Sue,

Another possibility: water. If mass 18 is high, you'll get a drift in 
results for reference peak injections, quite separate from anything 
happening in the GC.

Paul


At 09:52 AM 28/04/2006, you wrote:
>Thanks Paul:
>
>I have cleaned the injector per manufacturers recommendation, but
>omitted the brushing step since it didn't look too bad. So perhaps it
>needs a more rigorous cleaning.
>
>It does seem as though the combustion column is to blame. I did get
>two decent runs of this standard before observing mass dependent
>fractionation. Prior to these last tests the new combustion oven was
>oxidized at 940C for 30 minutes followed by overnight at 540C.  I'll
>go ahead and reoxidize at each temperature longer. Perhaps the
>reaction at 940 was not complete enough.
>
>Thanks again for your input!
>Sue
>
>On Apr 28, 2006, at 11:40 AM, Paul Eby wrote:
>
>>Sue,
>>
>>Two possibilities come to mind: your combustion oven, while new,
>>might not have enough oxygen. As the run progresses, oxygen is
>>depleted and results get worse.
>>
>>The other potential source might be your injector. If it isn't hot
>>enough, or it's dirty, you could get a mass dependent fractionation.
>>
>>Paul Eby
>>University of Victoria
>>
>>
>>At 09:29 AM 28/04/2006, you wrote:
>>>Hello All:
>>>
>>>I am seeking any possible hints to solve a problem that has developed
>>>in our GC-IRMS system.
>>>
>>>At the moment the suite of standards we run for d13C of FAMEs in
>>>order to check for consistency no longer exhibits a constant
>>>correction factor across all the compounds in each standard
>>>chromatogram which range from 14:0 through 23:0 in size. In fact we
>>>are seeing greater depletion in 13C for the longer (or later RT)
>>>FAMEs by as much as a few per mil relative to the first compounds
>>>(14:0).  I have been trying to isolate the variables and this is what
>>>I've found so far:
>>>
>>>1. Flow is constant throughout the entire GC run (2 mL/min).
>>>2. CO2 injected throughout the run remains constant in it isotopic
>>>composition so does not exhibit the depletion see in compounds that
>>>must be combusted.
>>>3. I figured it must be the combustion column so it has been replaced
>>>and we are again seeing the same problem.
>>>4. Peak shape looks pretty good with widths remaining within 15-20s
>>>and fairly sharp ratio traces.  The ratio traces do become a little
>>>broader with the last few more deplete compounds but not particularly
>>>worse than runs where we've gotten good consistent values.
>>>
>>>We are using an Agilent (HP) 6890 GC with a split/splitless injector
>>>in splitless mode.  I have changed the septa (done frequently
>>>anyway), liner, and the gold seal and washer.  The GC column is very
>>>new but also has been cut (`0.5m) to see if reactive sites may have
>>>been a problem.  We are using the metal Valco cross piece and have
>>>been now for years without a problem, could this have developed
>>>active sites that may yield such symptoms?
>>>
>>>Thanks for any information anyone might have regarding this
>>>problem...I've been at this for some time now and am in danger of
>>>pulling out all my hair!
>>>
>>>Best regards,
>>>Sue
>
>Susan Ziegler
>Associate Professor
>Department of Biological Sciences
>University of Arkansas
>Fayetteville, AR  72701
>479.575.6944 (lab 6342)
>479.575.4010 (fax)
>http://www.uark.edu/~susanz/