I am using Hitachi S3000N VP-SEM with EDX equipment to investigate the composition of deposited Cu and CuW and Ni films on lightly doped Silicon substrates (Phosphorous).
Unexpectedly, there is a peak of Sn (at 3.4 KeV) coming out in the three measurements (I have a separate sample for each material).
What is the cause of Sn appearing ?
Hi Ahmed,
How thick is your films? Because, SEM-EDS is not an accurate method to analyse the composition of the films that are less than 4um. Instead you can use RBS to measure the composition.
Is there any possible contamination during deposition of these films? I am not sure for the appearance of Sn peak but analysis using RBS, LAICPMS and XRD (if the films are crystlline) may confirm if there is any Sn contamination.
It could be Si sum peak. Usual artifact at too high count rates.I do not have tables at hand, but you can check it it is close to 2*Si(K) energy.
Hi Radha,
The films are ~100-200 nm . I'm sure that there is no tin contamination.
Hi Vladimir,
could you please elaborate what is "sum peak"? Can you provide references or keywords I can search with?
Hi Ahmed,
If your samples are too thin, then EDX is not suitable for chemical composition analysis. Try RBS (Rutherford Backscattering) or LAICPMS.
Hi Vladimir,
The k-alpha peak for Si is 1.74keV and it correlates with 2*Si energy. So possibly it might be an artifact due to high count rates.
Hi Ahmed,
Vladimir and Radha are right; it is the sum peak ('simultaneous' detection of two photons) at high countrate. Please see in your first spectrum another example: the tiny 2,6keV peak; it is the coincidence peak of the 0,9keV one and the 1,7keV peak. So you suffer from to high countrate. This effect is called 'pile-up'.
No real Sn; don't worry.
Yes, it's sum pick, as mencioned. If your film tick is less than 5 µm, is better use other method as Radha wrote above.
I would discard all "advises" about thin films. EDS is good for qualitative analysis of thin films (and for quantification with proper software also). You just need an experienced operator.
I've been Thinking about Vladimir said. Excitation energy of Cu La = 0.930 KeV and Ni La = 0.851 KeV, so difference between them is 0.079 keV. If you put your EHT = 5.0keV, penetration will low down , but resolution of EDS is aprox. 0.130 keV. I'd like to hear ideas of Vladimir how to do.
@I. J. Sayeg
As you pointed it, low voltage EDS will not work for Cu+Ni. We are limited here with K-lines, which is good enough for qualitative analysis. For quantification special software is needed ("thin film analysis" or something like that). It is usually an option which is not included in basic EDS software package and should be bought separately.
Thank you Vladimir, I've found easily subject after your suggestion.
- Badges
- Science topic
- Similar topics
- Counseling Psychology
- Confusion