Acquity PDA detector - downward baseline drift problem

<p>Hello,

We recently got an Open Architecture UPLC in our Lab.
So far, we are very pleased with the machine, but I have a question concerning the PDA trace.

What we are finding is that the PDA trace is almost always drifting downwards over the course of the run.

This is not normally a problem, however sometimes the drift downwards is too severe, and it seems, that the baseline drops below the lower detector limit i.e. we actually loose some raw data.
This can prove problematic if, for example there are some small peaks present later in the run, as the data for these may be permanently lost.

We usually extract the 225nm wavelength for our reports, but we have seen this problem for the PDA trace as well.

It is disappointing that we can actually loose raw data in this way – after all the mobile phases we are using are not unusual.
It seems that the lower (i.e. below 0AU) limit of this detector is verity limited!.

We have checked that the column is fully equilibrated before making an injection.
The detector cell was changed but this made no difference.

So my question is : can anything be done to lower the lower limit of detection for the PDA detector, so that we don’t loose raw data.
Or, are there any other suggestions?

We are using the highest grades (18.2 Mega Ohm Water, and LCMS or gradient grade solvents & buffers).


Conditions:

PDA 210 – 400nm. Extract to 225nm for reports.

Gradient is 2 to 98% B over about 1.3 mins.

Mobile phase A1 = 0.1% Formic acid in water – for acid pH runs.

Mobile phase A2 = 10mM Ammonium Formate in water – for neutral pH runs

Mobile phase B = Acetonitrile – no buffer.





</p>

Comments

  • A couple of things come to mind when I asked an expert, here's his comments

    (1) Because the gradient goes to a more optically clear mobile phase, there is a downward shift. Adding the formic and buffer to the organic may very well help in this regard but will obviously have to use B1 and B2 for ACN with formic acid or ammonium formate.

    (2) The other piece is with regards to viewing chromatograms in MassLynx. By default the baseline is at zero so that you don't see data below that. On the chromatogram window under the display pulldown, go to View... and make sure the radio button for baseline at lowest point is checked. That will hopefully solve the problem.

    Let us know if we are on the right track.

    Thank you,

    Liz

  • Hello again,

    I have put my comments below yours:

    "Because the gradient goes to a more optically clear mobile phase, there is a downward shift. Adding the formic and buffer to the organic may very well help in this regard but will obviously have to use B1 and B2 for ACN with formic acid or ammonium formate".

    I agree that the ACN will be more optically clear, but I am tring to avoid having the buffer in the B line, as it is fairly insoluble in ACN, and I want to avoid it crashing out.

    I would have thought that the PDA should have been able to handle these combinations of mobile phases as they are not unusual.

    "The other piece is with regards to viewing chromatograms in MassLynx. By default the baseline is at zero so that you don't see data below that. On the chromatogram window under the display pulldown, go to View... and make sure the radio button for baseline at lowest point is checked. That will hopefully solve the problem".

    I have the setting on "baseline at lowest point".

    The problem is that at the actual baseline goes below the lowest level of what the detector can pick up, so there is no signal to see anymore!.

    Regards.....

  • Perhaps you could post a picture because this statement is a little confusing "The problem is that at the actual baseline goes below the lowest level of what the detector can pick up " . Are you saing the baseline is drifting -4 AU? If that is the case, something is broken.

  • Hello,

    I will need a little time to get an example chromatogram, as this issue has been very intermittent, and I have not seen it for a while. I think its very sample dependant (maybe down to combinations of wavelenghts....).

    The machine is open access, so what I will do is ask the group to keep an eye out for this issue, and once we have an example I will post it.

    Interestingly, the community Acquity Method development tutorial no 3 recommends an option using 20mM Ammonium Formate as A, and 100% ACN as B - similar to us.

    So it looks like this combination of solvents should be fine with an Acquity PDA.

    Regards,

    A.

  • Hello,

    - here's an example - ESI, PDA, ELSD all give expected traces.

    But, extracted 225nm loses the baseline towards the end. No info is available for the last 0.5mins. Possibility of small peaks being missed.

    I have also seen this happen with the PDA trace itself.

    The Chromatogram display baseline view is set to "lowest point".

    Cheers,

    A.

  • So it seems that, yes, the baseline drifts down 4AU, which is the the limit. So now the question is why. To answer this question I afraid I have to ask a few more questions.

    1. Does this always happen at the same time?

    2. If you run blanks, does this happen.

    3. If you were to inject 1/4 of that sample you posted, does this still happen?

    4. Can you post the method parameters (all of them) - In your masslynx project folder, in the acqudb folder, find your detector method (something like methodname.wpda). Open it with wordpad, save it as a text file. Edit any information out of that text file which you might consider sensitive. Post it here so I can have a look at it.

    This would be a good start.

    AA

  • Hello AJA,

    this will take some time to dig this up.

    Will get back as soon as can.

    Cheers,

    A.

  • AJA has ome excellent suggestions, it would be also interesting to run without autozero enabled, but lets see what your results indicate.

    LIz

  • Hello Liz,

    - thanks for the update.

    Very busy at moment - will try and find some time to look at this when I get the chance. Sorry.

    Regards,

    A.

  • Hello AA,

    I have tried to answer your questions below - see italics:

    1. Does this always happen at the same time?

    No. I have seen traces where the whole extracted 225 nm trace lies below the view.

    Other instances where there is a gradual lowering of the baseline until it disappears.

    2. If you run blanks, does this happen.

    No.

    3. If you were to inject 1/4 of that sample you posted, does this still happen?

    No.

    I have asked around the lab, and although people have also seen what I have described, it would appear that its quite a rare occurance.

    So, the problem here is that it seems to occur only every now and again.

    In other words I cant, right now, reproduce this effect. Everything is now behaving normally.

    4. Can you post the method parameters (all of them) - In your masslynx project folder, in the acqudb folder, find your detector method (something like methodname.wpda). Open it with wordpad, save it as a text file. Edit any information out of that text file which you might consider sensitive. Post it here so I can have a look at it.

    I have attached the .wpda from the "dye261009" chromatogram.

    I'm guessing you wanted more info than that, but do I attach the rest as is, or do they need converting into text files?

    Regards,

    A.

  • Some interesting clues but no "smoking gun".

    I would say that this is sample related in some way. Is it possible that with some of these concentrated samples that there is still material eluting off the column after the run time is over? If there is, the detector would autozero with absorbers in the cell, and as the next sample runs, the absorbance would drop into the negative. Just a thought.

    Also, you have the UV blocking filter on, set it to off, I don't think this is the issue, but it is a simple thing to do. For your application it doesnt need to be on.

    AA

  • Thank you AA.

    Its possible its sample related - we are open access and "anything goes" so to speak.

    I have tried to find more info on the "UV blocking filter", but cant seem to see any info in the ML help files.

    I assume it simply ensures there is no signal collected below 210nm?

    Either way I will turn it off....

    Cheers,

    A.

  • When the UV blocking filter is enabled, wavelengths below ~210 nm are prevented from entering the flowcell. You would enable this feature if you are running compounds (some classes of proteins in particular) that may undergo photo-oxidation in the flow cell in combination with MS (you can see the oxidation states with the MS and it can make spectral deconvolution difficult in those cases).

    AA