ARTIFACTS IN EXPERIMENTAL AND SIMULATED SOLID-STATE H-2 NMR POWDER SPECTRA

Two types of artifacts in solid-state H-2 NMR echo spectra that may be confused with spectral changes resulting from molecular motion of a d euterated phenyl ring undergoing rr flips are investigated. The first type of artifact arises from improperly determining the initial point of the FID (i.e., the echo maximum). If sampling begins after the top of the echo, the spectrum has the appearance of resulting from a log-G aussian distribution of correlation times (i.e., flip rates). A specif ically deuterated sample of 1,4-bis(1,3-octadecadiynyl)benzene is used to illustrate this effect. The second type of artifact results from c omputational limitations in the simulation of 2H NMR echo spectra. To accommodate corrections necessary to mimic experimental conditions in simulated spectra, it is easier to work in the time domain. Recursion has been suggested as a means for efficiently propagating the simulate d FID. However, finite computer precision results in the ''blurring'' of the long-time components of the FID and thus the subsequent omissio n of fine spectral structure from the simulated spectra. An alternativ e method is suggested in which the FID is propagated using recursive c ycles. This method retains much of the efficiency of the recursive met hod while retaining the fine spectral structure. (C) 1995 Academic Pre ss, Inc.