[Analytical2015]

Hello people,
I am trying to validate an analytical method on binary UPLC system for related substances and after months of work literally creating the method from scratch (sourcing suitable columns with the correct diamater and internal chemistry, adjusting the baseline to accomadate a sizeable number of impurities, ensuring they do not co-elute or overlap with each other at different concentrations, using a suitable buffered mobile phase with an appropriate pH etc), I have come across a considerable number of issues which have now almost all been resolved to create a suitable method, but I am having issues with the baseline and although I strongly suspect this issue is related to the mixing of my mobile phase/contamination within the needle wash or injector mechanism (i have eliminated practically everything else), I would appreciate any advice from the knowledageble people here.
Unfortunately I dont have time to go back and start over the method from scratch or to engage in lengthy elimination exercises like changing machines, column types or sourcing different mobile phase raw materials as my method needs to be wrapped up, sealed and delivered to the higher-ups very soon or it could be curtain time for me at end of year review.. Deep breath....

The Facts. (includes images of typical and problem chromatograms)

Intermittent baseline noise (Solvent front to approx. 5 minutes)
Mobile Phase A: Acetonitrile
Mobile Phase B: (Acetonitrile: 20mM Sodium dihydrogen orthophosphate di-hydrate buffer (pH 2.9)) 5:95
100 mm phenyl-hexyl column at 50C

Images below from a sample set. Injection 1, 2, 3 (Start of run where problem arises)
As you can see, injection 2 is a mess.
Starting conditions are 18:82 (A: B) changing to 22:78 over 7.5 minutes – curve of 2.
There are several other compounds that elute over 40 minutes which do not pose any problems. After 40 minutes there is a 5 minute equilibration step at 18:82 (A: B).
We increased our gradient mixer from 50mcl to 100mcl which dramatically reduced noise in this region; however this issue has arisen and only occurs intermittingly.
I think the issue arises when my sample is injected onto the column (Sample diluent is 30:30:40 MeOH: ACN: Water). It’s a bit late in the day for me to make any drastic changes, but if anyone could suggest anything to aid mixing/equilibration?
Should I try a 10 mM buffer?

I could alter the constituents of the mobile phases and alter the gradient accordingly if anyone thinks that may help. It’s a real problem that it only happens intermittently. Could it be a faulty mixer? What would cause intermittent poor baseline which sometimes manifests at the end of a gradient only to spill over and adulterate the following injection  (we have 3 impurities which elute from 1 - 3 mins), then followed by a typical injection? Room is temperature controlled so no fluctuations possible. And advice is very helpful. Thanks. Here are the images. First and third image typical, second problematic.



 

[Furanone]

First off, I am guessing you are using a UV detector based on the AU as absorbance units. If so, UV are sensitive to air bubbles, and this could explain why they only happen sporadically. Trace impurites should not be such large broad interference peaks.

Check your injector settings -- try setting a slower injection transfer speed (eg. 0.3 ml/min) and also you should be able to set the injection volume higher (for my system it is called loop filling coefficient). This will consume more of your sample in the vials but it should help get rid of any air bubbles in the fixed volume loop.

Not sure about your system, but with mine I can set a lower injection loop volume and actually see the the air bubbles in the PTFE tubing that makes my fixed volume injection loop, and I can test the dramatic effect that air bubbles have on some detectors.

[Arkcon]

If you believe this is caused in some way by the needle wash or some loop contamination, by all means switch those to something compatible with your separation.  You generally need an aqueous needle wash with some solvent to prevent microbe growth.  People tend to use very strong needle washes -- pointlessly, needle wash has to wash your sample off of steel, not elute sample off an absorbent.  You can run some needle wash cycles and syringe purge cycles until you're satisfied.

Another problem is seal wash.  If its been neglected, the seals may dry and shrink, and the seal wash may mix occasionally with eluent.  Make sure that wash is weaker than your eluent, or drop it entirely, if your solvent salt concentration isn't too high.

By all means, drop your buffer strength by half.  If that solves your baseline problem, you'll have narrowed down the source of the contamination.  Although you may have already determined that you need this buffer level for proper peak shape, at least you'll know.

[Babcock_Hall]

Others here are far more expert in liquid chromatography than I am.  But I have heard it said that mixing MPA and MPB can be difficult when they are unlike solvents.  Could you include 5% buffer in the acetonitrile in MPA (making it the mirror of MPB)?  I am not certain that this would help, but perhaps someone else will comment on this idea.

[How Dont]

I would agree with Babcock_Hall that the constituents of your mobile phases is worth looking at. That you had the insight to try a 95:5 mix in MP - B suggests to me that you may have already tried 5:95 in MP - A, but the buffer crashed out? That can happen, but if you haven't tried already, give it a go, and assess.

Also, if you can reduce the buffer strength like you say, the constituents may combine more readily. Does it even matter if a solution is cloudy as long as it works? I'm not sure to be honest.

Could you look at a gradient mixer of even greater volume?

If it is the mixer, its likely that the majority of time - under the current set up - the mixing is sufficient, giving you the good chromatography of injection 1 and 3, however occasionally insufficient mixing occurs - possibly due to slight pressure fluctuations - and when your sample is injected the poor chromatography of injection 2 is observed. Although the fact you have a 5 minute equilibration, and the issue still occurs is a strange one.

What is the situation leading up to the mixer? pumps, tubing, is there anything you can improve here?

Issues that appear randomly - but regularly (if that makes sense) can be troublesome as they are not always repeatable so can be difficult to investigate especially with a looming deadline approaching. But you seem to be almost there with your method so try not panic.

Discuss with colleagues who are more familiar with your setup, it maybe something they have seen before.

[How Dont]

Any updates on the method OP? would be interested to hear if you resolved the issue.