In a class I was teaching recently, I had a question come up that has me puzzled, and I wonder if any of the CF participants can shed light on the subject. The chemist was using an HP 1100 with Diode Array set to monitor a single wavelength at 214 nm. She was running A = 25 mM phosphate, pH 3 and B = ACN with a 5-80%B gradient (I forget the gradient time) on a reversed phase column. She reported _negative_ baseline drift during the gradient. This is contrary to anything I've ever seen -- usually the drift is in a positive direction. She used the same pump, mobile phase, column with a Kratos detector and saw the normally expected baseline rise. I asked attendees if anyone else had seen it and one other person had seen the same behavior.

Anyone know what's happening? I'd sure like to know.

Additional questions to ask:

1. Was the drift linear or non-linear (i.e., did the drift track the gradient profile)
2. Since a PDA was used, was the drift more or less pronounced at longer wavelength.

What I'm looking for here, as you have probably surmised, is RI drift. Depending on the cell geometry and alignment, it is quite possible for an RI shift to let *more* light get through rather than less.

In Diode array detector apart from measuring wavelength (214) is an reference wavelength (usually about 500nm) for compensation lamp drift. Try to work with reference off, then DAD is true single wavelength dedetector.

This is a common problem with the HP 1100 DADs. Your chromatogram is based upon two things: first, the monitoring wavelength; second, the reference wavelength. If the reference wavelength is stable, then your monitored wavelength will be stable. But if your reference wavelength changes (as it might during a gradient), the monitored wavelength changes in the OPPOSITE direction. You also might observe inverted peaks. Fix this problem by choosing a reference wavelength that is unaffected by the gradient AND the analytes. Best of luck!