Print

Print


Hi Matt,

Do you know it's CO? Not N2? What are the argon backgrounds?

Andy

On Wed, Jan 11, 2017 at 10:08 AM Matthew Lachniet <[log in to unmask]>
wrote:

> Hi all,
>
> I have a question for TC/EA users.
>
> Over the past four weeks I've been resurrecting from the dead my TC/EA for
> water analysis after no use for three years.
>
> The last major issue among a long list of them is that I've experienced
> very high CO backgrounds (28/29/30) of >500 mV that don't decrease over
> time. H2 backgrounds are very low to be essentially negligible. As I
> understand it from the ISOGEOCHEM archives, the glassy carbon tube (GCT)
> can react with the ceramic tube to produce CO and Al2. Indeed, when I
> removed the GCT, which has only seen about 20 water injections since
> installation, the hot zone had large amounts of reaction products, and I
> found ash particles in my bottom feed adapter.
>
> My question is: how can the generation of CO be avoided? Both the ceramic
> tube and the GCT are new, and I had the same results with an old ceramic
> tube/new GCT combo. The high backgrounds are preventing usable data for
> d18O, but the d2H data are fine. Both tubes are straight and have no bends
> or deformities that would cause contact between them.
>
> Other background: the Helium supply is clean. Backgrounds are very low
> when I turn on He dilution in the ConFlo, and backgrounds disappear when I
> close the needle valve to mass spec, so I'm pretty sure the CO is coming
> from the TC/EA. I have done extensive leak testing and found none. I've
> baked out the GC at 150+ degrees many times with no appreciable effect on
> backgrounds. The high backgrounds start to appear when reactor temperature
> is above 1300 C. Though I don't have all the specs off the top of my head,
> the parameters for He flow rate are the same as previously successful runs.
> I've removed and reset the GCT/ceramic tube combo about three times to
> ensure there wasn't some leak, and each time the high backgrounds return.
>
> Any advice would be appreciated.
>
> Thanks,
>
> Matt
>
>
>