“…28 The integral has been performed over the FSDP, 0.3-2.1 Å Ϫ1 , 29 which gives ⌬D/DӍ1Ϯ0.5%, where the error is mainly due to sample-to-sample variation. This ratio is comparable to the experimental observations of the photoexpansion, ⌬V/VӍ0.5%, 8,9,12 despite the rough estimation procedure. In detail, we must take into account x-ray changes in the whole wave number, specifically in smaller wavenumber regions, the proportionality factor between ⌬V/V and ⌬D/D, 22 and so forth.…”