on-line pre-concentration in RP HPLC

[Christine]

Hi there!

I've been observing a phenomenon using an LC-ESI-MS/MS method with
RP18 column (2x55 mm, 3µm) to quantify a small molecule (Mr ~ 350,
cyclic amino acid with a rather lipophilc substitution at the hetero-
N) which I have trouble explaining properly:


Now, in our experience we reach better LLOQ when using relatively
large injection volumes (50µl; flow 350µl/min). In one series of
experiments I've analysed increasing injection volumes (5 to 50µl) of
a 1nM solution of the analyte - once with installed column and once
without the column (dead volume free connection). In the series
without the column, I observe increasing peakwidths and therefore
also
increasing peakareas, but the signal intensity stays about the same
(meaning S/N; max. intensity ~600cps). However, when I perform the
same series with the column, not only the peakareas increase but also
the signal intesity (linearly with increasing injection volume) and
therefore I reach a better LLOQ (--> S/N increases significantly;
max.
intensity ~2400cps).


I concluded that I have somewhat of a pre-concentration effect on the
column (ESI is described in the literature as a concentration
sensitive detector). However, I cannot fully explain why I see that
because I use the same solvent composition to dilute my sample and
for
the eluent (MeOH/buffer = 50/50). Does the lipophilic interaction
between the analyte and the column material cause such a pre-
concentration? I mean, if I had used water as the sample solvent and
the said eluent it would be clear. But with the conditions used, I
would rather expect band broadening and not a pre-concentration
effect.


Has anyone observed something similar? And can someone help me with a
"theoretical explanation" for this?


Thanks a lot!!
Christine

[Christopher R. Lee]

Direct injection is alwayse a waste of time, because a nanogram per litre of
something in the matrix can (depending what it is, e.g. a surfactant)
enhance or suppress ionisation.

A column doesn't remove all risk of matrix effects, but the situation is
much more under control.

Of course, if you need to improve your detection limits, you would run a
monotonic or stepwise gradient. Very large injections are common in, for
example, water analysis.

Regards


"Christine" <***@hotmail.de> a �crit dans le message de news:***@q69g2000hsb.googlegroups.com...
Hi there!

I've been observing a phenomenon using an LC-ESI-MS/MS method with
RP18 column (2x55 mm, 3µm) to quantify a small molecule (Mr ~ 350,
cyclic amino acid with a rather lipophilc substitution at the hetero-
N) which I have trouble explaining properly:


Now, in our experience we reach better LLOQ when using relatively
large injection volumes (50µl; flow 350µl/min). In one series of
experiments I've analysed increasing injection volumes (5 to 50µl) of
a 1nM solution of the analyte - once with installed column and once
without the column (dead volume free connection). In the series
without the column, I observe increasing peakwidths and therefore
also
increasing peakareas, but the signal intensity stays about the same
(meaning S/N; max. intensity ~600cps). However, when I perform the
same series with the column, not only the peakareas increase but also
the signal intesity (linearly with increasing injection volume) and
therefore I reach a better LLOQ (--> S/N increases significantly;
max.
intensity ~2400cps).


I concluded that I have somewhat of a pre-concentration effect on the
column (ESI is described in the literature as a concentration
sensitive detector). However, I cannot fully explain why I see that
because I use the same solvent composition to dilute my sample and
for
the eluent (MeOH/buffer = 50/50). Does the lipophilic interaction
between the analyte and the column material cause such a pre-
concentration? I mean, if I had used water as the sample solvent and
the said eluent it would be clear. But with the conditions used, I
would rather expect band broadening and not a pre-concentration
effect.


Has anyone observed something similar? And can someone help me with a
"theoretical explanation" for this?


Thanks a lot!!
Christine

[Christine]

Post by Christopher R. Lee
Direct injection is alwayse a waste of time, because a nanogram per litre of
something in the matrix can (depending what it is, e.g. a surfactant)
enhance or suppress ionisation.
A column doesn't remove all risk of matrix effects, but the situation is
much more under control.
Of course, if you need to improve your detection limits, you would run a
monotonic or stepwise gradient. Very large injections are common in, for
example, water analysis.
Regards

Thanks for the reply!! But see, there is my problem: I'm fully aware
of the mischief which the matrix effect can cause in LC-ESI-MS/MS and
of course I also checked on that in my final method where a biological
sample is injected (and there I see signal suppression despite the
column chromatography).

However, in the experiment I explained, I use a "pure solvent"
solution (meaning isokratic eluent = solvent) - so I weighed my
compound which was bought from RBI (purity >99%) and diluted it to the
desired concentration using the HPLC-grade eluent. Do you think this
last 1% (see purity) can cause such a matrix effect?

The other problem, as I explained, is that the solvent and the eluent
have the same composition. In the water analysis you are mentioning,
the online pre-concentration occurs because they inject water (low
strenght solvent) and elute using some kind of organic modifier
(higher solvent strength) on an RP-column. This causes the organic
analyte to "adsorb" in the "head-volume" of the lipophilic column
material until the surrounding water is "displaced" by the eluent and
the organic analyte is moved on in a pre-concentrated band. So
compared to the water analysis I still cannot see how to explain
theoretically what I observe in my experiment.

Any other ideas?
And thanks again for fruitful thoughts!!

Christine