Modeling Detection Limits in CalcZAF

[Probeman]

I sometimes find the Model Detection Limits and Counting Times dialog in CalcZAF to be very useful. The idea came from John Fournelle, but how well does it compare to other methods, for example the Love and Scott method used by PFE for single point analysis?

Let's take some real intensities for Ti Ka in SiO2 and see how they compare...

Clicking the Run | Model Detection Limits menu pulls up the Model Detection Limits and Counting Times dialog in CalcZAF as seen here with the parameters filled in with values from our measured standard (TiO2) and unknown (SiO2) intensities:



Which yields around 2.3 PPM single point detection for 99% confidence (3 sigma).

Note that these intensity values are from PFE's aggregate mode where all 5 spectrometers are "aggregated" together (both standard and unknown intensities) to provide better sensitivity, though they can be calculated separately as well as seen here:

Detection limit at 99 % Confidence in Elemental Weight Percent (Single Line):

ELEM:       Ti      Ti      Ti      Ti      Ti
   266  .00072  .00049  .00047  .00081  .00068
   267  .00072  .00049  .00047  .00081  .00069
   268  .00072  .00049  .00047  .00081  .00069
   269  .00072  .00049  .00047  .00080  .00068
   270  .00072  .00049  .00047  .00080  .00068

AVER:   .00072  .00049  .00047  .00080  .00068
SDEV:   .00000  .00000  .00000  .00000  .00000
SERR:   .00000  .00000  .00000  .00000  .00000


By checking the Use Aggregate Intensity Mode checkbox in PFE, the intensities can be aggregated (which are the values we used for our detection limit modeling in CalcZAF) as seen here:

Detection limit at 99 % Confidence in Elemental Weight Percent (Single Line):

ELEM:       Ti      Ti      Ti      Ti      Ti
   266  .00025  .00000  .00000  .00000  .00000
   267  .00025  .00000  .00000  .00000  .00000
   268  .00025  .00000  .00000  .00000  .00000
   269  .00025  .00000  .00000  .00000  .00000
   270  .00025  .00000  .00000  .00000  .00000

AVER:   .00025  .00000  .00000  .00000  .00000
SDEV:   .00000  .00000  .00000  .00000  .00000
SERR:   .00000  .00000  .00000  .00000  .00000


As you can see we obtain 2.5 PPM using the Love-Scott method. What is the difference?  The CalcZAF detection limit model does not account for matrix effects since we may not know the composition yet!