INTRODUCTION -- First a few general points about magnetic field gradients and shimming -- Then gradients and their associated shims are classified into various categories, and the NMR line distortions caused by category are described. Additional shim protocols are given which describe menu choices, etc.
This information, tell which shim(s) are most badly in need of adjustment merely by looking at the shapes of the NMR lines in a spectrum for the FID. The information can be useful to evaluate spectra obtained elsewhere. This is followed by a listing of which shims must be adjusted under various circumstances. A recommended procedure for routine shimming is then described, followed by a list of common problems. Finally, several sections are devoted to additional information needed only by the NMR spectrometer manager. Users must return shims to the same quality with the reference standard.
The variation of magnetic field with position is a GRADIENT. The frequency of an NMR line is directly proportional to the strength of the magnetic field. If the strength of the magnetic field varies by some percentage over an NMR sample, the resonance frequency will vary by the same percentage, which will smear out the NMR lines and possibly distort their shape. To see a proton line as sharp as 0.3 Hz on a 300 MHz spectrometer, the field must vary over the sample by less than (0.3 Hz / 300,000,000 Hz), or one part in 109. Since the diamagnetic susceptibility of organic solvents is of the order of 10-7, gradients in the NMR probe will change of the order of 100 parts in 109 when sample tubes are removed or inserted.
Replacing one sample with another containing a different solvent (or filled to a different height -- or inserted into the spinner by a different amount -- or in a slightly different NMR tube, or changing the sample temperature, etc.) can cause a noticeable change in the NMR line shape and width. Paramagnetic samples obviously cause much larger changes.
To create a homogeneous (no variation in the magnetic field) field over the sample, distinctively-shaped coils of wire (SHIM COILS) are placed in the vicinity of the sample, and currents ( SHIM CURRENTS) are passed through them to create various gradients of any desired strength. The idea is to adjust the shim currents so as to cancel any gradients in the NMR sample as accurately as possible. This procedure is called SHIMMING THE MAGNET. In order to obtain satisfactory spectra, any operator of an NMR spectrometer must be able to shim out the relatively small gradients caused by changing the sample and the sample temperature. This must be done fairly accurately, or poor spectra will be produced. This routine shimming can be simple, fairly fast, and consistently accurate if done properly. It can be time consuming and produce inconsistent results if done wrong. PROPER SHIMMING STRATEGY IS IMPORTANT, BECAUSE IT HELPS GET A GOOD SHIM FAST.
In principle shimming is simple -- it is an adjustment of the shim currents until the gradients are cancelled out (or minimized) over the NMR sample. In practice, it’s a bit more complicated and tedious. Remember many shim controls interact, so a bad setting of one shim control will prevent finding the exactly correct setting of another. A related problem is false optima: There are some combinations of shim settings which are not good, and changing any one shim by a small amount actually makes things worse.
The most commonly-used criterion for shimming is the height of the lock-signal level and it is insensitive to some shims. In these cases it may not point the way to the best possible shim, and is wrong under some conditions. It helps to be able to identify which shims need to be adjusted, how much and in what direction and in what order to adjust them, and to know when to quit.
CLASSIFICATION OF GRADIENTS AND SHIMS Gradients can have different shapes. For example, if the strength of field is independent of X and Y, but increases linearly in the +Z direction, that’s a Z gradient -- same shape as a P z atomic orbital. An XY gradient is independent of Z, and has the shape of a D xy atomic orbital. A FIRST-ORDER GRADIENT produces a linear variation of magnetic field strength with position, and it’s shaped like a P atomic orbital. SECOND-ORDER GRADIENTS produce quadratic variations of field strength, and are shaped like D orbitals. Third-order gradients produce cubic variations, and are shaped like F orbitals, and so forth. Any 3 independent first-order gradients can be combined to produce a firstorder gradient which points in an arbitrary direction. This is closely related to the fact that 3 orthogonal atomic P orbitals constitute a closed shell. Similarly, 5 independent second-order gradients can combine to produce an arbitrary second-order gradient -- and atomic D orbitals come in 5 different flavors. 7 third-order gradients form a complete set (like the F orbitals), 9 fourth-order gradients, etc."SPINNING" GRADIENTS vary only along the axis about which the NMR tube is rotated (Y for an iron-core electromagnet, Z for a superconducting magnet). The NMR line distortions caused by spinning gradients do not change when the spinner is stopped or started. Normally, SPINNING GRADIENTS ARE ADJUSTED WITH THE SPINNER ON. "NON-SPINNING" GRADIENTS are everything else. They all depend in some way on at least one of the coordinates perpendicular to the spinning axis. The resulting line distortions are visible in all their glory when the spinner is off. When the spinner is on, these line distortions are greatly reduced, but now the non-spinning gradients produce spinning sidebands. NON-SPINNING GRADIENTS MUST BE ADJUSTED WITH THE SPINNER OFF. LOW ORDER gradients are 1st and 2nd order. HIGH-ORDER means 3rd order or higher. ODD ORDERS are 1st, 3rd, 5th, etc. EVEN ORDERS are 2nd, 4th, etc. The 2nd order spinning gradient is sometimes called CURVATURE ("C"). The shim controls on spectrometers can be divided into eight different classes, depending on the type of distortion which each gradient produces in the NMR lines, and depending on whether or not spinning the sample reduces the distortion.
DISTORTIONS PRODUCED BY MAGNETIC-FIELD GRADIENTS The shape of the NMR line(s), shows which gradients are bad. In general, odd-order gradients cause symmetrical distortions of NMR lines, and even-order gradients produce skew distortions. Low-order gradients distort the entire NMR line, while high-order gradients distort mainly the bottoms ("feet" or "tails") of the lines. The higher the order of the gradient, the lower distortion level. In the case of even-order gradients, the shape of the NMR line will tell which direction to move the shim-control knob to improve things. The skew (or "tailing") moves from one side of the NMR line to the other as an even-order shim is moved from a bad setting, through the correct setting, and further on to a bad setting on the other side. On any given spectrometer, the "line-skew" response to a particular even-order shim will always be in the same direction.
How to select a “shim-map” for gradient autoshimming Bracketed commands [test] represents a standard VNMR button. The gradient auto-shimming will currently only work on the Varian 300. Prior to starting, first load initial shims (type bestshim), lock, and shim on both z1 and z2. Also adjust the first five xy shims: x1, y1, xz, yz, and xy. Then iterate between x1 & xz the between y1 & yz as these are strongly coupled. Remember that when adjusting the ‘xy’ shims, the spinner must be off. Next, query the pulsed field gradient amplifier control by typing pfgon? Make sure this parameter is set to pfgon=’nny’. If not, type pfgon=’nny’. If there is a change it from ‘nnn’ to ‘nny’ go back and touch up the z1 shim. 1. Type gmapsys in the Input Window (Run gradient auto-shimming, set parameters, map shims) 2. Click [shim maps] 3. Click [shimmap files] the [Current Mapname] A message “current mapname is : xxx.fid” will be displayed in the output window. If “xxx.fid” corresponds to the shimmap required for the currently installed probe (see the note next to the monitor) skip to step number 4 below. • Click [cd to userdir] This will change the group “vnmrsys/gshimlib/shimmaps” directory. • Click [cd to systemdir] This will change to the “/export/home/vnmr/gshimlib/shimmaps” directory. • Highlight desired shimmap file, e.g., hcn.fid. Again, there should be a note taped nearby telling which shimmap to use. • Click [load shimmap] 4. Click [return] 5. Click [set params] 6. Click [gradient, nucleus] 7. Click [pfg h2] 8. Type nt=8 if using CDCl3, otherwise type nt=4 (if using 501, double value for nt) 9. Click [return] 10. Click [find gzwin] 11. As soon as experiment starts, type sa(‘nt’) then press enter ↵ 12. Type gmapsys 13. Enter the number of z shims to adjust by setting gzsize. Example, to adjust the first six shims, type gzsize=6. This will adjust z1 through z6. 14. Click [autoshim on z] • If the fitting step fails, this often means that the sample volume is either too small and/or the sample is improperly positioned. It is recommended to use 0.7 mL total volume, although volumes as low as 0.4 to 0.5 mL often work. 15. When done, tweak the “xy” shims, again. 16. Load a new set of parameters and restart spinner if desired. When performing a 2D experiment leave the spinner off. 17. To save data, make certain the directory is correct: “/export/home/vnmr/gshimlib/shimmaps”. Type gohome to automatically change back to the user's home directory.
