ESI Response not stable

[bunnahabhain]

Hi everybody

I have a very annoying issue regarding the stability of response in ESI ionization. See the data table, tells more than words:

Code: Select all

Inj. No.	Response	recovery
10		134101.063	97.7%
11		137028.422	100.0%
21		154454.938	114.1%
22		160899.297	119.4%
39		181004.703	136.3%
40		182717.781	137.7%
57		116918.531	84.2%
58		116183.773	83.7%

Recovery is calculated reom a quadratic calibration curve. Injections 10 and 11 are part of the calibration curve, later injections are from the same vial.

Any idea what could have happened? If I see such a trend (between injections 10 and 40), is my calibration curve reliable? I don't think so.
This is not a one-time observation, this happens very frequently on this machine.

Details on the method: Separation of ethanolamine in 30 mM NH4formate / ACN gradient on a Primsep column. Retention time and pressure are stable. MS is a Quattro Ultima Pt operating in SIM mode, ESI+. There is a little salt buildup on the baffle plate, but in my opinion this should cause a downwards trend of sensitivity rather than an upwards trend, and especially not such a sudden drop in the middle of a sequence.
"Funny" detail: Without cleaning the source, using the same eluent and samples, the signal was back to 100 % and remained stable over a sequence of similar length.

Greetings,
Jörg

[lmh]

I'd say that this level of variability is quite to be expected with electrospray. It's compound-dependent as well as instrument-dependent, and it's just a feature of ESI - and the reason for using an internal standard if available, and randomising sample order in all circumstances. Yes, it's a bit strange when an instrument improves as it gets dirtier, but not completely illogical. After cleaning, peak areas are certainly likely to vary more rapidly with successive runs than they will after 20/30 runs, because the system is all nice and clean and each new run makes a disproportionate change in dirtiness. Once everything is happily saturated with an even level of dirt, it can be that not much more will change (until it gets so filthy it fails). But dirt can have varying effects. An extreme example I had some years ago was someone who was running a method with 10% formic acid. As I remember, this was actually led to a slight increase in sensitivity in negative mode, but completely destroyed positive mode sensitivity. I think the reason was that it lined the capillary from spray-chamber to instrument with formate, and turned it into a sort of gas-phase ion-exchange column, mopping up all positive ions. Negative ions probably got through slightly better than normal because the walls were repelling them? Perhaps I'm unnecessarily defeatist, but I now regard electrospray efficiencies as one of the mysteries of the universe. Perhaps in your case there is an initial accumulation of "good" (helpful") dirt, followed by a longer-term, sample-dependent accumulation of generally unhelpful dirt??? Or perhaps it's just random fluctuation.