(Invited) Efficiency Limits for Hydrogen and Formate Production via Fully-Integrated Photoelectrochemical Devices

Monday, 29 May 2017: 10:00
Churchill C2 (Hilton New Orleans Riverside)
K. T. Fountaine (California Institute of Technology, NG Next, Northrop Grumman) and H. J. Lewerenz (California Institute of Technology, Joint Center for Artificial Photosynthesis)
In this talk, I will discuss design considerations for fully-integrated photoelectrochemical devices, supported by a multi-parameter sensitivity analysis of device efficiency and a discussion of parameter interdependency, and framed with experimental results. Both water-splitting and carbon dioxide-reducing devices will be discussed.

In the photovoltaics community, the detailed balance limit, which describes device efficiency as a function of semiconductor bandgap(s), is widely accepted as the gold standard to which all cell efficiencies are compared and by which technological viability is gauged. Conversely, photoelectrochemical devices do not have an equivalently elegant and fundamental limit due to their complex multicomponent nature. The multiple components of a photoelectrochemical device, primarily the photodiodes and the catalysts, force a multi-parameter efficiency analysis in which the parameters are often inter-dependent. This complex efficiency analysis, coupled with the unforgiving voltage cutoff set by the reaction potential results, results in a fickle design problem.

In this talk, I will present a unified framework for fully-integrated water-splitting and carbon dioxide-reducing device performance, which can be generalized to photoelectrochemical devices and used to contextualize other reported efficiency limits under specific conditions. To do so, I will first define analytic equations that govern the current-voltage characteristic curve and the efficiency of a photoelectrochemical half cell (consisting of a photodiode coupled to an electrocatalyst) and a variable-junction photoelectrochemical device.[1] A comparison to real experimental devices demonstrates the validity and applicability of these equations.[2] Second, I will present limiting efficiencies under specific ideal and realistic conditions (both ideal and experimentally realistic) for single, dual and triple junction photodiode units. Subsequently, I will introduce a sensitivity analysis by considering the effects of five parameters – semiconductor absorption fraction, semiconductor external radiative efficiency, series resistance, shunt resistance and catalytic exchange current density – to illustrate and analyze the effects of imperfect light absorption, charge transport, and catalysis on the limiting efficiency and the corresponding optimal semiconductor bandgap(s). The analysis for single junction photoelectrochemical devices reveals that catalyst performance is the most critical parameter to realizing a reasonable efficiency, whereas devices consisting of more than one photovoltaic junction in series depend more equally on all parameters. A discussion of the origin of the discrepancy between the limits discussed herein and reported experimental water-splitting and carbon dioxide-reducing efficiencies will be employed to contextualize this analysis and underscore its ability to provide insight not only into the design process and the advantages of different design motifs but also into analysis of the primary factors limiting device performance. Finally, these experimental examples will also be employed to demonstrate the interdependency between the five representative parameters in real photoelectrochemical devices. Specifically, the tradeoff associated with catalyst loading due to its inverse effects on absorption and catalysis will be discussed, among others.[3]

[1] Fountaine et al., Nat Comm (2016).

[2] Shaner, Fountaine et al., APL (2013).

[3] Fountaine et al., APL (2014).