Line Widths, Lineshape, and Sampling Considerations

In Section 1.4 we found that the uncertainty principle establishes spectroscopic time scales. It also has a significant impact on the halfwidth of signals. One version of Eq. (1.6) 

As we saw in Section 3.2, the intensity of an NMR signal is determined in part by the number of nuclei giving rise to the signal. We normally use solution samples to obtain reasonable relaxation times and halfwidths. It is typical to dissolve approximately 5-10 mg of a substance of average molecular weight (say, 200 daltons) in 0.4 mL of deuteriated solvent. Using special microtechniques, it is possible in some cases to obtain H spectra on as little as 10 pg of sample  

When picking a solvent there are several considerations. First, and most obvious, the solvent must be inert, and it must dissolve the sample Of course, if your sample is sparingly soluble or if there is only a fraction of a milligram of sample available, you can usually resort to signal-averaging techniques to obtain the desired S/N. Besides that, it is best if the solvent does not contain any nuclei of the isotope to be examined. 

Make sure you use thoroughly cleaned and dried NMR tubes, and never assume that a new tube, fresh out of the box, is clean to NMR standards. 

At this point it would be useful to review many of the important topics covered in the first three chapters by discussing a few of the actual techniques used to measure relaxation times. 

1 s acquisition time would provide a resolution of (4.1 s) = 0.24 Hz. If we require a resolution better (smaller) than 0.24 Hz, either we need more computer RAM to accommodate a longer acquisition time or we have to decrease our spectral width window. 

Because the FID signal decays due to spin-spin relaxation, there is a time limit beyond which further monitoring of the FID provides more noise than signal. The usual compromise is to have a short enough dwell time to cover the spectral width and a long enough acquisition time to provide the desired resolution, consistent with computer memory limitations. 

After collecting data from one pulse, we must wait for the nuclei to relax to equilibrium. A total of at least 3T is usually adequate, but part of this is spent as acquisition time. Thus, the additional pulse delay time (/ ,), the time between the end of data acquisition and the next pulse, is given by 

Since H nuclei normally exhibit T values on the order of 1 s, we need essentially no additional delay after 4 s acquisition time. For C and other slow-to-relax nuclei, however, substantial delay times are sometimes required. Following the pulse delay, the sample is irradiated with another pulse, and the data acquisition sequence begins anew. 

Lest you worry that you will have to supply each parameter in every pulse sequence each time you take an NMR spectrum, you can relax. Modem spectrometers are menu driven. For most common nuclei all you have to do is select the nucleus and the computer assigns the appropriate pulse sequence parameters for most normal cases. However, when you are dealing with an unusual structure or an uncommon nucleus, you may have to do some experimenting with the parameters to get the best spectroscopic data. 

---