Hi everybody !
My problem is a really bad peak shape: the peaks look as if the top of them was turned downwards. Now I learned that there could be two causes:
1. a void on the head of the column
2. contaminations
For the first case I tried to solve it by close fixing the column: nothing doing, still the same peak shape.
Then I tried to wash the column, even against flow direction. Still the same...
Now my question is:
How can I check whether my guard column (or worse my prep column) is really contaminated and needs to be replaced, or it is not and I'm looking up at the false place ?
I'd really appreciate any help because I don't want to order a new guard column if I'm not sure that I really need a new one...(question of budget...)
Thank you very much in advance.

You might also have a problem with pH. Make
sure the pH of the solution that you are
injecting is the same as the pH of your mobile
phase at the beginning of the run. If you want
to check column contamination, run a blank.

I checked the pH thing, the pH of my solution and that of my mobile phase are about the same...
Perhaps important to know:
The bad shapes only appear above a certain peak height, below this height the peak shapes are normal...

Is there a possibility of solvent mismatch? Are samples in same solvent as mobile phase? If not (if organic is higher in samples) this couldbe a possible cause of problems. Also, if it only happens with increased concentration, it sounds like the column could be overloaded. What is your abs when peak shape goes bad? As far as checking the guard, just remove it for an injection and see if the peak shape is still bad.

Sounds like an overload. You may be exceeding your column's sample capacity.

To me this sounds like a standard case of detector overload - hmm actually, it could also be the solvent mismatch suggested by Scott....
Anyway, Scott asked the right question: what are the units of your detector signal? If it is UV, and you are at 1 or 2 AUFS, it is the detector.

It occurs with a signal of more than 1 Volt (about 1.8 AU) height, but I had signals higher than that before (up to 2 Volts) and there the peak shape was totally normal:
The detector (Prostar 320, Varian) has integrated a dual-pathlength flowcell, so when absorbance continues to increase, the detector automatically should switch to operation on the shorter pathlength and to a extended limit (up to 10 AU).
Could it then be that the detector isn't able anymore to switch from normal to extended limit?
(a communication problem ?)
To Scott: the samples are quite the same, with a little bit more organic; but again: that was never a problem before either...

As for the detector I am not familiar with it but it again rings of what you said earlier, if it wasn't a problem before then why now? But in my experience, 1.8 AU is far beyond where I would feel comfortable with abs, and also in my experience linearity is hard to come by in the higher abs.

Getting back to overload, this could be a situation where the problem could still be column related. A new column has a certain capacity which certainly will decrease over time. In other words, maybe you are overloading your column at its present capacity? Not sure though. When I have had problems like you describe, column replacement is my first option.

Good luck.

If you have eliminated pH and solvent
compatibility, it does sound like a problem of
column overload. A much more affordable
option to replacing the column is to use a
smaller sample loop (or you can easily make
one). If this is not an option, dilute your
samples. It sounds like you might be
extending your standard curve past the
instruments limits.

According to the description that you get the phenomenon only at the highest load (and not below this) and the fact that the detector operates at 1.8 AUFS, I am now completely convinced that you have detector overload. You ask, why you did not see this before at the same signal height? The problem is that the detector automatically subtracts the background. If the background of your mobile phase is low, you may be able to go to 3 AUFS without a problem. If your background is high, your detector may go blind already at 0.5 AUFS - and you get funny signals like the one that you are describing. What is all the stuff that you have in your mobile phase, and what is the wavelength?
Unfortunately, I am not familiar with the trick used in your detector, where the detector decides on his own to switch to another cell. Seems like a cute trick, but I can't see how this would work in the middle of a chromatogram...

To all the previous responders:
I think Sybille is talking about PREPARATIVE HPLC, where sample overload is common practice. I would not worry too much, provided the separation is the same as before.

Oh, I'm sorry, I really did talk about prep. HPLC, didn't I mention before ?
Separation is the same as before, but it's sometimes very difficult whether it's just this peak shape phenomenon or two unseparated peaks...
And I have to admit that it makes me become uncertain because it never happened before...
But I will further check the idea of overloading..

Thank you all very much for your good advices and comments.