ARTIFICIAL PHOTOSYNTHESIS: MEMBRANE SUPPORTED ASSEMBLIES THAT USE SUNLIGHT TO SPLIT WATER.

Investigators

Nathan S. Lewis, Professor, Professor, Chemistry, Bruce S. Brunschwig, Member of the Beckman Institute Shannon Boettcher, Postdoctoral Scholar, Chemistry Michael Walter, Postdoctoral Scholar, Chemistry James Maiolo, Graduate Research Assistant, Chemistry Elizabeth Santori, Graduate Research Assistant, Chemistry Josh Spurgeon, Graduate Research Assistant, Chemistry Emily Warren, Graduate Research Assistant, Chemistry Harry A. Atwater, Professor, Applied Physics, Chemistry Andrew J. Leenheer, Graduate Research Assistant, Materials Science Seokmin Jeon, Graduate Research Assistant, Chemistry, Imogen Pryce, Graduate Research Assistant, Chemical Engineering Eyal Feigenbaum, Postdoctoral Scholar, Applied Physics Harry B. Gray, Professor, Chemistry Kyle M. Lancaster, Graduate Research Assistant, Chemistry Bert Lai, Graduate Research Assistant, Chemistry Keiko Yokoyama, Graduate Research Assistant, Chemistry A. Katrine Museth, Graduate Research Assistant, Chemistry Bryan D. Stubbert, Postdoctoral Scholar, Chemistry

Abstract

We are developing a photoelectrochemical system that uses sunlight to drive the splitting of water into H2 and O2. Our approach is to build a device that physically separates the reduction and oxidation reactions onto opposite sides of a flexible photoelectrochemical membrane. We are developing three distinct primary components, the photoanode, the photocathode, and the product-separating electrical/ion-conducting membrane. Each can be optimized separately prior to assembly into a complete water-splitting system, allowing the replacement of individual components by more efficient ones as the research progresses. This design incorporates two photosensitive semiconductor/liquid junctions that will collectively generate the 1.7-1.9 V at open circuit necessary to support both the oxidation of H2O (or OH-) and the reduction of H+ (or H2O). The membrane will consist of two semiconductor rod array structures connected back-to-back with ohmic contacts. Taken together the semiconductor (SC) arrays will straddle the hydrogen (p-type SC) and the oxygen (n-type SC) evolution potentials. The development of highly ordered arrays of micron or nano-sized crystalline Si rods grown using inexpensive starting materials has provided new platforms for studying the hydrogen evolving reaction at the photocathode. Motivation for studying microrod geometries includes orthogonalization of the directions of light absorption and charge carrier collection, which lowers the material purity requirements, and increases

the surface area for water reduction. Research has examined Si arrays as the photocathode in the photoelectrochemical cell. The band-edge potential of p-type silicon photocathodes is sufficiently negative to reduce water/protons to dihydrogen; however kinetic limitations prevent efficient hydrogen production on bare silicon. Deposition of metal catalysts (Pt) has been used to accelerate the hydrogen evolving reaction in such systems. Our work has also focused on the development of new semiconducting materials to serve as photoabsorbers, fabrication of highly structured arrays of radial junction semiconductors, attachment of catalysts, and development of methods to electrically connect the two photoelectrodes and the desired membrane assembly.

Introduction

The development of high aspect ratio semiconductor structures allows materials with low minority-carrier diffusion lengths to be used in solar energy conversion devices. In a radial p-n junction, light is absorbed along the axis of the structure, while carriers are collected in the much shorter radial direction. This design thus relaxes the purity constraints on the semiconductor material, making inexpensive materials viable for use in high efficiency solar cells. To date little work has been done toward the use of these materials for photoelectrochemical systems. The photoelectrolysis of water is a system of interest, because it can be used to produce a chemical fuel, H2.

Background

The materials and energetic criteria for efficient water splitting include the ability to generate > 1.23 V of potential to drive water splitting, band-edge positions that straddle the H2O reduction and H2O oxidation redox potentials, band gaps that allow absorption of the solar spectrum, charge transfer rates sufficient to support the solar flux incident on the assembly, materials that are stable in aqueous oxidizing and reducing environments, and a membrane that supports the mechanical stability and allows electrical contact between the semiconductor arrays as well as providing for ionic mobility through the membrane. The stability, energetics, kinetics, or systems comprised of earth-abundant materials are being actively pursued.

The proposed device structure is designed as a “photoelectrochemical” diode that uses two semiconductor rod array structures connected back-to-back via an ohmic contact. Together, the two semiconductors must provide electrons (p-type SC) and holes (n-type SC) with the potential to reduce water to hydrogen and and oxidize water to oxygen, respectively. One recent report of similar structure uses tungsten oxide as the photoanode and amorphous Si as the photocathode connected back-to-back in a similar manner.[1] This device could photoelectrochemically split water, however a bias was required to the drive the oxygen/hydrogen evolving reactions. We are investigating the modification of a number of metal oxide semiconductors such as WO3, TiO2, and ZnO for use as photoanodes in order to shift their band gaps into the visible portion of the spectrum. Further our research is confined to earth abundant materials, to allow production of a system that is scalable with respect global energy demand.

Results

1. Membrane-Bound Semiconductor Nanorod Arrays for Photoelectrochemical Water Splitting

Our work on the photoreduction of water has focused on the development of p-type Si photocathodes. Silicon is an abundant and well characterized material that has appropriate band energetics for this reaction, and under illumination is cathodically protected from corrosion. The kinetics of hydrogen evolution are slow on unmodified p-Si, but the addition of a catalyst on the Si surface allows H2 to be produced at low overpotentials. To date, work has focused on the use of metal nanoparticle catalysts that have high exchange current densities for the hydrogen evolution reaction (e.g. Pt).

Fabrication of p-Si Wire Arrays: The vapor-liquid-solid (VLS) mechanism provides a robust technique to create highly ordered arrays of p-Si microwires. These wires have been doped p-type by the introduction of BCl3 gas during the VLS growth process. Measurements on individual rods have shown that doping density can be controlled by varying the flow rate of BCl3 [2]. The material properties of the wires have been improved by the use of Cu (instead of Au) as the VLS growth catalyst, because Cu has a much higher degradation threshold in Si photovoltaic devices [0].

Figure 1: Schematic of fabrication of VLS p-Si wire arrays

Significant progress has been made towards commercially viable processes for the fabrication of semiconductor microwire arrays embedded in flexible polymer layers. Specifically, arrays of Si rods have been embedded in polydimethylsiloxane (PDMS), an inexpensive material that is highly transparent to visible light. The array is then removed from the rigid crystalline Si growth substrate, resulting in a composite material that merges the benefits of single-crystalline silicon with the flexibility of a polymer (Figure 2A). With this technique, inorganic solar cell absorber materials with the potential to achieve high efficiency can be prepared by high-temperature processing, and then transformed into a flexible, processable form. Addition to the PDMS mix of a low boiling-point silicone monomer that evaporates during the curing stage results in thin films in which with the majority of the Si wires protrude from the polymer.

Furthermore, to reduce the expense of the single crystal Si wafer required to grow well-defined, highly uniform wire arrays, we have developed a technique to reuse the wafer as a template for additional growth after the array is peeled off in PDMS. The combination of the polymer-embedding and wire re-growth techniques provides a processing scheme in which inorganic semiconductor microwire arrays can be readily fabricated and incorporated into durable organic layers that will be integrated into the proposed water splitting membrane.

Testing of electronic properties of p-Si Wire arrays: The electronic properties of the p-doped Si rod arrays have been studied in an aqueous electrochemical cell with the methyl viologen (MV)2+/+ redox couple. Cells have yielded with good open-circuit voltages, Voc, (0.36 V) and high fill factors (~0.6). The efficiencies are ~ 2.6%, but cells with efficiencies approaching 10% efficiency should be obtainable by increasing the rod length and/or increasing the filling fraction in order to absorb all of the incident radiation.

We have also demonstrated reasonable efficiencies using polymer-embedded wire arrays. The backs of the peeled arrays were electrically contacted using Au, and the films were then tested in the same electrochemical cell described above. Under 45 mW/cm2 of 808 nm illumination, overall efficiencies of 1.8% percent have been achieved, with a Voc

Figure 2: Silicon Carpet. (A) Cross-sectional SEM image of a Si microwire array embedded in PDMS and peeled from the growth substrate. (B) Optical image demonstrating extreme flexibility of the silicon microwire array.

of 384 mV, indicating that these membrane devices retain their electronic performance when removed from the substrate.

Photoelectrochemical H2 Production: Hydrogen production from VLS wire arrays has been realized by decorating the surface of the wires with platinum nanoparticles. To date, we have had the most success in achieving a homogeneous catalyst distribution through the use of a galvanic displacement reaction from metal salts in dilute HF solutions (Figure 3A). This method produces sub-100 nm diameter particles, which should minimize the effect that the low barrier height between p-Si and Pt has on the interfacial energetics [0]. These Pt-modified silicon rod arrays generate hydrogen under solar illumination with VOC ~250 mV and efficiencies approaching 1% (Figure 3C). We are currently working to optimize the silicon band-edge positions relative to the hydrogen redox level, to improve the photovoltage and hence improve this efficiency.

Figure 3: p-Si rod array photoelectrochemical devices. (A) Pt functionalized rods, (B) photoelectrochemical performance with reversible redox couple (the

vertical line at -0.3V indicates the thermodynamic hydrogen reduction potential).

Figure 4 – Electrochemical current (top), bubble radius and calculated amount of gas (bottom) for a video of three bubbles that grew and detached.

2. Optical Microscopy of Water Photoelectrolysis Traditional measurements of gas evolution during water photoelectrolysis rely either on measurement of the overall electrochemical current or capturing macroscopic quantities of gas. However, the current may not fully reflect the photoelectrolysis rate due to parasitic corrosion reactions or chemical reactions in the electrolyte other than water splitting, and neither technique measures the local reaction rate on the surface in a spatially resolved manner. By monitoring the nucleation rate and growth rate of gas bubbles at different areas on the semiconductor surface, insight into the active catalytic sites, and important surface features, can be obtained. We have used optical microscopy to image and quantify gas production via the bubble growth on a model water-splitting semiconductor, strontium titanate. Commercially available single-crystals of (100) oriented SrTiO3 are first annealed under reducing conditions to achieve n-type doping. Electrodes were then fabricated by contacting SrTiO3 with an In/Ga eutectic, Ag paste, and a Cu wire that was then sealed in epoxy. The electrochemical cell was operated in a two-electrode configuration with a small +500 mV bias applied versus a Pt gauze counter electrode in a 1.5 M KOH(aq) electrolyte. Intense above-bandgap radiation from an Ar-ion laser at 351 and 364 nm was focused to a 20 µm diameter spot on the sample surface, so that a single oxygen bubble nucleated at the laser spot, and the bubble growth was recorded on video. A quantitative analysis is shown in Figure 4. The bubble growth video is analyzed frame-by-frame with a “particle analysis” software routine to find the bubble radius as a function of time. Assuming that the bubble is approximately spherical, filled entirely with oxygen at standard temperature and pressure, and obeys the ideal gas equation of state, the number of oxygen molecules in the bubble is calculated based on the radius, and was found to grow linearly with time. Thus a quantitative measure of the reaction rate is possible by watching the bubble growth under local illumination. In this example, the reaction rate measured by watching the bubble is calculated as 1.8x1013 O2 molecules/sec, and the rate calculated from the electrochemical current is approximately 1.1x1013 sec--1, a remarkably close agreement. The discrepancy may be due to effects of increased local temperature or water vapor in the bubble. We have verified that the bubble grows only when the sample is illuminated with above-bandgap radiation, by chopping the laser at about 2 Hz. As

before, the bubble growth rate during illumination matches that expected from the electrochemical current, and without illumination the bubble is static. Additionally, the reaction rate is found to increase with increased laser intensity though the quantum efficiency (transferred electrons/incident photons) decreases as shown in Figure 5, indicating that the surface turnover rate is likely the limiting step. Future experiments will locally pattern catalytic features on the semiconductor surface to measure the comparative bubble growth rate with different amounts or types of catalyst particles or features.

Figure 5. Oxygen production rate and quantum efficiency for varied illumination intensities. 3. Water Molecular Adsorption and Dissociation on GaN

Photocatalytic water splitting reactions using wide bandgap inorganic

semiconductors are one of the candidates for a clean hydrogen gas source. A huge number of semiconductors have been widely studied. However, based on the fact that many of them can not generate H2 gas from water without additional bias voltage, GaN is one of the promising candidates with optimal band-edge energies straddling the hydrogen and oxygen evolution redox potentials. For successful achievement of photocatalytic or photoelectrochemical applications of GaN, improvement of the efficiency and stability through elucidation of reaction mechanisms is important.

To develop molecular-level control and understanding of water splitting reactions in GaN-based photoelectrochemical devices, we performed ab initio Density Functional Theory (DFT) calculations of an H2O/wurtzite-GaN(0001) system. These theoretical investigations demonstrate that a water molecule prefers to adsorb on top of Ga atoms at the surface at sub-monolayer coverage (Figure 6). Furthermore, the molecular water adsorbate thermodynamically prefers to dissociate to form HO- and H- species on top of Ga atoms (Figure 7a and b), which is consistent with previous ultra-violet photoemission spectroscopy (UPS) results [5]. We also performed geometry optimization of an adsorption structure in which a H atom is located at the interstitial position of the sub-surface Ga-N bond forming HO-Ga and H-N bonds (Figure 7c). Our simulation results demonstrated that the insertion of a H species produces a 1.703 eV

higher-energy product, implying that this reaction is energetically unfavorable. Based on Bermudez’s argument, we then optimized the adsorption geometry that described the complete dissociation of molecular water to form single atomic adsorbates such as H-Ga, O=Ga (or Ga-O-Ga), and H-Ga; that is, we evaluated the system assuming that the surface HO species further dissociates to create an atomic O adsorbate species at 200o C [1]. However, the adsorption energy in our calculations did not indicate that the further dehydrogenation step was thermodynamically favorable. We suspect that the unstable adsorption structure results from lateral interactions between Ga-O-Ga and H-Ga adsorbates, due to the small supercell size in the simulations.

Although our simulation results re reasonably consistent with previous experimental results, the simulations require more development. In particular, future calculations should account for surface reconstruction. The Ga-polar GaN(0001) surface, which has been observed in experiments, is known to have four surface reconstructions, including (2×2), (5×5), (6×4), and “1×1.” Secondly we need to consider transition state structures and need to find energy profiles along each reaction pathway.

Currently, we are gearing up to study water dissociation on GaN experimentally by studying surface reactions on GaN via scanning tunneling microscopy imaging of the surface adsorbate structures. Results such as those described above obtained using density function theory simulations will be compared to experimental STM images, to specify the site for water dissocation on GaN.

Figure 6. Initial structures; a water molecule located at a) atop, b) bridge, c) fcc, or d) hcp sites of the GaN(0001) surface unit cell. Structure e), f), g), and h) are the final structures after geometry optimizations of a), b), c), and d), respectively. The adsorption energies of each final structure are -1.220, -1.202, -1.210, -1.211 eV, respectively.

a) b) c)

Figure 7. Structures after geometry optimization when HO and H species were initially located on a) neighboring atop sites, b) bridge and atop sites, and c) atop and Ga-N bond interstitial sites. Structures after geometry optimization when H, O, and H were initially located on d) atop, bridge, and atop sites, respectively, and e) three neighboring atop sites.

d) e)

4. Protein Electron Transfer and Molecular Catalysis – Fundamental Studies En Route to Carbon Mitigation Technologies

Nature has developed the machinery to carry out the two most important chemical reactions on our planet: photosynthesis, the solar powered oxidation of water, and respiration, the reduction of oxygen to water. We have embarked on a decades long exploration of the chemistry that governs these vital processes, focusing on the transfer of electrons in protein and model environments and leading to greater insights into the clean generation and consumption of carbon free fuels.[5] Our GCEP funded projects currently focus on site-directed mutagenesis of enzyme active sites that markedly alter redox properties, extremely long range electron transfer across membrane proteins as a model for photosynthesis, and mechanistic studies of molecular electrocatalysts for the reduction of carbon dioxide to syngas components.

Background Redox Tuning of Protein Active Sites The capacity for precise tuning of active site reduction potentials in folded polypeptide environments has afforded nature the freedom to access myriad chemistries despite the limitations of elemental bioavailability.[6] The reduction potential of Cu(II) in Pseudomonas aeruginosa azurin is sensitive to substitutions within both the inner and outer type 1 copper coordination spheres.[7] This sensitivity makes it possible to tune the potential over 0.5 V higher than the aqueous redox couple.[7e] We seek to demonstrate the feasibility of redox tuning at enzyme active sites as a proof of concept for extension of this approach systems with the machinery needed for catalytic water oxidation.

Membrane Electron Transfer Photosynthesis begins with electron transfer (ET) across the thylakoid membrane, mediated by various aromatic amino acids along the ET pathway through an electron hopping mechanism. Considerable time and effort have been devoted to understanding ET in metal-modified metalloproteins, leading to the construction of a tunneling timetable (Figure 8) that establishes a millisecond to hundred-microsecond range for single-step electron transfer over 20 Å distances, well below the time required to support the function of many natural redox systems. Multi-step electron transfer has since been proposed to explain the short timescales of long range electron transfer.[9]. We have recently demonstrated multi-step electron tunneling through tryptophan (Trp) on an engineered system utilizing Pseudomonas aeruginosa azurin mutants. The electron traverses a Re-Cu distance of 19.4 Å on the nanosecond timescale, two orders of magnitude faster than predicted from the timetable in Figure 8.

To continue this work, we are examining multi-step electron tunneling across a biological membrane whose function can mimic that of the more complicated redox processes in photosynthesis, namely outer membrane protein A (OmpA). Ruthenium and rhenium complexes will be installed into this metal-free protein to model the origin and destination of ET. This event will begin with the electron donor folded inside a membrane mimic, small unilamellar vesicle (SUV). With the ET acceptor folded outside the SUV, Trp residues will be probed as intermediates along the hopping pathway. Time-resolved laser spectroscopy will be employed to monitor the electron transfer kinetics along this enormous 4 nm (40 Å) distance.

Molecular CO2 Reduction Catalysts Among our efforts to understand and ultimately improve catalytic water splitting and other transformations related to clean solar-generated fuels for a sustainable energy economy, we have used electrochemical and photochemical methods to better understand the performance of an efficient, highly-selective nickel macrocycle that electrocatalytically reduces CO2 to CO and water, and as a complement to these areas, investigate the mechanistic details of catalytic H2 production mediated by cobalt complexes in water. We expect these efforts to lead to improved design standards for electrocatalysts that can be covalently linked and/or transformed into photovoltaics and other light-absorbing materials for solar-generated fuels. The parent system of interest to our group has been [Ni(cyclam)]2+, which was shown by Sauvage and co-workers[10]to operate as an electrocatalyst for CO2 reduction over a range of conditions in mildly acidic aqueous media with remarkable selection for reaction with CO2 (> 106 vs. H2O or H+) at a reactive, electrode generated Ni1+ species. Coupled with a solar powered water splitter as a carbon free H2 source, this scenario can be envisioned as an environmentally friendly means for recycling CO2 via well-established syngas chemistry.

Results Redox Tuning of Enzyme Active Sites Exploration of the precise tuning of active site reduction potentials in folded polypeptide environments has continued by demonstrating the feasibility of active site tuning towards more positive redox potentials amenable to catalytic water oxidation and

Figure 8. Timetable for single-step electron tunneling in Ru-modified proteins: azurin (●); cyt c (○); myoglobin (Δ); cyt b562 (� ); HiPIP (◊); and Fe:Zn-cyt c crystals (� ).

sufficient stability to withstand the strongly oxidizing conditions necessary for O2 production at a biological site.

The reduction potential of Cu(II) in Pseudomonas aeruginosa azurin is sensitive to substitutions within the inner and outer type 1 copper coordination spheres, [8] and this sensitivity makes it possible to tune the potential > 0.5 V more positive than the aqueous redox couple. [8e] Robust azurins with relatively high reduction potentials in the absence of soft ligands have now been prepared for the C112D protein,[11] where the cysteine (C = Cys) to aspartate (D = Asp) substitution is made to avoid the unwanted irreversible oxidation of the Cys thiol group.[12-14] Mutants possessing charged axial ligands were generated, as well as a mutant lacking axial coordination from position 121 to evaluate the potential range of these robust scaffolds.

Electronic absorption spectra of the Cu(II) and Co(II) C112D/M121X (X = methionine (M = Met), glutamate (E = Glu), histidine (H = His), leucine (L = Leu)) azurins in pH 7 aqueous solution are shown in Table I. The M121L substitution results in a red shift of the Cu(II) ligand field (LF) absorption to 800 nm; consistent with this decrease in LF splitting, there is also a small red shift of the imidazole to Cu(II) ligand-to-metal charge-transfer (LMCT) band. The Cu(II) LF band of C112D/M121L Cu(II) azurin is much sharper than that of C112D. This could indicate decreased site reorganization in the case of a single transition comprising the d-d band, though further characterization is ongoing to dissect this spectroscopic feature. The Co(II) C112D/M121L LF system near 600 nm is virtually identical with that of the single mutant, confirming that the inner-sphere electronic interactions involve mainly the equatorial ligands.

The coordination geometries that can be inferred from examination of the Co(II) spectra of C112D/ M121X (X = M, E, H) are supported by 77 K X-band EPR data of Cu(II) analogues.[15] The C112D/M121E and C112D/M121H proteins possess axial hyperfine splittings comparable to that of the single mutant. Concomitant with the lack of significant anisotropy in equatorial g-tensors, these data suggest similar axial coordination environment to that of the C112D mutant in which the X121 side-chain weakly coordinates the metal. The C112D/M121L EPR spectrum is particularly interesting. The spin Hamiltonian parameters differ from the other azurins, displaying a substantial increase in gz and decrease in A||, consistent with a shift toward tetrahedral site geometry.[16] Furthermore, there is considerable g-tensor anisotropy attributable to dz2 mixing into a dxy ground state.[17,18] Together these features suggest that C112D/M121L Cu(II) is in a unique electronic environment.

The electrochemical properties of Cu(II) C112D/M121X (X = M, L) azurins (Table I) were investigated via redox titration (with P. aeruginosa cytochrome c551) or, where possible, adsorbed onto [CH3(CH2)8SH and HO(CH2)8SH] mixed self-assembled-monolayer (SAM) modified gold electrodes.[19,20] The C112D/M121L and C112D/M121E Cu(II/I) midpoint potentials are more positive than C112D, likely a result of weaker overall ligand fields. From the position of the LF band in the C112D/M121E spectrum, we would have predicted an even larger upshift were it not for the negative

charge of the carboxylate. Earlier work has shown that a carboxylate interaction with Cu(II) in azurin leads to a ~100 mV decrease in reduction potential.[8e]

Our work to date on C112D/M121L azurin demonstrates that type 2 copper can acquire type 1 character with tetrahedral coordination geometry at the metal center in the absence of sulfur ligation. Properties such as a relatively small A|| value and high reduction potentials (relative to C112D, and by extension, Cu(II) aquo ion) until now have been associated almost exclusively with blue copper centers.[8e,8h,21,22] We suggest that electron donation from the equatorial ligands will reduce the positive charge on Cu(II) in the hydrophobic axial environment, effectively mimicking the covalency attributable mainly to Cys thiolate ligation in a blue protein.

Membrane Electron Transfer Initial studies have focused on whether Ru- and Re-labeled OmpA will fold into the SUV. We have expressed a mutant containing a single Cys mutant, where the ET donor/acceptor label would be folded in the SUV. Preliminary circular dichoism and steady-state fluorescence spectroscopy results are consistent with successful folding of

Table I: Electronic absorption, EPR, and electrochemical data for Cu(II) azurins. a Parenthetical values represent molar extinction coefficient (M-1cm-1) calculated by titration of apoazurin with CuSO4 (± 5%). b EPR parameters simulated using SpinCount. Parenthetical values represent g or A strain uncertainty (ref Error! Bookmark not defined.). c mV vs NHE, pH 7.

X Cu(II) λmax (nm)

Co(II) λmax (nm) gx gy gz

A⊥ (10-4 cm-1)

A|| (10-4 cm-1)

Cu(II/I) Eº1/2

Met 754(100)

518 (210), 560 (200), 610 (280),

630 (650, sh)

2.07(1) 2.05(1) 2.31(1) 1.58(4) 151(1) 180

Glu 875 (70) 569 (270), 531 (330), 600 (330)

2.06(3) 2.070(1) 2.32(3) 2.15(1) 151(1) 270

His 657 (70)

549 (210, sh), 577 (240), 608 (280), 630 (170)

2.06(2) 2.06(5) 2.30(3) 4(2) 165(1) 305

Leu 798(100)

519 (170), 559 (170), 609 (220),

628 (170, sh)

2.11(1) 2.05(1) 2.38(1) 15.07(3) 101(1) 280

labeled OmpA into the SUV. The plasmid for the full transmembrane ET mutant has been constructed, and the molecular biology required for labeling studies is complete. The protein expression system and the ruthenium labeling protocol have also been optimized.

Molecular CO2 Reduction Catalysts Our goals in this project have included determining the number of N–H moieties

required on the cyclam ligand for efficient binding and catalysis, understanding the lifetime and binding strength of transient Ni(CO)n(cyclam) species formed, and building bimetallic complexes that test the limits of catalytic performance in this system. Our work in the past year suggest that removing the N–H moieties in the cyclam ring and adding peripheral H–bond donors (Figure 9) for putative Ni–C(O)(O…H) interactions[23] does not seem to greatly inhibit CO2 binding. However, the dozens of complexes examined for these purposes do not efficiently reduce CO2 electrocatalytically. The emergent trend is consistent with at least one N–H moiety on the parent cyclam ring (close to the Ni–C(O)2 center) being required for electrocatalytic reduction of carbon dioxide.

Figure 9. Representative examples of peripheral H–bond donors without core N–H bonds. These complexes do not appear to reduce CO2 at the Ni(cyclam) center.

Since the 2008 annual meeting, this work has been extended to include bimetallic Ni2 complexes. Unfortunately, this endeavor has offered little improvement in electrocatalytic performance (Figure 10), suggesting that this research direction will not lead to significant improvements in electrocatalysis. We were optimistic that a solid foundation for the salient features underlying the remarkable selectivity of Ni(cyclam) could ultimately be developed and translated into synthetically tunable catalysts operating at lower overpotential with improved activity and without significant loss of selectivity. However, our cumulative results suggest that a new direction in electrocatalytic CO2 reduction chemistry is needed.

Advances in mechanistic studies of nickel cyclam derivatives that probe various aspects of CO2 binding and activation have been modest. We have addressed many of our goals in this project, including the role and necessity of H–bond donor/acceptpr interactions (directly on the cyclam macrocycle), and building bimetallic complexes that test the limits of catalytic performance in this system. Our results with dozens of complexes suggest that this approach will not lead to further clarification of mechanistic pathways or significant improvements in electrocatalysis and that a new direction for electrocatalytic CO2 reduction chemistry is needed.

Figure 10. Cyclic voltammograms showing the electrocatalytic response of an ethylene-bridged [Ni2(N-CH2-cyclam)2]4+ complex in the presence (blue) and absence (red) of CO2 (2 mM [Ni2], pH 7 (NaPi), 0.1 M NaTFA). This response occurs at slightly less negative potential than monometallic Ni(cyclam) but with diminished activity for CO2 reduction and considerably enhanced (undesired) H2O reduction.

Future Plans Redox Tuning of Enzyme Active Sites Experiments in understanding these unique coordination environments at biological copper sites are ongoing. Specifically, preliminary 1H NMR spectroscopic studies are promising as to further elucidation of the structural composition of these tetrahedral metal binding sites in Cu(II) and Co(II) azurin. We expect continued progress in understanding the effects of rational site-directed mutagenesis in these “Type 0” copper proteins with regard to electronic structure, redox properties, and, ultimately, ways in which these results may be applied to other ongoing interests in water oxidation and biological fuel cell design.

Membrane Electron Transfer The next phase of the project will involve site-specific labeling of OmpA cysteine residues with ruthenium and rhenium complexes. After the full transmembrane electron transfer mutant has been expressed and selectively labeled, time-resolved laser spectroscopy will be employed to investigate whether tryptophan sites can indeed be mediate multi-step electron transfer.

Molecular CO2 Reduction Catalysts Our cumulative results in this area suggest that a new direction in electrocatalytic

CO2 reduction chemistry is needed to move this important problem closer to realistic solutions. We will approach CO2 reduction from a different direction, namely, to activate CO2 at a metal center in an η1-OCO end-on fashion that leaves the central carbon atom susceptible to reaction with nucleophiles (e.g., a nearby metal hydride species). This binding mode is rare and therefore presents many challenges for the upcoming funding period. We also plan to include mechanistic investigation of similar electrocatalysts that effectively mediate H2 evolution from water in our future GCEP efforts, with an end goal

16

14

12

10

8

6

4

2

0

x10-6

Curr

ent (A

)

-1.4-1.2-1.0-0.8-0.6-0.4-0.20.0

Potential (volts) vs. SCE

[Ni(cyclam)- N-CH2CH2-N-(cyclam)Ni]

4+

2 mM, Argon atmosphere 2 mM, saturated with CO

2

of lowering the overpotential for H2O reduction to a range amenable to solar powered H2 generation.

Conclusions

If successful, this work would enable the use of the most abundant energy resource, the Sun, to directly produce fuel. This approach would thus enable mitigation of the drawback of solar energy systems to date, which do not deal with the intermittancy of the resource in a satisfactory fashion. This system would thus provide an integrated approach to the capture, conversion and storage of sunlight, providing a globally scalable approach to renewable energy generation.

Publications

1. James R. Maiolo III, Brendan M. Kayes, Michael A. Filler, Morgan C. Putnam, Michael D. Kelzenberg, Harry A. Atwater, Nathan S. Lewis, “High Aspect Ratio Silicon Wire Array Photoelectrochemical Cells”, J. Am. Chem. Soc., 2007, 129(41), 12346.

2. Joshua M. Spurgeon, Harry A. Atwater, and Nathan S. Lewis, “A Comparison Between the Behavior of Nanorod and Planar Cd(Se, Te) Photoelectrodes”, J. Phys. Chem. C, 2008, 112, 6186.

3. James R. Maiolo III, Harry A. Atwater, and Nathan S. Lewis, “Macroporous Silicon as a Model for Silicon Wire Array Solar Cells”, J. Phys. Chem., C, 2008, 112(15), 6194.

4. Joshua M. Spurgeon, Katherine E. Plass, Brendan M. Kayes, Bruce S. Brunschwig, Harry A. Atwater, and Nathan S. Lewis, “Repeated Epitaxial Growth and Transfer of Arrays of Patterned, Vertically Aligned, Crystalline Si Wires from a Single Si(111) Substrate”, Appl. Phys. Lett., 2008, 93, 032112.

5. Katherine E. Plass, Michael A. Filler, Joshua M. Spurgeon, Brendan M. Kayes, Stephen Maldonado, Bruce S. Brunschwig, Harry A. Atwater, and Nathan S. Lewis, “Flexible Polymer-Embedded Si Wire Arrays”, Adv. Mater., 2008, 9999, http://dx.doi.org/10.1002/adma.200802006.

6. Jordan E. Katz, Todd R. Gingrich, Elizabeth A. Santori, and Nathan S. Lewis, “Combinatorial Synthesis and High-Throughput Photopotential and Photocurrent Screening of Mixed-Metal Oxides for Photoelectrochemical Water Splitting”, Energy Environ. Sci., 2009, 2, 103.

7. Lancaster, K. M.; Yokoyama, K.; Richards, J. H.; Winkler, J. R.; Gray, H. B. High-Potential C112D/M121X (X = M, E, H, L) Pseudomonas aeruginosa Azurins Inorg. Chem. 2009, 48(4), 1278-1280.

8. Yokoyama, K.; Leigh, B. S.; Sheng, Y.; Niki, K.; Nakamura, N.; Ohno, H.; Winkler, J. R.; Gray, H. B.; Richards, J. H. Electron Tunneling through Pseudomonas aeruginosa Azurins. Inorg. Chim. Acta 2008, 361, 1095.

9. Stubbert, B. D.; Winkler, J. R.; Gray, H. B. Mechanistic Investigations of Aqueous Electrocatalysts for H2 Evolution and CO2 Reduction. Gordon Research Conference on Electron Donor-Acceptor Interactions, Salve Regina University, Newport, RI, August 3-8, 2008.

10. Yokoyama, K.; Lancaster, K. M.; Sheng, Y.; Nakamura, N.; Ohno, H.; Leigh, B. S.; Niki, K.; Winkler, J. R.; Richards, J. H.; Gray, H. B. Mimicking protein-protein electron transfer: Electron

tunneling through mutant azurins on mixed-SAM gold electrodes. Gordon Research Conference on Electron Donor Acceptor Interactions, Salve Regina University, Newport, RI, August 3-8, 2008.

11. Stubbert, Bryan D.; Winkler, Jay R.; Gray, Harry B. Aqueous electrocatalysts for the conversion of solar energy to fuels. Abstracts of Papers, 236th ACS National Meeting, Philadelphia, PA, United States, August 17-21, 2008, INOR-148.

12. Yokoyama, K.; Lancaster, K. M.; Sheng, Y.; Nakamura, N.; Ohno, H.; Leigh, B. S.; Niki, K.; Winkler, J. R.; Richards, J. H.; Gray, H. B. Electron tunneling through mutant azurins on mixed-SAM gold electrodes. 3rd Joint Symposium on Bio-Related Chemistry, Tokyo Institute of Technology, Tokyo, Japan, 9/20/08.

13. Stubbert, B. D.; Brunschwig, B. S.; Winkler, J. R.; Gray, H. B. Investigations into the Mechanistic Details of Inorganic Electrocatalysts for H2 Evolution and CO2 Reduction in Water. Osaka University GCOE Forum 2008 on Bio-Environmental Chemistry, Milton Marks Conference Center, San Francisco, CA, December 8-10, 2008.

14. Stubbert, B. D.; Brunschwig, B. S.; Winkler, J. R.; Gray, H. B. Mechanism According to Medium: Homogeneous Electrocatalysis Relevant to Clean Solar Fuels. Gordon-Kenan Research Seminar on Renewable Energy: Solar Fuels, Ventura Beach Marriott, Ventura, CA, January 31-February 1, 2009.

15. Stubbert, B. D.; Brunschwig, B. S.; Winkler, J. R.; Gray, H. B. Mechanism According to Medium: Homogeneous Electrocatalysis Relevant to Clean Solar Fuels. Gordon Research Conference on Renewable Energy: Solar Fuels, Four Points Sheraton, Ventura, CA, February 1-6, 2009.

16. Yokoyama, K.; Nakamura, N.; Ohno, H.; Leigh, B. S.; Niki, K.; Winkler, J. R.; Richards, J. H.; Gray, H. B. Electron tunneling through Pseudomonas aeruginosa azurins on SAM gold electrodes. Gordon Research Seminar on Bioinorganic Chemistry, Four Points Sheraton, Ventura, CA, USA, January 31, 2008.

17. Stubbert, B. D.; Brunschwig, B. S.; Winkler, J. R.; Gray, H. B. Mechanisitic Interactions of Inorganic Catalysts for the Production of Solar Fuels. Gordon Research Conference on Inorganic Reaction Mechanisms, Galveston, TX, March 8-13, 2009.

18. Stubbert, Bryan D.; Winkler, Jay R.; Gray, Harry B. Inorganic catalysts for the production of solar fuels. Abstracts of Papers, 237th ACS National Meeting, Salt Lake City, UT, United States, March 22-26, 2009, INOR-726 (also selected for Sci-Mix session).

19. Gray, H. B. Powering the Planet with Solar Fuel. Nature Chemistry 2009, 1, 7.

References

1. Miller, E. L.; Marsen, B.; Paluselli, D.; Rocheleau, R., Optimization of hybrid photoelectrodes for solar water-splitting. Electrochemical and Solid State Letters 2005, 8 (5), A247-A249.

2. Putnam, M. C.; Turner-Evans, D.B.; Kelzenberg, M.D.; Boettcher, S.W.; Lewis, N.S., Atwater, H.A., “BCl3 In-Situ Doping of Vapor-Liquid-Solid Grown Si Wires. MRS Conference Proceedings, 2009.Rossi, R.C. and N.S. Lewis, Investigation of the size-scaling behavior of spatially nonuniform barrier height contacts to semiconductor surfaces using ordered nanometer-scale nickel arrays on silicon electrodes. Journal of Physical Chemistry B, 2001. 105(49): p. 12303-12318.

4. Davis, J.R., Rohatgi, A., Hopkins, R.H., Blais, P.D., Raichoudhury, P.; McCormick, J.R., Mollenkopf, H.C.,. Impurities in Silicon Solar Cells. IEEE Transactions on Electronic Device. 27 (4), 677-697, 1980

5. Bermudez, V.M. and J.P. Long, Chemisorption of H2O on GaN(0001). Surface Science, 2000. 450(1-2): p. 98-105.

6. (a) Gray, H. B. Powering the Planet with Solar Fuel. Nature Chemistry 2009, 1, 7. (b) Eisenberg, R.; Gray, H. B. Preface on Making Oxygen. Inorg. Chem. 2008, 47, 1697-1699.

7. (a) Gray, H. B. Biological inorganic chemistry at the beginning of the 21st century. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 3563-3568. (b) Zong, C. H.; Wilson, C. J.; Shen, T. Y.; Wittung-Stafshede, P.; Mayo, S. L.; Wolynes, P. G. Establishing the entatic state in folding metallated Pseudomonas aeruginosa azurin. Proc.Nat. Acad. Sci. U.S.A. 2007, 104, 3159-3164. (c) Malmstrom, B. G.; Wittung-Stafshede, P. Effects of protein folding on metalloprotein redox-active sites: electron-transfer properties of blue and purple copper proteins. Coord. Chem. Rev. 1999, 186, 127-140.

8. (a) Garner, D. K.; Vaughan, M. D.; Hwang, H. J.; Savelieff, M. G.; Berry, S. M.; Honek, J. F.; Lu, Y. Reduction Potential Tuning of the Blue Copper Center in Pseudomonas aeruginosa Azurin by the Axial Methionine as Probed by Unnatural Amino Acids. J. Am. Chem. Soc. 2006, 128, 15608-17. (b) Ralle, M.; Berry, S. M.; Nilges, M. J.; Gieselman, M. D.; van der Donk, W. A.; Lu, Y.; Blackburn, N. J. The Selenocysteine-Substituted Blue Copper Center: Spectroscopic Investigations of Cys112SeCys Pseudomonas aeruginosa Azurin. J. Am. Chem. Soc. 2004, 126, 7244-7256. (c) Berry, S. M.; Gieselman, M. D.; Nilges, M. J.; van Der Donk, W. A.; Lu, Y. An Engineered Azurin Variant Containing a Selenocysteine Copper Ligand. J. Am. Chem. Soc. 2002, 124, 2084-5. (d) Berry, S. M.; Ralle, M.; Low, D. W.; Blackburn, N. J.; Lu, Y. Probing the role of axial methionine in the blue copper center of azurin with unnatural amino acids. J. Am. Chem. Soc. 2003, 125, 8760-8768. (e) Pascher, T.; Karlsson, B. G.; Nordling, M.; Malmstrom, B. G.; Vanngard, T. Reduction potentials and their pH dependence in site-directed-mutant forms of azurin from Pseudomonas aeruginosa. Eur. J. Biochem. 1993, 212, 289-296. (f) Chang, T. K.; Iverson, S. A.; Rodrigues, C. G.; Kiser, C. N.; Lew, A. Y.; Germanas, J. P.; Richards, J. H. Gene synthesis, expression, and mutagenesis of the blue copper proteins azurin and plastocyanin. Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 1325-1329. (g) Canters, G. W.; Gilardi, G. Engineering type 1 copper sites in proteins. FEBS Lett. 1993, 325, 39-48. (h) Gray, H. B.; Malmstrom, B. G.; Williams, R. J. P. Copper Coordination in Blue Proteins. J. Biol. Inorg. Chem. 2000, 5, 551-559. (i) Gray, H. B., Malmstrom, B. G. On the relationship between protein-forced ligand fields and the properties of blue copper centers. Comments Inorg. Chem. 1983, 2, 203-209.

9. Shih, C.; Museth, A. K.; Abrahamsson, M.; Blanco-Rodriguez, A. M.; Di Bilio, A. J.; Sudhamsu, J.; Crane, B. R.; Ronayne, K. L.; Towrie, M.; Vlcek, A.; Richards, J. H.; Winkler, J. R.; Gray, H. B. Tryptophan-accelerated Electron Flow Through Proteins. Science 2008, 320, 1760-1762.

10. Beley, M.; Cellis, J.-P.; Ruppert, R.; Sauvage, J.-P. Electrocatalytic reduction of carbon dioxide by nickel cyclam2+ in water: study of the factors affecting the efficiency and the selectivity of the process. J. Am. Chem. Soc. 1986, 108, 7641.

11. Lancaster, K. M.; Yokoyama, K.; Richards, J. H.; Winkler, J. R.; Gray, H. B. High-Potential C112D/M121X (X = M, E, H, L) Pseudomonas aeruginosa Azurins. Inorg. Chem. 2009, 48(4), 1278-1280.

12. Faham, S.; Mizoguchi, T. J.; Adman, E. T.; Gray, H. B.; Richards, J. H.; Rees, D. C. Role of the active-site cysteine of Pseudomonas aeruginosa azurin. Crystal structure analysis of the CuII (Cys112Asp) protein. J. Biol. Inorg. Chem. 1997, 2, 464-469.

13. Mizoguchi, T. J.; Di Bilio, A. J.; Gray, H. B.; Richards, John H. Blue to type 2 binding. Copper(II) and cobalt(II) derivatives of a Cys112Asp mutant of Pseudomonas aeruginosa azurin. J. Am. Chem. Soc. 1992, 114, 10076-10078.

14. Mizoguchi, T. J. Probing the role of the active-site Cysteine of Azurin by site-directed mutagenesis. Ph.D. Thesis, California Institute of Technology, 1996.

15. Golombek, A. P.; Hendrich, M. P. Quantitative analysis of dinuclear manganese(II) EPR spectra. J. Mag. Res. 2003, 165, 33-48.

16. Bencini, A.; Gatteschi, D.; Zanchini, C. Anomalous spin Hamiltonian parameters of pseudotetrahedral copper(II) complexes. ESR spectra of copper(II)-doped dichlorobis(triphenylphosphine oxide)zinc(II). J. Am. Chem. Soc. 1980, 102, 5234-5237.

17. Gewirth, A. A.; Cohen, S. L.; Schugar, H. J.; Solomon, E. I. Spectroscopic and theoretical studies of the unusual EPR parameters of distorted tetrahedral cupric sites: correlations to X-ray spectral features of core levels. Inorg. Chem. 1987, 26, 1133-1146.

18. Fittipaldi, M.; Steiner, R. A.; Matsushita, M.; Dijkstra, B. W.; Groenen, E. J. J.; Huber, M. Single-crystal EPR study at 95 GHz of the type 2 copper site of the inhibitor-bound quercetin 2,3-dioxygenase. Biophys. J. 2003, 85, 4047-4054.

19. Fujita, K.; Nakamura, N.; Ohno, H.; Leigh, B. S.; Niki, K.; Gray, H. B.; Richards, J. H. Mimicking Protein-Protein Electron Transfer: Voltammetry of Pseudomonas aeruginosa Azurin and the Thermus thermophilus CuA Domain at w-Derivatized Self-Assembled-Monolayer Gold Electrodes. J. Am. Chem. Soc. 2004, 126, 13954-13961.

20. Yokoyama, K.; Leigh, B. S.; Sheng, Y.; Niki, K.; Nakamura, N.; Ohno, H.; Winkler, J. R.; Gray, H. B.; Richards, J. H. Electron Tunneling through Pseudomonas aeruginosa Azurins. Inorg. Chim. Acta 2008, 361, 1095-1099.

21. Solomon, E. I.; Szilagyi, R. K.; George, S. D.; Basumallick, L. Electronic Structures of Metal Sites in Proteins and Models: Contributions to Function in Blue Copper Proteins. Chem. Rev. 2004, 104, 419-458.

22. Solomon, E. I. Spectroscopic Methods in Bioinorganic Chemistry: Blue to Green to Red Copper Sites. Inorg. Chem. 2006, 45, 8012-8025.

23. Fujita, E.; Creutz, C.; Sutin, N. Brunschwig, B. S. Inorg. Chem. 1993, 32, 2657-2662.

Contacts

Nate Lewis: nslewis@caltech.edu Bruce Brunschwig: bsb@caltech.edu

Harry B. Gray: HBGray@caltech.edu Harry A. Atwater: haa@caltech.edu