Functional Nanostructures from Sol–Gel Synthesis Using Keggin Polyoxometallate Phosphotungstic Acid as a Precursor

Subjecting phosphotungstic acid solutions to low pH in combination with introduction of polyvalent cations led to the formation of nanostructured microspheres of approximately 2 μm in size, as shown by scanning electron microscopy, which were almost insoluble and resistant to degradation at neutral and high pH. These microspheres were composed of secondary nanospheres with diameters around 20 nm as revealed by transmission electron microscopy and atomic force microscopy. Investigations of the crystal structure of a potential intermediate of this process, namely, acidic lanthanum phosphotungstate, [La(H2O)9](H3O)3[PW12O40]2(H2O)19, showed a tight network of hydrogen bonding, permitting closer packing of phosphotungstic acid anions, thereby confirming the mechanism of the observed self-assembly process. The new material demonstrated promising electrochemical properties in oxygen evolution reactions with the high stability of the obtained electrode material.


■ INTRODUCTION
The internationally recognized goal to reduce evolution of greenhouse gases has set focus on the development of efficient fossil-free technologies for energy production. 1 The main source of usable energy for the earth is solar light and thus solar energy conversion is one of the most addressed topics in modern research in designing and producing new materials. 2,3n attractive alternative to fossil fuel-based processes is hydrogen energy, primarily the production of electricity with the aid of fuel cells, 4 but also the use of hydrogen in reduction reactions such as the recently proclaimed HYBRIT technology for "green synthesis" of steel from iron ore. 5 A common feature of hydrogen energy technologies is the need for high-purity hydrogen gas required in large volumes.Its production is possible either via costly purification of hydrogen obtained via the water−gas shift reaction from natural gas or biogas 6 or via highly energy-demanding electrolytic water splitting. 7An attractive alternative to electrolysis is the use of photocatalytic or electrocatalytic water decomposition.In these approaches, the energy costs for hydrogen gas can be significantly reduced while not compromising its quality.The challenge, however, lies in the need for expensive components used in the making of such catalysts.Typical photocatalysts for water splitting are nanoparticles (NPs) of semiconductor oxides or chalcogenides in combination with noble metals. 8The efficient electrocatalysts applied so far also commonly contain platinum group metal-based NPs either as oxides (RuO 2 9 or IrO 2 10 ) or together with platinum NPs. 11−23 The challenge in the creation of such related nanostructures is the relatively high reactivity and solubility of WO 3 and its derivatives in both acidic and basic media. 24,25Here, in particular, phosphotungstic acid has attracted attention as a possible photocatalyst 26,27 and potentially electrocatalyst 21 .However, this compound in its hydrated form is highly soluble in water, which hinders its application.
Phosphotungstic acid is a well-studied representative of the polyoxometallate (POM) family of compounds.It is of the Keggin type, the smallest metal oxide nanoparticle of approximately 1 nm in diameter, which has a highly ordered structure consisting of a central heteroatom inside a cage of 12 transition metal atoms in their highest oxidation state, all  connected by oxygen bridges, giving the general formula XM 12 O 40 z− (X = P, Si, Ge, As, Sb; M = W, Mo; z = 3 or 4).We have previously reported complexes of Keggin POMs and oligopeptides, in an effort to model the interactions between NPs and proteins at the nanoscale. 28,29Some of these complexes displayed POM−POM contacts, approximately 3 Å in length, likely resulting from hydrogen bonds.As these were observed at very low pH, protonation of the POM likely shielded the charge, allowing for direct interactions.Acidic conditions are necessary when working with Keggin POMs, as one drawback of phosphotungstic acid, in particular, is its instability at neutral and alkaline pH. 24,25Thus, retaining POM intact at a higher pH could potentially strongly expand its suitability as a catalyst.
In the present study, we utilized the Keggin POM as a precursor in a sol−gel process.The sol−gel is a process where separation of a solid phase, usually metal oxide, from solution occurs via nucleation in the form of NP species resembling POMs.This creates a colloid solution (i.e., sol), allowing subsequent aggregation without growth, forming a colloid solid − gel. 30The Keggin POM species appear to act as such NPs and self-assemble into secondary particles, analogous to selfassembled particles formed by a sol−gel.The process can be described through a La Mer diagram, 31 illustrated by a graph in Figure 1A.The conditions required for their formation in this case appear to be highly acidic media, heat, and the presence of polyvalent cationic species in addition to the POM opening for neutralization of the POM charge and subsequent aggregation.Using this approach, we were even able to isolate and characterize a chemically individual intermediate in this selfassembly process�a lanthanum salt of POM.The produced hierarchical self-assembled POM structures were extremely poorly soluble under acidic and neutral conditions and demonstrated exceptional stable activity in electrochemical water splitting.

■ RESULTS AND DISCUSSION
In a recent study, we found that increasing acidity and ionic strength (cation concentration) in solutions of Keggin POMs together with peptides resulted in the formation of compounds with lower peptide-to-POM ratios, where metal cations and, most importantly, oxonium ions became incorporated into the resulting structures. 32Highly charged cations facilitated formation of "acidic" POM derivatives that are usually included in the composition of the product.In the present study, we attempted to use extremely acidic conditions with pH < 0 and relatively high concentrations of highly charged cations such as Ti(IV), Zr(IV), Ce(IV), and La(III) on heating with continuous stirring.As expected, this approach resulted in all cases in hierarchical self-assembly of POMs with formation of spherical aggregates several micrometers in size.The analysis of the particles showed a hierarchical structure with three levels of organization (Figure 1).First, the POM "nuclei" make contact at similar distances via hydrogen bonds.Second, nanospheres made from POMs of approximately 20 nm in diameter are formed.Third, assemblies of these nanospheres form ternary particles of up to 2 μm in diameter.
Scanning electron microscopy (SEM) images revealed spherical particles ranging from 500 nm to 2 μm in diameter (Figure 2A).A typical size distribution for Ti(IV)-derived material (i.e., nanospheres/particles) is shown in Figure 2C.Debris of broken spheres were present in the unwashed samples.EDS analysis showed mainly tungsten, with traces of the metal cations present during synthesis, such as titanium and potassium (Figure 2E and Table TS1).
X-ray powder diffraction (XRPD) of freshly obtained materials was consistent with the known pattern of hydrated Keggin-type phosphotungstic acid H 3 PW 12 O 40 •21H 2 O (Figure 2B). 33Drying of the samples resulted in broadening of the peaks and weakening of the peak intensity, with a fully X-ray amorphous product on longer storage.These transformations were likely caused by the loss of water molecules from the material that became amorphous while preserving the overall morphology at all levels.
The nitrogen adsorption/desorption isotherms for the PW NP sample are shown in Figure 2D.The shape of the isotherms corresponds to characteristic type I, typical of microporous solids having relatively small external surfaces, the limiting uptake being governed by the accessible micropore volume rather than by the internal surface area. 34The Brunauer−Emmett−Teller specific surface area (S BET ) was found as 100.6 m 2 •g −1 and the Langmuir surface area was found as −125.9m 2 •g −1 for the PW NP sample.The Barrett− Joyner−Halenda desorption cumulative surface area and volume of pores between 1.7 and 300 nm were 23.0 m 2 •g −1 and 0.019 cm 3 •g −1 (corresponding to 8.75 vol % of pores), respectively.
Atomic force microscopy (AFM) investigations showed that the spheres were made up of fairly uniform secondary particles (Figure 3), although larger than individual POMs, which are primary ones in this sol−gel process, analogues of what in metal−organic sol−gel are called micelles templated by the self-assembly of ligands (MTSALs). 35,36This implies, at the first step, aggregation of the Keggin POMs in the tens of nm size, which then form a tertiary aggregate in the μm range.The secondary particles, as shown from the X-ray diffraction (XRD) data, were originally formed as crystallites of the H 3 PW 12 O 40 •21H 2 O phase.Their growth, however, is appa- rently impeded by adsorption of highly charged cations on their surface, which drastically decreases solubility of the material and permits aggregation of these secondary particles.
The hierarchical composition of the spheres was also observed under transmission electron microscopy (TEM), where tertiary spheres of approximately 2 μm across (Figure 4A) are made up of secondary spheres in the tens of nanometers in diameter (Figure 4B,C) that are presumably composed of individual units of hydrated phosphotungstic acid.
Similar structures, made of potassium phosphotungstate, have been reported by Yan et al., 26,37 although these were produced by simple coprecipitation through dropwise addition of KCl solution to a POM solution.The latter were significantly smaller and singular in composition and structure rather than the hierarchical structure observed in this work.It can be speculated that they were formed by the action of analogous driving forces with their ground in charge compensation of the POM structural units.The cation content and the solubility, crucial for the application as stable electrocatalysts, are, however, drastically lower in the new material reported here.
The thermal stability of the produced material was investigated via both thermogravimetric analysis (TGA) (Figure S1) and by preparative experiments.The weight loss occurred in several steps in the interval 120−500 °C, being associated initially with the loss of the different forms of water content.Elemental analysis of the residues showed that the ratio of W to P did not vary significantly in the samples taken at different temperatures (σ = 0.78), and thus, no loss of the phosphorus content could be observed.
For dynamic light scattering (DLS) and zeta potential, three measurements were taken using distilled water as a medium, and the mean and standard deviation are shown in Supporting Information Table TS2.As the particles were fairly large, the standard deviation of the zeta potential was near 5 in all cases.
Single-Crystal X-ray Structure of the Isolated Intermediate.Crystals isolated from the synthesis using La(III) nitrate were triclinic centrosymmetric with a P1̅ space group, containing two phosphotungstate anions, one lanthanum ion, and 31 water molecules in an asymmetric unit, Z = 2 (Figure 5, Table 1).The composition of the material can thus be formulated as 19 .The large amount of water forms an extensive hydrogen bonding network throughout the crystal.The shortest contacts lie between water at 2.16 and 2.7 Å.At a bond length of 2.7−2.8Å, contacts exist between both water molecules and water to bridge POM oxygen.Between 2.8 and 2.9 Å, there are a number of bonds mainly between water and between water and terminal POM oxygen.In this range, there are also contacts between water and bridging POM oxygen, as well as POM−POM contacts.The longest contacts above 2.9 Å lie between water, POM oxygen and water, or two adjacent POMs.
The extensive contacts between POMs suggest that they are protonated at this pH (Figure S2), allowing for hydrogen  bonding between the POMs.Though the protons are not visible in the structure, the bond distances are consistent with H-bonds.Similar phenomena have been observed with phosphomolybdic acid previously. 29The water molecules participate mostly in four hydrogen bonds per molecule, which is typical for the structure of liquid water.We can observe two POMs per La ion, necessitating the need for other cations (e.g., protons) to contribute to charge neutralization.This structure can be seen as a "snapshot" of an intermediate in the process by which the spheres form, the next step being the accumulation of a larger number of POM units and transfer of the cations to the surface of the hydrated phosphotungstic acid crystallites.
Electrochemistry.Phosphotungstic acid has been used as an electrocatalyst at different pH values from 0 to 7 for the OER in water splitting.The catalyst is highly dependent on the pH of the electrolyte and LSV measurements give the lowest overpotential at acidic conditions, see Figure 6.At pH 0 (0.5 M H 2 SO 4 ), the overpotential was as low as 286 mV for the formation of O 2 (g) bubbles (OER).At pH 3 (citric acid/ sodium citrate buffer), the overpotential was 308 mV and at pH 7 (phosphate buffer), the catalyst showed an overpotential of 397 mV.The formation of oxygen bubbles becomes more and more pronounced with increasing potential.The determined faradaic efficiency was quite high, exceeding 90% and slightly increasing in time (see Supporting Information Table TS3−5).
From the cyclic voltammograms obtained at pH = 7, a redox peak was observed at 0.72 V, confirming the W 6+ /W 5+ redox state, while at pH = 0, two redox peaks were observed at 0.72 and 0.50 V, confirming both W 6+ /W 5+ and W 5+ /W 4+ redox states, respectively. 15The shift in the redox peak confirms the influence of phosphorus on the redox potential of tungsten, as phosphorus incorporation enhanced the acidity. 19he Tafel slopes calculated from the linear sweep voltammetry (LSV) measurements showed the efficiency of an electrode to produce current in response to a change in the applied potential.The Tafel slope increases with pH in the tested range: 83 mV/dec at pH 0, 86 mV/dec at pH 3, and 96 mV/dec at pH 7. Chronoamperometry (CA) was performed to evaluate the stability of the catalyst over time.The current density was observed during 24 h and found to be stable at an applied potential of 0.4 V (vs RHE) (Figures S3−S5 and Table S3).

■ CONCLUSIONS
Exploiting the factors leading to a decrease in the potential surface charge of phosphotungstic acid nanocrystals allowed the development of nanostructured microspheres of this material by simple sol−gel synthesis.Determination of the crystal structure of the potential intermediate product�an acidic salt of La(III) cations�provided additional clues to how the self-assembly process occurs.Sol−gel-produced waterinsoluble phosphotungstic acid was demonstrated as a  potential candidate for an electrocatalyst for the OER process at acidic conditions showing a very low overpotential in 0.5 M H 2 SO 4 (pH = 0) electrolyte with great stability.
Experimental Section.All chemicals were purchased from Sigma-Aldrich and used without further purification.The particles were produced by the general procedure of using 30 mL of a solution containing 1 mM of the cation Ti(IV) from potassium titanium oxide oxalate hydrate, or La(III) from lanthanum nitrate with pH near 0 (<0.1).To this was added 1 mM of phosphotungstic acid, upon which a precipitate formed.The precipitate was filtered either immediately or after the solution was evaporated to a near volume of 10 mL in a water bath held above 90 °C.The precipitate was washed with Milli-Q water and collected by centrifugation.Crystals of La and phosphotungstic acid were prepared by dissolving 0.4 g of lanthanum nitrate in 30 mL 2 M HCl and heating the solution in a > 95 °C water bath.To this was added 3 g of phosphotungstic acid and the solution was evaporated to approximately 12 mL, without stirring, at which point large cube-shaped crystals formed.Upon cooling, small X-ray quality crystals formed.The crystals were stable under the mother liquor, but upon drying, they degraded into a white powder.Safety concern: using relatively concentrated acidic solutions at near boiling water temperature is associated with a risk of stench of highly corrosive and irritating liquid�the use of gloves and safety goggles is compulsory on operation.
SEM and energy-dispersive X-ray spectroscopy (EDS) samples were immobilized on carbon tape and characterized using a Hitachi FlexSEM-1000 II.EDS spectra were analyzed using an Oxford Instruments EDS analysis system operated by the Aztec software.
For TEM observations, dispersions of the sol−gel were deposited on holey carbon grids (PELCO 50 mesh grids: pitch 508 μm; hole width 425 μm; bar width 83 μm; transmission 70%) and observed using a Philips CM/12 microscope (Thermo Fisher Inc.) fitted with a LaB 6 gun and operated at 100 kV.Negative TEM films were scanned using an Epson Perfection Pro 750 film scanner.
BET specific surface area and pore volume were determined from nitrogen adsorption/desorption isotherms at −196 °C (Micromeritics ASAP 2020 surface area and porosity analyzer).The samples were degassed at 120 °C for 12 h before measurements.
DLS and zeta potential were done by suspending spheres in distilled water and analyzing them on a Malvern Panalytical Zetasizer Nano analyzer, equipped with a red (362.8nm) laser.Data was processed using the Zetasizer Ver.7.12 software.
For AFM, samples were characterized using a Bruker Dimension FastScan atomic force microscope with a Nanoscope V controller in the ScanAsyst mode using a FastScan-B AFM probe (silicon tip, f 0 :400 kHz, k:4 N/m, tip radius: 5 nm nominally) and a scan rate of 1−3 Hz.Data was processed using Bruker NanoScope Analysis.
Preparative TGA was performed using a Nobertherm LE/6/ 11/P300 furnace and an FA2204B electronic balance.Approximately 160 mg of PW spheres was placed in two crucibles and heated to 120, 170, 220, 270, 320, 400, and 500 °C.At each point, the weight of one crucible was recorded, and from the other, a sample was taken for the EDS analysis.
Single-crystal XRD data was collected on a Bruker D8 QUEST ECO instrument and processed using the Apex4 software.A total of 2424 frames were collected.The total exposure time was 2.02 h.The frames were integrated with the Bruker SAINT software package using a narrow-frame algorithm.The integration of the data using a triclinic unit cell yielded a total of 57 188 reflections to a maximum θ angle of 28.00°(0.76Å resolution), of which 22 324 were independent (average redundancy 2.562, completeness = 98.7%,R int = 6.52%,R sig = 8.17%) and 17 582 (78.76%) were greater than 2σ(F 2 ).The final cell constants of a = 14.008(3)Å, b = 15.140(3)Å, c = 22.386(5) Å, α = 88.978(5)°,β = 89.228(4)°,γ = 80.885(5)°, and volume = 4686.7(18)Å 3 are based upon the refinement of the XYZcentroids of 9879 reflections above 20 σ(I) with 4.512 < 2θ < 71.58.Data were corrected for absorption effects using the multiscan method (SADABS).The ratio of minimum to maximum apparent transmission was 0.307.The calculated minimum and maximum transmission coefficients (based on the crystal size) were 0.0370 and 0.0870.The structure was solved and refined using the Bruker SHELXTL software package, using space group P1̅ , with Z = 2 for the formula unit, H 65 LaO 113.50 P 2 W 24 .The structure loses water extremely easily, which likely creates multiple defects reflected in low precision in determination of the electron density, in spite of using lowtemperature data collection.This generated a B-alert for large residual electron density.The B-alerts for isolated oxygen atoms are actually misleading because these atoms are actually water molecules invoked into a network of hydrogen bonding.The location of hydrogen atoms was impossible to discern because of the challenges in obtaining the correct electron density map.The final anisotropic full-matrix least-squares refinement on F 2 with 1268 variables converged at R1 = 5.96% for the observed data and wR2 = 16.30% for all data.The goodness-of-fit was 0.990.The largest peak in the final difference electron density synthesis was 9.579 e − /Å 3 and the largest hole was −6.835 e − /Å 3 with an RMS deviation of 0.749 e − /Å 3 .On the basis of the final model, the calculated density was 4.588 g/cm 3 and F(000), 5631 e − .The full list of bond distances and angles is available in Supporting Information Tables S7−S9.
All electrochemical experiments were performed at room temperature.The experiments were performed in a threeelectrode system using an SP-50 potentiostat (Biologic).The phosphotungstic material was tested as an electrocatalyst at different pH from pH 0−7 for the OER in water splitting.The catalyst was then mixed with carbon black to enhance the conductivity and subsequently deposited on the graphite felt (loading = 0.2 mg/cm 2 ).A three-electrode setup was used to perform the electrocatalysis experiments; a catalyst-loaded graphite working electrode, Pt-mesh as a counter electrode, and Ag/AgCl as a reference electrode.CV, LSV, and CA were performed.OER tests were carried out in a single-compartment electrolytic cell with different electrolytes of 0.5 M H 2 SO 4 (pH = 0), citric acid/sodium citrate buffer (pH = 3), and phosphate buffer (pH = 7).For cyclability, 200 cycles of CV were performed, and the working electrode saturated after 20−30 cycles of activation.The iR drop was directly compensated for by the potentiostat (with 82% compensation).The potentials recorded were finally calibrated in relation to the reversible hydrogen electrode (E RHE ) by using the equation: E RHE = E Ag/AgCl + 0.059 × pH.To minimize the capacitive current, the scan rate for the LSV curve was 10 mV/ s.The overpotential (η) of HER was calculated by using the equation: η = E RHE − 1.23, after reduction of the redox potential of oxygen, E O2/O2− = 1.23.The Tafel plots were obtained by transforming the LSV curve into log(j) vs E. All experiments were performed twice to check reproducibility.The faradaic efficiency was evaluated via control of the gas evolution.The instrumental setup and procedure details are reported in the Supporting Information (Figure S7 and Tables S3−S5).

Figure 1 .
Figure 1.Hierarchical assembly of phosphotungstate spheres by the sol−gel process, illustrated in relation to the La Mer concept graphically (A) and structurally (B).Individual POMs aggregate to form nanoparticles approximately 20 nm in diameter, which in turn assemble into ternary particles of up to approximately 2 μm in diameter.

Figure 2 .
Figure 2. (A) SEM image of the Ti-derived spheres at 5000× magnification.(B) XRPD pattern of the spheres.(C) Size distribution of the particles in A ranges from 0.8 to 2 μm, with an average of 1.4 μm.(D) Nitrogen adsorption/desorption isotherms for PW NPs.(E) EDS spectrum of the spheres.Tungsten and oxygen were the most abundant, while traces of potassium and titanium were also detected.

Figure 3 .
Figure 3. AFM images of the spheres.Particles in the micrometer range (A,B) can be seen at low magnification and their nanosized composition at high magnification (C,D).

Figure 4 .
Figure 4. TEM image of the spheres.The large tertiary particle (A) is made up of secondary particles of nanosize (B,C).

Figure 6 .
Figure 6.Electrochemical measurements at pH 0, 3, and 7, where phosphotungstic acid has been evaluated as an electrocatalyst for the OER in water splitting.(A) LSV.(B) CV curves obtained after initial stabilization.(C) pH dependence of the overpotential and Tafel slope.

Table 1 .
Details of Data Collection and Refinement for the Phosphotungstate−Lanthanum Structure a Details of data collection and refinement can be obtained free of charge from the Cambridge Crystal Structure Database at https:// www.ccdc.cam.ac.uk/structures/citing, deposition no.CSD-2307589. a