Results of screening articles for problematic Tauc plots.

We searched the full text of articles for the string ‘tauc’, returning 356 articles for which SEM images had brandmarks identifiable by OCR and no detected errors in SEM misidentification (brandmarks_id_correct), 128 articles for which SEM images had brandmarks identifiable by OCR and the SEM instrument was not declared in text (brandmarks_no_id), and 160 articles for which SEM images had brandmarks identifiable by OCR and the SEM instrument was incorrectly declared in text (problematic). We randomly sampled a further 300 articles that had SEM images with no identifiable brandmarks and featured ‘tauc’ in the full text (sem_no_brandmarks). We manually screened each of these articles for whether they actually featured Tauc plots (has_tauc) and whether at least one of these Tauc plots determines band gap at the wrong x axis (not_at_x_axis). (XLSX)