Testing EUV/X-ray Atomic Data for the Solar Dynamics Observatory

TL;DR

This study evaluates the accuracy of atomic data in the CHIANTI database for interpreting solar EUV/X-ray spectra, revealing significant underestimations in flux predictions for key AIA passbands and proposing empirical corrections.

Contribution

The paper assesses the completeness and accuracy of atomic data in the CHIANTI database for SDO EUV/X-ray bands using Procyon spectra, highlighting discrepancies and suggesting empirical adjustments.

Findings

Spectral models reproduce soft X-ray and EUV spectra well.

Models underestimate flux in the 50-130A range by factors of 1.5 to 7.

Observed flux in AIA bands at 94A and 131A can be underestimated by factors of 3 and 1.9.

Abstract

The Atmospheric Imaging Assembly (AIA) and the Exteme-ultraviolet Variability Experiment (EVE) onboard the Solar Dynamics Observatory include spectral windows in the X-ray/EUV band. Accuracy and completeness of the atomic data in this wavelength range is essential for interpretation of the spectrum and irradiance of the solar corona, and of SDO observations made with the AIA and EVE instruments. Here we test the X-ray/EUV data in the CHIANTI database to assess their completeness and accuracy in the SDO bands, with particular focus on the 94A and 131A AIA passbands. Given the paucity of solar observations adequate for this purpose, we use high-resolution X-ray spectra of the low-activity solar-like corona of Procyon obtained with the Chandra Low Energy Transmission Grating Spectrometer (LETGS). We find that while spectral models overall can reproduce quite well the observed spectra in the soft X-ray range ll < 50A, and at the EUV wavelengths ll >130A, they significantly underestimate the observed flux in the 50-130A wavelength range. The model underestimates the observed flux by a variable factor ranging from \approx 1.5, at short wavelengths below \sim50A, up to \approx5-7 in the \sim 70-125A range. In the AIA bands covered by LETGS, i.e. 94A and 131A, we find that the observed flux can be underestimated by large factors (\sim 3 and \sim 1.9 respectively, for the case of Procyon presented here). We discuss the consequences for analysis of AIA data and possible empirical corrections to the AIA responses to model more realistically the coronal emission in these passbands.