The New U.S. EPA Method TO-15A blog series – Part 3: Use CLEAN air on a CLEAN analytical system!

From the start of this blog series, I have been teasing you with how to take your canister blanks from the following red trace down to the blue trace:

In this post, I shall finally stop dragging it out. BUT FIRST… (Come on now, you had to have known this was coming) I shall digress a little to point out that last time we hopefully convinced you to use AIR when analyzing AIR, and how this specific detail may impact your canister blank concentrations. This time we are going to talk about how we MUST (yes, not optional in my opinion) use CLEAN air on a CLEAN analytical system (i.e., preconcentrator, autosampler rack, etc.) before we worry about anything else. You will notice how we have not even begun to talk about canister cleaning details like cycles, temperatures, sweep gas, storage time, etc. That is all coming in due time (i.e., future posts), but none of that matters until we address the following. Last time we mentioned a prevailing theme of “garbage in = garbage out.” While we keep that in mind for today, let us forget about the zero air challenge for section 9.4.2 (page 32). In fact, let us back up a couple of pages in TO-15A and go to section 9.3 (page 30). Like I have said before, customer comments/inquiries make for the best posts. And after the last post, Melanie M. drops a comment in our blog, which includes the following excerpt:

“My question to you is do we need to test the autosampler? Section 9.3 My understanding is this is only for New autosamplers?”

It is as if we paid Melanie to drop that comment when she did, as the content and timing was perfect. YES Melanie, you need to test the autosampler in my opinion. Not just upon initial setup. In fact, you set me up for the following soapbox speech:

To kick off your journey to 20 pptv blanks, I suggest you take your fill gas (air obviously) and connect it to one of the autosampler ports and sample your nominal volume of air (e.g., 250 mL) through the entire analytical system (i.e., autosampler, preconcentrator, transfer lines, GC-MS, etc.) just like a canister blank. Now, section 9.3 of TO-15A says:

“This procedure should be conducted at installation prior to initial use of the instrument. This basic evaluation does not require establishing calibration or determining quantitative results to assess potential positive bias.”

I have the following to say about this:

1. This procedure need not be limited to post-installation and prior to initial use. I advocate using this as the first diagnostic evaluation on your road to 20 ppt_v clean canister blanks. It also makes for a nice routine evaluation of your air source and analytical system.
2. I could not disagree more on the non-calibrated evaluation. In fact, I had to soften the wording of this sentiment for the published version of this post. I advocate that you calibrate the system first, which for the record may require a secondary or tertiary calibration throughout this process, as you may have to make changes as you track down leaks/contamination. I could see a strong argument to start out qualitative, but then work your way towards a calibration. But regardless, calibrate the system for Pete’s sake. Since you are reading this post, it is safe to assume you know 20 ppt_v represents a significant challenge for canister blank cleanliness. In particular there is not a lot of room for error here. So, to start making cleanliness determinations on our air source and analytical system based on a qualitative assessment seems nonsensical.
3. All of the above will also force you to ensure you have adequate analytical sensitivity to verify cleanliness down to 20 pptv. There is not a lot of logic in running air/system evaluations on an analytical system/method capable of only seeing 50 pptv, is there?

Excuse the digression. So by carrying out the exercise outlined in section 9.3 you can verify the cleanliness of your fill gas and entire analytical system. You also probably just knocked out section 10.1.1 of TO-15A as well. Talk about killing two birds with one stone! You literally knocked out almost half of the canister blank contributions we touched on last time. Until you can demonstrate 20 ppt_v or less for target analytes by way of the above exercise, you need not worry about how clean your canisters are. In fact, that is why section 9.3 comes before 9.4. Although obvious for many, I feel compelled to point this out, as so much of the buzz appears to be focused on sections 9.4’s zero air challenge. I am concerned that this obvious, but hugely vital step found in section 9.3 may be overlooked. So thank you for your question Melanie!

Now about that teaser c-gram. What you see in the above c-gram represents the following scenario. I was looking at some canister blanks and having some higher than expected concentrations, which were not allowing me to achieve the 200 ppt_v blank cleanliness level outlined in TO-15 (notice, no “A”). After chasing my tail around with canisters, the cleaning oven, the transfer lines, etc… I finally performed the same exercise from section 9.3 and I generated the red trace in the above c-gram, which obviously was less than stellar. After going up one side and down the other on the autosampler, preconcentrator, and GC-MS all roads lead to my source gas of all things. I am still not sure what happened to this date, but I do know the fix. So without any further teasing, we put our lab air on the following Parker Zero Air Generator.

BAM! We went from red trace to blue trace overnight. And no, this is not a shameless plug to sell the above unit, as we were able to achieve comparable results with other means. For example, we used a Molecular Sieve S-Trap to achieve a comparable cleanup as with the generator. The problem is that the more economical approaches like the S-Trap do not handle the flow and volume desired for our whole laboratory operation. In addition, the restriction is too much for the higher flows we desired. And lastly, the volume is limited, so that you have to regenerate the trap more often than most care to do (at least me). So, we landed on the generator, which clearly works well. Note that the above c-gram represents a 400 mL sample of house air pre-generator (red trace) and post-generator (blue trace). It is also important to note that this is a TIC, which means the following:

1. It will show compounds not included in your target analyte list. Which can be of real value, as we will find out in another follow-up post in this series.
2. It is not representative of the true results. So, we have included the appropriate concentrations in the table below:

I am sure you have already made the observation that we only cleaned up our house air and verified our analytical system down to the old TO-15 level of 200 ppt_v and even missed the mark on 3 of the compounds. And you would be correct, as this work was conducted back in 2017. The goal of the current post was to demonstrate the importance of a clean air source and a clean analytical system. For the record, we did indeed track down the source of contamination for those 3 compounds (i.e., post gas generator), so they were no longer problematic. However, at this point in time I am going to hand over the virtual baton to Jason Hoisington (yes, same first name and darn close last name, just to keep you on your toes), where he walks us through similar exercises as section 9.3 and 10.1.1 to get us from TO-15 (200 ppt_v) down to TO-15A (20 ppt_v). After all, that is what you really came here for in the first place. I shall return with more discussions around storage time, sweep gas, etc… so stay tuned.

In the meantime, keep in mind that “garbage in = garbage out” and until you verified a CLEAN air source on a CLEAN analytical system, NOTHING ELSE MATTERS for achieving clean canister blanks!