Print

Print


Geoff,
        Bill makes several good points, which I can amplify with experience
with the 253.  The signal linearity we all are used to is based on
specifications for signal intensities between ca. 2 and 8 V measured on the
Delta, 251 or 252; Finnigan has been a little slow to establish similar
standards for very intense ion beams measurable on the 253.  We have found
little evidence for non-linearity of the major ion beams on this instrument
between 10 and 20 volts, but significant non-linearity in measurements of
very weak ion beams for doubly-substituted molecules (e.g., 12C18O18O)
collected on 10^12 Ohm resistors.  While we do not fully understand the
cause of this non-linearity, it appears to be influenced by instrument
tuning and cleanliness, and it can drift with time.  It is clearly not a
software issue, as the effect can be traced back to the rawest form of data
the machine gives you (voltages for each cycle).  The tests you have run
demand exceptionally linear response of every component of the instrument
(faradays, source fractionations, etc.) and it should not be surprising
that they fail at some level under the most extreme conditions. I also
recommend that you confirm the instrument meets specs over normal ranges in
signal intensity (ca. 5 V) and for reference gases with modest differences
in composition (ca. 10 per mil), and then slowly work your way toward the
more extreme conditions you wish to explore.  As Bill says,
composition-matching your reference and sample is the easiest way to reduce
these effects without modifying the instrumental conditions or sample size.
Our experience with the 253 has been that this simple measure effectively
removes non-linearities in the more sensitive of our measurements.  Best
regards,

John


>Geoff,
>
>I have a MAT 251 and Delta+ XL, each instrument has different tuning
>characteristics and isotopic / sample size linearities.  You should tune
>each instrument for maximum linearity, or know what the voltage range is
>for an isotopic linearity that is acceptable for you.  The new Isodat
>software versions have linearity measuring utility programs, we generally
>can get fairly good linearity from 0.5 to 7 volts in continuous flow on the
>Delta, and even lower voltage ranges on the 251 with a dual inlet
>system.  Linearities will change over time as the source gets dirty and the
>filament wears, you may want to check this periodically, especially if your
>calibration standards shift or get more variable over time.
>
>It was not clear from your message, are you measuring voltages from 10 to
>40 volts?  If so this is well above any voltages we have tried to
>measure.  You might repeat the measurements in the 1-7 volt range and see
>if you get better results.  You are doing it the right way with tank CO2,
>but you might want to make sure you are not contaminating the CO2 by what
>ever method you get it from the tank to your IRMS.  You also have to be
>careful of the isotopic difference between your machine standard and
>samples.  If you are using tank CO2 which is generally fairly negative, and
>are trying to measure carbonate samples which are near 0 per mil, the
>isotopic difference between the two gases may be large and this will
>decrease your precision.
>
>Hope this is useful,
>Bill
>
>At 03:10 PM 5/14/2003 -0700, you wrote:
>
>>We recently installed a new Finnigan MAT253. After the install, we
>>discovered that the isotopic results being reported by ISODAT were varying
>>as a function of sample size. I would normally chalk this up to incomplete
>>sample combustion or acidification, but these samples were coming directly
>>from our CO2 standard tank. Finnigan has been totally unable to even
>>suggest what the problem might be or how to pursue it. I think the problem
>>is in how ISODAT (ver. 1.5) is processing the output from the 253, but its
>>not at all clear from the manual how to troubleshoot such a problem. There
>>is a -0.6 per mil reduction in d18O and a +0.3 per mil increase in d13C on
>>CO2 samples from the same tank ranging between 10K mV and 40K mV in size.
>>Any advise or tales of similar experiences would be greatly appreciated.
>>
>>
>>
>>
>>
>>Thanks,
>>
>>Geoff
>>
>>
>>
>>
>>
>>Geoffrey Garrison, Associate Researcher
>>
>>Department of Earth and Space Sciences University of Washington
>>63 Johnson Hall, Box 351310
>>
>>Seattle, WA  98195
>>
>>(206)616-8763, office
>>
>>(206)543-6327, lab
>>
>>[log in to unmask]
>>
>>
>
>Dr. William J. Showers
>Dept of Marine, Earth & Atm Sciences
>Box 8208
>North Carolina State University
>Raleigh NC 27695
>
>(919) 515 - 7143 Office
>(919) 515 - 7802 Fax
>(919) 515 - 3689 Lab
>(919) 515 - 7911 Field Lab
>[log in to unmask]
>
>Visit our RiverNet Web Site
>http://rivernet.ncsu.edu
>
>If entropy is increasing,
>where does it come from?