Chromatography Forum

Hi,
I have seen this phenomenon before, but have never been able to get control of it except with "equilibration" of the system with lots (10-20) equilibration injections. I am trying to measure glutamine using TFA and HFBA in MPA on RP column. My first few injections have good rsd, then there is a period when the peak starts to get bigger, say over the 6-9th injections. After, the signal has increased about 2x the original intensity it stabilises for the remainder of the run. Does anyone have an idea as to what is happening here??
Thanks,
Cliff

We are also experiencing similar problems when analysing a basic hydrocarbon mix (a lubricant). Would be very keen to find the route of the problem!

Hello,

I can't get you any solution but to inquire blindly about the issue, i suppose that you have in mind that the original response in the firsts injections could be the erroneous part, despite the initial rsd. Then, perhaps it is after the 6-9 injections, when you say the concentration (when the colum is actually equilibrated/active sites ?¿!?) is again stable, the more real concentration of the standard.


Ricard

Hi,
I have seen this phenomenon before, but have never been able to get control of it except with "equilibration" of the system with lots (10-20) equilibration injections. I am trying to measure glutamine using TFA and HFBA in MPA on RP column. My first few injections have good rsd, then there is a period when the peak starts to get bigger, say over the 6-9th injections. After, the signal has increased about 2x the original intensity it stabilises for the remainder of the run. Does anyone have an idea as to what is happening here??
Thanks,
Cliff

All I know is that TFA, HFBA etc are never good news with LC-MS. Are you sure they are absolutely necessary for your method? They can be responsible for significant matrix effects which is why you might be seeing some sort of induction period. Can you still get adequate separation using more LC-compatible modifiers? Or could you use lower concentrations of modifiers?

If not and TFA is absolutely crucial, there are various "TFA Fix" formulas where you do post-column addition of something that reduces the matrix effects.

I think the phenomenon about which you are speaking is typical of LC/MS and not really the result of mobile phase modifiers. I operate a couple of mass specs (API 3000 & 4000) and never start a sequence without at least a half dozen injections and usually give it more. There is nothing worse than seeing a split curve with beginning stds all around 85% and trailing ones 115%. I use to think it was because I am analyzing different compounds practically every day and changing columns and mobile phases continually but have found it really doesn't make any difference. I rarely use mobile phase modifiers and see the same phenomenon regardless. If you are coming from the world of HPLC/UV where this sort of thing doesn't happen it can be frustrating and you begin to think something is wrong but I don't think there is. If I am going to have samples at 1:00 the instrument is up and shooting standards a couple of hours before that and I have a steady signal, or if loading a new method and I don't have the time to equilibrate the instrument I will submit at a minimum 5 high stds and 5 low stds before submitting the sequence.

I rarely use mobile phase modifiers and see the same phenomenon regardless.

I would think this exacerbates the problem. My experience suggests that in the absence of intentional low-level volatile mobile phase modifiers, ESI-MS is super-sensitive to signal enhancement and suppression from trace levels of ions such as sodium from glassware, acids and salts that are injected on the column and elute over a wide chromatographic window, etc. So if you are doing a long sequence with no modifiers, each successive injection could be introducing ersatz "modifiers" into the system, and it doesn't reach equilibrium until a certain number have gone in.

AFAIK the optimal modifier level is somewhere in the 0.1-10 mM region, which is just enough to give consistent signal (and have a more significant effect than the impurities) without excessive suppression.

I think I am in agreement with Kenn, you should always run a system suitability prior to a sample sequence and never vary the way in which you run it. Personally I believe the LCMS sources require longer equilibration times compared to the chromatography we are used to with UV detection systems. I think that simply having a flow into the source is insufficient equilibration, you need to actually put voltage through the system as if you were running a sample. 5x high and 5 x low seems to do the trick but can be extremely time consuming if you run time is longer that 5 minutes.

best regards

RJH

I experience likewise......but may be slightly different...gradient 20 - 90 % MeOH, ammonium formate pH4.1
I measure compound X and Y with respective X-D4 and Y-D4 as internal standards(D4 = 4x deuterium labelled isotope)
(the sample is a buffer pH 7.4 PBS like: 2 mM phosphate, 140 mM NaCl, Magnesium, Potassium and Calcium ~ 1-2 mM)
If I prepare a run with system suitability test at the beginning in 1 ml. vials, the X/X-D4 and Y/Y-D4 ratio looks okay RSDratio = 2.3 %
Then I have the calibration curve (in 150 µl vials inserted in the 1 ml vials)
3 injections of QC3, three inj. of QC2, three inj of QC1 (150 µl vials)
calibration curve measurement 2
injection of QC3, QC2, QC1
calibration curve measurement 3
injection of QC3, QC2, QC1
calibration curve measurement 4
injection of QC3, QC2, QC1

The QC's are each calculated back on the calibration curve run just before. After the final calibration curve (~injection 50 the QC's get into the 15% accuracy level). From injection 60 the QC's come into the 5-10 % accuracy level.

When the ratio of compound X vs. Compound Y is used throughout the same analysis run (in which it is not possible to calculate back, because of the same concentrations of X and Y in the standard solutions and QC's) the overall RSD's are : QC's : 2.3%, cal.curves : 1-10 ng/ml 4.0 %
whereas in the above setting more than 20 %

So .... depending on the internal standard, signal is highly variable and decreasing or very stable.