Elucidating Ti dopant effects in hematite photoanodes via high-throughput combinatorial screening

Abstract

Hematite (α-Fe2O3) represents a promising candidate for photoelectrochemical water splitting, but its practical application is hindered by poor charge transport and sluggish surface kinetics. Despite extensive research on titanium doping as a performance enhancer, the underlying mechanisms remain debated, with widely varying optimal concentrations reported across different fabrication techniques. Here, we demonstrate how high-throughput combinatorial approaches can rapidly elucidate complex structure-function relationships that would be challenging to identify through conventional methods. We first optimize titanium content using a concentration gradient sample, then leverage this optimal composition to create two complementary sample types: one with titanium co-sputtered throughout the hematite film and another with titanium sequentially deposited as an overlayer. By subjecting both sample types to similar temperature gradients, we create a powerful comparative framework that reveals how the same elements behave differently based on initial spatial distribution, yet converge toward similar structures at higher temperatures. This parallel temperature evolution allows us to precisely identify the conditions where performance dramatically changes and target these specific regions for in-depth structural characterization. Our findings reveal that Ti enhances photoactivity through two pathways: modifying surface properties to improve charge transfer efficiency and acting as a substitutional dopant to increase bulk charge separation. Structural analysis confirms titanium incorporates into hematite up to a thermodynamic solubility limit of approximately 4-5% Ti/(Ti + Fe), with excess titanium segregating to form surface TiO2 phases that become detrimental beyond optimal thickness. This comparative gradient approach represents a powerful strategy for simultaneously optimizing materials while unlocking fundamental mechanistic insights that are difficult to gain by means of conventional single-parameter studies.