CHAPTER 1: The Potential Contribution of Photoelectrochemistry in the Global Energy Future

This chapter is directed at a realistic assessment of the role of photoelectrochemistry in the future global energy scenario. We start with a brief history of the development of photoelectrochemical solar energy conversion devices and then extrapolate to the future to offer an opinion about the contribution that photoelectrochemical devices and processes might provide.

We limit this discussion to devices that employ semiconducting materials to absorb solar energy and convert it to photoexcited charge carriers that are harnessed to produce either electrical power or chemical fuels. Photoelectrochemistry implies that the semiconductor is immersed in a solution and its properties are investigated in the dark and under illumination. The modern era of photoelectrochemistry began at Bell Laboratories during the early development of semiconducting materials for use in electronic devices. Bell Laboratories’ researchers immersed various semiconductors such as Ge¹  and TiO₂²  and measured their electrochemical response in the light and dark and reported on their photocorrosion and even, in the case of TiO₂, the ability to evolve oxygen from water when illuminated with light of greater energy than the band gap. Other researchers, most notably Heinz Gerischer, began to publish experiments, models and theories to explain the energetics and kinetics of dark and photoinduced charge transfer at semiconductor electrolyte interfaces.^3–5  However it was not until the energy crisis of the early 1970s, and the paper by Fujishima and Honda,⁶  that the connection between semiconductor photoelectrochemistry and solar energy received wide attention. Although oxygen evolution was observed as early as 1968 when illuminating a rutile electrode in solution,²  the application of this concept to water photoelectrolysis was first pointed out by Fujishima and Honda in a series of experiments that used the n-type semiconductor rutile form of TiO₂.⁷  While rutile is stable under illumination in aqueous electrolytes, its large band gap (3.0 eV) restricts its utilization to the UV portion of the solar spectrum and thus limits its ultimate efficiency. It should also be pointed out that the conduction band of rutile is not negative enough to reduce water, and so a “pH bias” was used in early work, where the oxygen-producing side of the cell was basic with respect to the hydrogen-producing electrolyte.⁸  The simplicity of the Fujishima and Honda experiment, illuminating a rutile crystal electrode with UV light in an electrolyte to produce hydrogen and oxygen directly, and the energy crisis of the early 1970s, set off a flurry of further research in photoelectrochemistry aimed at solar energy conversion. Later work using a device with heterojunctions of III-V materials as photoelectrodes considerably increased the visible light conversion efficiency of direct water photoelectrolysis but at increased cost and decreased lifetime due to corrosion in aqueous electrolytes.⁹ 

In terms of solar energy conversion efficiency, the most successful devices that followed were not water-splitting devices but rather photoelectrochemical photovoltaic cells. Here, illumination of the semiconductor/electrolyte junction drove a reversible redox reaction such as sulfur/polysulfide with no net change in the chemical composition of the electrolyte, instead of photoelectrolysis where chemicals are consumed. These early cells used various semiconducting materials, such as GaAs,¹⁰  CdSe,¹¹  Si,¹²  and MoSe₂,¹³  and reached efficiencies as high as 14% in laboratory cells and in some cases achieved respectable efficiencies with polycrystalline materials due in part to the spontaneous production of a conformal junction by the redox electrolyte.^14,15  The stability of some of these devices was quite good, especially for MoSe₂ and related materials,¹³  but due to both the unproven long-term stability and issues related to encapsulation of the often corrosive liquid electrolytes employed, they never became serious alternatives to solid-state solar cells. These issues, along with the increased supply and lower cost of oil in the period between 1985 and 2000, reduced both the interest and the funding for research and development of photoelectrochemical energy conversion devices. A photoelectrochemical photovoltaic device did emerge in 1991, the nanocrystalline TiO₂ dye sensitized solar cell,¹⁶  which promises to become a contender in low-cost thin-film photovoltaic solar cell market. This cell exploited some of the main advantages of semiconductor liquid junctions; the junction forms spontaneously and is conformal even when high-aspect ratio or porous nanocrystalline networks are involved. This cell has spawned an enormous amount of research, partly because it is rather easy to construct an inexpensive device with a respectable efficiency without sophisticated equipment. However, even if this device is improved and becomes a large commercial success, it will not solve the major problem of transitioning the world to a renewable energy economy. The problem is that over 80% of current energy use is fuels,¹⁷  and so a method to convert solar energy, by far the most abundant renewable energy resource, to storable chemical energy is sorely needed. Therefore, there has been renewed interest in photoelectrolysis, culminating in the recent establishment of a “solar hub” by the United States Department of Energy to develop a device that can perform the photoelectrolysis of water within five years of its foundation in 2010.¹⁸ 

The photoelectrolysis of water to form molecular hydrogen and oxygen is an obvious and direct way to store solar energy as fuel.^19–21  Hydrogen represents stored energy in the form of chemical bonds, bonds that can directly or indirectly react with oxygen to release energy. The hydrogen can also be used to upgrade biofuels or coal to pure hydrocarbons and for the production of ammonia. At present, hydrogen is mainly produced by steam methane reforming using natural gas as the feed stock. This releases a large amount of CO₂ as a by-product and is clearly a non-sustainable process.

By contrast, hydrogen produced from photoelectrolysis of water would be a sustainable process using two of our most abundant resources, water and sunlight, to produce hydrogen and then regenerating the water when the hydrogen is consumed, releasing the stored energy.

Hydrogen can also be obtained, along with molecular halogen, by the photoelectrolysis of haloacids. The advantage of this process is that the electrochemical oxidation of halides is kinetically much faster than the oxidation of water, resulting in less overpotential loss in the reaction. Indeed, in the 1970s, Texas Instruments spent tens of millions of dollars on the development of a very clever photoelectrolysis device to store solar energy by splitting hydrogen bromide.²²  The energy in the stored hydrogen gas and bromide/tribromide electrolyte could be recovered as electrical energy by recombining the hydrogen and bromine in a hybrid fuel cell/redox battery. While, strictly speaking, this device was not a photoelectrochemical system, since the semiconductor p/n junctions were insulated from the solution by a thin metal film, the Texas Instruments device showed nonetheless that the direct photoelectrolysis concept is scalable.

A large-scale operating water photoelectrolysis facility producing meaningful amounts of hydrogen would rapidly saturate the commercial demand for pure oxygen. Oxidation reactions, other than water oxidation, should be considered if there is a market for them, since there is much less demand for pure oxygen than there is for hydrogen. The only other commodity chemical currently produced on a very large scale by aqueous electrolysis is chlorine via the chlor-alkali process,²³  but again the demand for chlorine would be quickly met by any system producing enough hydrogen fuel to make an impact on world energy demand. However, photoelectrolysis technology could get an initial revenue boost from the sale of chlorine and hydroxide obtained from the photoelectrolysis of brine.

Approximately 44% of the world’s population live within 150 km of coastal areas, thus the use of seawater as a feedstock in the photoelectrolysis process is an attractive option. However, the bulk of the anode reaction in seawater electrolysis produces chlorine even though oxygen is thermodynamically preferred. To avoid chlorine evolution from seawater, photoelectrolysis would then require desalination, an expensive and energy intensive process. The alternative is to identify an electrode material that would evolve oxygen from seawater without the concurrent evolution of chlorine. Manganese oxide-based anodes with a high current efficiency for oxygen evolution in preference to chlorine evolution have been reported.²⁴  However, the application of these MnO_x-based electrodes is limited due to their lack of stability, low voltage efficiency and an unknown mechanism for the preferential evolution of oxygen. Integrating such an electrode material into a photoelectrochemical cell would be a challenge.

Any efficient photoelectrolysis material must utilize the solar spectrum effectively and generate sufficient photovoltage to drive the water splitting reaction. Either a single or tandem semiconductor electrode system may be used.⁶  The photovoltage must be greater than the thermodynamic value for the difference between the water oxidation and water reduction potentials (1.23 V at room temperature). Ideally, the semiconducting material(s) should have some catalytic activity for hydrogen or oxygen production so as to minimize the photovoltage in excess of 1.23 V that is required to overcome the electrochemical overpotentials needed to drive the water oxidation and reduction reactions at the desired rate. Current densities (photoelectrolysis rates) of the order of 15–25 mA cm⁻² at illumination intensities around one sun are needed for both single and tandem photoelectrode systems. In addition, a single semiconductor photoelectrode must have a valence band edge that is more positive than the water oxidation potential and a conduction band edge that is more negative than the water reduction potential so that the photogenerated holes and electrons have the necessary electrochemical driving force for the water-splitting reaction without needing to supply additional bias voltage. If a dual photoelectrode system is used, the conduction band of the p-type material must be negative of the water reduction potential and the valence band of the n-type material must be positive of the water oxidation potential. Stability of photoelectrolysis electrodes is crucial for a viable system since the significant capital investment needed for the support structures, electrolyte handling and gas handling systems requires that the system last for many years. Since the photoelectrode will be continually immersed in an electrolyte and illuminated with direct solar radiation, a long lifetime is a daunting challenge that must be adequately addressed even in the basic research stages. It is primarily for their thermodynamic stability that metal oxide semiconductors arguably hold the most promise for constructing a stable photoelectrolysis system.

For standard photovoltaic devices, the efficiency is determined by the product of the open circuit photovoltage, the short circuit current and the fill factor. However, for PEC devices, the only variable in the efficiency equation is the current.

Unlike solar cells, for photoelectrolysis cells (PEC) the potential is fixed to the chemical potential of the product, in this case hydrogen. The only variable in the efficiency calculation is the rate of the production of hydrogen – the current. Other methods including the “energy saved efficiency” have been used to inflate reported water splitting efficiencies, and these methods are reviewed in a prior publication.²⁵ 

As stated above, either a single semiconductor electrode or a two-semiconductor electrode photoelectrolysis system can be envisioned. The single illuminated electrode system will need a band gap greater than about 1.7 eV in order to supply the photovoltage needed for water photoelectrolysis, given that photovoltage of about 2/3 the band gap is often the maximum obtainable from a good semiconductor material operating at its maximum power point. A single photoelectrode system would need to be attached to a catalytic electrode (Pt-catalyzed or a platinum mimic) that can accomplish the complementary water splitting half-reaction. A two-semiconductor system, with both p-type and an n-type semiconductor photoelectrodes, can be configured in several ways. In one configuration, the electrodes could be placed with the larger gap material absorbing the higher energy portion of the solar spectrum in front of the smaller band gap material of the opposite majority carrier type in which the wavelengths transmitted through the large band gap material would be absorbed in the smaller gap material, as shown in Figure 1.1. The two materials must absorb nearly equal numbers of solar photons since the current through each semiconductor must be matched for maximum efficiency. Recombination at the ohmic contacts sums the two photovoltages. The current would be less than a single band gap cell, but the efficiency could be considerably higher because the device will absorb more than twice the number of solar photons since there is considerable photon flux in the red and near IR region of the solar spectrum. This configuration has the advantage that one glass substrate, with a transparent conducting layer on both sides, could be used for both materials, reducing the total system costs. One must also design the device to have low ohmic losses as well as to have minimal gas cross-over between the two compartments. This will require either an ion selective membrane or frits that have high ionic conductivity to reduce ohmic losses. A schematic of a cross section of such a device is shown in Figure 1.1 and a front view of the electrode is shown in Figure 1.2. Another configuration would be to place the n- and p-type materials side by side, rather than putting a larger band gap material in front, allowing full-spectrum sunlight illumination of both electrodes. In this case, an n-type and p-type electrode of the same material or two different oppositely doped smaller band gap materials could be paired as long as the sum of their photovoltages at maximum power was greater than the ∼1.8 eV needed for water photoelectrolysis and the photocurrents in both electrodes were matched. If two materials were used, the band gaps would have to be close to each other, since they need to absorb nearly the same flux of solar photons in order for matched photocurrents and maximum efficiency. Two stable materials would be needed for this configuration to work; a p-type material with a conduction band at least 0.8 eV negative of both the n-type material conduction band and the hydrogen potential, and valence band offsets of about 0.8 V. Two substrates would also be required, but they need not be transparent (as in the case of the stacked system discussed above and shown in Figures 1.1 and 1.2). A separator that conducts ions and inhibits gas mixing such as a membrane or a frit would also be required. Transparent substrates could be used in this case, since back illumination of both gas-evolving electrodes would eliminate all of the backscattering of light from bubbles that prevents it from reaching the photoelectrodes. The result is that the number of possible material systems is broadened and cheaper or more abundant materials may be utilized. Table 1.1 summarizes some of the advantages and disadvantages of the different photoelectrolysis configurations discussed above. Systems that contain buried junctions, where charge separation is remote from the electrolyte, are really solid state solar cells in series with metal electrodes, and they are not considered here since they are essentially equivalent to hooking up photovoltaic cells to an electrolysis cell, albeit without the wires.

Additional losses are associated with the non-ideal placement of the band edge energies. Poor or no overlap of band edge energies with the water redox half reactions can significantly increase the energy losses or prevent the reactions from occurring. An approach to deal with the band-edge mismatch is to modify the surface with charged species to control the band edge energetics. Some work has been done on modifying the surface,^26,27  but this area is still relatively unexplored.

At this point it would be useful to discuss the possible band gaps necessary for the as yet unknown materials to be used in an efficient solar photoelectrolysis device. The band gap values should be used as a guide for researchers’ efforts on the discovery and optimization of these materials. Previous publications have considered limiting efficiencies for PEC water-splitting devices, and these are well covered in a recent review.²¹  Calculations of theoretical efficiencies all hinge on assumptions of efficiency losses, which are primarily the overpotential losses for the hydrogen and oxygen half reactions. The efficiency loss estimates are typically based on the losses observed in electrolysis cells. However, electrolyzers operate at relatively high current densities, whereas at solar intensities, the maximum current at AM1.5 would range from around 30 mA cm⁻² to perhaps up to 100 mA cm⁻² for a single gap photoelectrolysis cell operating under mild solar concentration of around 3–5 suns (or half these values for a tandem cell). Figure 1.3 details the two-electrode voltage required to split water at current densities up to 130 mA cm⁻².²⁸  Even with these low current densities, high catalytic activity is still a necessity. Poor catalysis can allow charges to build up at the semiconductor/electrolyte interface, causing the band edges to unpin, producing a situation where the photogenerated carriers at the band edges are unable to accomplish the water splitting reaction and instead recombine. Bansal and Turner showed the impact of catalysis on band edge movement for the hydrogen evolution reaction in acid.²⁹  At 30 mA cm⁻², the overvoltage loss (H₂ and O₂), under this rather ideal condition, is only about 100 mV. Even with the current for a 3–5×sunlight concentration system, the overvoltage losses shown here are less than 250 mV. This data is for smooth platinum electrodes, whereas nanostructuring or dispersing the Pt or other catalysts can reduce the local current density and thereby further reduce the overvoltage losses.

Clearly, it would be useful to know the approximate band gaps of materials that would be needed to provide high efficiency photoelectrolysis. Weber and Dignam analyzed this problem and came up with an optimum efficiency for a tandem PEC system of about 22% if a high fill factor and overpotential losses of 250 mV are assumed with band gaps near 1 eV and 1.8 eV for the two semiconductors.³⁰  In his chapter in this book, Nozik gives a series of calculations showing the solar-to-hydrogen efficiencies as a function of catalytic activity assuming an ideal semiconductor system (single junction or tandem). An assumed overvoltage loss of 0.4 volts is reasonable and results in a possible efficiency of 30% with a bottom cell at ∼0.8 eV and a top cell at ∼1.6 eV. Nozik and Hanna have also calculated the increased photoelectrolysis efficiencies possible from carrier multiplication (more than one electron per high energy photon) if these effects can be exploited in a practical device.³¹  Carrier multiplication has been demonstrated in a quantum dot (QD) sensitized and a QD bulk absorber device.^32,33  However, new semiconducting materials must be discovered to make this a reality. Systems that convert any extra photovoltage to electrical power can also be designed, but the electrical power produced would most likely be more expensive than if produced from a conventional photovoltaic device, and given that this is the case, squeezing out enough photovoltage to overcome the overpotentials needed to photoelectrolyze water will be challenging enough.

To achieve an efficient, stable and affordable photoelectrochemical photoelectrolysis system, new semiconducting materials with band gaps much smaller than have currently been explored need to be discovered, developed and optimized.³⁴  Semiconducting oxides have the best chance of fulfilling the stability criteria since they can be thermodynamically stable, especially to the valence band holes that need to have the potential to oxidize water. Rocks are a good example of materials that are quite stable over many years in the presence of an electrolyte and sunlight, and most rocks are oxides. Given the choice, it will be far easier prevent corrosion in a thermodynamically stable photoelectrode than to kinetically stabilize it with a thin corrosion barrier that carriers can tunnel through. Since there are many millions of possible metal oxides that could be produced from the 60 or so metals in the periodic table mixed in various stoichiometries, combinatorial techniques would be highly useful for quickly producing and screening these combinations for semiconducting behavior. Several combinatorial techniques for accomplishing this task have been reported^35–37  and reviewed,³⁸  and so they are not be discussed in detail here. However, it is worthwhile to mention that the scaling up of a combinatorial search has been accomplished at the Joint Center for Artificial Photosynthesis (JCAP),¹⁸  and due to the high throughput that they have achieved, the focus may soon shift from the discovery of the new oxide semiconductors to understanding and optimizing the properties as well as configuring the newly discovered materials for photoelectrolysis reactions. The latter task will be more challenging and time consuming than the discovery phase, and many researchers will be needed to investigate these many potential materials. Combinatorial techniques may still be useful for the optimization of material growth morphologies and doping densities, as was recently demonstrated for improving iron oxide photoanaodes.³⁹ 

Another approach is to use density functional theory with high performance computing to calculate the electronic properties of candidate materials and to suggest alloys that should have the necessary properties.^40,41  While a number of possible alloy systems have been suggested using this approach, none have yet shown success. Solving this so-called inverse design problem is still in its early stages, but theory will still be important in helping to understand the role of defects and dopants in the new materials.⁴² 

Often materials are synthesized and tested for photochemical water splitting activity either as colloidal solutions or as powdered slurries using a sacrificial donor or acceptor. While this method may be useful for screening materials for water splitting activity, in our view it is unlikely that homogeneous colloidal solutions or slurries will be useful for a practical water splitting system. Reasons for this include: (i) the products are not produced in separate compartments, resulting in highly explosive mixtures of hydrogen and oxygen (an accident resulting in an explosion would likely terminate any large water splitting project); (ii) even if the explosive mixture can be handled with complete safety, energy is required to separate hydrogen and oxygen, reducing the overall efficiency of the water splitting process; (iii) since catalysts for hydrogen and oxygen production from water are also generally catalysts for the recombination of hydrogen and oxygen, illumination of such systems will eventually result in a photostationary state where forward and back reactions have equal rates and no more net water splitting can occur. In most publications reporting the use of semiconducting particles, this photostationary state is evident as a fall-off in the rate of hydrogen and oxygen production, whereupon the researchers need to purge the system and restart to obtain the initial faster gas production rate. The exchange current in the photostationary state will be related to the solubility of the gases in the electrolyte and the rate at which they are removed from the reactor.

In systems where the electrodes are not or cannot be separated (e.g. semiconductor particle systems), sacrificial reagents are often used in the photoreactors to scavenge either holes or electrons in order to avoid undesirable back reactions and poor efficiency for hydrogen or oxygen production. The most common examples of hole-scavenging agents introduced into solution are alcohols (usually methanol), amines (usually triethanolamine or EDTA) or sulfite salts. Electron scavengers such as the easily reduced Ag⁺ have also been added. Electron or hole scavengers can be useful to study one of the water splitting half reactions without complications associated with the kinetics of the other half reaction (although using a three electrode potentiostat to study a single photoelectrode achieves the same end), but these additives are not viable for any practical system for sustainable energy production since they are not available in the large quantities that would be needed and/or they are much more valuable than the hydrogen produced.

Having been through the basics and some efficiency calculations and various device configurations, one can ask at this point: what is the minimum solar-to-hydrogen (STH) conversion efficiency for a commercially viable PEC water-splitting device? Since the system must cover land area to collect sunlight, it is clear that the size of the plant will be directly related to the commercial viability, since a lower efficiency system will cover significant land area, meaning higher costs for land acquisition, longer piping for collecting the hydrogen and for distribution of the water feedstock. All these factors will increase the balance of plant costs and result in a commensurate increase in the cost of the produced hydrogen. Ultimately it is the price per kilogram of the produced hydrogen that will determine which PEC system can be used to produce sustainable hydrogen and, similar to PV, for photoelectrolysis the cost of that hydrogen is determined by the solar insolation, the STH efficiency, the lifetime, the cost of the photoconverters and the balance of plant. In an attempt to put realistic numbers on the cost of PEC produced hydrogen and how that relates to conversion efficiency, cell costs and lifetime, a technoeconomic analysis of PEC systems is needed. One recent chemical engineering analysis commissioned by DOE looked at four different engineering designs and calculated the costs of the produced hydrogen. The designs were (i) single-bed particle systems, (ii) dual-bed particle systems, (ii) flat plate PEC and (iv) low solar light concentration systems (<10×).⁴³  The projected price of hydrogen was the lowest for the single bed system and highest for the flat plate systems, ranging from ∼$1/kg to ∼$10/kg. However, the single bed system would produce a stoichiometric mixture of hydrogen and oxygen, which was clearly unacceptable to the chemical engineers doing the analysis since one accident would endanger operators, produce a negative public backlash and destroy a capital investment. The dual bed system requires a selective redox relay with redox properties that would be almost impossible to obtain, leaving the flat plate and the concentration systems as the only realistic options.

The main takeaway from this analysis is that the solar-to-hydrogen conversion efficiency has the largest impact on the hydrogen price, and that the efficiency required is much higher than previously thought acceptable for a viable device. A 10% system has generally been considered as an acceptable efficiency, but the technoeconomic analysis shows that the STH efficiency needs to be greater than 15% and in some cases, depending on the systems costs and lifetime, greater than 20%. (US Department of Energy website) One caveat though is that the hydrogen cost goal has been set by the US Department of Energy match that of hydrogen from steam methane reforming, ∼$2/kg. A more realistic metric would be affordability, factoring in the costs of the impacts of greenhouse gases emitted by processes that use fossil fuels. A simple comparison would be gasoline at $4.00/gallon and a 30 mpg fuel economy. The equivalent price for the same distance per kg H₂ would be ∼$8/kg, since fuel cell vehicles have much more efficient fuel utilization. The $8/kg cost must include ∼$1/kg for compression and delivery, so the price of the produced hydrogen should be closer to $6/kg. Nonetheless, it is apparent that the efficiency of a viable PEC water splitting system must be better than 15% in order to achieve an acceptable price for the hydrogen. Another lesson form this analysis was that low solar light concentration (<10×) could produce economic benefits, given that it may lower the costs of other components such as collector area and piping. Another attractive aspect of PEC hydrogen production is that the gases can be produced at higher than atmospheric pressure, reducing or eliminating the capital costs for compressors to compress the gas for pipeline distribution. Thermodynamically a ten-fold pressure increase is possible with a 59 mV addition to the overvoltage, which should be factored in to the band gap analysis presented above.

Carbon dioxide reduction using solar energy is often touted as a route to renewable solar fuels with no net carbon footprint, where any carbon that is either removed from the atmosphere or captured and reduced before entering the atmosphere will not add to the atmospheric CO₂ levels. Although the multi-electron, multi-proton chemistry involved in CO₂ reduction provides an interesting and challenging chemical problem and at first glance appears to be “carbon neutral”, in our view it does not currently make sense as a strategy for reducing carbon dioxide emissions in the context of climate change when the big picture is considered. We examine this premise in the following section.

The first consideration for CO₂ reduction to impact climate change is the scale. Coal production in Wyoming alone is 425 million tons/year, resulting in annual emission of nearly a billion tons of CO₂ after burning in coal-fired plants (cf. Figure 1.4 – total US production is about 3 times this). An industrial scale process to reverse the impact of just Wyoming’s coal production would dwarf even the current largest industrial chemical process, the production of sulfuric acid, by a factor of 25. If one considers reducing CO₂ to a liquid fuel using renewable energy thinking that there would be value added, the logical source of CO₂ would be a fossil fuel burning power plant since it is very concentrated in the stack, whereas it takes energy to extract and concentrate it from the air. The largest source of CO₂ in the US (as well as many other countries in the world) is from coal burning plants, and so we qualitatively examine the energy balance of this approach.

At first sight, photoelectrochemical CO₂ reduction looks attractive. The raw materials needed (CO₂ and sunlight) are free, and one might even get a subsidy to utilize or prevent the release of the CO₂. However, this thinking only considers a small part of the entire coal burning system (the stack) as shown with the red box in the upper right of Figure 1.4. In order to do a proper analysis as in thermodynamics, one must put the box around the whole system, as in the larger green box in Figure 1.4. It is then apparent that there is a considerable release of CO₂ from burning fossil fuels in machines for mining the coal, shipping it to the power plants in the locations far from the mine (e.g. Wyoming to the Midwest in the US or from Australia to China) and then pulverizing the coal before combustion. The coal is then burned to produce heat that is converted to electricity at an efficiency of about 33% in a modern coal-fired plant (the efficiency of older plants can be as low as 20%). Along with CO₂, burning coal also results in the emission of sulfur dioxide, mercury, other heavy metals and radioactive metals. Huge volumes of coal ash, which contains toxic metals, must be disposed of by shipping it to off-site (and out of sight) locations using more fossil fuel (although some of the ash can be used in building materials). These wastes have resulted in widespread water pollution in the US since their disposal is not federally regulated. However, focusing on CO₂, we see that at least 3 equivalents of CO₂ are emitted to the atmosphere to produce one equivalent of electrical energy. However it is much more than three when considering the previously mentioned fossil fuel-based energy inputs into the whole coal burning to electricity cycle. Therefore, it appears to make no sense to use diffuse, expensive to harvest sunlight to reverse the combustion reaction even to produce methanol, the easiest liquid fuel to synthesize. This is especially true when considering that the solar converter will be at most 20% efficient and that it takes six electrons per CO₂ molecule to convert it to methanol with efficiency losses in all of the electrochemical and chemical steps. The cost and stability of the catalysts also needs to be considered. Instead, renewable energy should be installed to replace the electricity from as many coal-fired power plants as possible “up front” before any attempt is made to undertake any large scale CO₂ reduction. In other words, the cost of generating electricity from solar photovoltaics and preventing a given amount of carbon emissions is much lower than the cost of capturing and reducing CO₂ after burning a fossil fuel. Implementation of photovoltaic (PV) arrays to replace the electrical output of coal-fired plants would pay off by reducing CO₂ emissions by probably a factor of four or more per unit of electric energy produced as well as by eliminating toxic emissions and reducing the number of deaths from coal mining that averages several thousand per year world-wide. Stated another way, to deploy an artificial photosynthetic system to totally reverse the CO₂ output of a typical 300 MW coal-fired plant, the land and solar collector area needed would be at least three or four times that needed to simply replace the electrical output of a 300 MW coal plant with photovoltaics. In practice, the area would be much more since it would be difficult to drive the complex chemistry at the same efficiency as direct PV electricity production (10% would be impressive). Solar conversion economics is driven by the capital costs of producing large areas of collectors, so the capital costs for conventional PV would then probably be at least six times less than needed to construct a system to convert the CO₂ to liquid fuel when considering the both the efficiency and the complexity of a complete CO₂ reduction installation that needs investment in gas handling, purification systems and catalysts. Further efficiency improvements of the solar PV systems will be gained if they are decentralized, unlike the conventional power plants, to be near the consumption of the power, resulting in considerable reductions of the very high transmission losses. Of course solar PV can only replace the peak output of a coal plant during daylight hours, and other renewable power or energy storage would be needed on the grid to replace the coal-fired electricity during the night where wind energy may be more available. This discussion is most pertinent to the many coal-fired plants that operate at high capacity during daylight hours and scale back output during darkness.

If a liquid fuel source with a minimal carbon footprint is required, it makes more sense to use the abundant coal resources, gasify the coal and do well-established high-temperature reactions using hydrogen produced from sunlight by the photoelectrolysis of water to reduce the gasified coal to methanol by the reaction sequence:

Despite the fact that it still produces net carbon dioxide emission, this approach has several advantages. First, the photoelectrolysis of water is much easier and more fully developed than the multielectron reduction of CO₂. Secondly, the high-temperature gas-phase catalytic chemistry is well known and fairly well optimized. Also, only one equivalent of CO₂ is emitted into the atmosphere from burning or using the methanol in a direct methanol fuel cell, as opposed to much three to as much as six when one tries to reverse the combustion in a coal burning plant. The gasification route may still be a preferred method of liquid fuel production even if water is used as the reductant by employing water gas shift chemistry after gasification.

In this approach, a minimum of 3 CO₂ molecules are released (considering that even these well-studied catalytic processes are not 100% efficient) after use of the liquid fuel, and ideally the gasification/fuel producing plant could be located near a geological formation that is acceptable for CO₂ sequestration. Therefore, any large-scale CO₂ reduction scheme for mitigation of atmospheric CO₂ emissions makes absolutely no sense until virtually all coal-burning plants are replaced with carbon-free power such as wind, nuclear or solar.

Considering possible sources of CO₂, while the reserves of fossil fuels are large, they are finite and there are climate change issues to be considered and so ultimately, to be sustainable we must get our CO₂ from the air. While >400 ppm CO₂ has a potent impact as a global warming gas, it still is rather dilute and would require great expenditures of energy and capital to collect and concentrate it for reaction. Biomass is nature’s way of collecting CO₂ from the atmosphere and generating reduced carbon that can be converted into fuels. However, typical bio-fuels have significant oxygen atom content, and pure hydrocarbons are preferred as fuels. Again, a better approach would be to pyrolyze biomass and use the pyrolysis oil as feed stock combined with PEC hydrogen for hydrocracking/hydroforming to produce carbon-neutral fuels of the sort currently used in the transportation infrastructure. Huge amounts of hydrogen produced from methane are already used in the petroleum refining industry, and photoelectochemically produced hydrogen from water would also eliminate CO₂ emissions associated with this industry. Of course eliminating deforestation and replanting of forests is the most near term and effective way of CO₂ mitigation since photosynthesis by self-replicating trees is low tech and effective.

Photoelectrochemically generated hydrogen from water could also be useful in another large industrial process that releases CO₂, nitrogen fixation via the Haber Bosch process. This process already consumes 3–5% of the world’s natural gas and 1–2% of the world’s energy supply, and therefore renewable hydrogen would have large global impact.

In summary, stable photoelectrochemical water splitting systems are the preferable target for photoelectrochemical energy conversion and, if made cheaply and with a high enough efficiency to be a cost-effective renewable hydrogen source, could realistically impact large-scale future energy conversion and storage. In order for this to happen, researchers must concentrate their efforts on discovering and improving stable materials that are capable of efficient solar photoelectrolysis. Since the hydrogen economy is the only long-term sustainable renewable fuel-based energy alternative, the question is not if but when will the world begin its transition to this energy carrier. In addition to research and development, barriers, to this conversion include the availability and low cost of fossil fuels, especially coal and recently natural gas, and the lack of a hydrogen infrastructure as well as political and sociological issues. Perhaps some of the latter issues can be addressed by educating people about the true costs of continued exploitation of non-renewable reduced carbon resources, such as climate change, land and ecosystem destruction, air and water pollution and their detrimental impact on human health and wellbeing. One hopes that this awareness will inspire countries to shift their subsidies from non-renewable to sustainable energy systems such as solar driven hydrogen production. To quote Sinclair Lewis, “It is difficult to get a man to understand something when his pay check depends upon him not understanding it.”

Some first steps that have been taken are exemplified by the recent increases in government funding of solar energy research and by the establishment of several centers focused on solar hydrogen production, including the Joint Center for Artificial Photosynthesis (JCAP) in California and the Korean Center for Artificial Photosynthesis (KCAP) in Korea as well as solar fuel research projects in the European Union, China and Japan. We can hope that this investment will result in research breakthroughs that will put us on the path to a world economy based on clean and sustainable energy and that photoelectrochemically produced hydrogen will be a large or at least contributing part of that economy. In Winston Churchill’s words, “People and countries can do the right thing once they have exhausted all other alternatives.”

Funding from the Division of Chemical Sciences, Geosciences, and Biosciences, Office of Basic Energy Sciences of the U.S. Department of Energy through Grant #DE-FG02-05ER15750 is acknowledged.