Proton conducting electrolytes composed of chondroitin sulfate polysaccharide and citric acid

Abstract Novel electrolytes composed of chondroitin sulfate A (CSA) and citric acid (CA) have been prepared using a clean, safe, and fast route. These electrolytes exhibit different physical-chemical properties, depending on the amount of CA. For X > 82.3%, where X is the mass ratio, in %, of CA/(CA + CSA), whitish polycrystalline powders result. Lower amounts of CA leads to the production of translucent, amorphous films, sticky for X  = 75.6 and 82.0, brittle for X


Introduction
According to the United Nations (UN), environmental and energy issues are two of the major challenges human society has to face. These problems have become so complex and taken on a global scale that, for the last decades, a call for game-changing technology that can lead to sustained development has consistently been issued [1]. In its latest iteration of the global sustainable development agenda, the UN has adopted 17 goals intended to prompt action from human agents in areas of critical importance for the sustained development of both human society and the planet [2]. Energy-wise, the UN goal aims to "ensure access to affordable, reliable, sustainable and modern energy for all" by focusing on three main areas: energy efficiency, renewable energy and energy access [2]. Access to reliable and sustainable energy sources has prompt the development of new technologies capable of addressing both the human need for energy as well as minimize the ecological consequences associated with its use. In this sense, the choice for chemistries that can minimize the ecological footprint and/or are easily integrated in a full circular economy needs to be prioritized.
In this context, biorenewable natural products, such as polysaccharides and proteins, are increasingly being held as ideal materials for the production of solid polymer electrolytes (SPEs), both from an economic and an environmental point of view [3][4][5]. The interest in SPEs [6,7] has been steadily increasing, due to their applications in various solid state electrochemical devices, such as batteries, sensors, fuel cells, solar cells and electrochromic devices [8][9][10][11].
Polysaccharides are a versatile and rich family of macromolecules whose structure and thermodynamics have been extensively investigated. Polysaccharides comprise some of the most abundant biopolymers on Earth [5,12]. Polysaccharides, such as starch, heparin, cellulose, chitosan, hyaluronic acid or agar, among others, have been blended, doped, combined or otherwise manipulated with suitable substrates to produce ion conducting materials [12][13][14][15][16][17][18][19]. A popular strategy for the production of novel SPEs based on polysaccharide host polymers is to dope them with ionic liquids (ILs). The underlying objective is the combination of the polysaccharides' properties with other desirable properties present in the ILs, namely their high thermal stability, wide electrochemical potential window, moderate viscosity, good redox stability, plasticizing and lubricating properties [13,14,20]. However, ILs have been found to possess some disadvantages: (1) they may be more toxic than first expected [21][22][23] and (2) IL leakage problems have been reported [24]. These drawbacks motivated the development of another type of materials with physical properties and phase behavior similar to those of ILs. These materials, classed as deep eutectic solvents (DESs), possess a melting point lower than those of their individual constituents and low enough to be used as solvents. Because they are synthesized from eutectic mixtures of Lewis/Brønsted acids and bases derived from natural sources (e.g., aminoacids, carboxylic acids, sugars, urea and choline), DESs are economically cheaper and environmentally friendlier than conventional ILs [25][26][27][28][29].
Following our previous works on polysaccharide-based electrolytes [13,14,31], we turned our attention to chondroitin sulfate (CS) as a potential SPE due to its anionic character and benign nature. CS is a naturally occurring linear polysaccharide (Fig. 1a) that is particularly abundant in the connective tissue of mammals, but also in invertebrates and even some bacteria. Similarly to what is found in the case of other biopolymers extracted from living organisms, CS chains exhibit significant structural variation in terms of molecular mass, arrangement of disaccharide units, sulfation pattern and charge density, among others [32,33]. However, all forms of CS share common structural features, including assemblages of disaccharide units mainly composed of β(1,4)-D-glucuronic acid and β(1,3)-N-acetyl-D-galactosamine, incorporating hydroxyl, amide and carboxyl groups, and held together by glycosidic linkages (Fig. 1a) [32,33]. Sulfonic groups are often attached in either the C-4 or C-6 positions of the galactosamine residue. These two forms of CS are usually known as CSA and CSC, respectively. Other disaccharide sulfation patterns have also been reported which differ in terms of number (ranging from non-sulfated disaccharide units to disulfated disaccharides) or position (the second sulfonic group appearing in either the galactosamine residue or the glucoronic acid moiety). CS is known to be highly stable, hydrophilic, easily chemically modified, and exhibits low toxicity and good biodegradability [34]. Furthermore, electrolyte complexes can be formed with inorganic acids [35], metal cations [35] and cationic biopolymers, such as chitosan [36][37][38][39][40]. Curiously, until now, research conducted with this polyanion has been mostly restricted to therapeutic applications, such as tissue engineering [34,41,42] and drug delivery [39,41,43].
Inspired by the DES concept, we combined CSA with citric acid (CA) (Fig. 1b), an organic polyprotic acid naturally occurring in citrus fruits. CA is known for its ability to establish hydrogen bonds [44] and form low-temperature transition mixtures (i.e. DESs) with other organic entities, namely choline, various sugars or proline [30,45]. Also of note is the increasing use of CA as a plasticizer [46,47] and a cross-linker [44,46], including in the preparation of proton exchange membranes [47]. A similar strategy was employed in the preparation of SPEs composed of chitosan and carboxylic acids, such as oxalic acid [48,49], acetic acid [50] or maleic acid [16]. It is necessary to highlight that CSA has remained practically unexplored as a promising host polymer candidate for SPEs, in spite of the significant presence of sulfonic groups. Interestingly, just one work on the proton conductivity of a composite prepared from CSA, chitosan and hydroxyapatite was found in the literature [51]. The proton conductivity values registered by this composite are higher than those reported for Nafion [51]. The SPEs here described were synthesized by means of a clean, safe and cheap procedure, and their structure, thermal and electrochemical properties were investigated. These electrolytes will be henceforth noted as CA:CSA(X), where X is the mass ratio, in %, of CA/(CA + CSA) (15.2 ≤ X ≤ 93.7).
In a typical experiment, a solution of CA (3.00 g/6.00 mL H 2 O) was added to an aqueous solution of CSA (4% wt.) and stirred for about 10 min. Different amounts of CA were added, ranging from 0.12 to 6 mL, depending on the specific mass ratio of CSA/CA intended. In all cases, the total volume varied from 9 to 11 mL, depending on the specific CSA/CA ratio used (Table S1 in the Supplementary Materials). The pH of the solution was checked before and after CA addition, and the resulting solution was carefully poured onto custom-made Teflon molds (5 cm inner diameter a rough surface due to use) and allowed to dry. After a drying period of 4-5 days during which the solution was left to evaporate in air, the resulting materials were collected and analyzed.

Methods
Fourier transformed infrared spectra obtained in attenuated total reflectance mode (ATR-FTIR) spectra were recorded with an IRaffinity-1S Shimadzu spectrometer equipped with a horizontal unique reflection  ATR accessory (Golden Gate). The sample was placed over the diamond crystal with the help of a spatula. All spectra were obtained with 128 scans between 4000 and 400 cm −1 with a resolution of 2 cm −1 . The Differential Scanning Calorimetry (DSC) curves of the samples were recorded at a heating rate of 10°C min −1 under a flowing argon atmosphere. Each sample was placed in a 40 µL perforated aluminum pan and the thermogram was recorded using a Mettler DSC 821e.
Thermogravimetric analyses (TGA) and additional DSC measurements were carried out in a DTA/TGA Netsch STA 449F3 thermal analyzer. In a typical experiment, samples were heated from room temperature to 700°C in alumina (Al 2 O 3 ) crucibles under a nitrogen (N 2 ) atmosphere (50 mL min −1 purge; 20 mL min −1 protective flow), at a heating rate of 10°C min −1 .
Polarized Optical Microscopy (POM) images were recorded in an OPTIKA B-600 Pol microscope. The images were obtained through a digital camera 8 Mpixel Digital Photo and were analyzed with Optika Vision Pro software.
Solid state Nuclear Magnetic Resonance (NMR) experiments were conducted at CICECO -Aveiro Materials Institute, using a Spectrometer Bruker Avance III 400 wide-bore (400 MHz, 1 H Larmor frequency). A 4 mm double-resonance Magic Angle Spinning (MAS) probe was employed at 100.6 MHz ( 13 C) Larmor frequency. Samples were spun in ZrO 2 rotors using a spinning rate of 12 kHz. 13 C Cross Polarization (CP)/MAS NMR spectra were recorded using a ramp step (varying from 100 to 50% in amplitude using 100 points), with a recycle delay of 5 s, a contact time of 2.0 ms and 1 H 90°excitation pulse of 3.20 μs. All chemical shifts are quoted in parts per million (ppm) from tetramethylsilane (TMS).
pH measurements were obtained with a HANNA pH20 pH-meter at room temperature. All buffer solutions employed in the calibration of the pH-meter were obtained from VWR Chemicals.
The ionic conductivity (σ) of the samples was determined under real conditions by locating an electrolyte disk between two 10 mm diameter ion-blocking gold electrodes (Goodfellow, > 99.95%) to form a symmetrical cell. The electrode/sample/electrode assembly was secured in a suitable constant volume support and installed in a Büchi TO51 tube oven with a K-type thermocouple placed close to the electrolyte disk to measure the sample temperature. Bulk conductivities of the samples were obtained during heating cycles by using the complex plane impedance technique with the Autolab PGSTAT-12 (Eco Chemie) equipment between room temperature and 110°C and at approximately 15°C intervals. The ambient relative humidity (RH) was lower than 30%.
The ionic conductivity of membranes was also determined at room temperature (~26°C) and under variable RH (30-98%) by impedance spectroscopy using an Agilent E4980A Precision LCR meter. The through-plane (TP) configuration was adopted and the membranes were placed between two carbon cloth gas diffusion layers (~0.5 cm 2 ) and graphite plates with gas channels. The samples were mounted on an appropriate support and placed in a climatic chamber made in-house, where variable relative humidity (RH) conditions were achieved using different saturated salts solutions. The RH was constantly monitored close to the sample with a humidity sensor. The impedance spectra were collected between 20 Hz and 2 MHz with a test signal amplitude of 100 mV. Spectra were analyzed with ZView (Version 2.6b, Scribner Associates) to assess the ohmic resistance (R), which was normalized to the samples geometry to calculate the conductivity through the usual formula σ = L(RA) −1 , where L is the distance between the in-plane silver stripes (or the thickness of the membranes) and A is the crosssection area of the membranes (or the surface area of the electrodes).
The Atomic Force Microscopy (AFM) analysis was performed in an AFM CSI Nano-Observer equipment (Scientec) in tapping mode using a super sharp Si HQ:NSC19/FORTA probe with a frequency resonance of 60 kHz and a spring constant of 0.3 N m −1 . In order to improve the quality of the images, flattening and elimination of line noise tools were applied using the Gwyddion 2.52 software. Because the samples produced were dried in air, the surface of the film was exposed to N 2 prior to AFM analysis to clean any dust and fibers that might have been trapped due to air exposure. As the film was rather sticky, additional precautions were undertaken to avoid the insertion of indentations or film stretching. The film was deposited on double tape glue that was previously homogenized on the support to avoid the formation of bubbles. The AFM analysis was performed on the surface which had been exposed to air during the drying process. Indeed, due to the nature of the film itself, the surface in contact with the mold mimicked its patterns (e.g., scratches caused by prolonged use, including removal of films with spatula) (Fig. S1 in the Supplementary Material). Therefore it was decided that the analysis of the surface exposed to air would provide a deeper knowledge of the topographical characteristics of the film, regardless of the mold used.
Scanning Electron Microscopy (SEM) images of the top (exposed to air during drying) and bottom (in contact with the Teflon mold) surfaces, as well as the cross-section, of the samples were obtained at 20 kV on a Hitachi S-3400 N type II microscope (Hitachi, Chiyoda, Japan) equipped with a Bruker x-flash 5010 at high vacuum. For the crosssection analyses, the cut was made at cryo-temperatures. All the samples were coated with gold in a rotating holder to ensure homogeneous distribution. Elemental mapping of the samples was performed by Energy Dispersive Spectroscopy (EDS) X-ray analysis. The acceleration voltage and intensity of the beam current were chosen taking into consideration the chemical composition of the samples. In addition, both parameters were optimized to maximize high count and low deadtime. The acquisition time for a satisfactory resolution and noise performance was 60 s. As the films were flat, the directionality was neglected in the measurements.

Results and discussion
CSs are low charge density structures which include several functional groups prone to protonation (Fig. 1a): most notably an amide group, a carboxylic group and a sulfonic group, but also an ether and hydroxyl groups, meaning that the same structure features both weak and strong acidic groups [32,37]. In our working conditions, the measured pH of a 4% aqueous solution of CSA was 6.0, indicating that both the weakly acidic carboxylic group (pKa 3.5-4.6) and the strongly acidic sulfonic group (pKa 2.6) were deprotonated [32,52].
The addition of a concentrated CA solution to the CSA aqueous solution led to an immediate drop in pH, and the concomitant appearance of some turbidity. However, this turbidity ultimately disappeared (Table 1 and S1 in the Supplementary Material). Different combinations of mass ratios of CSA and CA, henceforth designated by CA:CSA(X) where X indicates the specific mass ratio used, afforded different end products, from powders to films (Table 1). Films were obtained for a wide range of X values, ranging from 15.2 to 82.3%, from solutions with pH values as low as~2.1 (Table 1 and Fig. S2 in the Supplementary Material). These films exhibited different properties, depending on CA concentration. Thus, low concentrations of CA (X < 43.6 and pH > 3.0) afforded brittle films, probably due to the insufficient protonation of CSA by CA (pKa 1 = 3.13, pKa 2 = 4.76 and pKa 3 = 6.40) [53]. For 43.6 < X < 70.1 (2.55 < pH < 2.89) non sticky films were produced. Sticky flexible films were obtained for 70.1 < X < 82.3 (2.07 < pH < 2.29). These results point out a trend between the flexibility of the film and the increasing amount of CA, as a result of the formation of new intermolecular hydrogen bonds which occurs at the expense of the existing intramolecular interactions of CSA [54]. Moreover, this trend confirms the CA role as a plasticizer in the formation of SPEs [46,47].
The incorporation of higher amounts of CA (X > 83.4%, pH ≤ 2.0) led to the gradual precipitation of a white powder. Given the observed pH of the parent solution, a strong interaction between the CA and the anionic polysaccharide was expected, in a similar manner as reported for other SPEs obtained with CSA at low pH [55].
3.1. Spectroscopic characterization of the CA:CSA electrolytes 3.1.1. 13 C CP-MAS NMR spectroscopy 13 C NMR spectroscopy is known to provide relevant information about the structure of CSA and related polysaccharides, such as the pattern of sulfation, conformation of the polysaccharide or even hydrogen bonding [56,57]. The 13 C CP-MAS NMR data collected for both the starting materials are presented and discussed in the Supplementary material. The 13 C CP/MAS NMR spectra of the CA:CSA electrolytes exhibit signals pertaining to both CSA and CA ( Fig. 2 and Fig. S4 and Table S2, Supplementary Material). At X = 60.8 the 13 C CP/MAS NMR spectrum is still dominated by the CSA resonances. However, the carboxyl resonance, centered at 176.0 ppm, appears as a broad, asymmetric signal, with new shoulders at 179.4 and 173.6 ppm due to the carboxylic groups present in the CA moiety (Fig. 2). A new broad signal located at 43.7 ppm emerges, due to the CA methylene carbons (Fig. S3). As higher amounts of CA are progressively added to the CSA biopolymer, several changes occur until the spectra are entirely dominated by the characteristic resonances of CA (at X ≥ 82.3). The sharpening of the signals seen for X ≥ 82.3, indicate the high structural order of the powdered material, in contrast with the broad signals produced by films (for X < 82.3).
Two different regions are of particular interest: the region between 182 and 170 ppm (Fig. 2), corresponding to the carbonyl groups, and the region between 110 and 90 ppm, related to the glycosidic bonding (Fig. S4).
In the carbonyl region (182 < ppm < 170), the CSA signal at 176 ppm is visible up to X = 60.8%, with higher amounts of added CA leading to spectra dominated by the resonances of the CA carboxylic groups (Fig. 2). In the spectra of CA:CSA electrolytes with 82.3 ≤ X < 100, three prominent resonances are observed at 177.4, 174.1 and 173.6 ppm, which do not coincide with the characteristic resonances of the CA precursor, observed at 179.4, 176.0 and 174.6 ppm (Fig. 2). Fisher et al. associated those resonances with CA monohydrate [58]. In addition, weak signals at 179.2, 176.2 and 174.6 ppm are seen in some spectra, indicating the presence of a secondary phase consisting of anhydrous CA, as suggested by the spectrum of the pure precursor and by the literature data [58]. As the 13 C CP/ MAS NMR spectra for the CA:CSA electrolytes with X > 82.3 are dominated by the resonances of CA in its various forms, we can conclude that for these particular formulations, there is an excess of CA present in the collected powders. In X > 82.3, the predominance of the CA monohydrate signals over the CA anhydrous signals strongly suggests that the powders here obtained are mostly constituted by CA monohydrate with a secondary phase consisting of CA:CSA(X) electrolytes, which could be justified by the considerable CA molecular aggregation arising from the hydrogen bonding interaction between both CA and water [59].
Over the whole X range investigated, no signals above 180 ppm were detected. The absence of these signals, which are related with the formation of sodium citrate [58], indicates that the sodium counter-ions present in the raw CSA sample did not complex with CA.
In the region between 110 and 90 ppm (Fig. S4), the broad signal observed at 102.8 ppm, for CSA, gradually shifted into a duplet at X = 75.6, as a consequence of a change in the chemical environment of the glycosidic bridge [60,61]. According to previous works [60][61][62], two resonances are expected in this region of the CSA 13 C NMR spectrum, respectively at 106.5-105 and 103.5-102 ppm, due to the (1-4)β-linked C-1 carbon (peak b) and (1-3)-β-linked C-1 carbon (peak e), respectively. In this work, the same resonances appear at 104.2 and 101.2 ppm. The envelope that contains both signals remains centered in the same area as the starting CSA material (Fig. S3 and Table S2). Hamer et al. observed no significant shift of these signals when CSA was acidified to a pH~1. [60]. In a study involving the interaction between chondroitin sulfate-6 (CSC) and chitosan, the authors observed a shift of the resonance associated with the anomeric carbon from 104.7 to 99.3 ppm, which was attributed to the electrostatic interactions between both electrolytes [38]. Given that the difference between CSC and the CSA used in this work is simply the location of the sulfonic group (which in our case is located at the C4 position of the N-acetylgalactosamide ring, instead of the C6 position of the same ring), this might have led to a smaller upfield shift than that observed by Chen et al. [38], since the presence of sulfonic groups in neighboring atoms is known to induce higher chemical shifts [57].
From the gathered evidence, we deduce the nature of the interaction between CSA and CA is probably electrostatic, as observed for other materials containing CSA [37,38,63,64], especially at low pH (≤2). It has been argued that at low pH, a competition exists between the protonation of the polyanionic sites of CSA and their complexation with the Na + cations [37,38], which might result in the condensation of the acidic and sodium cations around the CSA backbone [65,66].

ATR-FTIR spectroscopy
The ATR-FTIR spectra of the CA:CSA electrolytes, and of pure CSA and CA are reproduced in Figs. 3 and S5 (in the Supplementary Material). Relevant ATR-FTIR data are collected in Table S3. Of  particular interest is the spectral interval of 1800-600 cm −1 , found to be the most useful for the identification of the vibrational modes of CSA and CA [67,68]. The analysis of these ATR-FTIR spectra is complicated by the fact that vibrational modes of different origins appear superimposed in most of the spectral regions.
In the spectrum of CSA, the asymmetric stretching vibration of the carboxylate groups (ν as COO − ) emerges as a very strong band centered at 1607 cm −1 with two shoulders at 1641 and 1551 cm −1 (Fig. 3), due to the Amide I and Amide II vibrations, respectively [68][69][70][71][72]. According to the literature, the shift of the CSA band at 1611 cm −1 to lower wavelengths (1607 cm −1 ) is indicative of the presence of hydration waters, interacting with the carboxylate groups through hydrogen bonds [72].
The asymmetric stretching vibration of the carboxylic groups (ν as COOH) in CSA is expected between 1740 and 1720 cm −1 [37,69,71]. In the ATR-FTIR spectrum of CA the ν as COOH mode is observed at 1743 and 1693 cm −1 (Fig. 3). These bands are assigned to the central carboxylic group and to the terminal carboxylic groups, respectively [73][74][75].
When CSA and CA are combined, changes in the shape, number and wavenumber of the carboxylate bands can be seen in this particular region of the spectra, indicating strong interactions between both starting materials. As expected, as the amount of CA added to the CSA solution increased, the maximum of the band envelope due to the asymmetric stretching vibration of the carboxylate groups, ν as COO − , mode of the polymeric moiety gradually shifted to higher wavenumbers, first with a band centered at 1708 cm −1 (for X ≤ 60.8) and then with a band located at~1727 cm −1 (for X ≥ 75.6) (Fig. 3). This shift, along with the disappearance of the ν as COO − band at 1607 cm −1 , is indicative of the gradual protonation of the electrolyte. At X ≥ 75.6, two new bands are clearly seen at 1755 and~1685 cm −1 (Fig. 3). The band located at 1685 cm −1 is attributed to the CA monohydrate [74].
The band at 1755 cm −1 is due to the presence of the dimeric form of CA [44,75], probably due to the establishment of hydrogen bonds between the carboxylic groups of each monomer [44,75]. The presence of this particular form of CA is the only to be expected, given the low working pH (which favors hydrogen bonding) and is in agreement with the 13 C CP/MAS NMR data.
Sulfonic characteristic modes [68][69][70]72] are found at 1224 cm −1 (with a shoulder at 1261 cm −1 ), associated with the asymmetric stretching vibration of the sulfate groups ν as (OSO 3 − ) mode, and at 851 and 720 cm −1 , due to the asymmetric stretching vibration of the carbon-sulfate bond, ν as (CeOeS) (Fig. S4). While the bands derived from the CSA carbon backbone remain unchanged in all the CA:CSA samples prepared, a shift of −18 cm −1 (from 1224 to 1206 cm −1 ) was observed (Fig. S5) in the ν as (OSO 3 − ) band as the amount of CA increased. Servaty et al. argued that the occurrence of the shoulder located at 1261 cm −1 might be due to non-symmetric water-mediated interactions between the Na + counterions and the sulfonic group [72]. As increasing amounts of CA were added to the CSA solution, the overall pH of the solution decreased. The pH thus approached the pKa for -OSO 3 H (about 2.6), [eOSO 3 H] ≈ [eOSO 3 − ] and changes in the local chemical environment around the sulfonic groups occurred.

Morphological studies of the prepared electrolytes
As reported earlier, the combination of CSA and CA led to the formation of powders and films, depending on the value of X. Given that these two series of materials have very different properties, the discussion that follows henceforth will be focused exclusively on the films. The results obtained for the powders are reported in the supplementary material. Films were formed at room temperature in a period of 4-5 days, following the addition of a concentrated CA solution to a 4% aqueous CSA solution. These translucent films were produced for all combinations of CSA and CA where X < 82.3% (Table 1).
The POM images confirm that CSA is practically amorphous (Fig. 4A  and B). However, when CA was added to the solution, isolated birefringent entities emerged in all the films, indicating the presence of crystalline regions (Fig. 4C-H).
The SEM images of the top and bottom surfaces, as well as the cross section of films with X < 60.8 are represented in Fig. 5. The smoothest surfaces were obtained for the CA:CSA(X) samples with X = 60.8 (Fig. 5D) and 43.6 ( Fig. 5D). At higher X, the surfaces became more irregular, culminating with clearly visible cracks in X = 82.3 (Fig. S6 in the Supplementary Material).
The corresponding EDSs mapping images for CA:CSA(X) films with X = 43.6 and 60.8 are shown in Fig. 6. The distribution maps for the more concentrated films can be seen in Fig. S7 in the Supplementary Materials. The obtained profiles for the distribution of S and N atoms (present only in CSA) demonstrate that the polymer is homogeneously distributed in the analyzed area. Moreover, for these films, the profile for the Na + cations matches that seen for the sulfonic group, suggesting that for CA-poor films where X ≤ 60.8, the sulfonic groups are bound to the Na + cations.
The full scale view of the film placed on the AFM analysis support shows that the sample CSA/CA(60.8) is homogenous (Fig. 7E) and the film topographical analysis revealed a medium roughness of 2.99 nm, which is consistent with the film macroscopic appearance. A close analysis revealed pore-like structures of different sizes and with a medium depth of 50 nm throughout the film (Fig. 7A-D). Moreover, it is also possible to see non-uniform structures (bright angular structures on the topographical images) that seem to have developed as the film dried out, since they are in different planes. These structures did not  suffer any detectable morphology alteration by the action of the cantilever passage and their size range from 0.88 to 2.60 μm in length. Taking into consideration the film formation and the characteristics showed by the AFM analysis, it is quite possible that these angular structures are indeed crystallites formed during the solvent evaporation process, as previously described by [59].

Thermal stability
According to the TGA data (Figs. 8 and S9 in Supplementary Material), thermal decomposition of all the CA:CSA(X) electrolytes with X ≤ 82.3 invariably occurs in two steps: (1) loss of water below 120°C (5-10% weight loss) and (2) major thermal degradation with onset at 120-140°C.
As the pH decreases, the strength of the electrostatic interaction first increases, and then, at very low pH values, decreases due to the increment of the repulsion of the protonated groups. A decrease in the water content of the solids is an indication of a more compact CSA conformation [63,76,77,78], which is prone to increase any repulsion effect on the protonated groups. On the other hand, Volpi et al. pointed out that the presence of acidic compounds, such as the CA used in this work, leads to an acid-catalyzed hydrolysis of the glycosidic linkages at moderate temperatures [79]. Such a mechanism might help explain what is happening here. Moreover, the comparison of the degradation onset values of the CA:CSA(X) compounds with those found for DESs obtained from combinations of simple sugars and organic acids [26,28], further suggests a thermal decomposition mechanism affecting primarily the CSA structure, as no gain in thermal stability was observed when a simple sugar is replaced by a polysaccharide, such as CSA.

Ionic conductivity
The Arrhenius conductivity plot of CA:CSA(X) electrolytes with X = 43.6, 60.8 and 82.3, shown in Figs. 9A and S10 (in Supplementary Material), for data, demonstrates that at room temperature CA:CSA(43.6) has a lower ionic than the polysaccharide precursor, whereas the remaining two films tested (X = 60.8 and 82.3) feature a proton conductivity higher than CSA. These results, along with those observed for the powdered samples ( Fig. S14 in Supplementary Material), indicate that the proton conductivity enhancement effect due to the presence of CA is only observable in a limited range of X. Low and high X values give rise to electrolytes with poorer ionic conductivity.
The effect of composition can be rationalised admitting chemical interaction between the mixture components. Indeed, CA is expected to interact with the anionic groups present in CSA.
As the CA molecule possesses two terminal carboxylic groups, the interaction probably occurs through cross-linking between the carboxylic groups of CA and the anionic groups present in the polysaccharide (carboxylate and sulfate groups, for low and high CA content, respectively) (Fig. S11, in Supplementary Material). Similar reaction schemes have been proposed for electrolyte systems formed between chitosan and organic acids [47][48][49]80]. The addition of CA to the CSA allows the formation of intermolecular interactions between CSA and CA at the expense of the existing intermolecular interactions between the amide and the carboxylic groups of the polysaccharide, resulting in the enlargement of the distance between the CSA chains with the increase of CA content and in an increase of the tensile strength of the electrolyte [47].
The ionic conductivity data can be explained in terms of the content of CA present in the electrolytes.
As indicated in Section 3.1, a freshly prepared solution of CSA had a pH of 6.0. The addition of CA leads to a decrease of the pH, and thus should lead to an increase of ionic conductivity. However, with CA acting as a cross-linker (Fig. S11), the low content of CA might not be enough to improve on the mechanical properties of the CSA films. The lower ionic conductivity values observed for CA:CSA(43.6) can thus be attributed to a low content of CA, which leads to the formation of films with poor mechanical properties and low intermolecular interactions between CSA and CA [47,65,66]. In the same way, the enhancement in ionic conductivity observed for CA:CSA(X) electrolytes with X = 60.8 and 82.3 can be explained with the higher amount of CA which led to   , and a remarkably different morphology than that observed for the remaining electrolytes of this series. This particular sample, which seems to be at the transition between the formation of the film and the powders formed at higher CA content, still possesses enough amorphous content, as observed by POM measurements, which is believed to be responsible for the proton conductivity. The temperature dependence of the ionic conductivity is irregular (Fig. S10 in Supplementary Material). The ionic conductivity increases for temperatures up to 50°C, which coincides with the CSA gel-sol transition (Fig. S8 in Supplementary Material). Above this transition temperature, the ionic conductivities decrease, probably due to the loss of mechanical properties. In addition, the CA:CSA(82.3) sample exhibits the highest proton conductivity at 50°C.
The conductivity values measured at room temperature and under variable relative humidity (RH) are plotted in Fig. 9B. The proton conductivity of the membranes is highly humidity-dependent, with values ranging from 1.4 × 10 −7 S cm −1 for CA:CSA(82.3) at 30% RH to 3.7 × 10 −2 S cm −1 for CA:CSA(60.8) at 98% RH, confirming that these membranes are highly hydrophilic. As expected, and in accordance with the results regarding the effect of temperature (Fig. S10 in Supplementary Material), the electrolyte with the lowest CA content (X = 43.6) shows the lowest proton conductivity values for RH higher than 30%. At 98% RH the conductivity of CA:CSA(43.6) reaches 2.1 × 10 −4 S cm −1 , which is one order of magnitude lower than that measured for CSA. At RH higher than 50% the CA:CSA(60.8) film presents the highest proton conductivity. One would expect higher proton conductivities for CA:CSA(82.3), since according to EDS data all the sulfonic groups at the surface are protonated. However, both the spectroscopic and the morphologic data point to the presence of a secondary phase consisting of pure CA. The lower values here observed, which are also in accordance with the ones seen in Fig. 9A and S10, are consistent with the occurrence of secondary processes such as the CA dimerization observed in Section 3.2, and of a lower intermolecular   interaction between the CA and CSA than the one observed for CA:CSA(60.8).
The higher conductivity of CA:CSA(60.8) may be probably related to the formation of a system with an amorphous structure and improved dimensional stability (crack-free), that contributes to the overall increase of proton mobility [47]. Such apparent inconsistency may also result from the viscous deformation of the membranes at high humidity, leading to uncontrolled changes of the geometric factor (e.g. the decrease of membrane thickness) used to estimate the conductivity. In fact, after the measurements at 98% RH the samples were found in a gel-like film, a fact that confirms the hydrophilic character of the materials and is a strong evidence that changes of the sample geometry may indeed occur during the measurements. These observations indicate that the CA:CSA(X) electrolytes may be not suitable for application in hydrogen/oxygen fuel cells, since they are not stable enough to withstand high humidity environments. The materials proposed here remain, however, suitable for other electroactive applications, where they can act as the ion-conducting layer in electrochromic materials or even batteries.

Concluding remarks
While CSA is widely researched for biomedical applications, seldom work has been done in exploring its applicability as electroactive material. And yet, the presence of sulfonic groups makes it an interesting option for these applications. This study aimed at understanding how CSA could function as a host polymer for PEs.
The combination of CSA with CA afforded materials shown to possess different properties, according to their specific mass ratio: whitish polycrystalline powders at high CA content, or transparent amorphous films at lower CA content. CA is shown to work as a cross-linker, interacting with the anionic groups present in CSA. The optimal CA mass ratio was observed for CA:CSA(60.8), with lower and higher CA mass ratios yielding PEs with worse mechanical properties and lower ionic conductivities.
The results reported here indicate that, although the present systems cannot be directly applied in a hydrogen/oxygen fuel cells, they can still be combined with an inorganic matrix in order to minimize their hydrophility. This study opens interesting prospects of research for CSA-based materials as electrolytes for solid state electrochemical devices, as long as their operation does not lead to the formation of water and does not require ionic conductivities as high as those needed in the case of batteries and fuel cells. Consequently, the present electrolyte system would be suitable, for instance, for electrochemical devices. Moreover, although the materials here reported were obtained by combination of CSA with CA, this strategy could be virtually extended to any combination between an anionic polysaccharide (e.g., CS, hyaluronic acid or heparin) and any carboxylic acid, thereby offering new research possibilities. Indeed variables, such as the number of anionic groups and chemical composition of the monomeric unit, in the case of the polysaccharide, and the number of available protons and the geometry, in the case of the carboxylic acid molecule, can be explored in order to further customize new films.

Data availability
The raw/processed data required to reproduce these findings cannot be shared at this time as the data also forms part of an ongoing study.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.