I am doing LC-MS (LC-QToF) of a phosphonate compound (nitrilotris methylene phosphonic acid), for which I add derivatization agent trimethylsilyl diazomethane to each sample and wait 2hours before running them on the instrument. I do not have any mass label internal standard for the compound, so I am using caffeine as the internal standard. I am using C18 column. In the beginning I was getting disturbed peaks with a tail from the start of the runs, then for two weeks I got clean peaks. However, I started getting the same type of disturbed peaks again. What could cause the disturbed peaks in LCMS? I have attached a screenshot of the disturbed peak I am getting. My internal standard peak also looks like this.