The morphology of semiconductor photoelectrodes significantly affects the performance of photoelectrochemical devices. Complex anisotropic morphologies are important for overcoming performance limiting bulk transport properties of semiconductor materials, but are often also an unintended outcome of the fabrication process. A better understanding of morphology-induced transport limitations of photoelectrodes is needed.

We used a coupled experimental-numerical approach to quantitatively characterize morphologically-complex photoelectrodes (nano- to micrometer thick semiconductor-films composed of mesoscopic structural units with nano-scale structural details). We utilized a 3D-microscopy method, FIB-SEM tomography with a high resolution of 4x4x4 nm³, to obtain a grey value array representing the photoelectrode morphology. The digital structure was segmented based on trainable machine-learning algorithms to subsequently quantify performance-related morphological parameters.

We applied this method to two distinct photoelectrodes, different in structure, composition, and scale: i) a particle-based lanthanum titanium oxynitride electrode with a film thickness of a few micrometers, and ii) a ‘cauliflower-like’ structured hematite electrode with a film thickness of a few hundred nanometers. The digitalized morphology of each of the two films was used to quantify specific surface, mean feature dimensions, and film homogeneity. Further, the structural characteristics in the meso and nano-scale, including the shape and orientation of these structural details, were quantified.

The digitalized photoelectrode morphologies were then used in direct pore-level simulations to understand transport around and in the semiconductor. Charge carrier generation rates in the semiconductor phase were calculated by an electromagnetic wave propagation simulation based on spatially resolved material density profiles. The generation rates were mapped onto the semiconductor-electrolyte interface and limitations in the diffusive ion transport in the electrolyte were investigated with a finite volume solver.

The FIB-SEM tomography, with its high, nanometer-scale resolution, reveals precise structural information of semiconductor films at the submicrometer scale. The methodology proofs to be applicable to various photoelectrodes and provides a unique insight into their morphologies. The analysis of the 3D-data obtained allows for the qualitative and quantitative assessment of performance-related morphological parameters, and can characterize and identify limiting transport phenomena in the structure in order to guide the morphology and fabrication of optimized photoelectrodes.