Lock

Even the best magnets have slight field drifts (~0.25 Hz/hour).

Since chemical shifts are field dependent, the field/frequency ratio of the particular
spectrometer must be constant during measurement.

=> LOCK CHANNEL

Modern spectrometers use a heterolock system (detect proton, use another nucleus- a
heteronucleus), this nucleus is almost always deuterium.
2
H is spin 1, resonating at ~76 MHz on a 500 MHz spectrometer for proton.

When the sample is locked, the spectrometer records the deuterium reference frequency
constantly to offset small changes in the magnetic field.

The deuterium signal is normally the deuterium in the solvent. Generally, samples are
dissolved in deuterated solvents. Mixed protonated/deuterated solvents are also
acceptable for the purpose of locking (commonly NMR data on macromolecules is
acquired in 90% H2O/10% D2O)

Spectra can be acquired without lock (in 100% protonated solvent) but field drifts cannot
be accounted for. If there are even slight changes in the magnetic field (such as slight
temperature changes), acquiring unlocked spectra will not work well. Even without
environmental changes, magnets will drift. Over time, the magnetic field gradually
decreases. Good magnets drift at rates of ~0.1 Hz/hour. Normally, the field drift
decreases over time; for the first few weeks after installation of a magnet, the drift might
be 1-2 Hz/hour, but after a few months be down to 0.5 Hz/hr. The drift of the magnet will
cause the frequency of the resonances to apparently shift over time.
 Optimizing the Lock Signal
The lock signal is a constant observation of the deuterium signal of the sample in
dispersion mode (to be discussed later).

The lock signal has 4 adjustable parameters:

1) Center of the Field (Z0)

The center of the magnetic field is very dependent upon the solvent used. The frequency
of Z0 is adjusted so that the lock frequency of that solvent is the center, at its maximum.
There is correlation between the center of the field and the shims

2) Lock Gain
The lock gain affects the amplification of the lock signal. There is no actual effect on the
sample when lock gain is changed. If lock gain were high, the signal-noise of the lock
would be low; if lock gain were too low, the amount of lock signal would be too low.

3) Lock Power
Lock power affects the amount of RF power used to stimulate a deuterium signal. Like
the lock gain, the lock signal goes up and down upon change of the lock power. The lock
power will increase the signal to noise of the lock as more power stimulates more signal.
However, there is a limit to how high lock power can be set. In general, the lock power
can be increased as long as the lock level increases with increasing lock power. At some
point, more power input into the sample causes no increase in lock level. This
phenomenon is referred to as saturation of the lock. Until saturation, the lock power
stimulates relaxation processes of the sample. Additional lock power after saturation will
not increase the lock level, but actually cause it to fluctuate or even decrease. If the lock
power is too high, additional power will be dissipated as heat, which in turn heats the
sample. Chemical shifts, shimming, tuning, relaxation, etc. are all temperature dependent.

4) Lock Phase
The lock phase is the phase of the lock signal received by the console. The phase of the
lock signal is normally optimized to give a maximum lock signal. The phase allows for
the lock to signal to be completely on resonance. By adjusting the lock phase on the
signal received, the shims (to be discussed shortly) affect the signal observed and thus the
setting of the lock phase. The optimal adjustment for lock phase is so the lock signal is
minimum amplitude at the center of the field. However, on a Varian display, the signal is
seemingly at a maximum with optimal lock phase.
Diagram of Lock Circuit
 Shimming
Shims are small magnetic fields to cancel out errors in the static magnetic field. An 11.7
Tesla (500 MHz) magnet does not have a perfectly homogenous magnetic field, and
small differences (1 in 107) in 500 MHz is at the level of 0.1-1.0 Hz which is certainly
relevant.

Ideally, the magnetic field is adjusted to be most sensitive around the small volume of the
sample inserted into the probe.

The shims then adjust the magnetic field to be most sensitive, most homogeneous. The
practical result is sharper resonances and resonances that are more symmetrical,
Lorentzian in line shape.

Difficulty of adjusting the field increases with increasing magnetic field, increasing
resonance frequency of a nucleus and increasing bore of the magnet. Thus, proton on a
wide-bore 900 MHz would be the most difficult scenario currently.

Shims are physically printed coils wrapped around a cylinder inserted into the magnet.

There are two levels of shims: cryo shims and room temperature shims

Cryo-shims are superconducting shim coils that are adjusted upon installation of the
magnet.

Room temperature shims are the shims to be adjusted by the user.

Room temperature shims can be broken down into two categories: those aligned with the
vertical axis of the magnetic field (Z1, Z2, Z3, Z4, Z5, Z6, Z7, Z8) and those aligned with
the horizontal axes of the magnetic field (X1, Y1, XZ, YZ, XY, ZXY, X2Y, Y2X, X3,
Y3, X2Y2...)
 Parameters Affecting Shimming

Or why the shims are terrible when you take over the instrument! It is rarely the
instrument's fault that your shimming is terrible!

1) Probe- Shimming is VERY probe dependent, even higher order shims- X3Z2 etc.
ALWAYS load a shim file for the proper probe.

2) Solvent- Dielectric constant, viscosity, etc., have large effects on shimming, mostly
lower order shims (Z1, Z2 etc.).

3) Temperature- Change temperature 20°, change the shims. Particularly, Z2, X, Y, Z1,
XZ, YZ are affected.

4) Sample- Various reason for sample effects, including concentration, paramagnetics,

residual impurities, especially ions, particulate matter, fuzz, dust, etc....

5) Sample Height- The center of the field obviously shifts with the height of the sample,
so the shimming changes as a direct correlation with sample height (one reason there is a
depth gauge).

6) NMR Tube- Different tubes, different thickness of glass, quality of glass, scratches on
the glass.
 Axial Shims
The Z shims are aligned with the Z axis of the magnetic field.

Z shims are commonly referred to as the Spinning Shims (Spinning will be discussed
later) or Axial shims (aligned with the long axis of the magnet and magnetic field)

The number in the shim refers to a power, e.g. Z1 = Z1, Z2 = Z2 ...

Thus, each shim has a characteristic shape, Z1 is linear, Z2 is a parabola ... The curves
that the shim gradients have can be used to optimize them automatically using gradients

The amount of shim coils is different for different magnets, usually higher field magnets
have more shim coils than lower field magnets, because higher field magnets are more
sensitive to small changes
 Radial Shims and Spinning

Radial shims are any shim with an X or Y in them as they are aligned with the X or Y
axes

Radial shims can be somewhat averaged out by spinning the sample

Spinning
For 1D data, samples are spun in the magnet by use of a small air turbine (the spinner)
with compressed air

Spinning improves the field homogeneity because the nuclei see an average of the
magnetic field, not the static inhomogeneous magnetic field

Samples are normally spun at a rate of ~20 Hz

Spinning produces spinning side bands, the side bands are on both sides of each
resonance at a frequency difference corresponding to the speed of spinning. Thus,
changing spin rate moves the side bands.

Radial shims are approximately averaged out by spinning, axial shims are not.
 Measuring Shimming
Ideal shimming means perfect line shape and narrow line width, but to acquire a
spectrum each time you change a shim is impractical and time consuming.

Thus, shimming on the instrument is often accomplished by measuring something other

than the spectrum each time you change a shim- usually the deuterium lock signal is
used; the other option is using the Free Induction Decay (FID) (the FID will be discussed
later), but the lock signal is easier.

The lock signal or lock meter is essentially a measure of the intensity of the deuterium
solvent resonance.

Sharper line, better line shape => more intense signal

Intensity of lock signal is an indication of a sharper deuterium line or a better line shape
or both; since the goal is good line shape and narrow line width, lock signal is an
indication of shimming.

Thus, the easiest way to shim is adjust a shim and watch the lock level go up or down;
keep adjusting until that shim is optimized.

Additionally, watching the lock level is faster than acquiring a proton spectrum or
shimming on the FID as the relaxation of deuterium is faster than that of proton
(relaxation will be discussed later). If relaxation is slow, when a shim is changed, the
response to the change of shims will be slow compared with how fast you can change the
shim value and how fast the NMR console can convert that change to a change in
magnetic field of the shim. Thus, it is preferable to follow the lock signal than the proton
spectrum initially as the change of shim value by the user is nearly in real time to the
response by the lock level. However, even the deuterium signal of some solvents has long
enough relaxation that the response to a change in shims is not instantaneous (such as
methanol or acetone, most notably of standard solvents).

After adjusting shims on the lock, acquire a 1 scan 1H 1D, and study a sharp, singlet
signal- TMS or TSP are good; solvent signals are also often good choices. The singlet
should be a single peak, without humps, symmetric, and as narrow as possible at half-
height (~0.3 Hz is often good on the 500 MHz instrument).
 Adjusting Shims
In general:

1) The lower the number, the more important the shim is for a good spectrum (Z1 most
important, Z8 the least important).

2) The lower the number the more sensitive it is, changing Z1 by 32 units has a much
greater effect on the homogeneity of the magnetic field on the sample than changing Z5
by 32 units.

3) If changing a shim by a large amount has little effect, then that shim is far from
optimal. Shims should be most sensitive when they are closest to correct.

4) Each shim is not completely independent of all other shims. Optimizing one causes
another to become something far from ideal. Some pairs are much more dependent upon
each other than others (Z1, Z2 particularly).
 Adjusting the Axial Shims
In general, odd numbered axial shims affect line width, even numbered axial shims affect
line shape
 Axial Shims on the 500
TMS, EthylBenzene in CDCl3 (Resonance shown is TMS)

Well-Shimmed, Starting Point (0.26 Hz) Z1 (+16) Z1 (-16)

Z1 (-64) Z1 (-64), Reshimmed Z2, Z3, Z4 on lock

Lock Phase (+64), Lock Level dropped 30% Lock Phase (+64), reshimmed on lock

Z2 (-16) Z2 (-64) Z2 (+64)

Z4 (+64) Z4 (-128) Z4 (+300)

Z4 (+300) Reshimmed Z1, Z2, Z3 on lock Reshimmed Z4 on lock

 Adjusting the Radial Shims
1) Radial shims can only be adjusted with the sample NOT spinning because they are
averaged out by spinning

2) When spinning, poorly adjusted low order radial shims cause spinning sidebands
First order radial shims (X, Y, XZ, YZ) cause 1st order side bands
Second Order radial shims (XY, X2, Y2) cause 2nd order side bands

Non-Spinning
 Radial Shims on the 500

Well Shimmed Spinning Non-Spinning

X and Y slightly off Spinning Non-Spinning

X and Y far off Spinning

3) High order radial shims (X3, Y3...) cause broadening, even for spinning samples

X3 Y3 off non-spinning spinning

X3 and Y3 far off, non-spinning spinning

(Lock 25% of normal) (Lock ~80% of normal)

4) Low order radial shims (X, Y, XZ, YZ) are sample (solvent) dependent and very
temperature dependent.

5) High order radial shims are mostly only probe dependent, should not normally need to
be adjusted except when changing probe (which is why there are saved shim files for
probes).
 Pulsed Field Gradients

Pulsed- A short time period

Field- The B0 magnetic field
Gradient- A variation in some quantity with respect to another.

Pulsed Field Gradient = A short change in the magnetic field with respect to distance.
Normally applied in the z-direction although x and y pulsed field gradients exist as well.

PFGs have a time, a power, and a shape. The time is usually 100's of µs or a few ms, (gt1,
gt2 etc. is normally the parameter name on Varian spectrometers for gradient time 1, 2,
etc.). The power, reported in the literature in Gauss/cm, is converted to a DAC unit; the
conversion is spectrometer dependent (different hardware = different DAC to Gauss/cm
conversion). The shape is the shape of the pulse, to be discussed later.

For PFGs to be used, the instrument must have a gradient amplifier that creates the
pulses, a connection to the probe (in the case of a Varian, the current for the PFGs is
transmitted through the upper stack of the probe), and a probe with gradient coils. Not all
probes made for instruments have gradient coils, as the gradient coils usually limit the
temperature range of a probe (normally gradient coils cannot handle low temperature).

On recent NMR instruments (post ~1995), this is normal hardware; the routine
application of PFGs is easily the most significant development in the NMR field in the
last ~15 years. Few NMR experiments developed in the last 5-10 years do apply PFGs.
Nearly all previously developed NMR experiments have now been rewritten with PFGs.

The control of when PFGs are turned on and off are controlled by the software (pulse
sequence); Varian has an additional software override parameter (pfgon for pulsed field
gradient on, which is set to 'nny' for z-gradient, 'yyy' for x, y, and z-gradients).
 Gradient Shimming

Shims are required to eliminate inhomogeneity of the magnetic field, as a magnet does
not have a perfectly homogenous magnetic field throughout the magnet. Thus, different
parts of the sample will see a slightly different field. The shims of the magnet are
supposed to make up for this problem. Gradient shimming uses magnetic resonance
imaging to get an image of the magnetic field, then adjust shims according to the image.
Since the purpose of shims is to make the field equal in all parts of the sample and the
goal of imaging is to observe a picture of the sample, combine the two and an image of
the magnetic field can show where the shims need to be adjusted.

All spins that contribute to a single line should have the same Larmor frequency except
that bad shimming causes slight differences- this is then translated to phase difference.
Gradients have the ability to separate out these differences as a function of distance from
the top (or bottom) of the sample for z-axis gradient shimming. In principle, x and y-axis
gradient shimming should work, but Varian spectrometers cannot do that yet.

First, an image of the probe is acquired, a reference map essentially, the gradient
shimming is done by trying to optimize response in a gradient echo experiment.

Errors in the local magnetic field cause phase differences between protons with
equivalent Larmor frequencies as a function of time in the x-y plane and distance in the
sample tube:
Gradient echo experiments are used to provide the spatial encoding of the phase errors as
a function of τ.

1D image of the sample:

2 minutes of gradient shimming is amazing!:
 Practical Aspects of Gradient Shimming

Gradient shimming uses pulsed field gradients or homospoil gradients to image each of
the Z shims (Z1, Z2, Z3, Z4, Z5, Z6) and compare to a map of the magnetic field
(shimmap). Gradient shimming can be an efficient way of achieving good shims, but can
easily result in poorer shims if done incorrectly. Gradient shimming can be done on either
proton or deuterium and pulsed field gradients or homospoil gradients. Unless your
sample is in protonated solvent or you have alot of sample, you are better off using
deuterium. Pulsed field gradients should be used on the probes with them. All probes for
the Mercury 400, Inova 500 and Inova 600 have gradients; none of the other Chemistry
Department instruments has the ability to accomplish gradient shimming. The gradient
autoshimming routine takes approximately a few minutes if there is an adequate map of
the field (shimmap). The shimmap can take as much as 3-5 minutes to complete. The first
time you use gradient shimming, you will need to have your own shimmap and your own
shimmap directory. Shimmaps are probe and instrument dependent, but only loosely
related to the sample. Some types of tubes are different enough that a shimmap for that
tube is required. Although solvents have some effect, normally acquiring a shimmap for a
particular solvent is not required. Standard shimmaps can be found in the
/vnmr/gshimlib/shimmaps for the Chemistry Department Inova 600, Inova 500 and
Mercury 400.

First the shimmap is created. Each shim a characteristic profile Z1 = line; Z2 = parabola
etc. Ideally, all shims should be easily identifiable with no visible noise.
Next, a deuterium spectrum is acquired with a varied delay. The second spectrum should
be lower in intensity than the first. The spectrum is essentially a 1D NMR image of the
deuterium in the sample.

The computer compares the image to the map of the shims, then adjust shims
accordingly.

The data is plotted as a frequency versus field plot. The sample is lying horizontally
across the plot in the center. Ideally, the plot should be flat across. The computer will
continue to acquire spectra until the fitted plot is nearly linear.
How to Gradient Shim Using Pulsed Field Gradients (PFG Probes)

To accomplish gradient shimming for the first time:

Turn spinner off. (the macro should turn spin off anyways). Note the new lock level after
turning off the spinner. If the lock drops alot, shim quickly X1, Y1, XZ, YZ (how much
is a standard drop spin to non-spin is solvent dependent, and instrument dependent; on
the 500, expect ~5-10% drop in CDCl3, on the 400, expect 10-15% in CDCl3).

type in vnmr window: gmapsys

lots of things will happen, parameters will change, messages will appear... After the
computer seems to have settled, click on Shimmaps button. Click on Shimmap files
button. Click on CD to System directory button, should be on the far left side. If you do
not see that button click on CD to User Directory. Then, you should see CD to System
directory button, click on it. Highlight file that starts with asw on Merc400 or PFG_SW
on the 500, click on Load Shimmap and Parameters. Click on Return button.

If you are on the 400: Set nt = 8 for CDCl3, or nt = 2 for any solvent with alot of
deuterium (benzene, D2O, acetone, DMSO, methanol). For methanol or acetone, set d1 =
5. Then, click on Autoshim on Z button. The gradient shimming should finish in 2-4
iterations; the lock should have increased after the gradient shimming is done. If for some
reason the gradient shimming does not succeed, abort acquisition (this rarely happens
now); if this does occur check the lock level, if it is lower than when you started, click on
Set Shims button, then click on Starting Shims button as gradient shimming failed and
probably made the shims worse.

If you are on the 500: Set nt = 16 for CDCl3, or nt = 4 for any solvent with alot of
deuterium (benzene, D2O, acetone, DMSO, methanol). For methanol or acetone, set d1 =
5. Set pw = 800 for any solvent. Set gzsize = 4 (shims Z1-Z4). Then, click on Autoshim
on Z button. The gradient shimming should finish in 2-4 iterations; the lock should have
increased after the gradient shimming is done. If for some reason the gradient shimming
does not succeed, abort acquisition (this rarely happens now); if this does occur check the
lock level, if it is lower than when you started, click on Set Shims button, then click on
Starting Shims button. After that finishes, you could go back and set gzsize = 5 (shims
Z1-Z5) and redo the Autoshim on Z (most of the time this helps, but sometimes, the
gradient shimming fails for gzsize = 5).
 Solvents

How to choose a solvent:

1) Solubility- The most important criteria is solubility; higher concentration of sample

leads to less signal averaging necessary, faster acquisition of data. Ideally, solubility
should be to ~millimolar concentrations.

2) Deuteration- Solvents need to have at lest 10% deuterium, preferably 100% deuterium;
thus, choose a solvent that is less expensive for deuterated solvent (although you could
add 5% Deuterium of one solvent to mix with another protonated solvent, the protonated
solvents resonances would have to be suppressed).

3) Temperature- Obviously choose a solvent that is a liquid at the temperature that you
want to work at.

4) Solvent Resonances- Even in supposedly 100% deuterated solvent, residual protonated

solvent remains as well as dissolved water resonances. Choose a solvent that does not
resonances that overlap with the resonances of the sample being studied.

5) Structural Relevance- If you are using NMR for three-dimensional structure

determination, choose a solvent that the structure is relevant in. (Don't study a molecule
that is only active in high salt in CDCl3.)

6) Standard Sample- Setup (Shimming, tuning) of a spectrometer is set to a specific

sample, using the same solvent or at least a similar solvent will help as shimming and
tuning will be closer to correct on your sample (the standard in all of our current
spectrometers is CDCl3).

7) Viscosity- Viscous solvents cause slower tumbling, and thus broader resonances.
Rough viscosity of common solvents at room temperature:
viscous (bad): DMSO, pyridine, water
moderately viscous: benzene, chloroform, toluene
non-viscous (good): acetone, acetonitrile, methanol

8) Exchangeable Protons- Solvents with exchangeable protons will exchange with the
sample. If the solvent is 100% deuterium, then nearly all exchangeable protons in the
sample will become deuterium. Even in 90% protonated solvent, the exchangeable
protons in a sample can be broadened due to exchange with solvent.

9) Chemical Shifts- Chemical shifts can change from one solvent to another. Sometimes
this is good, by removing overlap, other times bad, creating overlap or simply not
correlating well with previous data.
 Sample Preparation
Samples should be free of dust and other particulate matter

Samples should be of sufficient height to fully cover the coil in the probe, so that none of
the lip of the solvent or the bottom, curved part of the tube is within the coil.

For high precision measurement of NOEs or dynamic parameters, samples might have to
be degassed, to remove dissolved Oxygen (Oxygen is paramagnetic, and shortens
relaxation times).

Sample Concentration should be high enough to allow for rapid acquisition of data, but
not too high to cause a significant increase in viscosity which slows molecular tumbling
or overwhelm the receiver. Samples that have overly intense signals will have broad
lines, and be hard to shim. When detecting insensitive nuclei, high concentration is
important for detection in a reasonable length of time. For a 500 MHz spectrometer,
10-20 mM is high enough concentration for most experiments detecting proton (but
probably a little low in concentration for 13C detect experiments, and way too low for 2H
detect experiments in natural abundance).
 Tuning the Probe to the Sample
In order to detect the proper nucleus, the probe must be set on the correct frequency or
tuned to the proper frequency.

The detection coil in a probe is a piece of wire that forms a coil around the sample sitting
in the probe.

The transmitting coil and receiver coil are often the same coil in modern spectrometers.

To transmit the full power of the radio frequency field into the sample and to fully
amplify the signal of the receiver, the impedance of the coil must be matched with the
transmitter and receiver.

To match the coil to the receiver and transmitter, there are 2 capacitors, mounted inside
the probe near the coil. One changes the frequency of the circuit, the other changes the
impedance.

The object in tuning is to minimize the power reflected back to the radio frequency
generator from the probe at the proper frequency

The two capacitors are dependent upon each other, like the Z1, Z2 shims.
 What Effects does Poor Tuning Have?
1) Primarily loss of sensitivity, signal-noise goes down

2) Longer 90° pulse

When is Tuning Necessary?

1) Detect a nucleus other than the standard nucleus, normally that means something other
than proton or carbon.

2) When sample conditions don't match the standard sample (not CDCl3).

3) When experimental conditions don't match the standard (e.g. not room temperature).

4) Any time that long experiments are acquired, as a poorly tuned probe will lead to
lower signal-noise or a longer time to acquire equivalent signal-to-noise.

What Sample Conditions are Relevant?

1) Primarily solvent. Dielectric constant of solvent has great effect on the inductance of
the coil

2) Concentration of sample. High concentration sample are often different as the solvent
is no longer the only contributor

3) Ions. Salt concentration has a great effect on tuning

What Experimental Conditions are Relevant?

1) Temperature. Probe tuning is very temperature dependent.

2) Nuclei for direct/indirect detection. Probes often have multiple coils, and consoles
have multiple channels to pulse and detect on. Tuning a specific nucleus on one channel
is not necessarily the same as on another channel.
 Effects of Tuning on Spectra

Table of Tune Meter Readings on 500 MHz Instrument when Tuned to Standard
(0.1% Ethyl Benzene in CDCl3)
1 13
Solvent H C
CDCl3 1 1
DMSO 101 576
D2O 95 572
Deutero-Benzene 6 353
Methanol 13 539
Acetone 49 505
30% menthol/70% CDCl3 1 1
13
C Spectrum of Standard (CDCl3 solvent resonance)

Well-Tuned (1 scan, 1 second)

Probe Tuned to DMSO (1 scan, 1 second) (64 scans, 3 minutes)

 Temperature Regulation

Temperature regulation has several purposes in NMR.

1) Dynamics- Exchange rates are affected by temperature. As the temperature is lowered,

exchange (or rotation) slows. Depending upon the time regime for exchange that you are
in (slow, intermediate, fast), temperature can be increased or decreased to either improve
spectra or better measure exchange rate.

2) Resonance Dispersion- Chemical shifts, particularly of protons that exchange with

solvent, are affected by temperature. Changing temperature of the experiment can affect
the dispersion of resonances.

3) Linewidth- Resonances sharpen as temperature is increased for molecules that tumble

slowly in solution (mostly macromolecules).

4) Fluctuating temperature in 2D experiments (and some 1Ds) will cause noise,

particularly in t1 (the indirect detect dimension). This is not a big problem for short 2Ds
(10-20 minutes), but with 2Ds acquired for hours, t1 noise from temperature fluctuation
can be a significant problem.

Instrumental effects of changing temperature:

1) Tuning of the probe- The tuning of the probe, particularly on the proton channel is
very sensitive to temperature. Normally, changes of ~>10˚C is enough to warrant
retuning of the probe.

2) Shimming- Some of the shim coils are also temperature dependent, particularly Z1,
Z2, X and Y. Since the linewidth of the deuterium signal is often smaller at higher
temperature, the lock level will increase as temperature is increased, at least after
reshimming.

Temperature regulation parameters:

temp: temp is the parameter to change to set the temperature at a particular value. Typing
temp = 25 followed by su will set the temperature to 25˚C.

vttype: vttype is a parameter to prevent accidentally changing temperature. If you set

vttype = 0 then you cannot change the temperature so typing temp = 35 su will do
nothing. To change temperature, set vttype = 2

sethw: If vttype =2 and typing temp = 25 then typing su does nothing, the VT might need
to be reset (essentially turned off then on). To do this with software, type sethw('vt',
'reset').