PFAS in Air, Part 5: Resin Extraction

In previous blog posts, I’ve talked about the draft OTM-45 method; the chromatography we were able to achieve; the calibration requirements and results; and the resin cleaning method. In this post I’ll cover the Accelerated Solvent Extraction (ASE) method I used as an alternative to the shake-out extraction covered in OTM-45, which provides significant time savings and extraction improvements.

Air sampling methods that use sorbents such as polyurethane foam (PUF) or XAD-2 type styrene-divinylbenzene (SDVB) resins often use Soxhlet extractions for sample preparation. A Soxhlet extraction involves continuous refluxing of hot solvent through solids in order to extract the target compounds from it. This is often done over a long period of time. TO-13A for example calls for an 18-hour extraction. OTM-45 opts for a simpler approach of simply shaking out the sample in the extraction solvent. While this allows for a simpler setup with no specialized glassware or heating, it is a less efficient extraction process. To compensate for this the shakeout in OTM-45 is even longer than a Soxhlet extraction, done in two rounds of 16-hour shakeouts for a total of 32 hours. Each shakeout is done with 180 mL of 5% ammonium hydroxide in methanol, for a total of 360 mL of solvent use. For the first XAD trap (container #3 in Figure 1) the solvent used is the back-half rinse (container #4 in Figure 1), which rinses the sampling train between the particle filter and XAD trap. If the back-half rinse is less than 360 mL, then additional methanolic ammonium hydroxide is added. The breakthrough XAD trap (container #7 in Figure 1) is still extracted in methanolic ammonium hydroxide, but none of that has been used to rinse the sample train. This means that the total sample train extraction uses at least 360 mL of non-rinse solvent for the resin extractions, and 720 mL of total solvent.

Figure 1 – OTM-45 sample train and container identification.

In contrast to the time and solvent intensive shakeout, the use of ASE has a number of advantages. ASE uses heated solvents at high pressures to give a very quick and efficient extraction. The ASE method I used could extract a sample in 45 minutes using less than 100 mL of solvent, a large improvement over the shakeout method. The ASE can’t use the back-half rinse as a solvent though, since it uses central solvent reservoirs for all samples, so the rinse would have to be combined with the ASE extraction solvent during blowdown. Still, that is 200 mL of non-rinse solvent used over both resin fractions (i.e., containers #3 and 7) vs. 360 mL of non-rinse solvent from the shakeout, or 560 mL of total solvent for the ASE (360 mL rinse solvent plus 200 mL ASE solvents) vs. 720 mL total for the shake extractions. Table 1 shows the ASE parameters used. A 4:1 mixture of methanol:ACN was used since pure methanol gave poor recoveries across the board (results not shown), and too much ACN lead to poor recoveries (results not shown) for a handful of compounds.

Dionex ASE 350 Extraction Parameters

• Pressure — 1500 psi
• Temperature — 120°C
• Heating time — 6 minutes
• Statis time — 15 minutes
• Cycles — 2
• Rinse volume — 60%
• Solvent — 4:1 Methanol:Acetonitrile

Table 1 – ASE extraction parameters

Extraction time and solvent use improvements don’t mean anything if the extraction itself isn’t equivalent or better though, so how does the ASE compare performance-wise to the shakeout? The most dramatic difference is in the raw area counts, as shown in Figure 2. The ratio of the ASE:Shakeout areas range from 142% up to 5495%, with most being between 200-500%. This shows that the ASE seems to extract more PFAS from the resin matrix than the shakeout.

Figure 2 – Ratio of ASE:Shake extraction areas.

Due to the isotope dilution method used, the shakeout method still had recoveries in a similar range as the ASE method, although the ASE outperformed the shakeout overall. Figure 3 shows the results of the two extractions. The outliers with very high or low recoveries tended to be compounds that I didn’t have isotope standards for, so the poor accuracy is likely due to a mismatch in the extraction efficiencies of the target compounds and non-identical internal standard compounds. However, the better ASE extraction efficiency means that compounds that recovered poorly in the shakeout such as 3:3 FTCA and the Me and Et-FOSAs had much better recoveries using ASE.

Figure 3 – Spike recoveries for ASE and shake extractions

The ASE extraction had better precision overall than the shakeout as well. 40 of 49 compounds met the Many of the failures were compounds that lacked exact isotope dilution standards for, so it’s likely that results would be improved with an improved selection of isotopes. The % RSDs for the later eluting compounds tended to be much higher on the shakeout than with the ASE, as shown in Figure 4. This may indicate that the shake extraction has less consistency with longer chain PFAS compounds.

Figure 4 - %RSD for ASE and shake extractions

This all shows that the ASE extraction is a good alternative to the shakeout method outlined in OTM-45 being faster, using less solvent, and giving better results overall.