Hi, Nima,

Typically, the EDS overall accuracy is approximately 1wt% and the sensitivity is approximately 0.1wt% for all elements it is capable of detecting. However, from my experience at least, there is a lot of inconsistency when you try to measure C content on steels, especially low carbon steels. In many SEMs there appears to be an unavoidable contamination of the internal walls of the vacuum chamber with carbon, especially if you inspect organic material as well. You can get carbon peaks on everything, including aluminum alloys, etc. I know a lot of people just disregarding the carbon data (some just put “sample coated with carbon” by default), or account for a known bias (the bias can be of a few percents sometimes). Of course it all depends on what type of SEM you have, with or without vacuumed antechamber, etc. There are instruments that can clean the chamber and therefore make your measurements more accurate, but after a relatively short while you need to clean it again.

Fluorine and sodium don’t have the same problem as carbon, but be aware of the F K alpha overlap with Mn L alpha and Fe L alpha. Sodium overlaps with Zn, but since you don’t have Zn in the steel composition, you don’t have to worry about.

Hi, Nima,

Typically, the EDS overall accuracy is approximately 1wt% and the sensitivity is approximately 0.1wt% for all elements it is capable of detecting. However, from my experience at least, there is a lot of inconsistency when you try to measure C content on steels, especially low carbon steels. In many SEMs there appears to be an unavoidable contamination of the internal walls of the vacuum chamber with carbon, especially if you inspect organic material as well. You can get carbon peaks on everything, including aluminum alloys, etc. I know a lot of people just disregarding the carbon data (some just put “sample coated with carbon” by default), or account for a known bias (the bias can be of a few percents sometimes). Of course it all depends on what type of SEM you have, with or without vacuumed antechamber, etc. There are instruments that can clean the chamber and therefore make your measurements more accurate, but after a relatively short while you need to clean it again.

Fluorine and sodium don’t have the same problem as carbon, but be aware of the F K alpha overlap with Mn L alpha and Fe L alpha. Sodium overlaps with Zn, but since you don’t have Zn in the steel composition, you don’t have to worry about.

Hi Nima, EDS is not accurate for low atomic number elements, C, B, N, O, in new EDS system this detection efficiency is improved. But is not recommended as technique for the Qualy/Quantyfication of very low atomic number elements. 

thanks guys, you provided some useful information,

some of my results showed 20 wt. % C when I conducted a point analysis on an intermetallic or an inclusion! and sometimes even around 50 wt. % when doing an area analysis! that drove me crazy!

Ok, so what is the recommended alternative to measure C and O accurately?

You need to do regular chemistry for bulk chemistry, if you can avoid the microscopes. For carbon you can try ICP (Inductively Coupled Plasma). For oxygen you could try one of the LECO instruments dedicated for oxygen, nitrogen, and hydrogen.

Otherwise you can try WDS instead of EDS, if your microscope is equipped with it (WDS is much more precise). Also XPS (x-ray photoelectron spectroscopy), but it's an entirely different instrument. Or AES (Auger electron spectroscopy), but it's for surfaces only and is semi-quantitative (although more precise than EDS).

Roxana, could you pls provide a reference that Na overlaps with Zn?

Meanwhile, I found some information. Feel like to share here;

Looking at the respective X-ray lines of Na and Zn, the Na Kα might overlap with Zn Lα, but the Zn Kα is further away so you should be able to differentiate them by looking at the presence of Zn Kα. However, to be able to see Zn Kα, you need an accelerating voltage of at least 20kV.

Na: Kα 1.0410

Zn: Kα 8.6313, Lα 1.0118

Credits go to "Min Hao Goh" who has provided the above information for me.

Hello, Mohsen,

This is an excellent paper I refer to a lot:

"Newbury, D. E. (2009), Mistakes encountered during automatic peak identification of minor and trace constituents in electron-excited energy dispersive X-ray microanalysis. Scanning, 31: 91–101. doi: 10.1002/sca.20151".

Each time I read it I find something useful. You are right, the confusion Na/Zn is at low energies amounts, I should have stated that.

Hello Nima Mossayebi ,

Could you find an explanation for this weird behavior? I am also seeing a huge C peak when studied in a metal system, a bit shifted though.