COSY Processing

gCOSY is acquired as a Magnitude spectrum so the data is not phase sensitive- phase = 1. Thus, there is no phase correction and the apodization should be a sine like function (sine, sine-squared)

1) In most cases, default gCOSY processing will be fine in MestrReNova, but potentially you can improve it with adjusting Zero Filling , Linear Prediction , and Apodization

2) Zero-Filling. There is normally no need to zero-fill in F2 but if you acquired less than 2K points, zero-filling to 2K might help resolution some. in F1, you want to zero-fill to at least 1K, and probably 2K. This should lead to some resolution improvement.

3) Linear Prediction. There is no need to linear predict in F2. In F1, predicting 2 to 4 times the actual data points acquired will likely improve resolution. The software may default to predicting to the end of the zero-fill, but that might be too many. It is suggested to limit to the linear predicted points.

4) Apodization. The optimum apodization function is most often a sine-squared function in both dimensions which yields moderately sharp resonances with minimal truncation artifacts. You can observe the shape of the function by turning on Full View under View Menu and then selecting Fid Preview at the bottom of the Apodization Menu. For a Sine Squared function in both dimensions, either choose Sine Square 0° or Sine Square II 0.0% 50%. These 2 should be equivalent. For a sharper resonance, you could switch to a Sine Bell of Sine Bell II with the same 0° or 0.0% 50%. Alternatively, change the % values of the Sine Square II and observe the change of the apodization function and result in the data.

5) Other options. The reduce t1 noise option under Processing menu can reduce the t1 noise across the spectrum. The t1 noise is noise across the spectrum in the t1 (F1) dimension off of intense peaks.

6) Symmetrize. Since COSY data should be symmetric on either side of the diagonal, the Symmetrize option under the Processing menu will look for cross peaks on both sides and make the data equivalent on both sides. However, this can potentially eliminate real signals that are more easily observed on one side of the diagonal and amplify noise that is coincidental on both sides. In general, this option is not recommended, but it can make the data look substantially cleaner.

7) Display options. One option that can appear to improve resolution or signal is to adjust contours. Under the Edit Menu then P roperties, there are options to change the Display including traces on the spectrum. One option that can improve the look of the plot is to increase the number of contours displayed, and decrease the spacing between contours. On the option 2D, just above Traces there are choices for Positive/Negative Contours, Scaling and Line Width. Increasing the number of contours can help, but drawing speed will slow as you incre ase that. More useful is setting the Scaling to ~1.3- the optimum value depends upon number of contours and the signal. Typically a number i n the 1.1 to 1.5 range is best. The Line Width can also improve darkness on the screen.