Using SEM-EDS analysis I worked with samples covered with a very thin layer of carbon, so I recommend you carbon sputtering or other carbon deposition techniques.

We usually use sputtered gold but, depending on the apparatus, it is possible to analyze samples without additional coat

Thank you very much for your answers.

I think I should tell you more details about my situation. I have a ceramic sample, it was penetrated by some molten metal through its pore and grain boundary. So I am trying to use EDS to analyse to investigate this kind of penetration. The pore size of my ceramic is very small, around 20~200 nm, so I think I will confuse between the pores and the gold particles depositing on the ceramic surface by sputtering method. In addition, I am wondering that if my coating layer is very thin (less than 5nm), is it possible that the electron beam can penetrate the coating layer to reach the ceramic/metal surface and give me the accurate result of analysis (no peak of coating elements like Au, Os…). Otherwise, I must try the carbon coating method (this method is not available at my lab). Could you give me your advise?

Best regards

We usually use a very thin layer of gold also for porous oxides and it works fine

Usually thin layer of Au (Au-Pd) is mostly transparent for electrons and X-rays at 15 kV, but you will have small peaks of coating metal in your spectrum. If elements of interest in your specimens do not overlap with peaks from coating material you can get pretty decent spectrum (of course in case of publication you’ll need to emphasize origin of Au (and Pd?) peaks). If you have spare specimens you can try it. But of course carbon coating is preferred for EDS.

Try to use as low voltage as possible for increasing spatial resolution. Try BSE if your metal has higher Z than ceramics. BSE have much less noise and much higher spatial resolution than X-rays.

As an X-ray guy (;-)) I have somehow to insist on EDX (EDS) - backscattered electrons give of course some Z-information, but they are not quantitative. So if you are interested in the type of atoms incorporated in the pores and grain boundaries, you should always try EDX. But as Vladimir mentioned, try to use a coating that does not interfer with the materials you are looking for! And in this situation (i.e. if you coat your sample under in-situ conditions (i.e. without contact to air between coating and EDX - measurements or SEM investigations), any conductive material as coating may help. good luck, Dirk

Thank you very much Vladimir and Dirk for your very useful information!

Son, I believe your question is still related to your specimen of infiltrated with metal ceramics. If so (I insist :)) BSE can give more information on distribution of metal.

Thank you, Vladimir. Yes, I am still working with my metal-ceramic samples.

Dear All, I have a question related to this topic discussion. Is it possible to run qualitative EDX without a coating?

It depends... Can you be a bit more specific about your task?

Hi Luca !

In high vacuum mode with traditional SEM , you need to evacuate the electrons coming from the beam. Now without coating it still possible to perform qualitative analysis (but not quantitative) if:

- you are using a Field emission gun (FEG) microscope that allows you to work with low tensions (up to 3kV). Caution, in that case you need a specific detector and all the elements are not available...

- you are working in low vacuum or environmental mode but in that case, you have to be aware about the skirting effect.

I suggest you to refer to this book (in French):

'Microscopie électronique et microanalyses', groupement national de microscopie electronique à balayage et de microanalyses, François Brisset, EDP sciences ISBN: 978-2-7598-0082-7

Thanks Samuel , If we understand we can make meaningful EDS qualitative analysis with environmental SEM at low tension (3kV) without coating on a cement paste , that’s right?

If so, which EDS detector is compatible in this mode? Which elements can be no detected with respect to the mode with high kV with coating on a cement paste? Thanks

Dear Son Thanh Nguyen,

As we know that sputter coaters are a great option for imaging applications. Adding a metal coating such as (Au or Pt) allows the user to image a nonconductive sample at higher voltages. In general, higher voltages coincide with a smaller effective beam diameter (higher resolution). A thin metal coating not only provides a conduction path, but a way to dissipate heat buildup and minimize beam damage with beam sensitive samples. These heavy metal coatings also provide a higher secondary electron yield, which can improve signal to noise.

One disadvantage may be absorption of low energy electrons and X-rays. This may impact sensitivity of backscatter electron (BSE) imaging especially in areas with low atomic number. Also, the metal X-ray lines could interfere with the X-ray lines inherent to the sample being analyzed. If very high resolution imaging is of interest, the user may be better off with finer coatings and more exotic metals, such as Iridium or Osmium. Carbon evaporation will also provide a conduction path for a non-conductive sample to be imaged at high voltages. Interfering X-ray signals are not typically an issue and low absorption will allow for better BSE signal to the detector. Therefore, if BSE imaging or X-ray analysis is of interest, this coating method has its advantages.

So if I summarize my above discussion related to coating, we can say that if the intention is to perform some analytical characterization such as EDS or intensive imaging with your BSE detector then carbon evaporation is a good choice. For high magnification imaging with low atomic number samples, adding a metal coating has its advantages.

Best Regards