major contamination trouble

[Roberto]

Hi all,

My lab is experiencing a major contamination problem. Since the beginning of this year, we observed several late eluting peaks while running a H2O-MeOH gradient on C18 columns. We did lots of cleaning the whole HPLC system, as recommended by WATERS (we used 50% nitric acid, then lots of H2O, then MeOH, MeCN and isopropanol, with small injections of DMSO). The peaks disappeared, but one problem still remains: the baseline of a chromatographic run starts climbing to a "platform-like" shape at the end of the gradient (in 100% MeOH). Therefore, we were wondering if this could be a water contaminant that remains retained in the column and is eluted in 100% MeOH.

Do anyone has any idea what this could be and how to clean the HPLC and columns?

I will very much appreciate your help.

[aceto_81]

First you have to find where the root of the problem is: the water or the MeOH.

Suppose you have a gradient going from 0% MeOH to 100% MeOH in 10 minutes.
First run this gradient to cleanup.
Now run an isocratic part of 5 minutes 100% water. This should accumulate contaminants from the water.
Run your gradient.
Run again an isocratic part, but for 20 minutes, 100% water.
Run your gradient again.
If you compare the gradient results, contamination comming from your water should be much bigger in the gradient following the 20min isocratic part.

If not, your water isn't the source of the contamination.

Try to figure out what's the source of your contamination, and tell us, then we (maybe) can find some solutions

Kind regards

Ace

[placetoo]

Greetings Roberto,

I would not be too surprised to discover that this is caused by your water. It is, relatively, easy to confirm or deny the water hypothesis. Simply run a few gradients (maybe three) where you hold at initial conditions for an increasing period of time. If the problem is caused by your water, the distortion, at the end of the gradient, will increase the longer you hold at initial conditions.

You did not mention the length of your gradient, but the increments in initial hold should be made long enough to yield unambiguous results. Let's say your gradient is 15 minutes in duration. I would increase the hold, at initial conditions, by 1 minute on each trial. The longer the gradient, the longer you will have to increase the initial conditions hold to see an effect.

If it is the water, you don't have too many options. The best solution is to obtain a better source of water. Either change out the cartridges on your water system or purchase LC grade water. Otherwise, you may have to adjust to the "new" profile of your gradient or have your data system stop collecting before the platform emerges so it is not so annoying.

Regards,

placetoo

[Rob]

Best to eliminate some simple things first. The fact that the baseline increases and stays up as a plateau could be caused by detection of the MeOH. What wavelength are you using?

[tom jupille]

Gradient baseline issues come up often enough that we have put a brief (15 minutes) "mini-seminar" up on the LC Resources web site:

http://www.lcresources.com/more_resourc ... hp?f=3&t=5

You do have to register to access it (registration is free).

[zokitano]

I tend to agree with Rob. Maybe the plateau that you're observing is due to the UV absorption of MeOH if you're using low UV detection wavelength

[Roberto]

Firstly I would like to thank you all for your very nice suggestions. As aceto_81 suggested, we have ran gradients by increasing the amount of water at the gradient "starting". Effectively, the gradient plateau increased significantly. Therefore, we were very suspicious that the water was the main contamination source. We currently use a double filtering system: a Millipore RIOS-MilliQ. We changed all the necessary cartridges (somewhat expensive), and performed the LC "experiments" again, and we observed the same problem, i.e., by increasing the amount of H2O at the beginning of the gradient time, a plateau of an increasing size starts to appear near to 100% MeOH. So, we eliminated the water contamination hypothesis..
Although Placetoo suggestion appears to be a good one, we have to run the gradient up to 100% in order to elute all the compounds we want to analyze.
Rob, we are not only using an UV detector, but simultaneously an evaporative light scattering detector (ELSD) and a mass spectrometry detector (MS). Our system is a LC-PDA-ELSD-MS. The plateau and the signals appear in all three detectors, so it is not a problem related with the wavelength of the UV detection only.
Dear Tom Jupille, great tutorial! I really enjoyed it.
As for a contamination (seems to have some plasticizers), how to remove them efficiently?
Another question: how can I include a picture in my posts?

[ksharp]

refer to this note for attaching pictures:

viewtopic.php?t=2617

[Roberto]

Firstly I would like to thank you all for your very nice suggestions. As aceto_81 suggested, we have ran gradients by increasing the amount of water at the gradient "starting". Effectively, the gradient plateau increased significantly. Therefore, we were very suspicious that the water was the main contamination source. We currently use a double filtering system: a Millipore RIOS-MilliQ. We changed all the necessary cartridges (somewhat expensive), and performed the LC "experiments" again, and we observed the same problem, i.e., by increasing the amount of H2O at the beginning of the gradient time, a plateau of an increasing size starts to appear near to 100% MeOH. So, we eliminated the water contamination hypothesis..
Although Placetoo suggestion appears to be a good one, we have to run the gradient up to 100% in order to elute all the compounds we want to analyze.
Rob, we are not only using an UV detector, but simultaneously an evaporative light scattering detector (ELSD) and a mass spectrometry detector (MS). Our system is a LC-PDA-ELSD-MS. The plateau and the signals appear in all three detectors, so it is not a problem related with the wavelength of the UV detection only.
Dear Tom Jupille, great tutorial! I really enjoyed it.
As for a contamination (seems to have some plasticizers, maybe?), how to remove them efficiently? Please, take a look at the chromatogram and at the MS of the major peal (Rt 11.3 min.) below

[aceto_81]

Hi,

could you also provide your gradient?

Ace

[Roberto]

Hello Ace,

The gradient is:
a) 1 minute at 100% H2O;
b) then linear gradient of MeOH in H2O during 20 minutes;
c) then 5 minutes at 100% MeOH;
d) then reequilibrate during 5 minutes.
The MS was obtained using electrospray ZQ 2000 (Micromass).

Thank you for your help.
Roberto

[aceto_81]

Strange, I had expected a linear gradient from 0 to 100% MeOH at 10 minutes or so...

Are you sure your GPV is working correctly?

Ace

[Roberto]

Dear Aceto,

You are right. I checked with one of my analysts. I was wrong. The gradient starts with 100% H2O then moves directly to 100% MeOH in 10 minutes. Then it runs at 100% MeOH during 10 minutes. Then goes back to 100% H2O and re-equilibrates during 10 minutes.

[aceto_81]

Then the gradient looks quit normal to me, but if you find the plateau too high, you can try with another type of MeOH.
I think the raise of the baseline is a due to MeOH.

gl

Ace