by
lmh » Tue Oct 24, 2023 12:59 pm
The logic, I believe, is that even if you've established that the instrument was working reliably by running a beautiful calibration curve before you injected the sample, it might have gone wrong as soon as it finished the last calibration point and before it ran the sample! Therefore if you want to be sure that the sample has been measured correctly, you need some sort of standard or quality control sample run after the sample, to check that all is still okay.
Also, of course, if your method was validated on the assumption that you were always averaging two calibration points, you'd be using it outside its original specifications if you failed to have two calibration series, so the bracketing is still necessary.