Recently, all our gradient HPLC runs are having these down-slope "smile face" baselines. In the past, the gradient baseline were all up slope "normal looking". We suspected our milliQ water was getting bad so we replaced filter cartrige, but "smile faces" remains. We are using HP 1100, and mobile phases are combinations of ACN, water or phosphate buffer (pH6.6) with detection at 220 or 280. This phenomenon was observed in several gradient methods. Any suggestions?

Is the problem worse at 280 compared to 220? If so, I'd suspect RI mismatch as the root cause of the problem. Since the problem appeared abruptly (if I interpret your post correctly), I'd look at cell alignment as the proximate cause.

-- Tom Jupille / LC Resources Inc.

yes, it is worse at 280 than 220. However, we used 3 HP1100 and 1 Waters Alliance systems, and observed the same type baseline. I suspect either water or ACN was the cause.

Hmmmm . . .
Well, the wavelength dependence argues in favor of RI mismatch (longer wavelength light refracts more than shorter), but the fact that the problem arose abruptly on multiple instruments argues against it. You've already changed the water system cartridge, so that's (probably) out (you might borrow some water from another lab that uses a physically different system just to be sure). I'd say the next best bet is the buffer, but basically you're now faced with two options:

a. live with it (if you data system can handle the drift).

b. try changing one thing at a time (buffer salt, ACN, column, pumping system, etc.) until you find which change makes the problem go away.

Not fun, but if it were, they wouldn't pay us to do it :-)

Good luck!

-- Tom Jupille / LC Resources