Purge and trap - loss of methylene chloride

Discussions about sample preparation: extraction, cleanup, derivatization, etc.

13 posts Page 1 of 1
Good afternoon,

We're currently using a Tekmar 3000 with a Vocarb 3000 trap for the extraction of the following volatile organic compounds :

Chlorure de vinyle / Vinyl chloride
Dichlorométhane / Dichloromethane
Trihalométhanes / Trihalomethanes:
Bromoforme / Bromoform
Chloroforme / Chloroform
Bromodichlorométhane / Bromodichloromethane
Dibromochlorométhane / Dibromochloromethane
Tetrachlorure de carbone / Carbone tetrachloride
Benzène / Benzene
1,2-Dichloroéthane / 1,2-Dichloroethane
Trichloroéthylène / Trichloroethylene
Toluène / Toluene
Tétrachloroéthylène / Tetrachloroethylene
Ethylbenzène / Ethylbenzene
Xylènes totaux / Total Xylenes
1,4-Dichlorobenzène / 1,4-Dichlorobenzene
1,2-Dichlorobenzène / 1,2-Dichlorobenzene

This is an old obsolete system that's always worked well, but for the past year or so we've been having problems with the loss of a large portion of the methylene chloride somewhere during the extraction step (the signal is about 95% lower than expected) from historical results. The problem has progressed to the point that we can't analyse client samples because even trace amounts of methylene chloride are quantified as reportable due to the quality of the quantification curve.

The purge and trap extraction method uses most of the parameters recommended in EPA method 524.3 (some slight modifications to the desorb temperature and time), but we've tried significantly different parameters with success for all VOCs except methylene chloride who's signal is always greatly diminished compared to the other compounds.

Injection of a standard directly into the GC column or placing the standard directly on top of the trap gives the expected signal. Extraction of a sample from the sparge unit (either in water or a standard placed directly into the unit) gives us the much smaller signal. We've replaced the both the trap and sparge unit without improvement.

Finding replacement parts for that unit is getting costly, so we decided to purchase a refurbished unit from an independant supplier to "replace everything at once" (Tekmar doesn't refurb these units anymore). Anyhow, we're been testing the new unit for a week or so and we have the exact same problem!

Tekmar tech support gave us some leads to follow, but nothing has been successful. We're pretty certain that active sites are the culprit, but we just can't locate the source. The only things the 2 units have in common is the water (which we've tried eliminating without success), the trap (which was replaced without success), the sparge unit (which was replaced without success) and the teflon O-rings used to install the sparge unit (though Tekmar tells us those do not come in contact with the sample so they shouldn't be the cause of the problem).

Tekmar recommended we clean the moisture trap (please note that this is something we removed from our old system long ago and replaced with a small length of tubing), so we removed the trap from the refurbed system, pumped some methanol through the trap and dried it in the oven for about 2 hours at 120 degrees C (we couldn't find a recommended procedure for cleaning this part).

Anyhow, this leaves us with Tekmar's final recommendation which is cleaning the mount, but we want to do this right if we're going to remove it.

As anyone experienced a situation similar to this one, or does anyone have any ideas on how to isolate/resolve the problem?

Best regards,
Mathieu
A small update : We tried injecting the sample directly on top of the trap and purging a clean water sample to see what would happen to compounds that bypass most of system except the trap and we did observe loss of sensitivity for the dichloromethane.

This is interesting, but we have replaced the trap and historical data shows that we've never had normal sensitivity for methylene chloride.

Any advise?
Can you inject a standard into your inlet and bypass the entire P&T system?
Are you using 84 as your quant ion?
Steve Reimer wrote:
Can you inject a standard into your inlet and bypass the entire P&T system?
Are you using 84 as your quant ion?


Thank you for your reply.

When bypassing the P&T system, there is no loss of sensitivity.

Historically, we've used 83 as the quant ion. We tried 84 without improvement, but we currently use 49+83. The signal is quite adequate with those ions when we don't purge the sample.

Tekmar suggested that the trap heater might be defective, but I find it hard to believe that this would be the case for both instruments.

They also suggested trying a new trap, but we replaced the trap back in late 2016 following a PT failure for Methylene chloride that started this whole process and the problem persisted.

Sigma suggested removing the teflon ferrule if there was one in the system (which was the case for the refurbished system - I replaced it with a Vespel/graphite ferrule this morning, the only type we had that could withstand temperatures above 350C) and if this didn't improve recovery, baking the trap at 360C for a few hours to see if we can get rid of whatever would be causing the degradation/loss of methylene chloride. This is what I'm trying now, and unless someone has an interesting suggestion or hypothesis, it's my last attempt before giving up since I can't justify additional troubleshooting at this time.
Are any of your other compounds higher than normal in response?

I am wondering if something was breaking down to methylene chloride and now isn't, since methylene chloride is one compound I have never had a problem with in the last twenty years. Usually anything that would adsorb methylene chloride would also cause a lot of problems with things like Bromoform which is one of the compounds I lose first.

If the trap temperature is not getting cool enough maybe you are losing it to breakthrough, but I would think vinyl chloride would go first. To test this you can purge for half the normal time and see if any responses increase.
The past is there to guide us into the future, not to dwell in.
Ahh... Looks like we've had a breakthrough (pun intended).

By cutting the purge time to 6 minutes, the signal for methylene chloride is as good as a direct injection into the GC. I only had time for one injection, so I'll see if I can reproduce the signal a few times tomorrow.

Tekmar recommends a 385mL total purge volume so we've been purging for 11 minutes at a flow of 35-40 minutes (except that I'm measuring a max flow of 28mL/min on this refurbed instrument, but I haven't been able to locate a leak). If we're getting breakthrough with this total volume, what do you think would be the proper corrective action? A lower purge temperature? A higher gas flow for a shorter purge period?

I tried a few different programs and none gave significant improvement so I'm wondering if there might not be a problem in the trap heater that gives an erroneous temperature.
I usually purge at 40ml/min for 10 minutes (400ml purge volume) with a 1 minute dry purge(440ml total purge volume). I have my purge temperature set at 38C, so the purge won't start until the trap cools down to at least that. If you keep the lab near 70F then it should not take long to cool down. If the temp is higher, then the trap can break through sooner. Purge and trap is really the same thing as GC, the analytes travel through the packing as the flow passes through the trap, so the colder it is the less distance the analytes will travel. I used to have a Tekmar 2000 with the cyro unit on the trap, that gave the best peak shapes ever because the analytes stayed in a very tight band on the trap, once they started using pack pressure to focus the analytes we didn't need the cryo.

On the Tekmar 3000 you should have a trap backpressure regulator, if you are not getting backpressure it cold be allowing the methylene chloride to pass through the trap. 4-8psi backpressure should be plenty to focus the analytes.
The past is there to guide us into the future, not to dwell in.
The trap backpressure is set to the factory default pressure of 4 psi during the purge so I was consulting the manual to see how to adjust the value and there was a line indicating that the back pressure should be set lower than the GC column hear pressure. I figured I'd better check this before proceeding and turns out the initial/standby column pressure is set at 3.9 psi. After a 2 minute hold, it rises to 4.6 psi at a rate of 0.2 psi/min and then to 11.0 psi at a rate of 0.5 mL/min. Could this somehow contribute to the loss of the methylene chloride? Considering the different flow charts for the P&T, I don't think this would be the case but it does mean that I can't raise my back pressure without modifying my GC parameters.

Thnking back, the reduction/loss of signal for methylene chloride dates back to when we replaced our old Saturn II with a new instrument. I always found it odd that the signal had dropped on an instrument that was much higher quality, but I never really took the time to understand how the different parameters affected the signal so I didn't really pursue the matter.

Anyhow,
Never had a 3000, according to Supelco you should not use a flow greater than 250 cc/min. On my Stratum ( using Teklink) the factory default for bake is 400 cc/min!
You might consider changing traps more frequently as I've never used that trap for more than 8 months before they give up.
Mathieu_CZRI wrote:
The trap backpressure is set to the factory default pressure of 4 psi during the purge so I was consulting the manual to see how to adjust the value and there was a line indicating that the back pressure should be set lower than the GC column hear pressure. I figured I'd better check this before proceeding and turns out the initial/standby column pressure is set at 3.9 psi. After a 2 minute hold, it rises to 4.6 psi at a rate of 0.2 psi/min and then to 11.0 psi at a rate of 0.5 mL/min. Could this somehow contribute to the loss of the methylene chloride? Considering the different flow charts for the P&T, I don't think this would be the case but it does mean that I can't raise my back pressure without modifying my GC parameters.

Thnking back, the reduction/loss of signal for methylene chloride dates back to when we replaced our old Saturn II with a new instrument. I always found it odd that the signal had dropped on an instrument that was much higher quality, but I never really took the time to understand how the different parameters affected the signal so I didn't really pursue the matter.

Anyhow,


Sorry I have been away for a while.

If the problem started when you switched instruments it could be related to the difference in tuning of the instrument. If the sensitivity for a given mass is different between two instruments then the response will be different for any analyte using that mass. I run two 5973s, a 5975 and a 7000 for volatiles and each one will tune differently and the response factors are all different even though the BFB tune checks look similar. If you switched from an ion trap to quadrupole instrument then the sensitivities to different analytes could be very different.
The past is there to guide us into the future, not to dwell in.
Even if breakthrough has been identified as the problem, we haven't been able to determine a proper corrective action.

We compared P&T settings between different methods and while there are some differences, most parameters should play a limited role in the problem we’re observing. We tried various purge flow rates and times and we’re observing the breakthrough at a volume between approximately 145mL (maximum signal) and 174mL (complete loss). EPA 524.3 has a minimum allowable purge volume of 240mL, and while we’re not obligated to follow EPA guidelines in Canada, we try to limit changes between our method and the reference method. We would consider lowering our purge volume only as a last resort since we haven’t measured the effect on other compound’s uncertainty and detection limits. We’ve also installed a new trap, but the breakthrough volume was the same.

We also tried modifying injection port and purge and trap pressure settings to check for improvement, but without success. We tried a constant pressure of 6 psi setting instead of a constant flow setting of 2.0mL (which gives us a pressure of approximately 3.8 psi, lower than the trap pressure setting) and also tried different trap pressure settings to slow the breakthrough, but there was little improvement in breakthough volume. Below are injection port parameters for our method and 2 reference methods, and there are significant differences. We haven’t modified those parameters yet because we’re not observing loss of signal if we bypass the trap, but I thought it was worth listing here in case anyone has some ideas.

CZRI : injection temp = 120C, column flow = 2.0mL/min, Split ratio = 20:1, initial column temp - 35C

EPA 524.3 : injection temp = 200C, column flow = 0.9mL/min, Split ratio = 30:1, initial column temp - 45C

MA.403-COV 1.1 : injection temp = 220C, column flow = 1.8mL/min, Split ratio = 4:1, initial column temp - 35C

At this point, any ideas are welcome!
The injection port won't affect breakthrough at all, so you will have to focus on the purge and trap if breakthrough is the problem, which it appears to be.

Trap temperature and trap back pressure during purge would be the two variables other than flow rate and flow volume that would change the trap breakthrough. I never had a problem with methylene chloride but at one time about 20 years ago Supelco had a problem with one of the trap materials that caused a total loss of trichlorofluoromethane on the VOCarb3000 and it took a few months to get it resolved, maybe it is something similar happening now. Maybe you could try the VOCarb4000 and see if it gives better trapping of the methylene chloride.

http://www.teledynetekmar.com/resources ... 08260C.pdf

Here is a comparison of a these traps and another from Tekmar the #9 that might work, though I would avoid the tenax/silica gel/charcoal because of the massive water retention of that older design.
The past is there to guide us into the future, not to dwell in.
Is there any way you can inject the standard mix on top of the trap during the purge cycle? This would confirm break through.
You have established that there is no loss from the trap to the column, what about from the top of the trap backwards? Inject the mix any place you can from the trap back to the sparge vessel. Eventually you will find where the loss is.
13 posts Page 1 of 1

Who is online

In total there are 5 users online :: 0 registered, 0 hidden and 5 guests (based on users active over the past 5 minutes)
Most users ever online was 200 on Fri Nov 10, 2017 4:15 pm

Users browsing this forum: No registered users and 5 guests