University of Ottawa NMR Facility Web Site

Please feel free to make suggestions for future posts by emailing Glenn Facey.

Wednesday, October 31, 2007

T1 Measurements and Estimation

T1 relaxation time measurements are usually done with a simple 180 -tau -90, inversion recovery pulse sequence (see figure). Tau is varied from a small value to a large value and a nonlinear regression is carried out to fit the best T1 value.These measurements can be very time consuming. One can get a reasonable estimate of the T1 much more quickly. Follow these simple steps:

1. Call up the pulse sequence "t1ir1d" (Bruker) or "s2pul" (Varian).
2. Set p1 and p2 (Bruker) or PW and P1 (Varian) to the 90 degree and 180 degree pulses , respectively.
Set the recycle delay, d1 (Bruker and Varian) to something you believe is much longer than the T1.
3. Set tau to a very small value (3 microseconds for example). Tau is d7 on a Bruker spectrometer or d2 on a Varian spectrometer.
4. Collect a spectrum and phase it such that all peaks are negative (one scan is often enough for protons). Store the phase correction.
5. Repeat step 3. increasing d7 (Bruker) or d2 (Varian) until the peak of interest is nulled. If the peak is negative, tau is too short. If it is positive, tau is too long.
6. The T1 of the peak of interest is the tau value for the null divided by the natural log of 2.

Tuesday, October 30, 2007

Spectra Acquired with a Sweeping Field

Have you ever seen spectra like those in the top and middle traces below and wondered what was the problem?
These spectra were acquired while the magnetic field was sweeping. This will happen on a Bruker spectrometer if you do not bother locking the field (or fixing the field when running unlocked). On rare occasions in automation, if there is difficulty in locking your sample, I have seem the spectrometer run a spectrum while sweeping the field. This however is very infrequent.

Monday, October 29, 2007

Quadrature Images

Quadrature images are caused by an imbalance in the magnitude of signals in the x and y channels of the receiver. Such an imbalance will create a quadrature spike in the Fourier transformed spectrum (see entry for October 11, 2007) as well as quadrature images. An example of quadrature images is shown below for the 1 scan proton spectrum of a 10% ethylbenzene in CDCl3. Quadrature images are small reflections of large signals in the spectrum about the center. They are much less of a problem with newer instruments than they were with older ones. They become smaller as more scans are collected due to receiver phase cycling. You should be aware of quadrature images if you are searching for very small signals in the presence of very large ones.

Friday, October 26, 2007

Why Are Some of My HMQC Correlations Doubled?

HMQC and HSQC signals are 1H detected with 13C decoupling. The sequences essentially throw away all of the signals for protons attached to 12C and retain all of those for protons attached to 13C. The coupling between 1H and 13C is used to establish a correlation but while the 1H signal is being detected, the 13C is decoupled. Occasionally you may find that some of your HMQC/HSQC correlations appear to be doubled or have a much lower than expected signal-to-noise ratio. Often this is due to poor 13C decoupling and is a problem for correlations at the extremes of the 13C chemical shift scale. At moderate magnetic field strengths, 13C "garp" decoupling is efficient over about 150 ppm or so. If the correlations are near the limits of the center of the 13C axis plus or minus (150 ppm/2), then you will find a loss in signal-to-noise ratio and perhaps some signal doubling due to partial 13C decoupling. Below is an example of the partial HMQC spectrum of 3-heptanone. In the left panel the 13C offset was set near the 13C resonances. In the right-hand panel the 13C offset was set at 110 ppm and the terminal methyl signal is poorly decoupled.

If you encounter this problem, center the 13C offset closer to the correlation which is poorly decoupled.

Thursday, October 25, 2007

New Solids NMR Spectrometer for the University of Ottawa

Congratulations are in order for Dr. David Bryce whose successful CFI proposal will fund a new solids NMR spectrometer for the University of Ottawa. Way to go Dave!!

To Spin Or Not To Spin?

For liquids NMR one can average out the transverse inhomogeneity of the magnetic field by spinning the sample tube about the direction of the magnetic field. This will lead to sharper lines and a higher signal-to-noise ratio however artifacts called spinning sidebands will appear at multiples of the spinning frequency on either side of an NMR line. The intensity of these artifacts depends on the degree transverse inhomogeneity of the magnet, the quality of the NMR tube and the NMR probe. The decision to spin depends on how well the transverse shims of the magnet (i.e. x, y, xy, (x2-y2).... etc..) are set, and your tolerance for spinning sidebands. In a very well shimmed magnet it may not be worth spinning at all. In a poorly shimmed magnet spinning will provide a substantial line narrowing as well as a huge boost in the signal-to-noise ratio. For 2D experiments, one should never spin as sidebands will introduce undesirable artifacts in the spectra. Below are two comparisons of spinning vs non spinning in a well shimmed and a not-so-well shimmed magnet. What do you think? Should you spin?

Wednesday, October 24, 2007

Measurement of 13C-19F Coupling in a 1H-13C HMBC

It is often difficult to find the 13C signals in fluorinated organics by direct 13C observation. It may sometimes be easier to find these signals by way of a 1H-13C HMBC spectrum if there are nearby protons in the molecule. As shown below, it is even possible to measure the 13C-19F coupling constants.

Tuesday, October 23, 2007

Measuring 1H Line Shape - The Line Shape Sample

It is important for NMR spectroscopists to know how well their magnets are shimmed and how well their probes are performing as far as line shape and width are concerned. A quantitative specification is needed so one magnet-spectrometer-probe configuration can be compared meaningfully to another. We use a standard sample of 1% chloroform in acetone-d6. The resonance of CHCl3 is examined. The width of the line (in Hz) is measured at 50% of the line height, 0.55% of the line height and 0.11% of the line height. 0.55% of the line height was chosen as it is the height of the 13C satellites. The specification is usually stated as x/y/z where x = the width at 50%, y = the width at 0.55% and z = the width at 0.11%. Often just y/z is given. These numbers are often quoted with and without spinning the sample. In the example below from our Bruker AVANCE 400 spectrometer on a non-spinning sample, the line shape was 0.27 Hz / 5.28 Hz / 8.72 Hz.

Monday, October 22, 2007

Eliminating t1 Noise in 2D-Homonuclear Data

2D experiments often have t1 noise (not to be confused with T1 relaxation). This noise is evident as "stripes" perpendicular to the directly detected domain (see the left panel of the figure below). Several factors such as instrumental instability and temperature instability can lead to t1 noise. In homonuclear experiments one can remove the t1 noise by symmetrization of the data about the diagonal. In this process all regions symmetric about the diagonal are compared. The region having the largest signal is thrown away and the region with the smallest signal is put on both sides of the diagonal. This procedure will retain all symmetric signals (i.e. cross peaks) and eliminate t1 noise. Below is an example of the 1H COSY spectrum of ethylbenzene.

One should use this method with care as artificial cross peaks will appear for uncoupled signals with excessive t1 noise. Before symmetrization, one should look for the smallest real off-diagonal signal. Make a mental note of the signal. Symmetrize the spectrum and then scale it such that the smallest real off-diagonal signal noted above is the smallest signal in the symmetrized spectrum.

On a Bruker spectrometer one can symmetrize the data with the commands "sym" and "syma" for magnitude and phase sensitive data, respectively. On a Varian spectrometer the command is "foldt".

Sunday, October 21, 2007

Excellent Resource for Solids NMR

If you are a beginner at solids NMR, you should take a look at this EXCELLENT link from Durham University.

It is full of valuable and practical knowledge.

Friday, October 19, 2007

I Only Have 0.5 mg of Sample. What Can I Do?

Here are your options on collecting NMR data when you have only a limited amount of sample.

1. Use a standard setup and get more sample or collect more scans. Yeah yeah..... I know ..... if you could do this you wouldn't have a problem.

2. Reduce the volume of the solution keeping the same mass and lift the NMR tube such that all of the sample is in the receiver coil. This is not a very desirable option as the magnetic susceptibilities of the solution inside the coil and the air outside the coil are very different and will result in a severe distortion of the magnetic field. It is very difficult to shim the magnet to compensate for this effect and the resolution and signal-to-noise ratio in your spectrum will suffer greatly.

3. Reduce the volume of the solution keeping the same mass and use a "Shigemi" tube. These tubes have plastic parts which go above and below the sample which is centered in the receiver coil. The plastic parts are chosen to have the same magnetic susceptibility as the solvent used to prepare your solution.

4. Reduce the volume of the solution keeping the same mass and use an NMR tube of a reduced diameter with a probe designed for such a tube. This is an option we have at Ottawa U. on the Varian INOVA 500 for which we have a 3 mm probe.

5. Use a Cryogenically cooled probe. These probes have cryogenically cooled coils and electronics which dramatically reduce the amount of noise thereby increasing the signal-to-noise in the spectrum. They are very expensive to both purchase and maintain and not currently an option at Ottawa U. Maybe we can ask Santa.....

Thursday, October 18, 2007

Proton Probe Tuning for 13C Detected Experiments

I am sometimes asked: Do I have to tune the proton channel of the probe for my 13C detected experiments? The answer is yes. If you are just running a simple 13C with proton decoupling then the efficiency of the decoupler will depend on the tuning of the proton channel. If the proton channel of the probe is very badly tuned, you may see broadening and even splittings in your carbon signals. The signal-to-noise ratio will also suffer. For experiments like DEPT or INEPT, the proton tuning is critical as these sequences require proton pulses of specific flip angles. If the proton channel of the probe is not tuned and matched then the flip angles for the proton pulses in the sequence will be less than they should be and you will not get the results you expect. In the figure below is an example. The bottom trace is a 13C DEPT-135 spectrum of menthol with the proton channel of the probe properly tuned. The spectrum in the upper trace is also a DEPT-135 spectrum of menthol run under the same conditions except that the proton channel of the probe was detuned. It looks more like a DEPT-90 as the proton pulses are closer to 90 degrees than 135 degrees.
Remember to tune the proton channel!

Wednesday, October 17, 2007

The Proton Decoupled 31P NMR Spectrum of Triphenyl Phosphate

I was asked about the origin of the small peaks in the 31P NMR spectrum of triphenyl phosphate posted to this blog on October 4, 2007. The small peaks are 13C satellites. The phosphorus is coupled to the 13C in the ipso (J=7.47 Hz), ortho (J=4.98 Hz), meta (J=0.93 Hz) and para (J=1.32 Hz) positions on the aromatic rings. The coupling constants were measured in a high resolution 13C NMR spectrum. The figure below shows the proton decoupled 31P NMR spectrum with the couplings color coded. The satellites due to the meta and para couplings overlap. The displacement of the doublets is due to the isotope effect (see blog entry for September 13, 2007).

Tuesday, October 16, 2007

U of O NMR Facility Blog Receives a Positive Review

The University of Ottawa NMR Facility Blog was started less than two months ago. Even though it was intended for the NMR users of the University of Ottawa, it is being read by others interested in NMR as well. It has received a positive review from Ryan Sasaski of Advanced Chemistry Development. Check it out here.

Ryan's NMR Blog is a very valuable resource for those interested in ACD's NMR prediction software as well as other issues related to NMR.

Thank you Ryan!

Monday, October 15, 2007

The Consequences of Setting the Receiver Gain Too High

The receiver gain of an NMR spectrometer is much like the volume control on a radio. When it is set too high there will be distortion in the NMR signal. The FID will be clipped near the beginning of the signal. The Fourier transform of this distorted signal is a distorted NMR spectrum. The figure below shows what this distortion looks like.
Many spectrometers will calculate the receiver gain automatically however you should be aware that this automatic calculation is not always perfect and that the receiver gain may have to be set manually. On a Varian spectrometer the receiver gain is the "gain" parameter. On a Bruker spectrometer the parameter is "rg". In both cases higher numbers mean a higher receiver gain.

Friday, October 12, 2007

Nyquist Fold-Backs in HMBC Spectra

The digital filters of modern NMR spectrometers have not only improved the signal-to-noise specification quoted by instrument companies but they have also eliminated Nyquist fold-backs for signals outside of the spectral window. This is not the case however for the indirect dimension of 2D experiments. In the left panel of the figure below is a properly recorded HMBC spectrum of 3-heptanone. In the right panel is a spectrum acquired with the 13C spectral window set too small. The carbonyl correlations are folded into the low frequency end of the 13C axis.
If you inadvertently collect a spectrum with a Nyquist fold-back you can still calculate the correct chemical shift, as the signal will be the same number of ppm away from the wrong end of the axis as it is outside of the correct end of the axis.

Thursday, October 11, 2007

Quadrature Spikes and How to Get Rid of Them

Have you ever seen a small sharp peak in the exact center of your NMR spectrum? This is called a quadrature spike and is due to a small offset in the two channels used for quadrature detection. It can sometimes be confused with a real NMR peak. You can get rid of this peak very simply by reprocessing your data. In XWINNMR or TOPSPIN do the following:
Type "bc_mod qfil"
Type "bcfw 0.02"
Type "efp" to reprocess your data

This will null a region 0.02 ppm in width at the center of the spectrum. In the unlikely event that you have a real signal at the exact center of the spectrum, you should not use this technique as it will null out the real signal as well. Below is an example of removing such a spike in a 13C NMR spectrum.

Wednesday, October 10, 2007

Why is the Tuning Display Unstable When I Tune the Probe?

On the Varian INOVA 500 you may notice that the tuning display is unstable while you are tuning the probe. On the Bruker spectrometers, this may show up as ripples or waves on the "wobble" tuning curve. This phenomenon is due to Q modulation in the probe as a result of spinning the NMR tube. The tuning of the probe is dependant on the orientation of the NMR tube within the probe. This problem is much worse for cheap NMR tubes or NMR tubes which have been damaged or lack perfect radial symmetry. The effect can also be seen for solids samples spinning in MAS probes.

To avoid the problem, simply stop spinning the sample.

Tuesday, October 9, 2007

13C NMR of Fluorinated Organics

Many NMR users (of the organic variety) come to believe that all 13C resonances are singlets as the most common way to collect the data is with proton decoupling, which indeed reduces most 13C resonances to singlets. It is a surprise to some to find multiplicity in the 13C resonances when a nucleus such as 31P or 19F is present in the molecule. Below is the aromatic region of the proton decoupled 13C NMR spectrum of alpha, alpha, alpha-trifluoro-p-tolualdehyde (the aldehyde peak is not shown). The assignments are given by the colors in the figure. Note the multiplicities in the 13C resonances due to the 1, 2 and 3 bond J coupling to 19F.

Some people find it surprising that the trifluorinated methyl group comes in the aromatic region of the spectrum. Fluorinated carbons are often difficult to find in 13C spectra with low signal-to-noise ratios as the signal is spread over multiple lines and can be buried in the noise.

Friday, October 5, 2007

The Width of Your Water Line - Radiation Damping

The width of the water resonance in a proton NMR spectrum depends critically on the amount of water present. When the concentration of H2O is very low the NMR resonance is very narrow. When the concentration is very high the width is many times greater. The reason for this is that the strong magnetization of the water signal induces currents in the NMR coil which generate magnetic fields which broaden the line. This phenomenon is called radiation damping. The width of the water line is a function of the strength of the water signal which depends on the amount of water, probe tuning, field strength, coil size etc... Radiation damping can also affect the symmetry and phase of the peak. Below are the 500 MHz proton NMR spectra of two samples of H2O / D2O with different concentrations of H2O. In both cases the magnet is well shimmed.

Thursday, October 4, 2007

The Effect of Proton Decoupling on the Signal-to-Noise Ratio for 31P NMR

I have seen some students collect 31P NMR data without proton decoupling. Their reason was that there were no directly bound protons to the phosphorus. What they have not taken into account are the remote protons which have small couplings to the phosphorus. There are usually many of these remote protons and the cumulative effect of the couplings is to broaden out the 31P resonance. This broadening causes a severe loss in the signal-to-noise ratio. The upper trace of the figure below is a single scan 31P spectrum of triphenylphosphate where the closest protons are 4 bonds away from the phosphorus atom. The lower trace was run under identical conditions with the proton decoupler turned on. In this case, the gain in signal-to-noise using proton decoupling is 17.

Wednesday, October 3, 2007

Proton NMR Assignment Tools - The D2O Shake

You can identify exchangeable protons in your proton NMR spectrum with a very simple technique called a "D2O shake". For some reason this simple technique, used frequently years ago, seems to be used so much less today. All you do is run a 1H NMR spectrum of your sample then put a single drop of D2O in the tube, shake it, and run another spectrum. The exchangeable protons will exchange with the deuterium in the D2O and disappear from the spectrum. The D2O does not have to be miscible with the solvent.

Below is a partial proton spectrum of menthol in CDCl3. The bottom trace is the spectrum before the addition of D2O and the top trace after 1 drop of D2O was shaken with the sample. The -OH peak has clearly disappeared.

Tuesday, October 2, 2007

Improve Your Chemical Shift Resolution Without Going to Higher Fields

Deuterated chloroform is a very inexpensive solvent suitable for many organic compounds, however, the proton spectra of compounds dissolved in CDCl3 often have overlapping peaks complicating the interpretation of the data. One solution is to use a higher magnetic field strength but, what if you are already using the highest available field? One method used a great deal in the past seems to have been forgotten - the use of an aromatic solvent like benzene or pyridine. These solvents induce shifts in the resonances of the solutes often providing dramatically better resolution. Below is an example of the 500 MHz spectrum of tropolone. This compound undergoes fast exchange between the two tautomers indicated in the figure. The bottom trace is the spectrum in chloroform-d and the top trace is spectrum of the same sample in benzene-d6. The exchangable proton is broad and not shown in the figure. The improvement in resolution in the top trace compared to the bottom trace is remarkable.

The next time you are fighting a resolution problem, try using benzene-d6 as your solvent!

Monday, October 1, 2007

Establishing the Lock for Paramagnetic Samples

Paramagnetic samples often have very broad lines and lines at unusual chemical shifts. The paramagnetic solute may also affect the width and frequency of the deuterium resonance signal of the solvent used to establish the lock. For this reason an "autolock" and "autoshim" procedure may not work. In such cases the lock signal must be searched for manually. If you are not sure how to do this, please seek help in the NMR lab.