I have prepared a sealed DMSO-d6 capillary tube (3mmOD) to use as an internal reference for an undergraduate prac using Evans method.

When I run the capillary with DMSO-d6 only in the outer tube (ie no sample) there is a small difference in the shift of the residual DMSO solvent peak of the outer tube vs the capillary (~0.005ppm, 300MHz). The difference is not great enough to effect the result of the experiment but I would like to be able to explain the discrepancy to the students.

Due to the large number of samples it is impractical to run samples manually therefore we use an autosampler, but it is too expense to buy the number of coaxial inserts required, so I have used 3mmOD economy tubes (Wilmad) for the capillaries with Teflon tubing used as a spacer to hold the capillary vertical in the tube. While this works well it is not held as firmly as a coaxial insert would be.

Would the difference in shift be due to the capillary not being absolutely vertical/wobbling or because the capillary tube is lower quality glass than the outer tube (Wilmad 528-PP), ie more paramagnetic impurities, or some other phenomenon?

Any answers and/or easy to follow references explaining this observation would be greatly appreciated.