Comparative study of PFAS treatment by UV, UV/ozone, and fractionations with air and ozonated air

Xiaodong Dai a, Zongli Xie b, Brian Dorian c, Stephen Gray *d and Jianhua Zhang *d
aShengli College, China University of Petroleum, Dongying, 257061, China
bManufacturing, CSIRO, Australia
cThe Environmental Group Ltd., Australia
dInstitute for Sustainable Industries and Liveable Cities, Victoria University, Melbourne, Australia. E-mail: stephen.gray@vu.edu.au; jianhua.zhang@vu.edu.au

Received 16th August 2019 , Accepted 22nd September 2019

First published on 30th September 2019


Per- and poly-fluorinated alkyl substances (PFAS) and their derivatives are persistent organic chemicals that result in significant adverse human health and environmental effects. In this paper, UV, ozonated air fractionation, air fractionation and combined UV/ozone treatment for PFAS removal were studied and compared. The feed water was synthesized from firefighting foam, which contained multiple per- and poly-fluorinated alkyl substances. Combined UV and ozone equipment patented by EGL were used to compare the performance of the different treatment methods, in which a two-unit rig was employed as a benchtop setup and a nine-unit rig using UV/ozone combined treatment was employed as a pilot rig. It was found from the benchtop tests that the PFAS removal efficiency was improved with the increase in air and feed flow rates in the UV/ozone combined treatment. The highest PFAS removal efficiency achieved was 87% at an air flow rate of 30 L min−1 and a feed flow rate of 1.4 L min−1 (10 min residence time). The different treatment techniques, including UV only, ozonated air fractionation, and air fractionation, were compared with the UV/ozone combined treatment at the residence time of 20 min. UV alone removed 16.8% PFAS, which showed the worst performance of all tests. The UV/ozone combined treatment removed 73% PFAS, and air fractionation was able to achieve 81% PFAS removal. Ozonated air fractionation showed the best PFAS removal efficiency, which was more than 95%, as a result of the enriched OH radicals in the gas bubbles. For the nine-unit pilot rig, the UV/ozone combined treatment achieved about 79% PFAS removal. However, foam fractionation occurred during the treatment, which led to approximately 4% removal of PFAS based on the mass balance. Therefore, the PFAS removal contributed by UV and ozone combined treatment was 75%, which was similar to the result from the benchtop rig. All treatment resulted in a concentration increase in at least one type of short-chain PFAS. When foam fractionation with gas bubbles occurred in the treatment, it was easier to remove perfluoroalkyl sulfonate (PFSA) than perfluoroalkyl carboxylate (PFCA) because PFSA is more hydrophobic than PFCA, which gives it more affinity for gas bubbles. In addition, in comparison with long-chain PFAS it is much more difficult to remove the short-chain PFAS by fractionation technologies, due to the partition factor declining exponentially with the reduced carbon number.



Water impact

PFAS are persistent chemicals of concern and pose a significant risk to human health and the environment. In this study, some advanced treatment technologies capable of commercialisation and generating less secondary waste were studied and compared for treating wastewater containing PFAS. This research provides a valuable reference for the employment of the existing technology in PFAS-contaminated water treatment.

1. Introduction

PFAS products have been used since 1940s, and their roles as synthetic pollutants emitted into the natural environment by various pathways were hardly considered by environmental chemists, toxicologists and food chemists. The first observations of organic compounds containing fluorine in humans were made at the end of 1960s but they did not attract great attention.1 At the beginning of the 2000s, the works conducted by Moody and Field2 and Giesy and Kannan3 stimulated significant interest in these anthropogenic pollutants. They considered the environmental implications of the use of firefighting foam with PFAS and the global distribution of perfluorooctanoate in wildlife.

Australia's national industrial chemicals assessment body concluded that perfluorooctane sulfonate (PFOS) and perfluorooctanoic acid (PFOA) are persistent, bioaccumulative, and toxic substances that undergo long-range transport in water and air and are transferred between different media. In 2017, the Australian Government Department of Health regulated the PFOS/perfluorohexane sulfonate (PFHxS) and PFOA concentrations of 0.07 μg L−1 and 0.56 μg L−1 in drinking water, respectively.4 In 2017, the USEPA proposed the unregulated reference concentrations of PFOA and PFOS to be 0.07 μg L−1 in drinking water.5

The primary technologies for PFAS removal from water resources can be divided into two categories:

I. Separation of PFAS from treated water, and destruction of the concentrated PFAS in a separate process.

II. Direct destruction of PFAS in situ in the treated water.

At present, the technologies in category I are widely applied in large-scale water treatment plants and include membrane filtration, adsorption, ion exchange and foam fractionation. Membrane filtration such as reverse osmosis (RO) and nanofiltration (NF) can effectively remove PFAS. However, the treatment cost by filtration is considerable for plants treating large volumes of water. Furthermore, a PFAS concentrated stream (approximately 10% of the treated water volume) is produced from RO and NF processes and needs further treatment. Currently, adsorption is the most popular technology used by water utilities. Adsorbents include powdered activated carbon (PAC) and granulated activated carbon (GAC), which are not generally regenerated. The cost of incorporating these adsorbents into existing water treatment processes can be high because a contact column is essential and the adsorbent needs to be continually purchased. The disposal of the adsorbents is also costly,6 usually requiring high-temperature incineration. Ion exchange resin can also remove PFAS efficiently.7 However, the exhausted ion exchange resin is difficult to regenerate, since it produces high salinity PFAS containing wastewater; if no regeneration is undertaken, considerable solid waste would be produced. Foam fractionation has also been used for the separation of the PFAS from wastewater since PFAS also act as surfactants. The running cost of foam fractionation is low and produces only small amounts of highly concentrated PFAS-containing water. However, the removal efficiency is about 80% and is not very effective when dealing with short-chain PFAS.

For all treatment methods of category I, the destruction of concentrated PFAS is still required. Plasma incineration followed by calcium hydroxide absorption is commonly used to convert PFAS into low hazard compounds.8

Category II methods focus on the direct destruction of PFAS, avoiding the concentration step. However, perfluorinated surfactants do not exhibit the preferred reactive sites for reaction with ˙OH radicals generated during the advanced oxidation. The abstraction of fluorine from a carbon atom is thermodynamically unfavorable because the F–OH bond has a dissociation energy at least 216 kJ mol−1 lower than that of the C–F bond (CF3F 552 kJ mol−1, R–CF2–F 352 kJ mol−1, R,R′–CFF 508 kJ mol−1).9 Furthermore, the electron density of the ionic head group (e.g., carboxylate and sulfonates) is reduced by perfluorination, hindering electron transfer reactions. The kinetic rate constant for the reaction of trifluoroacetic acid with ˙OH has been estimated to be less than 1 × 106 M−1 s−1,10 indicating a slow rate of reaction with practically no reactivity of ˙OH towards trifluoroacetic acid. PFOS is resistant to different advanced oxidation treatment methods (Fenton, UV/H2O2, ozonation in alkaline solution, and peroxone processes (O3/H2O2)) over 120 min at room temperature.11 Hori et al.12 found that the addition of hydrogen peroxide in the UV treatment method has a detrimental effect on the degradation of PFOA, as compared to direct UV-photolysis (radiation source: xenon-doped mercury lamp), which supports the persistence of PFCs in the presence of ˙OH. It was reported that 83% PFOS was removed under the conditions of 5 h UV dosing with a molecularly imprinted polymer-modified TiO2 catalyst.13

The combination of ozone and ultraviolet radiation (UV) could achieve better water treatment than the ozone alone, due to the promoted formation of hydroxyl radical intermediates.14,15 In the UV/O3 process, the dissolved ozone is split by UV irradiation (λ < 300 nm), and reacts with water to form thermally excited H2O2 that subsequently decomposes into two ·OH.15 Houston Research Inc. showed that ozone/UV enhances the oxidation of complexed cyanides, chlorinated solvents, pesticides that contribute to COD and BOD.16–19 In conventional ozonation, organic compounds are not completely oxidized to CO2 and H2O in many cases.20 In some reactions, the intermediate oxidation products remaining in the solution may be as toxic as or even more toxic than the initial compound.21 Completion of oxidation reactions, as well as oxidative destruction of compounds immune to unassisted ozone, can be achieved by supplementing the reaction with UV radiation. Many organic contaminants absorb UV energy in the range of 200–300 nm (ref. 22) and decompose due to direct photolysis or become excited and more reactive with chemical oxidants. Therefore, the combination of UV/ozone might be able to achieve effective PFAS removal.

Category II technologies are not directly favoured by most large-scale water treatment plants because strong oxidants/radiation at high doses are required to decompose PFAS, and the PFAS concentration in the wastewater is generally at the ppb to ppt level.23,24 Thus, it is necessary to combine both treatment methods in categories I and II. Through the methods in category I, the PFAS could be concentrated into a much smaller volume than the original PFAS, which will significantly reduce the load of the subsequent destruction treatments in category II. Therefore, minimizing the total rejected volume (containing high concentration PFAS) from category I would be critical to reducing the overall operation cost. A multifunctional rig patented by EGL was tested for PFAS removal. It can be used to treat the water by UV, ozonation, fractionation and their combinations.25 The performances of different treatments were also compared in this study.

2. Experimental

2.1. Benchtop rig tests

A schematic drawing of the benchtop UV/ozone rig (Environmental Group Limited, Australia) is shown in Fig. 1. The benchtop rig consists of two reaction chambers connected to two ozone generators. There is one ozone bubbler stone connected to the ozone generator and two UV lamps (λ = 254 nm, Cnlight®, China) in the reaction chamber. The air from a compressor can be partially ozonated in the ozone generator by an ozone lamp (λ = 185 nm, Cnlight®, China) in the chamber before bubbling into the reaction chambers. A peristaltic pump was used to supply the feed into the benchtop rig. The total volume of the two reaction chambers was 14 L. The UV intensity was measured by UV intensity meters (210–280 nm, Run-Goal Technology Co. Ltd. China), which showed different readings for the same amount of UV radiation due to the difference in the distances to the UV lamps. The air flow rates were monitored by an air flow meter (Siargo®, M5700 Series, China). The PFAS containing feed was synthesized by mixing waste firefighting foam (Cleanaway, Australia) containing about 6.3 mg kg−1 PFAS with tap water to achieve PFAS concentrations between 3 and 10 μg L−1.
image file: c9ew00701f-f1.tif
Fig. 1 Schematic diagram of the benchtop rig.
2.1.1 Blank test. Blank tests were conducted with the tap water to achieve the operation baseline. The UV intensity and ozone residuals were measured without the interference from the added PFAS-containing firefight foam. The feed flow rates were set at 0.35 and 1.4 L min−1 to obtain residence times of 40 and 10 min, respectively. The air flow rates during the tests were varied at 10, 20 and 30 L min−1, respectively.
2.1.2 System PFAS adsorption test. An adsorption test was conducted to check whether PFAS would be adsorbed in the benchtop testing rig. The original PFAS wastewater was pumped through the rig at a flow rate of 0.7 L min−1 (residence time = 20 min). The feed was sampled at the beginning of the test, and the effluent was sampled after one hour.
2.1.3 UV/ozone combined treatment for PFAS wastewater. In the UV/ozone combined PFAS wastewater treatment tests, both the ozone and UV lamps were turned on. The feed flow rate was set at 0.35, 0.7 and 1.4 L min−1 to achieve the residence times of 40, 20 and 10 min, respectively. The air flow rate was varied in the range of 15–30 L min−1.
2.1.4 Ozonated air fractionation, air fractionation and UV for PFAS wastewater treatment. As can be seen from the schematic diagram of the benchtop rig in Fig. 1, it is possible to have UV radiation, air fractionation and ozonated air fractionation/ozone oxidation involved in the treatment. Therefore, it is necessary to identify the dominant PFAS removal mechanism in the treatment. All the tests were conducted at the feed flow rate of 0.7 L min−1 (residence time of 20 min). The feed was sampled at the beginning of the test, and the effluent was sampled at the end of the test (after one hour).

In the UV test, only UV lamps were switched on, and there was no air bubbling into the reaction chambers.

In the air fractionation tests, only air was bubbled into the reaction chambers at an air flow rate of 20 L min−1, and both the UV and ozone lamps were off.

In the ozonated air fractionation tests, the ozone lamps were on, and the UV lamps were off. The air at a flow rate of 20 L min−1 was partially ozonated by the ozone lamps and bubbled into the reaction chambers.

2.2. Pilot UV/ozone rig test

The scale-up of the engineering process may result in great performance differences relative to the benchtop tests.26 Therefore, a pilot test was also conducted under the optimum conditions identified from the benchtop tests. In comparison to the benchtop rig, the pilot rig included nine reaction chambers and sixteen ozone generators. All the reaction chambers were equipped with two UV lamps, but three of them were not connected to the ozone generators, which is different from the benchtop rig. The total volume of the pilot rig was 63 L. Therefore, the feed flow rate was set at 3.15 L min−1 to achieve a 20 min residence time. The feed was prepared by dissolving 100 g firefighting foam into 200 L water. The experiment lasted for one hour. The feed was sampled at the beginning of the test, and the effluent was sampled at the end of the test.

2.3. Analysis

The PFAS in the water was analyzed by ALS, Australia with the LC-MS-MS method EP231X, and the detection limit was in the range of 0.01–0.1 μg L−1. The total PFAS concentration is the sum of the concentrations of 28 PFAS analytes, which are listed in Appendix 1.

The ozone residual in the reaction chamber was analyzed by the Hach DR/800 with the AccuVac® method. Triplicate measurements were conducted for the same tests, and the mean value is reported in this paper.

3. Results and discussion

3.1. Benchtop UV/ozone combined tests

3.1.1 Blank tests with tap water for identification of the maximum UV radiation and ozone residuals. Blank tests were carried out by using the tap water that was used to dilute the PFAS containing firefighting foam. The purpose of these tests was to understand the baseline performance of the combined UV/ozone system in the absence of PFAS. Besides, the blank tests also helped to identify the variations in ozone residuals and UV intensity with the feed and air flow rates, as the reference for the subsequent PFAS water treatment. Table 1 shows the ozone residuals and the UV intensity of the blank tests in reaction chambers R1 and R2 at feed flow rates of 1.4 and 0.35 L min−1 respectively, which are also the theoretical maximum values achievable for subsequent PFAS tests under similar conditions. At the higher feed flow rate of 1.4 L min−1 (10 min residence time), the total ozone residuals in R1 and R2 were almost the same and in the range of 0.18–0.20 mg L−1; this difference is within the error range. However, the UV intensity declined by about 17% in R1 as the air flow rate was increased to 30 L min−1, due to the enhanced deflection/reflection of UV beams by the increased air bubbles.27 At the lower feed flow rate of 0.35 L min−1 (40 min residence time), the maximum total ozone residuals in R1 and R2 were 0.2 mg L−1 when the air flow rate was 20 L min−1. The measured UV intensity also reduced to approximately 12% at an air flow rate of 30 L min−1, in comparison with that of an air flow rate of 10 L min−1 in R1.
Table 1 Blank tests: ozone residual and UV intensity
Feed flowrate (L min−1) Air flowrate (L min−1) UV intensity (μW cm−2) Ozone residual (mg L−1) Total
R 1 R 2 R 1 R 2
a The different readings are due to the different installation distances to the UV lamps.
1.4 12 100 44 0.08 0.12 0.20
1.4 20 100 45 0.11 0.08 0.19
1.4 30 83 43 0.09 0.09 0.18
0.35 10 84 43 0.07 0.06 0.13
0.35 20 80 43 0.14 0.06 0.20
0.35 30 74 44 0.06 0.04 0.10


It is also worth mentioning that in general, the total ozone residuals at high feed flow rates were higher than those at low feed flow rates. This is because of the high liquid turbulence at high flow rate, encouraging the ozone transfer from the gas phase to the liquid phase.28 Furthermore, in R1, the UV intensity was also higher at a feed flow rate of 1.4 L min−1, as compared to that at 0.35 L min−1 with the same air flow rate, because finer and more numerous air bubbles created at higher feed flowrates will produce more uniform UV scattering.29

3.1.2 System PFAS adsorption test. All the detectable PFAS in the system adsorption tests are shown in Table 2. The PFAS concentrations in the effluent were found to be almost the same as the feed, thus, there was no detectable PFAS adsorption by the benchtop rig in our tests.
Table 2 Analytical results for the system adsorption test
Sample PFPeS (μg L−1) PFHxS (μg L−1) PFHpS (μg L−1) PFOS (μg L−1) PFPeA (μg L−1) PFHxA (μg L−1) PFHpA (μg L−1) PFOA (μg L−1) PFAS (μg L−1)
Feed 0.05 0.38 0.07 2.3 0.02 0.18 0.04 0.11 3.15
Effluent 0.04 0.37 0.08 2.4 0.02 0.14 0.04 0.10 3.19


3.1.3 UV/ozone combined treatment for PFAS wastewater. In the UV/ozone combined tests, both UV and ozone lamps were switched on, and the ozonated air was bubbled into the reaction chambers. Fig. 2a and b show the total PFAS concentration in the treated water and the total PFAS removal efficiency under different feed and air flowrates, respectively. Total PFAS removal efficiency was calculated using eqn (1) as below:
 
image file: c9ew00701f-t1.tif(1)
where Re is the percentage of PFAS removed from the treated water, Ct is the concentration of PFAS in the treated water, and Cf is the PFAS concentration in the feed.

image file: c9ew00701f-f2.tif
Fig. 2 Treatment at different feed and air flow rates.

The maximum total PFAS removal was 87% at an air flow rate of 30 L min−1 and feed flow rate of 1.4 L min−1. At the same air flow rate, better PFAS removal was achieved at higher feed flowrate, owing to the higher UV intensity and ozone residual obtained at higher feed flowrate in the blank tests as shown in Table 1. Therefore, it can be concluded that the high ozone transfer rate and UV intensity are able to enhance the PFAS removal.

Table 3 shows the concentrations of four major per- and poly-fluoroalkyl substances in the feed and treated water. It can be seen that after the ozone/UV combined treatment, all PFAS concentrations were reduced, except for perfluorohexanoic acid (PFHxA). Moreover, the increases in the feed and air flow rates could facilitate the PFOS and PFOA removals under all the operating conditions. However, the feed and air flow rates had less influence on PFHxS removal. On the other hand, the PFHxA concentration increased under almost all the treatment conditions, and up to a 50% increase in PFHxA was observed when 97.3% PFOA removal was achieved. Therefore, it can be confirmed that the UV/ozone combined treatment could convert PFAS into PFHxA.30 It seems that PFHxS could also be generated from other PFAS sources in the treatment process, but it could not be demonstrated unequivocally based on the available analytical results. After the treatment under the optimum treatment conditions (air flowrate = 30 L min−1, feed flowrate = 1.4 L min−1), the total concentration of PFHxS (0.48 μg L−1) and PFHxA (0.22 μg L−1) was 0.70 μg L−1 greater than the total concentration (0.60 μg L−1) of PFOS (0.52 μg L−1) and PFOA (0.08 μg L−1), although their total concentration in the feed was only one third of that of the PFOS and PFOA. Therefore, to achieve high total PFAS removal, the PFHxA and PFHxS concentrations should be considered, as they could not be easily removed by the UV/ozone combined treatment and PFHxA appeared to be generated by the process.

Table 3 Operation parameters and analytical results
Sample Feed flowrate (L min−1) Air flowrate (L min−1) PFHxS PFOS PFHxA PFOA
C (μg L−1) Re (%) C (μg L−1) Re (%) C (μg L−1) Re (%) C (μg L−1) Re (%)
Feed 2.40 5.40 0.16 2.20
S1 0.35 15 0.44 81.7 2.30 57.4 0.18 −12.5 0.60 72.7
S2 0.35 20 0.46 80.8 1.70 68.5 0.19 −18.8 0.80 63.6
S3 0.35 30 0.48 80.0 1.10 79.6 0.18 −12.5 0.26 88.2
S4 0.7 15 0.70 70.8 2.20 59.3 0.18 −12.5 0.12 94.5
S5 0.7 20 0.58 75.8 1.60 70.4 0.18 −12.5 0.40 81.8
S6 0.7 30 0.60 75.0 0.72 86.7 0.19 −18.8 0.18 91.8
S7 1.4 15 0.58 75.8 1.50 72.2 0.16 0.0 0.24 89.1
S8 1.4 20 0.92 61.7 0.28 94.8 0.24 −50.0 0.06 97.3
S9 1.4 30 0.48 80.0 0.52 90.4 0.22 −37.5 0.08 96.4


3.1.4 UV, air fractionation and ozonated air fractionation for PFAS wastewater treatment. The purpose of the tests was to verify the overall PFAS removal efficiency and to identify the removal efficiency of specific types of PFAS under different treatment methods provided by the benchtop rig.

Fig. 3 shows images of the observation windows of the UV/ozone combined test, air fractionation test, ozonated air fractionation test, and UV only test. In Fig. 3a and d, no foaming was observed; however, foaming occurred in both fractionation processes where air and ozonated air were bubbled through the reaction chambers, as can be seen from Fig. 3b and c. Since foaming in the reaction chambers is a sign of fractionation due to existing surfactants (PFAS), no foam being observed in the UV/ozone combined and UV tests indicated that the PFAS removal in both processes was due to the destruction rather than foam fractionation.


image file: c9ew00701f-f3.tif
Fig. 3 Foaming at different operations.

The analytical results from the various processes are shown in Table 4. Feed 1 was used for the air fractionation test, and S1 was the treated water by air fractionation. Feed 2 was used for both UV and ozonated air fractionation tests, and S2 and S3 were the respective treated waters. The removal efficiencies of all detectable PFAS containing different carbon numbers are shown in Fig. 4. Since the detection limit was 0.02 μg L−1 for the perfluoropentane sulfonate (PFPeS) and perfluoropentanoic acid (PFPeA) containing five carbons, it is impossible to make a solid conclusion based on the analytical method, although some PFPeA removal was found during the ozonated air fractionation. Therefore, the removal efficiency of the PFAS containing five carbons is not shown in Fig. 4. Furthermore, since some concentration changes after treatment were less than the detection limit of 0.02 μg L−1, those changes were considered in the analytical error range and are not discussed further.

Table 4 Analytical results for the feed and treated water
Sample Operation PFPeS (μg L−1) PFHxS (μg L−1) PFHpS (μg L−1) PFOS (μg L−1) PFPeA (μg L−1) PFHxA (μg L−1) PFHpA (μg L−1) PFOA (μg L−1) PFAS (μg L−1)
Feed 1 0.05 0.38 0.07 2.3 0.02 0.18 0.04 0.11 3.15
S1 Air fractionation 0.03 0.06 <0.02 0.3 0.02 0.15 0.02 0.01 0.59
Feed 2 0.04 0.3 0.06 2.8 0.02 0.16 0.04 0.09 3.51
S2 UV 0.05 0.4 0.05 2.1 0.02 0.16 0.04 0.1 2.92
S3 Ozonated air fractionation 0.03 <0.05 <0.02 <0.04 <0.02 0.14 <0.02 <0.01 0.17



image file: c9ew00701f-f4.tif
Fig. 4 PFAS removal by air fractionation, UV and ozonated air fractionation (residence time = 20 min).

It was found that UV treatment could remove 16.8% PFAS in total (Fig. 4), which was mainly associated with PFOS removal (Table 4). After the UV treatment, 25% PFOS was removed from 2.8 to 2.1 μg L−1 in the treated water, but the PFOA concentration did not change. However, the PFHxS concentration increased by approximately 33% from 0.3 to 0.4 μg L−1. Since only PFOS and/or PFOA were used in the firefighting foam, and the variation of perfluoroheptane sulfonate (PFHpS) from 0.06 to 0.05 μg L−1 is in the range of the detection limit and is negligible, it can be concluded that the newly formed PFHxS was mainly from PFOS degradation.

Based on research from Vecitis et al., perfluoroalkyl carboxylate (PFCA) should have a faster photolytic degradation as compared to perfluoroalkyl sulfonate (PFSA)31 However, it can also be seen in Table 4 that PFCA such as PFOA, PFHxA or PFHpA did not show obvious changes after the UV treatment, in comparison to the degradation of PFOS (PFSA). Yamamoto, et al. found that one of the photodegradation pathways of PFOS is through transformation into PFCA.30 It was also found that 80% of the total PFAS in the feed was in the form of PFOS as shown in Table 4. Therefore, even if the PFCA degraded faster than PFSA, it could be possibly replenished by the photodegradation of the PFOS in this study; thus, it would not be possible to make any conclusion as to whether the PFCA had been destroyed by UV treatment in our study.

From Fig. 4, it can be seen that air fractionation removed about 81.3% PFAS in total, and 80% PFHxS, approximately 100% PFHpS, 87.0% PFOS, 16.7% PFHxA, 50.0% PFHpA, and 91% PFOA were removed from the treated water. Since air is not a strong oxidant, foam fractionation should be the dominant separation mechanism for the PFAS removal. For the PFSA, the removal efficiency did not show a clear relationship to the molecular weight. However, the removal efficiency for PFCA declined more than 80% as the carbon number was reduced from 8 to 6, because the lower lipophilicity of the shorter alkyl chain PFCA led to its lower affinity for the air bubbles.32,33

The total PFAS removal efficiency was about 95% by the ozonated air fractionation, as shown in Fig. 4, which is the highest among all the tested treatment methods. After the treatment, no PFSA, PFOA and PFHpA were detected in the treated water. However, only 12.5% PFHxA was removed from the treated water. Based on the blank tests, the maximum total ozone residuals in both reaction chambers was 0.2 mg L−1, which could not oxidize the PFAS in a contact time of 20 min based on bond energy theory.11 Therefore, the dominant PFAS removal mechanism should be ozone-boosted foam fractionation. When ozone is dosed into water, hydroxyl radicals are generated, which are strong electron binders.34,35 As shown in Fig. 5, there will be an OH radical rich layer around the interface of the gas bubbles, due to the ozone diffusion from the gas phase into the water phase. Since strong hydroxyl radical electron binders can bind to the negatively charged hydrophilic SO3/COO ends of the PFOS/PFOA based on affinity theory,36–39 more PFAS can be concentrated around the interface of the gas bubbles, in comparison with air fractionation. Therefore, the ozonated air fractionation can achieve higher PFAS separation efficiency than the air fractionation.


image file: c9ew00701f-f5.tif
Fig. 5 Schematic diagram of the proposed PFAS affinity for the gas bubble.

PFHxA is highly persistent and mobile in the environment as a short-chain PFAS.40 Based on the results, it can be seen that none of the tested methods can remove PFHxA as effectively as other PFAS substances. The UV/ozone combined treatment method even caused an increase in the PFHxA concentration in the treated water, when PFOA was present in the feed water. Therefore, ozone-enhanced fractionation or air fractionation could be used to lower the PFOA concentration before applying the UV and ozone combined technology to destroy any PFAS passing fractionation treatment. Although the possibility exists that UV and ozone combined technology could destroy the PFHxA, it could not be confirmed based on our test results and further research is necessary to determine if it is possible and at what rate it might occur.

3.2. Pilot-scale UV/ozone combined tests

It can be seen in Fig. 6 that the foaming only occurred in the last two reaction chambers. During the 1 h experimental period, 189 L water was treated, and 2.5 L of water was collected from the overflow line due to the foaming in the last two reaction chambers.
image file: c9ew00701f-f6.tif
Fig. 6 Foaming in the pilot rig.

From Table 5, after the combined UV/ozone treatment, the mass of all short-chain PFCA increased, while the mass of all short-chain PFSA decreased. The 6[thin space (1/6-em)]:[thin space (1/6-em)]2 fluorotelomer sulfonic acid (6[thin space (1/6-em)]:[thin space (1/6-em)]2 FTS) that was not detected in both the feed and effluent was found in the foam. Therefore, it can be concluded that the advanced oxidation could facilitate the formation of short-chain PFAS.

Table 5 PFAS concentrations in the feed, treated water and foam
Sample PFPeS (μg L−1) PFHxS (μg L−1) PFHpS (μg L−1) PFOS (μg L−1) PFPeA (μg L−1) PFHxA (μg L−1) PFHpA (μg L−1) PFOA (μg L−1) 6[thin space (1/6-em)]:[thin space (1/6-em)]2 FTS (μg L−1) PFAS (μg L−1)
Feed 0.04 0.24 0.03 1.20 <0.02 0.14 0.03 0.07 <0.05 1.75
Effluent 0.02 0.09 <0.02 0.05 0.02 0.15 0.03 0.01 <0.05 0.37
Foam 0.09 0.90 0.13 4.10 0.03 0.23 0.13 0.30 0.06 5.97


In Fig. 7, the removal efficiency of PFAS during the pilot rig operation by the UV/ozone combined treatment was calculated using eqn (2).

 
image file: c9ew00701f-t2.tif(2)
where Dpfas is the percentage of PFAS destroyed by the combined UV and ozone treatment; Cfoam and Ce are the PFAS concentrations in the foam and treated water (effluent), respectively, and Vfoam, Ve and Vf are the volumes of foam, effluent and feed, respectively.


image file: c9ew00701f-f7.tif
Fig. 7 Percentage of PFAS removed from the effluent, and PFAS destroyed by the combined UV and ozone treatment (residence time = 20 min, air flow rate = 160 L min−1).

It can be seen from Fig. 7 that about 79% of PFAS in total was removed from the treated water (effluent), in which 75% was destroyed by the combined UV/ozone treatment similar to the result achieved by the benchtop rig at the same residence time. Meanwhile, with the same carbon number, it is easier to remove the PFSA than the PFCA from the feed by either destruction or fractionation, which is attributed to the hydrophobicity or lipophilicity difference between PFCA and PFSA. With the same alkyl chain length, PFCAs are more hydrophilic than PFSAs.32,33 Therefore, the affinity of the PFSA molecule for the water-bubble interface is higher than that of the PFCA molecule.41 As a result, it would be easier for the PFSA to access the oxidant in the bubble and to partition into the foam as compared to PFCA. Furthermore, the concentration of the short-chain PFCA including PFHxA and PFHpA increased due to the UV/ozone combined destruction, consistent with the benchtop results.

To assess the influence of carbon number on the foam fractionation, the partition factors of the PFAS with different carbon numbers were calculated by eqn (3) using data in Table 5 and shown in Fig. 8.

 
image file: c9ew00701f-t3.tif(3)
where λ is the partition factor between the effluent and the foam.


image file: c9ew00701f-f8.tif
Fig. 8 Distribution factors vs. PFAS carbon number.

It was found that the partition factor has a nearly exponential relationship with the PFAS carbon number, due to the octanol–water and air-water partition coefficients increasing exponentially with perfluorinated chain length.42 Therefore, partitioning the shorter-chain PFAS from the bulky feed into the foam would be more difficult as compared to that of the long-chain PFAS by the ozonated air fractionation treatment.

4. Conclusions

The performances of ozonated air fractionation, air fractionation, UV, and combined UV/ozone treatments in PFAS removal were studied and compared. The benchtop tests were conducted at a hydraulic residence time of 20 min and air flow rate of 20 L min−1, and the conclusions were made as follows:

• Ozonated air fractionation achieved the best performance, in which about 95% PFAS was removed from the treated water,

• UV showed the worst performance of 17% PFAS removal,

• Air fractionation removed 81% of PFAS, and

• UV/ozone combined treatment achieved 73% removal of PFAS at a residence time of 20 min, which was also confirmed in the pilot rig test.

In comparison with the air fractionation, the ozonated air fractionation had the higher PFAS removal efficiency because of the affinity of OH radicals for the negatively charged hydrophilic SO3–/COO– ends of the PFOS/PFOA.

It is relatively easier to remove PFSA than PFCA by both the air and ozonated air fractionations and the UV/ozone combined treatment, which is attributed to the involvement of gas bubbles and the greater lipophilicity of PFSA.

When PFAS destruction is involved, the short-chain PFAS concentration will generally increase after the treatment. Both the air and ozonated air fractionations could not effectively partition the short-chain PFAS from the contaminated water into the foam, due to the exponential decline of the partition factor with the reducing carbon number of the PFAS.

Appendix 1

PFASPerfluorobutane sulfonic acid (PFBS)
PFASPerfluoropentane sulfonic acid (PFPeS)
PFASPerfluorohexane sulfonic acid (PFHxS)
PFASPerfluoroheptane sulfonic acid (PFHpS)
PFASPerfluorooctane sulfonic acid (PFOS)
PFASPerfluorodecane sulfonic acid (PFDS)
PFASPerfluorobutanoic acid (PFBA)
PFASPerfluoropentanoic acid (PFPeA)
PFASPerfluorohexanoic acid (PFHxA)
PFASPerfluoroheptanoic acid (PFHpA)
PFASPerfluorooctanoic acid (PFOA)
PFASPerfluorononanoic acid (PFNA)
PFASPerfluorodecanoic acid (PFDA)
PFASPerfluoroundecanoic acid (PFUnDA)
PFASPerfluorododecanoic acid (PFDoDA)
PFASPerfluorotridecanoic acid (PFTrDA)
PFASPerfluorotetradecanoic acid (PFTeDA)
PFASPerfluorooctane sulfonamide (FOSA)
PFAS N-Methyl perfluorooctane sulfonamide (MeFOSA)
PFAS N-Ethyl perfluorooctane sulfonamide (EtFOSA)
PFAS N-Methyl perfluorooctane sulfonamidoethanol
PFAS N-Ethyl perfluorooctane sulfonamidoethanol
PFAS N-Methyl perfluorooctane sulfonamidoacetic acid
PFAS N-Ethyl perfluorooctane sulfonamidoacetic acid
PFAS4[thin space (1/6-em)]:[thin space (1/6-em)]2 Fluorotelomer sulfonic acid (4[thin space (1/6-em)]:[thin space (1/6-em)]2 FTS)
PFAS6[thin space (1/6-em)]:[thin space (1/6-em)]2 Fluorotelomer sulfonic acid (6[thin space (1/6-em)]:[thin space (1/6-em)]2 FTS)
PFAS8[thin space (1/6-em)]:[thin space (1/6-em)]2 Fluorotelomer sulfonic acid (8[thin space (1/6-em)]:[thin space (1/6-em)]2 FTS)
PFAS10[thin space (1/6-em)]:[thin space (1/6-em)]2 Fluorotelomer sulfonic acid (10[thin space (1/6-em)]:[thin space (1/6-em)]2 FTS)

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

Funding from the Environmental Group Limited and Australian Department of Industry, Innovation and Science are gratefully acknowledged. Pilot and benchtop testing rigs were supplied by the Environmental Group Limited. All Intellectual Property relating to this project and equipment design are covered by the EGL's existing Patent.

References

  1. D. R. Taves, Evidence that there are Two Forms of Fluoride in Human Serum, Nature, 1968, 217, 1050 CrossRef CAS PubMed.
  2. C. A. Moody and J. A. Field, Perfluorinated Surfactants and the Environmental Implications of Their Use in Fire-Fighting Foams, Environ. Sci. Technol., 2000, 34, 3864–3870 CrossRef CAS.
  3. J. P. Giesy and K. Kannan, Global Distribution of Perfluorooctane Sulfonate in Wildlife, Environ. Sci. Technol., 2001, 35, 1339–1342 CrossRef CAS PubMed.
  4. Final Health Based Guidance Values for PFAS for use in site investigations in Australia, Australian Government Department of Health, April, 2017 Search PubMed.
  5. The third unregulated contaminant monitoring rule, USEPA, 2017 Search PubMed.
  6. Recommendation on Perfluorinated Compound Treatment Options for Drinking Water, New Jersey Drinking Water Quality Institute, USA, 2015 Search PubMed.
  7. I. Ross, J. McDonough, J. Miles, P. Storch, P. Thelakkat Kochunarayanan, E. Kalve, J. Hurst, S. S. Dasgupta and J. Burdick, A review of emerging technologies for remediation of PFASs, Remediation Journal, 2018, 28, 101–126 CrossRef.
  8. K. H. Kucharzyk, R. Darlington, M. Benotti, R. Deeb and E. Hawley, Novel treatment technologies for PFAS compounds: A critical review, J. Environ. Manage., 2017, 204, 757–764 CrossRef CAS PubMed.
  9. G. W. Olsen, D. C. Mair, T. R. Church, M. E. Ellefson, W. K. Reagen, T. M. Boyd, R. M. Herron, Z. Medhdizadehkashi, J. B. Nobiletti and J. A. Rios, Decline in perfluorooctanesulfonate and other polyfluoroalkyl chemicals in American Red Cross adult blood donors, 2000− 2006, Environ. Sci. Technol., 2008, 42, 4989–4995 CrossRef CAS PubMed.
  10. B. des Umweltbundesamtes, Referenzwerte für Perfluoroctansäure (PFOA) und Perfluoroctansulfonsäure (PFOS) im Blutplasma, 2009 Search PubMed.
  11. A. M. Calafat, L. L. Needham, Z. Kuklenyik, J. A. Reidy, J. S. Tully, M. Aguilar-Villalobos and L. P. Naeher, Perfluorinated chemicals in selected residents of the American continent, Chemosphere, 2006, 63, 490–496 CrossRef CAS PubMed.
  12. H. Hori, A. Yamamoto, K. Koike, S. Kutsuna, I. Osaka and R. Arakawa, Photochemical decomposition of environmentally persistent short-chain perfluorocarboxylic acids in water mediated by iron (II)/(III) redox reactions, Chemosphere, 2007, 68, 572–578 CrossRef CAS PubMed.
  13. Y. Wu, Y. Li, A. Tian, K. Mao and J. Liu, Selective Removal of Perfluorooctanoic Acid Using Molecularly Imprinted Polymer-Modified TiO2 Nanotube Arrays, Int. J. Photoenergy, 2016, 2016, 10 CrossRef.
  14. W. H. Glaze, J.-W. Kang and D. H. Chapin, The chemistry of water treatment processes involving ozone, hydrogen peroxide and ultraviolet radiation, 1987 Search PubMed.
  15. D. B. Miklos, C. Remy, M. Jekel, K. G. Linden, J. E. Drewes and U. Hübner, Evaluation of advanced oxidation processes for water and wastewater treatment – A critical review, Water Res., 2018, 139, 118–131 CrossRef CAS PubMed.
  16. R. Garrison, C. Mauk and H. Prengle Jr, Advanced ozone oxidation system for complexed cyanides, Ozone for Water and Waste Water Treatment, Int. Ozone Assoc., Cleveland, OH, 1975, 551–577 Search PubMed.
  17. H. Prengle Jr, C. Hewes and C. Mank, Oxidation of refractory materials by ozone with ultraviolet radiation, in Proc. Int. Symp. Ozone Technology, 1975, pp. 224 Search PubMed.
  18. H. Prengle Jr, C. Mauk and J. Payne, Ozone-UV oxidation of pesticides in aqueous solution, in Technical paper presented at the IOI Meeting, Cincinnati, OH, 1976 Search PubMed.
  19. H. Prengle Jr, Evolution of the Ozone/UV Process for Wastewater Treatment, in Symposium on Wastewater Treatment & Disinfection With Ozone, Cininnati, Ohio, 1977 Search PubMed.
  20. T. Commentary, Advanced Oxidation Processes for treatment of industrial wastewater, An EPRI Community Environmental Center Publ, 1996 Search PubMed.
  21. R. Munter, Advanced oxidation processes–current status and prospects, Proc. Est. Acad. Sci., Chem., 2001, 50, 59–80 CAS.
  22. M. S. Díaz-Cruz, M. Llorca and D. Barceló, Organic UV filters and their photodegradates, metabolites and disinfection by-products in the aquatic environment, TrAC, Trends Anal. Chem., 2008, 27, 873–887 CrossRef.
  23. R. Loos, R. Carvalho, D. C. António, S. Comero, G. Locoro, S. Tavazzi, B. Paracchini, M. Ghiani, T. Lettieri and L. Blaha, EU-wide monitoring survey on emerging polar organic contaminants in wastewater treatment plant effluents, Water Res., 2013, 47, 6475–6487 CrossRef CAS PubMed.
  24. F. Xiao, T. R. Halbach, M. F. Simcik and J. S. Gulliver, Input characterization of perfluoroalkyl substances in wastewater treatment plants: source discrimination by exploratory data analysis, Water Res., 2012, 46, 3101–3109 CrossRef CAS PubMed.
  25. I. Cummins, Fluid treatment apparatus, WO2010/139001Al, 2014.
  26. M. Zlokarnik, Scale-up in chemical engineering, John Wiley & Sons, 2006 Search PubMed.
  27. Y. Fan, N. V. Lafferty, A. Bourov, L. V. Zavyalova and B. W. Smith, Study of air-bubble-induced light scattering effect on image quality in 193-nm immersion lithography, SPIE, 2004 Search PubMed.
  28. J. Villadsen, J. Nielsen and G. Lidén, in Bioreaction Engineering Principles, Springer US, Boston, MA, 2011, pp. 459–496,  DOI:10.1007/978-1-4419-9688-6_10.
  29. Y. Fan, N. Lafferty, A. Bourov, L. Zavyalova and B. W. Smith, Air bubble-induced light-scattering effect on image quality in 193 nm immersion lithography, Appl. Opt., 2005, 44, 3904–3911 CrossRef PubMed.
  30. T. Yamamoto, Y. Noma, S.-i. Sakai and Y. Shibata, Photodegradation of perfluorooctane sulfonate by UV irradiation in water and alkaline 2-propanol, Environ. Sci. Technol., 2007, 41, 5660–5665 CrossRef CAS PubMed.
  31. C. D. Vecitis, H. Park, J. Cheng, B. T. Mader and M. R. Hoffmann, Treatment technologies for aqueous perfluorooctanesulfonate (PFOS) and perfluorooctanoate (PFOA), Front. Environ. Sci. Eng. China, 2009, 3, 129–151 CrossRef CAS.
  32. N. A. Fernandez, L. Rodriguez-Freire, M. Keswani and R. Sierra-Alvarez, Effect of chemical structure on the sonochemical degradation of perfluoroalkyl and polyfluoroalkyl substances (PFASs), Environ. Sci.: Water Res. Technol., 2016, 2, 975–983 RSC.
  33. J. M. Conder, R. A. Hoke, W. d. Wolf, M. H. Russell and R. C. Buck, Are PFCAs Bioaccumulative? A Critical Review and Comparison with Regulatory Criteria and Persistent Lipophilic Compounds, Environ. Sci. Technol., 2008, 42, 995–1003 CrossRef CAS PubMed.
  34. D. M. Chipman, Electron affinity of hydroxyl radical, J. Chem. Phys., 1986, 84, 1677–1682 CrossRef CAS.
  35. M. S. Elovitz and U. von Gunten, Hydroxyl Radical/Ozone Ratios During Ozonation Processes. I. The Rct Concept, Ozone: Sci. Eng., 1999, 21, 239–260 CrossRef CAS.
  36. L. Dogliotti and E. Hayon, Flash photolysis of per [oxydi] sulfate ions in aqueous solutions. The sulfate and ozonide radical anions, J. Phys. Chem., 1967, 71, 2511–2516 CrossRef CAS.
  37. Z. Yang, R. Su, S. Luo, R. Spinney, M. Cai, R. Xiao and Z. Wei, Comparison of the reactivity of ibuprofen with sulfate and hydroxyl radicals: an experimental and theoretical study, Sci. Total Environ., 2017, 590, 751–760 CrossRef PubMed.
  38. R. G. Pearson, Ionization potentials and electron affinities in aqueous solution, J. Am. Chem. Soc., 1986, 108, 6109–6114 CrossRef CAS.
  39. P. S. Surdhar, S. P. Mezyk and D. A. Armstrong, Reduction potential of the carboxyl radical anion in aqueous solutions, J. Phys. Chem., 1989, 93, 3360–3363 CrossRef CAS.
  40. S. Brendel, É. Fetter, C. Staude, L. Vierke and A. Biegel-Engler, Short-chain perfluoroalkyl acids: environmental concerns and a regulatory strategy under REACH, Environ. Sci. Eur., 2018, 30, 9 CrossRef PubMed.
  41. B. P. Binks, M. Kirkland and J. A. Rodrigues, Origin of stabilisation of aqueous foams in nanoparticle–surfactant mixtures, Soft Matter, 2008, 4, 2373–2382 RSC.
  42. S. Rayne and K. Forest, Congener-specific organic carbon-normalized soil and sediment-water partitioning coefficients for the C1 through C8 perfluoroalkyl carboxylic and sulfonic acids, J. Environ. Sci. Health, Part A: Toxic/Hazard. Subst. Environ. Eng., 2009, 44, 1374–1387 CrossRef CAS PubMed.

Footnote

Electronic supplementary information (ESI) available. See DOI: 10.1039/c9ew00701f

This journal is © The Royal Society of Chemistry 2019