Hi,

I'll try to do XPS on the same sample to see whether Si peak also appears there. If XPS shows the presence of Si peak, then it means there is some silicon incorporation/contamination in your sample.

Hi Jea Cho, thank you for your comment. Unfortenately, there is some minor problems with the XPS instrument in my laboratory and it requires few weeks to be fixed. I tried to investigate the sample with XRD, but the XRD results do not show any Si-related peak...

If you have an older detector, you may get a small silicon internal fluorescence peak from Si dead layer of Si-Li detector.

Hi Son,

I'm not too sure how helpful will the XRD be, as XRD will be able to reveal the presence of Silicon-based compounds only if you have some degree of crystallinity.

For example, if you have an amorphous glass of Si-based compounds embedded in your sample, XRD will not show any peaks related to it, thus you will not be able to extract any useful information from XRD spectra related to Silicon.

Assuming your sample is contaminant-free, then I think what Vladimir wrote above is most likely what is happening to your sample (which I just learned too).

Thank Vladimir and Jea,

Maybe you are right, there are possibly some problems with my detector... To be sure about that, now I am preparing another sample with the same conditions. I will tell you the results after testing the new sample with EDS.

Hi,

You peak appears indeed very small in intensity. When I worked in the past

with EDS and see a lot of Si samples, sometimes some other elements

have peaks that appear at the same Energy or very near : have you consider

the eventuality that what You see could be, i.e. W (tungsten)? I apologize

if this is an almost trivial consideration, but your question is very interesting.

Vladimir has right, You can simply put any pure aluminum stub or pure copper sample and check wit the same conditions. System peaks like Si should appear in the same way. This peak is really small and usually is below quantification limit. In common work I would skip this peak.

Hello Son,

instead of searching for pure metals which can be difficult, why not just to use a simple coin. The composition of coin metal is always available in internet. For example, I always do it with 1 euro cent. It is the simplest way to detect a problem.

Good luck.

Hello Son,

looks to me like a small silica grease contamination on your sample. this can happen when handling the sample with `dirty ` tweezers. silica is a component of liquied soap and can be easiliy transmitted by hands touching tweezers, which is then covered.

Bert

Dear Dr. Son Thanh Nguyen,

Did you polished the sample using silicon carbide or colloidal silica?

If so, the Si peak may be derived from the particles containing silicon.

If you analyze a darker particulate observed in BSE image with EDS point analysis, you may detect Si. 

Sincerely,

Tohru Awane,

Fukuoka, Japan

Finding Si peak in your spectrum may be because of these reasons:

a) Using antiquated or un-calibrated Si-Li detector in the EDS system.

b) Contamination of the surface of the sample by abrasive materials (SiC) or surrounding dust.

c) Wrong interpretation of spectrum: some element have over lap peaks such as Mo and S

Si peak overlap secondary peak of ytrium. It may be the reason but, for network security reason, I can not load your file to check more precisely.

If your sample surface was polished with Si-colloidal liquid and was not carefully cleaned by ultrasonic "shaker" , you may have residual Si film  or Si particles.

Due to the carbon film on the Cu-grid and the silicon substrate on the mounting base small, intense peaks corresponding to C-Kα1 and Si-Kα are also seen in the EDS spectra.

The question is a little bit old (September 28, 2013). However for later notice

EDS spectral artifacts are:

- Internal fluorescence peak

- Sum peaks

- Escape peaks

https://myscope.training/#/EDSlevel_1_7

Mehrdad Khatami

Mehrdad Khatami, the appearance of Au peaks might be due to the coating of your sample with gold as a result of charging effects.