Impact of Nb(V) Substitution on the Structure and Optical and

3 days ago - Department of Chemistry, North Carolina State University , Raleigh , North ... Department of Chemistry, Wake Forest University , Winston-...
0 downloads 0 Views 9MB Size
Article pubs.acs.org/IC

Cite This: Inorg. Chem. XXXX, XXX, XXX−XXX

Impact of Nb(V) Substitution on the Structure and Optical and Photoelectrochemical Properties of the Cu5(Ta1−xNbx)11O30 Solid Solution Brandon Zoellner,† Shaun O’Donnell,† Zongkai Wu,‡ Dominique Itanze,§ Abigail Carbone,† Frank E. Osterloh,‡ Scott Geyer,§ and Paul A. Maggard*,† Inorg. Chem. Downloaded from pubs.acs.org by UNIV AUTONOMA DE COAHUILA on 05/09/19. For personal use only.



Department of Chemistry, North Carolina State University, Raleigh, North Carolina 27695-8204, United States Department of Chemistry, University of California, Davis, Davis, California 95616, United States § Department of Chemistry, Wake Forest University, Winston-Salem, North Carolina 27109, United States ‡

S Supporting Information *

ABSTRACT: A family of solid solutions, Cu5(Ta1−xNbx)11O30 (0 ≤ x ≤ 0.4), was investigated as p-type semiconductors for their band gaps and energies and for their activity for the reduction of water to molecular hydrogen. Compositions from 0 to 40 mol % niobium were prepared in high purity by solid-state methods, accompanied by only very small increases in the lattice parameters of ∼0.05% and with the niobium and tantalum cations disordered over the same atomic sites. However, an increasing niobium content causes a significant decrease in the bandgap size from ∼2.58 to ∼2.05 eV owing to the decreasing conduction band energies. Linear-sweep voltammetry showed an increase in cathodic photocurrents with niobium content and applied negative potential of up to −0.6 mA/cm2 (pH ∼7.3; AM 1.5 G light filter with an irradiation intensity of ∼100 mW/ cm2). The cathodic photocurrents could be partially stabilized by heating the polycrystalline films in air at 550 °C for 1 h to produce surface nanoislands of CuO or using protecting layers of aluminum-doped zinc oxide and titania. Aqueous suspensions of the Cu5(Ta1−xNbx)11O30 powders were also found to be active for hydrogen production under visible-light irradiation in a 20% aqueous methanol solution with the highest apparent quantum yields for the 10% and 20% Nb-substituted samples. Electronic structure calculations show that the increased photocurrents and hydroen evolution activities of the solid solutions arise near the percolation threshold of the niobate/tantalate framework wherein the Nb cations establish an extended −O−Nb− O−Nb−O− diffusion pathway for the minority carriers. The latter also reveals a novel pathway for enhancing charge separation as a function of the niobium−oxide connectivity. Thus, these results illustrate the advantages of using solid solutions to achieve the smaller bandgap sizes and band energies that are needed for solar-driven photocatalytic reactions.

I. INTRODUCTION

significantly reduces the band gap of these p-type metal oxides by ≲1.0−2.0 eV. As reported for NaCu(Ta1−xNbx)4O11, increasing the level of substitution of Nb(V) cations causes the emergence of a new lower-energy conduction band comprised of the Nb 4d orbitals in place of the higher-energy Ta 5d orbitals.3 In this solid solution, the bandgap size decreases from ∼2.7 to ∼1.8 eV upon moving to the compositions richest in Nb(V). However, relatively few studies have investigated the impact of Nb(V) substitution and atomic-site disorder on photoelectrochemical properties, such

Intense research efforts on metal−oxide semiconductors have focused on understanding how to optimize their bandgap sizes and band energies for increased visible-light absorption while conserving the ability to thermodynamically drive the watersplitting half-reactions, i.e., for hydrogen and oxygen production. One approach utilized by the Maggard research group has been to modify the energies of the valence and conduction bands through the mixing of two or more metal cations, such as for the solid solutions Li1−xCuxNb3O8,1 (Na1−xCux)2Ta4O11,2 and NaCu(Ta1−xNbx)4O11.3 Substitution of Cu(I) cations for the alkali metals lithium and sodium, or the substitution of Nb(V) cations for Ta(V) cations, © XXXX American Chemical Society

Received: January 31, 2019

A

DOI: 10.1021/acs.inorgchem.9b00304 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Table 1. Refined Compositions and Unit Cell Parameters from Powder Diffraction Data for Each of the Cu5(Ta1−xNbx)11O30 Samples x (loaded)

chemical formula (refined)

a = b (Å)

c (Å)

volume (Å3)

goodness of fit

0 0.1 0.2 0.3 0.4

Cu4.98Ta11O30 Cu4.94(Ta0.91Nb0.09)11O30 Cu5.04(Ta0.82Nb0.18)11O30 Cu4.91(Ta0.68Nb0.32)11O30 Cu4.98(Ta0.64Nb0.36)11O30

6.229 6.229 6.230 6.231 6.232

32.542 32.545 32.550 32.554 32.558

1093.6 1093.8 1094.1 1094.4 1095.0

1.41 1.28 1.21 1.38 2.21

and the crystal structure for Cu5Ta11O30 was used as the initial model for the refinement. First, the background was refined using the ChebyChev function with four parameters. Next, the unit cell was refined to correct the peak position and intensities. Peak profiles were adjusted by refining the crystallite size and microstrain. Once the peak shapes were a reasonable fit, the atomic positions were refined in multiple sets starting with tantalum/niobium, then copper, and finally oxygen. The atomic positions of symmetry-equivalent tantalum and niobium were constrained to each other. Additionally, the fractional occupancies of tantalum and niobium were refined with the constraint to keep the total site occupancy equal to 1. After all of the parameters were stabilized within the refinements, the UISO values were refined. The chemical compositions from the refined structures are listed in Table 1. Scanning Electron Microscopy. A JEOL SM 6010LA scanning electron microscope was used to collect scanning electron microscopy images of the powdered samples. A 10 or 20 kV accelerating voltage was used along with a secondary electron imaging detector. Ultraviolet−Visible (UV−vis) Spectroscopy. UV−vis diffuse reflectance spectra (DRS) were recorded using a Shimadzu UV3600 instrument with an integration sphere and a slit width of 32 nm. Approximately 20−30 mg of powder was spread evenly and pressed into a barium sulfate plate background. The collected reflectance data were transformed using the Kubelka−Munk function. Tauc plots of [F(R) × hν]n versus photon energy were used to determine the onset energies of the allowed direct (n = 2) and indirect (n = 1/2) transitions. The direct and indirect bandgap energies were determined by extrapolating the linear portion of the Tauc plots to the baseline fit. Photoelectrochemical Measurements. Polycrystalline films of Cu5(Ta1−xNbx)11O30 (0 ≤ x ≤ 0.4) were prepared by suspending ∼15 mg of each sample in a tert-butanol (30 mL)/water (5 mL) mixture through sonication and drop casting onto an ∼1 cm2 area of an FTO slide (TEC-7). The film was smoothed using the doctor blade method and annealed under vacuum for 3 h at 500 °C (∼25 mTorr). Photoelectrochemical measurements were carried out in a custom Teflon cell and a CH Instrument model 620A electrochemical analyzer. A high-pressure Xe lamp at 250 W with an AM 1.5G filter produced a 100 mW/cm2 photon flux at the electrode surface, as confirmed using a Si photodiode detector (Oriel company). A 0.5 M Na2SO4 solution with a pH adjusted to ∼6.5−7.5 was used to conduct both linear-sweep voltammetry and chronoamperometry experiments. Suspended Particle Photocatalysis. A 1% (by weight) amount of platinum metal was deposited on the surfaces of the metal oxides for each composition. Each material was stirred in a solution of chloroplatinic acid (1 mg/mL) and irradiated with both visible- and ultraviolet-light wavelengths for 5 h from a high-pressure Xe lamp. The collected powders were washed with deionized water and dried. A mass of ∼50 mg of each sample was loaded into a cylindrical quartz tube that was filled with ∼65 mL of a 20% methanol/deionized water solution. A high-pressure Xe arc lamp operating at 1000 W was set up with an IR-water filter and a 420 nm cutoff filter. The light was collimated, and the power density at the site of the reaction vessel was measured as 300 mW/cm2. Gas was collected and measured volumetrically in an L-shaped tube with a known diameter. The contents of the headspace of the reaction tube were collected with a syringe and injected directly into a gas chromatograph (SRI 8610C). Atomic Layer Deposition. Ultrathin layers of aluminum-doped zinc oxide (AZO) and TiO2 were deposited on top of polycrystalline films of Cu5(Ta1‑xNbx)11O30 using a thermal ALD system (Gemstar-6

as their photocatalytic activities as suspended powders or their p-type photocurrents as polycrystalline films under visible-light irradiation. Cationic substitution has been investigated within other chemical systems, but relatively few studies have focused on the investigation of photocathode materials.4,5 Copperbased photocathodes generally suffer from low photoelectrochemical performance and poor overall stability, while mixed-metal oxide solid solutions provide an opportunity to improve both of these disadvantages as described in a recent review.6 Polycrystalline films of the p-type Cu5Ta11O30 semiconductor were previously reported to have relatively large cathodic photocurrents (∼1−3 mA/cm2) under visible-light irradiation.7 While its conduction and valence band energies are suitable for driving the half-reactions of water oxidation and reduction, its bandgap size of ∼2.6 eV limits the percentage of visible light that can be absorbed from the solar spectrum.8,9 Decreasing the band gap by adjusting the conduction band position with the solid-solution Cu5(Ta1−xNbx)11O30 potentially offers an effective strategy for increasing overall light absorption. Herein, we report on the synthesis of a range of solid-solution compositions for Cu5(Ta1−xNbx)11O30 and the extent of the impact of the Nb(V) substitution on its structure, bandgap size, surface photovoltage, activity for hydrogen production and the stability and size of its cathodic photocurrents.

II. EXPERIMENTAL METHODS Materials. The reagents Cu2O (99.9%, Alfa Aesar), Nb2O5 (99.9985%, Alfa Aesar), and Ta2O5 (99.99%, Acros Organics) were purchased and used without further purification for the synthesis of Cu5(Ta1−xNbx)11O30. Ammonium hydroxide (concentrated, EMD) was used to wash the products. Deionized water and methanol (99.9%, Fisher Chemical) were used for the 20% methanol solutions. The electrolyte solution for the photoelectrochemical studies was made with Na2SO4 (99.0%, Alfa Aesar), H2SO4 (concentrated, Fisher Scientific), and deionized water. Synthesis. Solid-state reactions were performed using the appropriate stoichiometric mixtures of Nb2O5 and Ta2O5 along with a 10% molar excess of Cu2O to prepare Cu5Ta11O30 and solidsolution compositions of Cu5(Ta1−xNbx)11O30, for x = 0.1, 0.2, 0.3, 0.4, and 0.5. The limit of the solid-solution composition is reached at x = 0.5, as described below. The reactants were mixed, ground together within a mortar and pestle for 30 min, and sealed under vacuum in a fused-silica ampule (∼50 mTorr). The mixed reactants were heated at 900 °C for 48 h and subsequently washed in NH4OH to remove any remaining Cu2O impurities. Characterization. Powder X-ray Diffraction. The products were characterized using powder X-ray diffraction techniques with an INEL diffractometer using Cu Kα1 (λ = 1.54056 Å) radiation from a sealedtube X-ray source (35 kV, 30 mA). The diffracted X-rays were collected using a curved position-sensitive detector (CPS 120). Rietveld Refinements. Samples of Cu5(Ta1−xNbx)11O30 (x = 0, 0.1, 0.2, 0.3, and 0.4) were sent to the 11-BM Argonne National Laboratory so high-resolution powder X-ray diffraction data could be collected. The collected diffraction data were loaded into GSAS-II, B

DOI: 10.1021/acs.inorgchem.9b00304 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 1. (a) Unit cell of Cu5Ta11O30 viewed along the [010] direction. (b) Section showing the local coordination environments of the connecting layers of TaO7 and TaO6 with linearly coordinated copper. Bench-top ALD system). Deposition of AZO was carried out using the precursors diethylzinc (Sigma-Aldrich, ≥52 wt % Zn basis), trimethylaluminum (Sigma-Aldrich, 97%), and 18.2 Milli-Q H2O for the zinc, aluminum, and oxygen sources, respectively. The substrate temperature was held at 175 °C, while the precursors were kept at room temperature. Zinc and aluminum precursors were both held in the ALD chamber for 1.0 s after a 22 ms pulse, followed by a 28.0 s N2 purge (UHP 300). Next, the oxygen precursor was held in the ALD chamber for 1.0 s after a 22 ms pulse, followed by a 28.0 s N2 purge. Aluminum-doped ZnO was deposited by intercalating one cycle of trimethylaluminum and water after 20 cycles of diethylzinc and water. The growth rates were, as determined by X-ray reflectivity (XRR) on a reference Si wafer with a native oxide, 1.613 and 1.191 Å per cycle for ZnO and Al2O3, respectively. Deposition of TiO 2 was carried out using titanium(IV) isopropoxide (Sigma-Aldrich, 97%) and 18.2 Milli-Q H2O for the titanium and oxygen sources, respectively. The substrate temperature was held at 200 °C, and the titanium precursor was heated to 65 °C. Titanium(IV) isopropoxide was held in the ALD chamber for 4.0 s after a 100 ms pulse, followed by a 23.0 s N2 purge. Water was kept at room temperature before being held in the ALD chamber for 4.0 s after a 1.0 s pulse and then purged with N2 for 28.0 s. The growth rate, as determined by XRR, for TiO2 was 0.317 Å per cycle. Surface Photovoltage Spectroscopy. Surface photovoltage spectroscopy (SPS) measurements were conducted using a vibrating gold Kelvin probe (Delta PHI Besocke) mounted inside a home-built vacuum chamber (420 nm), hydrogen evolution was observed at an initial rate of ∼4−6 μmol h−1. After 4.5 h, hydrogen evolution slowed and became negligible. However, the powders could be reactivated with similar rates at least twice more by drying at ∼90 °C and using a fresh aqueous methanol solution, as shown in Figure 12a for Cu5Ta11O30. The average hydrogen production over the three trials for all of the compositions shows the total amounts of gas produced are relatively similar, as shown in Figure 12b. Under full solar irradiation, the 40% Nb sample showed the highest photocatalytic activities for hydrogen production, consistent with the photocurrent data

0.1, i.e., 10% Nb, solid-solution composition reaches an IPCE at 500 nm of ∼0.4% and at 580 nm of ∼0.16%. The energies of both of these photon wavelengths are lower than the bandgap size of the nonsubstituted Cu5Ta11O30 of ∼2.6 eV, which shows no photocurrents at these wavelengths. The 40% Nb samples show significantly smaller IPCEs of 0.06% and 0.02% at wavelengths of 500 and 580 nm, respectively. No photocurrents were observed below their bandgap sizes at 700 nm (1.77 eV) for either composition. The higher photocurrents of the 10% Nb sample for the visible-light energies likely arise from its higher conduction band edge as well as the greater dispersion in the extended niobate connectivity, i.e., just crossing the percolation threshold for the charged carriers, which functions to separate and funnel the excited electrons to the more extended parts of the framework at lower energies. Under full solar spectrum irradiation, however, the higher photocurrents are observed for the 40% Nb composition that more efficiently utilizes the full ultraviolet and visible-light wavelength range. These electrodes were not found to exhibit the linear behavior of 1/C2 (C is the capacitance) versus applied potential that is required for a Mott−Schottky-based analysis of their band energies. The formation of copper(II) oxides at the surfaces of Cu2O and other Cu(I)-based oxides has been shown to both enhance the photocurrents and act as a protecting layer between the main oxide and the electrolyte solution.29−35 Previously, nanoislands of CuO were shown to form at the surfaces of Cu5Ta11O30 when heated in air at temperatures of >350 °C. These induced a greater number of p-type defects and enhanced the photoelectrochemical performance and stability.7 To study this effect for Cu5(Ta1−xNbx)11O30, films of the material were heated in air at 550 °C for 1 h and the photocurrents was measured under the same conditions that were used for the original films. Powder X-ray diffraction of the samples shows no significant change in the overall crystallinity or phase composition, as shown in Figure S2. Scanning electron microscopy images show the formation of CuO nanoislands on the edges of the particles, as shown in Figure S4. A noticeable increase in the photocurrents for all of the compositions was observed over the potential range from 0.1 to −0.1 V versus SCE (Figure 9c). The increased photocurrents can be attributed to the type II band offset of the conduction and valence bands between the Cu5(Ta1−xNbx)11O30 phase and surface CuO islands.6,7 The excited electron from the bulk can be injected into the CuO islands and thereby inhibit charge recombination. This effect has been observed with other copper-based oxides, as well, including Cu2O and Cu3VO4.30,37 The CuO islands at the surfaces of the particles change the colors of the films to black and allow a greater portion of light to be absorbed.7 Diffuse reflectance spectroscopy of the powders after heating in air shows the absorption of photons from the CuO islands as well as the absorption from the bulk compounds (Figure S5). Furthermore, the extrusion of Cu(I) cations creates a greater number of p-type defects within the film as Cu(I) cations from the bulk material migrate out to the surface to react with the oxygen in the air.7,11,36,37 Repeated linear sweeps of the same films show nearly identical photocurrent production, which represents a more stable electrode surface with the electrolyte under irradiation, as shown in Figure S6. Chronoamperometry measurements at 0 V versus SCE for 1000 s were conducted to further test the stability of the material (Figure 9d). Films were held under dark conditions for the first 60 s to observe the I

DOI: 10.1021/acs.inorgchem.9b00304 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Next, a 1% (by mass) addition of platinum metal was deposited onto the surfaces of each sample via standard photodeposition techniques. Platinum ions from chloroplatinic acid are reduced to metal islands over the particles’ surfaces under light irradiation in an aqueous methanol solution. Energy dispersive spectroscopy confirmed the presence of platinum at the surfaces of the particles at the expected low percentage, as listed in Table S1. The platinized particles were suspended in a 20% methanol solution and irradiated under the same conditions as previously used. As plotted in Figure 12c, hydrogen was evolved at generally faster rates over the first 1.5 h compared to the rate of the nonplatinized particles. However, the rate of hydrogen production slowed after approximately 8−10 μmol had evolved (Table 4). Again, the photocatalytic activity for hydrogen evolution could be restored by replacing the methanol solution and by drying the powders in air. After a sample of 40% niobium with platinum was irradiated for three cycles, the collected powders were imaged via SEM and compared to the original products, as shown in Figure S9. A majority of the particles remain unchanged, while there are some additional irregular features that appear after repeated light exposure and hydrogen production. Each of the Cu5(Ta1−xNbx)11O30 samples was also heated in air at 550 °C for 1 h to investigate the impact of the CuO surface nanoislands on their photocatalytic activities. Photocatalytic activity was found for each sample upon its irradiation under the same conditions that were used in the previous tests. Overall, roughly twice as much hydrogen was produced over the first 1.5 h compared to the amounts in all of the other trials (Table 4). In addition, the averaged initial rates for all of the compositions doubled between the nonprocessed powder and the powders heated in air. These results are consistent with the observation of significantly higher cathodic photocurrents being produced after heating the polycrystalline films in air under the same conditions. Similarly, the rate of hydrogen evolution was found to diminish after irradiation for approximately 2 h. The decrease in hydrogen evolution over time is a common issue faced for many visible-light photocatalysts. In some cases, the photocatalytic materials may decompose, while in other cases, the reaction conditions may become unsuitable to allow for continued catalysis.41,42 As hydrogen is produced from the methanol solution, the aqueous methanol is typically sequentially oxidized to formaldehyde, formic acid, and carbon dioxide. A positive result from the Tollens test confirms the presence of formaldehyde and/or formic acid after the solution was irradiated with visible light. To test their impact on hydrogen evolution, formaldehyde, formic acid, and sodium formate were added to separate 20% methanol solutions in an excess of the calculated moles that could be produced during the average hydrogen evolution reactions. The addition of the oxidation byproducts of methanol did not prevent the production of hydrogen. Similar hydrogen rates were observed between the original and modified conditions. To further test the impact of the changes within the solution, a solution of methanol that was previously used was added to an untested sample of Cu5(Ta0.9Nb0.1)11O30 (heated in air). Hydrogen was produced with only a minor decrease in the initial rate and overall production as compared to when the same material was suspended within a freshly prepared methanol solution. Next, repeated hydrogen evolution reactions were conducted on the same powder while only the reaction solution was changed. The mixture was irradiated until no more gas was generated.

Figure 11. Linear-sweep voltammetry (top) and chronoamperometry (bottom) data of Cu5(Ta1−xNbx)11O30 (x = 0, 0.1, 0.2, 0.3, and 0.4) after coating with aluminum-doped ZnO (5 nm) and TiO2 (5 nm) under visible-light irradiation in an aqueous 0.5 M Na2SO4 solution with a pH of ∼6.7, with the light chopped off between 500 and 600 s.

for the same compositions. The measured apparent quantum yields (AQYs) of the x = 0.1 and 0.2 samples under only visible-light irradiation at 435 nm (and lower intensities) were ∼0.0034% and 0.0038%, respectively, as shown in Figure S8. However, the 10% Nb sample had more consistent hydrogen evolution over the 5 h period, while the efficiency for the 20% sample decreased over time. For other compositions, the amount of hydrogen was too small to be detected in the experimental determination of the AQY. This is consistent with the formation of an infinitely extended O−Nb−O−Nb− O− connectivity that forms at even these low concentrations according to percolation theory. In addition, at these lower Nb concentrations, there is a greater thermodynamic driving force for charge separation from regions of the structure that still have a less extended niobium−oxide connectivity to other regions of the structure with greater niobium−oxide connectivity that constitute the lower energies of the conduction band edge. The similarities of their photocatalytic activities and AQYs are also consistent with the photoelectrochemical measurements mentioned above for the nonair-annealed samples; i.e., the photocurrents with little or no applied potential were comparable for all samples. J

DOI: 10.1021/acs.inorgchem.9b00304 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 12. (a) Plot of hydrogen production vs time for pure Cu5Ta11O30 in 20% methanol over three repeated trials. Plots of hydrogen production vs time for the Cu5(Ta1−xNbx)11O30 (x = 0, 0.1, 0.2, 0.3, and 0.4) solid-solution compositions (b) averaged over three trials, (c) with 1% Pt photodeposited over their surfaces, and (d) after each sample had been heated in air at 550 °C for 1 h.

irradiation. No additional hydrogen was formed during the subsequent irradiation. However, particles that were dried at ∼90 °C were found to regain a performance similar to that seen in the previous experiments over multiple trials. Previous results from the work on CuNb1−xTaxO3 provided evidence through XPS and SEM that copper(I) could be reduced to copper metal at the surfaces under similar conditions.11 The stabilities of the bulk crystalline powders were confirmed using powder X-ray diffraction over multiple irradiation cycles (Figure S10). No impurity peaks or shifts in peak positions or intensities were detectable. Taken together, the data presented above suggest that the majority of the rapidly slowing hydrogen evolution (and its subsequent recovery) arises from the buildup of hydrogen in solution and the reduction of Cu(I) cations at the surface to copper metal. Thus, both the solution and surface properties of the photocatalyst play a role in slowing the hydrogen evolution over time. Samples that are heated at 550 °C have smaller amounts of Cu(I) cations at their surfaces owing to their oxidation to the CuO islands. This inhibits the corrosive reduction of the bulk phase and allows for higher and more stable photocurrents.

Table 4. Amounts of Hydrogen That Evolved over the Initial 1.5 h Irradiation under Visible Lighta Cu5(Ta1−xNbx)11O30 composition x x x x x

= = = = =

0 0.1 0.2 0.3 0.4

bare powder (μmol)

1% Pt (μmol)

heated at 550 °C (μmol)

6.0 7.2 5.4 6.2 7.7

7.7 7.0 8.8 5.1 9.9

13.0 17.7 15.6 12.9 14.0

a

Reaction conditions: 50 mg of sample, volume of 65 mL, irradiation intensity of ∼300 mW/cm2.

Each new solution allowed hydrogen to be generated, but the initial rates of production slowly decreased over the series of tests. Changes in the oxidation state of Cu(I) cations at the surfaces may also lead to the particles no longer being able to drive the reduction of protons to hydrogen, as it is well-known that Cu(I)-based oxides are susceptible to cathodic and anodic photocorrosion.43 The excited electrons can reduce the copper ions at the surface to a metal and create a passivation layer. For example, reoxidizing the copper metal was found to restore the original performance of the Cu(I)-based delafossite CuRhO2.44 To test this idea, pure oxygen was bubbled through the reaction mixture of the powder and 20% methanol after K

DOI: 10.1021/acs.inorgchem.9b00304 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

IV. CONCLUSIONS The Cu5(Ta1−xNbx)11O30 (0 ≤ x ≤ 0.4) solid-solution members were investigated as p-type semiconductors for their optical, photoelectrochemical, and photocatalytic activities for hydrogen evolution. Solid-state synthetic techniques were used to achieve the substitution of niobium for tantalum cations within the crystalline structure at up to at least x = 0.4. A large shift in the bandgap size (∼2.58 to 2.05 eV) corresponding to the increase in niobium content was measured, which was a result of bandgap transitions arising from the filled copper 3d orbitals and the relatively lowerenergy niobium 4d orbitals, comprising the valence and conduction bands, respectively. Electronic structure calculations show the extended −O−Nb−O−Nb−O− regions of the structure constitute the lowest energies of the conduction band edge, which arises at Nb concentrations as low as ∼11− 22% according to percolation theory results. This is consistent with the large change in the optical bandgap sizes at relatively low Nb concentrations of 10−20% but significantly smaller changes at 30% and 40%. Photoelectrochemical experiments conducted under visible-light irradiation produced cathodic currents, indicative of p-type semiconductors. Films of the nonprocessed materials showed clear composition-dependent differences in photocurrents as the applied bias became more negative. The maximum photocurrents were approximately −0.57 mA/cm2 for x = 0.4, i.e., Cu5(Ta0.6Nb0.4)11O30, with an applied voltage of −0.48 V versus SCE. The x = 0.1 composition exhibited IPCE values at 500 and 580 nm of 0.4% and 0.12%, respectively. Their stability was increased by heating the films in air at 550 °C for 1 h to form surface CuO islands and by adding protecting layers of AZO and TiO2. These modifications allowed for a stable photocurrent without any detectable decay over time. All solid-solution compositions showed activity for hydrogen evolution as suspended powders. The production of hydrogen increased after the powders had been heated in air at 550 °C and produced a maximum of ∼16−18 μmol over 2 h for the x = 0.1 and 0.2 compositions. Thus, these results demonstrate the effectiveness of utilizing solid solutions to achieve smaller bandgap sizes with band energies for driving stable and more efficient hydrogen production from sunlight.



Paul A. Maggard: 0000-0002-3909-1590 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS Use of the Advanced Photon Source at Argonne National Laboratory was supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under Contract DE-AC02-06CH11357. Support for surface photovoltage spectroscopy measurements was provided by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under Grant DOE-SC0015329.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.9b00304.



REFERENCES

(1) Sahoo, P. P.; Maggard, P. A. Crystal Chemistry, Band Engineering and Photocatalytic Activity of the LiNb3O8-CuNb3O8 Solid Solution. Inorg. Chem. 2013, 52 (8), 4443−4450. (2) Palasyuk, O.; Palasyuk, A.; Maggard, P. A. Site-Differentiated Solid Solution in (Na1‑xCux)2Ta4O11 and its Electronic Structure and Optical Properties. Inorg. Chem. 2010, 49 (22), 10571−10578. (3) Palasyuk, O.; Maggard, P. A. NaCu(Ta1‑yNby)4O11 Solid Solution: A Tunable Band Gap Spanning the Visible-Light Wavelengths. J. Solid State Chem. 2012, 191, 263−270. (4) Kato, T.; Hakari, Y.; Ikeda, S.; Jia, Q.; Iwase, A.; Kudo, A. Utilization of Metal Sulfide Material of (CuGa)1−xZn2xS2 Solid Solution with Visible Light Response in Photocatalytic and Photoelectrochemical Solar Water Splitting Systems. J. Phys. Chem. Lett. 2015, 6 (6), 1042−1047. (5) Hisatomi, T.; Okamura, S.; Liu, J.; Shinohara, Y.; Ueda, K.; Higashi, T.; Katayama, M.; Minegishi, T.; Domen, K. La5Ti2Cu1−xAgxS5O7Photocathodes Operating at Positive Potentials During Photoelectrochemical Hydrogen Evolution Under Irradiation of up to 710 nm. Energy Environ. Sci. 2015, 8 (11), 3354−3362. (6) Sullivan, I.; Zoellner, B.; Maggard, P. A. Copper(I)-Based pType Oxides for Photoelectrochemical and Photovoltaic Solar Energy Conversion. Chem. Mater. 2016, 28 (17), 5999−6016. (7) Sullivan, I.; Sahoo, P. P.; Fuoco, L.; Hewitt, A. S.; Stuart, S.; Dougherty, D.; Maggard, P. A. Cu-deficiency in the p-type Semiconductor Cu5‑xTa11O30: Impact on its Crystalline Structure, Surfaces, and Photoelectrochemical Properties. Chem. Mater. 2014, 26 (23), 6711−6721. (8) Fuoco, L.; Joshi, U. A.; Maggard, P. A. Preparation and Photoelectrochemical Properties of p-Type Cu 5Ta11O 30 and Cu3Ta7O19 Semiconducting Polycrystalline Films. J. Phys. Chem. C 2012, 116 (19), 10490−10497. (9) Palasyuk, O.; Palasyuk, A.; Maggard, P. A. Syntheses, Optical Properties, and Electronic Structures of Copper(I) Tantalates: Cu3Ta7O19 and Cu5Ta11O30. J. Solid State Chem. 2010, 183 (4), 814−822. (10) Jahnberg, L. Structure of the Copper(I) Tantalam Oxide, Cu5Ta11O30. J. Solid State Chem. 1982, 41 (3), 286−292. (11) Zoellner, B.; Stuart, S.; Chung, C.-C.; Dougherty, D. B.; Jones, J. L.; Maggard, P. A. CuNb1‑xTaxO3 (x < 0.25) Solid Solutions: Impact of Ta(V) Substitution and Cu(I) Deficiency on their Structure, Photocatalytic, and Photoelectrochemical Properties. J. Mater. Chem. A 2016, 4 (8), 3115−3126. (12) Choi, J.; King, N.; Maggard, P. A. Metastable Cu(I)-Niobate Semiconductor with a Low-Temperature Nanoparticle-Mediated Synthesis. ACS Nano 2013, 7 (2), 1699−1708. (13) Shannon, R. D. Revised Effective Ionic Radii and Systematic Studies of Interatomic Distances in Halides and Chalcogenides. Acta Crystallogr., Sect. A: Cryst. Phys., Diffr., Theor. Gen. Crystallogr. 1976, 32 (5), 751−767. (14) Boltersdorf, J.; King, N.; Maggard, P. A. Flux-mediated Crystal Growth of Metal Oxides: Synthetic Tunability of Particle Morphologies, Sizes, and Surface Features for Photocatalysis Research. CrystEngComm 2015, 17 (11), 2225−2241.

Powder XRD patterns for Cu5(Ta1−xNbx)11O30 (x = 0, 0.1, 0.2, 0.3, and 0.4) after Rietveld refinements, powder XRD, SEM images, UV−vis DRS and LSV after samples had been heated in air at 550 °C for 1 h, additional surface photovoltage spectra, apparent quantum yield measurements for the x = 0.1 composition, and X-ray reflectivity measurements of coated films (PDF)

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Frank E. Osterloh: 0000-0002-9288-3407 Scott Geyer: 0000-0003-4422-3313 L

DOI: 10.1021/acs.inorgchem.9b00304 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

(34) Han, J.; Zong, X.; Zhou, X.; Li, C. Cu2O/CuO Photocathode with Improved Stability for Photoelectrochemical Water Reduction. RSC Adv. 2015, 5 (14), 10790−10794. (35) Yang, Y.; Xu, D.; Wu, Q.; Diao, P. Cu2O/CuO Bilayered Composite as a High-Efficiency Photocathode for Photoelectrochemical Hydrogen Evolution Reaction. Sci. Rep. 2016, 6 (1), 351581. (36) King, N.; Sahoo, P. P.; Fuoco, L.; Stuart, S.; Dougherty, D.; Liu, Y.; Maggard, P. A. Cu-deficiency in the p-type semiconductor Cu1‑xNb3O8. Chem. Mater. 2014, 26 (12), 2095−2104. (37) Sahoo, P. P.; Zoellner, B.; Maggard, P. A. Optical, Electronic, and Photoelectrochemical Properties of the p-Type Cu3‑xVO4 Semiconductor. J. Mater. Chem. A 2015, 3 (8), 4501−4509. (38) Morales-Guio, C. G.; Tilley, S. D.; Vrubel, H.; Grätzel, M.; Hu, X. Hydrogen Evolution from a Copper(I) Oxide Photocathode Coated with an Amorphous Molybdenum Sulphide Catalyst. Nat. Commun. 2014, 5 (3059), 4059. (39) Miyata, T.; Minami, T.; Tanaka, H.; Sato, H. Effect of a Buffer Layer on the Photovoltaic Properties of AZO/Cu2O Solar Cells. Proc. SPIE 2005, 6037, 603712. (40) Kandiel, T. A.; Ivanova, I.; Bahnemann, D. W. Long-term Investigation of the Photocatalytic Hydrogen Production on Platinized TiO2: An Isotopic Study. Energy Environ. Sci. 2014, 7 (4), 1420−1425. (41) Wang, Z.-L.; Yan, J.-M.; Wang, H.-L.; Ping, Y.; Jiang, Q. Pd/C Synthesized with Citric Acid: An Efficient Catalyst for Hydrogen Generation from Formic Acid/Sodium Formate. Sci. Rep. 2012, 2 (1), 5981−5986. (42) Yang, X.; Pachfule, P.; Chen, Y.; Tsumori, N.; Xu, Q. Highly Efficient Hydrogen Generation from Formic Acid Using a Reduced Graphene Oxide-Supported AuPd Nanoparticle Catalyst. Chem. Commun. 2016, 52 (22), 4171−4174. (43) Chen, S.; Wang, L. W. Thermodynamic Oxidation and Reduction Potentials of Photocatalytic Semiconductors in Aqueous Solution. Chem. Mater. 2012, 24 (18), 3659−3666. (44) Gu, J.; Yan, Y.; Krizan, J. W.; Gibson, Q. D.; Detweiler, Z. M.; Cava, R. J.; Bocarsly, A. B. P-type CuRhO2 as a Self-Healing Photoelectrode for Water Reduction Under Visible Light. J. Am. Chem. Soc. 2014, 136 (3), 830−833.

(15) Arney, D.; Maggard, P. A. Effect of Platelet-Shaped Surfaces and Silver-Cation Exchange on the Photocatalytic Hydrogen Production of RbLaNb2O7. ACS Catal. 2012, 2 (8), 1711−1717. (16) McLamb, N.; Sahoo, P. P.; Fuoco, L.; Maggard, P. A. Flux Growth of Single-Crystal Na2Ta4O11 Particles and Their Photocatalytic Hydrogen Production. Cryst. Growth Des. 2013, 13 (6), 2322−2326. (17) Boltersdorf, J.; Wong, T.; Maggard, P. A. Synthesis and Optical Properties of Ag(I), Pb(II), and Bi(III) Tantalate-Based Photocatalysts. ACS Catal. 2013, 3 (12), 2943−2953. (18) Lin, H.; Maggard, P. A. Copper(I)-Rhenate Hybrids: Syntheses, Structures and Optical Properties. Inorg. Chem. 2007, 46 (4), 1283− 1290. (19) Yoshino, M.; Kakihana, M.; Cho, W. S.; Kato, H.; Kudo, A. Polymerizable Complex Synthesis of Pure Sr2NbxTa2‑xO7 Solid Solutions with High Photocatalytic Activities for Water Decomposition into H2 and O2. Chem. Mater. 2002, 14 (8), 3369−3376. (20) Kato, H.; Kudo, A. Energy Structure and Photocatalytic Activity for Water Splitting of Sr2(Ta1‑xNbx)2O7 Solid Solution. J. Photochem. Photobiol., A 2001, 145 (1−2), 129−133. (21) Kresse, G.; Furthmuller, J. Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave basis set. Comput. Mater. Sci. 1996, 6 (1), 15−50. (22) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 1996, 77, 3865−3868. (23) Broadbent, S. R.; Hammersley, J. M. Percolation processes I. Crystals and mazes. Math. Proc. Cambridge Philos. Soc. 1957, 53, 629− 641. (24) Shante, K. S.; Kirkpatrick, S. An introduction to percolation theory. Adv. Phys. 1971, 20 (85), 325−357. (25) Melo, M. A.; Wu, Z.; Nail, B. A.; De Denko, A. T.; Nogueira, A. F.; Osterloh, F. E. Surface Photovoltage Measurements on a Particle Tandem Photocatalyst for Overall Water Splitting. Nano Lett. 2018, 18, 805−810. (26) Rodriguez-Perez, M.; Canto-Aguilar, E. J.; Garcia-Rodriguez, R.; De Denko, A. T.; Oskam, G.; Osterloh, F. Surface Photovoltage Spectroscopy Resolves Interfacial Charge Separation Efficiencies in ZnO Dye-Sensitized Solar Cells. J. Phys. Chem. C 2018, 122 (5), 2582−2588. (27) Sharma, G.; Zhao, Z.; Sarker, P.; Nail, B. A.; Wang, J.; Huda, M.; Osterloh, F. Electronic Structure, Photovoltage, and Photocatalytic Hydrogen Evolution with p-CuBi2O4 Nanocrystals. J. Mater. Chem. A 2016, 4, 2936−2942. (28) Nail, B. A.; Fields, J. M.; Zhao, J.; Wang, J.; Greaney, M. J.; Brutchey, R. L.; Osterloh, F. E. Nickel Oxide Particles Catalyze Photochemical Hydrogen Evolution From Water-Nanoscaling Promotes p-Type Character and Minority Carrier Extraction. ACS Nano 2015, 9 (5), 5135−5142. (29) Won, D. H.; Choi, C. H.; Chung, J.; Woo, S. I. Photoelectrochemical Production of Formic Acid and Methanol from Carbon Dioxide on Metal-Decorated CuO/Cu2O-Layered Thin Films Under Visible-Light Irradiation. Appl. Catal., B 2014, 158−159, 217− 223. (30) Lim, Y.-F.; Chua, C. S.; Lee, C. J. J.; Chi, D. Sol-Gel Deposited Cu2O and CuO Thin Films for Photocatalytic Water Splitting. Phys. Chem. Chem. Phys. 2014, 16 (47), 25928−25934. (31) Wang, P.; Ng, Y. H.; Amal, R. Embedment of Anodized p-type Cu2O Thin Films with CuO Nanowires for Improvement in Photoelectrochemical Stability. Nanoscale 2013, 5 (7), 2952−2958. (32) Ohta, K.; Suzuki, T.; Ohya, S.; Kaneco, S.; Katsumata, H. Electrochemical Reduction of CO2 in Methanol with Aid of CuO and Cu2O. Catal. Today 2009, 148 (3−4), 329−334. (33) Dubale, A. A.; Pan, C.-J.; Tamirat, A. G.; Chen, H.-M.; Su, W.; Chen, C.-H.; Rick, J.; Ayele, D. W.; Aragaw, B. A.; Lee, J.-F.; Yang, Y.W.; Hwang, B.-J. Heterostructured Cu2O/CuO Decorated with Nickel as a Highly Efficient Photocathode for Photoelectrochemical Water Reduction. J. Mater. Chem. A 2015, 3 (23), 12482−12499. M

DOI: 10.1021/acs.inorgchem.9b00304 Inorg. Chem. XXXX, XXX, XXX−XXX