Chemical approaches to artificial photosynthesis

In the early 1970s, the works by Fujishima and Honda (1) and Honda et al. (2) reported on the results of a now famous experiment. They showed that band gap excitation of anatase TiO₂ in a photoelectrochemical cell with a Pt counter electrode and an applied bias resulted in water splitting into hydrogen and oxygen. The timing of the result was impeccable. In 1973, the Organization of the Petroleum Exporting Countries (OPEC) declared an embargo on oil imports to the West, resulting in gasoline shortages and long lines at gas pumps. Suddenly, there was a pressing need for energy independence and new ways of providing for the energy-hungry economies of Western Europe, Japan, and the United States. The international research community responded. There was a short lived explosion of interest in converting sunlight into high-energy molecules by what we now call artificial photosynthesis to make solar fuels. Target reactions were water splitting into hydrogen and oxygen (1) and light-driven reduction of CO₂ by water to give CO, other oxygenates, or hydrocarbons. Methane is shown as the product in equation 2, but the ultimate target is liquid hydrocarbons to power our existing energy infrastructure (1 and 2):

The oil crisis eased in the late 1970s, and with it, the surge in research and the interest in new energy technologies. Time has passed; the world is a different place. Significant increases in oil prices and demand, fueled by increased consumption and growth in the rapidly developing economies of China, other parts of East Asia, and India, are changing the energy landscape. There is increasing momentum for developing new energy technologies driven by cost, national security considerations, and concerns about environmental damage from the use of hydrocarbon fuels. Hydrocarbons—oil, natural gas, and coal—dominate the energy scene, accounting for 85–90% of current energy use internationally. Although new hydrocarbon sources continue to be found and exploited, production costs are rising, and they will not last forever. There is also rising concern about environmental impact and global warming. Carbon emissions increased 5.9% in 2010 to the release of one-half of a billion tons of carbon. This jump was the largest in a single year since the Industrial Revolution and the largest percentage increase since 2003. Rising temperatures that lead to sea level rise pose an enormous economic risk, not just to isolated Pacific atolls but also to large stretches of the heavily populated US East Coast.

The dependence of the United States and world economies on hydrocarbons will extend into the future. It will be driven by a new mix of hydrocarbon sources that have been made economic by the increasing price of oil; natural gas and liquid fuels from fracking are the poster children of the moment. However, environmental concerns and longer-term access to adequate energy supplies on a continuing basis are problems that will not go away.

It is possible to look ahead to a new energy future: one that is driven by enhanced efficiencies and increasing use of renewable energy sources. It will be driven by higher costs, decreased availability of accessible hydrocarbons, and a growing recognition of the long-term economic impact of hydrocarbons on the environment. All sources of energy will play a role driven locally and regionally by availability and cost, with increasingly important roles for renewable energy sources—biomass, wind, and solar energy. When integrated with high-efficiency grid systems, they could bring the capacity to power the urban centers and industrial complexes of a growing world economy.

The ultimate renewable energy source should be the sun. It provides ∼10,000 times our current energy needs on a continuing basis day after day. However, the sun as a power source has two limitations. First, it is diffuse. To meet current US power demands with a 10% efficient solar device or devices, a collection area of ∼58,000 square miles would be required. At current semiconductor prices, that collection area would cost tens of trillions of dollars. It would also require a sophisticated grid system. In the United States, it is a long way from New York City to the high deserts of Nevada, New Mexico, and other western states, where there is sunshine in abundance. Even worse, the sun is intermittent, providing only ∼6 h of useful energy per day on the average worldwide. If solar energy is to become a primary energy source, it will have to be tied to energy storage during the day for power consumption at night at unprecedented scales.

The only reasonable solution is artificial photosynthesis with energy stored in chemical bonds as solar fuels. The goal of artificial photosynthesis is to mimic the green plants and other photosynthetic organisms in using sunlight to make high-energy chemicals but with far higher efficiencies and simplicity of design for scale-up and large-scale production.

This goal takes us back to the famous Honda–Fujishima experiment showing water splitting with TiO₂. The simplicity and promise of their result opened new research doors to using sunlight to make solar fuels. In 1974, another door was opened. It came with the experimental demonstration that the metal-to-ligand charge transfer excited state, Ru(2,2′-bipyridine)₃²⁺* [Ru(bpy)₃²⁺*], could undergo electron transfer quenching (scheme 1) (3, 4). The work by Rehm and Weller (5) had earlier shown electron transfer as a viable reaction for organic excited states, and the works by Vogler and Adamson (6), Zinato et al. (7), and Demas and Adamson (8) had suggested that Ru(bpy)₃²⁺* could undergo electron transfer quenching by Co(III) ammine complexes (6–8); however, their results were equivocal and could also be explained by energy transfer quenching.

As shown by the reactions in scheme 1 and flash photolysis measurements, visible excitation of the dπ⁶ Ru(bpy)₃²⁺ ground state gives the corresponding ³(dπ⁵π*_bpy¹) excited state. It was sufficiently long-lived, ∼1 μs, to undergo diffusional electron transfer quenching with added methylviologen dication, MV²⁺ (3). In the overall scheme, the energy of the incident light is captured and transiently stored in the excited state, and then, it is converted into the transient redox equivalents produced by excited state electron transfer.

Scheme 1

Excited state quenching results in conversion of 2.1 eV excited state free energy into 1.7 eV separated, transiently stored oxidative [Ru(bpy)₃³⁺] and reductive (MV⁺) equivalents. Given their redox potentials, with E^o′ = 1.3 V (vs. normal hydrogen electrode) for the Ru(bpy)₃^3+/2+ couple and E^o′ = −0.4 V for the MV^2+/+ couple, the quenching products are thermodynamically capable of water splitting over an extended pH range. The reaction in scheme 1 (3) gave birth to the concept of chemical approaches to artificial photosynthesis (9–11).

Without control of structure in an isotropic fluid, the transiently stored redox equivalents in scheme 1 (3) undergo diffusional back electron transfer, and the stored redox energy is lost as heat. To be useful, a light absorption–excited state quenching cycle must produce, separate, and store transient redox equivalents. It must be used at physically separated catalytic sites for driving the solar fuel half reactions, all before back electron transfer can occur.

To put this concept into perspective, at photosystem II (PSII) in natural photosynthesis, sunlight drives water oxidation by plastoquinone (PQ): 2 PQ + 2H₂O + 4 hν → 2 PQH₂ + O₂. A transmembrane pH gradient is also created for ATP production. The reductive equivalents in the reduced form, the hydroquinone plastoquinol, are transferred to PSI, where light-driven CO₂ reduction occurs in conjunction with the Calvin cycle.

Water oxidation and O₂ evolution occur at the reaction center in PSII (1, 12–24). Oxygen is evolved after sequential absorption of four photons through the four stages of the Kok cycle from S₀ to S₄. The Kok cycle is driven by antenna sensitization of the chlorophyll excited state ¹P₆₈₀. In the S₀ → S₁ stage, absorption of the first photon and production of ¹P₆₈₀ trigger a cascade of five coupled electron–transfer events, with time scales ranging from picoseconds to microseconds (Fig. 1). The net result is transfer of an oxidative equivalent to the oxygen evolving complex separated by ∼50 Å from the reductive equivalent at PQ as the semiquinone, PQ⁻(Q_B⁻) (Fig. 1). At this distance, back electron transfer is sufficiently slow to not interfere with the water oxidation cycle.

Fig. 1.

The sequence of events after antenna-sensitized formation of ¹P₆₈₀ (Chl_D1) in the initial step in the Kok cycle. D₁His190, Histidine190; glu, glutamate; OEC, 4 Mn containing oxygen evolving complex; Pheo_D1, pheophytin D1; Q_A, quinone A; Q_B, quinone B; Tyr_Z, tyrosine Z. Modified from refs. 12 and 13.

Photosynthesis in green plants is an inspiration but not an answer. It evolved slowly, and has remained the same for 2.4 billion years. It is amazingly complex, with five integrated, membrane-bound assemblies involving thousands of atoms bound in the thylakoid membrane of chloroplasts and is integrated with the Calvin cycle for CO₂ reduction. However, its low efficiency for biomass production of ∼1% is inadequate for the high-density power production needed to power our increasingly urban societies.

There is a need for new, more efficient solar solutions. They need to be inexpensive, with their materials in sufficient supply for vast collection areas; also, they need to minimize environmental impact and overcome complications from longevity, production, and distribution.

The basic principles are understood: light absorption, excited state electron transfer, separation of electron transfer-generated oxidative and reductive equivalents by free energy gradients, electron transfer activation of catalysts for multielectron, multiproton solar fuel half reactions, and separation and collection of the products. In semiconductors, band gap excitation gives electron–hole pairs directly, which are the delocalized versions of reductive and oxidative equivalents. Once formed, electron–hole pairs are separated by internal electric fields and directed to spatially separated catalytic sites on the semiconductor surface.

The direct band gap approach pioneered in the works by Fujishima and Honda (1) and Honda et al. (2) is remarkable for its simplicity, but the semiconductor is supposed to do it all—light absorption, charge separation, transport, and even catalysis—without degradation or decomposition. Pure chemical approaches have been proposed with the underlying theme of allowing the molecules to do the work. An example is shown in Fig. 2. Chemical approaches offer the advantage of using a modular approach (6), with individual components for light absorption, etc. investigated separately, maximized in performance, and integrated in appropriate architectures. They offer the advantages of rapid iteration by chemical synthesis and the ability to study individual components separately.

Fig. 2.

Molecular assembly for water splitting. In the absence of an antenna array, multilayer structures are required to achieve sufficient light absorption.

Contemporary interest in chemical approaches to artificial photosynthesis is apparent by the large number of reviews that have appeared in recent years (25–38).

The most straightforward approach to solar fuels is combined photovoltaics–electrolysis, a proven technology with known long-term device efficiencies. Light is absorbed by a stand-alone PV device with electrical or wireless connections to the electrodes of the electrolyzer. The overall efficiency of the device is the product of efficiencies for the PV device and electrolyzer. An example of the wireless approach appears in recent results reported in the work by Reece et al. (39), which used a triple-junction, amorphous Si solar cell with electrodeposited catalysts for water reduction (NiMoZn) and oxidation (amorphous cobalt oxide). This configuration is similar to the one adopted earlier in the work by Khaselev et al. (40), which used Pt catalysts in a wired arrangement. The work by Walter et al. (41) at the Department of Energy-sponsored Joint Center for Artificial Photosynthesis (JCAP) Solar Hub developed a wireless approach, in which at least one of the photoelectrodes incorporated semiconductor nanowires functionalized with catalysts.

The molecular assembly approach is elegant, but in the absence of a viable spontaneous assembly strategy, it requires multiple synthetic steps and is impractical. This result leads to a second guiding theme in solar fuels research—keep it simple—and the hybrid photoelectrochemical approach shown in Fig. 3. It illustrates a dye-sensitized photolectrosynthesis cell (DSPEC) based on a photoanode for water oxidation. DSPECs are closely related to dye-sensitized solar cells in that light absorption occurs at a light absorber (dye or chromophore)—a molecule, cluster, or quantum dot—bound to the surface of a high band gap semiconductor, typically TiO₂. The first example appeared in 1999 for the dehydrogenation of isopropanol (42), and the works by Youngblood et al. (43, 44) developed a DSPEC for water splitting based on IrO₂ as a water oxidation catalyst.

Fig. 3.

Dye sensitized photoelectrosynthesis cell (DSPEC) for water reduction of CO₂ to methane.

Light absorption and injection initiate a series of electron transfer events. For the example in Fig. 3, the injected electron is directed to a separate cathode for catalytic reduction of water or CO₂. A DSPEC device differs from a dye-sensitized solar cell device in using sunlight to drive spatially separated solar fuel half reactions rather than generate a photopotential and photocurrent.

In principle, multiple approaches and alternate architectures may contribute to a solar fuels future. The state of the art at the moment is still very much on the basic research side. However, there is growing comprehension of the importance of anticipating the future and beginning the process of translation of basic science results to the marketplace. The key to future success will be a productive iteration among basic research, translation, engineering, and economics in a highly integrated environment. The goal is to shorten the timescale for evolving a technology or technologies that could help transform our energy future.

An overview of the state of the art in chemical approaches to artificial photosynthesis is the theme of this account. The focus is on basic research and progress made in key areas that have the ability to contribute to an ultimate solution. Enjoy the volume, because there is much to come.

Architectures

Information on architectures is given in Grimm et al.: Step-by-step self-assembled hybrids that feature control over energy and charge transfer (45); Fukuzumi et al.: Formation of a long-lived electron-transfer state in mesoporous silica-alumina composites enhances photocatalytic oxygenation reactivity (46); and Megiatto et al.: Mimicking the electron transfer chain in photosystem II with a molecular triad thermodynamically capable of water oxidation (47).

Catalysis

Information on catalysis is given in Duan et al.: Highly efficient and robust molecular ruthenium catalysts for water oxidation (48); Valdez et al.: Catalytic hydrogen evolution from a covalently linked dicobaloxime (49); McNamara et al.: Cobalt-dithiolene complexes for the photocatalytic and electrocatalytic reduction of protons in aqueous solutions (50); Tanaka et al.: Similarities of artificial photosystems by ruthenium oxo complexes and native water splitting systems (51); and Chen et al.: Splitting CO₂ into CO and O₂ by a single catalyst (52).

Devices

Information on devices is given in Zhao et al.: Improving the efficiency of water splitting in dye-sensitized solar cells by using a biomimetic electron transfer mediator (53); Surendranath et al.: Interplay of oxygen-evolution kinetics and photovoltaic power curves on the construction of artificial leaves (54); and Xiang et al.: Evaluation and optimization of mass transport of redox species in silicon microwire-array photoelectrodes (55).

Mechanism

Information on mechanism is given in Farnum et al.: Visible light generation of I–I bonds by Ru-tris(diimine) excited states (56); Pool et al.: Acidic ionic liquid/water solution as both a medium and proton source for electrocatalytic H₂ evolution by [Ni(P₂N₂)₂]²⁺ complexes (57); Barroso et al.: Dynamics of photogenerated holes in surface modified α-Fe₂O₃ photoanodes for solar water splitting (58); Smieja et al.: Kinetic and structural studies, origins of selectivity, and interfacial charge transfer in the artificial photosynthesis of CO (59); and Vagnini et al.: Ultrafast photodriven intramolecular electron transfer from an iridium-based water-oxidation catalyst to perylene diimide derivatives (60).

Theory

Information on theory is given in Muckerman et al.: Calculation of thermodynamic hydricities and the design of hydride donors for CO₂ reduction (61); Horvath et al.: Insights into proton-coupled electron transfer mechanisms of electrocatalytic H₂ oxidation and production (62); and Lin et al.: Theoretical study of catalytic mechanism for single-site water oxidation process (63).

Photocatalysis

Information on photocatalysis is given in Tamaki et al.: Photocatalytic CO₂ reduction with high turnover frequency and selectivity of formic acid formation using Ru(II) multinuclear complexes (64).