Tuesday, February 7, 2017

Assigning a simple compound: ODCB

ODCB (o-dichlorobenzene) is a simple aromatic compound used as a standard for monitoring 1H resolution. In a 1H spectrum it produces just two peaks, due to the symmetry of the molecule, but it is not immediately obvious which nuclei cause which peaks.

The chemical structure of ODCB is shown below superimposed on an expansion of the 1H spectrum. The two peaks in the ODCB 1H spectrum are split by J-coupling but the patterns are too distorted, likely by strong coupling, to be of much use figuring out which is which.

A lot of work has been done developing empirical tables (e.g Pretsch et.al. 1) that allow one to calculate chemical shifts given the nature and position of the substituents. There are even online calculators like this one that enable quick calculation of 1H and 13C chemical shifts. Using this online calculator for ODCB gives predicted chemical shifts of 7.28 ppm for the protons closest to the chlorines and 7.16 ppm for the distal ones. The tables in Pretsch et.al. predict 7.27 ppm for the proximal protons and 7.16 for the distal ones. The difference between the predicted chemical shifts matches the difference between the observed peaks, and one could assume that in the experimental spectrum both peaks are moved downfield by the same amount, but this is an assumption and does not prove the assignments.

Moving to 2D NMR experiments, the 1H-1H COSY is not very helpful since it just shows a correlation between the two aromatic resonances. The 1H-13C HSQC shows 13C resonances at 129.3 and 131.6 ppm, which are again different from the chemical shifts predicted by the online calculator (129.7 ppm proximal, 127.5 ppm distal) and by Pretsch et al (128.5 ppm proximal, 128.0 ppm distal). The HMBC, shown below in red overlaid on the blue HSQC, shows correlations between both protons to all three types of carbon nuclei present. Unfortunately, in the HMBC there is no way to distinguish 2, 3 or 4 bond correlations, so again this is not very helpful.

The H2BC experiment was developed to show just two bond correlations2. It can be thought of as a HMQC-COSY experiment where the large 1JCH coupling is used to first record the chemical shift of a 13C nucleus attached to a 1H, then the 3JHH coupling is used to transfer magnetisation to a 1H on a neighbouring nucleus. This scheme will nearly always show just two bond correlations.

In the figure below the H2BC spectrum of ODCB in red is shown overlaid with the HSQC in blue. The upfield 1H resonance at 7.35 ppm show two H2BC correlations, while the downfield one at 7.57 ppm shows just one. The upfield resonance must therefore belong to the protons furthest from the chlorines as these protons would show H2BC correlations to the neighbouring proximal proton and to the neighbouring symmetry-related distal proton. The downfield resonance belongs to the proximal protons which show just one H2BC correlation to one of the neighbouring distal protons.

Another experiment to distinguish two bond correlations from longer ones is the 1,1-ADEQUATE3. This experiment first records the 1H chemical shift then uses 1JCH followed by 1JCC to transfer magnetisation to a neighbouring carbon before finally recording the 13C chemical shift. I tried recording a 1,1 ADEQUATE with NUS on the ODCB sample but did not see any signal at all. I used a 600 MHz spectrometer with a 5mm room temperature probe and ran the experiment using 512 scans for 830 minutes. In contrast, the H2BC was run on the same instrument and sample using 16 scans for just 85 minutes.

As it turned out the chemical shift calculation tables correctly predicted the relative order of the 1H and 13C resonances but the actual chemical shift values were not close enough to be unequivocal. The lesson learned is that calculations may give you a good idea as to what the assignments are but they are not accurate enough to be proof.

1. Ernö Pretsch, Philippe Bühlmann, Martin Badertscher
"Tables of Spectral Data for Structure Determination of Organic Compounds" 4th Ed. Springer Verlag, 2009 (ISBN 3540938095)

2. Nils T. Nyberg, Jens Duus, Ole W. Sørensen
"Heteronuclear Two-Bond Correlation: Suppressing Heteronuclear Three-Bond or Higher NMR Correlations while Enhancing Two-Bond Correlations Even for Vanishing 2JCH"
J. Am. Chem. Soc. vol. 127, no. 17, 6154-6155 (2005)

3. Matthias Köck, Rainer Kerssebaum, Wolfgang Bermel
"A broadband ADEQUATE pulse sequence using chirp pulses"
Magn. Reson. Chem., vol. 41, no. 1, 65-69 (2003).


  1. Nice post!

    How much sample did you use for the experiment? It is surprising to see that 1,1-ADEQUATE did not give a signal. The HD-1,1-ADEQUATE might help in this case since the lineshape might be too complicated.

  2. The sample was a Bruker standard that lists the concentration as 1% which works out to be 6.5mg in 500ul of CDCl3 or 88.8 mM.

    The 1,1-ADEQUATE I ran was homonuclear decoupled as well as NUS. My conclusion is that the ADEQUATE needs a cryoprobe.

    1. Ah, just noticed that it was a room temperature probe. No wonder it did not work for ADEQUATE. I guess a cryoprobe is very necessary.