Just an update regarding our pyrolysis issues on our Isolink (described below) just in case anyone ever sees a similar problem. Our sincerest thanks to Henry Holmstrand, Patrick Griffin, and Paul Eby who all replied with very helpful suggestions.
Based on our symptoms of changing delta values with successive injections and improvement in delta value accuracy following reconditioning of the reactor, we were all on the same general reasoning path that something had to be oxidizing the graphite out of the pyrolysis reactor. My first (easiest) move was to swap the He carrier gas tank (Henry Holmstrand's idea) to see if it was contaminated and introducing an oxidant to the reactor. Unfortunately, that had no effect, but a great idea! Patrick and Paul both suspected a leak. I had searched the pyrolysis reactor path NUMEROUS times with Ar and never saw a hint of a leak.
The key to that sentence is that I had ONLY searched the pyrolysis reactor pathway and had ignored my combustion reactor. As well as I thought I understood the Isolink, I clearly do not, because apparently a leak on the combustion reactor can find its way to the pyrolysis reactor...I still do not understand how this is possible unless the Valco valve is a bit leaky internally or the flow paths via the MCD are not what I thought they were. Well, in any case, isolating the combustion reactor solved the problem and our standards are now looking normal again (knocking on wood as I write this).
Thanks again for your suggestions Henry, Paul, and Patrick. Next time any of you are at the same conference with me please print this e-mail as your free beer coupon and I shall honor it.
I know listserv's are (or have?) largely gone out of vogue, but thank goodness for this one!!!!!
From: Stable Isotope Geochemistry <[log in to unmask]> On Behalf Of Henry Holmstrand
Sent: Saturday, December 8, 2018 6:29 AM
To: [log in to unmask]
Subject: Re: [ISOGEOCHEM] GC-IRMS Odd Behavior with dD
How about air impurity in the carrier gas?
Ute på vift
> 8 dec. 2018 kl. 02:37 skrev Patrick Griffin <[log in to unmask]>:
> My first reaction is an air leak somewhere upstream from the TC that is degrading the graphitization. I'm sure you see why that would fit, and so you have troubleshot around it with He sniffer and Ar overpressure on every conceivable fitting. The second is the magnet current regulator, which I know nothing about in the V but is straightforward to diagnose on the software side. The other explanations are harder to pin down... shitty for you, fun for us puzzles obsessives.
> How closely have you monitored the high temperature conversion device? My imagination has concocted a elaborate scenario where everything you've seen matches a nearly-but-not-quite dead heating element with a hairline fracture somewhere. That could either be the wise and helpful voice of experience or the bitterness and resentment of someone who has PTSD from a silicon carbide heating element (Roxanne, Seth, Marilyn?). I don't know about your device at all, but if you crack it open and find a variable resistor, it may be worthwhile to move it up a step. If you find the time to fully power down (unplug!) your device, checking the resistance along the element when cold will be definitive.
> An overnight trace of standard gas that bypasses the thermal conversion can rule out the magnet current regulator. I would also monitor the injection volume as a function of intensity and d2H since conversion efficiency is dependent on temperature and bond strength is a function of redox conditions.
> If conditioning returns the B4 standard to good values and good peak heights, consider overloading your injections with a fast-eluting compound that conditions the reactor on each run. Works a charm for me.
> Keep us updated.