Shoulder Checkmr. Mac's Virtual Existence

Even the best laid plans of NMR spectroscopists oft go awry. It may be as simple as a poorly positioned sample causing high-order shimming problems or as perplexing as persistent out-of-phase 3 Hz sidebands. Whatever their cause, these artifacts can ruin a spectrum or lead to an incorrect analysis. This page will attempt to list the most common artifacts in NMR spectroscopy that arise from incorrect sample preparation, improper instrument set-up, and processing problems. Below you will find spectra that contain a problem and the various causes for those problems as well as their resolution (if possible).

Please choose from the following list to find a resolution to your spectral artifact.

Problem: Asymmetrically broadened peaks with non-Lorenzian shape.

Cause: Misadjustment of Z2 shim.

Problem: Symmetrically broadened peaks with non-Lorenzian shape.

Cause: Misadjustment of Z1 shim.

Problem: Asymmetrically broadened peaks with multiple maxima.

Cause: Misadjustment of Z1 and Z2 shim.

Problem: Asymmetrically broadened peaks with multiple maxima.

Cause: Misadjustment of Z3 and/or Z4 shim.

Problem: Center glitch. Artifact at exact center of spectrum.

Cause: Slight imbalance in Quadrature detectors. Usually visible with low number of transients.

Spectrum 8: 4 Hz doped H₂0 sample.

Result: Poor autophasing, interference with 'real' signals, potential structure misassignment.

Resolution #1: Acquiring more transients will eliminate the problem. Take a minimum of 4 transients. More is better.

Problem: Receiver overflow causing significant distortions in spectrum.

Cause: Gain is set too high. Only should occur when autogain is turned off.

Problem: ADC (Analog-to-Digital Converter) overflow resulting in spectral distortion.

Result: Useless spectrum.

Resolution #1: Turn autogain on by typing gain='n' and reacquire the spectrum. If you get a message similar to 'gain driven to 0...', your sample is too concentrated and should be diluted to get a good spectrum. It could still be run by reducing the pulse width to 1 or less (e.g. type pw=1).

Problem: Rolling baseline.

Result: Difficult to interpret and impossible to phase.

Resolution #1: Reset the phasing parameters to zero by typing rp=0 lp=0, then type aph0 to perform automatic zero-order phasing. Finish up by clicking Phase, then click and hold on the rightmost peak and drag the mouse up or down to get proper phasing for the right peak (if you run out of room, click Phase again and continue). Now click and hold on the leftmost peak and drag mouse up or down to correct left phase. Redo right end if necessay. When completed, type ds and hit return.

Resolution #2: Reset the phasing parameters to zero by typing rp=0 lp=0, then click Phase, then click and hold on the rightmost peak and drag the mouse up or down to get proper phasing for the right peak (if you run out of room, click Phase again and continue). Now click and hold on the leftmost peak and drag mouse up or down to correct left phase. Redo right end if necessay. When completed, type ds and hit return.

Problem: Aliased or folded Peaks (compare to Spec. 6 and note the peaks around 0 ppm). This is more typical for carboxylic acids and certain hydrides.

Result: Peaks that are outside the spectral window will fold over to the opposite end of the spectrum by the amount they are outside the window. In the above example, the aromatic peaks that should be at 7.2 ppm are folded to ~0.5 ppm. This can lead to incorrect spectral assignment.

Resolution #1: Increase spectral window to accomodate all expected peaks and rerun experiment. This is done by typing setsw(upper ppm, lower ppm), where upper ppm is the upper limit for your spectrum and lower ppm is the lower limit. If you want a spectrum from 20 ppm to -10 ppm, type setsw(20,-10) then type ga.

Resolution #2: An alternative is to double the size of the existing spectral window by typing sw=sw*2. Rerun the experiment with ga to get the new spectrum.

Problem: All Peaks have multiple tops.

Result: All peaks with have many maxima and spectrum cannot be interpreted.

Resolution #1: Stop experiment if it is running. Open the LOCK interface by clicking acqi. Click on LOCK, turn LOCK OFF, move LOCK slider to find 'zero beat', click LOCK ON, reduce lock power to appropriate level (20-32 for CDCl₃, 5-20 for other solvents), click CLOSE (see Locking for more help). Rerun spectrum.

Resolution #2: If you do not have a lock solvent, open the LOCK interface by clicking acqi. Click on LOCK, turn LOCK OFF. Rerun experiment with lock off.

Problem: Baseline is crooked.

Result: Slanted spectrum and inaccurate integrals.

Resolution #1: Type f cdc dc to display full spectrum, clear previous drift correction, and apply new drift correction. If you have peaks at the edge of the spectrum, this may cause a more tilted baseline. In this case, either acquire spectrum with larger spectral window (type sw=sw*2 ga) and apply f cdc dc or expand region with no peaks at the edge and type f cdc dc.

Problem: Spinning sidebands at multiple of spinning speed.

Result: Additional peaks that make interpretation difficult.

Resolution #1: First make sure that your NMR tube is good. It should be rated for the field you are using and free from any defects. It should not of been stored in the oven. Once you have a good tube, open LOCK interface (click acqi =>LOCK) and turn off spinning. Increase LOCK GAIN and/or LOCK POWER to maintain lock (be sure not to saturate). Click SHIM and then right mouse click the arrow next to SHIM in the middle of the window. Choose first order non-spinning shims. Adjust X and Y. Now adjust X and ZX then Y and ZY. Finish with X and Y. Click on LOCK and start spinning, reduce LOCK POWER and GAIN to appropriate level. Click SHIM and right mouse click next to the middle SHIM and choose Z shims. Adjust Z1 and Z2.

Resolution #2: If resolution #1 does not work, you can either run the sample non-spinning or increase the spinning rate to 30-35. Do not increase it too high because this can cause vortexing and lead to bad spectra.

Problem: Poor digital resolution.

Result: Inaccurate peak shape, poor integration, missing fine structure.

Resolution #1: For standard small molecule proton NMR, the acquisition time should be at least 2 seconds and preferably 4 seconds. Check acquisition time by typing at?. If it is too small, type at=4 to set it to 4 seconds and rerun experiment. If acquisition time is good, you can increase the Fourier number by typing fn=4*np wft. This will set fn up to 4 times the number of acquired data points. All additional points will be zeros and add no additional noise. This technique is called 'zero filling'.

Problem: Poor integration. The aromatic signals (around 7.2 ppm) should integrate to 5.

Result: Inaccurate integration and possible wrong structural asssignment.

Resolution #1: The most common cause of this problem (assuming baseline is flat, phasing is good) is having a recycle delay that is too short. Check the current recycle delay by typing d1? (that's d and the number one). Note the value and then increase it by a factor of 3. Thus, if d1 is 1, increase it to 3 or even 5. Rerun the experiment. If integration is still wrong, increase d1 further and repeat experiment. If you want to be accurate, obtain your molecule's T1's (see T1 measurement opens PDF) and set d1 + at to be 5 times the longest T1 of your molecule.

Resolution #2: An alternative to #1 is too reduce the pulse angle according to the Ernst equation to get the best signal with the given d1 and known T1. This approach is not recommended because it is intended for obtaining the best S/N for a given recycle delay and T1 and will not allow for full relaxation between scans.

Problem: Truncation artifacts or so-called 'sinc wiggles'.

Result: Potential obscuring of peaks.

Resolution #1: This usually occurs with large peaks and a short acquisition time. Check the acquisition time by typing at?. If it is 2 or less, type at=4 to set to 4 seconds and rerun experiment. Otherwise, increase acquisition time by typing at=at*2. Rerun experiment.

Resolution #2: If it is small (much less than in example), add exponential multiplication by typing lb=0.3 wft (this is for proton data). You can use bigger lb values, but this will reduce resolution and in these cases it is best to use resolution #1.