The design of the NMR experiment does not lend itself for measurements at exactly determined temperatures. During experiments the NMR tube sits in an rf-coil and the proximity of any parts made out of metal – including a thermocouple – will interfere with the experiment. Therefore, the thermocouple used to measure and regulate the temperature of an air-stream surrounding the sample is sited at a distance from the sample. In addition it also does not help that it is virtually impossible to achieve a homogenous Temperature distribution along the NMR tube, which will lead to convections.
The only type of thermometer, which can be used to directly measure temperatures during the conduct of magnetic resonance experiments are fluoroptic thermometers. These thermometers detect the color change of specific dyes with temperature. Fluoroptic thermometers have been used in MRI studies, however for an NMR experiment one would have to deal with the signal from the thermometer (i.e. the dye and its solvent).
A direct way to directly measure the temperature in the NMR tube is through the temperature-dependent chemical shift of some solvents. For instance the chemical shift difference between the two peaks of methanol is temperature dependent and it can be used to calibrate the actual temperature in the NMR tube with respect to the temperature measured by the thermocouple.
If things are bad in solution they are worse for solid-state MAS experiments. Here the temperature is regulated for the stator, which is the ceramic housing surrounding the spinning MAS rotor. In between the spinning rotor and the temperature-regulated stator is a stream of ambient air, the bearing. This air stream is used to float the rotor within the stator. Again, it is possible to regulate the temperature of the sample within the MAS rotor through the chemical shift of a reference sample. In this case it is lead-nitrate, where the temperature dependence of 207Pb-NMR chemical shift is well characterized.