Gilles St-Jean wrote:
> 
> Hi all
>     In the 12 years I've been working here no one ever asked to do
> N-15.   Well odds were against me.  Someone asked.
>     The materials are plant samples where the 14N/15N ratios are to
> be determined at natural abundances.  The samples are from a growth
> chamber.
>      Instead of reinventing the wheel, or finding the blueprints for
> it, I am looking for the following info.
>      I have a Carlo-Erba with an off line prep system to capture the
> effluents. I am presently using it for 13C and 34S. I know about
> the Keljall prep methods, but I am not equiped for it.  What I would
> like is to find a way to trap the N2 in a sample holder from the Erba
> to take to the MS. I've heard of someone using a molecular sieve, but
> I have no idea what materials are used.  Any info would be greatly
> appreciated.
>     Also what is the best way to calibrate (ref gas, standards)????
> 
> Gilles
> Gilles St-Jean, U. d'Ottawa / Geologie
> 140 Louis Pasteur, Ottawa, Ontario, Canada, K1N 6N5
> Tel: 613 562-5800 ext: 6839(Bureau / Office)
>      613 562-5800 ext: 6836(Lab)
> Fax: 613 562-5192
Dear Gilles:

The ability to trap C,N and S gases combusted by Elemental Analyzers has
been available for some time by stable isotope ratio instrument
manufacturers such as VG Isotech (now Micromass) and Finnigan MAT.  The
double and triple traps were the result of this methodology.  These
traps allowed C and N samples to be combusted and then analyzed by dual
inlets.  They worked very well for C and reasonable for N samples.  The
difficulty with N samples is that they require molecular sieve or
silical gel to adsorb onto.  Analyzing C and N in this manner gave good
precision and accuracy for natural abundanc samples (ie. 0.1 per mil
external precision).

Recent advances by IRMS manufacturers (Micromass, Finnigan MAT and
Europa)allows C,N and S samples combusted by Elemental Analyzers (such
as available form Carlo Erba -Now CE Instruments) to be analyzed by
continuous flow (CF-IRMS).  External precision that can be expected are
the following: C <0.15 per mil, N <0.15 per mil with samples of approx.
50 micrograms.  This approaches the precision obtained by trapping but
much easier and faster.

The reason the above described trapping worked was because the samples
were trapped and released by software and as such the trapping/heating
cycles were identical from sample to sample.  Trying to reproduce the
trapping/heating cycles manually is not possible and most people that
try to manually trap samples have varying results.

Some people have mentioned that you should interface your EA to your
IRMS.  The interfaces available for you to do that would not give you
the external precision mentioned above.  Depending on the linearity of
your source you can expect the following: C <0.3 per mil, N <0.5 per
mil. In some cases better performances have been obtained but you cannot
predict that.

My recommendation is to not even try to trap your N samples but have
them analyzed by a laboratory that has current technology CF-IRMS
instruments.

Regards,
Nik Binder