Hi Carolyn--

We had a similiar problem with our Costech ECS.  It has a "T" off the He 
line that is not used in our setup, and that would eventually be a 
source of an air leak (or a virtual leak after changing He tanks) even 
though it was well capped.  We ended up removing the cap and putting on 
a fine capillary pig-tail so there is always a little bleed out of the 
line (<1mL/min). 

The Zero-Blank autosampler works well, but bits of ash or dust on the 
seals or misalignment of the purge valve o-ring can cause some 
problems--which I'm sure you know.  Rarely we've had to clean the seal 
between the carousel housing and the tube to the reactor.

We always see a little N2 blank from our O2.  We've stopped using the 
high purity O2 because the N2 contaminant was hard to pick up, yet was 
running at 200+ per mil so it still threw off our N2 readings, 
especially when the signal was small.  With standard O2, we tend to see 
a peak >5mV with ~100+ per mil value.

One other thing, occassionally we'll see a "virtual" leak from a 
sample--with trapped air slowly bleeding out of the tin.  Once the 
offending tin is run, there is a sudden drop in the baseline N2.

take care

gerry

p.s.  A bad He tank can also lead to rising N2 or Ar levels over time, 
though that doesn't seem to be the case here.  The reactor tube can also 
get a fine fissure in it, but you've already replaced those.

Carolyn L K Colonero wrote:

>Hi,
>
>We have been running dC and dN on our EA-IRMS (an NA1500 w/zero blank to a Delta
>plus XP) for about two months.  The first month went great; this month has been
>nothing but headache.
>
>The first month's samples were dC and dN on whole rocks (HCl/HF cleaned) and
>bitumens and fractions thereof (sats, aros, etc).  We took what dN results we
>could get, but there wasn't much N to be had in these.  Some of the whole rock
>sample sizes went as high as 20mg.  Key here though is that through all of
>these very large samples, the background mass 30 remained low, and the standard
>deviations for both the dC and the dN results were good.
>
>This month's samples are kerogens in whole rocks, also HCl/HF cleaned, also with
>low N concentrations, and many with some amount of pyrite.  These samples tend
>to be smaller in size, in the <10 mg range, but as we run a carousel we see the
>mass 30 rise until halfway through the batch, the dN becomes nonsense.  Worth
>noting is that the dC for all of these samples is good, both accurate and
>precise.
>
>Standards run in a batch without unknowns will stay good.  Our blanks are truly
>blank.  Initially I was sure that our Cu reactor was spent, but when I removed
>it, the Cu was only half used.  After another 100 or 200 samples I changed both
>reactors again with no improvement in the conditions.  I have also replaced the
>H2O trap, and baked out the packed column overnight.  I have tried adjusting
>the O2 (timing and amount), but we were already seeing a good flash combustion.
> Except for the increasing mass 30 background, our peaks are really quite lovely
>(symmetrical and tight).
>
>Any ideas what might be causing the increasing mass 30 background?  It
>invariably arrives after about 10 or so unknowns, making the remaining samples
>garbage.  The high 30 background will go down after a few hours, but standards
>run at this point are not as good, and any unknowns will cause mass 30 to rise
>rapidly.  The only recourse seems to be changing the reactors again, but the
>reagents are not used up.  Is the pyrite poisoning our reactors?
>
>Any thoughts or advice?
>Carolyn
>_______________________________
>Carolyn Colonero
>Lab Manager
>Massachusetts Institute of Technology
>Summons? Geobiology Lab
>E34-510
>42-44 Carleton St.
>Cambridge, MA 02142
>Office: 617-253-7850
>Lab: 617-324-4002
>
>  
>