The chemical composition of protoplanetary disks, especially in the molecular layer and above, is set by photoprocesses driven by stellar radiation fields. Chemical modeling approaches in the literature treat the radiation field with varying levels of sophistication with respect to the propagation of photons or the wavelength dependence of photoprocesses. We present a systematic investigation of how simplifications to the radiation field treatment impact our inferences of disk composition. Using a modified version of the Nautilus code applicable for the highly irradiated environment of disks, we evaluate the accuracy of six different methods of approximating the ultraviolet stellar spectra, and therefore the computed reaction rates, for six young star and disk models (using spectral energy distributions of the stars RECX 15, GM Aur, RU Lup, HD 135344B, MWC 480, and AB Aur). We compare the rates computed via each approximation, and we investigate differences in their modeled chemistry. We find that the approximations can lead to significant differences in the rates for photoprocesses like desorption, dissociation, and ionization. Of the approximations considered, low-resolution ultraviolet spectra provide rates consistent to within a factor of 2 from the fiducial high-resolution ultraviolet spectra. Otherwise, accuracy of the other approximations is highly dependent on the host star and molecule. More reliable predictions are then best achieved by either using a downsampled ultraviolet spectrum from the source in question, or a reasonable ultraviolet spectral template. Otherwise, we advise that underlying assumptions on the ultraviolet spectrum are carefully considered when modeling disk chemistry across the pre-main-sequence stellar distribution.
Pegues et al. (Thu,) studied this question.