-
- Posts: 1
- Joined: Tue Aug 27, 2019 7:51 pm
I'm having a problem with the presence of air during GCMS runs. Here's the background: I'm running a 6890 GC and a 5973 Turbo MS, using helium carrier gas. I just had Agilent do complete PM in May and all I've done since then parts-wise is change the inlet liner and septum. There have been one or two brief power outages that didn't appear to cause any damage, but no interruptions of gas flow. It sees very light use since I do low-throughput analyses and I'm still in the methods development stages.
First, I had an issue with high nitrogen coming up in tune evals/air and water checks. Around 10.5% nitrogen but <1% oxygen. Looking on chromforum, I discovered that the most likely cause was a leak somewhere before the O2 scrubber, or a contaminated tank. The first thing I did was switch the tank and let the gas already in the lines bleed out (we have helium tanks for the whole lab hooked up, and a main branch line runs through the ceiling with individual hookups at each instrument). After doing that and waiting a day, my air and water check was reading a normal level of N2 (2% or so), with 28 higher than 18 and 28 and 32 in a roughly 4:1 ratio.
Next, just to check and see what was coming in to the GCMS, I did some runs with no sample injection (i.e. not just a blank, but no injection at all), scanning from 10-100 and holding at 50°C. What I saw was an air peak, abundance of 30,000-40,000, and not just nitrogen but all atmospheric gases. The baseline was also air, but at a much lower level (5,000-6,000).
I've been doing SPME sampling, with a 2 minute splitless time at 1ml/min column flow, followed by a 50ml/min purge at 2 minutes and then 15ml/min gas saver at 2.25 minutes. The peak would begin to appear after the column dead time, at about 1.4 minutes, and continue until about 3.4 minutes (1.4 minutes after purge begins). I've also done runs scanning my normal 50-550 range, and while I obviously can't see the air, I do see a corresponding rise in column bleed (207 m/z) during the same time period.
I used an electronic leak checker, and did infact discover a couple of small leaks in the main helium lines (not in the GCMS itself), but fixing them has not changed what's happening. I feel like the leak must be at the inlet, since it's being "diluted" by the increased helium flows at the inlet after the splitless time ends, but the leak checker shows nothing (although I didn't remove any panels to check the inner workings; I just checked all visible connections and those inside the oven).
What confuses me especially is two things. First, I'm getting perfectly good tunes, tune evals, and air and water checks now. The amount of air seems only to be problematic during that peak, so diagnostics are not picking it up. Two, there is another GCMS in the lab, a 7890, and the exact same phenomenon is happening with that one- the same air peak at the same time. I've just been focusing my troubleshooting efforts on the 6890 since it's what I use.
Can anyone clue me in as to what's causing this?
One related thing I'd like to mention: in my air and water checks, peak 28 is consistently far higher than 18 (though they're both low relative to 69). I've found conflicting information online as to whether that's how it should be, or the opposite.
Thanks!