Issue with VOC analysis using Agilent Atomx P&T system

[tangaloomaflyer]

We have two Atomx purge and trap concentrator/autosamplers hooked up to 2 Agilent 6890/5973s. We have always had trouble on these sytems with sensitivity analysing VOCs (USEPA 8260) and have trouble achieving linearity for certain analytes.

My interstate colleagues claim to have fixed the issue (with assistance from Agilent technicians) by improving the sensitivity of the GCMS (using ultra-inert consumables), but to me this seems like the GCMS is compensating for the shortcomings of the concentrator?

There is an Agilent application note for 8260 on the system, but it requires BFB tuning that is not supported by 5973 MSDs. It also specifies ultra-inert septa and gold seal, plus an ultra-inert column. We have never had to use ultra-inert consumables to analyse VOC at 2ppb

Has anyone else had problems with this system?

[LindseyPyron]

**Warning - Commercial ahead***

I would encourage you to investigate the difference between the AtomX and the systems from EST Analytical (www.estanalytical.com). In general, our systems can run at much lower split ratios, improving sensitivity, due to our improved moisture control and lower dead volume on the desorption side of the instrument. I can send a couple of application notes demonstrating this if you want to get in touch at lpyron@estanalytical.com

[Bigbear]

That being said there is now a known issue with Agilent 5973's and 75's dealing with the water that gets transfered to the MS durring desorb.
There is a fix Agilent part #G7022Awhich goes for ~ $1000. If you are under contract you may get it under the contract as we did.
All I can say is that it works! We struggled for at least 2 years with a 5973 before installing the kit. Smooth sailing after.

[James_Ball]

There is also a paper done by Agilent where they study non-linearity of the higher boiling compounds. If you are having trouble with the Dichlorobenzenes or Trichlorobenzenes it helps. The trick is to reduce the Electron Multiplier Voltage as much as possible. We dropped ours back to about 1100 volts and are running with a threshold of about 10 counts and split ratio to 30:1 and now we have much better linearity for those compounds.

If you run like we do where you are have the instrument active and analyze about 120 samples in a little over two days, you will see a gradual loss of sensitivity but not more than 30% on the internal standard counts over that period. If you allow the instrument to rest for a day they come back pretty well. The loss is due to the water that builds up in the system. One thing we do to help with this is run the old style molecular sieve moisture traps on the rough pump like Agilent sold with the 5970, 71 and 72.

Another thing we found is that the ETP multipliers are less prone to the non-linear behaviour over concentration range than the K&M style.

Oh and I would add a plug for the EST concentrators too, I really like those

[tangaloomaflyer]

Lynsey et al, re EST concentrator: Thanks, I has a rep show us the specs and some app notes for one of their concentrators and we were impressed, however at the time weren't in the market for a new concentrator. I would like to trial an EST, but I'll have to work on our technical director. I would like to get the Atomx systems we are stuck with working!

Bigbear: I have purchased some of the components of the VOV kit, but haven't had chance to install them yet, and thought it might be futile if we can't do the BFB tune as specified in the app note (http://www.chem.agilent.com/Library/app ... 029\EN.pdf).

James Ball: Well, this is awkward, but I can't remember the analyes we are having problems with (I'll check at work tomorrow - I haven't worked on these instrumnents for a while). We have played around with reducing the mass spec sensitivity to improve linearity but we can't reach the detection limit on some of the problem compounds. At t would have to what point in the gas stream do you install the molecular sieve? The concentrators have moisture traps before the analytical trap, so if you are using it to reduce moisture in the sample gas stream desorbing from the trap is there any issue with retention of the analytes in the sieve? We don't see much moisture in the tune.

Thanks for the help.

[Bigbear]

The G7022A comes with a new BFB_autotune macro. It optimizes the emission current as well as everything else an autotune does. It works very well. We have been using this "fix" for about 18 months with no problems.
I think the moly sieve traps used to be installed at the rough pump.

[James_Ball]

Lynsey et al, re EST concentrator: Thanks, I has a rep show us the specs and some app notes for one of their concentrators and we were impressed, however at the time weren't in the market for a new concentrator. I would like to trial an EST, but I'll have to work on our technical director. I would like to get the Atomx systems we are stuck with working!

Bigbear: I have purchased some of the components of the VOV kit, but haven't had chance to install them yet, and thought it might be futile if we can't do the BFB tune as specified in the app note (http://www.chem.agilent.com/Library/app ... 029\EN.pdf).

James Ball: Well, this is awkward, but I can't remember the analyes we are having problems with (I'll check at work tomorrow - I haven't worked on these instrumnents for a while). We have played around with reducing the mass spec sensitivity to improve linearity but we can't reach the detection limit on some of the problem compounds. At t would have to what point in the gas stream do you install the molecular sieve? The concentrators have moisture traps before the analytical trap, so if you are using it to reduce moisture in the sample gas stream desorbing from the trap is there any issue with retention of the analytes in the sieve? We don't see much moisture in the tune.

Thanks for the help.

As Bigbear said, the molesieve trap is attached to the inlet of the rough pump. It works to keep moisture from entering the pump oil and shortening the life of the oil. It also helps somewhat if there is a leaky valve there as it can catch oil vapors that might try to work their way back up to the detector.

[tangaloomaflyer]

OK, but Bigbear said there is a problem with moisture from desorb reaching the MS - this, I presume is moisture that had sorbed onto the trap from the sample? Installing the VOC-kit would not reduce the moisture in the sample stream, but I guess it minimises the damage doen by the moisture.

It seems as though we need to improve the sensitivity of the GCMS to compensate fro the shortcomings of the concentrator not drying the sample gas flow effectively.

[James_Ball]

Honestly I think to compensate for the moisture we need to reduce the sensitivity of the instruments. I ran a 5970 and had very little problem with moisture, barely made a blip in the baseline back then. The 5971/72 were similar, saw a little disturbance in the baseline as the moisture came through and just a little curvature of the calibration lines on the heavier weight targets. On the 5973 I always had trouble with non-linearity with the original setup, then changing the drawout plate to the larger diameter(which reduces sensitivity)helped compensate. The 5975 is the same as the 5973. Now I have a 7000QQQ that I am trying to setup for volatiles along with other tests. This new one is very sensitive and when the moisture comes through from the desorb the baseline gets very high. Even when I set MS2 to filter everything below 45m/z some gets through. I am thinking about trying to use MS1 as the filter to remove background then use MS2 as the scanning quad for analysis and see how that works.

The best way to eliminate moisture is to use a 30 second to 1 minute desorb time, which the EPA disallows in Method 524.2 but has begun to allow if you move to Method 524.3. Method 624 is up to the interpretation of your auditing body, since it is actually a packed column GCMS method, it is always highly modified as it is currently run in pretty much all laboratories. Method 8260 is much more open to using the shorter desorb time as long as you get the same performance as the longer time.

The best moisture control device I ever used was on the old Tekmar 2000 which froze the water out with a Peltier cooled portion of tubing between the trap and transfer line. If set at -5C it would remove all of the moisture but you did lose sensitivity for the Trichlorobenzenes and Naphthalene.

[tangaloomaflyer]

Ah... I didn't think of it like that

I am in Australia so not bound by USEPA, but our method is based on 8260. We only have to answer to NATA and they are not particularly concerned about the nitty-gritty of the method as long as we meet acceptable quality criteria - so, I should be able to play around with the desorb time. We are running 7min purge and 3min desorb at the moment.

[Bigbear]

A trick that can be used to reduce desorb moisture if you must purge for 4 min is to up the split half way through the desorb using the gas saver option. So if with your split rate the total flow is 100cc/min, using gas saver to set the flow to 150 cc/min or so @ 2min would increase the split ratio and reducing the water transfered to the ms. The added benefit is the increased flow helps to "dry" out the transfer line between runs.

[James_Ball]

Ah... I didn't think of it like that

I am in Australia so not bound by USEPA, but our method is based on 8260. We only have to answer to NATA (National Australian Testing Authority) and they are not particularly concerned about the nitty-gritty of the method as long as we meet acceptable quality criteria - so, I should be able to play around with the desorb time. We are running 7min purge and 3min desorb at the moment.

This is one advantage most other countries have over the US. Most other countries have the policy that if you can meet or exceed the quality control and detection limit criteria of a test you can use that modification. Here once you make those changes if you are running anything other than 8000 series methods, you have to apply for an ATP(Alternative test procedure) and that can take up to 5 years to be approved for waste water and drinking water analysis. By then even newer technology has come along and you are ready to move to that instead of what just got approved.

[whalozan]

The best purge&trap concentrator I have used for water control is the OI 4660 Eclipse, their cyclone water management system really works and the IR heater does a great job. I was able to get down to 0.2 ug/L with a 5mL purge for most compounds.