How Accurate is the NOE?

The NOE is a very good qualitative measurement. If a NOE is observed, the distance between the 2 protons is within 5 angstroms. The NOE is routinely used in a semi-quantitative fashion: (Different spectroscopists may use slightly different ranges)

In many cases, the NOE is quantitated and reported as a percentage of the origional peak intensity without the NOE. When interpreting semi-quantitative or quantitative NOEs, you MUST consider the following:

  1. Magnetic Equivalence
  2. Spin Diffusion
  3. Complicated Modes of Motion
  4. Other Relaxation Mechanisms
  5. Strong Coupling
See also Frequently Asked Questions about NOE Experiments

1. Magnetic Equivalence

In many molecules, some protons are magnetically equivalent (indistinguishable by NMR, these protons generate 1 peak). These magnetically equivalent protons are best represented as a "pseudoatom". NOEs involving the magnetically equivalent protons represent distances to and from the pseudoatom. Add 0.5 angstroms to the maximum allowed distance of the semi-quantitative NOE ranges.

The lowest energy conformation of the molecule shows that one methylene proton may be very close to the H2' sugar proton. However, the two methylene protons generate just one NMR signal. A medium-intensity NOE is observed between the methylene signal and the the H2' proton of the sugar. Since the NMR spectrum cannot resolve the magnetically equivalent methylene protons, the NOE measures the distance from the methylene PSEUDOATOM to the H2' sugar proton.

2. Spin Diffusion

This is the biggest cause of misinterpreting the NOE. Consider three protons A, B, and C. If A generates an NOE at B, then the NOE at B can "diffuse" to C. For small molecules, the NOE between A and B is less than expected since some of the NOE has "diffused" from B. The NOE between A and C caused by the "spin diffusion" is negative. ALWAYS evaluate the geometry of your protons and consider if the NOE from A to B to C can be affected by spin diffusion.

Calculated NOE enhancements on irradiating A. Although B and B' are equidistant from A, the enhancement at B is drastically reduced by the efficient "diffusion" of the NOE from B to C. B' suffers no such loss, and receives the full NOE.

Variation in the calculated NOE from A to B on changing the angle a as shown. The direct A-to-B enhancement (affected by spin diffusion to C), the indirect spin diffusion from A to C to B, and the sum of the two effects are shown.

3. Complicated Modes of Motion

A. If the distance between two protons changes faster than 10 msec., the NOE is porportional to the weighted average of the distance during the experiment. Since the NOE depends upon (distance)-6, short distances are weighted VERY heavily compared to longer distances.

B. If the molecule exchanges conformations slowly, a proton may give rise to a different peak in each magnetic environment. These two peaks from the same proton will have an apparent NOE between them, which is always negative.

C. If one part of the molecule rotates more than 3 times faster than another part of the molecule, then NOEs from the different parts of the molecule will not have the same correlation with distance. Also, if the principle molecular axes have lengths differing by more than a factor of 3, then NOEs between protons aligned with one axis vs the other axis will not have the same correlation with distance.

The internal motion of a long aliphatic tail causes the effective tumbling rate for this tail to be much faster than for the bulky, rigid ring system. NOE intensities involving the flexible tail will be greater than NOE intensities among ring protons.

A DNA duplex has an axial and radial dimension. For a 20 base pair DNA duplex, two protons separated by 2.5 angstroms and aligned with the DNA axis have 17% less NOE intensity than two protons separated by 2.5 angstroms and aligned with the DNA radius.

4. Other relaxation mechanisms

In most cases, only paramagnetic and quadrupolar nuclei (such as oxygen) compete with the NOE. NOEs between protons near paramagnetic or quadrupolar nuclei will be less intense than NOEs from other protons in the same molecule. Remove paramagentic impurities from your sample.

Samples of small molecules can be de-gassed to remove dissolved oxygen by bubbling argon through the soltion.. Dissolved O2 decreases the NOE of solvent-accessible (surface) protons. Since most protons in small molecules are solvent-accessible, so they are equally affected by dissolved oxygen; qualitative and semi-quantitative relative NOE intensities will not change.

However, the decrease in quantitative NOEs can be estimated as follows:
0.1 / [linewidth at half height in degassed sample (where lb=0, digital resolution < 0.2 Hz/point)]
Thus, this error is only important for very sharp lines.

5. Strong Coupling

This problem is rare and usually unimportant. If the difference in chemical shifts between two peaks is less than 0.3 ppm and they are coupled to each other, these peaks are strongly coupled. An NOE transferred to one of the strongly coupled protons will affect the intensities of the other strongly coupled protons.

Frequently Asked Questions about NOE experiments

When should I use the 1D NOE vs the NOESY experiment?

Use the 1D NOE experiment when you need to identify through-space neighbors of 1 to 5 protons, or if you need very accurate, quantitative NOEs. Use the NOESY when you want to measure NOEs between more (or all) protons. Use the NOESY when you need to measure the NOE from a proton in a crowded or overlapped spectral region; the second dimension of the NOESY will separate NOEs from overlapped peaks.

How long should I set the NOESY mixing time?

The NOE is transferred between protons during the "mixing time" in the NOESY experiment. In general, set the mixing time near 400 msec for small molecules (< 1000 MW) and near 200 msec for large molecules (> 1500 MW). Longer mixing times allow the NOE to "build up" in intensity. However, mixing times that are too long can cause lots of spin diffusion, leading to innacurate NOEs. If your NOESY spectrum has small or missing NOEs, re-run your NOESY at a different mixing time.

When should I use the ROESY experiment?

Small molecules (< 800 MW) at room temperature tumble faster than 400 MHz in solution; their NOEs are positive. Large molecules (> 1500 MW) at room temperature tumble slower than 400 MHz in solution; their NOEs are negative. Molecules between 800 and 1500 MW at room temperature tumble at a rate at wihc the maximum possible NOE is small or zero. OF course, small molecules at low temperature or large molecules at high temperature may also tumble at a rate at which the NOE is small.

The ROE (rotating frame NOE) signal is always positive. The maximum possible ROE increases with increasing molecular weight. Therefore, in cases when the NOE is small or zero, the ROESY experiment is preferred. Also, if the large molecule is in slow conformational exchange, the exchange and the NOE of the NOESY spectrum will both generate negative peaks. Since ROESY peaks are always positive, the negative exchange peaks and the positive ROESY peaks can be easily separated.

Back to NMR Facility Users Guide
Back to NMR Facility Home Page
Send comments to chemnmr@indiana.edu
Last updated: April 2nd, 1998
URL: http://nmr.chem.indiana.edu/NMRguide/misc/noeacc.html
Copyright 1996, The Trustees of Indiana University