We are using a 5973 MSD on Volatile Water analyses. Here is the problem:

We have average Internal standard responses of 302000 663000 750000 and 515610 for the initial calibration (at 50ppb).


This held true for the first four runs.

However, after this there was a sharp increase in sensitivity with the responses more than doubling. The samples were mostly clean as they are drinking waters.

has anyone seen this before and know a solution?