Ethylene Oxide – Storage and Stability in Air Canisters

So, what exactly does this bias look like? If you were fortunate enough to attend the virtual National Environmental Monitoring Conference (NEMC) in 2020, you may have caught a talk I gave on this exact topic. For those of you who didn’t attend, you can find a copy of the presentation at https://apps.nelac-institute.org/nemc/2020/docs/presentations/pdf/8-4-20-Air%20Methods,%20Monitoring,%20and%20Technology-4.02-Hoisington.pdf. What we’ve found is that a positive bias in EtO seems to be tied to two things – the fill gas (humid air vs. dry air or nitrogen) and canister cleanliness.

Let’s start with the fill gas. We had received customer feedback that some canisters they had in use for EtO were checking out clean in the lab, but giving abnormally high results when used in the field. When we checked out the canisters in question in our lab we found some background EtO present, but only if they were filled with humid air. Dry air or an inert gas such as helium/nitrogen were non-detect for EtO. This matched with the customer experience, as they used nitrogen to clean and fill canisters for blank checks.

Table 1: Comparison of EtO blanks using different fill gases. Average of 3 samples.

We found that this background EtO contamination was something that occurred regardless of the canister typed used, whether it was a plain electropolished TO-Can or a SilcoCan. Testing 2 competitor canisters equivalent to our SilcoCans also revealed growth of EtO over time when filled with 50% RH air.

Figure 1: Graph of EtO growth for various canister types in 50% RH air. Average of 3 samples each, error bars are 1 standard deviation.

In case we haven’t made the point enough in our TO-15A blog, this really underscores the need to use humid air when qualifying canisters. The use of dry air or inert gas can mask possible issues with your canisters for EtO and other compounds, including some volatile sulfur compounds.

You’ll note that one of the competitors was significantly higher than every other canister tested, which brings us to the second point – canister cleanliness. When we look at the chromatograms of these canisters, as well as some customer canisters that we were told had issues with EtO bias we see that they have a significant amount of unknown contamination at the end of the chromatogram, much higher than the internal standards.

Figure 2: Comparison of canister cleanliness. Black – customer canister. Blue – Competitor canister. Red – Restek TO canister.

In addition, we had some customer canisters that had seen heavy use in the field and were now having cleanliness issues unrelated to EtO. When we tested those cans not only did they show a very high amount of contamination in the semivolatile range, they also had a high level of EtO, with blanks at 5.9 ppbv. After thoroughly cleaning the canister using a proprietary cleaning process, the semivolatile contamination was gone and the background EtO became ND.

Fig. 3 – Customer canister with heavy field contamination and 5.9 ppbv background EtO (top). Customer canister after cleaning, ND for EtO. Filled with 50% RH air, analyzed 7 days after filling.

This seems to indicate that EtO can be produced by oxidation of larger contaminants built-up in canisters, possibly catalyzed by the presence of water and the metal surface. Even canisters that have not seen field use, such as the ones used to generate the data in figure 1, show measurable EtO growth over time. If you are using selective ion monitoring (SIM) to improve your EtO detection limits and check your cleaned cans using only the SIM method you’ll likely miss the semivolatile contamination that seems to be correlated with high EtO growth, so the use of full scan MS is highly recommended when testing for canister cleanliness. If you are currently performing EtO analysis, or planning to do so in the future, then the use of humid air and full scan analysis will help you determine if you canisters are suitable for EtO.