Drop in Baseline, Software or Detector?

[CCRUtech]

Hello all,

I recently started a summer project measuring tocopherols extracted from barley samples. All was well until recently we changed procedure to deal with some co-eluting peaks.

Here are the basic specs to give you an idea about our set-up

RF10A detector
LC6A cromatograph
Gilson 234 autoinjector
Running class VP software

Originally using Pure Hexane buffer, but 2 months ago switched to a 2%dioxane 2% ethyl acetate solution which helped to seperate some important peaks.

At first a good chromatogram would have a solute peak at 700-1000 mV with our variables coming out at about half that amount.

Currently all of our samples are coming out odd. The best chromatograms still contain all the peaks at thier proper elution times but the solute peak is relativley small at ~25mV with the variables coming out at ~100 mV. The baseline also shifts to -10Mv after the solute peak.

We also randomly get chromatograms with a larger solute peak (~250mV) a straight baseline exactly at 0, and greatly reduced variable peaks. These look almost as if you took an old 'normal' chromatogram and raised the baseline then changed that value to zero. This data a useless since it eliminates important peaks and fails to give us even relative values to work with. (I can post pictures if anyone thinks it will help formulate an opinion)

We have retested several samples with varied results on identicle samples, meaning that the same sample would kick out both kinds of baselines desribed above.

My best guesses are a bad detector (lamp?) or somthing odd in the software method. If you have had similar problems please let me know!

Thanks

[tom jupille]

I don't have any hands-on experience with that detector or data system, but some general suggestions:

1. If the detector has an analog output and you can find a strip chart recorder, try hooking that up so that you can look at the output directly (before processing). That should quickly answer the hardware/software question.

2. If it's a hardware issue, it might be as simple as a marginal / corroded connection. Try unplugging/replugging all cables. If the problem persists (and you have more courage than good sense ), open it up and unplug/plug circuit boards.

[elsico]

I think first way is to check flow cell and clean if necessary. Second watch baseline with stopped pump and compare with page 9-3 of manual.
Really good idea to check this with recorder outputs but first unplug software cables.

[CCRUtech]

The general feeling among my superiors was that the lamp was bad (over 2000 hours on it, dont those explode!?) or the flo-cell had a bubble. I was of the opinion that either a lamp works, or it doesn't; no inbetween. We found no obvious bubble in the flo cell but cleaned it out, and replaced the lamp.

Turns out, the lamp we ordered didn't seem to fit as well as the old one. When the detector would boot up we would hear a rapid clicking like when the contacts of a switch are bad, swtching back to the old lamp was the only remedy.

I am convinced the original problem has somthing to do with the detector output, its little go-between box, or its connection to the computer. I plan to try the solutions tom suggested, as both issues elisco mentioned were addressed yesterday.

Any thoughts on the lamp? This system is old... older than me at 24. We will contact our supplier today and make sure the new bulbs are compatible. I didn't change the thing myself, so who knows...

one thing goes wrong, everything goes wrong

[ivanvins]

I will suspect a chemistry problem could be here - some fluorescent impurity accumulates on your column until breakthrough or some late eluting peak is present - the detector will loose sensitivity while compensating high background signal. The fluorodetectors usually zero before each run.

[CCRUtech]

But what tips you off (besides the chromatograph pattern)?. I have never worked with this kind of buffer before, is it harsh on columns?

If the problem continues after the current repairs, that would be a good explanation, but would set us back, we'd have to flush after every sample then i think.

Edit:: A chem problem could also help explain the 10-15 bad runs followed by 5 worse runs on our last few runs. Only, at one point we got consistently good results so there could also be a communication problem.


7/16 UPDATE::

After a few more short experiments and a few calls to our shimadzu supply guy we are only sure of one thing

It's the detector... Possibly the power supply which never occured to me. They are going to check it out this month. Meanwhile, we may swap it out for an unused RF-10 in another lab.

In responce to the chem problem, that also could have been it. Pyrogallol used in the extraction could build up in the column, but we are all very careful to re-spin any samples with the slightest hint of goopy brown syrup below the organic layer.

We still, however; get the same results (odd low baseline) even between cleanings and running blanks, even after checking the flo-cell and re-running.

Thanks for all the great suggestions folks!
Hopfully the last update is on its way