Relative coronal abundances derived from x-ray observations. III. The effect of cascades on the relative intensity of Fe xcii line fluxes, and a revised iron abundance

Observations of the helium-like and hydrogen-like resonance lines of N, 0, Ne, Na, Mg, Al, Si, S' and Ar below 25 A, made from the U.S. Air Force satellites OVI-lO and OV!-17, have previously been used to construct coronal models and to determine the relative abundances of these elements. Lines of Fe xvii are also prominent in the coronal spectrum below 25 A. The intensities of the lines of the 2p-3d transitions near 15 A have also been used in conjunction with these coronal models, and the assumption that the excited states are populated only by collisional excitation from the ground state ("coronal excitation"), to determine the Fe abundance. We have observed eight lines of Fe xvii which originated from the 2p53s, 2p53d, and 2s2p63d configurations in the coronal spectrum between 13 and 18 A. Many of the upper levels are populated by cascade so that the assumption of "coronal excitation" is not a valid one for most of these lines. Recent theoretical work on the spectrum of Fe xvii has taken account of the statistical equilibrium of the first 36 excited states in computing the intensities of the observed lines. Although the theoretical atomic rate constants for Fe xvii (especially the collisional excitation rates) cannot be as accurately calculated as those of the simpler ions, the relative intensities of the 2p-3d Fe xvii lines observed in the corona are now in good agreement with theoretical predictions. The measured flux in the lines from the 2s22p53s levels, which are populated in part by cascade from the 2p63l (1 = 1, 2, 3) levels, have been used to obtain the excitation-rate coefficients for the 2s2p631 3L levels, since the theoretical rate coefficients for these excitations are less accurately known than those for the 2p-31 transitions. Using this more complete theoretical model, and higher-resolution observations which allow a number of line blends to be resolved, we have revised our earlier calculation of the coronal abundance of Fe. We find an iron abundance relative to hydrogen of 26 x 10-6, which is in excellent agreement with the photospheric iron abundance. Subject headings: abundances, solar - atomic and molecular processes - coronal lines - line identifications - spectra, solar - X-rays, solar