EPA 8260 Optimization (GC/MS w/ Purge-Trap & Autosampler

[Ron]

The problem is not the VOCARB packing itself, but is related to the differences between the Tekmar and the OI purge and trap systems. The Tekmar uses a heater jacket around the trap for heating, and the temperature measured at the thermocouple in the heater jacket is higher than the actual trap temperature. OI uses direct resistive heating of the trap with a thermocouple welded to the trap, so the measured temperatre is the actual trap temperature, if the thermocouple is attached correctly, more on this later. Since trap life is related to temperature, a VOCARB trap will usually last longer and give more consistent performance if it is operated at about 25C lower temperature than in a Tekmar.

One of the big problems is the lifetime of a VOCARB trap in an OI. I don't know the current status of this issue, but a few years ago OI manufactured all of the traps for its instrments other than the VOCARB, which I believe is a #11 trap. Supelco had the VOCARB as a proprietary packing, and no other companies were allowed to manufacture the VOCARB trap. That isn't a problem with the Tekmar since it uses an external thermocouple to measure the trap temperature, but with the OI it is a big problem if the thermocouple is not welded to the trap correctly. IF the weld is not done properly the temperature measured is wrong, which can have a huge effect on trap lifetime and performance. When you inspect an OI trap the thermocouple wires should form a V which just touches the trap. If the wires cross above the weld, or if the two wires are actually welded separately the trap is actually much hotter than the measured temperature. The worst case I have heard of is one where the trap lasted two days, and was black on the outside when removed. Not all of the Supelco traps are bad, but it's hard to tell until you actually use one. I always have used an OI #10 trap for GCMS work.

The need for desorb preheat is related to trap heating rates. If the trap is heated rapidly there is no need for trap preheat. The OI has never really needed trap preheat, and it is much less important for newer Tekmars than for older ones.

[Caged Alpinist]

The traps I am using come from Supelco... In reading the app notes from both OI and Supelco, I see that a dry purge is necessary to reap the benefits of the VOCARB traps in respect to less water getting into the system. There are currently no dry-purge settings in the P/T program, I am adding a dry purge to the cycle today. I will also be ordering an OI #10 to compare results.

I really appreciate everyone sharing their experience with me!

A.W.

[LindseyPyron]

If you are getting overly frustrated, call our tech support line at 800-283-3510. We will be glad to help and will not charge you anything.

[Ron]

The water management on OI purge and traps is pretty effective, I don't know if dry purge will do much more than lengthen cycle time. Dry purge won't do anything on the OI #10 trap, with the silica gel in the trap you can't dry purge out water.

[Bigbear]

On first glance I would suggest a 3-4 min hold time at 35C so the gases can interact with the liquid phase. I would use a quicker ramp after that.
I use 35C for 4 min then 15C/min to 220. I am running 524 on the same GC/ms as you but with a DB-624 column of the same dimentions. I use a 0.6 ml/min flow. I am using a Tekmar Stratum with a #9 trap ( a Vocarb 3000 with an additional bed). I use a 2 min dry purge @ 100 cc/min and still get good recoveries for bromomethane ( if you dry purge for too long it can be blown through the trap!).
The shorter you can desorb the better, as after 1 min or so all that happens is more water gets to the source. I have run as little as 1 min desorb and get good recoveries but for my method the lowest EPA will accept is a 3 min desorb. Rumor has it that 524.3 will allow 1 min desorb.
I have found that this mass spec source can be made active with all the water being transfered to it. I was getting very poor linearity for my aromatics and nothing I tried worked except for cleaning the source.
I now have to clean every 150-200 samples. I monitor the fluorobenzene peak areas and when they approach 50% of my inital calibration post cleaning I clean again.

I long for my 5970 with a jet!!!!!!

Good luck
Bear

[Bigbear]

Some additional thoughts. I feel that your injection port is a little too hot. I use 150. If you can change your P&T temps I would set the zone where the 6-8 port valve and the transfer line to 150 as well.

A trick I was told for re-setting the column calibration of a 6890 ( after column change or multiple column clips). Set your oven to 40 and set the gc to the stop watch feature. Go in to manual tune and set a mass to 28, shut off the calibration valve. Set the system to repeat profile. Take a syringe and inject 1-2 ul air and start the stopwatch. Wait for the air to hit the detector, hit the stop watch. Do this 3-4 times and average the results.
In the gc software parameters go to:
Column
calibrate
at the lower right of the screen click on calculate length (if un- retained compound time is known)
type in the value calculated for the air
click OK
Then save your method.

[Caged Alpinist]

Yes, I think we are getting a bit of thermal degradation at the inlet. I also believe that I have finally identified the culprit behind the terrible separation. I found a tiny crack at the inlet assembly on the end of the xfer line from the P/T. I also believe that the xfer line may have had a crack or several cracks. Replacement parts arrive tomorrow, hopefully that improves things a lot.

Some other notes:

I switched to a straight liner rather than the flared. Peak shape improved significantly. I also removed the dry purge and cleaned the hell out of the P/T lines by filling the sparge tube with methanol and attaching a drain tube to the trap inlet fitting. Then I turned up the sample temp to 70 C and.... Instant MetOH steam cleaning cycle...

If the separation is still bombed out after replacing the inlet assembly and xfer line, I'm getting a hammer...

Cheers,

A.W.

[Omaejel]

With a fully optimized 5890 running an OI 4660 and an Archon autosampler we have gotten times between samples down to as low as 17 min. That would be fully optimized, but we generaly run in the 23min range with the older Tekmar P/T.

I hate anything longer than a 1min. desorb.

You should be able to get an 8260 run off within about 14 min on the GC, though your column is a little short. I would just toy with the ramp parameters and play with the column flow and see what you can achieve. I have never used that column.

We use a 30m .250 column and run 1ml/min through it.