1H with 19F Decoupling

The proton NMR spectra of compounds containing fluorine often show ¹H - ¹⁹F J coupling. This can of course provide useful information but is also sometimes seen as an unwanted complication. In such cases, it can be removed by ¹⁹F decoupling. An example of this is shown in the figure below. Depending on the NMR equipment used, observing ¹H [¹⁹F] spectra may be more difficult than for example observing protons with ³¹P, ¹³C, ¹¹B or even ²⁷Al decoupling. The difficulty arises because the frequencies of ¹H and ¹⁹F are very close to one another. A probe must be either doubly tuned on the high frequency channel to both ¹⁹F and ¹H or the broadband channel must be able to be tuned as high as ¹⁹F. In either case, additional filters are usually required.

Using the Nyquist Sampling Theorem to Obtain Higher Resolution Solid State NMR Spectra of Spin I = n/2 Quadrupolar Nuclei.

In favorable circumstances, the solid state MAS NMR spectra of spin I = n/2 quadrupolar nuclei can show a strong central transition (free of spinning sidebands) and the satellite transitions in an extended spinning sideband manifold. Although the central transition is very strong, it is often broadened out significantly by the second order quadrupolar interaction. This broadening often presents resolution problems when more than one site is present. The problem can of course be reduced by going to higher magnetic field strengths where the second order interaction is reduced. If this option is not available, it is sometimes advantageous to look at the satellite transitions which can be affected to a lesser extent by the second order quadrupolar interaction and therefore exhibit narrower lines. Such is the case for the first satellite transition for ²⁷Al. The problem is that since the satellite transitions are spread over a very large sideband manifold, any one sideband is of very low intensity. It is desirable to add the intensities of all of the sidebands into the 0^th order sideband. This is conveniently accomplished by taking advantage of the Nyquist sampling theorem and collecting the data in simultaneous mode without digital filtering. If the dwell time is made equal to a single rotor period (i.e. the spectral width is set to one half of the spinning speed) and the analog filter bandwidth is maximized, the sidebands of the satellite transitions will all fold into the 0^th. order sideband. If the magic angle is set very precisely, in the case of the first satellite transition of ²⁷Al, the intensity of this sharper line is greater than that of the central transition. Further, the central transition can be suppressed with a double quantum filter (Ashbrook and Wimperis, Journal of Magnetic Resonance, 177, 44 (2006)) to produce a clean spectrum with a much sharper line than the central transition. This is illustrated for the ²⁷Al MAS NMR spectrum of Al(acac)₃ in the figure below.

Temperature Calibration in an NMR Probe

Often it is desirable to collect NMR data at temperatures other than ambient temperature. Most NMR spectrometers are equipped with a variable temperature accessory. The user sets the desired temperature and the variable temperature unit regulates the temperature by continually adjusting the current in a resistive heater within the probe. The probe heater is inside a dewar into which a gas (air or nitrogen) is directed. The gas must be cooler than the desired temperature. It is heated by the heater and directed over the sample tube. The temperature is typically measured at the bottom of the NMR tube with a thermocouple. The set temperature may be different than the true temperature of the sample in the coil due to thermal losses, poor thermocouple calibration, or undesirable gas flow characteristics. Although the precision of the set temperature is typically 0.1 degrees Celsius, the accuracy (i.e. difference between the set temperature and the true temperature) can be several degrees Celsius. The problem is then to know the true temperature of the sample in the coil. For temperatures above room temperature, this can be done by collecting proton NMR spectra of ethylene glycol as a function of the set temperature. The chemical shift difference between the -OH and methylene protons is linearly dependant on temperature (Stefan Berger and Siegmar Braun, 200 and more NMR Experiments (2004), p. 146) and can be used to construct a calibration plot for the set temperature vs. the true temperature. The figure below shows such data collected as a function of temperature. Methanol shows similar behavior and is commonly used to calibrate temperatures below room temperature.

Backward Linear Prediction

Like forward linear prediction, backward linear prediction uses observed data to predict data which is unavailable. In the case of forward linear prediction, data is predicted at the end of the acquisition time in the observed domain (1D) or used to predict more slices in the indirect dimension of 2D datasets. Backward linear prediction, on the other hand, predicts missing or distorted data back to time zero (immediately after the observe pulse). The data immediately after the pulse may be unavailable or distorted due to a long receiver dead time, pulse breakthrough, or acoustic ringing. Backward linear prediction can recover broad features in a spectrum, solve baseline problems and recover phase information. It should be noted that if a broad signal has completely decayed before the collection of meaningful data, then backward linear prediction will not be able to predict the lost broad feature. An example of backward linear prediction to predict data lost during acoustic ringing is shown below.
Only the initial portion of the FID is shown.

Acoustic Ringing

When a pulse is applied to an NMR probe in a strong magnetic field, the oscillating rf current in the circuit induces mechanical (acoustic) oscillations in metal parts of the probe. These mechanical oscillations in turn generate rf signals detected by the coil. The so called "acoustic ringing" is seen in the FID and (depending on the specific frequency and probe) usually decays within several tens to hundreds of microseconds after the pulse. It is more of a problem at high fields and low frequencies and a particular problem when wide spectral widths (short dwell times) are employed. The acoustic ringing may impede the observation of very broad lines and cause baseline and phasing problems when observing sharp lines with large spectral widths. An example of the latter case is shown in the figure below.

Artifacts due to One-Way Helium Valve Oscillation

Many NMR magnets use a one-way valve on the helium exhaust to permit the helium gas to vent but not allow air to be sucked into the cryostat. Some of these valves have a mechanical oscillation when helium gas is vented. This oscillation can be seen as an artifact in NMR spectra and is more of a problem immediately after a liquid helium fill. This is illustrated in the figure below for the 500 MHz ¹H NMR spectrum of the singlet in ethyl acetate. All three spectra were acquired non-spinning with 1 scan.The left-hand panel shows an artifact-free spectrum acquired immediately before a liquid helium fill with the one-way valve in place. The middle spectrum was acquired immediately after a liquid helium fill with the one-way valve in place. The oscillation "sidebands" are approximately 10 times higher than the ¹³C satellites. For this particular valve on this magnet, the artifacts are 11.5 Hz from the from the NMR signal and disappear slowly over the course of approximately 24 hours. The right-hand spectrum was acquired immediately after a liquid helium fill without the one-way valve in place. Like the left-hand spectrum, it is artifact-free.

Interference from FM Radio Stations

FM radio stations transmit their signals at frequencies close to 100 MHz. Unfortunately this frequency is coincident with the resonance frequencies of several NMR active isotopes at commonly available magnetic field strengths. Under some circumstances, FM radio signals can interfere with NMR measurements. The figures below are solid state ²⁷Al and ⁶⁹Ga QCPMG NMR spectra taken in Windsor, Ontario in a 9.4 Tesla magnet. Both examples show interference from local FM radio stations. In the case of the ⁶⁹Ga, no NMR signals are observed.Thank-you to Joel Tang and Robert Schurko from the Department of Chemistry and Biochemistry at the University of Windsor for providing the data for this post.

Effect of 1H Tuning on the Signal-to-Noise Ratio in 13C NMR Spectra

Most liquid state ¹³C NMR data is collected with ¹H decoupling so that all of the ¹³C resonances (not coupled to nuclei other than protons) will appear as singlets. Since ¹³C is only 1.1% naturally abundant and often the chemist is limited by the quantity of sample, acceptable signal-to-noise ratios in ¹³C NMR spectra are often difficult to obtain in a reasonable period of time. Any hints to improve the signal-to-noise ratio are welcome. One way to ensure that the signal-to-noise ratio is what it should be, is to make sure that the proton channel of the probe is well tuned. A poorly tuned proton channel will lead to incomplete decoupling resulting in broad lines or residual splittings and hence a low signal-to-noise ratio. The reason for this is that when the proton channel is poorly tuned the sample receives less power from the proton channel of the spectrometer and the decoupler pulses in the decoupling scheme (usually WALTZ 16) are no longer calibrated for proper decoupling. This is illustrated in the figure below.For the spectrum in the left-hand panel, the ¹H channel was well tuned. In the right-hand panel, the ¹H channel was poorly tuned. One can see that the spectrum acquired with the poorly tuned ¹H channel has a lower signal-to-noise ratio. There are even more problems with pulse sequences (such as DEPT experiments) which require hard pulses from the ¹H channel. This sort of problem can also be seen in proton detected experiments with poorly tuned ¹³C channels when ¹³C decoupling is required (eg. HMQC/HSQC experiments)

Why are NMR Probes So Expensive?

It is not unusual to pay $50,000 - $120,000 for an NMR probe, depending on the configuration. Cryogenically cooled probes cost much more. Even the simplest NMR probes, which have very simple electronics, are very expensive. Here are some of the reasons why.

1. All of the components (including capacitors, coils, frame, shield, screws, adjustment rods, springs, supports etc...) must be nonmagnetic. These parts are often more expensive.
2. Since the market for NMR probes is quite small, specialty parts are manufactured in small batches making them more expensive.
3. Parts near the coil must be manufactured out of materials which will not give a background NMR signal or have huge magnetic susceptibility differences. These parts are often expensive.
4. Many parts must be machined to very strict tolerances. This is especially true for MAS probes but also true in high resolution liquids probes as well. The cost of very precise machining is very expensive.
5. Some materials which are desirable for NMR probes are very difficult to machine and require specialized expensive tools as well as highly paid and highly skilled machinists. An example of such a material is zirconia, used in many MAS probes.
6. A great deal of research and development on the part of the instrument company is invested in optimizing the design of the probe to improve rf handling, signal-to-noise ratio, lineshape, temperature handling etc... Instrument companies must compete with one another in this regard. These costs must be recovered.
7. Each probe must be individually NMR tested to ensure that it meets all of the specifications. This requires an NMR spectroscopist and an NMR spectrometer, neither of which come cheap.

Please note that I was not prompted by any sales person or instrument company to post this BLOG entry ...... and I do realize that instrument companies make a profit from the sale of NMR probes.

The Consequence of Locking on the Wrong Solvent

Most often, running an NMR spectrum in automation requires only two pieces of information from the user, namely, the NMR experiment to be performed and the deuterated solvent used to dissolve the sample. With such a high demand on the user (this is meant to be scathingly sarcastic), there are bound to be errors in the input. If the user declares the wrong solvent, one of two things may happen. Firstly, the spectrometer may not be able to establish a deuterium lock and will report an error and not run the sample. Secondly, the spectrometer may be able to establish a lock despite the fact that the deuterium signal is off resonance. If the lock is established, the field strength will be set to a value appropriate to put the declared solvent signal on-resonance. When a proton NMR spectrum is collected, the chemical shift scale will be incorrect by an amount equal to the proton chemical shift difference between the true solvent and the declared solvent. This is illustrated in the figure below.In this case the solvent was benzene-d₆. The lower trace shows the spectrum where the solvent was correctly declared as benzene-d₆. The middle trace shows the spectrum collected when the declared solvent was CD₂Cl₂. Note that the spectrum is shifted to lower frequency by (7.16 ppm - 5.32 ppm = 1.84 ppm). The top trace shows the spectrum collected when the declared solvent was DMSO-d₆. Note that the spectrum is shifted to lower frequency by (7.16 ppm - 2.50 ppm = 4.66 ppm). In this spectrum some of the resonances are outside of the spectral width and are lost, as the data were acquired with digital filtering. A similar thing can happen when a solvent with more than one deuterium signal (eg. THF-d₈, methanol-d₄, toluene-d₈ ...) is correctly declared. The spectrometer may lock on the wrong deuterium signal of the solvent. In this case the chemical shift scale of a proton NMR spectrum will be incorrect by an amount equal to the chemical shift difference between the two types of protons in the solvent.

Nyquist Fold-backs and the Mode of Data Acquisition

Some older Bruker NMR spectrometers allow the collection of the two quadrature channels of the FID either simultaneously (where complex pairs of points are collected at the same time) or sequentially (where a real data point is collected then an imaginary point etc...). In the case of simultaneous acquisition, a Nyquist fold-back will fold in at the far side of the spectrum, whereas for sequential data acquisition, it will fold in at the near side of the spectrum. This is illustrated in the figure below. The data were collected without digital filtering.

Thank you to Dr. Michael Lumsden, the NMR Facility Manager at Dalhousie University for suggesting this entry.

17O NMR for H2O vs. D2O

D₂O rather than H₂O is recommended for setting up an NMR spectrometer to observe ¹⁷O. This is because D₂O is isolated based on its higher mass. It has more than the natural abundance of ¹⁷O and in turn gives a larger NMR signal than H₂O which has only the natural abundance of ¹⁷O (0.037 %). The figure below shows the ¹⁷O NMR spectra of H₂O and D₂O acquired on the same volume of sample under identical conditions.

Simultaneous NMR of both 23Na and 51V

When one looks at an NMR spectrum, it is almost always assumed that the spectrum is composed of the resonances of only one nuclide. There are examples where this is not the case. One of these examples is shown below. The spectrum is that of sodium orthovanadate, Na₃VO₄ (which also forms Na₂VO₃(OH) in aqueous solution). The spectrum contains the resonances of both ²³Na and ⁵¹V and shows the ratio between the VO₄^3- and [VO₃(OH)]^2- ions.



This spectrum was acquired on a 300 MHz spectrometer.

Digital Filtering, Nyquist Fold-Backs and Signal-to-Noise Ratio

Modern NMR spectrometers use digital filtering to improve the quality of the data. Digital filtering is achieved in three steps:

1. Despite the spectral width requested by the user, the spectrometer oversamples the FID as if a very large spectral width was requested (i.e. short dwell time).

2. Depending on the requested spectral width, a digital filter is calculated and applied to the oversampled FID by the spectrometer.

3. The digitally filtered oversampled data is decimated according to the originally requested spectral width and then Fourier transformed to produce an NMR spectrum.

These steps are schematically illustrated in the figure below.Note that digital filtering applied as in the figure, will suppress Nyquist fold-back signals but more importantly, it will suppress noise from outside the requested spectral width from folding into the spectrum. This suppression of folded in noise represents a significant improvement in the signal-to-noise ratio compared to a spectrum acquired without the use of digital filters..

Nyquist Fold-Back Signals

The Nyquist sampling theorem states that an FID must be sampled at a rate at least twice the highest frequency in the FID in order to faithfully reproduce the correct frequencies in an NMR spectrum. In the FID, the highest frequency is plus or minus 1/2 the spectral width. If a resonance falls within plus or minus 1/2 the spectral width, it will be correctly represented in the spectrum. In the absence of digital filters, if a resonance is outside of the spectral width but within the analog filter band width of the spectrometer, it will still appear in the spectrum but at the wrong frequency (and often with a different phase than the correctly represented resonances). The figure below shows an example of this.The bottom trace is a properly recorded NMR spectrum. The top trace shows a spectrum of the same sample with the spectral width set smaller than necessary to capture all of the peaks. One can see that the resonance outside of the spectral width by delta f is folded into the other side of the spectrum by delta f. This phenomenon is also observed in the indirect dimension of a 2D data set as well as in magnetic resonance images.

Cross Polarization Using Ramped Pulses

Cross polarization is routinely used for sensitivity improvement when measuring the solid state MAS NMR spectra of spin I = 1/2 nuclei (X). This technique relies on dipolar coupling between ¹H and X as a means to transfer magnetization. When the dipolar coupling is small or averaged due to molecular motion or fast MAS, the Hartman Hahn matching condition required for cross polarization is very sensitive. This sensitivity means that even minor missettings or instrumental instability can lead to a dramatic (or complete) loss in signal. One way of overcoming this sensitivity is to use a ramped contact pulse on either the X or the ¹H channel as in the figure below.
The effect of using such ramped pulses is shown in the figure below where the dipolar coupling is averaged by fast MAS.

One can see that the matching condition is spread out over a broad area and the matching sidebands and troughs are no longer a problem. It is easy to see how using such a scheme makes the CPMAS experiment less susceptible to instrumental instability.

Modes of Heteronuclear Broadband Decoupling

In talking to many novice NMR users over the years, I have often been discouraged to learn that some do not even realize that they are using heteronuclear broadband decoupling when they run (for example) a ¹³C spectrum. They do not question why their signals are singlets or even realize why they may not be singlets if ¹H decoupling is not used. It is for this reason I have added this BLOG post. Heteronuclear broadband decoupling is the irradiation of one nuclide while observing another nuclide. It is normally applied using multiple pulse decoupling schemes such as WALTZ 16 or GARP. These schemes effectively allow the decoupling to be efficient over a range of frequencies (sometimes even the entire chemical shift range!) with a minimum of power. The figure below shows some common modes of heteronuclear decoupling.In the top trace, broadband ¹H decoupling is applied with a 100% duty cycle. This is most common for the collection of ¹³C NMR data. It is applied during the acquisition time to decouple protons and during the recycle delay to take advantage of any nuclear Overhauser enhancement. In the second trace, decoupling is applied only during the acquisition time to decouple protons but not during the recycle delay. This is called inverse gated decoupling and is often used when collecting NMR data for nuclei like ¹⁵N or ²⁹Si which have negative gyromagnetic ratios and hence negative nuclear Overhauser effects. In the third trace the decoupling is applied only during the recycle delay but not during the acquisition time. This is called gated decoupling and is used when a fully coupled spectrum is desired while still taking advantage of any nuclear Overhauser enhancement. The bottom trace shows an example of inverse gated decoupling applied to an X nucleus while observing protons. For examples of this, click here and here.