Is it related to the resolution of the UV-vis system? We have measured the diffuse reflectance spectra and then we converted it in absorption . Our system resolution is 0.05 nm
Your system's resolution is probably not the limiting factor. Instead, there are two likely error sources: 1) The bandgap determination is based on a linear fit. The standard errors of the components (slope, intercept) of your linear fit determine the overall error of the measured bandgap. 2) There may be variation within your sample. Repeat the measurement 3 (or better more) times at different locations on the sample and calculate the average and standard deviation.
I am assuming either 1) or 2) is dominating your experimental error.
maybe the references mentioned in https://www.researchgate.net/post/What_is_the_best_experimental_method_for_calculating_of_the_band_gap_of_a_semiconductor_material_what_is_it_based_on