90˚ 1H pulse

Proper pulse widths are critical in 2D experiments and some 1D experiments. A 90˚ pulse
is when the signal is maximum and positive, a 180˚ is when the signal is null, a 270˚
pulse is when the signal is maximum and negative, a 360˚ pulse is again a null. 1H pulse
widths are normally calibrated on 1 pulse (s2pul) experiments, looking for the null (180˚
or 360˚ pulse) since it is alot easier to see a null than a maximum signal.

The length of the 90˚ pulse is primarily dependent upon 4 factors:

1) Probe- the type of probe (inverse, broadband, etc.), the quality, the age, etc. Different
probes have different standard pulse lengths.

2) Tuning- a poorly tuned probe will have a long 90˚ pulse length.

3) Power of the pulse- Pulse power is a parameter in experiments (transmitter power,

tpwr on a Varian). Higher power leads to shorter pulse lengths.

4) Solvent- Although solvent effects are mostly incorporated into the difference in tuning,
there is still some effect of solvent on pulse length beyond tuning. In general, lots of salt
leads to long 90˚ pulse lengths.
To measure (calibrate) the 90˚ pulse, there are two methods the long method and the
quick method. In most cases the quick method is sufficient.

1) Pulse Array:
Set the sweep width (sw), acquisition time (at), center of the proton spectrum (tof)
appropriately in a standard 1D experiment (s2pul or std1H). Set the number of scans (nt)
= 1, as you do not want to have to acquire multiple scans. Next, set relaxation delay
between scans (d1) to ~5*T1 for the resonance with the longest T1 that you want to
follow (normally it is easiest to follow an intense isolated singlet such as TMS or solvent
peak, but a resonance of your molecule is ok as well, maybe better if you can easily
observe it with one scan), so 30-60 seconds for most solutions. Set the transmitter power,
the power of the proton pulse appropriately (tpwr), the standard value should be fine, but
remember you are measuring the pulse length at THIS power level, so it must be the
same as the experiment that you are actually acquiring! Now, array the proton pulse
width (pw). To array the pulse on a Varian you can type the command array, the
computer asks you the starting value (enter something that you think is well short of a 90˚
pulse), then how many steps you want in the array (10 or so, depends how long you want
to wait), and how much you want to increase the value by each time (4 or so, depends
upon how many steps in the array and what you think the pulse width will be).
Alternatively, you can manually type your array, e.g. pw = 3,6,9,12
Then, acquire. The data can be processed on a Varian by typing ft (or wft as usual) then
dssh to display the data in an array form.
2) Acquire a 1D with short pulse width (e.g. pw = 3). Then, simply guess at the 180˚ (or
360˚ pulse) and type pw = that pulse length or a short array of that pulse length. For
example, if you think that the 90˚ pulse ~8 µs, then type pw = 15.5, 16, 16.5 then acquire
(again d1 must be set to ~5*T1). If you determine 90˚ pulse this way, you might be sure
that this is the 180˚ or 360˚ that you are determining rather than the 540˚...
 Low Power Pulses

Some pulse sequences require pulses that are not at maximum or near maximum power.

Such pulses are used for many purposes including decoupling, small frequency range
saturation, and spin-locking.

For now, just accept that lower power pulses exist.

Why they are sometimes used:

1) Lower heat to the probe. The higher the power transmitted to the probe, the greater the
heat to the sample. Normally, high power is fine because the pulse length is on the order
of 10µs, and there is time to cool before the next pulse. Decoupling and spin-locking are
often on for long periods of time, milliseconds to seconds, which at high power could
cause extreme heating

2) Sometimes only smaller frequency ranges are required

Field Strength (Hz) = 1/(4*pulse with)

If you want to selective excite a particular resonance, you can apply a long pulse

How to calibrate them:

Same as with high power except, lower the transmitter power, set tpwr = lower power
(~43 for TOCSY or ROESY).
 T1 Relaxation

Magnetization is aligned along the z-axis, energy, an RF pulse, is inputted to the sample;
the sample magnetization is flipped into the x-y plane, aligned along the y-axis if the
pulse is a 90˚X pulse. The individual frequencies disperse due to chemical shift; the
receiver is turned on and a sine wave is detected. The sine wave is called the FID, and
over time, the amplitude of the signal gradually goes to zero.

Why does the amplitude of the signal decrease to nothing over time?

Relaxation. The spins relax to steady state.

The theory is that the spins will relax back to the z-axis exponentially over time. The
relaxation can be fit to the equation:

dMz/dt = (M0 - Mz)/T1

where M0 is the magnetization at equilibrium and T1 is the longitudinal or spin-lattice

relaxation time constant.

During T1, energy is transferred to the environment (also know as the lattice) from the
spins. This energy is dissipated as heat.

Thus if the magnetization is initially 0,

Mz = M0(1 - e-t/T1)

The exponential growth of the magnetization:

After a pulse:

Two obvious reasons you should care about T1 relaxation:

1) The longer the magnetization is in the x-y plane, the longer in time that signal will be
detected; the longer the FID will have real signal, the narrower the line in the frequency
domain of the signal observed (remember the discussion of acquisition time).

2) If the magnetization is still in the x-y plane, then it is not aligned along the z-axis, so if
you try to pulse again (acquire another scan), there will be less magnetization to pulse on
and thus observe in the next scan. If you continue to pulse before the magnetization has
fully returned to the z-axis, over time each successive scan will have less signal.

Thus, the T1 relaxation time is correlated to the time between scans, also referred to as
the relaxation delay, usually d1 on a Varian. Actually, the time between scans is d1 + at,
so the relaxation delay + the acquisition time. As we will see next, the optimal pulse
width, pw, is also correlated to the T1 relaxation time.
 Repetition Rate

Repetition rate is the amount of scans acquired per unit time, so scans/minute. More
scans/minute will give you higher signal-to-noise until the T1 relaxation is too slow to
allow for a higher repetition rate.

To achieve maximal signal in 1 scan, a 90º pulse should be used; to achieve maximal
signal in the next scan, a length of time = ~5*T1 should be used as a delay between scans
(acquisition time + relaxation delay or at + d1).

However, waiting that long is not necessary to achieve the highest signal-to-noise over
many scans. The highest signal-to-noise for many scans is waiting ~1.3*T1 when using a
90º pulse.

The actual highest signal-to-noise is achieved in a 1D by waiting less time between scans
but using something less than a 90º pulse according to the Ernst equation:

cos (αe) = e-t/T1

 Pulse flip angle vs. signal-to-noise

(t = time between scans, so t/T1 = 5 means that the time between scans is 5*T1)

Thus, 13C spectra should be acquired with something short of a 90º pulse. The default
parameters have the pulse set to ~30-45º. The parameter pw90 is a parameter not used in
actual experiments (although it affects the pw in some macros); the pw90 parameter
should be the approximate 90º pulse for that power for that nucleus on that instrument.
Compare the pw90 and the pw to estimate the pulse length. Also, the macro ernst will
calculate the optimum pw based upon the longest T1 and the delay time. The syntax is:
ernst(T1 estimate, 90º pulse width)
 How to measure T1

The most common method to measure T1 is inversion recovery.

 How to measure T1 on a Varian Spectrometer

Acquire a normal 1D with correct 90˚ pulse, process, phase correctly.

Type a macro:
dot1

The computer asks you questions, answer them as correctly as possible.

For shortest T1, guess ~0.1s for proton.

For longest T1, guess ~5s for proton.

For length of experiment time, it depends upon concentration, so most samples guess 0.1
hours.

Start acquisition, type:

go

The data can be processed with:

wft dssh

dssh = display stacked spectra horizontally

Then analyzed with the menu buttons:

T1 Proc then T1 Analysis

which are accessible through the Analyze and Exponential menu buttons.

The T1 Proc will select the resonances, and measure the intensity. The T1 Analysis will
fit the data to an exponential

To manually set threshold, type:

ds(12)
to display last spectrum. Then, set threshold as in standard 1D. Next, type:
fp
to measure the peak heights of the resonances
To analyze the data, measure the T1s, type:
t1s
This will give a list of T1 values only. Alternatively, for more information type:
t1
To print the data, highlighting the output and copy/paste to text editor works. Also,
obtaining a line list is useful to identify what T1 value goes with what resonance so type:
pll page
 T2 Relaxation

T2 relaxation is relaxation in the x-y plane, also known as transverse relaxation or spin-
spin relaxation.

T2 relaxation like T1 relaxation is assumed to be an exponential process.

In the vector model, T2 is a blurring of the vectors due to some of the individual vectors
in the sample precessing slightly faster or slower due to some inhomogeneity in the
sample, magnetic field etc. Another description is that the magnetization fans out or
dephases; there is no energy change in T2 magnetization, but the disorder of the sample
(no change in enthalpy, but a change in entropy).

Any loss of magnetization that causes loss of magnetization in the x-y plane contributes
to T2; thus, T1, relaxation back to the z-axis, actually contributes to T2. In other words,
T2 can = T1, but can never be larger than T1.

Since the FID is the decay of signal in the x-y plane after a pulse, and T2 is the loss of
signal in the x-y plane, the FID = T2 relaxation? Not quite, T2 is the intrinsic time
constant of relaxation, T2* is the observed T2 in the FID. Ideally, the acquisition time (at
on a Varian) would be set based upon the length of the T2*.

The intrinsic time constant (T2) would include factors specific to the molecule such as
chemical/rotational exchange, whereas the observed time constant in the FID (T2*)
would include inhomogeneity of the magnetic field, poor shimming for example.

Thus,

∆ν(1/2) = 1/(T2*)π

where ∆ν(1/2) is the resonance line width at half height.

 Spin Echoes

Spin Echoes are one of the key building blocks for multiple pulse experiments. The basic
concept is that after a pulse that flips magnetization into the x-y plane, magnetization fans
out in opposite directions (some at higher frequency, some at lower frequency) due to T2
relaxation; then, by pulsing with a 180˚ pulse, the magnetization is flipped and now the
higher frequency are behind the lower frequency, so if you wait the same length of time
after the 180˚ pulse as after the 90˚ pulse, then the magnetization vectors are all together
again. This is called refocusing of the magnetization. The regeneration of the
magnetization is a spin echo.
 Measuring T2

T2 is much harder to measure than T1. The simple way to do it would be to vary τ is the
standard spin echo experiment then measure the decay of amplitudes of the echo. This
does not work well, especially for long values of T2.

Problems with this method:

1) T2 is independent of any inhomogeneities in the magnetic field (T2* is the field

dependent time constant). Thus, molecules cannot diffuse from one place in the magnetic
field to another during the experiment, else the echo is no longer field independent. For
the echo to work properly, the speed of individual vectors must be the same before and
after the 180˚ pulse, and this will not be true if the molecule diffuses from one part of the
magnetic field to another. This is a big problem and long values of τ.

2) Spinning the sample will also interfere with measurements of T2 as molecules move in
the magnetic field during the experiment. The spinning must be either turned off or slow
compared to the values of T2 being measured.

3) π pulses are always a problem due to off-resonance effects and incorrect calibration.
The 180˚ flip has to be exact or the experiment will not work.

The normal way to measure T2 these days is to a CPMG T2 measurement which to some
extent corrects for the errors mentioned above.

CPMG T2 = Carr Purcell Meiboom Gill T2 measurement

(π/2)X - τ - πY - 2τ - πY - 2τ - πY - ...
 Why do the nuclei relax?

Relaxation is not a spontaneous process; relaxation is stimulated by

fluctuating fields at the appropriate frequency inducing spin transitions.

The appropriate frequency is the Larmor frequency for that nucleus. Thus, relaxation is
magnetic field dependent, and line-widths of nuclei are magnetic field dependent! Higher
magnetic field strength is not always better.

Time-dependence arises from various molecular motions including vibrations, rotations,

diffusion, tumbling in solution, etc. However, most motions are either fast or slow on the
NMR time scale (in this case, ~500 MHz, the Larmor frequency), except molecular
tumbling which does have frequency within 1-2 orders of magnitude of 500 MHz. The
molecular tumbling is the rate at which the molecule tumbles in solution. Obviously, the
rate of molecular tumbling is correlated to molecular size, molecular shape,
concentration, and the viscosity of the solvent. Smaller molecule, faster tumbling; more
spherical, faster tumbling; lower concentration, faster tumbling (less collisions between
molecules); less viscous solvent, faster tumbling.

The rotational correlation time (or simply referred to as the correlation time), τc is defined
as the average length of time it takes a molecule to rotate through one radian.

The spectral density function (J(ω)) is defined as the frequency distribution of the
fluctuating magnetic fields; essentially, the probability of finding a component of motion
at a given frequency (ω, in rad/s). Only when a component is at the Larmor frequency
does relaxation occur.

J(ω) = 2τc/(1 + ω2τc2)

slow motion = molecular tumbling is << than the Larmor Frequency (macromolecules)
fast motion = molecular tumbling is >> than the Larmor Frequency (small molecules)
intermediate motion = molecular tumbling is ~ the Larmor Frequency
 T1 and T2 Relaxation Times and Correlation Time

T1:

T1 is long for molecules in slow and fast motion as the probability of a

component matching the Larmor frequency is low. T1 is also field
dependent (dashed line) the Larmor frequency is field dependent.

T2:

T2 is ~T1 for molecules in fast and intermediate motion, but is short for
large molecules as T2 relaxation is also stimulated by slow molecular
motions.
 Primary Causes of Relaxation

1) Magnetic dipole-dipole interactions:

By far the most common mechanism of relaxation in liquids.

Intramolecular and intermolecular interactions can lead to dipole-dipole interactions. This

subject will be discussed in greater detail when discussing NOE, as NOE is a dipole-
dipole interaction.

Spin 1/2 nuclei relax off of other spin 1/2 nuclei nearby. Each spin 1/2 nucleus is
essentially a bar magnet. The orientation of all of the bar magnets are unaffected by
tumbling in solution as the bar magnets will always align with the magnetic field.
However, the positions of the bar magnets change as the molecule tumbles in solution.
There is obviously a distance term as bar magnets far away from each other will not
affect each other.
13
C relaxes primarily by intramolecular dipole-dipole interactions with the attached
protons. This is obviously a problem for quarternary carbons as they have no protons to
relax off of.

2) Relaxation due to unpaired electrons (paramagnetic relaxation):

Actually, this is an intermolecular dipole-dipole interaction as well, but it is special

because the one dipole is an electron with a gyromagnetic ratio that is 658 times larger
than a proton, which makes relaxation more efficient. Paramagnetic metal ions such as
manganese, chromium, and dissolved oxygen are common sources for unpaired electrons
that lead to intermolecular dipolar relaxation. Sometimes, paramagnetic species are added
deliberately to the sample to shorten relaxation times (referred to as relaxation agents, to
be discussed in more detail later). Shorter T1 relaxation times can allow for increased
repetition rate; in other words, more scans per minute. If the relaxation time is still on the
order of a few seconds, the NMR lines will remain sharp. Also, note that paramagnetic
species will cause paramagnetic relaxation which means there is less intramolecular
dipole-dipole relaxation, so less NOE (NOE will be discussed later).
3) Quadrupolar relaxation:

Relaxation by interaction with electric field gradients. Electric field gradients exist
wherever the molecular electronic environment is not symmetrical because of different
electronegativities of nearby atoms. The amount of relaxation due to quadrupolar nuclei
is dependent upon the quadrupolar moment. While only the quadrupolar nuclei
themselves are directly affected by quadrupolar relaxation, other nearby nuclei can be
affected indirectly.

For I = 1/2, the charge distribution of the nucleus is spherical; for quadrupolar nuclei, the
charge distribution is elliptical. The quadrupole moment is influenced by electric field
gradients about the nucleus, rather than symmetric electric fields. The electric field
gradient is affected by the tumbling in solution. If this occurs at the appropriate
frequency, flipping of the spin states is induced. Thus, the primary factors affecting
relaxation of quadrupolar nuclei are:

1) quadrupolar moment: the larger the quadrupolar moment, the faster the nucleus will
relax by interaction with the electric field gradient.

2) molecular tumbling: as tumbling in solution increases, the quadrupolar relaxation

should be less efficient since the interaction with the electric field gradient must occur at
an appropriate frequency. If the molecular tumbling were fast relative to the interaction,
then the quadrupolar relaxation would not occur often.
3) magnitude of the electric field gradient: the electric field gradient is dependent upon
molecular symmetry. The field gradient should be 0 for octahedral or tetrahedral
symmetries, and thus lines should be narrow. Usually, there is slight asymmetry, leading
to some quadrupolar relaxation.

11
B NMR spectrum

4) Scalar relaxation:

The most common instance is 2 strongly coupled spins: one spin can affect the magnetic
environment of another usually in exchanging systems such as -CHOH- fragments.

Nucleus A is in some sort of chemical exchange process that causes the JAX to change
over time, a time-dependent fluctuating field. Thus, the exchange process destroys and
regenerates a field on the other nucleus, affecting the relaxation time of the X nucleus.

A second type of scalar relaxation is correlated to quadrupolar nuclei (so it could be

considered quadrupolar relaxation). If nucleus A is coupled to nucleus X, and X is
quadrupolar. The JAX is always present, eventhough it is not necessarily observed. If the
relaxation of the X nucleus is fast relative to the coupling constant, the coupling is not
observed in the NMR spectrum of A. However, JAX is present and the fluctuating fields
affect the relaxation of A. The relaxation time of A is affected by X if 1/T2X >> JAX
5) Spin rotation relaxation:

If a molecule or group within a molecule undergoes transitions between one rotational

state and another, relaxation can be affected by the spin rotation relaxation. For small
molecules rotating on their own axes, the electronic cloud moves, creating fluctuating
magnetic fields. Spin rotation relaxation is more efficient when the molecule is
symmetric or for freely rotating methyl groups. This only matters when the rotation is
fast, then the rotation speed is proportional to the rate of relaxation induced; the
efficiency is actually increased by increasing tumbling rates. Thus, higher temperature
actually causes enhancement of spin rotation relaxation.

6) Chemical shift anisotropy (CSA):

The instantaneous chemical shift of a molecule depends upon orientation in the magnetic
field as the electron distribution in chemical bonds is anisotropic. The chemical shift gets
averaged out by molecular tumbling, but the difference between the chemical shift in one
orientation versus another is chemical shift anisotropy. The effect on T1 relaxation is
correlated to the rate at which the molecule tumbles in solution. CSA is a problem when
the chemical shift difference is large between orientations, the tumbling of the molecule
is slow, and the field strength is high as higher field leads to a greater chemical shift
difference in Hz.

Orientation dependence of chemical shift: there is a greater orientation dependence of

chemical shift for asymmetric environments. For example, there is little CSA for
octahedral geometries, but very large CSA for square planar geometries.

CSA is proportional to the square of the applied field, so there is a great detrimental
effect of higher field. Since CSA increases a square of the field, while dispersion is
improved only linearly with the field, data on some nuclei at higher field strengths is
actually worse, much worse.

Nuclei that have large CSA include: 19F, 31P, and many metals. Aromatic 13C (any sp2
13
C) can have large CSA values
 Other Causes of T2 Relaxation

1) Inhomogeneity in the magnetic field (actually only affects T2*). Poor shimming.

2) Chemical Exchange including proton transfers, conformational equilibria, and valence

tautomerism. Any event that leads to more than one distinct chemical shift of a nucleus
will be observed as T2 relaxation if the event causing the distinct chemical shifts is on the
order of molecular tumbling. If the event is fast compared to molecular tumbling, there
will little effect, as with molecular vibrations. If the event is slow compared to molecular
tumbling, 2 distinct chemical shifts will be observed (chemical exchange will be
discussed later).

3) Changes in local environment including molecular motion. Actually, this fits with the
explanation above. If there is molecular motion, nuclei will have different chemical shifts
dependent upon the local environment; for example, molecular motion that causes a
proton to change orientation relative to a nearby aromatic ring will experience a different
ring current effect. If the molecular motion causing this change in chemical shift is on the
nanosecond time scale, the resonance will be broadened. If the molecular motion is slow,
so much slower than nanosecond time scale, 2 distinct chemical shifts will be observed
(chemical exchange will be discussed later).

4) Slow molecular tumbling (macromolecules or any molecule in viscous solution). One

spin exchanges energy with a nearby spin, sometimes referred to as a flip-flop process.
The flip-flop process is stimulated by low frequency motions, such as tumbling for large
molecules.
 Spin Echoes and Chemical Shift

Chemical shift differences behave the same way as the fanning out due to T2 relaxation
with spin echoes. This is a useful property in that by using spin echoes, the spins of the
whole molecule can be affected at once without differential excitement due to chemical
shift.
 Spin Echoes and Scalar Coupling

The same refocusing with the spin echo does not apply to homonuclear coupled systems.
The system is the same through the 180˚ pulse that flips the magnetization. The
difference is that by inverting the signals of a scalar coupled system, the direction of the
precession is also reversed as the line that was precessing at +J/2 is now precessing
at -J/2. This reversal occurs because the nuclei which used to α neighbors, or ground
state neighbors (causing the +J/2) now have β neighbors or excited state neighbors
(causing -J/2).

Note that this does not apply to heteronuclear coupled systems because the 180˚ pulse
does not excite the X nucleus (unless simultaneous pulses are applied to 1H and X).