COSY experiments

The COSY experiment was one of the very first 2D NMR experiments developed and still remains an essential tool. The original experiment has several shortcomings, however, and numerous modifications have been proposed. Three of the most useful are the gCOSY, the DQF-COSY, and the CLIP-COSY. This post discusses the problems of the original experiment and how the new versions attempt to solve them.

gCOSY

The original COSY experiment was devised before probes with gradient capability become common. For coherence selection phase cycling was used instead of gradient pulses. This meant that at least two scans, and perhaps as many as sixteen, were required no matter how much sample was available. The gCOSY experiment incorporated gradients into the pulse sequence and allowed spectra to be recorded with just one scan.

Both the original COSY and the gCOSY produce spectra where the peaks cannot be phased. For this reason the spectra are normally processed in magnitude mode to make everything positive. Unfortunately, the lack of phasing means that the broad dispersive lineshape is retained, leading to peaks with very broad bases. This is a problem because the diagonal peaks, which are much larger than the crosspeaks, produce streaks in both dimensions which can obscure the crosspeaks. To narrow the diagonal peaks the data is typically processed with a harsh window function, an unshifted sine bell.

The spectra above are all from the same experimental data recorded on a sample of cholesteryl acetate in CDCl₃. Only the processing is different. All three plots are plotted at the same contour levels. Notice how the streaks produced by magnitude processing are removed by the sine bell, but the intensity of the signals is reduced as well. Parameters to record a gCOSY on the Skaggs NMRs are saved under the name UCSD_COSY.

DQF-COSY

The double quantum filtered (DQF) COSY improves upon the gCOSY by producing peaks that can be phase corrected. A double quantum filter eliminates all magnetization except for double quantum coherence, which is converted to antiphase magnetization just prior to detection. The signals produced this way can be be phased so that the dispersive peak shape is dispensed with and narrower peaks obtained. This narrows the diagonal peaks in particular and enables better resolution.

Unfortunately, the peaks in a DQF-COSY spectrum are anti-phase, rather than in-phase. Anti-phase means that the sign of the lines in a multiplet alternates between positive and negative. For example, in a doublet one line will be positive and the other negative. In the 2D representation a crosspeak between two doublets will have four lobes of alternating sign. For more complex splitting the crosspeaks become even more complex. In overlapped areas of a DQF-COSY spectrum it can be difficult to tell where the center of the peaks are. Also, DQF-COSY spectra are not easy to phase correctly.

The spectrum above shows the upfield region of a DQF-COSY recorded using the same cholesteryl acetate sample as the other spectra. Parameters for this experiment were the same as those for the gCOSY, except the number of scans was doubled. The plotting threshold was adjusted to compensate so that intensities in the images should be comparable. Parameters to record a DQF-COSY on the Skaggs NMRs are saved under the name UCSD_DQF-COSY.

CLIP-COSY

The CLean In Phase (CLIP) COSY is one of the more recently published modifications of the COSY. This experiment produces spectra that can be phased and provides in-phase peaks. This provides the resolution advantages of the DQF-COSY experiment with a simpler crosspeak representation. Gradients are used for coherence selection, as in the gCOSY, so the number of scans is only determined by the sample concentration. The CLIP-COSY should also work better with the common NUS reconstruction software since its peaks can be phased and are all positive.

The spectrum above shows the same upfield region of a CLIP-COSY on cholesteryl acetate as the other spectra. The parameters used to collect the spectrum were the same as those used for the gCOSY spectrum, and the plotting threshold is set to the same level. Parameters to record a CLIP-COSY on the Skaggs NMRs are saved under the name UCSD_CLIP_COSY.

References

1. Jeener, J
   Ampere International Summer School, Basko Polje, Yugoslavia.
   1971
   
   
2. Carpenter, TA, Colebrook, LD, Hall, LD and Pierens, GK
   Applications of gradient‐selective COSY and DQCOSY to brucine and gibberellic acid.
   Magn Reson Chem 1992 30: 768-773
   
   
3. Piantani U, Sorensen OW, Ernst RR
   Multiple quantum filters for elucidating NMR coupling networks.
   J Am Chem Soc 1982 104: 6800–6801.
   
   
4. Koos MR, Kummerlöwe G, Kaltschnee L, Thiele CM, Luy B
   CLIP-COSY: A Clean In-Phase Experiment for the Rapid Acquisition of COSY-type Correlations.
   Angew Chem Int Ed Engl. 2016 Jun 27;55(27):7655-9