Hi all,
I've been approached by a materials science researcher asking whether our EPMA (JEOL JXA 8230, W filament) can perform elemental analysis on a 50 nm ternary thin film deposited on a 4-inch (10 cm) Si wafer. They're interested in scanning across the wafer to see composition variations (of the metals Pt, Au, one other metal). Apparently, detection limits are not the main concern and even a 'large' spot size (around 1 mm) would be useful for their purposes.
I've tentatively said yes, but with the disclaimer that I need to look ionto this because I've never worked with thin films before, and no one in our facility has either. We work mainly with geological samples (minerals & glasses). However, we do have pure metal standards for Au and Pt (and other metals) and if broken up the thin film/wafer sample could fit into our sample holder.
I've seen there are a few thin film releated entries in the forum and since we have PfE (I'm still a beginner) I can in theory utilize the functions John has built in for this (multiple setups and volatges). However, I'm not sure of the fundamental limitations of analyzing thin films, how much effort this will be for what gain/result, how long it would take, etc.
Is a 50nm film too thin?
If not, would there be any detectable heterogeneity with such a thin film for these metals? What about the Si beneath overwhelming counts? Just not analyze for it?
If anyone has experience with this, or know who I could contact, I would really appreciate help. Firstly I'm keen to know if it is a definite no-go so we can advise the researchers. They had already checked with our Chemistry department for TEM, but their sample is too large for that method and they were referred on to me. Secondly, if it would work, what analyzer crystals, count times, etc. should I be using? (we have TAP, LIF, LIFL, PETJ, PETL)
Thanks,
Kerstin
Quote from: KerstinGruender on February 02, 2026, 03:36:55 PMIs a 50nm film too thin?
50 nm films are totally doable using multi-voltage EPMA using the method of Pouchou:
http://www.geology.wisc.edu/~johnf/g777/Misc/Pouchou-1993.pdf
In fact, it's been a good income stream for my lab! By using measuring k-ratios at multiple keVs, one can obtain both composition and thickness. I personally have utilized this method to measure 2 nm thick thin films with extended acquisition times for improved precision.
You can process the k-ratio data in either the STRATAGem or BadgerFilm software (BadgerFilm is free!):
https://smf.probesoftware.com/index.php?board=37.0
The key is to first model the k-ratios as a function of beam energy for the rough geometry and composition as seen here using STRATAGem:
(https://smf.probesoftware.com/gallery/395_03_02_26_9_04_45.png)
You can do the same in BadgerFilm. I modeled Au-Pt (and Se) because it's important to model (and analyze) all the elements in the film. The substrate is also measured to obtain thickness information.
Then pick 2 or 3 or 4 keV values that will produce the largest differences in k-ratios. This "leverage" is critical for analytical accuracy. For this material I would choose 10, 15 and 20 keV. Then acquire the data at these voltages in Probe for EPMA using the Multiple Samples Setups button in the Automate! window:
https://smf.probesoftware.com/index.php?topic=1498.msg11441#msg11441
Let me know if you have any more questions.
Thank you, John!!
I'll look into this and get back to you if more questions. I'm sure there'll be some, but I'll do more reading first :)
Cheers,
Kerstin