Since a couple of weeks I have to deal with the following problem:

very unexpected I dedected a peak or much better described as a "hill", cause it lasts for more than 10 minutes with slight slopes. This hill overlaps the real data. Other peaks are detected normally and retention times are unchanged!

Working with small peptides I use a C18 / Eurosphere 100 column. Solutions that I use for my gradient (5-95 for 40 min) are A: 0,1 % Triflouroacetic acid and B: 80 % Acetonitrile + 0,1 % TFA. In our laboratory this is the method of choice and we never had any problems before using more than five different chromatographs.

We already checked columns, the pumps, solutions, detector, optic insruments, chamber for mixing, lamp, injector - anything possible. We had two technicians from Shimadzu working on this problem without any satisfying results. Latest idea was, that filters might be infected with biological material, but this turned out to be wrong.

We get this hill with different columns ( only size changes), with different solutions, with different optical chambers, with exchanged chambers for mixing etc. We just canīt pin the source of trouble down.

Has anyone ever experienced the same or a similar problem and what was the clue?

Thank you in advance!

If you have a strongly retained peak slowly eluting through the column, it must come out at some point, typically as a broad hump or hill as you describe.

Try flushing the column with high organic to get this stuff off of the column.

I also belive that this is a late eluting peak. If it is important to identify or observe if there is growth in the peak than the method ought to be held at the higher organic concentration for a longer time until you see the peak comes out.

I don't think this is a late eluting peak because it is a gradient method and the increasing solvent strength usually prevents this type of behavior.

Perhaps a change in one of your reagents has cause a slight uv mismatch that wasn't present before. What wavelength are you using, have you tried new lots of all reagents, have you tried leaving out the TFA? Do the same solutions and column produce the same effect on a different HPLC, or is the problem hardware related? Need more info before we can be of much help. Good luck!

Dear Tom,

the wavelength used for measurements is 220nm. We left out TFA and surprisingly the results became better. But we still belive that it is not a problem caused by the solvents, because the same solvents worked before and still do on different HPLCs. We also changed every reagent, water and even pipettes including all glass bottles etc. The same solutions and columns (we also tried a polyencap) show best results on different HPLCs. Personally I do believe it is a hardware problem, but I donīt know what it could be. Right now a measuring is running without the Autoinjector. Here are some specifications of the HPLC (all Shimadzu): computer module - CBM-10A; detector - SPD-6A; pumps - LC-6A; system controller - SCL-6A; autoinjector - SIL-6A.

Do you see the same "hill" if you make several injections of the B solvent? If so, does it get better/worse/no change with subsequent injections? You might try this after rinsing the column with TFA in acetonitrile. If you don't see any peaks with the solvent injections, it would tend to indicate a late eluting peak from your samples. After rinsing the column, do you see the "hill" if you make a gradient run without making an injection? Has this method worked on this instrument in the past but you have recently encountered the problem? More questions than answers, sorry.

Dear Russ,

we already rinnsed the the column with different types of solvents - everytime with no positive effect. And we are absolutely sure, that this hill/hump canīt be caused by an eluting peak. To answer the other questions: yes, I do see a hump while running a gradient without an injection. Yes, this method has workd before and still does on any other HPLC in out laboratory. The last thing I did was to exclude the autoincector by connecting the column directly with the chamebr where the fluids are mixed. I also put the cilumn behind the detector just to leave it out and still have the same pressure conditions. Instead the column I used a capillary before the detector. The hump was still to be seen. The latest problem is now, that there is a detectable drift of the baseline.. I donīt know where that is coming from...

I`m sorry Axel,but the detector cell has been exposed to high pressure.The drift is caused by a leaking cell.
Try an acid wash to clean the system.Refer to the manufacturer`s manual.

Bill, I am interested in how you sermised that it was a leaky cell? How would a leak in the flow cell cause the baseline to drift?

I had seen a similar phenomina on machines. One machine would run the gradiant and produce a relativily flat baseline while another would give a large hump. We attributed the difference initially to the proportioning system not being exactly the same but now we belive that it is the wavelength of the machines are not exactly the same.

Axel,
If the column is placed after the detector,the pressure in the cell will likely exceed the pressure spec for the cell.The load for the pump is always before the cell.I was refering to your last post.A leaking cell means liquid on the outside of the cell,hence a drift or noise.
A problem gradient in the low UV is hard to trouble shoot because of the time involved to find the problem.Assume there was residual cleaner left on the glassware you used.
That contaminant is in the system and on the column.You would have to flush with large volumn of mobile phase to see if the `hump `is gone.
An acid wash will give you a new starting point for your troubleshooting.Column removed of course.
Checking the wavelenght is certainly important at 220nm. Does the `hump dissapear at a higher wavelength?

Hi, Bill!

After more than six weeks of "trouble-shooting" the hump ist still there. Even with the wavelength set to higher values the hump stays, it only flattens more.
Washing was done more than once and also with different solvents. If there should be really any contamination shouldnīt there any changing in the hump shape to be seen? We canīt...
I know this is really a tricky one and I would like to thank all of you for taking interest. Thanks.

It may be more trouble than it is worth (and for some systems may not be possible) but, have you tried moving components from one system to another? For example, move the detector from the "problem" system to a "good" one and vice versa. You might also be able to repeat this with the autosamplers. This might help pin down the source of the problem.

Dear Russ,

indeed we did exchange several components, just to see that they function very well on different systems... And we know it canīt be the autosampler, cause we left it out and still could see the hump. Right now we are testing the pumps. Another idea was (from Shimadzu), that a small gas bubble developes in the cell. I checked this out and could not confirm this theory.

If you are able to pin point the problem please post it. For I am very interested in the out come since we have seen similar problems here. It has never affected our chromatography quantitation of the analyte. Thanks

Have you tried swaping flow cells. This may validate the bubble or leak theory.

The flow cells have been cleaned and we also tried different new ones. The leakage we got was due to a pressure problem - the pressure we used for one of the tests was too high and so the cells teflon lagging gave up. My fault, but this was not the reason why we got this hump..!

Here comes the funny part! After checking everything I just took the solutions from our technician. We both use the same water, chemicals, pipettes, flasks, etc. . The procedure how we mix everthing was compared without any obvious differences. But with her solutions the HPLC is doing fine - with mine not?! Are we talking about bad karma or what?

I don't know if you filter your solvents but, we have had problems with people handling the solvent filters with their hands instead of with tweezers. This seems to be worse when the person smokes. Apparently it is possible to transfer some oils from hands to the filter which is then dissolved by the solvent. This may not be your problem, though.

This is quite intriguing. You must find out what the difference is.

Recent results: after we had to change some of the teflon laggings of both pumps, the hump did not disappear, but changed his shape. Quite surprisingly I also had a few runs where I injected some samples with the result that there was no hump anymore. We are not sure if the problem is caused by the pumps. In the nest days we will exchange them and see what happens if we use the same type of pump with this HPLC.

By the way: we donīt filter our solvents, cause it never seemed to be necessary.