Tubing material considerably affects measurement delays of gas-phase oxygenated per- and polyfluoroalkyl substances

ABSTRACT

Per- and polyfluoroalkyl substances (PFAS) are persistent environmental pollutants associated with negative health impacts. Assessments of tubing-related measurement bias for volatile PFAS are lacking, as gas-wall interactions with tubing can delay quantification of gas-phase analytes. We use online iodide chemical ionization mass spectrometry measurements to characterize tubing delays for three gas-phase oxygenated PFAS − 4:2 fluorotelomer alcohol (4:2 FTOH), perfluorobutanoic acid (PFBA), and hexafluoropropylene oxide dimer acid (HFPO-DA). Perfluoroalkoxy alkane and high-density polyethylene tubing yielded relatively short absorptive measurement delays, with no clear dependence on tubing temperature or sampled humidity. Sampling through stainless steel tubing led to prolonged measurement delays due to reversible adsorption of PFAS to the tubing surface, with strong dependence on tubing temperature and sample humidification. Silcosteel tubing afforded shorter measurement delays than stainless steel due to diminished surface adsorption of PFAS. Characterizing and mitigating these tubing delays is crucial for reliable quantification of airborne PFAS.

Implications: Per- and polyfluoroalkyl substances (PFAS) are persistent environmental contaminants. Many PFAS are sufficiently volatile to exist as airborne pollutants. Measurements and quantification of airborne PFAS can be biased from material-dependent gas-wall interactions with sampling inlet tubing. Thus, characterizing these gas-wall interactions are crucial for reliably investigating emissions, environmental transport, and fates of airborne PFAS.

Introduction

Per- and polyfluoroalkyl substances (PFAS) are a class of anthropogenic compounds used in numerous industrial applications and consumer products (Glüge et al. 2020; Kotthoff et al. 2015). The widespread use of PFAS stems from their high thermal and chemical stability. This relative inertness also lends to slow or negligible chemical degradation upon release to the environment. PFAS are therefore capable of building up in the environment as persistent pollutants (Kwok et al. 2013; Pickard et al. 2018). This stability also makes PFAS resistant to biological breakdown, leading to high bioaccumulation potential (De Silva et al. 2021; Lesmeister et al. 2021). Exposure to certain PFAS has been linked to various negative health impacts including cancer and liver disease (Espartero et al. 2022; Fenton et al. 2021; Gaballah et al. 2020; Sunderland et al. 2019). These health concerns have led to a recent increase in federal and state regulations and monitoring of environmental PFAS in the U.S. (U.S. Congress 2021; EPA 2021).

Many PFAS are sufficiently volatile under ambient conditions to exist as airborne gas-phase pollutants (D’Ambro et al. 2021; Riedel et al. 2019). Thus, the atmosphere serves as an important environmental transport and exposure route for these pollutants (Davis et al. 2007; D’Ambro et al. 2021; Galloway et al. 2020; Kirkwood et al. 2022). The abundance of volatile PFAS in household consumer products and building materials has also led to their detection in indoor air, thereby serving as an important exposure route in residential and occupational settings (Langer, Dreyer, and Ebinghaus 2010; Morales-McDevitt et al. 2021; Schlummer et al. 2013; Shoeib et al. 2005). Despite this, measurements of environmental PFAS have historically focused on condensed phase media, including water and soil. Previous efforts to measure gas-phase PFAS in ambient air and source emissions (i.e., emission stacks from industrial facilities) typically involve offline measurements, which are subject to low time resolution, and require external laboratory analysis of samples (Nakayama et al. 2019).

Recent studies have applied online chemical ionization mass spectrometry (CIMS) instrumentation toward real-time measurements of gas-phase PFAS (Riedel et al. 2019, 2021). When utilizing iodide (I⁻) reagent ions, CIMS is sensitive and selective toward fluorotelomer alcohols (FTOHs), and perfluorinated carboxylic acids (PFCAs) including perfluoro ether carboxylic acids such as hexafluoropropylene oxide dimer acid (HFPO-DA; C₆HF₁₁O₃) (Riedel et al. 2019, 2021). FTOHs are present in fluorosurfactants (Riedel et al. 2019), aqueous film-forming foams (AFFFs) (Riedel et al. 2019; Titaley et al. 2022), and various other consumer products (Buck et al. 2011; Gremmel, Frömel, and Knepper 2016; Liu et al. 2015). They may photochemically transform in the atmosphere to form PFCAs (Ellis et al. 2004). Long-chain PFCAs including perfluorooctanoic acid (PFOA; C₈HF₁₅O₂) have been utilized in manufacturing and other industrial applications, but their use has been phased out due to their adverse health impacts and environmental persistence (Wang et al. 2014). Consequently, short-chain PFCAs such as perfluorobutanoic acid (PFBA; C₄HF₇O₂) have recently seen increased use as substitutes for long-chain PFCAs (Wang et al. 2014; Chemical & Engineering News 2010). Another substitute for long-chain PFCAs is HFPO-DA, a component of GenX which has received attention as an emerging environmental pollutant (Chemical & Engineering News 2018). The EPA recently issued a drinking water health advisory for GenX chemicals, including HFPO-DA and its ammonium salt (EPA 2022).

Choice of tubing material used for sample inlet construction may substantially impact measurements of gas-phase analytes. Recent studies have demonstrated how sampled gas-phase organic compounds must establish gas-wall equilibrium with tubing, leading to delayed transmission and subsequent quantification (Deming et al. 2019; Liu et al. 2019; Pagonis et al. 2017). The nature of these gas-wall interactions, and resulting measurement delays strongly depend on the type of tubing material used (Deming et al. 2019; Pagonis et al. 2017). Organics sampled through polymeric tubing undergo absorptive partitioning into tubing walls (Deming et al. 2019; Pagonis et al. 2017). Resulting measurement delays are dependent on compound volatility, and independent of concentration and humidity (Deming et al. 2019; Pagonis et al. 2017). Sampling gas-phase organics through metallic or glass tubing result in adsorptive partitioning delays with a strong dependence on analyte concentration, functionality, and sample humidity (Deming et al. 2019; Pagonis et al. 2017). Additionally, tubing length and diameter, and sample flow rate can influence these measurement delays (Deming et al. 2019; Pagonis et al. 2017). The online, real-time mass spectrometry methods performed in these studies enabled researchers to probe the rapid (timescale ~minutes) gas-wall partitioning process driving these delays (Deming et al. 2019; Liu et al. 2019; Pagonis et al. 2017).

Analogous assessments of tubing-related measurement bias for gas-phase PFAS are lacking due to the paucity of real-time detection methods for these compounds. Characterizing these biases is essential for performing reliable measurements of gas-phase PFAS. To address this need, we use online CIMS measurements to characterize tubing-related measurement biases for 4:2 FTOH (C₆H₅F₉O), PFBA, and HFPO-DA—three environmentally relevant short-chain PFAS with distinct oxygenated functionalities (Figure S5). We demonstrate how choice of tubing materials, tubing temperature, and sample humidity influence gas-wall interactions between these oxygenated PFAS and the sampling tubing, leading to biased quantification of these analytes.

Experimental methods

Tubing delay experiments

We present a schematic of the experimental setup in Figure 1. We prepared permeation tubes of neat 4:2 FTOH (99% purity; Synquest Labs), PFBA (98% purity; Oakwood Chemical), and HFPO-DA (97% purity; Synquest Labs). We contained the 4:2 FTOH and PFBA permeation tubes in temperature-controlled oven (Dynacalibrator 190; VICI Metronics Inc.) held at 40°C, and the HFPO-DA permeation tube in a custom-built temperature-controlled oven (60°C). We delivered ~50 sccm dry ultra-high purity nitrogen (UHP N₂; Airgas) regulated via a critical orifice to each temperature-controlled oven to generate gas-phase standards of 4:2 FTOH, PFBA, and HFPO-DA.

Figure 1. Schematic of experimental setup.

We directed a mixture of these PFAS standards to a three-way perfluoroalkoxy alkane (PFA) solenoid valve (Cole Palmer). When the valve was configured to the “on” state, PFAS standards were mixed with a dilution flow of 3 sLPM ultra-zero grade air (UZA; Airgas) delivered via MFC mass-flow controller (MFC; Alicat Scientific). UZA dilution flow was either dry (<0.5% relative humidity; RH) or humidified (48. ± 1.% RH at 25°C) with a glass bubbler containing HPLC grade H₂O (Sigma Aldrich) (Figure 1). We monitored UZA dilution RH in-line using a Rotronic Hygropalm HP23AW. This diluted PFAS standard then flowed through 3 m temperature-controlled (30–120°C) tubing (0.476 cm ID, 0.635 cm OD; Clayborn Lab) comprised of either PFA, high density polyethylene (HDPE), stainless steel (type 316), or Silcosteel (Restek) into an online CIMS instrument to monitor sampled PFAS mixing ratios in real-time (Lee et al. 2014; Riedel et al. 2019). We provide CIMS measurement details in the following section. Diluted PFAS standard mixing ratios sampled by the CIMS were ~1 ppbv 4:2 FTOH, ~5 ppbv PFBA, and ~5 ppbv HFPO-DA. Sampled PFAS flow through the 3 m tubing was laminar (Reynolds number ≤800) and had a residence time of  ~ 1 s. Lengths of additional PFA tubing/connections (0.476 cm ID, 0.635 cm OD) in the sampling inlet construction were minimized (<30 cm) to mitigate additional measurements delays. When the solenoid valve was configured to the “off” state, PFAS standard flow was exhausted to a laboratory fume hood, and only dry or humidified UZA was sampled through the temperature-controlled tubing with the CIMS. We toggled this solenoid valve state to alternate between diluted PFAS standard sampling (i.e. tubing passivation), and UZA-only sampling (i.e. tubing depassivation) periods (Figure 1). We typically performed triplicate passivation and depassivation cycles lasting ≥2 hours for each experimental perturbation tested. We began experiments with instrumental zeros, where CIMS background signals were monitored while sampling UZA for ≥2 hours. Stainless steel tubing required extensive depassivation times (≥12 hours) before experiments to mitigate hysteresis. To assess potential effects of sampling, a PFAS standard mixture versus individual compounds on measurement delays, we repeated experiments involving PFA and stainless steel tubing materials using only 4:2 FTOH standard.

We calculated total measurement delay times (t_total) for PFAS standards as the sampling time required to achieve 95% of a measured equilibrium (i.e., steady-state) CIMS signal (S_eq) during tubing passivation periods. For PFA, HDPE, and Silcosteel tubing, we quantified t_total and S_eq for sampled PFAS standards using double exponential fits to CIMS signal time series data in Igor Pro 7 (WaveMetrics, Inc.). Previous studies involving real-time measurements of gas-phase analytes have reported success in using double exponential functions to parameterize measurement delays attributable to inlet and/or instrument surfaces (Liu et al. 2019; Sauer et al. 2021; Veres et al. 2008). We discuss t_total calculations for stainless steel tubing experiments in the Supplemental Information (SI). The interior surface of the CIMS ionization source is comprised of electropolished stainless steel, and therefore may act as a sorptive surface for sampled analytes (Krechmer et al. 2016; Liu et al. 2019). To decouple measurement delays attributable to the 3 m temperature-controlled tubing (t_tubing) from those caused by the instrument (t_instrument), we repeated PFAS standard sampling experiments under dry and humidified conditions while replacing the 3 m tubing with a relatively short length (~4 cm) of polytetrafluoroethylene (PTFE) tubing (0.476 cm ID, 0.635 cm OD). From this, we calculated t_tubing as the difference between t_total and t_instrument (Deming et al. 2019; Liu et al. 2019; Pagonis et al. 2017). We expect sorptive delays from the ~4 cm PTFE tubing and PFA solenoid valve (0.476 cm ID, ~4 cm additional sampling length) were insubstantial compared to 3 m tubing lengths.

Chemical ionization mass spectrometry measurements

We performed fast (1 Hz), online measurements of gaseous oxygenated PFAS using a long time-of-flight chemical ionization mass spectrometer (CIMS; Tofwerk AG and Aerodyne Research Inc.) utilizing iodide (I⁻) reagent ions (Lee et al. 2014). We generated I⁻ reagent ions by flowing ~2 sLPM ultra-high purity nitrogen (UHP N₂; Airgas) through a temperature-controlled oven (40°C) containing a custom-built iodomethane (CH₃I) permeation tube. The output gas mixture from the oven (containing N₂ + CH₃I) then flowed through a ²¹⁰Po ionizer (NRD) to yield I⁻ reagent ions, which subsequently entered the ion molecule reactor (IMR) of the CIMS. We sampled analytes into the IMR at a flow rate of ~2.1 sLPM, regulated via critical orifice. Analytes of interest (M) reacted with I⁻ reagent ions in the IMR to form iodide-analyte adducts ([I+M]⁻), which were subsequently detected in real-time via time-of-flight mass spectrometry. We calibrated CIMS instrumental responses toward 4:2 FTOH, PFBA, and HFPO-DA following Riedel et al. (2019). From these calibrations, we derived instrumental sensitivities of (1.0 ± 0.2) × 10⁴, (1.0 ± 0.2) × 10⁴, and (5. ± 1.) × 10⁴ Hz ppbv⁻¹ for 4:2 FTOH, PFBA, and HFPO-DA, respectively. We acquired CIMS mass spectra at 1 Hz using Acquility software (v2.3.13; Tofwerk AG) running in Igor Pro 7 (WaveMetrics Inc.). We acquired CIMS mass spectra between 30 and 1000 mass-to-charge (m/z) units. Mass spectral peak resolution (m/Δm) was ≥12000, and mass accuracy was typically <2 ppm.

We processed CIMS mass spectral data using Tofware (version 3.2.3; Tofwerk AG) running in Igor Pro 7. Briefly, this software performs mass calibration, mass spectral baseline calculation, and high-resolution peak fitting algorithms to extract CIMS signal time series from individual mass spectral peaks. We typically included I(H₂O)⁻, I(CH₂O₂), I(HNO₃), I₂⁻, I₃⁻, I(C₄F₇O₂H)⁻ and I(C₆F₁₁O₃H)⁻ as CIMS mass calibration peaks, as they were typically fully resolved, and persistent in instrumental background measurements. We performed CIMS mass calibrations using a three-parameter (p_x) fit (p₁*(m/z)^p₃+p₂) to convert ion time-of-flight (ns) to m/z. We performed all further analyses of CIMS data in Igor Pro 7, including CIMS data normalization and background signal subtraction. We ensured changes in CIMS signal intensity were not driven by changes in reagent ion signal by normalizing CIMS time series data to total reagent ion signal. Here, we multiplied CIMS signals by the ratio of the averaged total reagent ion (I⁻ + I(H₂O)⁻) signal during instrumental background measurements to the real-time total reagent ion signal during periods of analyte measurement (Bertram et al. 2011).

Results and discussion

We demonstrate measurement delays associated with stepwise changes in sampled PFAS concentrations in Figure 2. Sampling through PFA and HDPE tubing generally resulted in the shortest tubing delays (t_tubing) for all compounds, while stainless steel produced the longest delays (Figure 3a and S1a). Tubing delays were typically lowest for 4:2 FTOH, and highest for HFPO-DA (Figure 3a). We observed an analogous trend for instrument delays (t_instrument) (Figure S2). Instrument delays often contributed substantially to total measurement delays (t_total), accounting ≤60% of t_total for 4:2 FTOH, ≤87% of t_total for PFBA, and ≤76% of t_total for HFPO-DA. We briefly discuss tubing depassivation timescales in the SI, but focus our discussion of experimental results on comparing tubing passivation timescales.

Figure 2. CIMS signal time series for 4:2 FTOH (orange traces), PFBA (blue traces), and HFPO-DA (green traces) during (a) PFA tubing passivation sampling, (b) PFA tubing depassivation sampling, and stainless steel tubing passivation experiments under (c) dry conditions, and (d) humidified conditions. CIMS signals are normalized to signal magnitudes achieved at gas-wall equilibrium (i.e., S_eq). Sinusoidal oscillations in signal observed in panel (d) are due to the heating cycle of the temperature-controlled tubing (see also Figure S4c).

Figure 3. Tubing delay times (t_tubing) for 4:2 FTOH (orange markers), PFBA (blue markers), and HFPO-DA (green markers) calculated for 3 m of various tubing materials, tubing temperatures, and humidity conditions while sampling (a) the three-compound mixture (i.e., 4:2 FTOH + PFA + HFPO-DA), and (b) sampling 4:2 FTOH only. Error bars correspond to ± one standard deviation of calculated t_tubing. Horizontal dashed lines on plots visually indicate break in vertical axes. We provide a base-10 logarithm scaling of the vertical axis in Figure S1.

We report relatively short tubing delays for PFA and HDPE across all tubing temperature and humidification conditions tested, with t_tubing ranging from 2 to 7 s for 4:2 FTOH, 4–18 s for PFBA, and 100–500 s for HFPO-DA (Figure 3a). This suggests the different chemical makeup of these tubing materials (fluoropolymer versus polyethylene polymer) did not substantially influence gas-wall equilibrium of the sampled PFAS mixture. These t_tubing values increased with decreasing compound volatility (Table S1) (EPA 2018; Bhhatarai and Gramatica 2011; Krusic et al. 2005), and did not exhibit a clear dependence on sampled humidity, consistent with absorptive partitioning behavior expected between polymeric tubing material and gas-phase analytes (Deming et al. 2019; Pagonis et al. 2017). Interestingly, we did not observe a clear trend in t_tubing as a function of tubing temperature for PFA. Absorptive delay times of organic molecules are diminished with increasing compound volatility, and augmented with increasing mass of polymeric tubing available for absorptive partitioning to take place (Deming et al. 2019; Pagonis et al. 2017). While increasing tubing temperature from 30 to 120°C would increase volatility of sampled PFAS, we hypothesize this was accompanied by an increase in absorptive mass available due to thermally-driven changes to the physical properties of the polymeric tubing, resulting in a “canceling out” of these two temperature effects on absorptive delay times (Deming et al. 2019; Pagonis et al. 2017). Sampling only 4:2 FTOH through PFA tubing yielded similarly low t_tubing (<10 s) compared to the PFAS mixture, with no clear dependence on sampled humidity or tubing temperature (Figure 3b and S1b). We observed transient spikes in 4:2 FTOH and PFBA signals above their respective S_eq following initial passivation of PFA tubing. This is unlikely due to competitive displacement of 4:2 FTOH and PFBA by HFPO-DA, as transient spikes above S_eq also occur while sampling 4:2 FTOH only through PFA tubing under dry and humidified conditions (Figure S6). Rather, a sampling artifact, e.g., transient pressurization of sampling lines upon switching of the PFA solenoid valve state may explain this phenomenon.

We compare experimental t_tubing for PFBA and HFPO-DA through PFA tubing with values derived from the absorptive tubing delay model of Pagonis et al. (2017). This model relates absorptive gas-wall partitioning to compound saturation vapor concentration (C*), or the ambient concentration at which a compound exists equally in gaseous and condensed phases (e.g., absorbed into walls of polymer tubing) (Pagonis et al. 2017). We estimated C* values of 7 × 10⁷ and 4 × 10⁶ µg m⁻³ for PFBA and HFPO-DA, respectively, as model inputs. We report additional model details in the SI. Modeled PFBA and HFPO-DA absorptive delays through PFA tubing were considerably underestimated compared to experimental values (Figure S3). Reproducing experimental PFBA and HFPO-DA tubing delays in the model required lowering C* by approximately two orders of magnitude. We speculate this discrepancy was influenced by uncertainties in the structure-activity relationship method for estimating C* for PFAS (D’Ambro et al. 2021). Additionally, it is unclear to what extent aqueous solubility of these PFAS influenced experimental tubing delays, which the model does not explicitly account for. For instance, Liu et al. (2019). demonstrated that tubing delays of certain small polar molecules are driven by their solubility into thin films of water present on Teflon tubing surfaces, rather than by C*. However, a detailed mechanistic elucidation of these absorptive tubing delays lies beyond the scope of this work.

Sampling the PFAS mixture through stainless steel tubing resulted in prolonged measurement delays compared to PFA and HDPE. We visually compare time series from these experiments, and do not explicitly calculate t_tubing for sampled compounds (Figure 2c,d). Sampling under dry conditions resulted in volatility-driven competitive adsorptive behavior, consistent with previous observations of gas-phase organics sampled through stainless steel tubing (Figure 2c) (Deming et al. 2019; Pagonis et al. 2017). Here, we observe a transient increase in measured 4:2 FTOH above its gas-wall equilibrium level signal (i.e., S_eq) as it is competitively displaced by PFBA, followed by a slight displacement of PFBA by HFPO-DA. Sampling times of >3 × 10⁴ s would be necessary to mitigate tubing-related measurement biases of these compounds under these experimental conditions. Further, we observed negligible transmission (<1% of S_eq) of sampled PFAS during the first 5 × 10³ s of sampling. Sample humidification considerably decreased these measurement delays through stainless steel, due to the competitive adsorption of water molecules on active surface sites (Figure 2d) (Deming et al. 2019). Here, we observed a relatively rapid equilibration in 4:2 FTOH signal (<60 s), followed by a slower rise in PFBA and HFPO-DA signals. While notably faster than dry conditions, the humidified PFAS mixture still required >2.5 hours to fully establish gas-wall equilibrium with stainless steel tubing.

We quantitatively evaluate the effects of sample humidity and tubing temperature on stainless steel tubing delays while sampling 4:2 FTOH only (Figure 3b and S1b). We calculated t_tubing of 5.3 × 10⁴ s for 4:2 FTOH sampled through stainless steel tubing at 120°C under dry conditions. We observed <1% transmission of 4:2 FTOH in the first 1.2 × 10⁴ s of sampling (Figure S4). We did not perform a corresponding dry experiment at a 30°C tubing temperature, but expect a lower tubing temperature would further augment t_tubing due to a reduction in 4:2 FTOH volatility. Humidifying the sample greatly diminished this tubing delay, resulting in t_tubing of 100 s. Interestingly, we observe an increased t_tubing when raising tubing temperature from 30°C to 120°C under humidified conditions. We suspect this is due to the increased vapor pressure of water at higher tubing temperatures. The enhanced volatility of water would decrease its propensity to interact with the stainless steel surface, thereby reducing the strong surface-displacing effect of humidity on sampled PFAS.

We demonstrate the reversible nature of adsorptive partitioning of sampled PFAS to stainless steel in Figure 4. Here, we passivated the stainless steel tubing with 4:2 FTOH under dry conditions for varying periods of time (30–300 s), which coincided with insubstantial transmission (and therefore detection via CIMS) of 4:2 FTOH (Figure 4a,b). Following passivation periods, we depassivated the stainless steel tubing with humidified UZA (Figure 4a,b). Sampled water molecules competitively displaced the adsorbed 4:2 FTOH, leading to a spike and subsequent decay in measured 4:2 FTOH. The integrated 4:2 FTOH signal (i.e., total desorbed 4:2 FTOH) during depassivation directly correlated with passivation time (i.e., total sampled 4:2 FTOH) (Figure 4c). This indicates effectively all sampled 4:2 FTOH adsorbed to stainless steel tubing during these short passivation times, preceding competitive desorption by another species.

Figure 4. (a-b) CIMS signal time series for 4:2 FTOH (black traces) during systematic adsorption-desorption experiments with stainless steel tubing. Shaded blue periods correspond to 4:2 FTOH passivation under dry conditions, and shaded green periods correspond to 4:2 FTOH depassivation under humidified conditions as described in the main text. (c) Scatter plot demonstrating linear relationship between total (e.g., integrated) 4:2 FTOH CIMS signal, and passivation time. Red line corresponds to linear fit to data.

Tubing delays through Silcosteel were considerably shorter than stainless steel, but generally longer than PFA or HDPE (Figure 3a). The silicon-based coating on the tubing pre-passivates the stainless steel surface, presumably reducing adsorption and therefore measurement delays of sampled PFAS. Unlike stainless steel, the dependence of gas-wall equilibrium on sampled humidity is unclear for Silcosteel. We speculate the coated surface of this tubing also inhibits the adsorption of water molecules, thereby mitigating the humidity-dependent measurement delays observed for stainless steel tubing. However, temperature strongly impacted this gas-wall equilibrium for Silcosteel. We observed substantially lower t_tubing for all sampled PFAS at a tubing temperature of 120°C compared to 30°C during both dry and humidified experiments, supportive of volatility-dependent gas-wall interactions of sampled PFAS with the Silcosteel tubing.

Conclusion

Our results highlight the importance of selecting appropriate inlet tubing materials when sampling airborne PFAS. Adsorption to stainless steel tubing led to prolonged measurement delays/losses of sampled FTOHs and PFCAs. These adsorptive delays were substantially impacted by tubing temperature and sampled humidity. Polymeric tubing materials such as PFA or HDPE are potentially more suitable for these measurements. Silcosteel tubing, and potentially other “pre-passivated” adsorptive tubing materials may be a suitable alternative to stainless steel if the use of polymeric tubing for PFAS sampling is untenable. For instance, PFA and other fluoropolymer tubing materials may outgas fluorinated organic molecules (Brophy and Farmer 2015; Veres et al. 2010), which could potentially interfere with gas-phase PFAS measurements.

This work serves as the first in-depth assessment of tubing material-dependent measurement biases of airborne PFAS. Given the selective scope of this work, future studies could explore effects of additional tubing materials/dimensions, additional PFAS functional groups, and PFAS concentrations on these measurement delays. Further, more realistic atmospheric matrices containing variable humidity and other traces gases which may influence gas-wall partitioning of sampled PFAS should be investigated. Mitigating tubing-related measurement bias is crucial for reliably investigating emissions, environmental transport, and fates of airborne PFAS.

Acknowledgment

The U.S. EPA through its Office of Research and Development supported the research described here. It has been subjected to Agency administrative review and approved for publication but may not necessarily reflect official Agency policy. Any mention of trade names, manufacturers or products does not imply an endorsement by the United States Government or the U.S. Environmental Protection Agency. EPA and its employees do not endorse any commercial products, services, or enterprises. J.M.M. was supported by the Oak Ridge Institute for Science and Education (ORISE) Research Participation Program for the U.S. Environmental Protection Agency (EPA).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that support the findings of this study are openly available in ScienceHub at http://doi.org/10.23719/1527927, reference number D-0p2r.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/10962247.2023.2174612.

Additional information

Notes on contributors

James M. Mattila

James M. Mattila is an Oak Ridge Institute for Science and Education (ORISE) postdoctoral fellow at the U.S. Environmental Protection Agency Office of Research and Development. He received his PhD in Chemistry form Colorado State University in 2021. His research focuses on studying airborne PFAS using online mass spectrometry instrumentation.

Emily Y. Li

Emily Y. Li is a Research Chemist with the U.S. Environmental Protection Agency Office of Research and Development. Her current research focuses on developing methods for characterization and mitigation solutions for air pollutant sources to assess and control emerging air emissions.

John H. Offenberg

John H. Offenberg is a Research Chemist at the U.S. Environmental Protection Agency Office of Research and Development. His research focuses on photochemical transformations in the atmosphere.