The ratio of the area of the small peak to that of the larger is five. In the first instance the area of each peak is taken by integrating between the start of the envelope to the minimum between the peaks and from the minimum to the end of the envelope. The same can be achieved manually by dropping a perpendicular from the minimum to the baseline as shown and the area of each peak taken as that on either side of the bisector. It is seen that the division is fairly accurate, the first peak including a small portion of the area of the second peak and vice versa. The diagram also illustrates the process of peak skimming. Depending on the software available, the contour of the larger peak is projected under the smaller peak. The area of the smaller peak is taken as the difference between the area of the total envelope and the area of the first peak taken beneath the projected contour, which will give a fairly accurate result.
The accuracy of the skimming procedure depends on the method used for projecting the baseline. The method used in the diagram utilizes a linear extrapolation under the second peak and it can be seen that this leads to extensive error. Employing a second order polynomial or an exponential extrapolation would probably provide more accurate results. The weakness of this procedure lies in the fact that each unresolved pair of peaks provides a unique contour and, thus, any standard function that is employed for extrapolation purposes can only be an approximation.
Another problem is the area measurement of a peak superimposed on a sloping baseline or on the tail of a large overloaded peak. The elution of a small peak on the tail of a large peak is shown in figure 40.
Figure 40. The Elution of a Small Peak on the Tail of a Large Peak
It is seen that the shape of the small peak is extensively distorted and that the use of a linear function to project the curve of the major peak below that of the smaller is virtually useless. Again a second or third order polynomial or exponential function would certainly provide improved results. However, accurate values for the peak areas would probably be obtained by peak deconvolution. The first peak would be reconstructed using parameters obtained from the ascending portion of the peak that is free from contamination from the smaller peak. The area of the first peak would be obtained by direct integration, and that of the smaller peak, by the difference between the area of the total envelope, and that of the first peak. A skimming procedure is often used when measuring the area of peaks on a sloping baseline as shown in figure 41.
Figure 41. Peak Area Measurement on a Sloping Baseline
The slope is small, so a linear function might be used to extrapolate the baseline beneath the peak.
For baselines with greater slopes (e.g. that shown in the diagram) a second order polynomial will usually provide an accurate contour of the baseline below the peak. In manual measurement, the baseline beneath the peak can be quite accurately constructed using a set of French curves.
