2D spectra, particularly for homonuclear correlation, can show a variety of artifactual signals in the F1 domain. Common sources include carry-over of signal modulation from one transient to the next ("rapid pulsing artifacts") and systematic variations in room temperature ("parallel diagonals"). In both cases there is one very simple expedient which can greatly reduce the impact of these sources of error. Multidimensional data sets are almost invariably recorded by simply incrementing or decrementing evolution periods, largely for reasons of convenience and historical precedent. If instead the sampling of the evolution periods is carried out in random order, the perturbations responsible for the sharp F1 signals in the conventional experiment manifest themselves as t1 noise. Since the randomized acquisition redistributes coherent artifactual signals randomly in F1, the maximum artifactual signal is substantially reduced in the randomized experiment and no longer appears in the form of misleading distinct peaks.
Copyright 1999 Academic Press.