Low response for hormones in pos mode but not neg on API3000

[sassman]

I have an API3000 that we are using for hormone analysis in surface water samples. The machine is working great for the estrogens which are analyzed in MRM negative mode. My problem is with analysis of the androgens which are analyzed in MRM positive mode. I was seeing very low response for the androgens and had already tried doing a mass and resolution calibration, so I decided to try giving the front end a cleaning. After cleaning the orifice plate, skimmer, and Q0, the response was much (100X) better. I figured everything was OK again and started a batch. The first set of standards came out good, but after that the response quickly dropped back to what it was before the cleaning. My samples have been cleaned up using an SPE method designed to eliminate humics and other potential interferences, so they are fairly clean but probably still contain some junk. I have heard of Q0 “chargingâ€

[sam.pedraglio]

I cannot guess about the reason for this behaviour, the only explanation I have is the higher ionized background in pos than in neg mode, this could justify why you have a response reduction in pos and not in neg.

[sassman]

That would make sense except that up until a few weeks ago we got much better response in positive mode than we are seeing now, so I assume this is related to the instrument. The multiplier is nearing the end of its lifetime and is now at max voltage. I wonder if this could have something to do with it.

[sam.pedraglio]

I don't think so.
The multiplier is the same for neg mode and, usually, in the latter case its higher than for pos.

[Camisotro]

Just curious, what kind of mobile phase conditions are you using with the androgens? I know that testosterone, for example, can be quite finicky in positive ESI response with respect to additive concentrations.

[sassman]

My mobile phase is A: methanol/water (10/90) and B: acetonitrile. The gradient is from 40% B to 95% B over about 10 minutes. Same mobile phase (but slightly different gradient) we use for estrogens. We played around with adding formic acid, but we had problems getting rid of the formic when it was time to run estrogens in negative ion mode. This method was working fine for a long time, but only recently gave us this problem.

I spoke with our service engineer who said that sometimes a drop in signal is caused by charging on the quads due to contamination. However, this does not appear to be the case here since the signal drop happens after a number of runs. All of my standards seem fine in the beginning of the batch, but by the time I get to the fourth real sample my response for the IS has dropped 10X. I know that sounds like a problem with the samples, but the method has been working fine for over a year up until a few weeks ago.

Our service engineer seems to think its an autosampler problem. That doesn't seem right to me since no other methods are having the same problem. I guess it could be related to pressure changes, but the pressure isn't all that different from other methods. I have replaced the metering pump seal, sampling loop, and low pressure valve rotor (Shimadzu HT-A autosampler). Next up will be the needle and needle seal. Also replaced tubing from the injection port to the 6port valve and from the 6port valve to the column. At this point, I don't know what else I can do except to change my method. Any ideas?

[lmh]

just a thought: in negative mode I presume you're looking at M-H as the parent ion. In positive mode the ionisation obviously has more options: M+H, M+Na, M+K, M+NH4 etc., and what you get depends strongly on the chemistry of the analyte, and also on what adduct ions the analyte can possibly scavenge from the mobile phase.

Even very, very slight changes in mobile phase handling can completely change the predominant adduct, if you are unlucky. For instance, if you have an analyte that really likes to form sodium adducts, and you make really fresh solvent quickly, you might still see hydrogen adducts, but if the aqueous buffer is left overnight in a glass bottle, you'll get sodium instead.

What parent ions are happening? Is there a chance that your method is now looking for a low-abundance parent ion because a minor change in mobile phase/water handling has led to a different adduct being formed? Remember, this will be compound specific; other methods in positive mode may not be affected. If you've got a general problem in positive mode, others will be affected, and it ought to show up in instrument calibration.

[Bryan Evans]

We have some LC-MS (API4000) data for hormones:

beta-estradiol on Cadenza CD-C18:
http://www.imtaktusa.com/site_media/fil ... TI054E.pdf

steroid hormones & metabolites on Unison UK-C8:
http://www.imtaktusa.com/site_media/fil ... TI397E.pdf

[Camisotro]

My experience was that testosterone has a large MNa+ signal and a weak MH+ signal that could be helped by very low concentrations of ammonium acetate (0.5-5 mM) and sometimes low concentrations of formic acid too (less than 1%). With the right additive, the sodium adduct signal drops and the MH+ signal shoots way up.

So you might want to look back at additives and see if your signal is more reliable that way.

[sassman]

Thanks for the advice ctroster! It's strange that we have been running without formic acid for a long time with no issues. I was suspecting some kind of matrix problem and decided to give the formic acid a try. I added 0.1% to the aqueous portion of my mobile phase and the signal is now rock solid after 100 runs! I am guessing that either some change in our water supply (nanopure filtered) or our recent switch from LCMS grade to HPLC grade ACN is to blame. Good thing its working now because I am way behind. I have over 2000 samples waiting to run!

[Camisotro]

Thanks for the advice ctroster! It's strange that we have been running without formic acid for a long time with no issues. I was suspecting some kind of matrix problem and decided to give the formic acid a try. I added 0.1% to the aqueous portion of my mobile phase and the signal is now rock solid after 100 runs! I am guessing that either some change in our water supply (nanopure filtered) or our recent switch from LCMS grade to HPLC grade ACN is to blame. Good thing its working now because I am way behind. I have over 2000 samples waiting to run!

Glad to help! It could well be that a contaminant in your previous mobile phase (or stuck in the plumbing, or stuck in the ESI source) was in fact giving a matrix enhancement effect, which went away after a while. Now you can be assured that most of your signal strength is coming from a controlled factor rather than a random glitch. Hopefully the low concentration of formic will mean you don't have much trouble getting rid of it in negative mode.

(On a curious tangent, what is the difference between LCMS grade ACN and HPLC grade?)

[Rappoldbr]

Check the paper by Annesley - Methanol-Associated Matrix Effects in
Electrospray Ionization Tandem Mass Spectrometry, Clin Chem 2007. A shift in the manufacturing of ACN can be just as liable to the effects he demonstrated.

[Rappoldbr]

Check the paper by Annesley - Methanol-Associated Matrix Effects in
Electrospray Ionization Tandem Mass Spectrometry, Clin Chem 2007. A shift in the manufacturing of ACN can be just as liable to the effects he demonstrated.

[Rappoldbr]

Check the paper by Annesley - Methanol-Associated Matrix Effects in
Electrospray Ionization Tandem Mass Spectrometry, Clin Chem 2007. A shift in the manufacturing of ACN can be just as liable to the effects he demonstrated.

[Rappoldbr]

Check the paper by Annesley - Methanol-Associated Matrix Effects in
Electrospray Ionization Tandem Mass Spectrometry, Clin Chem 2007. A shift in the manufacturing of ACN can be just as liable to the effects he demonstrated.