T wo major technological challenges in the development of a sustainable, clean energy system are providing massive grid-scale energy storage and an ample supply of carbon-neutral, highenergy-density, transportation fuels. The development and deployment of massive, grid-scale energy storage is imperative for reliably and robustly compensating for the intermittency involved with the utilization of very large amounts of wind energy and solar energy. 1 Another challenge is that ~ 40% of current

more »

... al transportation fuel is consumed in uses for which electrification is technically difficult, if not impossible, such as in heavy-duty trucks, ships, and aircraft. 2 Exhaustive use of advanced biofuels might possibly supply adequate carbon-neutral transportation fuel for these uses, but could not then also fulfill the requirement for long term, massive, grid-scale energy storage. 3 Chemical fuels are desirable for energy storage because fuels are the most energy-dense storage medium known to man (other than the atomic nucleus), and could simultaneously provide a means to baseload at scale intermittent renewable energy resources while also fulfilling gaps in the need for high energy-density, carbon neutral, sustainable, transportation fuels. 4 Hence a clear rationale exists to develop technology options that involve the conversion of sunlight, by far the largest energy source, directly into chemical fuels. One approach to address both of these technology development imperatives involves the development of artificial photosynthesis. In artificial photosynthesis, sunlight is directly converted, without the use of (or the limitations of) living systems, into a useful chemical fuel. 5 Artificial photosynthesis has been pursued in the laboratory for over 40 years, since the observation by Fujishima and Honda that exposure of TiO 2 to sunlight effects water splitting to produce, with high quantum yields, H 2 and O 2 . 6 However, to date, no single light absorber for artificial photosynthesis has been shown to combine simultaneously three desired technological attributes: efficiency, affordability, and robustness. 7 A manufacturable artificial photosynthesis system also involves much more than a single photoelectrode, necessitating the incorporation of suitable catalysts, materials for the separation of the products, mitigation of undesirable effects of bubbles and flows, methods to manufacture the system at scale, and approaches to encapsulate the materials while maintaining facile reactant access and product egress. The requirements that are imposed on the components therefore depend intimately on the design of the whole system. Hence, optimal progress toward a viable solar fuels technology mandates a holistic, systems approach, to identify and then solve the research and development needs that historically have served as barriers to the development of a scalablymanufacturable solar fuels generator. Natural photosynthesis provides a complex, but elegant, blueprint for the production of fuels from sunlight. With only water, carbon dioxide, and sunlight as the inputs, solar energy is stored in the form of chemical bonds as the output of photosynthesis. However, natural photosynthesis has significant performance limitations at the systems level, such as saturation at approximately one-tenth the peak intensity of sunlight; relatively modest overall energy conversion efficiencies on an annually averaged basis (1% or less); the need to spend significant amounts of energy internally to regenerate the unstable enzymes and to resynthesize the highly exquisitely ordered and arranged molecular machinery of photosynthesis; and in general the production of a fuel that is not directly compatible with widespread use in existing energy systems. 8 Production of fuels directly from sunlight is thus inspired by natural photosynthesis, but has the mandate to provide far superior performance than photosynthesis. In this respect, "performance" is measured by the net annually averaged energy conversion efficiency to produce a useful chemical fuel in a scalable, cost-effective fashion. A fully artificial photosynthetic system would also not require arable land, potable water, or involve tradeoffs of land to be used either for food or for fuel production. It is clearly possible to construct a fuel-producing, man-made, solar energyconversion system that outperforms natural photosynthesis on an efficiency basis. For example, solar panels can be over 30% efficient in conversion of sunlight into electricity. 9 In turn, electrolyzers can take electrical energy and produce H 2 and O 2 from water at over 70% energy efficiency. 10 Hence, in combination, the sunlight to fuel (in this case solar-to-hydrogen, STH) energy-conversion efficiencies of a modular photovoltaic/electrolyzer combination system can be over 10 times greater than that of the fastest growing plants (on a yearly average). A goal of research in artificial photosynthesis is however not only to demonstrate high efficiency, but to develop a technology that is the basis (continued on next page) for a fully integrated solar fuels generator system that can simultaneously combine the three desired attributes of cost-effective scalability, robustness, and efficiency. A fully integrated artificial photosynthesis system is a complex assembly that will need to bridge many length scales, likely over as many as seven orders of magnitude (Fig. 1) . The requirements, outcomes, and success of the R&D at each scale length are intimately dependent on the requirements, success, and outcomes of the R&D needed to construct such a system at many other scale lengths. For instance, the requirements on the materials used for light capture on the nanoscale depend significantly on the form factor and architecture of the system developed as a prototype on the cm-to-m length scale. Similarly, success on the µm length scale for the production of fuels from sunlight will generally produce bubbles and flows of product that would, if not controlled or mitigated, degrade the performance of the system on the m length scale. Rapid progress therefore is facilitated by vertically integrated R&D efforts that simultaneously address a multitude of critical bottlenecks at a multitude of length scales, in a parallel, spiral development type of structure.