Information form NMR University Centre, Australian National University

(http://bloch.anu.edu.au/index.html)

All the really basic NMR questions you were too embarrassed to ask!

 

Answers

  • If you haven't used any NMR spectrometers at the NMR Centre before.
    Come down to room G48 or G42 in the Chemistry Faculties building, and ask Chris, Tin or Peta to arrange a training session for you. Even if you have used Varian spectrometers before, our setup here is likely to be slightly different, so in the long run you will save yourself time by having a short lesson.

(Top)

 

 

  • How much solvent volume should I use?
    To get good resolution you need at least 0.7 ml of solvent in a 5 mm NMR tube. If you have a limited amount of sample you can increase its effective concentration by reducing the solvent volume to 0.4 - 0.5 ml however you will need to spend more time shimming. If you have a sample that will give a proton spectrum in 15 minutes when disolved in 0.7 ml, there is little point reducing sample volume if it means you need to spend an extra 10 minutes shimming! On the other hand if you then want to run a carbon spectrum of that sample it would certainly be worth reducing the sample volume. For 13C HMQC or HMBC runs with very small amounts of material, a 4mm tube and rotor is available.

(Top)

 

  • What does the "ADC overflow" error message mean?
    The signal recieved from the NMR sample is first amplified by the reciever and then digitised by the analog to digital converter (ADC). If the signal is too strong for them to handle, either the receiver or ADC will "overflow", causing a RECEIVER OVERFLOW or ADC OVERFLOW message to be displayed. The acquired FID is likely to be clipped, resulting in a distorted spectrum. The solution is to use autogain (type gain='n' or on a Gemini GAIN=N) or to type in a lower value for the receiver gain. If overflow still occurs when the gain is set to zero, reduce the observe pulsewidth (PW) to half its present value. If overflow still occurs dilute your sample, or if the solvent signal is causing the ADC overflow use a solvent suppression technique.

(Top)

 

  • How do I shim / tune the spectrometer?
    First of all, let's get our terminology straight. Shimming is adjusting the resolution of the signal by optimizing the homogeneity of the magnetic field. Tuning is adjusting the impedance of the probe. A poorly tuned probe reflects a lot of the power of the pulses, so that what should be a 90 degree pulse is in reality only (say) a 50 degree pulse. Probe tuning does not affect the resolution, however the signal to noise of a standard spectrum will be worse. Also, experiments such as DEPT or COSY that rely on accurate 90 degree pulses may produce artefacts or not work at all. (Note however that it is possible to adjust the pulse width to give a 90 degree pulse on a poorly tuned probe). Poor shimming on the other hand, results in broad NMR resonances. People often talk about "tuning the resolution" which is where some confusion between shimming and probe tuning arises. Shimming is adjusting the homogeneity of the magnetic field, so that every part of the sample in the NMR tube experiences exactly the same field strength.

(Top)

 

  • OK, so how do I tune the probe?
    If you're using the broadband Gemini or Mercury spectrometers, you never need to tune the probe. The probes of these intruments are tuned at the factory, and further tuning is a specialised operation. For best results, you should tune all other spectrometers before acquiring a spectrum. Frequency, solvent and sample height all affect probe tuning. If you were running a set of similar samples in the same solvent, you might only bother to tune the probe before running the first spectrum. If however, half your samples were disolved in chloroform and half in D2O, you might run all of the chloroform samples and then quickly adjust the tuning after inserting the first D2O sample. Tuning involves setting up for the nucleus of interest and minimizing the reflected power shown on the meter on the magnet leg. Some recabling is required. Do not attempt to do this unless an NMR staff member has given you a lesson. This doesn't mean that Geminis have some great "automatic tuning" technology. It just means they are left in a state of tune that is good enough for the run-of-the-mill experiments they were designed for. On other spectrometers, tuning is necessary because
    • You can get the best possible tuning for your sample,
    • You may not know what nucleus the previous user left the probe tuned to, or whether he/she completely messed up the tuning,
    • More sophisticated experiments such as HMQC, HMBC etc. work best when the probe is tuned and short 90 degree pulses are requested.

(Top)

 

 

  • How do I tune for carbon or phosphorus?
    As mentioned above, if you're using a Gemini spectrometer, you don't need to tune the probe. First, check whether the probe you are using requires a tuning stick to be inserted. Tuning sticks are kept separate from the probe, and have a small capacitor on the end to change the tuning range of the probe. If a tuning stick is required, select the stick for the observe frequency and screw it gently all the way into the probe. You can find the observe frequency by setting up for the nucleus of interest and reading the value of sfrq from the dg display. Then make the cable connections for tuning, and adjust both the tuning and matching rods. These two tuning rods affect each other, so it is usually necessary to go back and forth between them to get a good minimum. There is a bit of a knack to it, so persevere! (Hint: make the tuning worse with one rod, then better with the other. Each dual operation should result in better tuning than before).
    Also note that if you are decoupling protons while observing carbon or phosphorus, it is a good idea to check the proton tuning. If the probe is poorly tuned to protons, some decoupler power may be reflected, resulting in an improperly decoupled spectrum. On the Inova spectrometers you can tune the observe and decoupler channels at the same time. On older spectrometers you need to set up for and tune protons, then set up for and tune carbon or phosphorus. (Hint: always tune the highest frequency first and the lowest frequency last).

(Top)

 

 

  • Which spectrometer should I use for carbon?
    When measuring carbon spectra, the main concern is usually signal to noise. You would expect higher field spectrometers to have a decisive advantage - for example a 500 Mhz spectrometer when compared to a 300 MHz spectrometer should have an advantage of (5/3) squared, or 2.8 times the signal to noise. However there are other considerations, including for example the type of probe. An indirect detection probe has the proton observe coil on the inside (that is, closer to the sample than the coil used for carbon). This improves the proton signal to noise, however if you use an indirect detection probe for directly observing carbon, the signal to noise will of course be worse than a standard probe which has the carbon coil on the inside. Regardless of the probe design, carbon and protons use different coils, and since the electronic circuit for the two nuclei is different it makes no sense to compare proton signal to noise on two instruments and extrapolate the results to carbon.

    Also, signal to noise tests are usually performed by collecting a single scan on a concentrated sample, however this does not give the best indication of the results obtainable on "real" samples where the sample is scanned for several hours. When a sample is repeatedly pulsed, the relaxation times of the various carbons must be taken into consideration. Nuclei take longer to relax at higher fields, so the gain in signal to noise is less than expected. Also note that carbons that do not have directly bonded protons (i.e. carbonyls and quaternaries) have much longer relaxation times than protonated carbons.

    In order to see how some of the spectrometers compare under "real life" conditions, a dilute sample was run for 256 scans on the Inova 500 (PFG indirect detection probe), broadband Gemini, and VXR300. A D1 delay of 1 second with a 45 degree pulse was used, and 16 dummy pulses were given to bring the system to a steady state before starting acquisition. The signal to noise ratios of three resonances were then measured.

    CDCl3
    CH3
    quarternary
    VXR300
    21.3
    16.5
    2.2
    Gemini BB
    16.9
    15.8
    4.3
    Inova 500
    29.9
    22.4
    5.9

    It can be seen that there is not a large difference in the signal to noise you can expect to see on these instruments. Also remember that

    • if there is not much sample available, you should reduce the amount of solvent. A 4-mm tube and rotor is available. This will allow you to use even less solvent than is necessary in a 5mm tube.
    • if you are interested in quarternary carbons, a longer D1 delay of 3 seconds or more is advisable.

(Top)

 

 

  • Why are some of the peaks in my APT missing?
    The APT experiment relies as much on the size of the 1JCH coupling as the number of attached protons to generate the spectral pattern. This is because the delays in the experiment are matched to the inverse of the size of 1JCH. If 1JCH is much larger than the default 1JCH of the experiment (usually set to 140 Hz which is the average of 1JCH for sp3 and sp2 carbons) then peaks will either disappear or appear with incorrect phase. Carbons that may show this behaviour are terminal ethynyl groups (1JCH = 250 Hz approx.), epoxide carbons (1JCH = 175 Hz), furan, pyrone and isoflavone carbons (1JCH = 200 Hz), 2-unsubstituted pyridine and pyrolle carbons (1JCH = 180 Hz) and 2-unsubstituted imidazole and pyrimidine carbons (1JCH > 200 Hz).

(Top)

 

 

  • I can't lock on.
    • You are using a deuterated solvent aren't you?
    • Can you see a lock signal? If not, make sure the lock is turned off, turn the lock power and lock gain to their maximum values, and look for a sine wave by adjusting Z0. If you find a sine wave, adjust Z0 until its frequency becomes zero. Then reduce the lock power (to avoid saturating the lock) and try to lock on.
    • If it loses lock as soon as you try to lock on, turn the lock off and adjust the lock phase as shown in the manual.
    • Is your tube spinning? It might not be spinning because you inserted the tube too quickly, causing it to break. Take the tube out and check that it is in one piece. While you have it out, use a depth gauge to check that the sample is centred in the probe.

(Top)

 

 

  • It won't shim.
    • Check the linewidth of the narrowest line in your spectrum. If there are some broad lines and some narrow lines, the broad lines are probably broad because they are undergoing chemical exchange, not because the resolution is poor. Broad lines may also be caused by quadrupolar broadening if your compound has a transition metal.
    • If you have not already done so, load the standard shims. You don't know what sort of state the previous user left the shims in! On the XL200, type homo and check that the shims agree with the values displayed on the screen. All other spectrometers in the NMR Centre have a macro rtss which loads the standard shims. This macro is equivalent to typing rts('stdshm') su. If you are not using the PFG probe on the Inova 500, you will need to read in a different shim set.
    • If it still won't shim, take the tube out and inspect your sample. Is the tube scratched? Is there anything floating in the sample? Is the sample centred in the coil? If you are using a small amount of solvent to improve the concentration, you may need to add some more solvent to make it easier to shim.
    • Do you have paramagnetic ions in your sample?
    • Have other people been getting poor resolution? If so, report it to a member of the NMR staff. If not, change NMR tubes, filter your sample, and try again. If changing tubes solves the problem, throw the old tube away.
    • Have you placed your NMR tube in an oven to dry? If so, throw the tube away as it has distorted. (Remember, glass is a liquid. It flows at high temperature). The correct way to dry a tube is via a stream of dry nitrogen through a glass wool filter.

(Top)

 

 

  • My tube broke when I inserted it into the magnet.
    During the day, phone an NMR staff member. Tell them the solvent and any hazards posed by your compound. After normal working hours tell the watchmen who will call in someone. Leave a note to warn others not to use the spectrometer. Remember - the more quickly you lower a tube into the magnet, the more likely it is to break! If necessary, use two hands on the sample eject button to make it easier to lower the tube slowly.

(Top)

 

 

  • I can't phase correct my spectrum.
    The aph (automatic phase correction) command usually does a good job of correcting the phase, and should be the first thing you try. Sometimes (for example in noisy spectra) the aph command is unable to correct the phase, and in these situations it often leaves lp at a high value (say one or two thousand). In these situations you will have to correct the phase manually. First a couple of obvious things: if you ran a DEPT or APT experiment or something similar, there will be some positive and some negative peaks, so don't try and phase them all positive! Similarly in a 1:1 binomial solvent suppression sequence, half the spectrum will be positive and half negative.
    Having established that you are not running an exotic pulse sequence that produces strange phases, the next thing to consider is foldback. Are you sure that you used a large enough spectral width when acquiring the spectrum? If one or more resonances occurred outside the observe region, the method used to digitise the signal results in these resonances appearing within the observed spectral width, but with a phase error. If in doubt, double or triple the spectral width, run the spectrum again, and see if the resonance that could not be phase corrected remains at the same chemical shift as before.
    To perform manual phase correction, proceed as follows:
    Type lp=0 rp=0. This sets the left phase and right phase to zero. On Varian spectrometers, "right phase" is the zero-order phase adjustment and "left phase" is the first-order phase adjustment. The zero-order phase affects the entire spectrum equally, while the first-order phase is frequency dependent. The zero-order phase should always be in the range -360° to +360° and the first order phase should also usually be in this range. If you have a first order phase correction of more than a thousand degrees, not only is it probably incorrect, but you will also probably be generating baseline roll. On Gemini or XL spectrometers, type QP to get into quick phase mode. On Sun based spectrometers or data stations, click the "phase" button with the mouse. Perform a zero-order phase correction on the largest peak as described in the manual for the spectrometer you are using. Now choose another peak some distance from the largest peak, and adjust the first-order phase. On Sun based systems you only get one shot at adjusting the zero-order phase - all subsequent corrections are made to the first-order phase, so there is no point clicking on the largest peak again. If you want to readjust the zero-order phase, get out of the phase-correction routine (by for example typing ds) then click on the phase button again. If for some reason a large first-order phase correction is required, it may be easier to choose a peak for the first-order adjustment that is close to the peak you used for the zero-order adjustment. On Sun based systems, set the phasing parameter to 100. This causes the effect of the phase values to be shown for the entire spectrum as you are making the adjustments, thus making it easier to see what you are doing.

(Top)

 

 

  • I need to run my spectrum at a higher field to get better resolution.
    No you don't! The resolution of a high field spectrometer may even be worse than a low field spectrometer. What a high field instrument has more of is dispersion. This means that resonances with different chemical shifts are further apart. Multiplets due to coupling will not show any improvement unless the higher field instrument separates overlapping multiplets with different chemical shifts, or the multiplet showed strong coupling effects at lower field. Some nuclei such as 31P may have worse resolution because of a property called chemical shift anisotropy which increases with field strength.

(Top)

 

 

  • There are no parameters for the solvent I want to use.
    If you're running a proton spectrum, set up for 1H / CDCl3, double the spectral width, run a quick spectrum, and put the two cursors around the spectrum. Then do a movesw and acquire the final spectrum. If you're running a carbon spectrum, set up for 13C / CDCl3, increase the spectral width by 20 percent, and run as normal. If your solvent has carbon nuclei which show up quickly, reference the solvent and check that the observed spectral range is correct.
    If you are running a phosphorus spectrum, set up for 31P / CDCl3, increase the spectral width by 20 percent, and run as normal. Then supply an NMR tube containing the solvent to the NMR staff so that they can set up H3PO4 referencing parameters for you.

(Top)

 

 

  • How can I suppress a strong solvent resonance in a proton spectrum?
    If the solvent signal is less than two or three times the size of the largest signal from your compound, it may not be worth bothering. On the Geminis, XL200 and VXR300, the usual method is to presaturate the solvent signal using the decoupler (instructions for doing this are in the folders near the spectrometers). Although it is simple, this method has the disadvantage that NH or OH protons that are exchanging with water also have their signals reduced or eliminated. Another method is the 1:1 binomial pulse sequence. The signals on one side of the solvent resonance are of opposite phase to the other side when this method is used. On the Inova spectrometers, the method of choice is watergate solvent suppression. Simply set the observe transmitter on the solvent position, type wgate and acquire a spectrum. The watergate sequence set up by the wgate macro uses hard pulses and therefore does not require pulse phases etc. to be optimised. The other watergate technique available uses shaped pulses. Simply type autowatergate, and wait while it automatically optimises the parameters and runs a final spectrum. Watergate is only available on the Inovas because it uses pulsed field gradients. If chemical exchange is very rapid, watergate may not be suitable, in which case a binomial pulse sequence is the best choice.

(Top)

 

 

  • What is nuclear spin?
    All nuclei carry a charge. In some nuclei this charge "spins", causing the nucleus to behave like a tiny bar magnet. This is why it aligns with or against the magnetic field of an NMR spectrometer. However unlike a bar magnet, the low energy state is aligned with the field and the high energy state is aligned against the field. Up to now we have been talking about nuclei with a uniform spherical charge distribution. These nuclei are said to have a spin of ½.  Protons, 13C and 31P are all spin half nuclei. Note that the most common isotope of carbon, 12C, has no spin and can therefore not be observed using NMR. Nuclei with a non-spherical charge distribution have a spin number I of 1, 3/2 or higher (in steps of ½ ), and are referred to as quadrupolar nuclei. Spin ½   nuclei have two orientations (with or against the field). Spin 1 nuclei have three orientations, spin 3/2 nuclei have 4 orientations, etc. Deuterium is an example of a spin 1 nucleus. Although deuterium is chemically the same as hydrogen, for the purposes of NMR it is completely different. For example a carbon spectrum of CDCl3 is a 1:1:1 triplet regardless of whether you turn on the proton decoupler. This is because the deuterium attached to the carbon can have three orientations, and occurs at a different frequency to protons.

(Top)

 

 

  • What is a double quantum coherence?
    When you put your sample in the magnet, all the spin half nuclei align either with or against the magnetic field. The population difference between these two orientations (known as the Boltzman distribution) is field dependent, and is determined by their energy difference. An NMR signal is observed when nuclei flip from one orientation to the other. This is a single quantum coherence. When two nuclei are coupled, they can flip together as though they were a single unit. If they flip in opposite directions, the flips "cancel each other out" (sort of) resulting in a zero quantum coherence. If they both flip the same way, you get a double quantum coherence. The frequency of a zero quantum coherence is between zero and a few kilohertz, so it is not directly observed. Similarly the frequency of a double quantum coherence is roughly twice the normal observe frequency, so that is not observed directly either. You can also have triple quantum coherences from groups of three coupled nuclei. The effect of double and triple quantum coherences can only be observed by inserting pulses or delays into a pulse sequence to convert them to single quantum coherences before acquisition of the NMR signal. Do not confuse double quantum coherences with coupling in a normal spectrum. A doublet for example, arises when there are two coupled spins, but only one of these spins flips.

(Top)

 

 

  • What are pulsed field gradients?
    Imagine if you could really mess up the Z1 resolution for a few milliseconds then restore it to its proper value during the course of the pulse sequence. This is an oversimplification, since pulsed field gradients do not use the normal shim circuits. A special PFG probe, and a PFG amplifier are necessary. By applying a gradient to the magnetic field, the top of the sample experiences a slightly different magnetic field to the bottom of the sample. Since magnetisation precesses at different rates in different fields, it is possible after a 90 degree pulse and a PFG of a few milliseconds to have the magnetisation vectors along the length of the tube pointing in all directions instead of nicely aligned along one axis of the rotating frame. Obviously if the magnetisation vectors are pointing in all directions, there is no net signal. The vectors are said to be dephased. If you now apply a PFG of opposite sign for the same time, you will rephase the magnetisation, and get your signal back. You could achieve the same thing by giving the dephased vectors a 180 degree pulse, then applying a PFG of the same sign. The other thing to be aware of is that double quantum coherences dephase at twice the rate of normal single quantum coherences, so by adjusting the strength or duration of pulsed field gradients, you can select single, double or triple quantum coherences. The "old fashioned" way of selecting certain types of coherences is to use elaborate phase cycles which cause the unwanted signals to cancel out on successive scans. The PFG method acquires only the desired signal on each scan, resulting in fewer artifacts and allowing fewer scans. The old method can be thought of as "cancellation of unwanted signals over time" whereas the PFG method can be thought of as "cancellation of unwanted signals over space" where "time" refers to successive scans, and "space" refers to the physical length of the sample in an NMR tube.

(Top)

 

 

  • What is the Nuclear Overhauser Effect?
    Glad you asked. Have a look at our (see NOE in C-13 NMR part)

(Top)

 

 

  • How do I run a quantitative spectrum?
    A quantitative spectrum is simply a spectrum where you can trust the integral ratios. In other words, if the integral of resonance A is twice the height of the integral of resonance B, you can say with certainty that resonance A is due to twice the number of nuclei as resonance B. Why do we use integrals? Because it is the area of the resonances that is proportional to the number nuclei. The height of a broad line may be less than that of a sharp line, but its area may be greater. How do we get accurate integrals? By ensuring that all resonances are equally excited, well digitised, and properly relaxed.
    • Equally excited : if the pulse power is not high enough, some resonances far from the observe frequency may experience a reduced flip angle, resulting in a smaller observed signal.
    • Well digitised : if the number of data points in the spectrum is too low, there will not be enough points to accurately define each resonance, resulting in inaccurate integrals (and peak heights).
    • Properly relaxed : resonances that are not fully relaxed give a weaker signal than fully relaxed resonances. The nuclei in your compound will not all relax at the same rate, so if you pulse too rapidly the quickly relaxing resonances will appear stronger than the slowly relaxing ones. To be sure of obtaining accurate integrals, you need to measure the relaxation times of your compound, and set a delay equal to 5 times the longest relaxation time. Fortunately it is easy to run an inversion - recovery experiment to measure relaxation times.
    It is harder to obtain quantitative carbon spectra, because carbon relaxes more slowly than protons, is less intense, and steps have to be taken to eliminate the Nuclear Overhauser Effect which builds up when protons are decoupled.

(Top)

 

 

  • What is digital resolution?
    Digital resolution is simply the separation in hertz between each data point in your spectrum. It has nothing to do with shimming! Say, for example, you set the number of points np to 32,768 and acquire a normal 1 dimensional FID. The number of points in the spectrum you see will be 16384, since half the data points are imaginary. Now if the spectral width (sw) is 6000, the digital resolution will be 6000/16384, or 0.366 hz per point. (Before you grab your calculator to measure your own digital resolution, note that the number of points in the spectrum is not always simply np/2. See the section below on the Fourier number). The Vnmr command to display the digital resolution is dres. If you place the cursor on a peak and type dres, two values will be displayed:
    • the linewidth which is the width of the peak at half-height, and depends on shimming, weighting functions and the natural width of the line. Also the
    • digital resolution, which is what this section is all about.
    The dres command may give a different linewidth value for every peak you put the cursor on, but the digital resolution value will always be the same, unless you change the Fourier number fn and do another Fourier transform. If the natural linewidth of a resonance is comparable to the digital resolution, the resonance may only be defined by one or two data points. If you expand a line like this, it will look more like a spike than a proper Lorentzian line. Consequently the height of the line may appear less then it really is, the integral will be inaccurate, and even the chemical shift value will be less accurate than it should be. Also, if the separation between two resonances is comparable to the digital resolution, they may appear as a single resonance in the spectrum, because no data point falls in the space between the tops of the two peaks.

(Top)

 

  • What is the fourier number?
    Mathematicians can do a Fourier transform of any number of points. NMR spectrometers speed things up by using the Cooley-Tukey fast fourier transform algorithm. As implemented on NMR spectrometers, this requires the number of points to be a power of two. So what happens if the number of points np is not a power of two? On Varian spectrometers this can be controlled by the Fourier number (fn) parameter. If it is used, fn can only be set to powers of 2, and the value of fn is the number of points that are actually used in the Fourier transform. If fn is less than np, some points on the end of the FID are not used in the Fourier transform. If fn is greater than np, the end of the FID is padded with zeros to increase the number of points. This is referred to as zero filling. Zero filling does not write extra zeros on to the end of the FID file on the disk where the FID is stored, it merely adds the zeros in memory just before doing the transform. It is also possible to set the Fourier number to n (not used). In this case, the spectrometer uses the first power of 2 which is higher than np when doing the Fourier transform. So for example if np was 16385 (that is, 214 + 1) it would use 32768 (i.e. 215) points for the Fourier transform.

(Top)

 

 

  • What is the relaxation time?
    It would be an oversimplification to say that the relaxation time is the time taken for a nucleus to relax to equilibrium. After a pulse, a nucleus relaxes toward its equilibrium value at an exponential rate. The value quoted as the relaxation time is actually the time constant of this exponential curve. It takes five time constants for the magnetisation to relax to 95% of its equilibrium value. There are two basic types of relaxation, T1 and T2. In the T1 process, the magnetization remaining along the z-axis relaxes back to its equilibrium value. This is also known as spin-lattice relaxation because relaxation occurs by the loss of energy from the excited nuclear spins to the surrounding molecular lattice. In the T2 process, the magnetization in the x-y plane fans out out until the net magnetization is zero. This is also known as spin-spin relaxation because it is due to the excited spins exchanging energy with each other.

(Top)

 

 

  • What NMR Simulation Programs are Available?
    • To simulate a normal (non-exchanging) spin system, you can perform the simulation using the same Vnmr program that you use for data processing. There are instructions in the folders. The first step is to decide what sort of spin system you have - AB, A2X, ABCXY etc. The letters are not important to Vnmr, so it doesn't matter whether you tell it that you have an ABC or an AMX system. Vnmr only needs to know the values of the chemical shifts and coupling constants.
    • To simulate a dynamic (exchanging) spin system, the program you use depends on the type of experiment you ran. If you ran a series of normal spectra at different temperatures, then you need to simulate the lineshape. This is done using the DNMR5 program. There is another program, dnmr5input, to help you create the input file for DNMR5. Instructions are in the folders by the Sun computers.
      If you ran a series of
      selective inversion pulse -- delay -- hard pulse -- acquire
      experiments to use magnetization transfer information to determine rate constants, we have a program provided by Prof. Brian Mann of Sheffield University that you can use to analyse your data.

(Top)