Counting times and LLDs for large area vs more conventional sized diffractors

[Malcolm Roberts]

Folks,
Anyone have a feel for the counting times for similar llds between large area and more conventional sized diffraction crystals. I'm looking for a rough ball park percentage figure...... like....cuts counting times by 10, 15, 20 % for a similar LLD.....
Cheers,
malc.

[Probeman]

Folks,
Anyone have a feel for the counting times for similar llds between large area and more conventional sized diffraction crystals. I'm looking for a rough ball park percentage figure...... like....cuts counting times by 10, 15, 20 % for a similar LLD.....
Cheers,
malc.

Hi Malcolm,
This is a simple problem in geometry, assuming that we are comparing crystals with similar focal circles.

It is analogous to the aggregate intensity feature in Probe for EPMA where on and off-peak intensities from spectrometers running the same element and x-ray emission line are combined for improved precision. In the aggregate method, let's say we are combining two same sized PET crystals from two spectrometers with the same focal circle. In this case our geometric efficiency doubles because our diffraction area doubles and hence our intensity doubles.

But precisions add in quadrature, so we need to take the square root of the doubling, so double the intensity and take the square root and we get ~1.4 which is a 40% improvement in precision or sensitivity.

Now as for larger area crystals the same math applies.  I don't know much about the sizes of JEOL crystals, but the Cameca large area crystals (e.g., LPET), are roughly 2.8 times larger in area than the standard PET crystals, which should produce an improvement in precision of about the square root of 2.8 or ~1.7 or so, so a 70% improvement in sensitivity.

Mike Jercinovic has an excellent treatment of these issues and also compares the PET/LPET and VLPET crystal sizes in this paper:

http://www.probesoftware.com/download/BoT&A_Jercinovic.pdf

[Malcolm Roberts]

Thanks John....... Most helpful. I managed to extract some crystal performance data comparing JEOL H, J and L PET on peak count rate and p/b for Mo La. Superficially, these numbers suggest that tighter radius and larger area crystal types provided no improvement in p/b, in fact the contrary. However, this seems to be a somewhat one-sided view of the problem, and certainly one that doesn't aid the salesman!!
Cheers,
Malc.

[Probeman]

Thanks John....... Most helpful. I managed to extract some crystal performance data comparing JEOL H, J and L PET on peak count rate and p/b for Mo La. Superficially, these numbers suggest that tighter radius and larger area crystal types provided no improvement in p/b, in fact the contrary. However, this seems to be a somewhat one-sided view of the problem, and certainly one that doesn't aid the salesman!!
Cheers,
Malc.

Hi Malcolm,
I would be most interested to see this data.  Can you can post your peak intensity and P/B values for the different crystals (in cps/nA with the keV and the materials you utilized)?  There is a thread already started here for this:

http://probesoftware.com/smf/index.php?topic=645.msg3793#msg3793

There's also this interesting comparison by Gareth Seward:

http://probesoftware.com/smf/index.php?topic=18.msg564#msg564

john