Synthesis of (Ga1–xZnx)(N1–xOx) with Enhanced Visible-Light

All Topics; Analytical; Applied; Biological; Materials Science & Engineering; Organic- .... provide useful structure–property studies of GZNO and po...
2 downloads 0 Views 4MB Size
Download PDF
Article pubs.acs.org/IC

Synthesis of (Ga1−xZnx)(N1−xOx) with Enhanced Visible-Light Absorption and Reduced Defects by Suppressing Zn Volatilization Dennis P. Chen and Sara E. Skrabalak* Department of Chemistry, Indiana University, 800 East Kirkwood Avenue, Bloomington, Indiana 47405, United States S Supporting Information *

ABSTRACT: (Ga1−xZnx)(N1−xOx) (GZNO) particles with enhanced optical absorption were synthesized by topotactic transformation of Zn2+/Ga3+ layered double hydroxides. This outcome was achieved by suppressing Zn volatilization during nitridation by maintaining a low partial pressure of O2 (pO2). Zn-rich (x > 1/3) variants of GZNO were achieved and compared to those prepared by conventional ammonoylsis conditions. The optical absorption and structural properties of these samples were compared to those prepared in the absence of O2 by diffuse-reflectance spectroscopy and powder X-ray diffraction methods. Notably, suppression of Zn volatilization leads to smaller-band-gap materials (2.30 eV for x = 0.42 versus 2.71 eV for x = 0.21) and reduced structural defects. This synthetic route and set of characterizations provide useful structure−property studies of GZNO and potentially other oxynitrides of interest as photocatalysts.

■

INTRODUCTION Solar water-splitting has the potential to provide a renewable and storable energy source for the ever-increasing global energy demand.1 At the heart of most artificial solar water-splitting processes are one or more semiconductors that absorb solar radiation and mediate photogenerated electrons and holes to their respective active sites to generate H2 and O2.2 The oxidative stability of most oxides in aqueous media make them attractive for use as photoanodes. The thermodynamics for the overall water-splitting reaction impose design constraints such that the positions of the conduction and valence bands (VB) must sit above and below the hydrogen reduction and water oxidation redox potentials, respectively.3 Although several oxide semiconductors such as RbPb2Nb3O104 and SrTiO35 have the ability to split water under UV-light irradiation, wide-band-gap materials leave much of the solar spectrum unharnessed. Lowlying O 2p states that comprise the VB of the majority of photocatalytically relevant oxides limit the design of oxides for the overall water-splitting reaction or to be coupled in a tandem device. In contrast, N is less electronegative than O, which will result in raised VB energies for nitride or oxynitride analogues.6 The shifts in the VB energies are associated with band-gap reductions7 and often do not compromise the oxidative stability8 of the semiconductor. Therefore, nitride and oxynitride semiconductors make for promising systems for watersplitting applications. Cocatalyst-loaded (Ga1−xZnx)(N1−xOx) (GZNO) displays the largest photonic efficiency9 among visible-light-driven water-splitting photocatalysts (5.9% for wavelengths between 420 and 440 nm) and exhibits longterm stability.6a,10 Despite its promising proof-of-concept performance, the photonic efficiency of GZNO is far from its theoretical limit. Unfortunately, conventional ammonolysis © XXXX American Chemical Society

routes yield inhomogeneous compositions and particle morphologies,11 which is prohibitive to structure−activity studies. In contrast, topotactic routes, as demonstrated here and by Wang et al.,12 have well-defined parent−daughter crystallographic relationships and can consistently yield Zn-rich (x > 1/3) phases for such studies. Conventionally, the synthesis of GZNO is accomplished by heating a mixture of Ga2O3 and ZnO under anhydrous NH3 at high temperatures (>800 °C).13 This nitridation approach proceeds through a spinel intermediate, which inherently limits the Zn/(Ga + Zn) ratio to 1:3.14 The highest Zn content reported for a GZNO phase prepared from binary oxides is x = 0.42, where the band gap (Eg) was estimated to be ∼2.4 eV.6a,15 In contrast, density functional theory (DFT) calculations suggest that the full GZNO solid solution range is compositionally accessible.16 Moreover, the band-gap minimum (Eg = 2.3 eV) is expected when x ∼ 0.5.16,17 Significant effort has been placed toward preparing GZNO phases with higher Zn content to enhance the light-harvesting ability of GZNO and to study the properties of its Zn-rich phases.12,18 High-pressure, high-temperature synthesis produces GZNO with x = 0.50 and 0.75, but the restricted volume in high-pressure syntheses severely limits the sample sizes, which makes its characterization challenging. While Zn fractions as high as 0.9019 and band gaps as low as 2.21 eV18c have been reported in nanoscale GZNO using low-temperature nitridations routes, high-temperature routes that may lead to Zn-rich bulk phases with different order−disorder16 states have not been widely investigated. Herein, the pO2 value present in the gas-phase composition of Received: December 10, 2015

A

DOI: 10.1021/acs.inorgchem.5b02866 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 1. SEM images of ZnGa-LDH (a), GZNO-NH3 (b), and GZNO-O2/NH3 (c). Segmented yellow and red lines outline preservation of the platelike particle morphology for GZNO-NH3 and GZNO-O2/NH3, respectively. The insets display Zn fractions as obtained from ICP-OES analysis. The yellow arrows indicate the presence of pores in GZNO-NH3 plates. The solid orange box depicts a high-magnification image of the area enclosed in the fragmented orange box, while the red arrows indicate the growth of pyramidal features in GZNO-O2/NH3 particles. Shown in panel d is a schematic of the crystallographic relationship of the LDH and wurtzite structures viewed down [0001]. TEM images of GZNO-NH3 (e) and GZNO-O2/NH3 (f) with the corresponding SAED patterns displayed as insets. was then adjusted to a pH of 8 with an aqueous solution containing NaOH (3 M) and Na2CO3 (1 M). The precipitates were aged in the mother liquor at 80 °C for 24 h and then filtered and washed with copious amounts of DI water. The sample was dried overnight under vacuum. LDH-derived (Ga1−xZnx)(N1−xOx) were synthesized from ZnGaLDH by heating 400 mg of gently ground ZnGa-LDH at either 800 °C for 10 h under a NH3 flow (860 mL/min) [referred to herein as GZNO-NH3] or a mixture of O2/NH3 (2 vol % O2, total 860 mL/ min) [referred to herein as GZNO-O2/NH3]. Caution! O2 and NH3 form a f lammable and explosive mixture above a critical O2/NH3 (∼4.5 vol % at 800 °C).22 Prior to nitridation, the tube furnace containing the starting material(s) was purged with Ar for 1 h while the sample was isothermally heated at 90 °C and subsequently purged with NH3 (200−250 mL/min) for 30 min to minimize the presence of water. Heating ramp rates were kept constant for all nitridation reactions at 15 °C/min while under the desired atmosphere at a flow rate of 860 mL/min. After heating, the sample was naturally cooled to room temperature under NH3 or O2/NH3 flow. All reactions were performed on precursor material that was loaded in a dense alumina combustion boat (AdValueTech). In a typical reaction obtained from flowing O2/NH3, the crude product consisted of the desired GZNO phase along with black ZnO/N microcrystals. The crude product was worked up by dissolving the ZnO microcrystals by incubating the powder in 4 M HNO3 for 30 min, which was followed by several centrifugation and DI rinsing steps (until the pH of the supernatant was ∼6). Characterization. Powder XRD. Laboratory powder XRD measurements were performed on a Panalytical Empyrean instrument equipped with Cu Kα radiation and an X’Celerator linear strip detector. Powder samples were loaded on a zero-diffraction Si holder.

the nitridation mixture is shown to reduce volatilization of Zn and leads to Zn-rich GZNO. This example illustrates the importance of gaseous reactants in addition to the use of high Zn/Ga and kinetically less-hindered precursors18a,c,20 to achieve high-Zn-containing GZNO. The optical absorption (Kubelka−Munk, KM) profiles of Zn-rich and Zn-deficient GZNO were analyzed to reveal Zn volatilization as a source of defects in GZNO.14,21 Spectroscopic methods were supported with structure models obtained from synchrotron X-ray diffraction (XRD). Taken together, this study outlines a new route to Zn-rich GZNO with enhanced visible-light absorption and provides insight into the origins of defects that may be detrimental to photocatalytic processes.

■

EXPERIMENTAL SECTION

Reagents and Materials. All chemicals were handled at atmospheric pressure. All chemicals were reagent-grade or higher and were used as received. Zn(NO3)2·6H2O (99%), Ga(NO3)3·yH2O (99.9%), and ZnO (Puratronic, 99.999%) were obtained from Alfa Aesar. A nominal value of y = 8.6 hydrates was obtained from thermal gravimetric analysis of Ga(NO3)3·yH2O salts. HNO3 (69−70%) was purchased from Mallinckrodt. NaOH pellets and anhydrous Na2CO3 were purchased from VWR and Sigma-Aldrich, respectively. Milli-Q (18.2 MΩ·cm−1) deionized (DI) water was used in all syntheses and workup procedures. Synthesis. Solution Synthesis of ZnGa-LDH (where LDH is layered double hydroxide) was adapted from Wang et al.12 A solution containing 0.15 M (metals basis) zinc and gallium nitrate salts was prepared by dissolving stoichiometric amounts of Zn(NO3)2·6H2O (10 mmol) and Ga(NO3)3·8.6H2O (5 mmol). The pH of the solution B

DOI: 10.1021/acs.inorgchem.5b02866 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry High-resolution synchrotron XRD data were collected at beamline 11BM (λ = 0.413832 Å) of the Advanced Photon Source (APS) at Argonne National Laboratory in 0.8-mm-diameter Kapton capillaries. Whole powder pattern refinements were carried out using TOPAS v4.1 (Bruker AXS).23 GZNO-O2/NH3 structure models were refined by the Rietveld method using a P63mc structure model. Occupancy parameters were fixed during the refinement. Nominal cation concentrations were obtained from inductively coupled plasma optical emission spectroscopy (ICP-OES) analysis, while the N(O)/Ga(Zn) ratios were set to unity. Isotropic displacement parameters, ⟨u2⟩iso, were refined for cation and anion sites. Microstructural parameters were refined using the double-Voigt approach,24 which is available in TOPAS “GUI mode”. Because of significant parameter correlation, Rietveld refinements of GZNO-NH3 structure models resulted in nonphysical structure models. For the purpose of a qualitative comparison, GZNO-NH3 structure models were refined using the Pawley approach. The refinement incorporated the Stephens anisotropic model,25 which was implemented in TOPAS “launch mode”. For a comparison of the line-width broadening, peaks obtained from a National Institute of Standards and Technology (NIST) certified Standard Reference Material 674b ZnO,26 GZNO-O2/NH3 and GZNO-NH3 powder patterns were fit analytically with a pseudoVoigt function. Electron Microscopy. Scanning electron microscopy (SEM) images were obtained with an FEI Quanta 600 FEG field-emission scanning electron microscope operating at 30 kV. It was interfaced with an Oxford Inca detector for energy-dispersive X-ray spectroscopy (EDS). Transmission electron microscopy (TEM) images and selected-area electron diffraction (SAED) were obtained on a JEOL JEM-3200FS transmission electron microscope operating at 300 kV. Spectroscopy. Fourier transform infrared (FTIR) measurements were performed on KBr pellets, and IR spectra were recorded on PerkinElmer Spectrum Two FTIR software. Diffuse-reflectance (DR) spectroscopy was performed on a Cary 100 UV−visible spectrometer equipped with a Cary 301 DR accessory. Typically, 20 mg of the powder sample is diluted (2 wt %) with ball-milled BaSO4 (Alfa Aesar [99.998%]). Ball-milled BaSO4 served as the 100% reflectance standard. The powder mixtures were loaded and pressed onto a cylindrical powder holder with the dimensions of 3 mm height x 17 mm diameter. X-ray photoelectron spectroscopy (XPS) measurements were performed using a PHI Versa Probe II scanning X-ray microprobe under ultrahigh-vacuum conditions and with a monochromatic Al Kα X-ray source. The C 1s peak obtained from adventitious C was calibrated at 284.8 eV and used as an internal standard. ICP-OES of ZnGa-LDH and selected GZNO samples was performed on a PerkinElmer 2000DV spectrometer.

of the nitridation process as crystallographic relationships become identifiable. Shown in Figure 1a are the ZnGa-LDH plates used as a precursor to GZNO. The faceting is consistent with a rhombohedral or hexagonal unit cell. The formula unit Zn0.67Ga0.33(OH)2(CO3)0.17·zH2O corresponding to a rhombohedral LDH phase was confirmed from ICP-OES analysis, FTIR (Figure S1 and Table S1), and Pawley refinement of an R3̅m unit cell fitted to synchrotron XRD data (Figure S2). The chief purpose of the refinement was to verify a single-phase starting material; therefore, a more rigorous structural refinement (Rietveld or whole powder pattern modeling) was not attempted. The morphological contrast between GZNO synthesized under pure NH3 versus O2/NH3 is exemplified in the SEM images of GZNO-NH3 (x = 0. 22) and GZNO-O2/NH3 (x = 0.42), as shown in Figure 1b,c, respectively. In both cases, the platelike features of the ZnGa-LDH microcrystals are retained. However, for GZNO-NH3, a large amount of Zn is lost during nitridation (from x0 = 0.67 to x = 0.22), which likely results in the predominating porous features in GZNO-NH3 particles. On the other hand, particles synthesized under the mixed O2/ NH3 atmosphere retained the majority of its Zn content (from x0 = 0.67 to x = 0.42). Volatilization of Zn requires Zn2+ to be reduced to its metallic state, so increasing pO2 in the system would thermodynamically prohibit the reduction of Zn2+. Morphologically, GZNO-O2/NH3 particles display pyramidal overgrowths along the basal plane of the platelets rather than extensive porosity. The overgrowth is suggestive of some form of restructuring because ammonolysis is a substitutive process and should not increase the overall mass of the particle. From a geometrical point of view, the (001) plane of ZnGa-LDH (R3̅m) lies parallel to the (001) plane of the hexagonal wurtzite phase of ZnO (P63mc) [Figure 1d]. Coincidentally, the pyramidal texture has been observed in solution-processed ZnO nano/microcrystals in which growth occurs most rapidly along the [001] plane.29 To investigate the crystallographic relationship between the ZnGa-LDH starting material and GZNO oxynitride, SAED patterns were obtained for GZNONH3 and GZNO-O2/NH3 (Figure 1e,f). SAED patterns taken along the ⟨001⟩ axis exhibit the 6-fold-symmetric, hexagonal wurtzite diffraction pattern. Topotactic transitions would result in the homogeneous distribution of Zn and Ga cations. This assertion is confirmed by STEM/EDS analysis because the Zn and Ga cations are shown to be evenly distributed at the nanoscale, as shown in the STEM/EDS-mapping images in Figure S3. The geometrical aspects are insufficient to deduce the transformation mechanism, and further investigation is necessary to describe the presence or absence of an intermediate phase(s) during nitridation. The crystalline phases present in the nitridation product were probed ex situ by powder XRD for ZnGa-LDH reacted under NH3 (Figures 2a and S4a) and under O2/NH3 (Figures 2b and S4b). Reactions performed under either atmospheric condition resulted in a single-phase product that could be indexed to a hexagonal wurtzite phase. The absence of a Ga2O3 phase in GZNO-O2/ NH3 may be due to slower oxidation kinetics in comparison to the formation of GaN.30 Moreover, no other crystalline phases were observed from products quenched at various time points of nitridation, other than the initial ZnGa-LDH and final GZNO phases. Unlike solid-state reactions between binary zinc and gallium oxides,14 nitridation of ZnGa-LDH does not form a

■

RESULTS AND DISCUSSION Suppression of Zn Volatilization. LDHs are layered compounds composed of alternating positively charged brucitelike layers and negatively charged anions (typically CO32−) and have the general formula [Zn1−yGay(OH)2]y+(An−)y/n·zH2O. The typical LDH phase contains a 2:1 divalent-to-trivalent cation ratio, making ZnGa-LDH phases an attractive Zn-rich precursor to GZNO. Because the maximum Zn/Ga in GZNO is restricted to the Zn/Ga of the crystal structure undergoing ammonolysis, verifying the topotactic relationship of the initial and final crystal structures is important so as to confirm the absence of intermediate phases containing low Zn/Ga. Although LDHs have been extensively studied as single source precursors toward spinel compounds,27 few studies report the topotactic conversion of a LDH to the hexagonal wurtzite phase.12 The compositionally stable28 2:1 ratio of Zn2+/Ga3+ LDH was selected as the precursor to Zn-deficient and Zn-rich GZNO, denoted as GZNO-NH 3 and GZNO-O 2/NH3 , respectively, on account of their differing ammonolysis conditions. Single-phase precursors simplify the interpretation C

DOI: 10.1021/acs.inorgchem.5b02866 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Spectroscopic Investigation. The surface valence states of constituents in the GZNO solid solution were investigated by XPS. Figure S6 shows the XPS survey spectra and highresolution scans for Zn 2p3/2, Ga 2p3/2, O 1s, and N 1s and VB regions of GZNO-NH3 and GZNO-O2/NH3. All of the photoemission lines display a shift toward lower binding energies with increasing Zn content, consistent with previous reports.12,31a The difference in the Zn content is evident from the greater Zn/Ga 3d peak intensities displayed by GZNO-O2/ NH3. In addition, the onset of the VB edge is shifted in the direction of lower binding energy in GZNO-O2/NH3 relative to GZNO-NH3. Chemical analysis by XPS (Table S2) reveals that the surface of GZNO-NH3 is rich in Ga, which is likely due to the presence of a GaOx layer.32 However, a difference in Zn/ Ga between the surface and bulk for GZNO-O2/NH3 is not apparent. These differences in the surface states may play an important role in charge-transfer processes. The optical responses of GZNO-NH3 and GZNO-O2/NH3 were determined by DR studies (Figure 3). The relative

Figure 2. Reaction time studies monitored by ex situ XRD. XRD patterns of the set of reactions for GZNO-NH3 (a) and GZNO-O2/ NH3 (b). The black patterns corresponds to the precursor ZnGaLDH, while the red and blue patterns belong to particles that were quenched at 700 °C (red) and 800 °C (purple) after 1 h, respectively. The blue patterns correspond to GZNO-NH3 and GZNO-O2/NH3 particles that were nitridated for 10 h. Reference patterns were obtained from ICDD PDF 01-076-3644, 01-071-0843, and 01-0702546, which correspond to rhombohedral ZnGa-LDH (black), spinel ZnGa2O4 (blue), and hexagonal wurtzite GaN (red), respectively.

Figure 3. DR spectra of GZNO-NH3 (black) and GZNO-O2/NH3 (red) with KM functions plotted linearly (a) and in the form a Tauc plot (b). Shown in panels c and d are the KM transforms of the GZNO-NH3 and GZNO-O2/NH3 DR spectra plotted semilogarithmically to highlight the Urbach region, respectively. In panels c and d, the red segment is the linear region from which the slope (shape parameter, EU−1) was obtained.

ZnGa2O4 spinel intermediate. A qualitative comparison of the pattern evolution for the GZNO-NH3 and GZNO-O2/NH3 Bragg reflections reveals that the (002) reflection sharpens at a slower rate for GZNO prepared under the mixed O2/NH3 atmosphere than under pure NH3. The difference in kinetics may be tied to the formation of the pyramidal features during the conversion to GZNO-O2/NH3. Interestingly, XRD and EDS analyses of GZNO-O2/NH3 prepared at various reaction times indicate that extending the nitridation times did not decrease the nominal Zn content (Figure S5). Moreover, no noticeable change was observed in the line profiles of the Bragg reflections beyond 8 h of nitridation. In the ammonolysis of ZnGa-LDH under O2/NH3, Zn loss predominates at an early stage of the reaction and remains relatively constant upon formation of the wurtzite phase. The suppression of Zn volatilization over long reaction times (>10 h) and at temperatures ≥800 °C while preserving the overall crystallinity of the material is unprecedented.14,18a,c,31 Taken together, geometrical and chemical analyses support a topotactic transformation, which favors both a more homogeneous distribution of cations and synthetic conditions in which the upper Zn/Ga limit can be clearly defined.

absorbance (αKM) obtained from a KM transform of the reflectance data (Figure 3a) shows that the absorption edge of GZNO-O2/NH3 is shifted further into the visible regime than GZNO-NH3. The red shift is consistent with the higher Zn content in GZNO-O2/NH3 than GZNO-NH3.6a,18 A similar behavior was observed for the direct band gaps (Eg,direct) calculated from the Tauc plot (Figure 3b), where Eg,direct values of 2.71 and 2.30 eV were obtained for GZNO-NH3 and GZNO-O2/NH3, respectively. The dramatic decrease in Eg,direct compared to the reported GZNO phases with x = 0.426a may stem from the increased chemical homogeneity in GZNO-O2/ NH3. Moreover, the Eg,direct value obtained from this study is in agreement with the Eg(x) behavior predicted by DFT calculations.17 Tailing in the preedge region is evident in the Tauc plots of both GZNO samples and consistent with Urbach absorption. D

DOI: 10.1021/acs.inorgchem.5b02866 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 4. Rietveld refinement of GZNO-O2/NH3 (a) and Pawley refinement of GZNO-NH3 (b) from synchrotron XRD data. Refinement was carried out in the Q ranges of 1.3−12.1 and 1.3−11.6 Å−1 for GZNO-O2/NH3 and GZNO-NH3, respectively, where the scattering vector Q is defined as Q = (4π sin θ)/λ (for elastic scattering). Markers represent experimental data, while the solid red line indicates the calculated pattern. The difference pattern is shown in gray below the experimental and calculated patterns. The high-Q region of the powder pattern is shown in the inset.

scattering cross sections of Zn(Ga) and O(N), occupancies were fixed and isotropic displacement parameters were refined. The calculated pattern based on the ZnO-type structure model and the experimental pattern are shown in Figure 4a. The refined structural and geometrical parameters from the fitting are tabulated in Table 1. Unlike GZNO-O2/NH3, the Zndeficient GZNO-NH3 exhibits anisotropic broadening that is visually evident in the diffraction pattern. A Rietveld refinement was not possible because the displacement parameters were too strongly correlated with anisotropic broadening and microstructural parameters. Therefore, a structure model restricted to P63mc symmetry was refined by the Pawley method. Zn-rich GZNO-O2/NH3 exhibits lattice parameters [a = 3.21964(2) Å and c = 5.20501(3) Å] larger than those of the Zn-deficient GZNO-NH3 [a = 3.20657(2) Å and c = 5.20227(2) Å]. Interestingly, the reported lattice parameters of nanocrystalline GZNO for the same compositions are larger than the parameters obtained from GZNO-NH3 and GZNOO2/NH3.18a,35 In the GZNO-O2/NH3 structure model, both cation and anion isotropic displacement parameters were smaller than the values reported by Yashima et al.36 for GZNO synthesized by the conventional solid-state method. This comparison is consistent with the notion that structural disorder is proportional to the extent of Zn volatilization relative to Zn/Ga in the initial crystal structure. In the solidstate synthesis of GZNO, the reacting crystal structure is a spinel phase (x0 = 1/3), which leads to a 64% loss in the Zn content for typical GZNO compositions, where x = 0.12. In the case of Zn-deficient GZNO-NH3, the volatilization of Zn is even more severe because the initial structure contains x = 0.67. Indeed, the disorder is reflected in the broadening in the diffraction pattern of GZNO-NH3. The increased apparent broadening in the diffraction pattern relative to GZNO-O2/ NH3 persists in GZNO-O2/NH3 samples that were subjected

Urbach tailing is the broadening of the onset of absorption, which reflects structural disorder that may be of chemical or thermal origin.33 The Urbach energy EU, which represents the phonon capture efficiency or the breadth of the optical transition, is given by34 α = α0 exp[(hν − E0)/E U(T )]

(1)

where α is the wavelength-dependent absorption coefficient and α0 and E0 are constants. Scaling the KM-transformed DR spectra logarithmically (Figure 3c,d) reveals the linear Urbach region, where EU was obtained for the GZNO samples. A comparison of EU values suggests that GZNO-NH3 (0.119 eV) exhibits a higher degree of disorder than GZNO-O2/NH3 (0.108 eV). The broader Bragg reflections in the diffractogram of GZNO-NH3 versus GZNO-O2/NH3 (Figure 2) are also indicative of less coherency in the long-range order of the sample. The increase in the structural disorder or loss of longrange order is likely a result of the volatilization of Zn during nitridation because similar observations were made from extended X-ray absorption fine structure (EXAFS) results by Rodriguez and co-workers.14 They reported higher Debye− Waller factors in GZNO samples that underwent longer nitridation reactions. By suppression of the Zn volatilization, GZNO particles with a Eg,direct value of 2.30 eV and less structural disorder were obtained. Diffraction Analysis of Structural Disorder. The structural refinement of solid solutions from powder data is often hampered by overlapping reflections at high scattering vectors, Q (= [4π sin θ]/λ). To minimize instrumental contributions to profile broadening, structural refinements of GZNO-O2/NH3 (Figure 4a) and GZNO-NH3 (Figure 4b) were carried out on high-quality powder XRD patterns obtained from synchrotron XRD. A structure model based on the ZnO (wurtzite) unit cell was refined by the Rietveld method for GZNO-O2/NH3. Because of the similar X-ray E

DOI: 10.1021/acs.inorgchem.5b02866 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

reflections from the GZNO-NH3 diffraction pattern indicates that the (0kl) and (h00) families of reflections display broader peak widths over the other hkl families. To confirm the relevant anisotropic strain parameters, Pawley fits with each in turn of the three Shkl parameters were fixed at zero. In addition, the Lorentzian mixing parameter, η, was refined through the same process. Naturally, Pawley refinements of the GZNO-NH3 structural model with S400 and S202 fixed at zero resulted in high residuals (Table S1). In the Stephens model, each crystallite in the sample is regarded as having its own lattice parameters, and the width of each reflection is expressed in terms of moments of this distribution. The distribution of lattice metric parameters is considered to be a result of anisotropic strain; therefore, the relatively large Shkl parameters suggest that the GZNO-NH3 samples have significant strain contributions along the (0kl) and (h00) planes.

Table 1. (a) Crystallographic and Microstructural Parameters for GZNO-O2/NH3, (b) Atomic Coordinates, Fractional Occupancies, and Isotropic Displacement Parameters, and (c) Selected Bond Lengths and Bond Angles (a) Crystallographic and Microstructural Parameters for GZNO-O2/NH3 formula unit = (Ga0.58Zn0.42)(N0.58O0.42) Z=2 Lattice Parameters space group P63mc a (Å) 3.21963(2) c (Å) 5.20501(3) c/a 1.6167 cell volume (Å3) 46.7278(5) Microstructure LVOL-IB, k = 1 65.2(8) ε0 0.1173(7) (b) Atomic Coordinates, Fractional Occupancies, and Isotropic Displacement Parameters site cation cation anion anion

atom

x

y

z

occupancy

1 2 /3 /3 1 0.42 Zn2+ 1 2 /3 /3 1 0.58 Ga3+ 1 2 /3 /3 0.3748 0.42 N 1 2 /3 /3 0.3748 0.58 O (c) Selected Bond Lengths and Bond Angles

bond

bond length (Å)

cat(1)−an(1) cat(1)−an(2) cat(1)−cat(1)

1.9508 1.9698 3.1981

■

CONCLUSIONS By the introduction of O2 to the nitridation reaction, Zn-rich GZNO-O2/NH3 was obtained from ZnGa-LDH. Nitridation of an identical precursor under anhydrous NH3 alone resulted in Zn-deficient GZNO-NH3, highlighting the importance of controlling the pO2 value of the nitridation reaction. While the effect of controlling the pO2/pNH3 of the nitridation mixture was only demonstrated for GZNO, we envision that this strategy can be generalized to other oxynitride systems, in particular to compositions sensitive to reduction, to achieve higher efficiency photocatalysts.37 Here, this change in the nitridation conditions enabled a high Zn content of GZNO to be achieved with increased visible-light absorption (Eg,direct to 2.30 eV at x = 0.42). Suppressing Zn volatilization also decreased the structural disorder of GZNO. GZNO-O2/NH3 exhibited lower Urbach tailing near the absorption edge and less peak broadening in the diffraction patterns. A topotactic route has been demonstrated as a method to suppress the structural defects of GZNO while broadening the absorption spectrum of the semiconductor. Taken together, the Zn-rich GZNO synthesized by the topotactic method has the potential to serve as an efficient photocatalyst for visible-light water splitting.

⟨u2⟩iso (Å2) 0.0052(3) 0.0052(3) 0.0052(2) 0.0052(2)

bond

bond angle (deg)

an(1)−cat(1)−an(1) an(1)−cat(1)−an(2)

109.32 109.62

to calcination under NH3 and in product collected from a sequential O2/NH3 and NH3 nitridation reaction (Figure S7). Elucidating the directional component of crystalline defects can provide insight into anisotropic charge-transport processes that are highly relevant to photocatalysis. Shown in Figure 5 is a

■

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.5b02866. Additional FTIR, chemical analyses, XRD patterns, XPS spectra, and tables of structural refinement parameters (PDF)

Figure 5. fwhm plotted as a function of Q for NIST 674b ZnO (filled circles), GZNO-O2/NH3 (open squares), and GZNO-NH3 (open triangles). Shown are the fwhm values obtained from pseudo-Voigt fits to individual peaks.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected].

plot of the full width at half-maximum (fwhm) of pseudo-Voigt functions fitted to fully resolved diffraction peaks as a function of Q, compared to GZNO-NH3, GZNO-O2/NH3, and a ZnO reference. Although the GZNO-O2/NH3 particles exhibit a larger profile broadening than the reference material, the increase in the fwhm across the Q values is more dramatic in GZNO-NH3. Furthermore, a comparison of the fwhm of the

Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. Notes

The authors declare no competing financial interest. F

DOI: 10.1021/acs.inorgchem.5b02866 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

■

Wang, L.; Bai, J.; Hanson, J. C.; Warren, J. B.; Muckerman, J. T.; Fujita, E.; Rodriguez, J. A. J. Phys. Chem. C 2010, 114 (4), 1809−1814. (c) Han, W.-Q.; Liu, Z.; Yu, H.-G. Appl. Phys. Lett. 2010, 96 (18), 183112−3. (d) Reinert, A. A.; Payne, C.; Wang, L.; Ciston, J.; Zhu, Y.; Khalifah, P. G. Inorg. Chem. 2013, 52 (15), 8389−8398. (19) Li, Y.; Zhu, L.; Yang, Y.; Song, H.; Lou, Z.; Guo, Y.; Ye, Z. Small 2015, 11 (7), 871−876. (20) Li, X. H.; Shao, C. L.; Wang, D.; Zhang, X.; Zhang, P.; Liu, Y. C. Ceram. Int. 2014, 40 (2), 3425−3431. (21) Maeda, K.; Teramura, K.; Domen, K. J. Catal. 2008, 254 (2), 198−204. (22) Harris, G. F. P.; MacDermott, P. E. Flammability and explosibility of ammonia; Hazards VI: IChemE Symposium Series 49; The Institution of Chemical Engineers: Rugby, U.K., 1977; p 29037. (23) Cheary, R. W.; Coelho, A. J. Appl. Crystallogr. 1992, 25 (2), 109−121. (24) Balzar, D. L. H.; Ledbetter, H. Adv. X-ray Anal. 1995, 38, 397− 404. (25) Stephens, P. W. J. Appl. Crystallogr. 1999, 32 (2), 281−289. (26) APS Beamline 11-BM Standards Data, https://wiki-ext.aps.anl. gov/ug11bm/index.php/Standards_Data (accessed 08/13/2015). (27) (a) Kobayashi, Y.; Ke, X.; Hata, H.; Schiffer, P.; Mallouk, T. E. Chem. Mater. 2008, 20 (6), 2374−2381. (b) Sun, G.; Sun, L.; Wen, H.; Jia, Z.; Huang, K.; Hu, C. J. Phys. Chem. B 2006, 110 (27), 13375− 13380. (c) Li, F.; Liu, J.; Evans, D. G.; Duan, X. Chem. Mater. 2004, 16 (8), 1597−1602. (28) (a) Grégoire, B.; Ruby, C.; Carteret, C. Cryst. Growth Des. 2012, 12 (9), 4324−4333. (b) Evans, D.; Slade, R. T. Structural Aspects of Layered Double Hydroxides. In Layered Double Hydroxides; Duan, X., Evans, D., Eds.; Springer: Berlin, 2006; Vol. 119, pp 1−87. (c) Thomas, G. S.; Kamath, P. V. Solid State Sci. 2006, 8 (10), 1181−1186. (29) (a) Zhou, X.; Xie, Z.-X.; Jiang, Z.-Y.; Kuang, Q.; Zhang, S.-H.; Xu, T.; Huang, R.-B.; Zheng, L.-S. Chem. Commun. 2005, No. 44, 5572−5574. (b) Shinagawa, T.; Shibata, K.; Shimomura, O.; Chigane, M.; Nomura, R.; Izaki, M. J. Mater. Chem. C 2014, 2 (16), 2908−2917. (30) Brendt, J.; Samuelis, D.; Weirich, T. E.; Martin, M. Phys. Chem. Chem. Phys. 2009, 11 (17), 3127−3137. (31) (a) Maeda, K.; Teramura, K.; Takata, T.; Hara, M.; Saito, N.; Toda, K.; Inoue, Y.; Kobayashi, H.; Domen, K. J. Phys. Chem. B 2005, 109 (43), 20504−20510. (b) Maeda, K.; Domen, K. Chem. Mater. 2010, 22 (3), 612−623. (32) (a) McDermott, E. J.; Kurmaev, E. Z.; Boyko, T. D.; Finkelstein, L. D.; Green, R. J.; Maeda, K.; Domen, K.; Moewes, A. J. Phys. Chem. C 2012, 116 (14), 7694−7700. (b) Phivilay, S. P.; Roberts, C. A.; Puretzky, A. A.; Domen, K.; Wachs, I. E. J. Phys. Chem. Lett. 2013, 4 (21), 3719−3724. (33) Waring, R. K.; Hsu, W. Y. J. Appl. Phys. 1983, 54 (7), 4093− 4096. (34) Kurik, M. V. Phys. Status Solidi A 1971, 8 (1), 9−45. (35) Lee, K.; Lu, Y.-G.; Chuang, C.-H.; Ciston, J.; Dukovic, G. J. Mater. Chem. A 10.1039/C5TA04314J. Published online July 28, 2015. http://pubs.rsc.org/en/content/articlehtml/2016/ta/c5ta04314j (accessed October 15, 2015). (36) Yashima, M.; Yamada, H.; Maeda, K.; Domen, K. Chem. Commun. 2010, 46 (14), 2379−2381. (37) de Respinis, M.; Fravventura, M.; Abdi, F. F.; Schreuders, H.; Savenije, T. J.; Smith, W. A.; Dam, B.; van de Krol, R. Chem. Mater. 2015, 27 (20), 7091−7099.

ACKNOWLEDGMENTS We acknowledge financial support from Indiana University and NSF CAREER Grant DMR-0955028. Access to the powder Xray diffractometer and X-ray photoelectron spectrometer was provided by NSF CRIF Grants CHE-1048613 and DMR MRI1126394, respectively. Use of the APS 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. Synchrotron data for GZNO samples (Figure 4b) were collected as a part of the 2015 Modern Methods in Rietveld Refinement and Structural Analysis (MMRRSA) workshop, a school supported by the U.S. National Committee for Crystallography and the American Crystallographic Association. A trial dongle for TOPAS v5 was provided by Bruker AXS as part of the 2015 MMRRSA workshop. We thank Dr. Saul H. Lapidus for his assistance with synchrotron powder XRD data collection at 11-BM and the instructors at the 2015 MMRRSA for their instruction and guidance. D.P.C. is grateful to Professor David L. Bish for his suggestions in analysis of the synchrotron XRD data and for the use of TOPAS v4.1. D.P.C. is thankful for the financial supported provided by the IU Siedle Materials Fellowship.

■

REFERENCES

(1) Lewis, N. S.; Nocera, D. G. Proc. Natl. Acad. Sci., U.S.A. 2006, 103 (43), 15729−15735. (2) Nielander, A. C.; Shaner, M. R.; Papadantonakis, K. M.; Francis, S. A.; Lewis, N. S. Energy Environ. Sci. 2015, 8 (1), 16−25. (3) Walter, M. G.; Warren, E. L.; McKone, J. R.; Boettcher, S. W.; Mi, Q.; Santori, E. A.; Lewis, N. S. Chem. Rev. 2010, 110 (11), 6446−6473. (4) Yoshimura, J.; Ebina, Y.; Kondo, J.; Domen, K.; Tanaka, A. J. Phys. Chem. 1993, 97 (9), 1970−1973. (5) Wrighton, M. S.; Ellis, A. B.; Wolczanski, P. T.; Morse, D. L.; Abrahamson, H. B.; Ginley, D. S. J. Am. Chem. Soc. 1976, 98 (10), 2774−2779. (6) (a) Maeda, K.; Domen, K. J. Phys. Chem. C 2007, 111 (22), 7851−7861. (b) Wu, Y.; Lazic, P.; Hautier, G.; Persson, K.; Ceder, G. Energy Environ. Sci. 2013, 6 (1), 157−168. (7) Maeda, K. Phys. Chem. Chem. Phys. 2013, 15 (26), 10537−10548. (8) (a) Zhong, M.; Ma, Y.; Oleynikov, P.; Domen, K.; Delaunay, J.-J. Energy Environ. Sci. 2014, 7 (5), 1693−1699. (b) Higashi, M.; Domen, K.; Abe, R. J. Am. Chem. Soc. 2012, 134 (16), 6968−6971. (9) Serpone, N. J. Photochem. Photobiol., A 1997, 104 (1−3), 1−12. (10) (a) Ohno, T.; Bai, L.; Hisatomi, T.; Maeda, K.; Domen, K. J. Am. Chem. Soc. 2012, 134 (19), 8254−8259. (b) Fabian, D. M.; Hu, S.; Singh, N.; Houle, F. A.; Hisatomi, T.; Domen, K.; Osterloh, F. E.; Ardo, S. Energy Environ. Sci. 2015, 8 (10), 2825−2850. (11) Hisatomi, T.; Maeda, K.; Lu, D.; Domen, K. ChemSusChem 2009, 2 (4), 336−343. (12) Wang, J.; Huang, B.; Wang, Z.; Wang, P.; Cheng, H.; Zheng, Z.; Qin, X.; Zhang, X.; Dai, Y.; Whangbo, M.-H. J. Mater. Chem. 2011, 21 (12), 4562−4567. (13) (a) Maeda, K.; Domen, K. Chem. Mater. 2010, 22 (3), 612−623. (b) Maeda, K.; Takata, T.; Hara, M.; Saito, N.; Inoue, Y.; Kobayashi, H.; Domen, K. J. Am. Chem. Soc. 2005, 127 (23), 8286−8287. (14) Chen, H.; Wen, W.; Wang, Q.; Hanson, J. C.; Muckerman, J. T.; Fujita, E.; Frenkel, A. I.; Rodriguez, J. A. J. Phys. Chem. C 2009, 113 (9), 3650−3659. (15) Maeda, K.; Teramura, K.; Lu, D.; Takata, T.; Saito, N.; Inoue, Y.; Domen, K. Nature 2006, 440 (7082), 295−295. (16) Li, L.; Muckerman, J. T.; Hybertsen, M. S.; Allen, P. B. Phys. Rev. B: Condens. Matter Mater. Phys. 2011, 83 (13), 134202. (17) Jensen, L. L.; Muckerman, J. T.; Newton, M. D. J. Phys. Chem. C 2008, 112 (9), 3439−3446. (18) (a) Lee, K.; Tienes, B. M.; Wilker, M. B.; Schnitzenbaumer, K. J.; Dukovic, G. Nano Lett. 2012, 12 (6), 3268−3272. (b) Chen, H.; G

DOI: 10.1021/acs.inorgchem.5b02866 Inorg. Chem. XXXX, XXX, XXX−XXX