FAQ 98-08.1 NOESY, ROESY and TROESY: what is the difference?
This is quite possibly the most often asked question... In one way they are all the same: all three techniques provide correlations based on through-space interactions of protons as opposed to correlations based on coupling interactions as found in GCOSY, HMQC and many more. In all three cases the relationship between the strength of the interaction and the distance between the nuclei is the same: 1/d^6 (d is the inter-nuclear distance between the protons). What is substantially different is how these correlations are obtained (different pulse sequences) and how the max. attainable peak intensities (peak volumes) depend on the size of the molecule and the frequency of the spectrometer. Please see also zero-NOE estimation guide added 2007-12-12.

The diagram illustrates this. For certain combinations of field (w_o, e.g. 500 MHz) and molecular size (expressed in the correlation time t_c), the NOE can be zero even when the protons are very close in space (this is often called the correlation time problem).

ROESY and TROESY do not have this undesirable feature: when the NOE is zero the ROE is +50%. The disadvantage of ROESY is that besides the desirable effects through space, also some undesirable contributions via couplings are observed, especially in strongly coupled systems (carbohydrates are one but not the only example!). The TROESY (a modified ROESY) minimizes these contributions, but it does not always fully eliminate them either. Based on years of experience, the TROESY is the most robust and versatile technique, provides stronger correlations than NOESY except for large molecules at high field (typically proteins) and very small ones at low field strengths and is therefore implemented on all our spectrometers.