(NH₄)₃PW₁₂O₄₀-H₃PO₄ composites as efficient proton conductors at intermediate temperatures

1. Introduction

Heteropolyacids (HPAs) have been recognized as the most proton conductive inorganic solids at near-ambient temperatures [[1],[2]]. Their proton conductivity, such as those from dodecatungstophosphoric acid (H₃PW₁₂O₄₀·29H₂O) and dodemolybdphosphoric acid (H₃PMo₁₂O₄₀·29H₂O) can approach > 0.1 S/cm at room temperature and under proper humidity [3]. This high conductivity is mainly attributed to the facile transportation of oxonium (H₃O⁺) or dioxonium cations (H₅O₂⁺) in the micro-channels of the crystal structure. One of the major problems that hinder the applications of these promising materials is the dehydration and concomitant loss of proton conductivity, which generally takes place above 40 °C in the normal conditions [3]. Another problem is the solubility of these HPAs whereby the mechanical strength of these materials was lost in the presence of liquid water. How to overcome these problems is a real challenge as proton conductivity at intermediate temperature, i.e. between 100 °C and 600 °C are ideal for fuel cell operations, particularly those used as power plants for automobiles. Two strategies have been proposed. One is to replace water molecules in the HPAs’ structure by other protonic moieties that can sustain high temperatures. This strategy has been realized by replacing water molecules with ionic liquids such as [BMIM][TFSI] [4]. High conductivity above 100 °C and under anhydrous conditions has been achieved. Nevertheless, the cost of ionic liquids is still too high to warrant large scale production. Another strategy is to form heteropolyacid salts which demonstrate reasonable conductivity at high temperatures and insoluble in water. Caesium based heteropolyacid salts have been investigated which shows encouraging results [5]. However, analogues such as ammonium based heteropolyacid salts, e.g. (NH₄)₃PW₁₂O₄₀ has not been investigated [6]. (NH₄)₃PW₁₂O₄₀ shows selectivity in catalyzing alkylation of amonia with methanol above 400 °C [7] and reduction of NO₂ at a temperature as low as 150 °C [[8],[9]]. It has surface area over 100 m² g^-1 which is remarkably high considering the heavy molecular weight (∼2931). Inumaru and Ito et al. found the “epitaxial self assembly” phenomenon of (NH₄)₃PW₁₂O₄₀ and categorized (NH₄)₃PW₁₂O₄₀ as “sponge crystal” since there are large residual space left between the constituent nanocrystallites [[10],[11]]. The high porosity provides ample rooms to accommodate secondary phases such as acids therefore opening possibilities to develop composite proton conductors. Mikhailenko et al. investigated the electrical properties of (NH₄)₃PW₁₂O₄₀ in dry Ar, 100% RH and in the presence of liquid water [12]. They found that the conductivity of (NH₄)₃PW₁₂O₄₀ was improved several magnitudes by introducing liquid water (10^-5-10^-4 S/cm in dry Ar and 3.3-8.0 × 10^-2 S/cm mixed with 10% water). Nevertheless, water is easy to evaporate at high temperatures and replacing water with stable acid would be a good choice to further improve the proton conductivity beyond 100 °C. In this paper, we introduce phosphoric acid to replace water and synthesized (NH₄)₃PW₁₂O₄₀-H₃PO₄ composites. The proton conductivity of (NH₄)₃PW₁₂O₄₀-H₃PO₄ composites at high temperatures was investigated. The high stability and insolubility of (NH₄)₃PW₁₂O₄₀ as well as its high uptake for H₃PO₄ signifies promising application as a proton conductor in the intermediate temperatures.

2. Experimental

2.1. Material synthesis

(NH₄)₃PW₁₂O₄₀-H₃PO₄ composites were synthesized by precipitation method: calculated amounts of H₃PO₄ (Aldrich) and NH₄HCO₃ (Aldrich) were mixed up and dissolved with deionized water simultaneously to obtain a transparent solution. The solution was then dripped into H₃PW₁₂O₄₀ solution under vigorously agitations in order to get homogeneous mixtures. White precipitations were produced immediately and the mixtures were heated at 200 °C to remove excess water. The resultant white powders were labeled as NP0, NP3, NP5 and NP10 according to the molar ratios of NH₄HCO₃/ H₃PW₁₂O₄₀/ H₃PO₄ = 3/1/0, 3/1/3, 3/1/5 and 3/1/10. The powders were stored at 90 °C for 12 h and collected for measurements.

2.2. Analytical procedures

Crystal structure and phase purity were examined by X-ray powder diffraction (XRD) technique on a Bruker D8 Focus diffractometer (Bruker, Germany) using CuK_α1 radiation (λ =1.5405 Å) and CuK_α2 radiation (λ =1.5444 Å) as incident radiation. 0.01° step size with duration of 0.1 s was adopted for data collection. The microstructures were examined under microscopic conditions using a field emission scanning electron microscope (FESEM, Hitachi S4800, Japan) and a transmission electron microscope (TEM, JEOL JEM-2100, Japan). Fourier transform infrared spectrum (FT-IR) was chosen to inspect the interior bonding conditions for all samples. KBr pellets were made for each sample and the IR spectrum was collected on the PERKIN-ELMER Spectrum GX FTIR systems from 400 cm^-1 to 4000 cm^-1 with a resolution of 4 cm^-1. Thermal analysis was carried out on a Rheometric Scientific TG 1000 M 1000 M + TA instruments SDT 2960 with heating and cooling rate of 5 °C/min under flowing air at a rate of 20 mL/min.

2.3. Conductivity measurements

The conductivity measurements were carried out by the AC impedance method over the frequency range from 1 MHz to 10 mHz with an amplitude of 10 mV on a Zahner electrochemical workstation. Pellets were used for measurements and were obtained by pressing powders under a pressure of 3 × 10³ kg/cm² (diameter 13 mm, thickness 4-5 mm). PTFE bonded carbon (Carbot Vulcan 72R) was dabbed on both sides of the pellets which served as electrodes. Carbon paper (Torory TGPH-090) was subsequently pressed on as current collector. The pellets were mounted into furnaces for impedance measurements. Data were collected from the highest temperature to the lowest on cooling after holding at least one hour at each temperature to reach equilibrium. Air and wet air (humidified at 20 °C with moisture content about 3%) were applied to create different atmosphere and to evaluate conductivity.

3. Results and discussion

3.1. Phase composition, thermal stability and morphology

All samples obtained are white powders. XRD analysis shows similar diffraction patterns (Fig. 1) to standard data (JCPDS: 00-050-0305) with cubic symmetries. There is a clear shift of all diffraction peaks towards low angles implying enlargement of unit cell parameters. The calculated unit cell parameters were listed in Table 1. The lattice parameters and lattice volume were enhanced by introducing phosphoric acid. According to Inumaru [10], the particles of (NH₄)₃PW₁₂O₄₀ were composed of (NH₄)₃PW₁₂O₄₀ nanocrystallites which self-assembled with the same crystal orientations and the microporosity of (NH₄)₃PW₁₂O₄₀ comes from the residual space between these nanocrystallites. Therefore, the increase of unit cell parameters can be rationalized by replacing water molecules with larger H₃PO₄ molecules in the residual space of (NH₄)₃PW₁₂O₄₀. However, there seems a saturation of unit cell enlargement at high H₃PO₄ content as can be seen from Fig. 2. This is probably due to a complete fulfillment of H₃PO₄ in the residual space of (NH₄)₃PW₁₂O₄₀.

View original graphic| Download| PPT slide

Fig. 1.   XRD patterns of NP0, NP3, NP5 and NP10, standard pattern for (NH₄)₃PW₁₂O₄₀ (JCPDS: 00-050-0305) is also included for comparison. Main peak around 26° is enlarged on the right. Dashed line is a guide for the eye.

View original graphic| Download| PPT slide

Fig. 2.   Unite cell parameters as a function of H₃PO₄ to (NH₄)₃PW₁₂O₄₀ molar ratio: (a) cell parameter a; (b) cell volume V.

Thermal analysis was carried out to investigate the dehydration and stability of these composites. For NP0, the analysis was continued from room temperature up to 800 °C. The results were shown in Fig. 3(a). According to Belanger et al. [8], ammonium was not released from the solid until the temperature is beyond 400 °C and considerable amount of ammonium was only released at a temperature higher than 550 °C. Therefore, the weight loss observed below 500 °C in Fig. 3(a) could be reasonably attributed to the loss of water. The thermogravimetric curves showed a two-step progress at this temperature range. The first step is a linear loss from room temperature up to around 118 °C, a weight loss of 1.48% was observed corresponding to 2.5 H₂O molecules per (NH₄)₃PW₁₂O₄₀ unit. The weight loss at this region was generally considered as desorption of physically absorbed water due to the hydrophilic properties of (NH₄)₃PW₁₂O₄₀. The second one is from 118 °C to 500 °C where 1.07% weight loss occurs, corresponding to 1.8 H₂O molecules per (NH₄)₃PW₁₂O₄₀ unit. The thermogravimetric curve at this region presented a smoother slope which could be ascribed to the loss of strongly absorbed water inside the internal structures of (NH₄)₃PW₁₂O₄₀. As the temperature kept increasing, a sharp loss took place and ended up around 612 °C. If ammonium was completely removed from the solid, a weight loss less than 1.7% should be observed theoretically. The experimental value (2.54%) was significantly larger than the theoretical one, suggesting a decomposition of other constituent elements such as oxygen. XRD patterns of samples after TG analysis confirmed these observations (Fig. 3(c)). The diffraction patterns were attributed to a pure W₁₂PO_38.5 (JSPDS card 00-041-0369) and a theoretical weight loss 2.6% can be calculated which is very close to the experimental one (2.54%). From TGA analysis of (NH₄)₃PW₁₂O₄₀, it is confirmed that the structure of (NH₄)₃PW₁₂O₄₀ can sustained up to 500 °C and water start to lose quickly even below 50 °C.

View original graphic| Download| PPT slide

Fig. 3.   (a) Thermogravimetric analysis (TGA) of NP0, (b) TGA of NP3, NP5 and NP10, (c) XRD patterns of NP0 after TGA and (d) XRD patterns of NP10 after TGA. Standard patterns for W₁₂PO_38.5 (JCPDS: 00-041-0369) and (NH₄)₃PW₁₂O₄₀ (JCPDS: 00-050-0305) are also included for comparison.

TGA analysis was then performed on (NH₄)₃PW₁₂O₄₀-H₃PO₄ composites and the temperature range was set from room temperature to 300 °C (Fig. 3(b)). The corresponding weight loss below 100 °C and from 100 °C to 300 °C was listed in Table 2. It can be seen from the table that weight loss below 100 °C was quite different from NP3 to NP10 due to various extent of hydration during the preparation process. However, the weight loss between 100 °C and 300 °C increased systematically. This weight loss could mainly come from the dehydration of phosphoric acid since only 0.99% was observed for pure (NH₄)₃PW₁₂O₄₀ whereas more than 2.53% was found in the composites. It has been realized that phosphoric acid could condense into pyrophosphoric acid in a temperature of 175 °C or even lower [13], therefore, the accelerated weight loss above 150 °C for all composites could mainly contributed to the condensation of phosphoric acid. It is worth mentioning that (NH₄)₃PW₁₂O₄₀ is quite stable in the presence of phosphoric acid at this temperature range. XRD patterns in Fig. 3d revealed that the composite was stable at high acid loadings (NP10) since the diffractions of (NH₄)₃PW₁₂O₄₀ phase were maintained after TGA analysis.

The morphologies of the composites and pure (NH₄)₃PW₁₂O₄₀ were examined under SEM and TEM conditions. Fig. 4 shows the SEM images of all sample powders. Spherical particles were observed in all cases. The particles sizes were approximately in the range of 100-500 nm which was consistent with literature [6]. According to Inumaru, the particles of (NH₄)₃PW₁₂O₄₀ were composed of nanocrystallites that self-assembled together in the same orientations and the morphologies of these particles were highly sensitive to the preparation temperatures: with spherical particles in the low temperatures (∼ 20 °C) and symmetric particles in higher temperatures (∼ 90 °C). In addition, they attributed this phenomenon to the various surface energy and/or growth rates between crystal planes. This speculation could be partly true due to the small constituent nanocrystallites (4-6 nm) and exceptionally high surface energy that could be envisaged. However, in our point of view, there could be another factor that might govern the outer shape of (NH₄)₃PW₁₂O₄₀ which is the hydrogen bonds. According to Okuhara et al. Caesium based heteropolyacid salts like Cs_3-xPW₁₂O₄₀ do not have such epitaxial growth phenomenon [14] and the particle dimensions calculated through XRD were significantly smaller (24 nm for Cs₃PW₁₂O₄₀) than (NH₄)₃PW₁₂O₄₀ (> 150 nm) prepared at the same conditions. This observation could not explained by surface energies since the constituent nanocrystallites were in the same dimensions (< 10 nm) and M₃PW₁₂O₄₀ (M=Cs, NH₄, Rb, K) shared the same crystal structures. Therefore, other factors could trigger the epitaxial growth of (NH₄)₃PW₁₂O₄₀ and hydrogen bonds are well-known for their direction and strength, it is quite likely that protons in ammonium formed the hydrogen bonds like N-H…O-W and lead to the translation and rotation of Keggin ions during the local dissolution-reprecipitation process. Fig. 5 shows the TEM pictures of NP10 where spherical particles with diameters varied from 150 nm to 500 nm were observed. Inumaru found that the BET surface area and micropore volume of (NH₄)₃PW₁₂O₄₀ decreased significantly at a high precipitation temperature due to the rearrangements and formation of rigid dodecahedra. It can be seen from Fig. 6 that the particles remained spherical after aging at 100 °C for 2 h and no rigid dodecahedra were observed.

View original graphic| Download| PPT slide

Fig. 4.   Field emission SEM images of freshly prepared sample powders: (a, b) NP0; (c, d) NP3; (e, f) NP5; (g, h) NP10.

View original graphic| Download| PPT slide

Fig. 5.   TEM images of NP10 at different magnifications.

View original graphic| Download| PPT slide

Fig. 6.   FT-IR spectra of all sample powders (85 wt% H₃PO₄ is also included for comparisons).

3.2. IR spectra

Fig. 6 shows the infra red absorption spectra of pure (NH₄)₃PW₁₂O₄₀ and all composites. The spectra of 85 wt% phosphoric acid were also shown for comparisons. It can be found that the composites and pure (NH₄)₃PW₁₂O₄₀ exhibit only minor differences, implying the structure of Keggin ions are preserved in the presence of phosphoric acid. The main differences between pure (NH₄)₃PW₁₂O₄₀ and the composites lie in the range 1100-1300 cm^-1 and 2600-3300 cm^-1. Pure (NH₄)₃PW₁₂O₄₀ shows almost no absorption in these two ranges. The composites show a broad shoulder absorption peak from 1100 cm^-1 to 1300 cm^-1 corresponding to vibrations of different phosphoric species (HPO₄^2- and H₂PO₄-) [15] and relatively stronger absorptions at 2600-3300 cm^-1 which is due to the ν(O-H) vibrations [16]. These absorptions could come from the phosphoric acid in the structure of (NH₄)₃PW₁₂O₄₀ if we compare them with 85 wt% phosphoric acid where much stronger and broader absorptions appeared around the same ranges. Park et al. pointed out that hydroxyl groups that have absorption around 2900 cm^-1 have higher proton mobility compared with the ones at higher wavelength [17]. Therefore, it is clear that introducing phosphoric acid into (NH₄)₃PW₁₂O₄₀ significantly enhances the intensities of absorption peaks at this region. Another factor worth noticing is that absorption peaks at 525 cm^-1, 795 cm^-1, 887 cm^-1 and 975 cm^-1 which could be assigned to ν_s(W-O-W), ν_as(W-O_b-W)and ν_as(W-O_c-W), and ν(W = O_t) [16] were broadened progressively with increasing phosphoric acid. This phenomenon could be a sign of interactions between phosphoric acid and oxygen ions in Keggin structure such as the formation of hydrogen bonds so that the symmetry of the Keggin ions was consequently lowered. The weak absorption peaks at 596 cm^-1, 1400 cm^-1 and 1079 cm^-1 could be assigned to δ(O-P-O), ν_as(PO_a) and deformation vibration of the ammonium ions respectively [[3],[16],[18]].

3.3. Proton conductivity

The conductivity of pure (NH₄)₃PW₁₂O₄₀ under different atmospheres was shown in Fig. 7. The values are calculated from the Cole-Cole plots obtained by AC impedance spectra at the intersection to the real axis. The pure (NH₄)₃PW₁₂O₄₀, however, was not a good proton conductor even in wet air and did not obey Arrhius behavior which could be ascribed to the continuous loss of water at high temperatures [[8],[12]]. According to Mikhailenko et al., mechanisms had been proposed for the explanations of proton transport in (NH₄)₃PW₁₂O₄₀. They attributed the conductivity to the transport of hydrated protons which was developed in the vicinity of NH₄⁺ sites in (NH₄)₃PW₁₂O₄₀ structures. In our experiment, (NH₄)₃PW₁₂O₄₀ showed a much better conductivity in wet air than in air which could be the water effects and become an evident sign of proton conductions. More specifically, the differences between wet air and air were enlarged from room temperature to 220 °C due to the thermal activation and subsided above 220 °C, probably due to the dehydrations [8]. The ammonium was not expected to release from (NH₄)₃PW₁₂O₄₀ below 400 °C, so the better performance in wet atmosphere could come from the strong interactions between ammonium ions and water molecules. This phenomenon gives rise to broad potential applications such as sensors etc. since more than two orders of magnitudes of differences in conductivity was achieved in different atmospheres.

View original graphic| Download| PPT slide

Fig. 7.   (a) Conductivity (σ) of NP0 in air and in wet air (humidified at 293 K) and (b) conductivity of NP3, NP5 and NP10 in air (T: temperature).

The composites, however, exhibited a much better performance on conductivity in air (Fig. 7). The highest conductivity is 0.14 S/cm for NP10 at 170 °C which is almost 4 orders of magnitudes higher than NP3 under the same conditions. It is concerned that phosphoric acid in NP3 was discretely distributed in the so-called “spongy crystal” of (NH₄)₃PW₁₂O₄₀ where continuous channels for proton conduction could not be fully established. However, in the case of NP10 where a huge amount of phosphoric acid was added, those internal space and channels inside crystals were filled and connected well with each other therefore a high conductivity was observed. The decreased conductivity at high temperature was attributed to condensation of phosphoric acid to polyphosphoric acid where large volume decrease occurs therefore reduce the continuous channels for proton transportations.

4. Conclusion

(NH₄)₃PW₁₂O₄₀ and (NH₄)₃PW₁₂O₄₀-H₃PO₄ composites were synthesized by precipitation method. Their phase compositions, thermal stability and morphologies were investigated. It was found that (NH₄)₃PW₁₂O₄₀ maintained its structures in the presence of phosphoric acid and the phosphoric acid was believed to be preserved in the micropores of (NH₄)₃PW₁₂O₄₀. SEM and TEM analysis showed that the particles of all samples were spherically shaped with diameters of 200-500 nm. FTIR spectra confirmed the maintenance of Keggin structures and strong interactions between phosphoric acid and (NH₄)₃PW₁₂O₄₀. The pure (NH₄)₃PW₁₂O₄₀ exhibited a poor conductivity in air at high temperatures but the conductivity was improved significantly in wet air due to the water promotion effect. The composites, however, demonstrated a much better conductivity especially at high acid loadings where continuous channels for proton conduction can be established. A conductivity of 0.14 S/cm was achieved at 170 °C for NP10. Such high conductivity gives rise to broad potential applications such as electrolyte in fuel cells and other electrochemical devices.

Acknowledgements

This work was supported financially by the National Natural Science Foundation of China (No. 21401142, 51972233), the Natural Science Foundation of Shanghai (No. 19ZR1459200), the National 1000-Plan Program, the Shanghai Science and Technology Commission (No. 14DZ2261100), the South Taihu Elite Project and the Fundamental Research Funds for the Central Universities.