Dear Sachin K. Yadav , it is difficult to make a statement without knowing the nature of the sample and the conditions of the TGA analysis (is there a magnetic field applied?).
Although it is hard to deduct what happened without knowing the nature of the sample, as previously said, maybe the problem relies on the "real" mass of the sample. TGA analysis normally measure how the weight of a sample decreases against temperature, but, if the initial mass is not the real mass of the sample, then everything you are seeing might be wrong, or might be awkward, as you observe. I tend to think that the mass of the sample was not stable when the analysis started. Normally, when setting up, the operator puts the sample in the equipment, and waits a couple of minutes until the weight of the sample is stable. If this was the case in your experiment, and all the cares had been taken when measuring, then I do not know what might have happened. Maybe your sample is reacting with the gas that is used during the analysis?
Hope this helps a bit at least!
PD: I don't know which TGA do you use, but if your TGA is from TA Instruments, the TA Universal Analysis software (free, downloadable) allows you to read the data file without using Excel. The format of the file is .001 I think. Maybe reading directly with that helps you to discover the problem!
This could be attributed to the side/over reaction of your sample with the purge gas. Moreover, the wrong sampling might be responsible for this unwanted happening. You should carefully handle while sampling & analyzing the considered specimens and take at least three raplica for better clarity.
I obtained similar results when I worked on electrospinned sample and I had a very low sample amount. By increasing the mass, the TGA curve returned "regular"