If we do some spectroscopy technique for only one sample, is it enough for data analysis or we have to do multiple samples from each treatment for getting authenticate results to publish in research paper?
Well, the unsatisfying answer is to reproduce as often as you can afford without losing too much time.
There are processes that are quite safe with respect to reproducibility while others aren't. If you have no benchmark on your method, I suggest characterizing everything repetitively in the beginning. If it turns out that the results are really similar and your process is reliable, you can start decreasing.
Regarding measuring the same sample: If your sample is homogenous, measuring XRD and FTIR once should do it since they are bulk methods and have an "intrinsic averaging" in them. For SEM, please be sure to acquire enough areas so you don't miss or overinterpret features. On the other hand, if you have e.g. a mineral sample that's obviously heterogenous, measuring XRD and FTIR in different locations may lead to different results.
This answer may have been somewhat unsatisfying but that happens.
Analytical investigation on identifying the minerals can be performed on an average of three samples for XRD and FT-IR, for SEM analysis the sample must be performed with EDX to correlate the XRF analysis and its chemical composition.