HI Patrick--

Sounds like a burning issue,  perhaps a timing problem with the O2 
introduction?  Our drop delay is around 9sec., but we've had to vary it 
on occassion.

I take it you've already swapped out O2 tanks and checked the N2 blanks 
from the tank itself, as well as mass spec linearity.  If you haven't, 
double check the Conflo III--make sure the capillaries are positioned 
properly in the open splits and move freely up a down.  For something 
that isn't related to your problem, we also had an apparent problem with 
an air leak coming in via the purge line--which was capped off since 
we're not set up to use it.  We replaced the cap with a piece of fine 
glass capillary tubing so that there is always a very slow bleed out of 
that line.

good luck

gerry

Patrick Inglett wrote:

>Greetings,
>
>We are having trouble with N analysis using a Costech EA interfaced to a
>Delta Plus XL via a Conflo III.  This problem may be similar to a listserve
>thread back in 2004 (see Item #009153) to which I could not determine if a
>solution was presented, so here's the problem today...
>
>Salt standards (N1, N11) run very well, but organic N standards (e.g., wheat
>flour, peach leaves) return higher than expected 15N.  The deviation worsens
>with increased weight of the organic standard with up to 10 permil 15N
>enrichment at mass 28 peaks of 4 volts.  Throughout this problem 13C has
>remained nearly perfect with all standards.
>
>In my own tests, I have added C sources (sucrose, cellulose) to N11 salt and
>identified that the problem seems to be related to the amount of organic C
>present (linear increase in 15N from mass 44 peaks of 1 to 8 volts). 
>Moreover, there are different effects from different C sources.  Cellulose
>caused the greatest effect (10 permil), while sucrose only had a slight
>effect (1 permil). 
>
>The system has worked reasonably well for more than 5 years, but this
>problem just appeared about a month ago.  We've tried changing all columns
>(numerous times), water trap, baking out the EA column, retuning, checking
>linearity, etc..  Also, increasing the O2 injection volume (from semimicro
>to macro setting) actually worsened the problem by an additional 1-2 permil.  
>
>In light of the sudden nature of the change, and the fact that C type is
>important in the effect, we're running out of answers.  Does anyone have any
>suggestions as to what might be happening?  We would be most grateful for
>any advice...
>
>Thanks in advance,
>
>Patrick Inglett
>University of Florida
>Department of Soil and Water Science
>
>  
>