Binder Influence on Polyantimonic Acid-Based Membranes’ Electrical Behavior for Low-Temperature Fuel Cells

Abstract

The development of innovative proton-conducting materials for low-temperature fuel cells (FCs) is, today, a central topic among the scientific community. Polyantimonic acid (PAA) is characterized by high conductivity and sufficient thermal stability; however, PAA-based solid membrane fabrication with high proton conductivity remains challenging. Additionally, PAA cannot be compacted into solid shaped electrolytes without a binder. In a previous work, using a fluoroplastic binder, the authors fabricated and investigated proton conductivity of bulk PAA-based membranes in the temperature range 25–250 °C. In the present research, the authors opted to use another binder, poly(vinyl alcohol), PVA (which already allowed to obtain PAA sensors with higher sensitivity to moisture, low hysteresis, and similar aging than the produced previously with the fluoroplastic binder), for fabricating new solid membranes. The sample’s structure and morphology were studied using diverse experimental techniques (Thermogravimetric analysis, X-ray diffraction analysis, etc.). Electrical Impedance spectroscopy, EIS, was used to assess the electrical response and respective time stability of the membranes; it also allowed the development of an equivalent model circuit to better interpret the samples’ electrical behavior and respective contributions. The samples with 20 wt% PVA content showed improved protonic conductivity and chemical stability up to 100 °C, when compared to previous prepared and reported ones using the fluoroplastic binder.

1. Introduction

The urgent need for green and stand-alone energy-related device technologies is helping to increment research in the field of proton-conducting materials for fuel cells (FCs) construction [1,2,3]. In fuel cell technology, polymeric membranes are most used for membrane electrode assemblage fabrication [2,4], once they are known by being easily processible at a low cost. Nevertheless, factors such as high gas molecular bonding and low chemical and thermal stability, can easily weaken membrane behavior [5]. Membrane electrode assemblies (MEAs) are one of the assemblage strategies used to construct FCs; they consist of solid membranes (electrolytes) placed between porous electrodes. When a proton conductor is used as eletrolyte, protons enter the solid membrane and move in the direction of the cathode; consequently, the eletrolyte material as well as its shape, must be carefully chosen considering the operation conditions of the fuel cell, like working temperature, type of fuel, moisture presence, etc. [6,7]. Allowing electrolytes to be used as membranes generally involves improving properties such as high conductivity and fast proton transport, temperature and structural stability, and chemical inertness, among others. Regarding protonic conduction, it is associated with OH⁻ or H₃O⁺ ion migration in aqueous environments (this usually forces operating at temperatures below 100 °C). However, the goal is to achieve conduction by free protons, H⁺, between stationary host anions, diminishing the need for molecular water in the structure and allowing the working temperature to rise. High-temperature proton-conducting polymers utilize a stationary host structure, upon which protons jump from site to site, added by sulfonation or phosphonation.

For instance, “Nafion” membranes, even if they display low conductivity at low water content and low mechanical strength at high temperatures [8,9], have started to be used in quite a few types of fuel cells [10,11], due to the found conduction properties and relatively high perm selectivity.

Nevertheless, operating temperature reduction has been effectively attained due to the availability of a wide range of candidate materials. System costs were also reduced, and long-term durability was improved. Enhanced grain boundary interfacial area due to nanometer-sized materials is described in diverse proton conductor reports available in the literature. For instance, Miyoshi et al. [12] described enhanced interfacial proton conductivity at low temperatures for yttria-stabilized zirconia. They found that when grain sizes were smaller than 100 nm, the interfacial protonic conduction was higher, fact ascribed to the absorbed water near the grain boundaries, which grew when grain size decreased. The same was observed by Maglia and his team [13] for Titania.

Consequently, solid proton conductors with a polymeric matrix are becoming widely used for MEA fabrication in fuel cell technology [2,4,14]. Tang et al. [8] reviewed long-term stability of FCs and concluded that the decomposition of the membrane is responsible for the decrease in the observed conductivity. As a consequence of the referred factors, several materials that have revealed to possess proton conductivity, like fluorite and perovskite-based inorganic compounds [15,16], and some acid salts with general composition of M_mH_n(AO₄)_(m+n)/2 (M = K, Rb, NH₄, Cs; A = S, Se) or of CsH₂(RO₄) (R = P, As) [17,18], have been assessed by researchers; however, regardless of exhibiting higher conductivities than others, for the majority of them, phase transitions occur that either lower the conductivity or reduce the chemical stability as temperature rises.

Regarding protonic conduction, one of the main contributions is due to the Grotthuss chain mechanism [19]. In it, see Figure 1, proton transport, which is an energy-activated hopping process, requires breaking and restoring oxygen/hydrogen, O–H, bonds; consequently, water molecules dissociate into hydroxyl groups (OH^-) and protons (H⁺), and while hydroxyl groups are incorporated into oxygen vacancies, protons form covalent bonds with lattice oxygen. In addition, the vehicle mechanism is also involved in the charge transfer process of protonic conductors [20], see Figure 1. In this other mechanism, protons do not migrate as H⁺ but as H₃O⁺, ${\mathrm{N}\mathrm{H}}_4^{+}$, among other formed ions, bonded to a “vehicle” such as H₂O, NH₃, etc. The “uncharged” vehicles move in the opposite direction. Proton hopping arises from one H₃O⁺ quickly rotating nearest to water molecules, being the overall conductivity, due to this mechanism, strongly dependency on the vehicle diffusion rate.

Recently, for usage as a proton-conducting electrolyte for low- and intermediate-temperature FCs, polyantimonic acid (PAA, Sb₂O₅∙nH₂O, 2 < n < 6) has been regarded as a candidate. Indeed, as a crystalline inorganic ion exchanger [22] it exhibits up to 400 °C of good chemical and thermal stability and above average proton conductivity (up to 10⁻¹ S/cm) [23,24,25]. In addition, it is insoluble and non-toxic. PAA antimony oxide in a cubic pyrochlore structure exhibits corner-sharing (SbO_6/2)⁻ octahedra connected in the anion skeleton, the reason for why it carries an excess negative charge [26]; the latter is balanced by an H⁺ or by an H₃O⁺ ion, or even by another interchangeable cation (Na⁺, Sr²⁺, Ag⁺, etc.). Furthermore, it is believed that the PAA dominant charge transfer mechanism is the Grotthuss one [27]; however, this assumption is not the consensus as electrochemical characterization data do not fully supports it. Still, reports about PAA property dependency on the preparation method are available in the literature [26,28]. Moreover, studies gathering evidence about electrical and transport properties of PAA have been conducted [27,28], all looking simultaneously for increased proton conductivity [29,30] or chemical structure organization with a wide temperature working range and decreased activation energy [31].

However, even if they possess the described ideal characteristics, pure PAA solid membranes are not easily obtained once they are mechanically unstable, and consequently easily brittle; therefore, it is mandatory the use of a binder to mold them [32,33]. In the literature, different ligands have been depicted, among which polytetrafluoroethylene (PTFE or Teflon^®) [34], polyvinylidene fluoride (PVDF) [28], and fluoroplastic [33] can be referred. In particular, the combination of PAA with PTFE in a work conducted by Vandenborre et al. [34] made it possible to evaluate the mechanical stability of the membranes up to 190 °C and to verify that they did not present corrosion problems or significant resistive losses when compared to Nafion™ membranes. It must be stressed that the literature mentions that only film-geometry membranes were obtained, except when poly(vinyl alcohol), PVA, was used. Yaroshenko et al. [31] fabricated, using PVA as binder (PVA content between 80 and 100 wt.%), bulk PAA decorated membranes. They found that proton transport was realized through the binder in the membranes with PAA content lower than 10 wt.%. In addition, and in the presence of moisture, even with a low PVA content in the membrane, the binder also contributed to the total conductivity.

In former work by the authors [35], the incorporation of PVA as a binder allowed an increase in the conductivity of the membranes that were evaluated as moisture sensors; the solid membranes displayed enhanced sensitivity to humidity, higher than the observed for the ones produced using a fluoroplastic binder (for similar contents of binders and PAA). The present work aimed to investigate the influence of PVA binder content on the structure and electrical properties of bulk PAA-based membranes and compare them to those fabricated using the fluoroplastic one [33]. The authors meant to obtain increased conductivity and chemical stability at low temperatures. Amounts of 10 and 20 wt.% of PVA were used to fabricate bulk PAA membranes and, simultaneously, to ensure direct contact between grains. Based on the Electrical Impedance Spectroscopy (EIS) assessment conducted, an equivalent model circuit interpreting the samples’ electrical behavior and respective contributions was developed; the model allowed us to identify and understand the conduction mechanisms contributing the overall observed electrical response. The samples with 20 wt% PVA content showed improved conductivity and chemical stability up to 100 °C.

2. Materials and Methods

2.1. Powder and Binder Preparation

PAA powder and binder synthesis were already reported in detail previously [35]. Nevertheless, a brief outline is given here for the sake of clarity.

Antimony (III) chloride (SbCl₃, Alfa Aesar, 99.0%, Kandel, GmbH, Karlsruhe, Germany) was used as a precursor for PAA preparation: SbCl₃ powder was dissolved into 12M hydrochloric acid (HCl, Alfa Aesar 36%, Kandel, GmbH, Germany). Then, an HCl solution was dropwise added to the SbCl₃ one at room temperature under magnetic stirring (200 rpm). Subsequently, a 6 wt.% solution of hydrogen peroxide, H₂O₂ (Alfa Aesar, Kandel, GmbH, Germany) was also added at room temperature in a dropwise fashion to the previous one, to achieve Sb(III) oxidization into Sb(V). Afterwards, the mixture was left under intensive stirring (400 rpm) for at least one hour to fulfill the oxidation of Sb (III). Subsequently, the obtained solution was heated up to 60 °C under intensive stirring (325 rpm) for 2 h, to accomplish the full decomposition of hydrogen peroxide. In sequence, tSbCl₅ solution hydrolysis was carried by gently adding to it a large volumetric surplus of deionized water (about 1:50) at room temperature under stirring in a dropwise fashion. The obtained white precipitate was left under the mother solution for crystallization at room temperature for 7 days, after which the solution was filtrated using a Büchner funnel. The filtered product was washed with distilled water until neutral pH was reached, i.e., the washing procedure was repeated until the presence of chloride ions was not detected in the washing waters tested containing the filtered product, using a reaction with a silver nitrate solution (AgNO₃, Alfa Aesar ACS grade, Kandel, GmbH, Germany). Finally, the obtained crystalline powder was placed on top of filter paper and dried in a muffle furnace at 60 °C until no weight variation was detected for at least 8 consecutive days; then, and before using it for molding the solid membranes, a short grinding procedure in a stainless-steel holder with two stainless steel balls, for 10 min, was executed for homogeneity purposes.

Regarding the binder used in this work, poly(vinyl alcohol) (Merck-Schuchardt 72000, KGaA, Darmstadt, Germany) was chosen. To avoid the formation of agglomerates, poly(vinyl alcohol) was gradually added to small portions of deionized water at room temperature under magnetic stirring; the obtained jellylike mixture was afterwards heated up to 80 °C, still under constant magnetic stirring, until no granules or agglomerates of poly(vinyl alcohol) were observed; finally, it was left to rest for 24 h, allowing for the elimination of bubbles produced during the agitation process.

2.2. Solid Electrolytes Preparation

Sample preparation methodology has also been previously depicted in the literature [35]. However, for the sake of clarity, it has been resumed here. Tablets with dimensions of 10 × 7 × 3 mm were produced in different proportions of PAA and PVA, namely 80:20 and 90:10 (having the respective amounts of each component been weighed and mixed), from here ahead, designated PP20 and PP10, respectively. Before molding, the solid membranes using the powder mixtures were placed inside a stainless-steel holder, and using stainless steel balls, they were grinded in a mixing mill for 10 min; this was performed with the purpose of achieving higher homogenization of each mixture. The solid samples were finally shaped by sequentially applying two uniaxial forces to the mold containing each prepared powder mixture; the first was a force of 70 kgf/cm² that was maintained for 1 min, while the second that was sustained for 30 min had an amplitude of 30 kgf/cm². Finally, the unmolded solid samples were left to dry for at least 5 days in ambient conditions

2.3. Structural Characterization of Samples

Structural characterization of the samples was performed using diverse techniques. Thermal properties and the consequent existence of phase transitions were investigated via thermal analysis in the temperature range of 30–400 °C (NETZSCH TG 209F1 Libra) in a synthetic air atmosphere with a heating rate of 10 deg./min. Structural analysis was performed utilizing in situ X-ray diffractometry, XRD, (to assess the evolution of PAA phase compositions) for the temperatures 25, 50, 75, 100, 150, and 200 °C (Philips X’pert, Malvern Panalytical, Malvern, United Kingdom, Bragg–Brentano geometry, using Co Kα irradiation, 1.78897 Å, scanning speed 0.025°/min, 2θ range 30–80°, equipped with a high-temperature camera Anton Paar HTK 16).

2.4. Electrical Response Measurements

To assess the sample’s electrical behavior with temperature, electrical impedance spectroscopy, EIS (Agilent 4294A, Agilent, Santa Clara, CA, USA), was utilized in the frequency interval between 400 Hz and 40 MHz, using a sinusoidal-shaped signal with a maximum amplitude of 0.5 V and no DC bias (for the sweeps, frequencies were defined using logarithmic variation in the defined test range). The used configuration was the four points one, where a single pair of electrodes is used by the equipment to simultaneously excite and measure the sample response; the integrated bridge allows to separate voltage and current circuits, allowing for a precise assessment of the impedance of the sample. It must be pointed out that the Agilent equipment has inbuilt Kramer-Krönig transformations and that, consequently, the obtained data integrity is assured. For the electrical measures, the samples were placed inside a closed chamber (possessing a volume of approximately 6.5 L) on top of a support equipped with measuring contacts. In the support holding the solid samples, two other elements are present, a type K thermocouple and two heating resistances, all connected to the temperature controller (GEFRAN model 1000, that runs a Proportional Integrative Derivative, PID, algorithm), used to regulate the desired test temperatures with an accuracy greater than 1 °C. All electrical measurements were carried out with a fixed flow rate of 5 L/h of dry air. Employing a mask, two electrodes were printed over the upper surface and on the extremities of each sample using gold ink (Gwent Group, Gold Polymer Electrode Paste C2041206P2) cured at 80 °C for 60 min in a muffle (Memmert, Model 200, GmbH + Co.KG, Schwabach, Germany). For all compositions, two specimens were used in the assessment, and for each one, five acquisition runs were performed for all test temperatures; the data displayed are always the mean values of the five runs for each measure for all samples of all compositions.

3. Results and Discussion

3.1. Structural Data Discussion

As already mentioned, the present work intended to prepare new PAA solid samples using PVA as a binder and compare their structural and electrical properties with samples composed of PAA and a fluoroplastic binder fabricated with similar binder content proportions (90:10 and 80:20 wt.%) named 90PAA and 80PAA, which were previously reported [33]. In a previously published study [35], the present study solid samples’ structure and morphology were analyzed and discussed using Scanning Electron Microscopy (SEM) and Transmission Electron Microscopy (TEM) techniques; overall, the PP20 sample was found to be the one possessing more adequate properties for reaching the purpose at that time, which was to look for enhanced moisture sensitivity. In summary, SEM images of the surfaces of the samples as well as their cross-sections revealed that the solid electrolytes were composed of nanosized PAA agglomerated particles. The comparison of the SEM images allowed us to further conclude that the increase in binder content induces (1) a decrease in the agglomeration degree of PAA particles in the samples and (2) an overall roughness and more uniform structure of the PP20 sample. Consequently, PP20 was further evaluated using TEM; from the images, it was possible to reckon the presence of particles with sizes in the nanometer range, consolidated with each other by the PVA binder. In addition, Electron Energy Loss Spectroscopy analysis (EELS) performed along with TEM allowed us to corroborate that the sample was mainly composed of antimony, Sb, mostly in the oxide form. The resulting comparison between both types of fabricated PAA solids using the two types of binders allowed us to demonstrate that the (1) PVA samples’ mesoporosity exhibited higher levels than fluoroplastic ones; (2) in general, but mainly for the case of PVA, the binder fills up the free spaces between the pure polyantimonic acid particles, allowing the direct transfer of protons through the bulk of the samples.

The present study produced and analyzed samples composed of PAA and PVA in the proportions 90:10 and 80:20 wt.%, named PP10 and PP20, respectively. Thermogravimetric (TG) analysis of samples PP10 and PP20 in the range of 30–400 °C is displayed in Figure 2a, while in Figure 2b the TG curves for 100PAA (pure PAA powder), 90PAA, and 80 PAA are depicted, obtained in the interval between 20 and 370 °C, sourced from a previous publication by the authors [33], for comparative purposes.

As can be seen, a gradual mass loss with increasing temperature is observed across all samples, which is characteristic of polymeric materials undergoing thermal degradation [36]. However, the rate and extent of mass loss varies between samples, as summarized in

Table 1. Mass losses calculated from the TG curves of 100PAA, 90PAA, and 80PAA samples, adapted from reference [33] and from powder, PP10, and PP20.

Sample Ratio (PAA:PVA Binder)	Δm (%) < 300 °C
	Fluoroplastic	PVA
100:0	14.8 [33]	11.1
90:10	14.7 [33]	12.4
80:20	14.5 [33]	15.9

. A significant tendency observable for the fluoroplastic is that increasing the binder content leads to reduced mass loss, indicating improved thermal stability. In contrast, the PVA-based samples (PP10 and PP20) exhibited the opposite behavior, where higher binder content resulted in more significant mass loss, suggesting a lower thermal resistance [36]. Specifically, the PVA-containing samples exhibit two distinct degradation stages: in the first, a sharp mass loss around 200 °C, corresponding to the breakdown of lower-molecular-weight fractions and water evaporation, while in the second, a more pronounced degradation above 300 °C marks the decomposition of the polymer backbone, which is more abrupt than that observed with the fluoroplastic samples [37,38,39].

In contrast, fluoroplastic samples display a more gradual mass loss and greater stability around 300 °C, which suggests that fluoroplastic enhances thermal resistance, slowing down the degradation process and making it more controlled. Based on the results, it is possible to conclude that PVA is less heat-resistant once there is a more significant mass loss in the molded solid samples.

However, an interesting observation arises when comparing the 100PAA and the present work synthesized PAA: both stand as pure powders yet exhibit different mass loss values of 14.8% and 11.1%, respectively. This discrepancy can be attributed to several factors: higher adsorbed water content by 100PAA, consequently retaining more water initially, will generate a more significant initial mass loss as water starts to evaporate at lower temperatures; or surface area effects—a higher exposed surface area facilitates dehydration, accelerating initial weight reduction [36,40].

The difference in the mass loss values for the different ratios is 3.7% for pure PAA, 2.3% for 90:10, and 1.4% for 80:20 (PAA/PVA binder proportions). Despite minor variations, all mass loss values remain below 5%, suggesting that neither PVA nor fluoroplastic significantly compromises the stability of the material. However, fluoroplastic enhances thermal resistance, as evidenced by its slower and more controlled degradation pattern.

Considering that, for the present study, up to around 200 °C, the samples containing a content proportion between PPA and binder of 80:20 in wt.% display steady and smaller mass loss than the remaining ones (similar to what occurred with the ones prepared using fluoroplastic as binder); once again that was the composition chosen for carrying the work, due to its predictable higher chemical stability up to that temperature (in line with the conclusions made for a previous work where overall, the PP20 sample was also the one found to be more adequate, even if, for the time being, the main goal was to look for enhanced moisture sensitivity).

In Figure 3, the XRD evaluation with varying temperatures for the PP20 sample is presented, while for comparison purposes, in Figure 4, a similar assessment for the 80PAA is reproduced [33]. For both samples independently of the binder used, through the X-ray patterns a slight shift in the peaks, <1°, is evident with temperature increase; this constatation suggests a reduction in the interplanar distance (Bragg’s Law), indicating a contraction of the crystal lattice with the temperature increase. Also, and again for both binder-type samples, it is possible to verify that some peaks suffer significant intensity attenuation for higher temperatures. In contrast, others even disappear in the case of the 80PAA sample (marked with an “x” in Figure 4), suggesting the occurrence of the material thermal decomposition or structural transition to an amorphous phase.

A more detailed analysis of the PP20 sample X-ray patterns in Figure 3 further permits us to confirm the previous constatation (for the (2θ) angle of 40.6°, the peak intensity gradually diminishes, with a noticeable reduction when comparing 25 and 200 °C), as well as to obtain some new insights: for instance, near the (2θ) angle of 58.9°, the visible evolution of the peak becoming broader and less intense suggests a transition to a less ordered phase, while in the (2θ) 76.0–76.6° angles range, the peak intensity decline indicates a partial crystallinity loss. The progressive weakening of the main characteristic peaks for higher temperatures signals either that the material is losing its crystalline structure, likely due to a shift toward an amorphous phase, or that thermal degradation of PVA is taking place, being more noticeable at 200 °C, where the peaks become significantly less defined, reinforcing the idea that above that temperature structural transformation may occur and chemical stability is no longer preserved. Once the objective is to fabricate solid PPA membranes for low-temperature FCs, its usage is not questioned for temperatures below 200 °C.

3.2. Electrical Behavior

In Figure 5a, represented is the electrical response of sample PP20 with varying temperatures assessed by EIS, while in Figure 5b, again for comparison, depicted is the one found for 80PAA [33]. All samples exhibit a Warburg-type contribution, indicating the existence of ionic conductivity once a characteristic linear response at low frequencies in the Nyquist plots is present. A closer look at the plots makes it possible to verify the existence of a well-defined arc attributed to both grain and grain boundary interactions. It is worth noting that their individual contributions may be more easily distinguished in the fluoroplastic-based sample than in the PVA one, once the arc radius is larger and smothered, indicating that their shape results from two different conduction contributions.

Except for the PP20 25 °C curve, all show a continuous increase in the real and reactive components of the impedance. It is also worth mentioning that, similarly to what was found for the fluoroplastic sample, PP20 presents a much more significant variation in the real and reactive components of the impedance for temperatures above 150 °C. In addition, for sample PP20, it is possible to verify that the impedance curves in the range between 25 and 75 °C are close or nearly overlap (in the case of the 25 and 75 °C ones), meaning its electrochemical behavior remains almost invariable within this temperature interval. Meanwhile, when comparing samples 80PAA and PP20, it is possible to verify that, in general, the fluoroplastic produced one possesses impedance component values reaching MΩ, which are almost one order of magnitude higher than the found for the sample obtained with PVA (impedance components magnitudes in the kΩ range); consequently, it can be concluded that in a temperature range up to 200 °C, the PVA binder ensures PPA chemical stability, as well as improved and more steady conduction performance. This fact highlights how the PAA-based material’s electrical response is conditioned by the binder used on solids fabrication, and it reinforces PVA as a great choice when looking for enhanced FC membranes that use PAA.

The previous assumption is confirmed by examining the activation energies, Ea, a function of the test temperature for the electrical measures performed in the first assessment of the membranes; see

Table 2. Activation Energy function of temperature.

Temperature (K)	ln (τp)	Activation Energy, Ea (eV)
298.15	−15.49	0.000
323.15	−15.63	0.011
348.15	−15.20	−0.025
373.15	−14.59	−0.082
423.15	−14.79	−0.072
473.15	−13.33	−0.252

. They were estimated via the Arrhenius plot seen in Figure 6, for which the relaxation frequencies for the visible semicircles in the Nyquist spectra were used. The slight variations observed, even if more evident for 200 °C, proves the membrane chemical stability, as shown by the improved and more steady conduction performance up to 200 °C.

Once the material’s electrical response lifetime is also an important parameter, an examination was conducted on possible aging effects on the PP20 sample. For that, EIS data were recorded for the same initial characterization temperatures across diverse time intervals, specifically, 1 week, 1 month, and 3 months, after obtaining the first EIS spectrums, designated from now on as day 0; see Figure 7. To enrich the discussion, a detailed comparative analysis is depicted in Figure 8, where the EIS spectrums are combined by test temperature. From the Nyquist plots, it can be confirmed that independently of the time-lapse for which the sweeps were taken, the general tendence is that both real and reactive components of the impedance increase with increasing temperature up to 200 °C (the exception is EIS data for day 0 assessment, where from 25 to 50 °C there is a slight decrease). Another constatation is that aging effects vary; indeed, for test temperatures up to 100 °C, aging is attenuated once there is an increase in the real and reactive components of the impedance, up to the 1-month measure, which then diminishes; however, above 150 °C, the trend is reversed, and the impedance components magnitude keep increasing after 1 month.

In addition, for all tested temperatures, independent of the time instant for which the seeps were performed, the slope of the visible Warburg component, which is associated with the protonic conduction contribution, keeps its value almost unchanged. Also, regarding time lapse assessment, for temperatures above 100 °C, an increase on the imaginary part of the impedance is observable, suggesting the development of more substantial polarization effects. This evolution likely reflects an increase in the distribution of pore sizes, which can significantly impact the material’s electrochemical properties [40]. These results highlight the significant impact of time and temperature on the impedance behavior of the PP20 sample. While higher temperatures generally lead to increased impedance components magnitude over time, certain lower temperatures (notably 50 and 75 °C) exhibit opposite tendencies, suggesting complex interactions between material degradation, structural changes, and ionic transport mechanisms. The impedance modulus values determined at 400 Hz along time and with varying temperature depicted in

Table 3. PP20 Impedance modulus at 400 Hz with varying temperature and time (initial—i, 1 week—1w, 1 month—1m, 3 months—3m).

Temp/Time	Re [Ω]	−Img [Ω]
25 °C_i	1.20 × 105	2.60 × 105
25 °C_1w	1.05 × 105	2.79 × 105
25 °C_1m	3.01 × 105	5.89 × 105
25 °C_3m	2.21 × 105	4.64 × 105
50 °C_i	1.06 × 105	2.40 × 105
50 °C_1w	2.39 × 105	3.18 × 105
50 °C_1m	6.22 × 105	6.54 × 105
50 °C_3m	4.48 × 105	5.68 × 105
75 °C_i	1.20 × 105	2.49 × 105
75 °C_1w	4.89 × 105	3.53 × 105
75 °C_1m	1.07 × 106	7.73 × 105
75 °C_3m	7.04 × 105	6.71 × 105
100 °C_i	1.46 × 105	2.39 × 105
100 °C_1w	6.62 × 105	3.75 × 105
100 °C_1m	8.79 × 105	4.92 × 105
100 °C_3m	1.01 × 106	7.62 × 105
150 °C_i	1.58 × 105	2.50 × 105
150 °C_1w	8.18 × 105	4.22 × 105
150 °C_1m	1.29 × 106	5.95 × 105
150 °C_3m	1.67 × 106	9.53 × 105
200 °C_i	5.17 × 105	3.74 × 105
200 °C_1w	1.03 × 106	5.75 × 105
200 °C_1m	1.52 × 106	7.71 × 105
200 °C_3m	2.84 × 106	1.47 × 106

further support the found behavior.

In summary, using PVA as a binder improved the structural and chemical stability and reduced aging effects at low temperatures, further profiling PAA usage as solid electrolyte membranes in FCs. Nevertheless, to better understand the electrical mechanisms behind the electrochemical behavior found in the studied composition, EIS data were used to develop an equivalent circuit model.

The one proposed, illustrated in Figure 9, consists of several key components: a Warburg element, two constant phase elements (CPE1 and CPE2), a resistor-capacitor (R//C) parallel network, and a geometric capacitance (C_GEO). This model was developed based on previous studies conducted by the authors, which analyzed the system’s behavior under varying humidity conditions while maintaining a constant temperature of 25 °C [33].

The Warburg component, characterized by its coefficient, represents the high ionic contribution from charge transport through the sample’s porous structure and binder material. This behavior has been well-documented in the literature, particularly in studies on membranes based on polyantimonic acid, where similar ionic transport effects have been observed [27,29,41].

The two constant phase elements, CPE1 and CPE2, account for charge diffusion phenomena linked to different interfaces within the system. CPE1 is primarily associated with the interaction between the electrodes and the sample material, as well as the mesoporosity of the solid sample. CPE2, on the other hand, reflects the influence of water retention within the membrane’s porous structure and its effect on the material’s conductivity. The presence of these elements implies that both the microstructural characteristics of the sample and the retained moisture play a significant role in the system’s electrochemical response. The resistor–capacitor parallel network (R//C) represents the electrical behavior associated with grain boundaries and bulk material effects. This element captures the influence of intergranular resistance and the capacitive response of the individual grains within the sample. Finally, the geometric capacitance, C_GEO, accounts for the purely geometric effects of the electrode placement configuration. Indeed, they are located at the extremities of the upper surface of the sample, creating a measurement setup analogous to a parallel-plate capacitor; consequently, the material effectively seems to serve as the dielectric medium between plates in this configuration.

The initial parameter values for the components in the proposed model were first estimated using EIS spectra analysis software, eisanalyser (Version 1.0), followed by their manual fine-tuning to achieve optimal accuracy.

Figure 10 illustrates a selection of Nyquist and Bode plots and respective fittings obtained using the proposed model for the PP20 sample as well as for comparison for 80PAA [33], while

Table 4. Fitted parameters for sample PP20 using the proposed equivalent circuit model.

T (°C)	R (Ω)	C (F)	QCPE1 (S)	nCPE1 (a.u.)	QCPE2 (S)	nCPE2 (a.u.)	CGEO (F)	Wcoeffi. (Ω × s−1/2)
25	4.12 × 102	1.03 × 10−10	7.25 × 10−6	8.04 × 10−2	3.85 × 10−9	8.84 × 10−1	3.38 × 10−12	6.33 × 10−10
50	5.20 × 102	4.97 × 10−10	7.58 × 10−6	9.13 × 10−2	4.01 × 10−9	8.89 × 10−1	3.51 × 10−12	5.29 × 10−10
75	3.48 × 103	2.17 × 10−10	6.56 × 10−6	9.54 × 10−2	4.08 × 10−9	8.83 × 10−1	3.87 × 10−12	7.23 × 10−10
100	1.32 × 104	3.81 × 10−11	6.09 × 10−6	7.48 × 10−2	4.03 × 10−9	8.91 × 10−1	4.15 × 10−12	1.33 × 10−10
150	4.54 × 103	7.72 × 10−11	6.93 × 10−6	3.55 × 10−2	3.87 × 10−9	8.88 × 10−1	3.53 × 10−12	4.24 × 10−10
200	1.05 × 104	1.62 × 10−11	1.57 × 10−6	6.50 × 10−2	2.15 × 10−9	9.20 × 10−1	6.01 × 10−12	4.52 × 10−10

summarizes the fitted parameters for the PP20 sample across the 25 to 200 °C temperature range. The close agreement between the Nyquist representations of the obtained experimental data and the modeled response confirms the effectiveness and reliability of the developed equivalent circuit. This consistency is further assured by looking at the Bode plots of both PP20 and 80PAA samples, where the proposed equivalent circuit model suitably fits the experimental data.

The illustrated examples demonstrate the model’s ability to represent impedance variations with temperature, allowing us to conclude that its consistency is maintained independently of the chosen binder.

Nevertheless, a final evaluation of the equivalent model circuit was carried for the samples prepared using PVA, in this case using again PP20 data. The standard deviation, σ, between the experimental and fitted data with varying frequency was calculated; see Figure 11. As can be seen, up to 10 × 10⁶ Hz, the estimated σ is lower than 3%, only assuming higher values for frequencies above 10 × 10⁶ Hz, due to fluctuations in the obtained experimental data for PP20, as perceived from plots in Figure 10; nevertheless, considering that PPA membranes will be used in FCs, their operation will be the low frequency range for which the model behaves quite well.

A closer inspection of fitted parameters using the model provides insights into the electrochemical response of the material with temperature. Indeed, from the data, it is possible to state that the structural stability of the material up to 100 °C seems to be kept once the capacitance of the (R//C) element, which accounts for grain boundaries and bulk material effects, exhibits a minor variation. At the same time, the resistive component displays a gradual increase (the typical variation observed for the grain resistance of semiconductors with temperature). Additionally, the low variations found for CPE1, CPE2, and Warburg contributions confirm that integrity. Indeed, the dependency of diffusion effects with temperature, which are dominant, is small, confirmed, for instance, by the line segment representative of the Warburg component visible in the Nyquist plots in Figure 5a, which keeps its overall amplitude and slope with temperature.

At 100 °C and above, the combined influence of the grain boundaries and material bulk effects (represented by the (R//C) element), as well as of the influence of water retention within the membrane porous structure (accounted for by CPE2) to the overall impedance of the sample increases, in accordance with the small structural changes observed through the situ X-ray patterns with temperature, where small shifts in some of the characteristics peeks are noticed. Indeed, while for the first ones, grain boundaries and material bulk effects, the resistive contribution increases with temperature, for the second, the influence of water retention within the membrane porous structure, which, above 100 °C starts evaporating, the CPE2 coefficient diminishes.

Finally, the small variations found for C_GEO also allow us to conclude about the slight influence of the electrode’s placement over the electrochemical response of the material temperature.

A brief comparison between the parameters fitted for the PP20 sample and those reported for 80PAA [33] confirms that the latter shows higher resistance, suggesting higher insulating properties. However, high resistance in fuel cell material membranes may limit ionic conduction, so it is unwelcome.

Resuming, the choice between PAA solid membrane fabrication using fluoroplastic or PVA as binders depends on the desired application: PVA-based materials are preferable for fuel cell membranes due to their lower resistance and better ionic conduction, while fluoroplastic-based ones offer greater thermal stability, which may benefit other high-temperature applications.

Table 5. Binder type and respective usage temperature.

Binder	Maximum Temperature	References
Polyvinylidene fluoride	70 °C	[28]
Polysulfone	120 °C	[5,28]
F-23 fluoroplastics	180 °C	[42]
Poly(vinyl alcohol)	200 °C	Present work
Polyacrylonitrile	277.8 °C	[43]

presents a brief comparison between data reported in the literature, and the present work for the maximum utilization temperature and respective binder used in PAA membranes fabrication is presented. From them it is possible to conclude that the binder chosen in the present work is one of the most resistant to the temperature, maintaining its integrity until around 200 °C. Only polyacrylonitrile is more heat resistant, starting to degrade at 277.8 °C.

4. Conclusions

In the current work, polyantimonic acid (PAA) samples using poly(vinyl alcohol) (PVA) as a binder were prepared, and their structural, thermal, and electrical properties were compared with previously prepared ones made with the fluoroplastic binder, earlier reported in the literature. The thermogravimetric analysis (TGA) demonstrated significant differences in thermal stability between the two binder systems. Fluoroplastic-based composites (90PAA and 80PAA) exhibited higher thermal resistance, decreasing mass loss as fluoroplastic content increased. In contrast, PVA-based composites (PP10 and PP20) showed increased mass loss with higher binder content, indicating lower thermal stability. The X-ray diffraction (XRD) studies further revealed that increasing temperature leads to peak shifts and intensity attenuation, suggesting lattice contraction and structural transitions. These effects were more pronounced in PVA-based samples, highlighting their tendency to transition to an amorphous phase at elevated temperatures.

EIS analysis indicated that all samples exhibited ionic conductivity, confirmed by Warburg-type contributions in the Nyquist plots. However, fluoroplastic-based composites demonstrated significantly higher impedance (reaching MΩ) compared to their PVA-based analogs (kΩ range), highlighting the impact of binder choice on electrical resistance and ionic transport. In addition, the aging studies revealed that PP20 exhibited an impedance that decreased over time at lower temperatures (up to 100 °C), suggesting potential structural modifications enhancing ionic conductivity.

Using an earlier proposed equivalent electrical model circuit, EIS spectra data were fitted to understand further the conduction mechanism behind the electrochemical behavior of the solid PVA-PAA samples.

Overall, the results highlight the significant impact of binder selection on the thermal and electrical properties of PAA-based materials. Fluoroplastic-based materials exhibit superior thermal stability and higher electrical resistance, making them more suitable for applications requiring stable electrochemical performance. While offering lower impedance and enhanced ionic transport, PVA-based materials suffer above 100 °C, with structural changes over time, impacting their long-term electrochemical stability. These findings provide valuable insights for optimizing the formulation of PAA-based materials and identifying which binders to choose for specific energy storage applications, such as in the case of fuel cells, making the choice dependent on the temperature of utilization.