Observation of Surface-Bound Negatively Charged Hydride and

Jul 6, 2015 - ... Ivy C. Fortmeyer , James Eujin Park , Tao Zhang , Kuo Liao , Jing Gu , Yong Yan , Travis W. Shaw , Esta Abelev , and Andrew B. Bocar...
0 downloads 3 Views 3MB Size
Download PDF
Article pubs.acs.org/JPCC

Observation of Surface-Bound Negatively Charged Hydride and Hydroxide on GaP(110) in H2O Environments Coleman X. Kronawitter,† Martina Lessio,‡ Peng Zhao,† Christoph Riplinger,§ Anibal Boscoboinik,∥ David E. Starr,⊥ Peter Sutter,∥ Emily A. Carter,§,# and Bruce E. Koel*,† †

Department of Chemical and Biological Engineering, ‡Department of Chemistry, and §Department of Mechanical and Aerospace Engineering, Princeton University, Princeton, New Jersey 08544, United States ∥ Center for Functional Nanomaterials, Brookhaven National Laboratory, Upton, New York 11973, United States ⊥ Helmholtz-Zentrum Berlin für Materialien und Energie GmbH, Institute for Solar Fuels, D-14109 Berlin, Germany # Program in Applied and Computational Mathematics and Andlinger Center for Energy and the Environment, Princeton University, Princeton, New Jersey 08544, United States ABSTRACT: Surface-bound species on GaP(110) formed upon interaction with water were investigated through experiment and theory. These studies are motivated by and discussed in the context of electrocatalytic and photoelectrocatalytic schemes for solar fuel production, including especially observations of selective CO2 reduction to methanol in acidified aqueous solutions of CO2 and nitrogen-containing heteroaromatics. Experimentally, surface-bound species over 10 orders of magnitude of pressure were spectroscopically identified in situ using synchrotron-based ambient pressure photoelectron spectroscopy. Ga 3d and O 1s core-level spectra indicate that the interaction of GaP(110) with H2O induces formation of a partially dissociated adlayer, characterized by the presence of both Ga−OH and molecular H2O species. Measurements of the P 2p core level indicate formation of a negatively charged hydride that irreversibly bonds to surface P in vacuum. The surface densities of the hydroxide and hydride species increase with increasing pressure (surface coverage) of water. Periodic slab calculations using density functional theory were used to study several relevant water configurations at 298 K on this surface. Consistent with earlier theoretical predictions at 0 K, the calculations confirm that Ga− OH, molecular H2O, and P−H species are thermodynamically stable on the GaP(110) surface under experimental conditions. Isobaric measurements at elevated pressures were used to probe the thermal stabilities of adsorbed species as well as the oxidation of surface Ga and P. The observation of stable surface hydride formation induced by interaction with water is especially notable given the critical role of hydride transfer to catalysts and CO2 during chemical fuel synthesis reactions in aqueous environments. It is hypothesized that the observed high stability of the hydride on GaP may contribute to its associated remarkable near-100% faradaic efficiency for methanol generation by solar-driven CO2 reduction in acidified aqueous pyridine solutions [J. Am. Chem. Soc. 2008, 130, 6342] because such stability is known to yield high overpotentials for the competing hydrogen evolution reaction.

1. INTRODUCTION

reduction system was described in a recent study from Bocarsly and co-workers, which reported highly selective solar-driven photoelectrocatalytic reduction of CO2 to methanol at p-type GaP photocathodes in pyridine(Py)-containing acidified aqueous solutions at cathodic underpotential.4 Electrochemical CO2 reduction experiments with other electrode5−7 surfaces have been reported, but high faradaic efficiency toward methanol (near 100%) is observed only when using p-GaP, and more recently with p-CuInS2,8 electrodes. Following the initial report of photoelectrocatalytic CO2 reduction with GaP, Bocarsly and co-workers5 suggested a homogeneous mechanism for the photoelectrochemical reduc-

Developing an efficient technology to create liquid fuels from renewable energy sources represents one of the greatest technical challenges facing society.1 Solar-driven electroreduction of CO2 to carbon-containing fuels is a promising strategy because photoelectrocatalysis can facilitate the energy conversion and charge-transfer processes necessary for these endoergic chemical bond transformations to occur. Ideally, a catalyzed CO2 reduction process will produce liquid fuels that can be stored on a large scale and utilized subsequently in a carbon-neutral cycle. In order to yield useful products and to occur at reasonable potentials (e.g., enabled by solar energy), electrochemical CO2 reduction requires the transfer of multiple electrons and protons in a concerted manner, as occurs in nature.2,3 A particularly noteworthy and promising CO2 © 2015 American Chemical Society

Received: June 4, 2015 Published: July 6, 2015 17762

DOI: 10.1021/acs.jpcc.5b05361 J. Phys. Chem. C 2015, 119, 17762−17772

Article

The Journal of Physical Chemistry C

In summary, thus far only the proposed mechanism involving adsorbed hydrides13 and adsorbed DHP14−17 is consistent with experimental observations and theoretical predictions made on GaP. First, experiments show that Py-derived species are required for this catalysis and that the product yield heavily favoring methanol is unique to GaP − hence adsorbed intermediates are involved.4,5 Second, theoretical calculations predict that the hypothetical PyH• adsorbed on GaP(110) spontaneously ionizes to form PyH+, which then would immediately desorb as it strongly prefers to exist in aqueous solution.16 Thus, in addition to homogeneous PyH• being excluded as the active cocatalyst,9−12 adsorbed PyH• and PyH+ are also ruled out, leaving DHP as the most likely adsorbate involved.15−17 However, no direct observation of intermediates on GaP has been reported as yet. Herein, we experimentally investigate the interaction of water with GaP(110) to begin interrogation of this mechanism. We use ambient pressure photoelectron spectroscopy (APPES) to follow the O 1s, Ga 3d, and P 2p core levels to study the interaction of water with GaP(110) over 10 orders of magnitude of pressure. Via measurements of the P 2p level, we observe formation of a negatively charged hydride that irreversibly bonds to surface P on GaP(110), consistent with earlier theoretical predictions.13 The measurements confirm that Ga−OH and molecular H2O exist on this surface in the presence of water.13 We also report further density functional theory (DFT) calculations that consider relevant water configurations on the GaP(110) surface and their thermodynamic stabilities at room temperature. Experimental observations of the hydride species afford further analysis and discussion herein on the role of surface hydrides in selectively catalyzing CO2 reduction.

tion of CO2 catalyzed by pyridinium (for the reported conditions of pH = 5.2, approximately half of the Py is in the protonated form pyridinium, PyH+). In this postulated mechanism, the first step consists of the one-electron reduction of PyH+ to the aqueous pyridinyl radical (PyH•). In the second step, PyH• reacts with CO2 to form a carbamate complex. The validity of this mechanism was later disproved by computational studies,9−12 which report that the calculated reduction potential for the homogeneous reduction of PyH+ to PyH• (∼ −1.4 V vs. SCE) is much more negative than the experimentally observed potential.4 In addition, the pKa of PyH• was found to be ∼27 in theoretical studies,10,11 which suggests that formation of the carbamate species is unlikely to occur, as it would require at least partial deprotonation of the PyH• radical in the transition state. Recent theoretical studies of the relative stability of waterand Py-derived species in solution and adsorbed on GaP(110) led to the hypothesis of an alternative heterogeneous mechanism for CO2 reduction involving adsorbed nitrogen heterocycles.13−17 These studies motivated our present investigation of the interaction of water with GaP(110). Specifically, a report by Muñoz-Garciá and Carter13 considered the interaction between water and GaP(110) at 0 K. They predicted that water dissociation is thermodynamically favored on this surface and produces hydride-like species adsorbed on phosphorus sites along with water and hydroxyl bound to gallium sites. Following this work, Keith and Carter14 predicted that dihydropyridine (DHP) could exist in solution under the electrochemical conditions of the Bocarsly experiments. In particular, the authors determined that the homogeneous twoelectron reduction potential of Py to DHP, Py + 2H+ + 2e− → DHP, is −0.72 V versus SCE,14 which is close to the applied operating potential. However, homogeneous DHP, like homogeneous PyH•, was ruled out as an electrocatalyst because of high predicted barriers for CO2 insertion to form carbamate. Therefore, on the basis of the predicted phosphorus-bound hydride species induced by dissociation of water mentioned above,13 Keith and Carter15,17 hypothesized that adsorbed Py could be reduced to DHP on GaP(110) by adsorbed hydridelike species and protons in solution. Keith et al.16 also predicted the stable adsorption of both Py and DHP on the GaP(110) surface at room temperature, which supports the idea that the two-electron reduction of Py should occur on the surface prior to further catalytic reduction of CO2. Alternative mechanisms for the reduction of CO2 catalyzed by Py-derived species have been proposed. In a mechanism suggested by Batista and co-workers12 relevant for Pt electrodes, PyH+ is reduced to Py, resulting in adsorbed hydride; in a subsequent step, the surface hydride is transferred to the CO2 molecule together with a proton from PyH+ to produce formic acid. Although this mechanism can explain experimental observations on Pt and features low activation barriers, the nagging question remains as to why PyH+ is required, given that this mechanism invokes only PyH+ as a proton donor; therefore, any Brønsted acid of similar acidity should be sufficient to reduce CO2. However, this is not the case; only nitrogen heterocycles have been effective as cocatalysts. Recently, an alternative homogeneous mechanism involving DHP was suggested.18 However, a homogeneous mechanism cannot explain the significant electrode dependence of the catalysis, namely why underpotentials and high selectivity are observed only at certain electrodes.

2. EXPERIMENTAL PROCEDURES 2.1. Ambient Pressure Photoelectron Spectroscopy. In situ APPES (or high-pressure X-ray photoelectron spectroscopy (HP-XPS)) experiments were performed at beamline X1A1 at the National Synchrotron Light Source (NSLS). The system19 operates over a wide range of pressures through use of a differentially pumped hemispherical analyzer (Specs Phoibos 150 NAP) equipped with a CEM 9 channeltron detector and a 300 μm entrance aperture to the first differential pumping stage. The analyzer entrance normal was oriented 70° from the incident X-ray beam and 20° from the surface normal of the sample. GaP(110) single crystals (Princeton Scientific) were cleaned in vacuum using a series of sputter−annealing cycles, using methods described in the literature.20 Each cycle consisted of 1 kV Ar+ ion bombardment followed by sample heating to 500− 550 °C. The cycles were repeated until no C 1s signal was detected. After similar treatments of GaP(110) in a separate chamber, a 1 × 1 LEED pattern was observed. The base pressure of the chamber was 3−4 × 10−9 Torr. We checked for radiation- and photoelectron-induced chemistry by (1) monitoring spectra over time, (2) monitoring spectra and changing the position sampled on the crystal, and (3) monitoring spectra and increasing the background pressure. No differences in the spectra induced by these changes were observed. The spectra in this article were signal-averaged over several scans, and the times to acquire the scans were 144 s (Ga 3d), 40 s (P 2p), and 360 s (O 1s). When photon energies were changed, the binding energies were calibrated using the Ga 3d peak, which originates from high kinetic energy photoelectrons and thus probes deeply into the substrate. All binding energies 17763

DOI: 10.1021/acs.jpcc.5b05361 J. Phys. Chem. C 2015, 119, 17762−17772

Article

The Journal of Physical Chemistry C Table 1. Relaxed Structures of All Water Configurations Considered in This Worka

a

Each water configuration is defined by the coverage (expressed in ML) and the degree of dissociation (D).

are referenced to EF = 0 eV binding energy. Spectra were fit to Gaussian−Lorentzian convolutions, and the full width at halfmaximum of each peak was fixed for a given spectrum. The integrated intensities of the spin−orbit split peaks of the P 2p level were fixed by the required 2:1 peak-area ratio. 2.2. Density Functional Theory. We performed DFT21,22 calculations with periodic boundary conditions to identify relevant water configurations on the GaP(110) surface and assess their associated thermodynamic stabilities. All calculations were performed using the software package VASP,23−25 the PBE exchange-correlation functional,26 and default projector augmented wave (PAW) potentials27 to describe nuclei and frozen core electrons (1s2s2p3s3p3d for Ga, 1s2s2p for P, and 1s for O). We described the rest of the electrons using a planewave basis with a kinetic energy cutoff of 800 eV. We used 6 × 4 × 1 k-point sampling based on the Monkhorst− Pack scheme.28 The Brillouin zone was integrated using the Gaussian smearing method with smearing width equal to 0.05 eV. For the calculation of an isolated water molecule, we only sampled the Γ point with a smearing width equal to 0.001 eV. Overall, with these computational parameters, the total energy was converged to within 1 meV/atom. We used the same model for the GaP surface as that used in previous work by our group (a seven-layer slab generated from a 1 × 2 unit cell of the GaP(110) surface, resulting in four GaP formula units in each layer).13 We modeled different water coverages, defined as the number of adsorbate molecules divided by the number of Ga adsorption sites per surface, ranging from 0.25 to 2 monolayers (ML). Adsorbates were symmetrically placed on both sides of the slab to eliminate artificial net dipoles. A vacuum space larger than 20 Å was inserted between slabs to avoid interaction of periodic images along the surface normal direction. During geometry optimization, we allowed all the adsorbate atoms and the four outermost GaP layers to relax while keeping the cell shape and volume fixed. We also kept the three middle layers frozen to capture the effects arising from the bulk regions of a semiinfinite crystal. Finally, we verified that our calculated structures

are minimum-energy structures by calculating the corresponding vibrational frequencies and verifying that none had an imaginary frequency. We calculated only the frequencies of the atoms that were allowed to relax. For the isolated water molecule, we also computed the frequencies associated with rotational and translational modes using the ideal gas, rigid rotor, and harmonic oscillator approximations. The vibrational frequencies were computed from a numerical Hessian calculated using displacements of ±0.02 Å and analytic gradients. We computed a phase diagram for water on the GaP(110) surface to determine the most stable water structures on this surface under experimental conditions (ultrahigh vacuum (UHV) and room temperature). To construct the phase diagram, we considered the variation of the room-temperature adsorption free energy of all the modeled water configurations (listed in Table 1) with respect to the water chemical potential shifting from water-poor to water-rich conditions (i.e., from low to high water pressures at the surface). Using the chemical potential, we can simulate different water conditions at the surface. Water-rich (water-poor) conditions at the surface are represented by gas (liquid) phase water in a reservoir in contact with the surface, because at room temperature gaseous (liquid) water is unstable (stable) in the reservoir. More information regarding the use of the water chemical potential in constructing the phase diagram can be found in previous work in which a similar diagram was computed for a different surface.29 We calculated the adsorption free energy, ΔΩ, dependent on the water chemical potential, μw, using the following equation: ΔΩ(T , p) =

1 (Gslab + n W − Gslab − nμw (T , p)) 2

Gslab+nW is the free energy of the GaP slab-water complex containing n water molecules, and Gslab is the free energy of the bare GaP slab. Both energies include zero point energy and thermal corrections at room temperature. n is the total number of water molecules adsorbed on both sides of the slab. μw is the 17764

DOI: 10.1021/acs.jpcc.5b05361 J. Phys. Chem. C 2015, 119, 17762−17772

Article

The Journal of Physical Chemistry C

Figure 1. Isothermal APPES spectra from GaP(110) after exposures to H2O in UHV and in situ for 1 × 10−7 Torr to 0.9 Torr H2O: (a) O 1s spectra and (b) normalized Ga 3d spectra. Figures on the left show data acquired at 4 × 10−9 Torr base pressure after exposing the sample to H2O at 1 × 10−7 Torr. Figures on the right show in situ measurements at the indicated pressures. Bold lines are experimental data, and fine lines are fitted component contributions.

chemical potential of water, which depends on temperature, T, and pressure, p. We introduce the factor 1/2 because as discussed above we model adsorbates on both sides of the slab, but we are interested in the adsorption properties of one surface only. The chemical potential μw is shifted from waterpoor to water-rich conditions using the parameter Δμw, as illustrated in the following equation:

electrons are sensitive to both the formal oxidation state and local chemical environment of the probed atom. The probing depth of the measured spectrum is defined by the inelastic mean free path of the ejected photoelectrons, and we used photon energies lower than those available in conventional laboratory X-ray sources to increase the surface sensitivity of our measurements. Figure 1a shows O 1s core-level spectra from GaP(110) at 38 °C after exposures of up to 76 langmuirs (1 langmuir = 10−6 Torr·s) of H2O in UHV conditions and in situ for H2O pressures up to 0.9 Torr. Although this temperature is higher than room temperature, we confirmed in independent experiments that the results presented herein are valid at room temperature as well. The interaction of H2O with GaP(110) is associated with a new feature in the O 1s spectra that grows larger with increasing exposure. After the 76 langmuir H2O dose, this feature does not increase further, i.e., the surface has reached saturation coverage of O-containing species at this temperature under UHV conditions. At low coverages, the O 1s spectrum in the range of 530−533 eV was deconvolved into two component peaks, which we assign to OH (531.6 eV) and molecular H2O (532.7 eV) adsorbed on the surface.31,32 The species corresponding to these two peaks remain on the surface in the presence of background H2O for all pressures examined. At 0.1 Torr, a new peak at 536.0 eV was

μw = Δμw + Gwgas

Ggas w is the free energy of a water molecule in the gas phase and includes zero point energy and thermal corrections at room temperature. When Δμw is equal to 0 eV, μw is the chemical potential of water in the gas phase, which corresponds to waterrich conditions at the surface; when Δμw is equal to −0.09 eV, μw is the chemical potential of water in the liquid phase describing water-poor conditions at the surface. The value −0.09 eV is the difference between the experimental values of the water chemical potential in the liquid and gas phases at room temperature.30

3. RESULTS AND DISCUSSION 3.1. Ambient Pressure Photoelectron Spectroscopy: Isothermal Conditions. X-ray photoelectron spectroscopy can be used to probe the chemical states of surface-bound species because changes in the binding energies of core-level 17765

DOI: 10.1021/acs.jpcc.5b05361 J. Phys. Chem. C 2015, 119, 17762−17772

Article

The Journal of Physical Chemistry C observed. This peak is attributed to photoelectrons originating from vapor phase H2O.32 At 0.4 Torr, a fourth peak at 533.7 eV, labeled *, was observed that we attribute to photoelectrons originating from interacting coadsorbed OH and H2O species. Above 0.1 Torr, the total signal intensity decreases because of photoelectron scattering away from the analyzer aperture. The assignments of OH and H2O to the peaks at 531.6 and 532.7 eV, respectively, are based on literature assignments: the adsorption of H2O and OH has been studied on a number of surfaces, including metals,31,33 oxides,32,34 and semiconductors.35 Zhang and Ptasinska recently studied the interactions of water with GaAs(100)35 and GaP(111)36 by high-pressure XPS. These papers assign four separate peaks associated with both O and OH bonded to surface Ga, which is based on methods in a prior study of GaAs(100).37 Herein, after consideration of available evidence and the limitations of the resolution of these particular measurements, we assign only OH and H2O peaks (use of additional peaks was not found to improve the overall fit in a statistically significant result). Figure 1b shows Ga 3d spectra excited by 300 eV photons in both the low-coverage UHV regime described above and in situ in the presence of an elevated pressure of H2O in the background. The substrate Ga 3d peak, in which the two spin− orbit split components are unresolved, is located at 19.5 eV. At and above a pressure of 1 × 10−7 Torr, a new peak at higher binding energy (ΔE = 1.1 eV) was observed. This peak grows with increasing H2O pressure up to 0.4 Torr. The energy of this peak relative to that for GaP is consistent with formation of a Ga−OH species. The presence of this species was not observable in Ga 3d spectra for the low-coverage regime within a level of high statistical significance for our fitting. This is due in part to contributions from subsurface Ga to the signal, the fact that not every surface Ga interacts with OH (see below), and that there is in general a very small chemical shift associated with the formation of Ga−OH. Figure 2a shows P 2p spectra taken on GaP(110) under the same conditions as above. For the clean GaP(110) surface, the P 2p region shows a partially resolved set of peaks from the 2p1/2 and 2p3/2 spin−orbit split components at 129.8 and 128.9 eV, respectively. The two peak contributions were strictly fit according to the required 2:1 area ratio of these peaks. For exposures up to 6 langmuirs H2O, no changes in the P 2p spectra were observed. Above this exposure and at higher background H2O pressures (above UHV), the spectra do change and it is necessary to include a new P 2p doublet at binding energy slightly higher than that of the substrate in order to fit the spectra. The peaks are detected with small signal intensity in UHV after exposure to water, and their intensities increase considerably when the surface is probed in situ at higher pressures (up to 0.4 Torr). An expanded view of these new peaks is shown in Figure 2b, which directly compares the GaP(110) surface when clean and in the presence of 0.4 Torr H2 O. After the pressure was raised to 0.9 Torr for approximately 5 min and then decreased by a factor of 9, the P 2p spectrum associated with the high-pressure condition remained, which shows that an irreversible reaction had taken place. The data in Figure 2 indicate the presence of a new phosphorus chemical state with a P 2p binding energy slightly greater than that of the Ga-bound P of the GaP(110) substrate. We assign the two new H2O pressure-induced spin−orbit-split peaks at 129.9 and 130.8 eV to surface P−H species. Previous work on H2S dissociation on InP38 clearly shows that the

Figure 2. Isothermal APPES spectra. (a) Normalized P 2p spectra from GaP(110) after exposures to H2O in UHV and in situ for 1 × 10−7 Torr to 0.4 Torr H2O. Figures on the left show data acquired at 4 × 10−9 Torr base pressure after exposing the sample to H2O at 1 × 10−7 Torr. Figures on the right show in situ measurements at the indicated pressures. (b) Expanded view of the P 2p spectrum from clean GaP(110) and from GaP(110) in an ambient of 0.4 Torr H2O, which is included to clearly show the contribution of the hydride peaks. Bold lines are experimental data, and fine lines are fitted component contributions.

presence of P−H on the surface of a phosphide semiconductor is associated with P 2p peaks located at binding energies only slightly higher than those of the substrate. This is consistent with the charge transfer (0.7e) from surface P predicted to accompany formation of the negatively charged hydride.13 17766

DOI: 10.1021/acs.jpcc.5b05361 J. Phys. Chem. C 2015, 119, 17762−17772

Article

The Journal of Physical Chemistry C The P−H and Ga−OH component peak areas increased as the H2O pressure was increased because in these conditions pressure is proportional to surface coverage. The Hertz− Knudsen equation describes the relationship between the flux of gas to the surface, Z, and the pressure, P: Z = NAP(2πMRT)−1/2, where T is the temperature and M is the molar mass of the gas. The surface coverage, θ, is equal to the product of the flux and the residence time, τ: θ = Zτ. The rate of dissociation is given by Rdiss = kdissθn; therefore, at low coverages, the rate of dissociation increases with increasing H2O pressure. There is a related mechanism that has been discussed in the literature that may further explain the origins of our observed increase in concentration of Ga−OH and P−H at high H2O pressures. Andersson et al.33 reported dissociation of H2O on Cu(110) that is driven by stabilization of adsorbates through H2O(donor)−OH(acceptor) complexes. The increased coverage of H2O on the surface that is associated with increased background pressure (see above) leads to a greater degree of hydrogen bonding among adsorbed H2O and OH. The H2O− OH bond is stronger than the H2O−H2O bond.31 Through Brønsted−Evans−Polanyi relations for water dissociation,39 a lower activation barrier for water dissociation results.33 This analysis from the literature provides a mechanism by which increased H2O pressure and therefore H2O coverage can result in an increase in amount of stable dissociation products. The experimental evidence presented in Figures 1 and 2 indicates that the interaction with H2O induces the formation of surface P−H, Ga−OH, and molecular H2O on GaP(110). These results are consistent with a number of observations of spontaneous H2O dissociation on related surfaces in recent years: GaAs(100), 35 GaP(111), 36 GaAs nanowires, 40 InP(100),41 and GaN(101̅0),42 as well as several recent theoretical studies on related III−V surfaces.43−45 We next performed periodic slab calculations using DFT to verify the thermodynamic stabilities of these species under experimental conditions. 3.2. Adsorbate−Surface Phase Diagram from Density Functional Theory. Our DFT results, presented below, are in very good agreement with the experimental observations: at room temperature and in water-poor conditions, Ga−OH, adsorbed molecular H2O, and P−H species are found to be thermodynamically stable on the GaP(110) surface. The stability of these species is shown by the computed H2O phase diagram on GaP(110) (Figure 3), which illustrates the thermodynamic stabilities of several water configurations for different water conditions at the surface. In order to build the phase diagram, we modeled several possible initial configurations of water, varying coverage and degree of dissociation. Coverages considered range from 0.25 ML (i.e., one water molecule per surface unit cell) to 2 ML (i.e., eight water molecules per surface unit cell), where coverage is defined as the number of adsorbate water molecules divided by the number of Ga adsorption sites per surface unit cell. For each coverage, we investigated an undissociated configuration and a dissociated configuration. Table 1 displays the configurations considered for each coverage along with its respective degree of dissociation. At 0.25 ML, the one water molecule in the unit cell is dissociated heterolytically by adsorbing the hydroxide on a surface Ga atom and the proton on a neighboring surface P atom. In this way, 100% dissociation is obtained. At 0.5 ML, two water molecules are in the unit cell. We investigated the effect of dissociating one molecule, which

Figure 3. Computed H2O phase diagram on GaP(110) at T = 298.15 K. ΔΩ is the adsorption free energy, and Δμw is the chemical potential shift from water-rich conditions (i.e., the chemical potential of gasphase water) as illustrated in section 2.2. Each water configuration is defined by the coverage (expressed in ML) and the degree of dissociation (D). Solid lines represent undissociated configurations, and dashed lines represent dissociated configurations.

corresponds to 50% dissociation, and then also adding an undissociated water molecule to this configuration, reaching 0.75 ML coverage with 33% dissociation. These three configurations were investigated to study whether coadsorbed water molecules assist dissociation. Finally, we modeled 1 and 2 ML coverages, both with 50% dissociation within the first monolayer. These two configurations were considered in order to investigate the effect of a solvating layer (i.e., the second water monolayer) on dissociation occurring in the first monolayer. Note that the dissociated configurations considered in this work have different geometries with respect to those modeled in previous work.13 In the phase diagram, for each value of water chemical potential (i.e., water conditions at the surface), the configuration with the most negative adsorption free energy represents the most thermodynamically stable one. Overall, the 1 ML, half-dissociated configuration (blue dashed line in Figure 3) represents the most stable water configuration for both water-poor and water-rich conditions. This configuration involves the Ga−OH, P−H, and adsorbed molecular H2O species observed in this work by APPES. In addition, in this half-dissociated configuration, Ga−OH and H2O species interact via hydrogen bonding (Figure 4), which was also observed here experimentally (peak labeled * in Figure 1a). In addition to the agreement between experimental and computational findings, the phase diagram yields three other important results. First, we observe that adsorption of water is always favored except for the 2 ML undissociated configuration and the 0.25 ML dissociated configuration when approaching water-poor conditions. In fact, only under these conditions do the lines corresponding to these configurations fall above the black horizontal line (0 ML) representing the clean surface (i.e., the adsorption free energy becomes positive and adsorption of water in these configurations is no longer thermodynamically favored). Second, inspection of the water phase diagram reveals that for coverages larger than 0.25 ML, water dissociation is always thermodynamically favored even in water-poor conditions. In fact, each dissociated configuration (dashed lines) is 17767

DOI: 10.1021/acs.jpcc.5b05361 J. Phys. Chem. C 2015, 119, 17762−17772

Article

The Journal of Physical Chemistry C

peak at 530.4 eV.35 When the sample is heated, evidence for this new oxide peak appears already at 80 °C, which was the lowest elevated temperature examined. The peaks corresponding to surface-bound OH and H2O remain at all elevated temperatures examined. The presence of molecular H2O adsorbed on the surface at such high temperatures seems remarkable; however, it is consistent with results from prior high-pressure photoelectron spectroscopy studies in the literature. Surface-bound molecular H2O at elevated temperatures in high H2O pressures has been reported previously on the surface of TiO2(110).32 In 0.375 Torr H2O, surface-bound molecular H2O was observed on GaP(111) up to 400 °C.36 P 2p spectra recorded at 0.1 Torr H2O (Figure 5b) indicate the presence at elevated temperatures of a new peak at higher binding energy (ΔE = 4.7 eV) than that of P in GaP(110). The intensity of this peak (also due to a partially resolved spin− orbit split doublet) increases with increasing temperature up to the highest temperature examined, 360 °C. The high binding energy of this peak is indicative of oxidation of surface P. An oxide peak at 21.3 eV appears and increases in intensity in the Ga 3d spectra at elevated temperatures (Figure 5c). This oxide peak is 0.8 eV higher in binding energy than that of the surface OH identified in the isothermal measurements (at 20.6 eV; Figure 1b). On the basis of the measured core-level shifts, we can put forward some description of the surface oxide using the available literature. We observe the difference in P 2p peaks between the surface oxide and GaP to be ΔE = 4.7 eV. Prior work on the oxidation of GaP47 reports that this shift for GaPO4 is 5.5 eV and for P2O5 is 5.8 eV. However, studies of P 2p core-level shifts associated with the oxidation of InP48 report an average of 4.7 eV for InPO4 and 6.6 eV for P2O5. The binding energy associated with a phosphine oxide is about 2 eV lower than that observed here.49 These comparisons suggest to us that the charge on oxidized surface P in our study is similar to that in GaPO4, although the structure of the oxide itself is likely unique to the surface. Given the coinciding observation of both Ga oxide and P oxide species, we currently favor assignment of a network structure that may contain bridging oxygen atoms between Ga and P, as has been suggested in recent studies of water adsorption on GaAs(100)35 and GaP(111),36 although we do not suggest a specific structure here. The oxides observed here at high temperature are unlikely to play an important role in the chemistry that motivates this report, because electrochemical CO2 reduction experiments take place at room temperature and under cathodic polarization. It has been observed in these conditions that metal oxide electrodes are reduced to metals.50 We can expect similar behavior for GaP. The thermal stability of the P−H species identified in Figure 2 can also be assessed using the isobaric P 2p measurements in Figure 5b. The areas of the peaks at 129.9 and 130.8 eV decrease with increasing temperature, up to 360 °C. At this temperature, the hydride peaks’ areas are decreased by 90%. The implications of this thermal stability for solar fuel reactions will be discussed below. 3.4. Role of Negatively Charged Hydrides in H+ and CO2 Reduction Reactions. Our experimental and theoretical results provide insights that aid discussions aimed broadly at the role of negatively charged surface hydride species in solardriven photoelectrochemical and photoelectrocatalytic CO2 reduction reactions. Without molecular catalysts, illuminated

Figure 4. Half-dissociated configuration of water on GaP(110) predicted by DFT calculations: the proton binds to a P adsorption site while the hydroxide binds to a Ga adsorption site. An undissociated H2O molecule interacts with the hydroxide via a hydrogen bond (represented by the blue dotted line).

more stable (lower free energy) than the corresponding undissociated configuration (solid lines), except for the 0.25 ML case. In previous work,13 water dissociation was predicted to be favored only for coverages larger than 2 ML. Here we find water dissociation to be favorable at much lower coverages (0.5 ML), resulting from finding a more stable adsorption geometry for the partially dissociated configurations than was found earlier. In particular, in our current structures, an adsorbed hydroxide deriving from water dissociation is stabilized by the presence of a coadsorbed, undissociated water molecule that promotes dissociation via hydrogen-bonding, as shown in Figure 4. The 0.25 ML anomaly is therefore easy to understand: dissociation is not favored at that low coverage because there are not enough water molecules on the surface to promote dissociation via hydrogen bonding. Third, relative adsorption free energy values of the 2 ML 50% D and 1 ML 50% D configurations suggest that at room temperature the formation of a solvation layer on top of the first adsorbed layer of water is not thermodynamically favored (i.e., the 2 ML 50% D configuration is higher in energy than the 1 ML 50% D configuration). This occurs because hydrogen bonds among the first and second monolayers are not strong enough to overcome the entropic contribution that drives the evaporation of the second monolayer at room temperature. In addition, this finding is consistent with the results of molecular dynamics simulations of liquid water in contact with metal surfaces reported by Limmer et al.,46 which shows the formation of a hydrophobic water layer on the surface. In summary, our computational findings are in excellent agreement with experimental results: we predict that all the experimentally observed species are thermodynamically stable at room temperature and in water-poor conditions. In addition, we find that water adsorption and dissociation is always thermodynamically favored on this surface if the Ga−OH species are stabilized by a hydrogen bond with an undissociated H2O molecule, with 1 ML being the most thermodynamically favored coverage under water-poor conditions. 3.3. Ambient Pressure Photoelectron Spectroscopy: Isobaric Conditions. Additional experiments were conducted to assess the thermal stabilities of these adsorbates as well as to study the oxidation of GaP(110) caused by the presence of H2O. Figure 5 shows isobaric APPES measurements of O 1s, P 2p, and Ga 3d spectra taken in situ at 0.1 Torr H2O from 38 to 360 °C. O 1s spectra (Figure 5a) indicate that increasing the sample temperature in the presence of 0.1 Torr H2O causes the growth of a surface oxide phase that is characterized by an O 1s 17768

DOI: 10.1021/acs.jpcc.5b05361 J. Phys. Chem. C 2015, 119, 17762−17772

Article

The Journal of Physical Chemistry C

Figure 5. Isobaric APPES measurements on GaP(110) in situ at 0.1 Torr H2O from 38 to 360 °C: (a) O 1s spectra, (b) normalized P 2p spectra, and (c) normalized Ga 3d spectra. Bold lines are experimental data, and fine lines are fitted component contributions.

intermediates. Possible intermediates include adsorbed CO, COH, CH−OH, CH2−OH, and O−CH3. These observations highlight the important roles that Hads species play in both the initial activation of CO2 molecules and additional hydrogenation steps to hydrocarbons and alcohols. The Hads species has also been proposed to be intimately involved in Py-catalyzed CO2 reduction, which was the primary motivation for this study. Keith and Carter showed that one possible source of the experimentally observed reduction wave associated with CO2 reduction in aqueous Py/PyH+ is the twoelectron reduction of Py to ortho- or para-dihydropyridine (oand p-DHP), Py + 2H+ + 2e− → DHP.14 This suggestion is relevant to our observation of negatively charged Hads in water environments: one possible formation mechanism for surfacebound DHP is hydride transfer to Py, which is bound to the surface via the N lone pair, followed by or concomitant with proton transfer from a Brønsted acid in solution.15,17 We show here that the interaction of water with GaP(110) yields a concentration of Hads that could facilitate such a reduction of surface-bound catalysts. Other proposed mechanisms explicitly emphasize the importance of Hads for Py-catalyzed CO2 reduction, and our work may also inform these analyses. In a related theoretical

GaP photocathodes under applied cathodic potential in aqueous CO2 solutions reduce CO2 to products including HCOO−, HCOOH, CH3OH, and HCOH.51−54 Alternate product distributions, including significant yields of CO, are observed in nonaqueous electrolytes.55 It is known that the addition of a small amount of water into organic solvents accelerates the hydrogenation of CO2 and therefore the formation of HCOO−/HCOOH. In this way, H2O is known to play an important role in the reaction by serving as a proton source.56 More specific observations concerning hydrogen can be made in the context of proton-coupled electron transfer. The carbon atom of the amphoteric CO2 molecule is susceptible to attack by nucleophilic negatively charged surface hydrides.57 This activation induces a negative charge on the oxygen atoms, which can be stabilized by a Brønsted acid.57 Although the role of the GaP surface in aqueous CO2 reduction reactions is not understood, some mechanistic reports are available for GaAs electrodes, and the reported conclusions are likely to be at least in essence applicable to GaP. Frese and Canfield58 studied pHdependent GaAs electrochemistry and concluded that the rate of CO2 reduction to methanol is limited by a surface chemical reaction involving adsorbed H (Hads) and unknown surface 17769

DOI: 10.1021/acs.jpcc.5b05361 J. Phys. Chem. C 2015, 119, 17762−17772

Article

The Journal of Physical Chemistry C study of Pt electrochemistry, Batista and co-workers12 suggested an acid-catalyzed reduction mechanism, in which the electrochemical reduction of PyH+ yields a surface-bound hydrogen, Pt−Hads, that facilitates CO2 reduction by protoncoupled hydride transfer, CO2 + Pt−Hads + PyH+ + e− → Py + Pt + HCO2H. A recent experimental investigation of this system by Lebegue et al.7 provides further evidence that the initial mechanistic role of PyH+ is to establish a concentration of Hads. This study used Ir, Pt, Au, and glassy carbon electrodes, materials with differing metal−hydrogen bond strengths, to assess the role of reduction of PyH+ to metal−Hads in catalysis of CO2 reduction. It was found that only Pt and Ir facilitate the formation of Hads and subsequently produce a catalytic interaction between CO2 and Hads. These studies have provided convincing evidence that the first mechanistic step of Pycatalyzed CO2 electroreduction on Pt, and the origin of the observed cathodic wave at −0.6 V in cyclic voltammetry studies, is the one-electron reduction of PyH+ to surface Hads. In this study, we do not consider the molecular cocatalyst, but this proposed mechanism from the literature is topical because it also emphasizes the importance of the Hads species. The efficiency and selectivity of catalyzed CO2 reduction processes may therefore be related to the nature of the negatively charged hydride on electrode surfaces. As indicated in ref 7, materials with intermediate free energies of H adsorption (Pt and Ir), which are also materials with a low overpotential for the hydrogen evolution reaction (HER),59 produce a catalytic interaction between PyH+ and CO2. Keith and Carter15 noted that materials with low HER overpotential should be active for CO2 reduction because they should also facilitate hydrogen transfer to CO2. GaP however is a high HER overpotential material. In order to understand the high selectivity toward methanol observed for Py-catalyzed CO2 reduction on GaP, it is useful then to consider the strength of the P−Hads interaction, of which we provide an experimental measure in this study, in the context of electrochemical HER. When limited by the Tafel reaction, 2Hads → H2, the most important descriptor for HER activity is the binding free energy of H.59 For metal electrodes, the relationship between HER exchange current density and free energy of H adsorption follows the Sabatier principle. Materials with intermediate H adsorption free energies, such as Pd, Pt, and Ir, exhibit the highest HER activities. The temperature at which H desorbs from surfaces should serve as a rough experimental indicator of the strength of metal−H bonds (rough because the energy associated with the desorption temperature overestimates the metal−H bond strength if there is an activation barrier for H2 dissociation). When a sample is heated in vacuum, the ratelimiting step for H2 desorption is typically the surface recombination reaction, 2Hads → H2. For example, on Ag(111), a poor HER catalyst that is associated with a low free energy of H adsorption, a saturation coverage of atomic hydrogen leads to H2 desorption at about −73 °C60 (again, a value that includes the added thermal energy required to overcome the H 2 dissociation activation barrier). The desorption peak for Pt(111), a good HER catalyst that is associated with an intermediate free energy of H adsorption, occurs below 130 °C.61 For W surfaces, which are poor HER catalysts and associated with large free energies of H adsorption, H2 desorption occurs at higher temperatures of 130 to 230 °C.62 In this context, the experimentally observed thermal stability for P−Hads on GaP(110) is very large, given

that this species is observable at an electrode temperature of 360 °C in 0.1 Torr H2O. Although the desorption temperature is an imperfect descriptor for activity, the measured thermal stability of the P−Hads species on GaP(110) is consistent with its associated high HER overpotential. The inability of GaP to catalyze H+ reduction is likely an important factor influencing the near100% faradaic efficiency for methanol generation from Pycatalyzed CO2 electroreduction using GaP. It has been observed in the literature that glassy carbon, which is also a poor HER catalyst, does not produce any cathodic current associated with CO2 reduction in similar conditions.5,63 However, there is a very important distinction in this case: glassy carbon is a poor HER catalyst because of its contrasting extremely weak interaction with H. The following observations in the literature further support the claim that the CO2 reduction selectivity of GaP relates to its inability to catalyze electrochemical HER. Because GaP is not catalytically active for HER in water, a portion of photoexcited electrons transported to the GaP−electrolyte interface participate instead in competing electrode decomposition reactions.64 In acidic conditions, therefore, GaP is not photoelectrochemically stable for HER. No activity is reported after 15 h in ref 65. In stark contrast, when photoexcited electrons in GaP drive the reduction of CO2 in aqueous Py/ PyH+, the electrode is stable for at least 30 h.4 Although both reactions are believed to require the participation of Hads, their associated electrochemical stabilities differ considerably. This contrast could originate from the fact that the execution of the electrochemical Tafel reaction, 2Hads → H2, requires Hads diffusion, whereas acid-catalyzed CO 2 reduction or Py reduction to DHP requires no surface diffusion.14 When participating in acid-catalyzed CO2 reduction reactions in the presence of Py/PyH+, photoexcited electrons in GaP could be efficiently shuttled to the catalyst or CO2, which prevents their involvement in competing electrochemical decomposition reactions.

4. CONCLUSIONS We have presented an experimental and theoretical investigation of the interaction of water with GaP(110). Experimentally we determine that exposure to water produces a partially dissociated surface layer consisting of Ga−OH, molecular H2O, and P−H species. The coverages of Ga−OH and P−H increase with increasing H2O pressure up to the largest pressure studied, 0.4 Torr. Periodic DFT calculations were used to identify relevant water configurations on the GaP(110) surface and calculate their ambient temperature thermodynamic stabilities. The calculations confirm the thermodynamic stability of the partially dissociated layer identified experimentally. Isobaric measurements at 0.1 Torr H2O pressure were used to probe the thermal stabilities of adsorbed species as well as the oxidation of surface Ga and P. P−H was observed on the surface up to 360 °C, which indicates a relatively high thermal stability for this species. Both surface Ga and P oxidize during temperature elevation in the presence of water vapor, beginning between 80 and 120 °C. These observations enable us to discuss the role of negatively charged hydrides in photoelectrocatalytic H+ and CO 2 reduction reactions and to further interpret related results from the literature. We hypothesize that the observed stability of P−H species contributes to the high faradaic efficiency for methanol generation associated with illuminated GaP in 17770

DOI: 10.1021/acs.jpcc.5b05361 J. Phys. Chem. C 2015, 119, 17762−17772

Article

The Journal of Physical Chemistry C aqueous solutions of CO2 and Py/PyH+. The strongly bound surface hydrides lead to low activity for the hydrogen evolution reaction, which directly competes with CO2 reduction under the reported conditions.

■

(12) Ertem, M. Z.; Konezny, S. J.; Araujo, C. M.; Batista, V. S. Functional Role of Pyridinium During Aqueous Electrochemical Reduction of CO2 on Pt(111). J. Phys. Chem. Lett. 2013, 4, 745−748. (13) Muñoz-García, A. B.; Carter, E. A. Non-Innocent Dissociation of H2O on GaP(110): Implications for Electrochemical Reduction of CO2. J. Am. Chem. Soc. 2012, 134, 13600−13603. (14) Keith, J. A.; Carter, E. A. Electrochemical Reactivities of Pyridinium in Solution: Consequences for CO2 Reduction Mechanisms. Chem. Sci. 2013, 4, 1490−1496. (15) Keith, J. A.; Carter, E. A. Theoretical Insights into Electrochemical CO2 Reduction Mechanisms Catalyzed by Surface-Bound Nitrogen Heterocycles. J. Phys. Chem. Lett. 2013, 4, 4058−4063. (16) Keith, J. A.; Muñoz-García, A. B.; Lessio, M.; Carter, E. A. Cluster Models for Studying CO2 Reduction on Semiconductor Photoelectrodes. Top. Catal. 2015, 58, 46−56. (17) Keith, J. A.; Carter, E. A. Correction to “Theoretical Insights into Electrochemical CO2 Reduction Mechanisms Catalyzed by Surface-Bound Nitrogen Heterocycles. J. Phys. Chem. Lett. 2015, 6, 568−568. (18) Lim, C. H.; Holder, A. M.; Hynes, J. T.; Musgrave, C. B. Reduction of CO2 to Methanol Catalyzed by a Biomimetic OrganoHydride Produced from Pyridine. J. Am. Chem. Soc. 2014, 136, 16081− 16095. (19) Tenney, S. A.; Lu, D. Y.; He, F.; Levy, N.; Perera, U. G. E.; Starr, D. E.; Muller, K.; Bluhm, H.; Sutter, P. Key Structure-Property Relationships in CO2 Capture by Supported Alkanolamines. J. Phys. Chem. C 2014, 118, 19252−19258. (20) Lee, B. W.; Ni, R. K.; Masud, N.; Wang, X. R.; Wang, D. C.; Rowe, M. Atomic-Structure of GaP (110) and (111) Faces. J. Vac. Sci. Technol. 1981, 19, 294−300. (21) Hohenberg, P.; Kohn, W. Inhomogeneous Electron Gas. Phys. Rev. 1964, 136, B864. (22) Kohn, W.; Sham, L. J. Self-Consistent Equations Including Exchange and Correlation Effects. Phys. Rev. 1965, 140, A1133. (23) 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, 15−50. (24) Kresse, G.; Furthmuller, J. Efficient Iterative Schemes for Ab Initio Total-Energy Calculations Using a Plane-Wave Basis Set. Phys. Rev. B: Condens. Matter Mater. Phys. 1996, 54, 11169−11186. (25) Kresse, G.; Hafner, J. Ab-Initio Molecular-Dynamics for OpenShell Transition-Metals. Phys. Rev. B: Condens. Matter Mater. Phys. 1993, 48, 13115−13118. (26) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 1996, 77, 3865−3868. (27) Blöchl, P. E. Projector Augmented-Wave Method. Phys. Rev. B: Condens. Matter Mater. Phys. 1994, 50, 17953−17979. (28) Monkhorst, H. J.; Pack, J. D. Special Points for Brillouin-Zone Integrations. Phys. Rev. B 1976, 13, 5188−5192. (29) Kanan, D. K.; Keith, J. A.; Carter, E. A. Water Adsorption on MnO:ZnO(001) - from Single Molecules to Bilayer Coverage. Surf. Sci. 2013, 617, 218−224. (30) Wiberg, E. Die chemische Affinität 2 Aufl.; Verlag de Gruyter: Berlin, 1972. (31) Andersson, K.; Ketteler, G.; Bluhm, H.; Yamamoto, S.; Ogasawara, H.; Pettersson, L. G. M.; Salmeron, M.; Nilsson, A. Bridging the Pressure Gap in Water and Hydroxyl Chemistry on Metal Surfaces: The Cu(110) Case. J. Phys. Chem. C 2007, 111, 14493− 14499. (32) Ketteler, G.; Yamamoto, S.; Bluhm, H.; Andersson, K.; Starr, D. E.; Ogletree, D. F.; Ogasawara, H.; Nilsson, A.; Salmeron, M. The Nature of Water Nucleation Sites on TiO2(110) Surfaces Revealed by Ambient Pressure X-Ray Photoelectron Spectroscopy. J. Phys. Chem. C 2007, 111, 8278−8282. (33) Andersson, K.; Ketteler, G.; Bluhm, H.; Yamamoto, S.; Ogasawara, H.; Pettersson, L. G. M.; Salmeron, M.; Nilsson, A. Autocatalytic Water Dissociation on Cu(110) at near Ambient Conditions. J. Am. Chem. Soc. 2008, 130, 2793−2797.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. A311 EQuad, Princeton, NJ 08544-5263. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This material is based upon work supported by the U.S. Department of Energy Office of Science, Office of Basic Energy Sciences under Award DE-SC0012455. M.L. and E.A.C. acknowledge financial support from the Air Force Office of Scientific Research under AFOSR Award FA9550-10-1-0572. Research was carried out in part at the Center for Functional Nanomaterials and National Synchrotron Light Source, Brookhaven National Laboratory, which is supported by the U.S. Department of Energy, Office of Basic Energy Sciences, under Contract DE-AC02-98CH10886. We thank Professor John Keith, Andrew Ritzmann, and Nima Alidoust for helpful discussions.

■

REFERENCES

(1) Lewis, N. S.; Nocera, D. G. Powering the Planet: Chemical Challenges in Solar Energy Utilization. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 15729−15735. (2) Costentin, C.; Robert, M.; Saveant, J. M. Catalysis of the Electrochemical Reduction of Carbon Dioxide. Chem. Soc. Rev. 2013, 42, 2423−2436. (3) Gagliardi, C. J.; Westlake, B. C.; Kent, C. A.; Paul, J. J.; Papanikolas, J. M.; Meyer, T. J. Integrating Proton Coupled Electron Transfer (PCET) and Excited States. Coord. Chem. Rev. 2010, 254, 2459−2471. (4) Barton, E. E.; Rampulla, D. M.; Bocarsly, A. B. Selective SolarDriven Reduction of CO2 to Methanol Using a Catalyzed p-GaP Based Photoelectrochemical Cell. J. Am. Chem. Soc. 2008, 130, 6342−6344. (5) Cole, E. B.; Lakkaraju, P. S.; Rampulla, D. M.; Morris, A. J.; Abelev, E.; Bocarsly, A. B. Using a One-Electron Shuttle for the Multielectron Reduction of CO2 to Methanol: Kinetic, Mechanistic, and Structural Insights. J. Am. Chem. Soc. 2010, 132, 11539−11551. (6) Seshadri, G.; Lin, C.; Bocarsly, A. B. A New Homogeneous Electrocatalyst for the Reduction of Carbon Dioxide to Methanol at Low Overpotential. J. Electroanal. Chem. 1994, 372, 145−150. (7) Lebegue, E.; Agullo, J.; Morin, M.; Belanger, D. The Role of Surface Hydrogen Atoms in the Electrochemical Reduction of Pyridine and CO2 in Aqueous Electrolyte. ChemElectroChem 2014, 1, 1013− 1017. (8) Yuan, J. L.; Hao, C. J. Solar-Driven Photoelectrochemical Reduction of Carbon Dioxide to Methanol at CuInS2 Thin Film Photocathode. Sol. Energy Mater. Sol. Cells 2013, 108, 170−174. (9) Tossell, J. A. Calculation of the Properties of Molecules in the Pyridine Catalyst System for the Photochemical Conversion of CO2 to Methanol. Comput. Theor. Chem. 2011, 977, 123−127. (10) Keith, J. A.; Carter, E. A. Theoretical Insights into PyridiniumBased Photoelectrocatalytic Reduction of CO2. J. Am. Chem. Soc. 2012, 134, 7580−7583. (11) Lim, C. H.; Holder, A. M.; Musgrave, C. B. Mechanism of Homogeneous Reduction of CO2 by Pyridine: Proton Relay in Aqueous Solvent and Aromatic Stabilization. J. Am. Chem. Soc. 2013, 135, 142−154. 17771

DOI: 10.1021/acs.jpcc.5b05361 J. Phys. Chem. C 2015, 119, 17762−17772

Article

The Journal of Physical Chemistry C (34) Yamamoto, S.; Kendelewicz, T.; Newberg, J. T.; Ketteler, G.; Starr, D. E.; Mysak, E. R.; Andersson, K. J.; Ogasawara, H.; Bluhm, H.; Salmeron, M.; et al. Water Adsorption on Alpha-Fe2O3(0001) at near Ambient Conditions. J. Phys. Chem. C 2010, 114, 2256−2266. (35) Zhang, X. Q.; Ptasinska, S. Dissociative Adsorption of Water on an H2O/GaAs(100) Interface: In Situ near-Ambient Pressure XPS Studies. J. Phys. Chem. C 2014, 118, 4259−4266. (36) Zhang, X. Q.; Ptasinska, S. Distinct and Dramatic Water Dissociation on GaP(111) Tracked by near-Ambient Pressure X-ray Photoelectron Spectroscopy. Phys. Chem. Chem. Phys. 2015, 17, 3909− 3918. (37) Epp, J. M.; Dillard, J. G. Effect of Ion Bombardment on the Chemical Reactivity of Gallium Arsenide(100). Chem. Mater. 1989, 1, 325−330. (38) Hung, W. H.; Chen, H. C.; Chang, C. C.; Hsieh, J. T.; Hwang, H. L. Adsorption and Decomposition of H2S on InP(100). J. Phys. Chem. B 1999, 103, 3663−3668. (39) Wang, G. C.; Tao, S. X.; Bu, X. H. A Systematic Theoretical Study of Water Dissociation on Clean and Oxygen-Preadsorbed Transition Metals. J. Catal. 2006, 244, 10−16. (40) Zhang, X. Q.; Lamere, E.; Liu, X.; Furdyna, J. K.; Ptasinska, S. Interface Chemistry of H2Oon GaAs Nanowires Probed by near Ambient Pressure X-Ray Photoelectron Spectroscopy. Chem. Phys. Lett. 2014, 605, 51−55. (41) May, M. M.; Lewerenz, H. J.; Hannappel, T. Optical in Situ Study of InP(100) Surface Chemistry: Dissociative Adsorption of Water and Oxygen. J. Phys. Chem. C 2014, 118, 19032−19041. (42) Shen, X. A.; Small, Y. A.; Wang, J.; Allen, P. B.; Fernandez-Serra, M. V.; Hybertsen, M. S.; Muckerman, J. T. Photocatalytic Water Oxidation at the GaN(1010̅ )-Water Interface. J. Phys. Chem. C 2010, 114, 13695−13704. (43) Jeon, S.; Kim, H.; Goddard, W. A.; Atwater, H. A. DFT Study of Water Adsorption and Decomposition on a Ga-Rich GaP(001)(2 × 4) Surface. J. Phys. Chem. C 2012, 116, 17604−17612. (44) Williams, K. S.; Lenhart, J. L.; Andzelm, J. W.; Bandara, S. V.; Baril, N. F.; Henry, N. C.; Tidrow, M. Z. First Principles Investigation of Water Adsorption and Charge Transfer on III-V(110) Semiconductor Surfaces. Surf. Sci. 2014, 622, 71−82. (45) Wood, B. C.; Schwegler, E.; Choi, W. I.; Ogitsu, T. HydrogenBond Dynamics of Water at the Interface with InP/GaP(001) and the Implications for Photoelectrochemistry. J. Am. Chem. Soc. 2013, 135, 15774−15783. (46) Limmer, D. T.; Willard, A. P.; Madden, P.; Chandler, D. Hydration of Metal Surfaces Can Be Dynamically Heterogeneous and Hydrophobic. Proc. Natl. Acad. Sci. U. S. A. 2013, 110, 4200−4205. (47) Mizokawa, Y.; Komoda, O.; Iwasaki, H.; Shen, D. H.; Nakamura, S. Interpretation of Apparent Chemical Shifts in XP Spectra from Oxide-GaP Interfaces. J. Electron Spectrosc. Relat. Phenom. 1983, 31, 335−354. (48) Pulver, D.; Wilmsen, C. W.; Niles, D.; Kee, R. Thermal Oxides of In0.5Ga0.5P and In0.5Al0.5P. J. Vac. Sci. Technol., B: Microelectron. Process. Phenom. 2001, 19, 207−214. (49) Katari, J. E. B.; Colvin, V. L.; Alivisatos, A. P. X-Ray Photoelectron-Spectroscopy of CdSe Nanocrystals with Applications to Studies of the Nanocrystal Surface. J. Phys. Chem. 1994, 98, 4109− 4117. (50) Ren, D.; Deng, Y.; Handoko, A. D.; Chen, C. S.; Malkhandi, S.; Yeo, B. S. Selective Electrochemical Reduction of Carbon Dioxide to Ethylene and Ethanol on Copper(I) Oxide Catalyst. ACS Catal. 2015, 5, 2814−2821. (51) Aurianblajeni, B.; Halmann, M.; Manassen, J. Electrochemical Measurements on the Photo-Electrochemical Reduction of Aqueous Carbon-Dioxide on p-Gallium Phosphide and p-Gallium Arsenide Semiconductor Electrodes. Sol. Energy Mater. 1983, 8, 425−440. (52) Flaisher, H.; Tenne, R.; Halmann, M. Photoelectrochemical Reduction of Carbon Dioxide in Aqueous Solutions on p-GaP Electrodes: An AC Impedance Study with Phase-Sensitive Detection. J. Electroanal. Chem. 1996, 402, 97−105.

(53) Noda, H.; Yamamoto, A.; Ikeda, S.; Maeda, M.; Ito, K. Influence of Light-Intensity on Photoelectroreduction of CO2 at a p-GaP Photocathode. Chem. Lett. 1990, 1757−1760. (54) Halmann, M. Photoelectrochemical Reduction of Aqueous Carbon-Dioxide on p-Type Gallium-Phosphide in Liquid Junction Solar-Cells. Nature 1978, 275, 115−116. (55) Ikeda, S.; Saito, Y.; Yoshida, M.; Noda, H.; Maeda, M.; Ito, K. Photoelectrochemical Reduction Products of Carbon-Dioxide at Metal Coated p-GaP Photocathodes in Non-Aqueous Electrolytes. J. Electroanal. Chem. Interfacial Electrochem. 1989, 260, 335−345. (56) Amatore, C.; Saveant, J. M. Mechanism and Kinetic Characteristics of the Electrochemical Reduction of Carbon-Dioxide in Media of Low Proton Availability. J. Am. Chem. Soc. 1981, 103, 5021−5023. (57) Benson, E. E.; Kubiak, C. P.; Sathrum, A. J.; Smieja, J. M. Electrocatalytic and Homogeneous Approaches to Conversion of CO2 to Liquid Fuels. Chem. Soc. Rev. 2009, 38, 89−99. (58) Frese, K. W.; Canfield, D. Reduction of CO2 on n-GaAs Electrodes and Selective Methanol Synthesis. J. Electrochem. Soc. 1984, 131, 2518−2522. (59) Skulason, E.; Tripkovic, V.; Bjorketun, M. E.; Gudmundsdottir, S.; Karlberg, G.; Rossmeisl, J.; Bligaard, T.; Jonsson, H.; Nørskov, J. K. Modeling the Electrochemical Hydrogen Oxidation and Evolution Reactions on the Basis of Density Functional Theory Calculations. J. Phys. Chem. C 2010, 114, 18182−18197. (60) Zhou, X. L.; White, J. M.; Koel, B. E. Chemisorption of AtomicHydrogen on Clean and Cl-Covered Ag(111). Surf. Sci. 1989, 218, 201−210. (61) Christmann, K.; Ertl, G.; Pignet, T. Adsorption of Hydrogen on a Pt(111) Surface. Surf. Sci. 1976, 54, 365−392. (62) Berlowitz, P. J.; Goodman, D. W. The Adsorption of H2 and CO on Strained Ni Overlayers on W(110) and W(100). Surf. Sci. 1987, 187, 463−480. (63) Costentin, C.; Canales, J. C.; Haddou, B.; Saveant, J. M. Electrochemistry of Acids on Platinum. Application to the Reduction of Carbon Dioxide in the Presence of Pyridinium Ion in Water. J. Am. Chem. Soc. 2013, 135, 17671−17674. (64) Butler, M. A.; Ginley, D. S. p-Type GaP as a Semiconducting Photoelectrode. J. Electrochem. Soc. 1980, 127, 1273−1278. (65) Brown, E. S.; Peczonczyk, S. L.; Wang, Z. J.; Maldonado, S. Photoelectrochemical Properties of CH 3 -Terminated p-Type GaP(111)A. J. Phys. Chem. C 2014, 118, 11593−11600.

17772

DOI: 10.1021/acs.jpcc.5b05361 J. Phys. Chem. C 2015, 119, 17762−17772