It is impossible to recommend kV and magnification, it depends on your task, specimens, and microscope. You can find a lot of publications by googling “SEM fossil wood”.

Generally, it is possible as Vladimir already mentioned, but you have to note that analysis strongly depends on the conductivity of your samples and their stability under high vacuum. Fossil wood should be relatively stable under high vacuum, but due to the large carbon content I wouldn’t recommend going to high magnification and high kV.

Depending on what you expect from the EDS analysis - trace elements other than carbon I would recommend up to 10kV for decent EDS spectra under low magnification - up to 5000x. Since this are samples of biological origin, in my opinion, typical morphology should be seen from 100x - 5000x magnification.

Fossilized wood is not a “biological specimen” in terms of specimen preparation and treatment. It should easily withstand high voltages; you can use any voltages you may need and any magnifications achievable with your SEM. No problems. You can use fractured specimens, or (it may be better for EDS and BSE) polished. For polishing specimens are usually embedded in a resin for ease of handling. I strongly recommend using not only secondary, but and backscattered electrons (BSE) for imaging; you can get pretty useful results, especially if specimen is polished. For EDS you should use 15-20 kV, otherwise many elements of mineral can be not detected. Special case – EDS of carbon. It is more difficult and usually requires lower voltages. If your specimens can be observed without any coating at 2-3 kV, you can use that mode for EDS of carbon.

Thank you for your answers.

Yes, the samples are quite stable.

They are unpolished, and unclean. We are not cleaning them for now, as cleaning will remove the micro-organisms off their surface. We detected some lithotrophe microorganisms, so our concern is not to lose that information as our first step is to record the pathology of the fossils. We are also not polishing nor embedding them for the time being, as we don't want to lose the microcavities and other pathology from their surface.

We checked some samples using 20kV and magnifications x70-5000. We used BSE and no coating; just carbon tape around the sample for conductivity.

In EDS we did detect some carbon, but very small percentages, so for time being we are ignoring it, as it could originate from the microorganisms. After we will have recorded the pathology, we will chemically clean the samples and then check them out again. EDS analysis showed a high content of Ti (around 70% and higher), Al (but this could also be deposits from the soil in which the fossils are burried in) and traces of metals such as Fe etc.

We spotted some papers on SEM for fossil observation, but they were mostly anatomy-orientated. Do you know any papers including EDS analysis for wood fossil? How likely is it that our samples include C, and at what percentages should we be expecting it?

I have not worked with wood fossils, but it looks like interesting thing.

There are some papers I found looking for “sem microanalysis wood fossils” in Google Scholar:

http://www.sciencedirect.com/science/article/pii/S0146638011001744

(see Supplemental data file).

http://jgs.lyellcollection.org/content/160/6/857.short

(only abstract)

http://www.sciencedirect.com/science/article/pii/S0040816605000911

I have not read these papers. Be careful: a lot of papers outside of usual field of quantitative X-ray microanalysis (material science, geology) contain erroneous results and conclusions.

As for quantitative analysis for C on non-polished specimens – forget about it. It will be wrong. Anyway, for good quantitative results for any element you need standards; standardless analysis on specimens like yours is useless. May be you will be able to compare different regions of your specimens (carbon content) if difference is big enough. For carbon better to use low voltages, 2-4 kV for better signal/background ratio. Another pitfall: using 20 kV on non-coated specimens definitely leads to charging, therefore effective accelerating voltage for EDS will be lower than applied (20kV). It is another source of errors in quantitative analysis. Better to stick to qualitative one and compare intensities at regions of interest (multiple measurements for good statistics).

Since you are interested in morphology of surface, may be imaging in secondary electrons (SE) will give you more information. But for SE of non-coated specimens you need to use low voltages, 1-3 kV (you need to find the best voltage experimentally). Do not use high voltage before a low one – specimen will be charged and SE images at low voltage will be awful.

You have an interesting investigation. Good luck.

I have used SEM/EDS in examining egg-shell from dinosaur eggs, the magnification I vary as I looked for and then focused on features (pores etc.). I used 10-20 kV

EDS was helpful in determining changing Mg concentration in the calcite from the upper-layer of the egg down to the lower layer.

John Paul Jones

MSU

Dear André,

I'm sorry for the delay in answering.

We are using carbon tape, so we do not have to coat the samples with carbon or gold. This way the technique is totally non-destructive as no alterations are caused to the specimens.

I have looked at many petrified wood specimens using SEM. I simply mount fractured pieces on aluminum stubs with a dab of quick-set epoxy from the local hardware store, and sputter coat samples with palladium or gold. Carbon  tape does not provide  very strong adhesion,and it doesn't matter that the tape is conductive, but the specimen itself is not, so electrical arc tracks can be a big problem for uncoated specimens. The biggest problem I've seen is that fossil wood is sometimes porous and absorbs moisture so  that it is hard to get a good vacuum pump down. I commonly dry samples at 110degrees C for  an hour or more before viewing.  Sometimes I clean my samples in water in an ultrasonic bath to reduce the amount of surface dust, and maybe do a final rinse with acetone or alcohol. 10 KeV gives good images for fossil wood, the grayscale tone range seems better than it is with higher voltage. But I sometimes use 15 KV, which gives a bit better resolution at high magnification, and also gives stronger peaks for EDS analysis. EDS elemental data made with a standardless ("ZAF"") software program is only semiquantitative, but it is great for showing the kind of mineralization (silica, calcium carbonate, calcium phosphate, etc..) and for revealing the presence of relict organic matter as indicated by a carbon peak. For EDS analysis, open up the spot size to a pretty big diameter, which will yield a much higher  X-ray count  rate. For viewing or photos, it's a simple matter to readjust to a small spot size.The Pd or Au coating does show up in the EDS spectrum, but it isn't really a problem since you expect it to be there.. And with coated samples in regular high vacuum mode, it is possible to get great images even at high magnification. I'm attaching a couple of papers that show some of my wood pictures.

George

Article Conservation of an Eocene petrified forest at Florissant Fos...

Article Mineralogy and geochemistry of late Eocene silicified wood f...

Dear George, thank you for your answer and your papers. The fossil wood you are looking at is quite different than the ones we are looking at. I met Herbert Meyer in Thailand last year and he showed me some pieces from one of your trees there at Florissant Beds. They are indeed very porous!!!! I will carefully read the papers you have uploaded for me and we will be in touch either here, or through private communication.