Dear David,

The carbon probably is coming on the surface during your EDX measurement. In your SEM chamber you generally have not ultra high vacuum, and therefore there are hydrocarbon molecules from e.g. pumping oils in the chamber. These hydrocarbon molecules adsorb on your samples surface and if the electron beam hits the surface the molecules are decomposed, leaving an amorphous carbon layer of some nm thickness. This is a well known phenomenon.

If you have an inlense detector on your SEM instrument and use very low electron energy (e.g. 1kV) you can even image this contamination layer. It is visibly as a rectangular dark area were you have made your last SEM scan.

Generally, you can not completely avoid this contamination. You can only minimize it by using oil free pumps, by using higher energy for your EDX measurement (because then you are less surface sensitive) , by using short measurement times, (because C-contamination increases linear with time) or by integrating over a larger area during your EDX measurement (because then the local electron dose will be lower and therefore C-contamination growth is slower).

Even on a completely clean surface, typically, you will always find some percent of carbon during EDX.

Best regards

Kirsten

I agree with Kirsten Ingolf Schiffmann . It is very common to find carbon at the surface of samples during surface analysis.

In addition, since you are using a new target material, I would recommend a long preclean (sputtering with the shutter closed in front of the substrate) before proceeding with the deposition.

You can check if your carbon coming from SEM.

Perform analysis. Do not move you beam/field of view for 10-20 min (so you'll let contamination to accumulate). Repeat analysis. Does %C increase?

-pre- sputtering of your substrate before the deposition.

Thank you everyone for your answers and help.

I do pre-sputter before depositions but only for 5 minutes, is this sufficient? What would you suggest as a sufficient pre-sputter time?

And what might be the typical percentage of C present from the EDX measurement, or a way to work it out?

I have asked this question because I think my results are quite high in terms of C content.

For my films there appears to be 30 - 50 at% C from the EDX. This varied with thickness, the thicker films showing the smaller at% C. If it is just the EDX analysis process introducing the C contamination then I think this would make sense as if each scan introduces the approximately the same amount of C, then the thinner films would show a higher ratio. However, these numbers still seemed high to me, but my experience here is limited. The films were 100 - 500 nm in thickness.

The older target I have was made up to be 81:19 at% Fe:Ga, which most of these film results come from, the recent EDX results were 66.3 : 13.9 : 19.8 , Fe:Ga:C.

Which again I thought was quite high.

This target is a couple years old and has been through some machining processes and a number of sputtering depositions over that time. And I thought it may have picked up C contamination over that time from one of the processes.

I have a new target which I plan on comparing analysis of to the first, I have sputtered from it twice, low deposition rate, got 40 nm of film down, which showed ~89 at% C.

I also had areas of the substrate layer that were masked during the sputtering process and had no film analysed also, the results I got back showed no significant trace of C for this, so my thought was something may be happening during the sputtering itself.

with which power?

" This varied with thickness, the thicker films showing the smaller at% C. "

What about your substrate? It is possible you are picking C radiation from it. Just run substrate analysis with no coating. Also you may want to run analysis of something like (cleaned) aluminum stabs for specimens (do they show carbon?)

EDX power is 20KV, sputter power is 60W

I have run EXD on the substrate, and a Ti target I have, both did not indicate the presence of C.

I now think it is the target itself that is probably the source, that it is actually FeGaC in composition, and the C was initially overlooked by the supplier in their analysis, they used ICP-OES for the metallic element and LECO for gas elements.

The EDX I have shows the Fe:Ga ratio is the expected 81:19 at% if C is excluded from the analysis. When the C is included Fe:Ga:C is 61:15:24 at%

Thank you all for your help. I have learn a lot from this discussion.

According to your information

1. the EDX results of your film, presence a larger percentage of C present, as the film thickness increases the relative percentage of C decreases.

2. the EDX of your target shows the presence of C on it.

3. to do pre-sputter before depositions but only for 5 minutes

4. the manufacturer document of the target does not indicate any trace of C.

Suggestion

1. check dip profiling present of C to your films and your target using XPS.

I guest at the subsurface of films should has higher present of C

due to at the begining the target surface still presence of C

2. If presence of C below the films surface deep enought, do a long pre-sputter

(the shutter still closed in front of the substrate) before the deposition.

Contacting the manufacturer they did not test for C, which is why it was not present in their analysis. And some additional reading up on ICP analysis indicated that this method is not suited for C detection. This is true for my first target also, recent EDX of this target shows FeGaC also.

For my first target, I saw the relative percentage decrease in C with the films.

For the new target I have only sputtered from it twice, ~1 hr 40 mins sputtering time, and the film showed high C content.

The EDX of the new target showed significant (approximately equal) presence of C on both sides of the target, sputtering side and non sputtered side.

My reasoning is that the new target has only been through its manufacture process and the sputtering. The sputtering affected only one side of the target, and should remove layers from the surface from that one side.

If the carbon is appearing as a results of the process there should be a a notable difference between the results for both sides of the target, and there is not. Additionally if it was primarily a surface contaminant the sputtering process should remove it and again show a difference between the two sides, and there is not.

As both target sides appear of similar composition, it's been through a limited number of processes, and analysis for C was not initially included, so I think it is reasonably to conclude that the target itself that is the source of C I am seeing.

I have also run an EDX on another older target, Ti, which has had limited usage and it did not indicate C, as well as on an area of the substrate where I deposited my film but was masked from the deposition, this also did not contain C.

A dip profile or way to compare C content for different depths is a good idea, however I am unsure whether XPS is available to me at the moment, and am reliant on EDX for composition analysis.

Mystery...

One more possibility: you have oil backstreaming in your sputter coater. Checking with Ti target in this case is questionable, because ionized Ti is a good absorbent of gases (it is used in ion pumps) and may clean vacuum during pre-sputtereing. It does not explain everything you've got, but neither other possible causes do. May be it's some combination?

And I somehow do not believe targets can contain large amounts of carbon. But of course I do not know what your targets are and from what manufacturer.

Get a few targets from different suppliers and compare EDS spectra. Carbon reading doesn't come from your SEM, since EDS doesn't show C on Ti target. As a "religious" precaution - put all new targets through O2 plasma cleaner prior to testing.

"Off the cuff" and pure guess, but nonetheless: new target could have been made from wrong alloy, or cut from a sheet of cold-rolled foil that became impregnated with oils (source of C) during rolling process.

P.S> Coat a sample by Ti sputtered from the target which itself doesn't show C in EDS - if you see C in sputtered film then it comes from coater (clean, check traps, etc...), if you see no carbon then it was coming from your new target, replace it and return to the vendor for a refund.

I agree with Kirsten Ingolf Schiffmann.