This is not directly an HPLC question, but I would appreciate any input. I run analysis of a single, small compound each day with a 4 point standard prepared more or less identically each day. Usually I experience about a 4% variation in the slope of the curve but recently it has been more, up to 15% variation. However, the regression is still very linear. The curve is made from area counts by HPLC and absorbance measured externally. I can tell from the analysis results, versus the expected results, that the standard is not representative. Any thoughts on why the slope is not consistant?

First thing I would do is make sure I was not exceeding the dtector linearity limits (maybe dilute both standards and samples and check).

Three things come to mind:
1) Your standard is degrading
2) Your lamp is getting old
3) The solvents are not of high quality or are coming from different lots/manufactures.

Or your pump is not working as well as it should

Im fairly certain that Im within the limits of the detector. Michelle, when you say the standard is degrading, is it possible that its degrading from the time I read the absorbance externally to the time that the sample is actually injected? The strange thing is that if I compare 2 regressions, the points at lower concentrations will have similar area counts/absorbance, but at the higher concentrations (usually) the area counts/ aborbance differ.

Try a simple method (eg caffeine or methylparaben) and check linearity for a few days, so you can check wether its your method or your system.

Why (and how) are you using 2 sources(area counts and external absorbance) for your stds. curve? Has the slope of both sources changed?

A Reid,
Your slope variation? The discussion assumes the slope always falls, that is, the amount of signal for a given concentration gets lower each day. Is that so?
Also is an implicit assumption that the decrease has occurred recently and the slope has been consistent up until now. Is that so?
Jim