(this opens a baseline points window containing a Return button)Integrating the spectrum means finding the area underneath the peaks that interest you. This peak area or integral is proportional to the number of hydrogens that create these signals. Therefore, integrals are useful only if we compare them to other integrals. Also, only the relative integral size is meaningful; two integrals with sizes of 1 and 4 tell us the same thing as two integrals with sizes of 0.2 and 0.8.
Reed College NMR Spectrometer disclaimer: NMR spectra for Chem 201 & 202 are gathered in a way that can easily distort the sizes of different peak patterns, making them artificially too large or too small (peak size distortion is an artifact of our "fast scan" procedure). These distortions do not affect integrals if all of the integrated peaks are distorted in exactly the same way, but this is rarely the case. Expect deviations in your integrals to be as large as 10-15% (routine) and even larger. For example, you might measure integral sizes of 0.9 and 3.3 for two peaks; these measurements are compatible with a molecular formula containing 1H and 3H (0.9:3.3 approximately equals 1:3) or a formula containing 1H and 4H (0.9:3.3 approximately equals 1:4).
SpinWorks integration involves four steps. Phasing should have been completed already (but you can rephase peaks at any time). Baseline correction is not usually necessary, and calibration is completely optional. So, if your baseline looks reasonably flat, and your patterns look reasonably well-phased, perform only step 3.
2. Baseline correction (optional)
Most spectra contain curved baselines. This curvature look like a broad, low hill stretching across the entire spectrum, or it might look like a sharp, steep twist at one end of the spectrum. In either case, baseline curvature will not usually affect integral measurements much compared to other distorting factors (see above). However, you may occasionally find a spectrum where some compensation for baseline curvature would be helpful, or you just might want to compensate for baseline curvature because that's the kind of careful person you happen to be.
Baseline correction is a procedure that flattens the baseline. First, you tell the program where the baseline is (you do this by clicking on several baseline points). Then, the program looks at the spectrum's height at these "baseline" points, draws a smooth curve that passes through these points, and subtracts this curve from your spectrum. The result is a corrected spectrum in which your selected points lie on the baseline (height = 0), and, you hope, curvature in the original spectrum has been reduced.
(this opens a baseline points window containing a Return button)3. Selecting integration regions
At this point, the integration dialog window opens.
Wait! Are you sure you know what you want to do? This section assumes that you want to integrate separate regions of your spectrum and compare these integrals. This is a useful, but potentially tedious, approach. However, a simpler approach is to integrate the entire spectrum and use a ruler to measure compare integrals for different regions. If you are willing to produce a single integral curve for the entire spectrum, do the following (otherwise skip ahead):
The next step is to define regions for integration. An integration region should consist of: 1) baseline on the left, 2) signals produced by a single type of H in the middle, and 3) baseline on the right. An ideal integration region is shown below. The region extends from the left edge of the blue integral curve to the right edge. If you view integration as a left-to-right procedure, then you can see that the integral curve starts low on the left, remains flat until it approaches an NMR peak, rises as it passes each peak (the degree of curve rise reflects the area under the peak), flattens between peaks, and finishes as a flat line again on the right.
To define an integration region:
4. Calibrate one integral curve (optional)
Table of Contents | 201_202 Home
created 2/2006 by Alan Shusterman