Random samples quench PID and blow out FID

[theavenger]

Heres a really strange problem at our commercial lab:

We run thousands of samples per year for BTEX/TPH (GRO) at our facility. (Benzene, Toluene, Ethylbenzene, Xylenes/ Total Petroleum Hydrocarbons, Gasoline Range Organics) This is to establish the degree of contamination of soil and water caused by leaking gasoline storage tanks and other spills.

We use a 16 place Tekmar volatiles autosampler with 5 ml sparge tubes, with a Tekmar LSC2000 purge and trap, which is cycled then automatically desorbed into a 30 meter 0.53mm, 0.25 film DB624 capillary column in an HP5890 with a tandem PID/FID detector. Helium as carrier gas, air and H2 as fuel for the FID. Perhaps one sample out of 100 quenches (loses sensitivity) the PID and blows out the flame on the FID. It makes no difference whether the sample has any gasoline contamination or not, nor how much sample we load. (NEVER more than 50uL of methanol for a soil) The retention time shifts out about 0.4 minute when the system is restarted, then sometimes after a few blank runs the RT returns to normal. Other times it never does without a carrier gas flow increase. I have run systems like this for over 20 years and have never seen anything like it. Blanks, standards, spikes never cause this effect. If thats not enough, the previous system before I recently installed this newer (to us) one did exactly the same thing! Our other BTEX volatiles system does too! I have yet to see a problem like this on our volatiles M.S. or on our direct inject diesel testing (FID) G.C.s however.

Clearly, this is most likely a sample problem, which is often a 50 uL aliquot of a 50/50 mix of soil/methanol per EPA 8021, 8020 etc. Water samples sometimes cause it too. It seems probable that either the sample itself, or the VIAL its extracted/stored in is the culprit. Tonight, I finally have a sample that trashed both BTEX G.C.s and will look at it on M.S. tomorrow to establish any abnormality, I hope. Sure won't help if its caused by something inorganic. Water has been ruled out by our MCM and then 5 minute dry purge cycle. New traps and columns have not helped either. Carrier gas traps have been replaced. The same helium works just fine on M.S. and with the other FIDs we use here. Exactly the same air and H2 work fine everywhere.

I will post a followup, but wanted to ask the kind members here if they have ever had a similar problem, or if they can even offer a reasonable theory about whats happening in our equipment, or what else for me to have a look at.

My boss is so frustrated that he wants to have an exorcism conducted here. (just kidding) We ARE pulling what little hair we have left, however.


Best regards,

Glenn

[theavenger]

G.C./M.S. analysis of the suspect sample Friday came up with absolutely nothing. The operator was mistaken about this particular sample affecting both units. The quench and blowout phenomenon occurred again Friday evening, seems to be much more frequent, and the mystery continues. The only thing not changed on this entiire system is the carrier gas regulator, which seems perfectly alright, but I'll be changing it anyway. I may also replace the entire gas line. I'll solvent rinse the line before installing it, of course. So many folks overlook that little source of contamination. We are truly grabbing for straws at this point.



Glenn

[chromatographer1]

Interesting problem.

The PID quench and FID flameout usually occurs because of a high flow pulse through the column and the detector of plug of material which disrupts the ionization process in both detectors. I suspect the problem may lie in the valve system of the injector-purge and trap-transfer line, especially when you state the retention times change SOMETIMES returning to the original time.

This indicates a buildup of a comtaminating material which breaks free or a partial blockage of gas flow which then release a pulse of gas or a bit of both.

I wonder if you are getting ice built up somewhere in the purge and trap?

Just some ideas to ponder and perhaps they will lead somewhere or perhaps not. But maybe they may suggest another solution of the problem in another way you think of yourself.

Good luck and keep the group informed of a solution if you find it.

best wishes,

Rod

[theavenger]

Thanks for your response. Yes, we have often referred to the problem as a pressure surge ourselves. I suppose even a small bit of graphite ferrule in the column or elsewhere could cause this problem. So far, no luck in that area, but we'll have another look, and think about the valving in the purge and trap, perhaps even try another one on the system. Thanks for your input. When we find the solution, I will certainly post it here.


Regards,

Glenn

[vballchemist]

Some third party servicemen routinely bypass the MCM. They see more problems stemming from the MCM than benefits from it. Just another possiblitiy...

[JI2002]

Can you post some details of the instrumentation including parameters of the P&T, GC and the detectors? I don't think sample is the cause of the problem. However, because all your systems have same problem, it's hard to think of anything else. The retention time shift indicates a injector pressure control or P&T desorbing problem. Don't know if you noticed anything abnormal during the desorb(both GC and P&T). I'm not sure what causes the malfunction of both detectors. Make sure the carrier gas flow rate is set up right. Also check the spectra of the baseline of your GC/MS system to see if helium is contaminated. Another question is what kind of trap you are using for the P&T?

[theavenger]

Thanks be glad to. I'm posting from home by memory, and am not the usual BTEX machine operator, but do have good knowledge of the setup. (I installed it)

Tekmar LSC2000 P&T: At trap temp of 30 degrees C or less, purge begins in a 5 milliliter sparge tube, runs 12 minutes then system goes to dry purge for 5 min, desorb preheat goes to 120 C, then to desorb at 250 C for 6 min. ( GC and data sytem start run at the beginning of desorb) then to Bake for 8 min at 260 C. I have deactivated the MCM and the bake gas bypass for now. The trap is a "J", (a.k.a. "BTEXtrap")purchased fron Supelco, designed for this unit, and the P&T temps set are it's recommended values. All heated zones are set at 100 degrees C, and all pass self-test and stablize. All fans work, unit passes leak test. Incoming helium pressure set at 20 p.s.i., per specs. This device seems to work perfect in every way.

Tekmar ALS16 does not seem to be an issue here. All heaters and valves function properly, unit operates normally.

HP5890 G.C: Column (DB624, 0.53 mm, 30 meters, 3 months old) is directly connected to the heated transfer line from the LSC2000. I am not certain of the exact helium flow rate, but the head pressure is 40 p.s.i.. No makeup gas is needed with a megabore column, so the end of the column goes directly into the HNU PI-52 (newer model) PID ( a non-destructive detector, of course) then the exhaust line from the PID is connected with deactivated capillary tubing to the FID. This particular type configuration has worked for us for 2 decades now. We have been in this business for quite some time. PID detector 220 degrees, FID 250 degrees Oven program: IOT 40 degrees, hold 2 min, ramp 6 degrees min to 160 degrees, hold 1 ramp 12 degrees min to FOT 200 degress C, hold 1. Total run time just over 27 min. All compounds of interest are off by 160
degrees, the rest of the oven ramp is just a column cleanup. The G.C.has cooled and is setting with ready light on when the next sample is ready to desorb.

Most runs are entirely normal, with clean, well-resolved peaks. Good looking chromatography on both detectors. As required by regs, we calibrate at 5 levels and get a good linear curve on all analytes. We do damned good BTEX work, frankly. We are now seeing that surge problem perhaps once a week. I may not be the sharpest knife in the drawer, nor the youngest, because it has stumped me so far.

I appreciate the comments and suggestions already given. Looking forward to any additional thoughts.

Regards,

Glenn

[JI2002]

Because it happens to all your systems, I would look for the things that are common to all of them. By looking at the parameters, they look ok to me except two temps: preheat temp 120C and P&T heated zones @100C.
Preheat temp should be 5C below the desorb temp which is 250C. I would increase preheat to 245C. I also recommend to increase the heated zone temps to 125C which are a liitle bit above B.P. of water. These may not be the culpit, but the change certainly won't make the situation worse. If you have a chance to see it when the problem happens, it will give you more clues.

[theavenger]

Thanks J.I. I recall that the 5 degree difference in preheat and desorb is correct, I don't know why we adopted the 120 degrees, but I will ask. It may have something to do with naphthalene degradation, another analyte in the 8021 methodology. I really like the 125 degree suggestion, to REALLY keep the water out of the system, and think that may have good potential as a solution to our woes. I will change those zones tomorrow, and followup in a few days. Who says you can't teach an old dog new tricks?

Regards,

Glenn

[arie2044]

Hi Avenger
I'm sirprised by the column head pressure in the GC, you write it is 40 psi, i'm affraid that this is a mistake. usually column of 30 m*0.53 mm diameter works very well in the range of 6-10 psi. with 40 psi inlet your flow in the column is very high, and it may cause the blowoff of the flame in presence of small particles or water.

[theavenger]

My apologies. The 40 p.s.i. is merely the helium tank pressure, the column flow rate is controlled elsewhere. The actual helium column
flow is 8 ml/min as measured this morning. The FID lights readily and is stable until we get the surge of whatever. Thanks for your response.

Glenn

[theavenger]

Remarkably, since we replaced all of the copper tubing distributing our carrier gas, there have been no further problems. This line was 7-8 years old and had horrendous corrosion and buildup inside in spite of our use of gas filters. Apparently it was partially plugging, then releasing a glob of God-knows-what sporadically. I appreciate the great input from everyone. Case closed.

[chromatographer1]

Much appreciated.

Sometimes the devil is in the details.

Glad to hear you solved the problem.