Identification of the Contributing Factors to the Photoelectric Conversion Efficiency for Hematite Photoanodes by Using Machine Learning

Combined Effect of Underlayer and Deposition Solution to Optimize the Alignment of Hematite Photoanodes

“…The parameter dependence of (R2, C2) and (R3, C3) on the bias voltage is shown to focus on the charge transport in the bulk hematite and the solution interface (the other parameters are shown in Figure S7). In the previous study, we figured out that the TiO 2 underlayer could reduce the standard deviation of the R1 value, but it did not affect the average values. In this study, we did not find a systematic difference in the R1 values depending on the sample types.…”

“…This variability was linked to certain features identified through analytical data. , It was also observed that some hematite samples exhibited no photoelectrode activity. Utilizing a categorization approach based on ML, we were able to not only detect these inactive samples but also trace their lack of activity to specific causes . A key factor identified was the resistivity at the hematite’s back contact on an FTO (fluorine-doped tin oxide) substrate.…”

“…Despite its potential, hematite performance is compromised by several factors: a relatively low absorption coefficient, inadequate majority carrier conductivity, and the limited diffusion length (<4 nm) of minority carriers, further exacerbated by a significant overpotential. − To overcome these limitations, considerable research has been directed toward optimizing materials. This has involved techniques such as doping, passivation, and enhancing crystal growth. ,− Nevertheless, a persistent challenge with hematite is the reproducibility of its preparation, , a topic we have previously discussed and sought to address by exploring methods to enhance reproducibility. , …”

Subtraction Descriptors in Machine Learning for Optimizing the Cocatalyst Effect of Cobalt Phosphate on Hematite Photoanodes

“…This result could indicate that the Co-Pi deposition process affects the surface morphology, causing the light scattering to be modified. Similarly, the UV–vis absorption in the UV region was also selected as a dominant descriptor for hematite, , and the absorption profiles in this region have a high correlation with the nanoscale morphology of the materials, − and this indication also suggests the surface morphology was modified with the Co-Pi loading. There is another possibility that the Co-Pi induces additional absorption around this wavelength region, and this has been suggested previously for a Co-Pi/TiO 2 sample, although we did not observe any specific peaks in this region after Co-Pi deposition in our case.…”

Enhancing photocatalytic efficiency with hematite photoanodes: principles, properties, and strategies for surface, bulk, and interface charge transfer improvement

“…13a and b). 112 The introduction of a TiO 2 underlayer beneath hematite in the study conducted by Luo et al 83 effectively addressed the issue of interfacial recombination between the substrate ( e.g. FTO) and hematite.…”