EPMA Life Hacks

[John Donovan]

This topic is for good ideas, tips, tricks in EPMA science for making your life easier and that don't fit into other existing topics.

Anette von der handt suggested it over the summer and I think it's probably good to have a "catchall" for such information.

[Probeman]

I posted this idea for using a tiny drop of isopropyl alcohol to adhere small flat samples to an Si substrate (or Al stub) a while back:



The idea being to allow the alcohol to "wick around" the edge and in dissolving hydrocarbon residues and then evaporating, such small flat samples will stick to a flat surface such as an Si substrate or Al stub without the need for gluing your samples using epoxy.

Previously we found that performing thin film analysis of multiple small cleaved pieces of Si weren't stable using just a 25mm carbon tab on an Al stub. The carbon tabs would move around and some pieces would go out of Bragg focus over time. Some relief can be had by utilizing the "hard" carbon tabs, but even better is the isopropyl "wick around" method.

The samples stay right where they are for the duration of the WDS measurement, and then when you want to re-use the Al stub, a tiny amount of pressure using a small needle is enough to dislodge the small pieces of Si wafer.

Recently I came across a company that could produce nice round pieces of Si wafer in various diameters including 25 mm, that can also be used in place of an Al stub. Here's a pic of a tray of 25 I bought that are cut to 25 mm so they'll fit right into my sample holder:



Then I adhere the small cleaved pieces of Si wafer to the large (25 mm) Si wafer each with a tiny drop of isopropyl (staying away from the edges), add a spacer ring equal to the thickness of the Si pieces, and then I'm automatically top referenced for analysis.

Just FYI.

Oh, I should mention that they cost about $200 for 25 wafers, but since they're totally re-usuable that seems like a good deal.

[Nicholas Ritchie]

How about this one?
If you want 1 nA, set your probe current to 0.9950 nA rather than 1.000 nA.  At least on my picoammeter, I get N digits of precision.  When I set the current to slightly more than 1 nA, I end up with one less significant digit than if I set it slightly below 1 nA.
Same goes for 10 nA, 100 nA etc.

[Probeman]

How about this one?
If you want 1 nA, set your probe current to 0.9950 nA rather than 1.000 nA.  At least on my picoammeter, I get N digits of precision.  When I set the current to slightly more than 1 nA, I end up with one less significant digit than if I set it slightly below 1 nA.
Same goes for 10 nA, 100 nA etc.

Good idea. If one doesn't have beam current regulation, this would help with monitoring stability.

On a slightly related SEM note, as you know, though others might not, don't believe your picoammeter readings if one is monitoring sample current from the stage and using "low vacuum" imaging mode (in an FEI instrument). The current flow from the water vapor to the stage will vary based on the ionization from different detector settings.  Of course performing quantitative measurements in "low vacuum" mode is fraught with many other issues as well.

[Nicholas Ritchie]

Another hack is use to collect EDS standard spectra:

Rather than collecting a single spectrum for 300 s, collect 5 spectra from 5 distinct points (usually selected at random (shotgun-style)) for 60 s each.  Then compare them analytically (https://search.proquest.com/openview/1ac95d227ead32a092137a9e61c71e20/1?pq-origsite=gscholar&cbl=33692) and/or visually to ensure that they agree. Remove any that don't and sum the remainders into a single spectrum to use as the standard.

* Ritchie, Nicholas WM. "Strategies for Standardizing EDS Measurements." Microscopy and Microanalysis 28.S1 (2022): 496-498.

[John Donovan]

Another hack is use to collect EDS standard spectra:

Rather than collecting a single spectrum for 300 s, collect 5 spectra from 5 distinct points (usually selected at random (shotgun-style)) for 60 s each.  Then compare them analytically (https://search.proquest.com/openview/1ac95d227ead32a092137a9e61c71e20/1?pq-origsite=gscholar&cbl=33692) and/or visually to ensure that they agree. Remove any that don't and sum the remainders into a single spectrum to use as the standard.

* Ritchie, Nicholas WM. "Strategies for Standardizing EDS Measurements." Microscopy and Microanalysis 28.S1 (2022): 496-498.

I completely agree that all analyses should be based on the average of replicate measurements and then evaluated statistically, including standards!

That is why the default number of automated point acquisitions for standard samples in Probe for EPMA is 5, though of course one can acquire more or fewer points:



I am reminded of a geology student I worked with long ago who had gotten his undergraduate degree in math, and who always insisted on measuring 25 points for each material. Because, you know, statistics:

Student's "t":

   # pts    d.f.    0.60    0.80    0.90    0.95    0.99
       2       1  1.3764  3.0777  6.3138 12.7062 63.6567
       3       2  1.0607  1.8856  2.9200  4.3027  9.9248
       4       3   .9785  1.6377  2.3534  3.1824  5.8409
       5       4   .9410  1.5332  2.1318  2.7764  4.6041
       6       5   .9195  1.4759  2.0151  2.5706  4.0321
       7       6   .9057  1.4398  1.9432  2.4469  3.7074
       8       7   .8961  1.4149  1.8946  2.3646  3.4995
       9       8   .8889  1.3968  1.8595  2.3061  3.3554
      10       9   .8834  1.3831  1.8331  2.2622  3.2498
      11      10   .8791  1.3722  1.8125  2.2281  3.1693
      12      11   .8755  1.3634  1.7959  2.2010  3.1058
      13      12   .8726  1.3562  1.7823  2.1788  3.0545
      14      13   .8702  1.3502  1.7709  2.1604  3.0123
      15      14   .8681  1.3451  1.7613  2.1448  2.9768
      16      15   .8662  1.3406  1.7531  2.1315  2.9467
      17      16   .8647  1.3368  1.7459  2.1199  2.9208
      18      17   .8633  1.3334  1.7396  2.1098  2.8982
      19      18   .8621  1.3304  1.7341  2.1009  2.8784
      20      19   .8610  1.3277  1.7291  2.0931  2.8609
      21      20   .8600  1.3253  1.7247  2.0860  2.8453
      22      21   .8591  1.3232  1.7207  2.0796  2.8314
      23      22   .8583  1.3212  1.7171  2.0739  2.8188
      24      23   .8575  1.3195  1.7139  2.0687  2.8073
      25      24   .8569  1.3178  1.7109  2.0639  2.7969
      26      25   .8562  1.3163  1.7081  2.0595  2.7874
      27      26   .8557  1.3150  1.7056  2.0555  2.7787
      28      27   .8551  1.3137  1.7033  2.0518  2.7707
      29      28   .8546  1.3125  1.7011  2.0484  2.7633
      30      29   .8542  1.3114  1.6991  2.0452  2.7564

https://skeptric.com/two-point-eight/