Hi, Jens,
the ion source (unfortunately) is not completely inert in terms of 
chemistry. This applies to surfaces which can for instance have an 
exchangeable layer and contribute to memory (eg the eta-effect). The 
beasts in the source are, however, freshly bombarded surfaces and the 
filament. The latter has a temperature of about 2300 C. In the Finnigan 
systems they are made from plain W, which likes to be oxidized to a 
certain degree, but it also and at the same time can form very stable WC 
(tungsten carbide). Hence, there are sources of carbon inside the ion 
source which can react with a pulse of oxygen to form CO2 (and CO+ from 
CO2+) and there is oxygen available for reaction with nitrogen to form 
N2O and NO2.
The freshly bombarded surfaces are chemically rather active areas which 
interact with anything that wants to cover them, ending in a dynamic 
equilibrium that can easily be altered quantitatively. Most prominent is 
the oxidation when O2 enters the source, leading to a loss of the latter 
and to a change in the O2/N2 ratio of air, which we measure. I am sure 
that the effects are active in most other applications as well. We for 
instance introduce a small continuous stream of CO2 into the ion source 
(about 0.2 V) for our laser ablation work on tree rings. This helps the 
long term stability a lot and might be related to this effect.
In general, every change of gas composition or number of ions in the ion 
source should be kept small. This was the trick of the old guys like 
McKinney who invented the dual inlet technique which simply balances the 
ion beams as precisely as possible in order to have the temporary and 
non-linear effects cancel out. Since the introduction and wide-spread 
success of the on-line technologies the underlying principles seem to 
have received less attention than they deserve.
Willi

Jens Dyckmans wrote:
> Dear Willi and others,
>
> the post scriptum in your posting triggered my interest since we are 
> apparently suffering from CO+ peaks where no CO2 should be 
> present.
> We are measuring N2 at fairly low concentrations (1000ppm) with 
> high O2 (50%) but N2O, CO2 etc should be eliminated by our 
> chemical trap/LN2 trap sequence in a GasBench-like setup.
> Do I correclty gather from your posting that CO2 (and CO?) could be 
> formed although no C is present in the sample?
> Sounds mysterious to be but could explain our problems (which also 
> dissapear when removing O2 from the sample).
>
> Thanks,
>
> Jens
>
>
>   

-- 
.....................................................................
Willi A. Brand, Stable Isotope Laboratory      [log in to unmask]
Max-Planck-Institute for Biogeochemistry (Beutenberg Campus)
Hans-Knoell-Str. 10, 07745 Jena, Germany      Tel: +49-3641-576400
P.O.Box 100164,      07701 Jena, Germany      Fax: +49-3641-577400
http://www.bgc-jena.mpg.de/
http://www.bgc-jena.mpg.de/service/iso_gas_lab/
.....................................................................