20 years running purge and trap GCMS and I am stumped

[James_Ball]

I use the autotune voltages. I have the Agilent water management kit (G7022A) wich has a BFB_autotune that is running 1211V.
I only calibrate to 0.5ppb. I get about 170000 cts. for 5ppb fluorobenzene.

Is the hardware upgrade in the kit the larger diameter opening drawout lens?

[Bigbear]

Yes it uses a 6mm draw out plate and a 1mm inlet liner as well as a 20MX0.18 ultra inert column. The best part is the BFB_autotune.

[James_Ball]

Yes it uses a 6mm draw out plate and a 1mm inlet liner as well as a 20MX0.18 ultra inert column. The best part is the BFB_autotune.

I have the larger drawout plate and a 1mm liner, have even used 0.75mm liners before. I currently use the 40mx0.18 Rtx-VMS column.

I can get enough sensitivity to run regular 524 samples at 100:1 but those UCMR samples have to be run at 30:1 split, even EST has that split ratio in their applications in order to hit the 30ppt MRL needed.

My original problem though doesn't seem to be water related since it is affecting the last few compounds to elute from the column. Also excessive water should not cause the massive carryover that I see with the original problem. It seems most likely to be something adsorbing the target compounds within the purge and trap itself. When clean tubing is used to replace what is present the problem goes away for a couple long sequences then returns shortly thereafter, no changes are done to the GCMS, only the purge and trap. Something is contaminating the system I just can not figure out what it is.

I put an indicating hydrocarbon trap on one of the instruments but it has not shown any indication of hydrocarbons entering through the helium supply, so I may have ruled that out. That would leave reagent water or purge and trap methanol that contains some kind of contaminant that will build up quickly on the sample path tubing. Reagent water is used to make the standards and blanks and to rinse the sparge tube between samples, while the methanol is used for internal standards and standard spikes. I wouldn't think the few microliters of methanol would contain enough of a high boiling contaminant to cause the problem so quickly.

The other thing could be sample preservatives. Ascorbic acid plus HCL or Sodium thiosulfate for the drinking water samples, along with a mix of HCL, or Ascorbic/HCL or Sodium thiosulfate for the waste water samples. The HCL would be the only thing common to both instruments, but that has been used for decades as a preservative for volatiles samples so the HCL itself I would not suspect, unless it has some contaminate in it I have not identified, but each sample only gets two drops of a 1:1 HCL/DI mixture.

The other culprit could be simply bad samples, probably ran on the soil purge pathway which would not have the foam sensor, but after changing out the sample pathway tubing that should go away.

Everything that makes sense can be pretty much ruled out, that is what is driving me nuts

[Bigbear]

Have you tried isolating the problem? After you get the strange result from the P&T try a direct inject of the same mass into the injection port.
Do you have a way to inject onto specific locations of your P&T to see where in the system losses happen? On my Tekmar Velocity I can put in an injection port of sorts (just a 1/8 " T with a septa in the side arm and 1/16" tubing in the other 2) this allows me to inject at desorb to test the transfer line, inject @ dry purge to test the trap ect. Tekmar has a steam cleaning procedure for their units, maybe OI has something like this.

[James_Ball]

Have you tried isolating the problem? After you get the strange result from the P&T try a direct inject of the same mass into the injection port.
Do you have a way to inject onto specific locations of your P&T to see where in the system losses happen? On my Tekmar Velocity I can put in an injection port of sorts (just a 1/8 " T with a septa in the side arm and 1/16" tubing in the other 2) this allows me to inject at desorb to test the transfer line, inject @ dry purge to test the trap ect. Tekmar has a steam cleaning procedure for their units, maybe OI has something like this.

On the regular Encon I can inject before and after the trap, on the Encon Evolution I can't. But I have narrowed it down to the contamination gathering in the empty trap tube that is the Moisture Reduction Trap. These units use an empty trap prior to the adsorbent trap to reduce the amount of moisture that will be desorbed. In a sense using a "cold spot" in the purge flow line held near room temp to condense out some of the moisture. It works very well, but what is happening is something is contaminating the empty trap that doesn't bake off at up to 260C. It is truly as if the empty trap becomes a chromatographic column since the low boiling analytes make it through to the analytical trap, but the high boiling analytes don't. It also works like a chromatographic column in that as you overload it, more of the analytes make it through, as you increase overall concentration of analytes the response of the high boilers increases exponentially. Analyze a single analyte at high concentration and it is not affected as much as when you analyze the full list of analytes. Amount of methanol has a small effect, but a standard with 90+ analytes at high concentration affects it more.

It was most pronounced on the Encon after I had a really bad sample foam. Worked for weeks trying to clean it, then had it refurbed with all the valves and internal tubing replaced along with the moisture trap. It worked like a charm for a few runs then quickly became contaminated again even though I was only running standards and blanks. Everything in the purge and trap was new, so the contamination must be coming from outside the unit, I just have not been able to eliminate it down to a single source of contamination yet. Put a small piece of 1/16" tubing to jumper the ports on the heated 8 port valve to bypass the moisture trap and everything works good, except for seeing more of the water/methanol hump in the chromatogram. I just need a lull in samples soon so I can actually trouble shoot instead of having to constantly run samples through it. Good to have the work, bad to not have time to troubleshoot

[Bigbear]

Have you looked closely at your sample rinse set up? I use plumbed in reverse osmosis water with a charcoal trap at my P&T. I've had strange things make it into the system when I let the charcoal go too long with out changing.
Do you use the same gas bottle for your gc/ms and p&t?

[James_Ball]

Have you looked closely at your sample rinse set up? I use plumbed in reverse osmosis water with a charcoal trap at my P&T. I've had strange things make it into the system when I let the charcoal go too long with out changing.
Do you use the same gas bottle for your gc/ms and p&t?

We have a tank farm that supplies the whole building with gasses, so the two use the same Helium source. That is one of my variables I am considering, if maybe we have gotten contamination in the gas lines at some point in time. We also just changed our water system so I am looking into what may be in that too. We did find that if we try to do Pesticide extractions with the DI water we get a lot of background peaks, but if we take the water after the RO and before the DI cartridges it is clean, I may start doing that with my water for volatiles as well.

[James_Ball]

I have been doing a little literature research on pyrolysis and char formation. Seems even glucose will form a char when pyrolyzed at 300C at about 30% by weight. If I have light hydrocarbons being trapped on the Moisture Reduction trap then heat it to 250C for bakeout, I wonder if that could be forming a tiny bit of char with each bake cycle, which over time becomes a layer of carbon trapped on the surface of the tube which is causing my problems. I read articles where they did pyrolysis tests on hydrocarbons, carbohydrates and coal. Drinking water samples should not have much hydrocarbons but we do know they contain some Total Organic Carbon(TOC), while my waste waters samples definitely do have large amounts of TOC present.

Going to have to do a little more study on this, but I am lowering my MORT bake temperature to 200C to see if that prolongs interval before problems occur.

[rv3]

Hi, I know this is a 3 year old thread but i have been having the same issue running an Archon to OI 4560 to 5890/5972. Was there ever a resolution? I had found that sonicating the sparge mount (ran out of ideas) and replacing the assay line from the 6-port to the water management unit would reduce the carry over for a bit; and oddly enough the connector from the carrier gas line into the Injection port, not any of the lines but the connector cleared it up on a different system (it was a zero dead-space connector from Restek). but i can't figure out a way to eliminate the problem, only delay its appearance.

[James_Ball]

One thing I have found with more research it that it seems to be related to the total mass of analytes purged not so much the amount of methanol. If you increase the level of the internal standard/surrogates then you decrease the range of total mass across the calibration standards. For example if you are using 5ppb internal standards and your curve is from 0.5ppb to 200ppb then you increase your internal standard concentration to 50ppb, when you add up the total mass of analytes including internal standards/surrogates then you can go from something like 1000x difference from lowest standard to highest standard to more like 400x (not exact ranges just an example) and when you do that the internal standard will go from almost doubling on the high standard to barely a 30% increase across the calibration range.

I discovered this when trying to work on a special project that needed a calibration range of 0.05ppb to 40ppb for the full Appendix 9 volatiles list. I began with my internal standard spiked at 5ppb, which is somewhat near the middle concentration, but I had terrible linearity in the later eluting targets and the 1,4-Dichlorobenzene-d4 doubled in area counts in the last couple standards. I used my normal internal standard mix which gave a concentration of 50ppb, and the calibration was then within 15% RSD and was giving linear responses and the internal standard was showing only about a 20% increase in response. This put the concentration of the internal standards higher than the highest target analyte concentration but it makes things run much better.

[Bigbear]

Having a weird issue as late. Using a Tekmar Stratum and Aquatek 70.
Every once in a while our CCS's last 4 cpds will be about 30% of all the rest.
Yesterday I cleaned the sample mount, flushed the tubing from the exit of the "water trap", and installed a clean sparge vessel and a new VOcarb 3000 trap.
Ran calibration curve. First 2 standards were fine, the remaining stds had the late eluters off by ! 70% ( the 2 tri chloro benzenes, hexachlorocyclopentadiene, and naphthalene).

I did 4 direct injections in a row with all cpds consistent in all injections.

Question for James, have you been able to clean your "dry trap"? Do you remove it, or flush in in place?

[cman26]

I ran into a similar issue a couple years ago using a Tekmar 3100. My last 4 compounds dropped off in sensitivity after running a particularly dirty sample from one of our clients. After fighting with it for a few days, what I wound up doing was flushing the MORT with around 50mL of hot (boiling) Methanol. After doing this, my last 4 compounds jumped way up in sensitivity and the instrument remained stable and had no more problems.

[Bigbear]

Thanks, that is one of the things left to try.

[Bigbear]

Thanks, that is one of the things left to try. MORT is the 3100's moisture control unit?

[Steve Reimer]

Use care with that boiling methanol. A fire will solve the equipment problem, but not in a good way!
When I have done flushes like that I use room temp methanol followed by hot water or 25% methanol in water.