Could any one explain why I experience a sharp drift in base line with my gradient runs using acetonitrile at UV 240nm. (35-85%ACN for 10mins)
What is the cutoff wavelength of ACN??
How do I determine my limit of detection with a "slanted baseline" on a computerised detector??
Your answer is welcome. Thanks

The UV cutoff for acetonitrile is 190nm, that is it has a UV absorbance of 1AU in a 10mm cell at 190nm.

The sharp drift in baseline is either due to the absorbance of acetonitrile at 240, which is negligible, or refractive index change of your mobile phase. If the base line is curved and peaks at about 50% ACN then it is due to refractive index. If it is linear and it is a maximum at 85% then it is due to absorbance.

You determine LOD just as you would if the baseline was not sloping. The sloping baseline just makes it more difficult to find adequate integration parameters.

If your drift is due to refractive index your only real option is to try a different detector. Make sure your cell is aligned correctly etc.

However, if the slope is due to absorbance, you can match the absorbance of the A and B components. Assuming that you are using water for the aqueous fraction and that your baseline is sloping up you can try adding sodium azide (or urea, nitrite, nitrate, etc) 20-200ppm to the 35% acn solution and see if you can find a concentration that will make the drift disappear.

Good Luck

Tom, it would be very instructive to find out how you derived those criteria to distinguish between ri and absorbance.

Will your software allow you to store and subtract
out a blank run? You would then be dealing w a
hoizontal (or almost horizontal ) baseline.

Hans,

Mostly just from experience, however I do recall reading as much in one of Snyder's books probably Introduction to HPLC.

I also recall reading an abstract from a Russian physical chemistry journal where they discussed the refractive index of binary mobile phases being parabolic functions of the mole fraction of organic. I'll see if I can dig it up.

Re: B. Buglio's suggestion: While it works in theory, often practice is trickier. Baseline subtraction works best when there are NO extraneous or excipient peaks in the blank. Otherwise, you'll get odd dips and inflections due to slight differences in peak RT's and magnitudes that could have a deleterious effect on integration. I've used the technique successfully, but tend to stay away from it when I'm doing quantitative work. I would look into your mobile phase first. Simply take a UV spectrum of each phase in a quartz cuvette blanked against an empty quartz cuvette and if your issue is a difference in MP absorbance, you'll know. If you don't see anything, go after the column, frits, etc...

Hope this helps a little!

Chris

Agree w Juddc regarding extraneous peaks however
problem, as stated, is with drifting baseline.
Also admit suggestion is actually a symptomatic
"fix": shouldnt be getting a sharp drift at 240nm
w ACN and Chris should identify the root cause of
problem. Having said all that I have had some
successes with baseline subtrac

Thanks for all your suggestions. I need time to digest them and I`ll get back. Anyway, this is my first time of presenting a question on this forum and am really impressed by the response. Thanks a million I`ll be back after digestion.

Chris,

The previous thread got me thinking about detectors. If your detector has the ability to use a reference wavelength this can greatly reduce the amount of drift.

Assuming that your detector wavelength is set to the lambda max of your analyte, just set the reference wavelength to the closest wavelength at which your analyte does not have appreciable absorption.

If you can tollerate a decrease in sensitivity you can even set the reference wavelength < the detector wavelength, pick the wavelength at the valley of your analytes UV sepectrum.

ACN/water gradients really shouldnt give significant baseline drift at 240 nm. The refractive index effect theory is fine, but really, most modern detectors (what model of detector were you using) deal with and compensate for these effects. My guess, based on the first message, is some kind of contamination, maybe from the water (very often the culprit). Contamination within the LC system itself (pump/injector) or even the column can cause this kind of thing as well.

Just a note if you follow the advice in the previous message (from Tom M.). You must be cautious selecting a reference wavelenght, incorrect selection can obscure peaks that you might be interested in. This is not usually a problem but I have seen it happen. As a matter of fact, I think there was a message on this board some time ago regarding that exact problem. Check the archives.

A bit late but...:
Thanks Tom, it would be nice if you did find some more info.
Back to the original question: Is it an ubdrift or downdrift (canīt recall mentioning of that)? Knowing that, would have allowed an initial assigment as to which solvent might be the culprit. Also, doesnīt ACN give rise to complexing with some buffers?
Donīt forget that a dirty column might also give a drift (up) with increasing ACN.
On integration: If you have sufficient "dirt" in a solvent to be above the linear range of your detector you will be in trouble. One can recognize this normally via excess noise. If your slope (drift) is almost as steep as that of the peaks an integration based on slope will have problems.
Too bad hardly anybody tells us about how the problem was solved. Please donīt be a ashamed if........
Hans

O. B. Rudakov and V. F. Selemenev/ Refractive Index of Binary Mobile Phases for High-Performance Liquid Chromatography // Russian Journal of Physical Chemistry. 1999. đ12. pp.2019-2022.

Hi,
Could someone explain to me why a base line drift suddenly appears after an injection? I am using a mobile phase of ACN:phosphate buffer (isocratic gradient) and a wavelengh of 230 nm.
Thanks...
Laura