HMBC (multiple-bond CH correlation) of codeine
This is a 2D experiment used to correlate, or connect, 1H and 13C peaks for atoms separated by multiple bonds (usually 2 or 3). The coordinates of each peak seen in the contour plot are the 1H and 13C chemical shifts. This is extremely useful for making assignments and mapping out covalent structure.
The information obtained is an extension of that obtained from an HMQC spectrum, but is more complicated to analyze. Like HMQC, this is an "inverse detection" experiment, and is possible only on newer model spectrometers. Acorn NMR's new JEOL Eclipse+ 400 is equipped to perform inverse experiments, and uses Z-gradients for improved spectral quality.
The time required for an HMBC depends on the amount of material, but is much greater than for HMQC, and can take from an hour to overnight.
The contour plot shown below is of 3.3 mg codeine in ~ .65 ml CDCl3. See also comparison to the HMBC spectrum of an 18 mg sample.
Normal 1D 1H and 13C spectra are shown along the edges. Peaks occur at coordinates in the 2 dimensions corresponding to the chemical shifts of a carbon and protons separated by (usually) 2 or 3 bonds. The experiment is optimized for couplings of ~8 Hz. Smaller couplings are observed, but their intensities are reduced. Compare to the spectrum obtained when the experiment is optimized for 4 Hz.
The experiment is designed to suppress 1-bond correlations, but a few are observed in most spectra. In concentrated samples of conjugated systems, 4-bond correlations can be observed. There is no way to know how many bonds separate an H and C when a peak is observed, so analysis is a process of attempting to assign all observed peaks, testing for consistency and checking to be sure none of the assignments would require implausible or impossible couplings.
Because of the large number of peaks observed, analysis requires several expanded plots. In this case, the spectrum has been divided into 4 sections, each of which is discussed below.
| The discussion below uses the numbering system shown at right. The numbers were assigned to peaks in the 1D 13C spectrum, starting downfield, moving upfield, and numbering each sequentially. This generates a unique identifier for each Carbon, even before knowing any assignments. |  |
In aromatic rings, the most common correlations seen in HMBC spectra are 3-bond correlations because they are typically 7-8 Hz, which is the value for which the experiment is optimized. The coupling constant is affected by substituents, so 2-bond correlations are also sometimes observed.
The red lines in the plot above show correlations from aromatic proton H-8 to aromatic carbons C-1 and C-6 (both are 3-bond couplings) and a weak correlation to C-2, a 2-bond coupling.
The other aromatic proton, H-7, has correlations to C-2 and C-4, both of which are 3-bond couplings.
The green lines in the plot above show correlations from proton H-9 to carbons C-1, C-3 and C-4 (all are 3-bound couplings). With the poor digital resolution of the spectrum in the carbon dimension (512 data points spread over 17kHz), the peaks for C-3 and C-4 run together because they are barely resolved.
The peaks indicated in red above are due to 1-bond coupling in CHCl3 solvent. Note that the pair of peaks don't line up with any H peaks, but are symmetrically located about the CHCl3 peak, with a separation equal to the 1-bond C-H coupling constant.
The other 2 peaks in this plot are H-7 to C-18 and H-9 to C-17.
The region in the above plot shows correlations between aliphatic Hs and aromatic Carbons. In the lower left corner is a peak showing the 3-bond coupling between the methoxy Hs and C-2, the aromatic carbon bearing the methoxy.
Two protons show correlations to the same 3 carbons. These are the geminal H-18 protons, showing coupling to C-7 and C-4 (both 3-bond couplings) and C-6 (2-bond coupling). The remaining peak is H-17 to C-4. Note that the corresponding coupling from H-17' is not observed. As with H-H couplings, the value of the 3-bond coupling constant is dependent on the dihedral angle.
The last quadrant is shown above, showing correlations between aliphatic Hs and Cs. The peaks are identified below, for each H, starting from the left. Carbons 14 and 15 are only 0.1 ppm apart, and are not resolved in the 2D spectrum.
H-11 to C-14 and/or C-15
H-18 to C-16, C-11
H-13 to C-17, C-14 and/or C-15, C-11
H-14 to C-13 and C-11
H-18' to C-11
H-17 to C14 and/or C-15, C-13
During the process of assigning HMBC peaks, it can be useful to indicate on the plot the positions of the 1-bond correlations. NUTS has the ability to do this, using the compare command. In the HMBC plot below, 1-bond HMQC peaks are indicated by X.
Acquisition Parameters:
512 complex points in direct dimension
128 t1 increments
8 scans
2 sec. relaxation delay
Total acquisition time: 35 min
Processing:
sine squared window function in both dimensions with 0 degree phase shift in t2 and 90 degree phase shift in t1
2x zero-fill in the indirect dimension
magnitude calculation (no phasing is required)
final data size 512 x 512
HMBC of codeine optimized for small couplings
See explanation of the HMBC experiment.
Long range C-H couplings cover a range of values, typically less than 10 Hz. The experiment is optimized for a specific coupling, and selection of that value is a trade-off. Correlations due to couplings of other values are reduced in intensity. To optimize for a specific J value, a delay in the pulse sequence is set to 1/(2J). For very small couplings, this delay becomes so long that much of the magnetization is lost to relaxation. Typically, the delay is optimized for J of 7-8 Hz, the expected value for aromatic 3-bond couplings. However, longer-range couplings can be observed if the delay is optimized for smaller couplings.
The contour plots shown below are of 8 mg codeine in ~ .65 ml CDCl3, in which the experiment was optimized for couplings of 4 Hz.
Compare to the HMBC spectrum optimized for 8 Hz.
Because of the large number of peaks observed, analysis requires several expanded plots. In this case, the spectrum has been divided into 4 sections, each of which is discussed below.
| The discussion below uses the numbering system shown at right. The numbers were assigned to peaks in the 1D 13C spectrum, starting downfield, moving upfield, and numbering each sequentially. This generates a unique identifier for each Carbon, even before knowing any assignments. | |
Compare to the same quadrant in the HMBC optimized for 8 Hz. Several additional 2-bond couplings are observed, as are 1-bond couplings for 7 and 8. The peaks indicated by red circles are 4-bond couplings: H-8 to C-4 and H-7 to C-1.
Compared to the 8 Hz-optimized spectrum, one additional peak is observed in this quadrant, indicated by the red circle. This is a 4-bond coupling from H-7 to C-15.
In this quadrant, three 4-bond peaks are indicated by red circles (H-18 to C-8 and C-1; H-18' to C-8). In addition, two 5-bond couplings are indicated by blue circles. These are H-18 and H-18' to C-2.
The only additional correlations in the quadrant not seen in the 8 Hz-optimized spectrum are from H-13' (just upfield of the intense N-Me singlet) to carbons 17, 14/15 and 11.
Acquisition Parameters:
512 complex points in direct dimension
128 t1 increments
64 scans
2 sec. relaxation delay
Total acquisition time: 4.5 hrs
Processing:
sine squared window function in both dimensions with 0 degree phase shift in t2 and 90 degree phase shift in t1
2x zero-fill in the indirect dimension
magnitude calculation (no phasing is required)
final data size 512 x 512
Codeine cough syrup supplier
ReplyDelete