1H-13C HSQC spectra are designed to show correlations between carbons and their directly attached protons. Interpretation of all the correlations observed in a HSQC spectrum is an essential step in characterizing a new molecule. It is not uncommon, however, to observe HSQC correlations over more than one bond. At first these long range correlations can be quite confusing, but if one is aware of the possibility of long range correlations being present, they can actually help the analysis.
The HSQC spectrum below was recorded using a standard sample containing 100mg/ml of cholesteryl acetate in CDCl3. The spectrum is plotted with a low threshold so that noise around the one bond correlations is significant, but it also allows several weak long range correlations to be observed. Three of the expected one bond correlations are labelled with black text; H6-C6, H3-C3 and H4-C4. Multiple bond correlations involving these resonances are labelled in red. Most of the multiple bond correlations are over two bonds, but there is a
three bond correlation between H4 and C6 that is propagated through a double bond.
These multiple bond correlations are much weaker than the one bond correlations and it is possible to set the threshold so that only one bond correlations are shown. However, using a low threshold to expose the multiple bond correlations allows one to confirm assignments by linking one bond correlations through a two bond correlation.
For example, starting with the H3-C3 peak at 4.62,74.0 ppm two peaks can be observed directly above at the same 1H chemical shift. The first of these has the same 13C chemical shift as the H4-C4 peak and is the two bond H3-C4 correlation. The complementary H4-C3 two bond correlation can also be observed directly below the H4-C4 peak and completes a rectangle linking the H3-C3 and H4-C4 correlations. The presence of these multiple bond correlations supports the assignment of the one bond resonances to adjacent carbons. Observing these multiple bond correlations in HSQC spectra allows resonances to be linked, as in a H2BC spectrum.
HSQC experiments select one bond correlations using a delay that corresponds to the 1JCH coupling constant. Typically the delay is set to correspond to a coupling of 145 Hz. Multiple bond CH correlations are typically in the range 0-10 Hz, so the likelihood of recording such a correlation with a delay corresponding to a 145 Hz coupling is small, but it is not zero. If a concentrated sample is used then weak multiple bond correlations are often observed, particularly between intense peaks.
One should take care when assigning these multiple bond correlations in HSQC spectra. Since the peaks are weak it is easy to mistake noise for a real correlation. Real multiple bond correlations should line up exactly with the one bond correlations. Looking at the H6-C6 peak in the bottom left of the spectrum, a peak to the right at the same 13C chemical shift can be seen. However, this peak has a 1H chemical shift that does not match any other resonance in the compound, and so is not a true correlation and is not assigned.