TECHNICAL NOTE #3

Advantages of the MAN background correction for Tunable Spectrometers

The Probe for Windows microanalysis software incorporates both an off-peak background correction and an on-peak background correction based on the mean atomic effect (i.e., the background continuum, often referred to as bremsstrahlung, is directly proportional to the mean atomic number, hereafter known as the MAN, of the sample).

The MAN background correction method in the Probe for Windows software should not be confused with other more simplistic methods used by other software. The MAN background correction in Probe for Windows is an iterative technique, using the actual calculated sample composition from the quantitative routines to calculate an exact MAN, based on a polynomial fit calibration curve and is highly quantitative down to minor element levels in most samples.

The MAN background correction, although originally implemented to provide an accurate background correction for fixed spectrometers or monochromaters, also provides many advantages for tunable spectrometers in the analysis of major and minor elements.

These advantages may be summarized as follows :

1. The use of MAN backgrounds requires much less time than off-peak background measurements, since no off-peak motion or counting time is required. This is especially true when a large (> 100) number of analyses are to be acquired. Another advantage is seen when analyzing beam sensitive or samples containing volatile elements. Because the use of MAN samples requires significantly less time, higher currents and more focused electron beams may be used.

2. Since off-peak position counts are not acquired with the MAN background correction, there are no off-peak interferences to be concerned with.

3. In analytical runs where the number elements analyzed is less than or equal to the number of spectrometers, no spectrometer motion at all is required, therefore spectrometer reproducibility is no longer a problem. In addition, because off-peak spectrometer motion is eliminated, wear and tear on the spectrometer motors and mechanisms is reduced.

In actual day to day practice, even when the fixed spectrometers are not being used, off-peak measurements almost never need be acquired except when analysis of trace element quantities are sought. In a typical silicate sample (z-bars approximately 10 to 20), for example, One can routinely measure down to several hundred ppm using the MAN background correction in the Probe for Windows software. Of course, for higher mean atomic number samples, the detection limit will be higher.

Let's examine these points four points briefly, beginning with reduced analysis times.

1. Reduced analytical acquisition times :

In a single pass (no more than one element per spectrometer), the total time required for an off-peak analysis is the count time on-peak plus the count time off-peak plus the time required for motion to and from the off-peak positions. In a typical 10 second count time analysis, this means that an off-peak analysis would require about 31 seconds while an MAN analysis, consisting of only the on-peak time, would require only 10 seconds. This is less than 1/3 the time required for an off-peak analysis.

Of course, the MAN calibration data needs to be acquired first and this does typically add about 4 or 5 minutes to the standard acquisition time for a single pass run using five MAN standards of five 10 second count points each. However, the MAN backgrounds tend not to drift so they generally only need to be acquired only for the first standardization. An additional savings in time can be achieved by using a calibrated setup run in which the MAN calibration was previously acquired.

In the case of more than one element per spectrometer, the time savings is almost as great. In a typical 10 element (two pass) run, the acquisition time for an off-peak corrected sample was 94.1 seconds, while the acquisition time for an MAN background corrected sample was only 48.6 seconds.

This time savings can be especially important when large numbers of data points are to be acquired as is the case for high resolution traverses or X-Y grid stage scans. The difference between 31 and 10 seconds per data point might not seem particularly compelling until multiplied by 3000 or so data points in a large scan. Then, the difference between 25 and 8 hours becomes more impressive.

2. Off peak interferences :

Another point is that off-peak measurements usually need to be carefully examined for off-peak interferences. That is, emission lines that fall on or near the positions chosen for the off-peak measurements. If the interfering line is of a sufficient intensity and the concentration of that element large enough, the subtraction of the background can result in a very negative on-peak intensity. Unlike on-peak interferences, off-peak interferences cannot be corrected for quantitatively. The only recourse is to move the off-peak measurement position.

However, since the MAN background correction requires only an on-peak measurement, interferences on the off-peak positions can be ignored. Of course, on-peak interferences can still be corrected for using the quantitative interference in Probe for Windows .

It should also be mentioned that the use of on or off-peak measured samples can acquired in any combination within a single run. That is, a Probe for Windows run can consist of all off-peak, all on-peak or a combination of both off-peak and on-peak background corrected samples. Note that when off-peak background corrected samples are acquired, Probe for Windows will perform off-peak measurements for the tunable spectrometers and on-peak measurements for any fixed spectrometer that may be present. However, when on-peak MAN samples are acquired, Probe for Windows will perform on-peak measurements for all spectrometers. In fact, Probe for Windows allows MAN samples to be acquired even if the MAN calibration has not be acquired or assigned, although a warning message will be given when the sample is analyzed.

3. Spectrometer motion :

Another point is, that because the use of MAN background corrected samples require no spectrometer motion (at least for runs with only one element per spectrometer), spectrometer reproducibility is no longer a concern. In fact, when extremely precise measurements for major elements are required, the use of a single pass MAN sample run is unequaled for stability.

Of course, spectrometer reliability (as for any mechanical device) is affected by the average duty cycle. The greater the use, the more quickly parts will wear out or require adjustment. Frequent use of the MAN correction might help postpone these inevitable occurrences.

It is hoped that this document has revealed a few of the more compelling reasons to take advantage of the MAN background correction method in the Probe for Windows software. Please contact me if you have any additional concerns or questions.