http://www.charfac.umn.edu/instruments/eds_on_sem_primer.pdf

http://www.charfac.umn.edu/instruments/eds_on_sem_primer.pdf

https://en.wikipedia.org/wiki/Energy-dispersive_X-ray_spectroscopy

Yes, the wt% is telling you about the relative concentration of the element in the sample if the results are normalised. If they are not normalised it gives the absolute concentration of the element. It can also be given as atomic %. See my previous to this in the link below.

https://www.researchgate.net/post/Can_you_tell_anyones_exact_definition_what_is_the_Atomic_and_Weight

He, answer Ian J Slipper  is corectly. You have to be careful what you analyze by EDS. If you analyze the light elements as (O,C,F, B...) result is obtained with a large measurement error. Unacceptable in the correct measurements for publication. If you examine the heavy elements obtained result is very accurate.

The results can be different, depending on probe current, on accelerating voltage. The electron beam can excite some spectral lines depending on few parameters. Also, the deepness of signal can be different, and if a larger volume provide signal, the elements with smaller content can disappear from spectra.

The wt. % does give in a way the concentration of atomic species in the sample. The molar (or atomic) ratio, however helps when you have a known single phase in the sample; you can infer the stoichiometry and the consistency of the obtained atomic ratios with the presence of that phase as a single phase in the sample

Thank you friends for your useful contributions  and answers. I quite appreciate. 

Basically, there has to be some relation between wt% and concentration, just find this relation. More closely related to the concentrations are at%. This is why I prefer EDS results in at% expression.

If you want to compare few analyses of samples then you have to keep on mind the poor accuracy of qualitative EDS analysis. Try to do the same measure from the same place on the sample and probably don´t get the same results exactly. This is because in principle the EDS analysis is very sensitive on many factors (beam parameters, topography of sample, several acquisition settings, electronics and external field noises, atomic number of element, detector type, and so on …).

So, in case you want to get the correct compositional results, you have to perform quantitative EDS analysis. However this procedure brings the other problems. You have to carry out ZAF correction, which is time consumable, and compare your spectrum with spectrum of standard (calibrant). I guess that you don´t own relevant standards (as most of researchers) because if you want to make precise comparison of sample and standard spectrum, you have to have the standard with the same composition as your sample, not only the pure element standards. After overcoming these procedures you can reach the theoretical accuracy of EDS analysis, which is approximately 1wt% (for the middle atomic number elements) measured on the flat and polished sample. Surface roughness degrade this value.

What I want to say is that EDS analysis results are not very reliable and you are just about rough idea of composition.

For thoroughly performed EDS analysis (quantification with standards, no light elements) wt% IS concentration. 

wt% stands for weight %, so it's indeed the realtive concentration of an element in mass (weight, it's the same when in %). Usually the software computes these concentrations by comparing the intensities of X-ray peaks of different elements, using the Cliff-Lorimer method (and k-factors). As someone already mentioned, soft X-rays (energy below 3 keV) can be absorbed and therefore, quantification of elements using soft X-ray lines could lead to errors, where the sample is thicker than the few hundreds of nm (this is known as the thin-film approximation). For a comprehnsive review of the Cliff-Lorimer method, its limitations and modern quantification techniques for EDS signal, I would advise the reading of Watanabe (2006) 10.1111/j.1365-2818.2006.01549.x

Best regards,

Daniele

Wt.% gives the concentration of the element in terms of the mass fraction of that element in the sample. The accuracy of the results depends on whether all elements in the sample (oxygen included if present) are included in the analysis. Atomic (or molar) ratios of the elements are related to the wt% through the atomic masses of the elements in the sample.

How to find the ratios of specific elements identified in SEM-EDS in order to properly identify an unknown? Thank you