Treatment of Per- and Polyfluoroalkyl Substances (PFAS)-Contaminated Hypersaline Brine by Membrane Distillation

Abstract

This research evaluates membrane distillation (MD) to treat problematic hypersaline per- and polyfluoroalkyl substances (PFAS)-laden waste streams. Results are shown for four commercially available membranes (unlaminated polytetrafluoroethylene (PTFE), polypropylene laminated PTFE, polyether ether ketone (PEEK), and polyvinylidene difluoride (PVDF)) that were tested with a model short-chain PFAS compound, perfluoropentanoic acid (PFPeA) at a concentration of 10 mg/L in the presence and absence of ion-exchange resin spent brine (10% NaCl). For each test, a new membrane with an area of 140 cm² was used, with a constant permeate temperature of 20 °C (cold) and varying feed temperatures of 50 °C, 60 °C, or 70 °C (hot). The unlaminated PTFE membrane demonstrated the best performance in treating the PFPeA-contaminated brine. The water flux through the unlaminated PTFE membrane was 50% higher than the flux through the PEEK membrane and 25% higher than that through the PVDF and laminated PTFE membranes. The laminated and unlaminated PTFE membranes achieved the highest rejection of NaCl and PFPeA (>99.7%) compared to 95 and 97% obtained by the PEEK and the PVDF membranes, respectively. During the 48-h extended experiments, the laminated PTFE membrane exhibited greater stability and mechanical strength than the other membranes, while the PEEK and PVDF membranes proved fragile. The laminated PTFE membrane was then selected for a 300-h experiment with ethanol cleaning cycles to test long-term durability. PFPeA caused reversible fouling in all tested membranes and reduced the membrane’s hydrophobicity; however, ethanol cleaning was effective in removing PFPeA, indicating that with further optimization, membrane distillation may be useful for concentrating PFAS for ultimate destruction or disposal.

1. Introduction

Per- and polyfluoroalkyl substances (PFAS) are synthetic compounds comprising thousands of chemicals characterized by carbon chains with fluorine atoms attached. The widespread use of PFAS in industrial manufacturing has resulted in their release into the environment and their presence at relatively high concentrations in drinking water, surface water, and groundwater. PFAS compounds have been detected in drinking water sources across hundreds of locations in the US.

Current scientific research suggests that exposure to certain PFAS may lead to adverse health outcomes. However, research is ongoing to determine how exposure to different PFAS can lead to various health effects, including carcinogenic and immunotoxic effects on humans. , U.S. Environmental Protection Agency (USEPA) promulgated National Primary Drinking Water Regulations for perfluorooctanoic acid (PFOA) and perfluorooctane sulfonic acid (PFOS). The regulation established maximum contaminant levels (MCLs) of 4 parts-per-trillion (ppt) each for PFOA and PFOS. Many water utilities are actively contemplating PFAS treatment strategies.

Due to the strong carbon–fluorine bond, PFAS compounds are highly resistant to thermal, biological, and chemical degradation and are poorly removed by conventional water treatment methods. After proper pretreatment, high-pressure membrane processes (i.e., reverse osmosis [RO] and nanofiltration [NF]), as well as granular activated carbon (GAC) and anion exchange resin (AXR), can be effective in removing PFAS. However, these methods generate waste streams with higher PFAS concentrations in liquid or solid states. Of the adsorptive technologies, AXR generally has much higher capacities for PFAS and smaller physical footprints than GAC; however, the current practice of discarding the spent AXR makes the optimal treatment choice dependent on cost, site-specific, and water quality considerations. The practice of using single-use AXR, followed by incineration of the spent resin, makes the AXR prohibitively expensive and environmentally unfriendly. Based on a life-cycle analysis, resin manufacturing is identified as one of the most critical factors contributing to the negative environmental impact of the ion exchange process. The costs of disposing of ion-exchange resin and the energy required for incineration are expected to exceed the costs of regeneration. A typical ion exchange system costs between $0.30 and $0.80 per 1000 gallons of treated water. Depending on the material, ion exchange resins may cost as much as ≥$1000 per ft³. The cost for offsite regeneration services ranges from $30 to $100 per ft³, depending on resin capacity, feedwater properties, flow rate, and regeneration frequency. Therefore, it would be desirable from a cost and environmental perspective if AXR regeneration could be acceptable to drinking water utilities and their state permitting agencies. However, PFAS-laden resin regeneration is typically conducted using an organic cosolvent and a brine solution (e.g., 20% methanol in 5–10 wt % NaCl brine). As a result, the AXR spent brine typically contains high salt concentrations in addition to the cosolvent and released PFAS, making it a problematic stream to treat before discharging.

Another potentially problematic PFAS treatment waste stream is the retentate from high-pressure membrane processes. In RO and NF membrane separation processes, 15–50% of the feedwater is retained as high-strength concentrated waste. A typical RO retentate from brackish water (i.e., with a total dissolved solids (TDS) concentration of 1000 to 15,000 mg/L) has a TDS concentration of 30 to 50 g/L. These highly concentrated waste streams represent a considerable cost and environmental challenge.

Ideally, the large volumes of AXR spent brine and concentrate produced by high-pressure membrane processes could be further treated to generate smaller volumes of highly concentrated PFAS solutions (see TOC Graphic). However, the high-TDS concentrations in these waste streams (>100,000 g/L) pose a significant challenge for post-treatment processes. Commercially available evaporative/crystallizers, including force circulation, draft tube baffle, and Oslo crystallizers, are the most suitable technologies for treating high-TDS brines but are limited by their high thermal energy requirements. The thermal energy demand for these systems is estimated at 246 kWh per m³ of recovered water, projected to be about 2 orders of magnitude higher than the energy required for RO seawater desalination (i.e., ∼2 to 4 kWh per m³). In comparison, membrane distillation (MD) is an energy-efficient technology for the desalination of highly saline brines, − and for industrial wastewater treatment. ,

Membrane distillation is a thermally driven separation process governed by the difference in vapor pressure across a hydrophobic membrane. The mass transfer of water vapor through the membrane is primarily driven by the vapor pressure difference resulting from the temperature gradient across the membrane. The MD process can operate at low-temperature differentials and utilize low-value heat sources (e.g., solar or waste heat from flue gases with temperatures below 100 °C), without needing a vacuum to reduce the boiling point of the treated brines. Therefore, the concentration of the treated stream has minimal impact on the MD process. Due to these potential advantages, this study evaluates membrane distillation as a sustainable post-treatment step for handling AXR and high-pressure membrane filtration waste streams in an energy-efficient integrated treatment manner. A detailed screening of commercially available membranes and a characterization of PFAS interactions with membrane surfaces are used to assess the fouling behavior of these PFAS-laden waste streams. To ensure practical applicability, membrane cleaning and flux restoration were also investigated.

2. Materials and Methods

2.1. Materials

Super-Q water with resistivity greater than 18.3 MΩ-centimeter (MΩ-cm) was used to prepare draw solutions and as the feed for the initial testing of all membranes. A model short-chain PFAS compound, perfluoropentanoic acid (PFPeA) (CAS Number 2706-90-3), and sodium chloride (NaCl) were obtained from Sigma-Aldrich (Milwaukee, WI). PFPeA has a molecular formula of C₅HF₉O₂ and was tested at a concentration of 10 mg/L in the presence and absence of 10% NaCl, using four commercially available membranes (Table ) obtained from Sterlitech Corporation (Auburn, WA).

1. Characteristics of Microfiltration Membranes Employed in MD Tests as Provided by the Vendor.

material	membrane ID	pore size (μm)	pH	water entry pressure (psi)
polytetrafluoroethylene	unlaminated PTFE	0.2	no limit	37
polytetrafluoroethylene	laminated PTFE	0.1	1–14	60
polyvinylidene	PVDF100	0.1	1–12	43
polyether ether ketone	PEEK100	0.1	1–14	29

2.2. Experimental Design

All MD experiments were conducted in replicates to ensure the reproducibility of the results. Baseline experiments were performed with 10 mg/L PFPeA and no added salt. These experiments assessed membrane performance in terms of permeate flux and rejection of PFPeA. Additional experiments were conducted with 10 mg/L PFPeA in the presence of 10% (w/v) NaCl to simulate an actual anion exchange waste stream. All feed solutions contained premeasured concentrations of PFAS, with and without salt. Temperature, conductivity, and pH were continuously monitored and recorded. The feed temperatures were 50 °C, 60 °C, or 70 °C ± 2 °C, and the permeate temperature was kept constant at 20 °C.

Experiments were conducted over extended periods to examine the interaction of PFAS with the membranes. The change in permeate mass was continuously data-logged using a Mettler Toledo MS8001TS/00 Laboratory Balance and WinWedge Data Acquisition Software. At the end of each testing cycle, used membranes were dried overnight in a clean fume hood and stored at 4 ± 1 °C to prevent membrane degradation due to bacterial or mold growth. These membranes were properly labeled with membrane ID, start and end dates, contaminants, and the volume of water vapor passed through the membrane. Samples of used and unused membranes were analyzed for hydrophobicity by obtaining their contact angle.

2.3. Experimental Setup

A schematic of the bench-scale MD experimental apparatus is shown in Figure . The two-compartment membrane holder is clear acrylic with active area dimensions of 9.7 × 14.7 cm (3.81 × 5.78 in). The top and bottom compartments were separated by the tested membrane, highlighted in black. Each membrane had an active cross-sectional area of 140 cm² (22 in²) and was sealed in the membrane holder using polypropylene PFAS-free O-rings. The feed and permeate solutions were circulated on the respective sides of the membrane holder at a fixed crossflow rate of 0.75 L/min using two-gear pump systems.

1.

Conceptual schematic diagram of the MD setup.

To mitigate PFAS adsorption onto the MD components, chemical-resistant pump heads with stainless-steel and polyether ketone wetted parts were used for feed and permeate transfer. Additionally, only silicon or high-density polyethylene tubes were used to transfer feed and permeate solutions from the stainless-steel tanks to the membrane holder. Stainless steel beakers were used as feed and permeate tanks. The permeate and feed follow the pathways outlined in Figure . The permeate beaker was placed on a balance, and the change in mass was continuously data-logged. A polyethylene tube connected the permeate beaker to the pump. This allowed the permeate to be pumped from the beaker through the cooler, then through one side of the membrane cell, and finally back into the permeate beaker. The feed beaker was placed on a hot plate and was heated to the desired temperature. A polyethylene tube connected the beaker to the pump and allowed the feed solution to be pumped through the respective side of the membrane cell. The feed was then pumped back into the beaker. During this time, the feed solution interacted with the membrane surface, and water vapor passed through the membrane and condensed on the permeate side of the membrane cell. The water flux was driven by the vapor pressure difference generated by the temperature gradient across the membrane and was continuously measured throughout the experiments. Polypropylene membrane spacers were used in the feed and the permeate solutions channels to promote mixing and prevent concentration and heat polarization. Due to the significant difference in concentration between the background salts (g/L level) and PFAS (μg/L level), salt crystals were expected to form and were removed before PFAS began to aggregate. The PFAS concentrations in samples collected from MD experiments, conducted with and without salts, were measured using a liquid chromatography-tandem mass spectrometry (LC/MS/MS) per the US EPA 1633 method (EPA Method, 2024).

Sterlitech’s Sepa Cell, with an active cross-sectional membrane area of 140 cm² (22 in²), was used for all experiments involving laminated and unlaminated PTFE membranes. In contrast, Sterlitech’s CF042 membrane test cell, with an active membrane area of 42 cm² (6.5 in²), was used for the PEEK and PVDF membranes to overcome membrane fragility. This decision was made after several failed attempts to run the experiment using the 140 cm² cell, as the PEEK and PVDF membranes were very fragile and broke within the first few minutes, causing the feed and permeate to mix. Once the smaller cell size was used, membrane tearing was no longer an issue with the PEEK or PVDF membranes.

3. Results and Discussion

3.1. Baseline Water Flux

The performance of the laminated, unlaminated PTFE, PVDF, and PEEK membranes using PFPeA and NaCl solutions were evaluated over a two-to-three-hour period. The experiments were terminated when approximately 65–70% recovery was achieved (i.e., 65–70% of the feed solution was transferred through the membrane) when the Sepa testing cell was used. The small CF042 membrane test cell experiments were terminated after 3 h. During this time, the feed PFPeA concentrations increased from 10 mg/L to 200 mg/L.

The membrane characteristics and feed temperatures significantly impact the performance of the MD process. Three main operational parameters were tested: feed temperature (50, 60, and 70 °C) and feed solution concentration (10 mg/L PFPeA with and without 10% NaCl). In contrast, the temperature on the permeate side was held constant at 20 °C.

Figure illustrates the permeate flux behavior of the different MD membranes when challenged with 10 mg/L PFPeA at various feed temperatures. Permeate flux increased as the feed temperature increased. The average water flux through the laminated PTFE increased from ∼14 to ∼25 LMH when the feed temperature increased from 50 to 70 °C due to the vapor pressure difference across the membrane. The same trend was observed with unlaminated PTFE, where the average permeate flux increased from ∼19 to ∼37 LMH when the feed temperature increased from 50 to 70 °C. The unlaminated membrane showed higher water flux than the laminated PTFE, which can be attributed to the larger average pore size compared to the laminated PTFE. The permeate flux behavior of the PEEK membrane was also consistent.

2.

Performance of the MD membranes at different feed temperatures (50 °C (a), 60 °C (b), and 70 °C (c)). The permeate inlet temperature was 20 ± 0.5 °C, and the feed contained 10 mg/L PFPeA with no added salt. Both feed and permeate were circulated at a constant flow rate of 0.75 L/min.

The dependence of permeate flux on temperature was expected since the vapor pressure, the driving force in this process, is a function of the solution temperature. The partial vapor pressure (P _m,f or P _m,p ) of water is exponentially related to the temperature, as described by Antoine equation (eqs and ()). T _m,f and T _m,p refer to the membrane temperatures on the feed and permeate sides, respectively. Temperature units are in degrees Celsius, and pressure units are in millimeters of Mercury.

$$P_{m,f}^{°}={\exp }^{23.19643816.44/(T_{m,f}-46.13)}$$ 1

$$P_{m,p}^{°}={\exp }^{23.19643816.44/(T_{m,p}-46.13)}$$ 2

Furthermore, increasing the temperature raises the kinetic energy of the water molecules, enabling them to overcome intermolecular forces in the liquid phase and evaporate at higher rates. This kinetic energy is directly proportional to the temperature of the molecules, leading to more rapid evaporation at higher temperatures. It is also important to note that PFAS compounds have very low vapor pressures.

3.2. Water Flux from Hypersaline Brines

Figure demonstrates the effect of feed concentration on permeate flux at different temperatures. The water flux presented in these figures was obtained from MD experiments conducted with 10 mg/L PFPeA in the presence of 10% NaCl, while the feed and permeate temperatures were kept constant at 50 or 60 °C. The average water flux through the laminated PTFE membrane decreased from ∼22 to ∼15 LMH (a 22% reduction) at 60 °C, compared to a ∼32% reduction in flux observed at 50 °C. The same trend was observed with the unlaminated PTFE and PEEK membranes. The average water flux through the unlaminated PTFE membrane was reduced by ∼25% at 60 °C and 20% at 50 °C. In contrast, the average water flux through the PEEK membrane was reduced by ∼31% at 60 °C and 27% at 50 °C.

3.

Performance of MD membranes for 10 mg/L PFPeA concentration at permeate inlet temperatures of 60 or 50 ± 0.5 °C in the absence ((a) 60 °C and (c) 50 °C) and presence ((b) 60 °C and (d) 50 °C) of 10% NaCl; both feed and permeate were circulated at a constant flow rate of 0.75 L/min.

This reduction in water flux can be attributed to the effect of concentration polarization, where the accumulation of salt molecules on the membrane surface obstructs vapor movement, resulting in additional resistance to mass transfer. Additionally, the higher salt concentration increases the boiling point of the feed solution because NaCl molecules form hydrogen bonds with water molecules. This can also be attributed to the colligative properties of the feed solution as the salt concentration increases. These molecules require additional kinetic energy to break their hydrogen bonds. Despite this, the MD process is still considered energy-efficient, as the inherited energy requirements are low. For seawater desalination and purification, the energy required ranges from 1 to 9 kWh/m³.

NaCl was successfully separated into crystals during the MD experiments (Figure ). Approximately 60–70% of the water was removed from the feed solutions before reaching the NaCl supersaturation, leading to NaCl precipitation/crystallization. The mass balance, obtained by comparing the PFPeA concentrations in solution and crystallized NaCl, showed that more than 97% of the PFPeA remained in solution.

4.

NaCl separation during MD experiments.

The PTFE and the PEEK membranes demonstrated over 99% PFPeA rejection, with no notable impact from the flux variation (Figure ). In contrast, PFPeA rejection using the PVDF membrane under the same conditions was slightly lower, at over 95% rejection.

5.

PFPeA rejection (%) obtained during the different MD experimental conditions (50 °C, 60 °C, and 70 °C) evaluating different membrane materials, permeate inlet temperature, and in the presence and absence of 10% NaCl. The PFPeA concentration was measured using an LC/MS/MS.

3.3. Interaction of PFPeA with the MD Membranes

Using a goniometer, the hydrophobic properties of the clean and used membranes were evaluated by measuring the contact angle of a water droplet on the membrane surface under different conditions. Figure shows the contact angle measurements of water in the presence or absence of PFPeA, NaCl, and ethanol. Membranes exposed only to PFPeA, without NaCl, are labeled as “PFAS.” Those exposed to PFPeA in the presence of NaCl are labeled as “PFAS+NaCl,” while membranes exposed to a solvent-cleaning cycle using a solution containing premeasured amounts of NaCl and ethanol are referred to as “PFAS+NaCl+Ethanol.” The solvent-cleaning cycle allows NaCl and ethanol to be pumped along the membrane’s feed side, desorbing the PFPeA molecules from the membrane surface.

6.

Impact of PFPeA on the membrane surface hydrophobicity and the efficiency of the cleaning/regeneration cycles. Contact angle analysis was performed on the membranes under different conditions, including Fresh, PFAS, PFAS+NACL, and PFAS+NACL+Ethanol.

The PTFE membrane was the most hydrophobic, with an average water contact angle of ∼130°, followed by the PEEK and PVDF membranes, with average water contact angles of 95 and 80°, respectively. The contact angle of all tested membranes decreased after interaction with PFPeA. Furthermore, adding salt increased the water contact angle of all tested membranes.

The attachment of PFPeA on the membrane surface may be similar to the adsorption of PFAS on certain types of ion exchange resins. Dual removal mechanisms have been suggested in the literature for PFAS removal by ion exchange resins. In this process, the hydrophobic carbon–fluorine tail of a typical PFAS molecule adsorbs to the hydrophobic backbone and cross-links of the resin. In contrast, the negatively charged hydrophilic head of the PFAS molecule is attracted to the positively charged ion exchange site. The presence of NaCl led to a decrease in the zeta potential of the membrane surface, thus reducing the electrostatic interaction with PFPeA.

The dual mechanism of PFAS attachment on membranes may be reversed by cleaning or regeneration cycles using brine solutions combined with organic cosolvents (such as ethanol or methanol). , A solution containing 10% ethanol and 10% NaCl effectively desorbs the anionic head and the hydrophobic carbon–fluorine of the PFAS molecule from the membrane surface. This process is effective due to the reversible nature of the electrostatic interaction between the NaCl and PFPeA and the desorbing capability of ethanol as a solvent rinse.

Figure shows an 11–16% decrease in hydrophobicity after use with PFPeA, and a 3–13% decrease in hydrophobicity after use with PFPeA and NaCl. However, membranes used in the presence of PFPeA, NaCl, and ethanol, indicating the use of the solvent-cleaning cycle, showed 96–102% regeneration of hydrophobic properties. This demonstrates that the membranes can be regenerated using a solvent-cleaning cycle in the presence of NaCl, as the ethanol and NaCl mixture in the feed solution restored the original membrane hydrophobicity.

The impact of ethanol on the performance of the MD membranes was evaluated. The water flux through the PTFE and PVDF membranes was unaffected compared to tests conducted under the same conditions without ethanol. Similarly, ethanol did not impact the rejection of the PFPeA and NaCl. However, the PEEK membrane was damaged when it came into contact with the ethanol/NaCl regeneration solution. Membrane wetting occurred when ethanol contacted the feed side of the membrane, allowing the feed solution to leak into the permeate solution.

3.4. Extended MD Testing (≥48 h)

Extended 48-h MD experiments with a feed temperature of 50 °C and a permeate temperature of 20 °C were performed (See Figure ). The average flux of each membrane was similar to the short-run tests conducted under the same conditions. As with the shorter MD tests, the permeate flux initially declined and reached a steady state, except when the PVDF membrane was used. The PFPeA rejection by the different membranes also remained stable over the 48 h. The unlaminated PTFE membrane demonstrated the highest average rate of flux. Membrane rejection of the PFPeA is further discussed in Section .

7.

Extended (∼48-h) testing of laminated PTFE, unlaminated PTFE, PEEK, and PVDF membranes for feed solutions containing 10 mg/L PFPeA at 50 °C. Permeate inlet temperature was 20 °C ± 0.5 °C; feed and permeate were circulated at a constant flow rate of 0.75 L/min.

Due to its high PFPeA rejection, mechanical strength, and resistance to ethanol, the laminated PTFE was selected for a 300-h test with regeneration cycles. The feed temperature was increased to 60 °C, and the permeate temperature was maintained at 20 °C. The average flux of the laminated PTFE membrane remained similar to the short-run tests performed under the same conditions. The permeate conductivity was continuously monitored. The membrane was rinsed by circulating a regeneration solution containing 10% NaCl and 10% ethanol. DI water was then used to thoroughly flush the membrane and the polyethylene tubing. The membrane was left to dry overnight, and the experiment without NaCl and ethanol was resumed the following day. The permeate flux remained relatively constant during the 300-h testing (Figure a). The PFPeA rejection values shown in Figure b confirm the NaCl and ethanol solution’s efficiency in restoring the PTFE membrane’s flux. PFPeA rejection increased after each regeneration cycle.

8.

Extended (∼300-h) testing of laminated PTFE membranes for feed solutions containing 10 mg/L PFPeA at 60 °C. (a) Water flux and (b) PFPeA rejection. Permeate inlet temperature was 20 °C ± 0.5 °C; feed and permeate were circulated at a constant flow rate of 0.75 L/min.

4. Conclusions

This work evaluated the performance of an energy-efficient integrated treatment process for simulated PFAS-contaminated ion-exchange resin, spent brine, or RO retentate waste streams. Four commercially available hydrophobic MD membranes (laminated PTFE, unlaminated PTFE, PEEK, and PVDF) were screened, and the long-term (∼300 h) membrane PFAS interactions were assessed for the laminated PTFE due to its effectiveness and durability. Membrane cleaning and regeneration using a NaCl and ethanol solution were also investigated. The laminated and unlaminated PTFE membranes achieved the highest rejection of NaCl and PFPeA (>99.7%) compared to 95 and 97% obtained by the PEEK and the PVDF membranes, respectively. The laminated PTFE membrane exhibited greater stability and mechanical strength than the other membranes, while the PEEK and PVDF membranes proved fragile. PFPeA was shown to cause reversible fouling in all tested membranes and reduce the membrane’s hydrophobicity.

NaCl was successfully separated into crystals during the MD experiments. The mass balance, obtained by comparing the PFPeA concentrations in solution and crystallized NaCl, showed that more than 97% of the PFPeA stayed in solution. The impact of ethanol on MD performance was also evaluated. Ethanol did not affect water flux through the PTFE and PVDF membranes. Similarly, ethanol did not affect the rejection of PFPeA and NaCl. The tested PEEK and PVDF membranes were fragile and were permanently damaged when in contact with ethanol, whereas the laminated and unlaminated PTFE membranes were not affected by ethanol.

The hydrophobicity of all tested membranes was reduced due to the attachment of PFPeA to the membrane surfaces. Cleaning or regeneration cycles with brine solutions combined with organic ethanol showed that this attachment was reversible. These cleaning cycles were surmised to be effective due to the reversible nature of the electrostatic interaction between the NaCl and PFPeA and the desorbing capability of the ethanol solvent rinse.

With further optimization, membrane distillation shows promise for recovering and concentrating PFAS-laden residual streams. This approach could make the final destruction and disposal of the PFAS more efficient.

CRediT: Hafiz H. M. Salih conceptualization, data curation, investigation, methodology, project administration, supervision, validation, writing - original draft, writing - review & editing; Erin Huggett data curation, investigation, methodology, writing - original draft, writing - review & editing; Craig Patterson conceptualization, funding acquisition, project administration, supervision, writing - original draft, writing - review & editing; John W Scott data curation, formal analysis, software, validation, visualization; Rendahandi Gune Silva data curation, validation; Tae Lee data curation, formal analysis, methodology, writing - original draft, writing - review & editing; Thomas F Speth project administration, supervision, visualization, writing - original draft, writing - review & editing; Mallikarjuna N Nadagouda conceptualization, formal analysis, funding acquisition, project administration, resources, supervision, writing - original draft, writing - review & editing.

DISCLAIMER This document has been reviewed in accordance with U.S. Environmental Protection Agency QA policy and approved for publication. 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.

The authors declare no competing financial interest.