boyd, 1978;Sikorski, Mo, Skjåk-Braek, & Stokke, 2007). This phenomenon was further investigated. Figure 2 shows the FTIR spectra for the different bionanocomposites. The X-series of composites displayed FTIR patterns very similar to those for XG-F and XE-F (Chen et al., 2020b), indicating inclusion of GO or rGO did not significantly alter the molecular interactions in the plasticised chitosan matrices. Compared with the X-series, the Y-series had blue shifts of the peaks originally at 1570 cm −1 and 1024 cm −1 . The peak at 1570 cm −1 is assigned to the N─H bending vibration from amine and amide II (Lawrie et al., 2007) and the one at 1024 cm −1 is attributed to the skeletal vibration of C─O stretching (Lawrie et al., 2007;Papageorgiou et al., 2010). Thus, PEC should have involved amine and amide groups and affected the saccharide ring structure.  (Lawrie et al., 2007;Papageorgiou et al., 2010)) became less intense, which is caused by the weakened PEC and hydrogen bonding between chitosan and alginate with the presence of [C2mim] [OAc]. With inclusion of GO, these changes induced by the IL were apparently suppressed since YE/GO-F showed an FTIR pattern similar to that for YG-F, YG/GO-F and YG/rGO-F. This is caused by the interactions of GO (negatively charged) with the IL (especially the [C2mim] + cation) and with the chitosan cation. rGO was also seen to counteract the effect of the IL on biopolymer molecular interactions, but to a much lesser extent than GO. Figure 3 shows the XRD curves for the different bionanocomposite films. All the X-series of composites displayed similar diffractograms to those for XG-F and XE-F (Chen et al., 2020b). For all these samples, there were three major peaks at 2θ of about 13.5° ((020) reflection, d-spacing = 0.76 nm), 21.7° ((100) reflection, 0.48 nm), and 27.2° ((110) reflection, 0.38 nm), attributable to the crystal lattice of chitosan (Kittur, Vishu Kumar, & Tharanathan, 2003). As the XRD pattern of the processed chitosan was completely different from that of original chitosan, the crystalline structure in the composites should be predominantly due to re-crystallisation (Chen et al., 2020b). Inclusion of GO/rGO did not change the recrystallised structure of plasticised chitosan.  All the Y-series of composites displayed a low degree of crystallinity as YG-F and YE-F did (Chen et al., 2020b). A predominantly amorphous structure should result from PEC between chitosan and alginate. Compared with YG-F, YG/GO-F and YG/rGO-F had slightly stronger peak intensities especially at 13.5° and 21.7°, suggesting that inclusion of GO or rGO increased the crystallinity of chitosan plasticised by glycerol. However, YE/rGO-F showed even weaker peak intensities than YE-   The Y-series of composites had Td unchanged relative to those for YG-F and YE-F (243 °C) (Chen et al., 2020b). For YG-F, YG/GO-F, and YG/rGO-F, the derivative-weight peak of alginate (peak temperature at 206 °C) should be overlapped with that of chitosan and was just about visible.

X-ray diffraction (XRD) analysis
Considering the Td values for unprocessed chitosan and alginate are 289 °C and 232 °C respectively (Chen et al., 2020b), complexation between chitosan and alginate dramatically resulted in decreased thermal stability of both polysaccharides. In contrast for the Y-series plasticised by glycerol, for YE-F (Chen et al., 2020b), YE/GO-F, and YE/rGO-F, the alginate peak became more prominent and

Mechanical properties
Representative stress-strain profiles from tensile testing (Figure S3) of the different bionanocomposite films indicates they were hard and tough. From these curves, the Young's modulus (E), tensile strength (σt), and elongation at break (εb) were calculated and plotted in Figure   6  and an increase in εb (68.2±1.7% and 69.8±9.2%, respectively). This suggests inclusion of GO or rGO may have improved the distribution of glycerol in the Y-matrix (especially in the alginate phase) and reduced polysaccharide chain interactions in amorphous regions and, thus, resulted in Young's modulus (MPa) Tensile strength (MPa) and εb = 64.5±12.0%), YE/rGO-F had similar mechanical properties, whereas YE/GO-F had lower E (323±75 MPa) but higher σt (29.1±4.9) and εb (89.3±8.6%). As discussed above, [C2mim] [OAc] disturbs PEC between chitosan and alginate whereas GO counteracts the effect of the IL, reflected in the enhancement in strength and ductility of YE/GO-F.
Despite these effects of GO or rGO on the tensile properties, Figure S4 shows that the Shore D hardness was not apparently influenced by their inclusion regardless of the matrix. The Shore D hardness was mainly influenced by the plasticiser especially, for the X-matrix. Figure 7 shows the θc0s, θc30s, and θc60s values for the different bionanocomposite films, as contact angle kept changing during the sessile measurement. Our previous study (Chen et al., 2020b) indicated that XG-F had θc0s = 102±6°, θc30s = 81±4°, and θc60s = 73±3° and XE-F had θc0s = 95±3°, θc30s = 74±4°, and θc60s = 70±4°. For the X-matrix regardless of plasticiser type, inclusion of GO or rGO did not cause notable changes in contact angle, i.e. all have similar surface hydrophilicity. In this regard, the surface hydrophilicity was predominantly determined by the polarities of chitosan and plasticiser groups on the film surface, which were not varied by GO or rGO.  (Chen et al., 2020d).

Contact angle
Compared with YG-F (θc0s = 98±6°, θc30s = 95±6°, and θc60s = 93±6°), YG/GO-F and YG/rGO-F did not show apparent changes in contact angle. However, for the [C2mim][OAc]-plasticised Ymatrix, the surface hydrophilicity was remarkably varied by inclusion of GO or rGO even at 0.75 wt% loading, which is surprising. Specifically, while YE-F had θc0s = 48±5°, θc30s = 36±5°, and θc60s = 33±5°, these values were dramatically increased to 84±6°, 80±7°, and 78±7° respectively for YE/GO-F whereas decreased to 31±5°, 20±4°, and 18±7° respectively for YE/rGO-F. The reduced surface hydrophilicity for YE/GO-F can again be, ascribed to the less-interfered PEC between the two polysaccharides by [C2mim] [OAc] with the presence of GO, as discussed above. In addition, the interaction of GO with the IL could also limit the binding of the IL with water, also contributing to decreasing surface wettability. In comparison, rGO was much less effective than GO to counteract the effect of the IL and, thus, the interactions between chitosan and alginate were still weak.
Meanwhile, rGO could dissociate some IL ions and/or the polysaccharide hydrophilic groups from interactions, which could then readily bind with water. In this regard, YE/rGO-F even had a greater surface wettability than YE-F.

Electrochemical impedance spectroscopy (EIS)
Figure 8 (a) shows the Nyquist plots of impedance (Z″ vs. Z′) for the different bionanocomposite films. Based on these plots, the Rb and σdc values calculated (Bonanos, Steele, & Butler, 2005) are listed in Table S1. Compared with XG-F (σdc = (3.93±0.70)×10 −5 S·cm −1 ) (Chen et al., 2020b), only XG/rGO-F showed an apparent increase in σdc, which should be associated with the intrinsic conductivity of rGO. XE/GO-F and XE/rGO-F had similar σdc values to that for XE-F ((6.85±0.78)×10 −5 S·cm −1 ). Given this result, the conductivity of the X-samples plasticised by [C2mim][OAc] could be mainly determined by the IL as a salt (Wang, Chi, & Mu, 2014). YG/GO-F and YG/rGO-F had σdc about twice that of YG-F ((2.80±0.31)×10 4 S·cm −1 ) (Chen et al., 2020b), suggesting inclusion of the 2D nanofillers contributed to the electrical charges (ions and dipoles) in the polysaccharide composite system. Compared with YE-F (σdc = (2.74±0.20)×10 −5 S·cm −1 ) (Chen et al., 2020b), YE/GO-F displayed similar σdc whereas YE/rGO-F had a higher value J o u r n a l P r e -p r o o f ((6.54±0.82)×10 −5 S·cm −1 ). Although GO could disrupt the interactions between the IL ions and the respective biopolymers, its interactions with the IL ions and the polysaccharides could limit the mobility of the electrical charges. The increased σdc value of YE/rGO-F could be derived from the conductivity of rGO, as well as the greater availability of IL ions and/or the polysaccharide hydrophilic groups, as discussed above.  (Chen et al., 2020b). In this regard, rGO contributed to the overall conductivity of the materials. shows that decreasing f led to an abrupt increase in ε′r, which could be ascribed to electrode polarisation and space charge effects (dipole moment) (Khiar, Puteh, & Arof, 2006;Navaratnam et al., 2015). Compared with the controls (Chen et al., 2020b), the bionanocomposites with GO had higher ε′r at low f (<50 Hz). And, inclusion of rGO had an even greater effect on ε′r at low f. In this regard, rGO could be more effective than GO at facilitating the accumulation of mobile ions. Moreover, the bionanocomposites displayed impressively high ε′r at 1 kHz (over 150) except for YE/GO-F (Table S1). For YE/GO-F, the strong interaction of GO with the IL ions and the polysaccharide polar groups could restrict the dipole moment.
Figure 8 (d) shows that for all the bionanocomposites, there was a well-defined peak in M″ at high f, indicating relaxation processes with distributed relaxation times (i.e. viscoelastic relaxation, or dipolar relaxation) (Fadzallah et al., 2014). Compared to XG-F whose M″ peak position was at about 1.4×10 5 Hz (Chen et al., 2020b), both XG/GO-F and XG/rGO-F had the peak moved to 1.7×10 5 Hz, indicating reduced relaxation time. In this regard, inclusion of GO or rGO increased the mobility of ions and associated dipoles. XE-F, XE/GO-F and XE/rGO-F had similar peak positions at about 1.7-1.8×10 5 Hz, suggesting no apparent effect of the nanofillers on the relaxation time for J o u r n a l P r e -p r o o f the X-matrix plasticised by [C2mim] [OAc]. In other words, in these three samples, the mobility of ions and dipoles could be mainly determined by the IL. On the other hand, while the peak position for YG-F was about 1.7×10 5 Hz, YG/GO-F and YG/rGO-F displayed a peak position at about 2.9-3.2×10 5 Hz. Compared with YE-F peak position at about 1.5×10 5 Hz, YE/GO-F and YE/rGO showed peak positions at about 1.7×10 5 Hz and 2.9×10 5 Hz, respectively. These results indicate increased mobility of ions and dipoles by inclusion of GO or rGO to the Y-matrix. In particular, the short relaxation time for YE/rGO-F corresponds to the disrupted interactions between the IL ions and polysaccharides, as discussed above. structure and properties was minor, most likely due to the dominant interactions between plasticiser and chitosan. However, for the [C2mim][OAc]-plasticised X-matrix, inclusion of GO or rGO increased ductility, with rGO being more effective, behaviour attributed to the GOs being capable of improving the distribution of this plasticiser in the chitosan matrix.
Thus, this work has shown the different ways in which these 2D carbon materials influence the structure and properties of polysaccharides and, in particular, the efficacy of GO to overcome the negative effects of the IL cation on PEC in polysaccharide materials. This information could be insightful for the design of various biopolymer composite systems where multiple interactions among components can be manipulated so as to tailor properties.