I would like to know composition wt% (silicon and aluminum oxide). Have done only SEM-EDX, is it okay to prove that (for publication) or necessary to go for XRF? From DEX received only elemental wt% of the powder sample.
It depends mostly on your paper. Does it require full quantification, or qualitative result is OK? For better quantification you need a WDXRF, for better EDS analysis you need oxide standards.
There are different ways to do EDS - if you want to be rigorous about it I would do an EDS analysis, find out what my percentage is, then see if you can find a sample that has approximately that concentration (or get two samples around that). Then you repeat the analysis with the three samples: The one with less silicon than your sample, the one with more Si than your sample, and then your sample - this should give you reasonable results.
In the absence of this you can of course do a standardless analysis, but understand that while the repeatability is reasonable, the absolute numbers may be off.
In principle, XRD is not necessarily better. You still need some form of standard to compare against. I think there is nothing wrong with publishing standardless SEM-EDS as long as you are specific what instrument and what software you used. If you can estimate the uncertainty of the numbers and publish them I think you should be fine. If course, having the standard is best.
You can transform wt% into at% very easily in your software (it is simply a click away) or you calculate it using formulas you can find in internet. They are related to the relative mass of the respective elements.
Oxides are not a real problem, but experimental data never match exactly, anyway which technique you are using. If you have only SiO2 and Al2O3 EDX should be enough because you only want confirm very stable phases. If you expect some dirty effects it becomes a bit more challenging, e.g. if you want characterize alumosilicates by the ration of Al/O and Si/O. Then it is better to go to WDX but also there you need some standards as usual for this technique.
The XRF technique would help you to get quantitative composition in terms of oxides, present in bulk of powdered sample. On the other hand SEM-EDX would be useful to find elemental heterogeneity present on the surface of the powdered sample exposed to SEM tool. In directly, if one performs XRD of the powder, the result is mineral peaks in terms of intensity corresponding to a mineral form. The presence of mineral in the powder would help you to anticipate the element present in the sample. Moreover, NMR technique would also help you to reveal Si and Al nuclei in the sample.