Lock failure and bad lineshape can result in several cases.
1. The sample length is too short.
The standard sample length for 5 mm diameter samples in the Agilent NMR is 2 inches. This will fill the coil with sufficient excess to maintain magnetic field homogeneity. If the sample is too short, the bottom of the tube and the surface between the liquid and air will disturb the homogeneity of the magnetic field around the sample, and will reduce the effectiveness of the gradient shimming, resulting in a bad lineshape. In the worst case, the lock will also fail. Please add enough solvent to meet the length requirement (approximately 500 - 600 microliters in a standard 5 mm diameter NMR tube).
2. The sample does not sit symmetrically to the center of the coil.
When the sample is a little short, you have to position the sample symmetric to the center of the coil. In depth gauge, the probe RF coil position is marked by dashed line, while the center is marked (CL) on the long horizontal line, as shown in Fig. 1. If the sample is not centered on the coil, the spectral lineshape will be bad and the lock may fail.
3. The sample is not all the way in the magnet.
In VT experiments, some users may set the VT gas flow rate too high. As the VT gas flows in from underneath the sample, a rate higher than 18 liters per minute will slightly lift sample, causing its position to deviate from the center of the magnet. Please do not forget to reduce the VT gas flow rate after VT experiments if you have changed it.
4. The gradient shimming failed.
A good lineshape needs re-shimming for each sample. The shimming method uses gradients and the gradient amplifier can sometimes fail. If the lineshape is very bad, but you are sure the sample is correctly positioned, look at the back of the console. On the top (Fig. 2) there is a unit “Pulsed Field Gradient Amplifier and Temperature Controller”. On the “Temperature” part, the “STATE” and “ERROR” lights should be off. If either light is on, reset the unit. Turn off the HEATER switch, wait for 10 seconds and turn it back on. If both lights “STATE” and “ERROR” are off, the gradient should now work.
If needed, you can make a brief test the gradient shimming. On the VnmrJ command line, type ‘gmapsys’, then click “Acquire Trial Spectra”. After the experiment finishes, there should be two square shape lines as in Fig. 3. If the lines are narrow as in Fig 4, the gradient shimming has failed.