Covalent triazine-based framework: A promising adsorbent for removal of perfluoroalkyl acids from aqueous solution

Environmental Pollution xxx (2016) 1e9

Contents lists available at ScienceDirect

Environmental Pollution journal homepage: www.elsevier.com/locate/envpol

Covalent triazine-based framework: A promising adsorbent for removal of perfluoroalkyl acids from aqueous solution* Bingyu Wang a, Linda S. Lee b, Chenhui Wei a, Heyun Fu a, Shourong Zheng a, Zhaoyi Xu a, Dongqiang Zhu a, c, * a b c

State Key Laboratory of Pollution Control and Resource Reuse/School of the Environment, Nanjing University, Jiangsu, 210046, China Department of Agronomy, Purdue University, West Lafayette, IN, 47907, USA School of Urban and Environmental Sciences, Peking University, Beijing, 100871, China

a r t i c l e i n f o

a b s t r a c t

Article history: Received 11 March 2016 Received in revised form 18 June 2016 Accepted 27 June 2016 Available online xxx

Perfluoroalkyl acids (PFAAs) are highly stable, persistent, and ubiquitous in the environment with significant concerns growing with regards to both human and ecosystem health. Due to the high stability to both biological and chemical attack, the only currently feasible approach for their removal from water is adsorbent technology. The main objective of this study was to assess a covalent triazine-based framework (CTF) adsorbent for removal from aqueous solutions of perfluoro C4, C6, and C8 carboxylates and sulfonates including the two C8s most commonly monitored, perfluorooctanoic acid (PFOA) and perfluorooctane sulfonate (PFOS). Adsorption affinity and capacity were quantified and compared to three commonly used sorbents: pulverized microporous activated carbon, single-walled carbon nanotubes, and Amberlite IRA-400 anion-exchange resin. CTF adsorbent exhibited pronouncedly higher adsorption affinity and capacity of PFAAs than other test sorbents. The remarkably strong adsorption to CTF can be attributed to the favored electrostatic interaction between the protonated triazine groups on the inner wall of the hydrophobic CTF pore and the negatively charged head groups of the PFAAs intercalated between the CTF layers. The homogeneous, nanosized pores (1.2 nm) of CTF hindered adsorption of a large-sized dissolved humic acid, thus minimizing the suppression of PFAA adsorption. Additionally, regeneration of CTF was easily accomplished by simply raising pH > 11, which inhibited the electrostatic adsorptive interaction of PFAAs. © 2016 Published by Elsevier Ltd.

Keywords: Perfluoroalkyl acids (PFAAs) Covalent triazine-based framework (CTF) Adsorption Electrostatic interaction Nanosized pore

1. Introduction Perfluoroalkyl substances (PFASs) are a group of synthetic compounds that have been widely used in metal plating, aqueous film forming fire-fighting foams (AFFFs), and surface coatings for over a half century (Carmosini and Lee, 2008; Lindstrom et al., 2011; Prevedouros et al., 2006). PFASs have raised serious environmental concerns because of their high persistence, global distribution, strong bioaccumulation, and potential toxicity (Houde et al., 2011; Lau et al., 2007; Post et al., 2012; Steenland et al., 2010). Residues of PFASs, particularly the two most widely used ones, perfluorooctanoic acid (PFOA) and perfluorooctane sulfonate

* This paper has been recommended for acceptance by Baoshan Xing. * Corresponding author. State Key Laboratory of Pollution Control and Resource Reuse/School of the Environment, Nanjing University, Jiangsu, 210046, China. E-mail address: [email protected] (D. Zhu).

(PFOS), are frequently detected in surface water, groundwater, and even drinking water (Fujii et al., 2007; Hansen et al., 2002; Moody and Field, 1999; Moody et al., 2001). PFOA, PFOS, and other shorter chained perfluoroalkyl acids (PFAAs) are endproducts of microbial ~ o, degradation of several precursor PFASs (Liu and Mejia Avendan 2013) (e.g., PFOS-based and fluoroteleomer-based surfactants in AFFFs) leading to their continual production in AFFF-contaminated groundwater (Place and Field, 2012). PFOA and PFOS concentrations in natural water samples are usually at trace ng/L levels (Hansen et al., 2002); however, in groundwater associated with military sites, concentrations as high as several thousand mg/L are being detected (Moody and Field, 1999; Moody et al., 2003). The U.S. Environmental Protection Agency (EPA) has established a provisional health advisory of 0.4 mg/L for PFOA, and 0.2 mg/L for PFOS in drinking water. It is thus imperative to develop effective treatment technology options for removal of PFAAs. PFAAs are chemically and biologically stable, and extremely

http://dx.doi.org/10.1016/j.envpol.2016.06.062 0269-7491/© 2016 Published by Elsevier Ltd.

Please cite this article in press as: Wang, B., et al., Covalent triazine-based framework: A promising adsorbent for removal of perfluoroalkyl acids from aqueous solution, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.06.062

2

B. Wang et al. / Environmental Pollution xxx (2016) 1e9

persistent, thus are not degraded in conventional water treatment systems that mainly rely on the metabolic function of microorganisms, but instead are generated from precursor PFASs (Sinclair and Kannon, 2006). A variety of treatment technologies based on photochemical (Hori et al., 2004), sonochemical (Moriwaki et al., 2005), electrochemical (Lin et al., 2013) or advanced oxidation process have been proposed to decompose PFAAs (Schroder and Meesters, 2005; Tang et al., 2012), but they are still in the early development stage and currently infeasible in practice. Thus, sorption to high-binding sorbents is currently the most feasible and economical approach to remove PFAAs from aqueous solutions, especially at low solute concentrations. Previous studies on sorption of PFAAs (particularly PFOA and PFOS) have mainly involved soils/sediments (Higgins and Luthy, 2006; Johnson et al., 2007), clay minerals and their surfactant-modified counterparts (Zhang et al., 2014; Zhao et al., 2014), activated carbon and carbon fiber (Yu et al., 2009; Zhi and Liu, 2015), carbon nanotubes (Deng et al., 2012; Zhou et al., 2012) and synthetic organic polymers including anion-exchange resins (Deng et al., 2010; Yu et al., 2009) and molecularly imprinted materials (Karoyo and Wilson, 2013). Sorption to these sorbents is primarily controlled by the specific electrostatic and/or ligand-exchange interactions between the anionic head groups of PFAAs and the corresponding interactive sites of the sorbents, as well as the nonspecific enthalpy-driven hydrophobic effect. Activated carbon has been proposed as an effective adsorbent for PFAAs in treatments of wastewater (Yu et al., 2012), drinking water (Rahman et al., 2014), and possibly groundwater (Oliaei et al., 2013). The relatively strong adsorption of PFOA and PFOS to activated carbon from pure water is attributed to its large specific surface area, high surface hydrophobicity, and high microporosity (Ochoa-Herrera and Sierra-Alvarez, 2008; Yu et al., 2009). However, the presence of municipal wastewater effluentborne dissolved organic matter has been shown to reduce PFOA and PFOS adsorption by almost an order of magnitude (Yu et al., 2009). In addition, the closed, irregular-shaped micropores of activated carbon may cause difficulty in desorption of preadsorbed adsorbates (Ji et al., 2009), thus limiting regeneration of the adsorbent. The covalent triazine-based framework (CTF), as first reported by Kuhn et al. (2008a), is an emerging class of non-swelling polymers synthesized by ionothermal trimerization of aromatic nitriles. Because of the very large specific surface area, homogeneous and rigid pore structure, high chemical and thermal stability, and relatively low cost (Bhunia et al., 2015; Kuhn et al., 2008a), CTF is considered for application in energy gas storage and catalytic support materials (Chan-Thaw et al., 2010; Hug et al., 2014). To date only a few studies have been conducted to investigate adsorption of environmentally relevant organic chemicals to CTF and focused mainly on anionic dyes and surfactants (Bhunia et al., 2015; Liu et al., 2012, 2013). In a recent study, Liu et al. (2012) found that CTF could strongly adsorb sulfonate-substituted monocyclic and bicyclic aromatic compounds because the adsorbate sulfonate group induces strong electrostatic interaction with the protonated triazine structure of CTF. Resulting from the homogeneous nanosized pores (1.2 nm), CTF demonstrated additional advantages of fast adsorption/desorption kinetics and complete adsorption reversibility for the test adsorbate. Thus, it seems reasonable to hypothesize that CTF is a promising adsorbent for the anionic PFAA molecules. However, few relevant studies have been performed thus far. This study investigated the possibility of using CTF as an effective adsorbent to remove perfluoro C4, C6, and C8 carboxylates and sulfonates, including PFOA and PFOS from aqueous solution. The batch adsorption isotherms to CTF were compared with those to three other commercial adsorbents: pulverized microporous

activated carbon, single-walled carbon nanotubes, and anionexchange resin. The effects of pH and dissolved humic acid on adsorption were further examined to probe mechanisms and key factors influencing adsorption. Moreover, desorption of preadsorbed PFOA or PFOS from CTF with pH-adjusted water or waterethanol solutions was measured to assess optimal regeneration potential of the adsorbent. 2. Materials and methods 2.1. Chemicals The adsorbates were perfluorooctanoic acid (PFOA, 98%, J&K Chemical) (Beijing, China), perfluorohexanoic acid (PFHA, 98%, J&K Chemical), perfluorobutyric acid (PFBA, 98%, Sigma-Aldrich) (St. Louis, USA), perfluorooctane sulfonate (PFOS, 98%, J&K Chemical), perfluorohexane sulfonate (PFHS, 98%, J&K Chemical), and perfluorobutane sulfonate (PFBS, 98%, Sigma-Aldrich). Their selected physicochemical properties are listed in Table S1 of the Supporting Information (SI). A soil-extracted humic acid was used to assess its effects on adsorption of PFAAs to different adsorbents. The extraction procedure and detailed structural characterization of the humic acid are described in our previous study (Sun et al., 2008). 2.2. Adsorbents Covalent triazine-based framework (CTF) was synthesized by ionothermal of trimerization of 1,4-dicyanobenzene (see details in SI). Its chemical and schematic structures are presented in Fig. S1. Microporous activated carbon (AC, Calgaon Carbon Co., USA), single-walled carbon nanotubes (SWNT, Nanotech Port Co., China), and anion-exchange resin Amberlite IRA-400 (Resin, Sinopharm Chemical Regent Co., China) were included as comparative adsorbents. They were commonly investigated as high efficiency adsorbents for PFAAs in the literature (Deng et al., 2010, 2012; Yu et al., 2009; Zhi and Liu, 2015; Zhou et al., 2012). The adsorbents of AC, SWNT, and Resin were pretreated to remove impurities according to the literature method (see details in SI). The CTF and AC samples after drying were ground and sieved (100 mesh) to obtain particles sized less than 0.15 mm. The particle size of the bundled aggregates of SWNT was determined as average hydrodynamic diameters by a laser particle size analyzer (Mastersizer 2000, MALVERN, UK), ranging from 400 to 500 nm. Based on the information provided by the manufacturer, the particle sizes ranged from 0.3 to 0.8 mm for Resin. 2.3. Characterization of adsorbents Transmission electron microscope (TEM) images of CTF, AC, and SWNT were collected on a Hitachi H-800 TEM (Hitachi, Japan). Xray diffraction (XRD) patterns of virgin CTF and CTF with preadsorbed PFOA (419, 687 and 1031 mmol/kg, respectively) or PFOS (318, 607 and 1338 mmol/kg, respectively) were collected from a Rigaku D/max-RA powder diffraction-meter (Rigaku, Tokyo, Japan) using Cu Ka radiation. N2 adsorption/desorption isotherms to CTF, AC, and SWNT were obtained on a Micrometrics ASAP 2020 (Micromeritics Instrument Co., Norcross, GA, USA) apparatus at
196  C (77 K) to determine their specific surface areas and pore parameters. To test the surface hydrophobicity/hydrophilicity of adsorbents, sorption/desorption isotherms of water vapor to CTF, AC, SWNT, and Resin were obtained on a DVS Intrinsic apparatus (Surface Measurement Systems Co., London, UK) at 25  C (298 K). The pKa of CTF was measured by the acid-base titration method using a Metrohm 877 Titrino Plus automatic titrator (Switzerland) (Konkena and Vasudevan, 2012).

Please cite this article in press as: Wang, B., et al., Covalent triazine-based framework: A promising adsorbent for removal of perfluoroalkyl acids from aqueous solution, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.06.062

B. Wang et al. / Environmental Pollution xxx (2016) 1e9

2.4. Batch adsorption Multi-concentration adsorption isotherm measurements were performed using 40-mL polypropylene centrifuge tubes equipped with polypropylene-lined screw caps. All isotherms were collected using individual chemicals. The aqueous solution containing the test PFAA at the desired concentration (prepared in MilliQ water, Purelab Ultra, USA) was preadjusted with NaOH to ensure the sample equilibrium pH equal to 6.0. Pre-weighed adsorbent was placed in tubes (15 mg of Resin and 10 mg of other adsorbents) and filled with 40-mL PFAA aqueous solution. Samples were mixed endover-end at room temperature for 3 days, which was long enough to reach apparent adsorption equilibrium according to the predetermined adsorption kinetics (see more details below). Afterwards, the samples were centrifuged at 4000 g for 10 min followed by supernatant sampling for analysis. To take account for solute loss from processes other than adsorbent sorption (i.e., sorption to the screw cap and tube wall), calibration curves were built separately from controls receiving the same treatment and conditions (e.g., concentration ranges of solutes) as the adsorption samples but without adsorbent. Calibration curves included at least 6 concentration levels over the test concentration range. Based on the calibration curves obtained, the adsorbed mass of solute was calculated by subtracting mass in aqueous solution from solute mass added. The initial concentration range was 9e624 mmol/L (3.8e259 mg/L) for PFOA, 111e772 mmol/L (59.9e415 mg/L) for PFOS, 23e695 mmol/ L (7.2e217 mg/L) for PFHA, 30e955 mmol/L (6.5e204 mg/L) for PFBA, 44e410 mmol/L (19.2e180 mg/L) for PFHS, and 18e737 mmol/ L (6e247 mg/L) for PFBS. Isotherm experiments were also conducted to measure adsorption of four low concentrations of PFOA (initially at 6.07, 9.11, 12.11 and 15.22 mg/L) and PFOS (initially at 6.13, 9.13, 12.06 and 15.13 mg/L) to CTF. At least duplicate samples were prepared for each concentration with the lower 4 concentrations having 4 replicates for each concentration. 2.5. pH and humic acid effects To assess the effects of pH, pH envelopes for PFOA and PFOS were constructed for each adsorbent using single-point adsorption data (initially at 0.20 mmol/L for PFOA and 0.24 mmol/L for PFOS) in triplicate at pH values from 4.0 to 11.8 (adjusted with HCl and NaOH). Single-point adsorption data (equilibrium pH equal to 6.0) in triplicate were also used to assess the effects of humic acid (10e40 mg/L). A bulk solution of humic acid was prepared by dissolving 50 mg of humic acid in 5 mL of 0.1 M NaOH followed by mixing with MilliQ water to reach an apparent concentration of 50 mg/L. The humic acid solution was adjusted to pH 6.0 with 0.1 M HCl, followed by filtration through a 0.45-mm membrane. The humic acid solution filtrate was diluted to reach apparent concentrations of 10, 20, 30 and 40 mg/L, respectively. Adsorption of humic acid to CTF was measured in no-PFAA controls using an initial humic acid concentration of 40 mg/L in triplicate and quantifying loss of humic acid to the adsorbent by measuring total organic carbon (TOC) (TOC 5000A, Shimadzu, Japan) before and after adsorption. 2.6. Desorption and regeneration of adsorbents To evaluate the reusability of the adsorbents, adsorption/ desorption of PFOA and PFOS was performed for four consecutive cycles. Each cycle was performed using a single-step, centrifugewithdraw-refill method. The sorption and desorption samples were mixed end-over-end at room temperature for 3 days (desorption kinetics presented in SI). The initial PFOA and PFOS concentrations were 84.3 mmol/L and 174.5 mmol/L, respectively. The supernatant after each cycle was replaced with fresh water or fresh 50/50 v/v

3

water/ethanol solutions. The adsorption experiments were performed at pH 3.4 (adjusted with HCl) and the desorption experiments at pH 11.8 (adjusted with NaOH). In between each cycle, adsorbents were washed with MilliQ water thoroughly as determined by achieving a near neural pH. Negligible amount of PFOA or PFOS was removed in the wash due to the low (< 6) pH of rinsing water (see more details below). All single-point adsorption and desorption experiments were performed in triplicate. 2.7. Analyses of PFAAs Quantifying PFAA concentrations was done using two different analytical methods for the low and high concentration ranges. The aliquots resulting from the four lowest PFOA and PFOS initial concentrations (mg/L range) were analyzed using an Agilent 1100 highperformance liquid chromatography (HPLC) equipped with a Waters BEH C18 column (2.1  50 mm, 1.7 mm in particle size) coupled with a AB Sciex API 4000 tandem mass spectrometer (MS) as previously developed (Ma and Shih, 2010). PFAA concentrations in all other samples were analyzed using a Waters Alliance 2695 HPLC equipped with a Waters 432 conductivity detector and a Waters XBridge-C18 column (4.6  250 mm, 5 mm in particle size). Isocratic elution was performed under the following conditions: 0.02 M ammonium acetate/acetonitrile (45/55 v/v) at 1 mL/min. The mobile phase was sonicated for 30 min and filtered through a 0.22-mm PVDF membrane (Anpel Scientific Instrument, Shanghai, China) prior to use. More details on the quality assurance and quality control (QA/QC) for PFAA analysis are described in SI. 3. Results and discussion 3.1. Characterization of adsorbents The TEM analyses show that the pore sizes of CTF were smaller than that of AC, and SWNT formed bundled aggregates (Fig. S2). The XRD patterns of the CTF with and without preadsorbed PFOA/PFOS are presented in Fig. S3. The peak with 2q at 7.2 is indexed to the (100) diffraction, reflecting the crystalline triazine-based framework with hexagonal pores (Kuhn et al., 2008a). The broad peak at 26.2 is assigned to the (001) diffraction, characteristic of the interlayer aromatic sheets. The information of specific surface area and pore volume of the four adsorbents is listed in Table 1. CTF had a specific surface area of 1270 m2/g, much higher than that of SWNT (468 m2/g) and AC (825 m2/g). The specific surface area of Resin was not determined due to its gel-type polymer matrix. SWNT was dominated by mesopores (2e50 nm in size), while CTF and AC were dominated by micropores (< 2 nm). The pore size distribution profiles (Fig. S4) further indicate that AC had a very wide pore size distribution, with the majority of the pores smaller than 1.0 nm. Alternatively, CTF had a very narrow pore size distribution centered at 1.2 nm, consistent with the XRD results. The adsorption affinity of water vapor negatively correlates with the surface hydropho th et al., bicity of adsorbents (Karanfil and Dastgheib, 2004; To 2012). At a given relative humidity, the water vapor adsorption to Resin and CTF was approximately one order of magnitude higher than that to AC and SWNT (Fig. S5), reflecting the much higher surface hydrophobicity of Resin and CTF. This is reasonable given that the surface of Resin and CTF are highly charged, while the surfaces of AC and SWNT are dominated by highly hydrophobic graphitic carbons. The more hydrophilic nature of CTF is expected to substantially mitigate the hydrodynamic sheer force and lower the energy cost when the material is used as specific sorbent for column cleanup and separation. Within the pH range of 2e10, the surfaces of SWNT and AC were negatively charged, as verified by the zeta potential-pH relationship determined in our previous

Please cite this article in press as: Wang, B., et al., Covalent triazine-based framework: A promising adsorbent for removal of perfluoroalkyl acids from aqueous solution, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.06.062

4

B. Wang et al. / Environmental Pollution xxx (2016) 1e9

Table 1 Elemental compositions (dry weight-based), surface area, and pore volume parameters for different adsorbents. Adsorbent

C%

H%

N%

Surface areaa (m2/g)

Vmicb (cm3/g)

Vmesc (cm3/g)

Vtd (cm3/g)

CTF AC SWNT Resin

74.99e 85.27 91.61 60.24

3.15e 1.82 0.93 8.67

21.86e 0.46 0.10 4.75

1270 825 468 NAf

0.44 0.35 0.15 NA

0.19 0.19 0.37 NA

0.63 0.54 0.52 NA

a b c d e f

Determined by N2 adsorption using the Brunauer-Emmett-Teller (BET) method. Micropore volume, calculated using the DubinineRadushkevich method. Mesopore volume, calculated by Vt
Vmicro. Total pore volume, determined at P/P0 ¼ 0.97. Calculated from the chemical structure of CTF, (C8H4N2)n (Bhunia et al., 2015). Not available as the N2 adsorption/desorption method is not applicable to a gel-type resin.

study (Ji et al., 2010). Due to the low water suspensibility of CTF and the hygroscopic expansion nature of Resin, zeta potentials of these two adsorbents were not measured. The triazine groups of CTF had a pKa of 6.5 (obtained from the acid-base titration curve, results presented in Fig. S6). 3.2. Adsorption kinetics Fig. 1 compares adsorption kinetics of PFOA and PFOS among different adsorbents, plotted as adsorbed-phase concentration (qt, mmol/kg) as a function of time (t, h). At a given time, the adsorbed concentration to CTF and AC was consistently higher than that to SWNT and Resin. In contrast to CTF and AC with fixed pore structures, SWNT contained mainly non-fixed porous interstices originated from the coagulation of individual carbon nanotubes, which might impede the diffusion of bulky, flexible molecules, leading to a prolonged course of adsorption (Ji et al., 2010). The slow adsorption kinetics to Resin can be attributed to the hindered diffusion of the flexible PFAA molecules within the gel-type polymer matrix. 3.3. Adsorption isotherms Fig. 2 compares adsorption isotherms of PFOA and PFOS among different adsorbents, plotted as adsorbed-phase concentration (q, mmol/kg) vs. aqueous-phase concentration (Ce, mmol/L) at equilibrium. Fig. 3 compares adsorption isotherms of PFAAs of varying chain length to CTF and AC. All isotherms with the possible exception of PFOA adsorption to Resin are nonlinear; therefore, the adsorption data were fitted with two commonly used nonlinear

sorption models: Freundlich model, q ¼ KFCne , where KF (mmol1
n Ln/kg) is the Freundlich affinity coefficient and n (unitless) is the KL Ce Freundlich linearity index; Langmuir model, q ¼ 1þK Qmax , where L Ce KL (L/kg) is the Langmuir affinity coefficient and Qmax (mmol/kg) is the maximum adsorption capacity. The model fitting parameters are summarized in Table 2 along with the adsorption distribution coefficients (Kd) calculated at selected Ce values. Both models generally fit the sorption data well (as indicated by the R2 values). The Langmuir model facilitates an estimation of maximum adsorption capacity, which is useful when assessing adsorbent for contaminant removal. Additionally, from mechanistic point of view the Langmuir model describes adsorption to limited, homogeneous sites, which is particularly suitable for CTF. Among the four test adsorbents, CTF exhibited the highest adsorption affinities and Qmax for both PFOA and PFOS. However, PFOS had the smallest affinity to Resin whereas PFOA adsorption to Resin fell between AC and SWNT (Fig. 2 and Table 2). The lower PFOS adsorption to Resin compared to PFOA is likely caused by the stronger steric hindrance of the larger-sized sulfonate group when the adsorbate molecule diffuses within the gel-type polymer matrix to approach the charged quaternary ammonium group (Yu et al., 2009). Within the concentration range examined (Ce ¼ 1.3E
04e182 mg/L for PFOA; Ce ¼ 1.0E
03e230 mg/L for PFOS, combining isotherm data presented in Fig. 2 and the low-concentration isotherm data), the measured Kd to CTF is in the order of 103e105 L/kg for PFOA and PFOS. Likewise, the adsorption affinity and capacity of the shorter chained PFAAs to CTF are both higher than the corresponding values to AC (Fig. 3), which is currently all that has been used in practice. The adsorption isotherm data demonstrate that CTF is a

CTF AC SWNT Resin

qt (mmol/kg)

qt (mmol/kg)

300

800

(a)

200

100

(b)

600 400 200 0

0 0

20

40 t (h)

60

80

0

20

40

60

80

t (h)

Fig. 1. Batch adsorption kinetics expressed as adsorbed concentration (qt, mmol/kg) at a given time (t, h) to different adsorbents for (a) PFOA and (b) PFOS.

Please cite this article in press as: Wang, B., et al., Covalent triazine-based framework: A promising adsorbent for removal of perfluoroalkyl acids from aqueous solution, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.06.062

B. Wang et al. / Environmental Pollution xxx (2016) 1e9

5

CTF AC SWNT Resin 1400

(a)

1200

600

q (mmol/kg)

q (mmol/kg)

800

400 200

(b)

1000 800 600 400 200

0 0.0

.1

.2

.3

.4

0 0.0

.5

.1

Ce (mmol/L)

.2

.3

.4

.5

Ce (mmol/L)

Fig. 2. Adsorption isotherms plotted as adsorbed-phase concentration (q) vs. aqueous-phase concentration (Ce) at equilibrium to different adsorbents for (a) PFOA and (b) PFOS.

PFBA/CTF PFHA/CTF PFOA/CTF PFBA/AC PFHA/AC PFOA/AC

1400

(a) PFCAs

600

1200

500

1000

q (mmol/kg)

q (mmol/kg)

700

PFBS/CTF PFHS/CTF PFOS/CTF PFBS/AC PFHS/AC PFOS/AC

400 300 200 100 0 0.0

(b) PFSAs

800 600 400 200

.2

.4

.6

.8

Ce (mmol/L)

0 0.0

.2

.4

.6

.8

Ce (mmol/L)

Fig. 3. Comparison of adsorption isotherms to CTF and AC plotted as adsorbed-phase concentration (q) vs. aqueous-phase concentration (Ce) at equilibrium for (a) Perfluorocarboxylic acids (PFCAs) and (b) Perfluorosulfonic acids (PFSAs).

promising adsorbent for PFAAs in terms of both affinity and capacity, as compared with other test adsorbents. 3.4. pH effects on PFOA and PFOS adsorption The pH envelopes constructed for PFOA and PFOS adsorption to the different adsorbents are presented in Fig. 4. Trends with pH varied for adsorbents as well as between PFAAs. For both PFOA and PFOS, the adsorption to CTF decreased pronouncedly with increasing pH with the most substantial decrease observed for PFOA adsorption, which decreased by an order of magnitude from pH 4.0 to 11.8. A similar decrease in adsorption of PFOS to CTF occurred but not until pH~10. The pH dependence of adsorption to AC, SWNT, and Resin was much less prominent with negligible effects (e.g., Resin) to no more than a factor of 3 decrease with increasing pH. The very strong pH dependence on adsorption to CTF indicates the importance of electrostatic interactions. With increasing pH, the degree of protonation of the CTF triazine

groups decreases, thereby reducing available cationic sites for adsorption of PFAAs. The decreased adsorption with pH observed for AC and SWNT can also be attributed to electrostatic interactions, but in this case it is enhanced electrostatic repulsive force resulting from increasing negative charges on the adsorbent surface (Ji et al., 2010). Likewise, the quaternary ammonium groups in Resin have a permanent positive charge, thus no pHdependency is expected. The reason why PFOA and PFOS showed different patterns of pH dependent adsorption was not very clear. It was probably due to the difference in charge and size of adsorbate molecules and/or different degrees of interlayer spacing increase of CTF upon adsorption of adsorbate molecules (see more details below). 3.5. Humic acid effects on PFOA and PFOS adsorption PFOA and PFOS adsorption with increasing humic acid concentration relative to the absence of humic acid is summarized in Fig. 5.

Please cite this article in press as: Wang, B., et al., Covalent triazine-based framework: A promising adsorbent for removal of perfluoroalkyl acids from aqueous solution, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.06.062

6

B. Wang et al. / Environmental Pollution xxx (2016) 1e9

Table 2 Freundlich model parameters (KF and n ± standard deviation), Langmuir model parameters (KL and Qmax ± standard deviation), adsorption coefficients (Kd) calculated at given equilibrium aqueous-phase concentrations (Ce) and measured Ce ranges for adsorption of perfluoroalkyl acids (PFAAs) to different adsorbents. PFAA

Adsorbent

KF (mmol1
n L/kg)

n

R2

KL (L/kg)

Qmax (mmol/kg)

R2

Kd,0.01a (L/kg)

Kd,

PFOA

CTF

850 ± 50

0.30 ± 0.03

0.935

27 ± 5

650 ± 30

0.920

2.1Eþ04

PFOS

AC SWNT Resin CTF

620 ± 30 280 ± 30 1270 ± 70 1670 ± 90

0.40 0.49 0.94 0.37

± ± ± ±

0.03 0.06 0.06 0.03

0.970 0.920 0.983 0.950

21 ± 2 12 ± 3 3.8 ± 0.2 15 ± 3

470 ± 10 190 ± 20 800 ± 20 1330 ± 80

0.985 0.940 0.983 0.922

AC SWNT Resin CTF AC CTF AC CTF AC CTF AC

1150 ± 30 1220 ± 50 350 ± 30 820 ± 60 520 ± 30 760 ± 90 490 ± 30 430 ± 20 240 ± 20 430 ± 20 230 ± 20

0.33 0.57 0.41 0.67 0.69 0.58 0.66 0.24 0.48 0.46 0.32

± ± ± ± ± ± ± ± ± ± ±

0.01 0.03 0.05 0.07 0.05 0.06 0.04 0.03 0.05 0.04 0.04

0.985 0.983 0.859 0.931 0.973 0.923 0.964 0.883 0.928 0.940 0.920

13 ± 1 4.0 ± 0.5 6±2 1.6 ± 0.7 1.5 ± 0.7 6±1 2.6 ± 0.5 12 ± 2 5±1 4.2 ± 0.8 2.1 ± 0.5

960 ± 20 1120 ± 80 330 ± 40 1200 ± 300 750 ± 230 560 ± 70 500 ± 50 430 ± 20 240 ± 20 470 ± 40 170 ± 10

0.979 0.975 0.793 0.902 0.950 0.943 0.974 0.919 0.954 0.934 0.860

PFHA PFHS PFBA PFBS

Kd,0.3c (L/kg)

Ce ranged (mg/L)

7.0Eþ03

2.0Eþ03

9.9Eþ03 2.9Eþ03 1.7Eþ03 3.0Eþ04

3.8Eþ03 1.3Eþ03 1.5Eþ03 1.1Eþ04

1.3Eþ03 5.2Eþ02 1.4Eþ03 3.6Eþ03

2.5Eþ04 8.8Eþ03 5.3Eþ03 3.7Eþ03 2.2Eþ03 5.3Eþ03 2.3Eþ03 1.4Eþ04 2.6Eþ03 5.2Eþ03 5.2Eþ03

8.6Eþ03 4.4Eþ03 2.0Eþ03 2.2Eþ03 1.3Eþ03 2.7Eþ03 1.4Eþ03 4.2Eþ03 1.1Eþ03 2.2Eþ03 1.8Eþ03

2.6Eþ03 2.0Eþ03 7.1Eþ02 1.2Eþ03 7.6Eþ02 1.3Eþ03 7.4Eþ02 1.1Eþ03 4.5Eþ02 8.2Eþ02 5.2Eþ02

4.9e182 1.3E
04e4.0E
03e 1.7e173 2.5e127 4.9e227 12e230 1.0E
03e3.4E
03e 28e207 35e189 49e246 17e172 4e144 11e125 37e155 14e183 5e145 25e220 2e130

0.05

b

(L/kg)

a,b,c

Concentration specific sorption coefficients (Kd,C) calculated from the Freundlcih model adsorption parameters: (Kd ¼ KFCn
1 ). e At Ce ¼ 0.01 mmol/L. b At Ce ¼ 0.05 mmol/L. c At Ce ¼ 0.3 mmol/L. d Measured in high-concentration isotherm experiments except where noted. e Measured in low-concentration isotherm experiments. a

CTF AC SWNT Resin

12000

(a)

5000

10000

4000

8000

Kd (L/kg)

Kd ( L/ k g)

6000

3000 2000

(b)

6000 4000 2000

1000 0

0 4

5

6

7

8

9

10 11 12

pH

4

5

6

7

8

9

10 11 12

pH

Fig. 4. Single-point adsorption coefficients as a function of pH on different sorbents for (a) PFOA and (b) PFOS. Error bars, in some cases smaller than the symbols, represent standard deviations calculated from three replicates.

For both AC and SWNT, humic acid greatly suppressed PFOA and PFOS adsorption with 31e48% and 95% (as measured by the decrease of single-point adsorption coefficient, Kd), respectively, at the highest humic acid concentration (40 mg/L) investigated. It is well recognized that dissolved humic acids can foul adsorption of organic chemicals to microporous activated carbon through direct competition for adsorption sites and pore blockage (Kilduff et al., 1996). Likewise, relatively strong adsorption was observed for humic acid to AC and SWNT (Kd ¼ 2070 ± 70 L/kg for AC, and Kd ¼ 3000 ± 400 L/kg for SWNT). In contrast, the effect of humic acid on adsorption to CTF was not significant with only a decrease in sorption of 13% for PFOA and essentially no measureable change for PFOS. The molecular sizes are approximately 0.8 nm (length)  0.36 nm (width) for PFOA and 1.04 nm  0.40 nm for PFOS (calculated in the gas phase by Chem 3D). Therefore, the PFOA and PFOS molecules would be able to enter the homogeneous,

nanosized (1.2 nm) internal pores of CTF readily. On the other hand, adsorption of humic acid to CTF was found to be very low (Kd ¼ 20 ± 3 L/kg). Lack of humic acid interaction with CTF is most likely due to size-exclusion in that the large-sized humic acid molecules are blocked from entering the nanosized internal pores of CTF. The negligible interference of humic acid to CTF adsorption further reflects on the superior nature of this adsorbent for PFAAs given that humic materials are commonly present in natural waters and industrial/wastewater effluent. Surprisingly, the presence of humic acid slightly increased adsorption of PFOA and PFOS to Resin by 36% and 12%, respectively. This unexpected observation is possibly due to the adsorbed humic acid (Kd ¼ 400 ± 100 L/kg) causing a slight expansion of the Resin’s gel-type polymer matrix, thereby increasing PFOA and PFOS access to more adsorption sites. The larger increase of the slightly larger PFOS molecules compared to the PFOA molecules supports this hypothesis.

Please cite this article in press as: Wang, B., et al., Covalent triazine-based framework: A promising adsorbent for removal of perfluoroalkyl acids from aqueous solution, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.06.062

B. Wang et al. / Environmental Pollution xxx (2016) 1e9

7

1.4 (a) 1.2

1.4 (b) 1.2

1.0

1.0

Kd/Kd0

Kd/Kd0

CTF AC SWNT Resin

.8 .6 .4

.8 .6 .4

.2

.2

0.0

0.0 0

10

20

30

40

0

Humic acid (mg/L)

10

20

30

40

Humic acid (mg/L)

Fig. 5. Effect of increasing humic acid concentrations on the adsorption ability of different sorbents for (a) PFOA and (b) PFOS. The y-axis is adsorption in presence of humic acid (Kd, L/kg) relative to adsorption in the absence of humic acid (Kd0, L/kg). Error bars, in some cases smaller than the symbols, represent standard deviations calculated from three replicates.

3.6. Desorption and adsorbent regeneration

Adsorption/Desorption ratio (%)

Fig. 6 displays the percentage of PFOA and PFOS adsorbed form pH 3.4 water and desorbed by either pH 11.8 water (Fig. 6a and b) or a pH 11.8 50/50 v/v water/ethanol solution (Fig. 6c and d) to virgin and regenerated CTF in four consecutive cycles. Based on what was learned from the pH envelopes, pH was used to optimize adsorption (pH ¼ 3.4) and subsequent desorption (pH ¼ 11.8) by CTF. The

100 95 (a) 90 85 80 75 70 65 60 55 50

percentage adsorbed by CTF decreased by only 12% for PFOA and 11% for PFOS over the four adsorption-regeneration cycles, whereas the percentage desorbed of 80e90% remained nearly constant between cycles. Much lower percentages of PFOA and PFOS could be desorbed from AC (approximately 40%) and from SWNT (approximately 50%) (see Fig. S8). Desorption of PFOA and PFOS from CTF was further increased to approximately 95% using high pH aqueous ethanolic solutions (Fig. 6c and d). The pH-sensitive desorption

100 95 (b) 90 85 80 75 70 65 60 55 50 cycle 1 cycle 2 cycle 3 cycle 4

cycle 1 cycle 2 cycle 3 cycle 4 100 95 (d) 90 85 80 75 70 65 60 55 50

100 (c) 95 90 85 80 75 70 65 60 55 50 cycle 1 cycle 2 cycle 3 cycle 4

cycle 1 cycle 2 cycle 3 cycle 4

Fig. 6. Adsorption and desorption ratios of PFOA/PFOS to virgin and regenerated CTF in four consecutive cycles. Adsorption from water at pH 3.4 followed by desorption in water at pH 11.8 for (a) PFOA and (b) PFOS. Adsorption from water at pH 3.4 followed by desorption in 50/50 v/v water/ethanol solution at pH 11.8 for (c) PFOA and (d) PFOS. Black bars are for the percentages adsorbed of the adsorbate added in each cycle and the gray bars are the percentages desorbed of the total amount of adsorbate adsorbed. Error bars represent standard deviations calculated from four replicates.

Please cite this article in press as: Wang, B., et al., Covalent triazine-based framework: A promising adsorbent for removal of perfluoroalkyl acids from aqueous solution, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.06.062

8

B. Wang et al. / Environmental Pollution xxx (2016) 1e9

property of CTF provides another benefit over other sorbents such as activated carbon requiring pyrolysis in regeneration, which could be more costly and lead to liquid and solid secondary waste streams. 3.7. Adsorption mechanisms PFAAs contain a negatively charged hydrophilic head group and a hydrophobic-oleophobic perfluoroalkyl chain. Accordingly, a variety of mechanisms might be involved in adsorption of PFAAs in response to the surface properties (such as charge and hydrophobicity) of adsorbents. The surfaces of AC and SWNT are dominated by graphitic carbons, which are highly hydrophobic and have large electronic polarizability. Adsorption of PFAAs to AC and SWNT is expected to be mainly driven by hydrophobic effect, which is a combination of entropic gradient and van der Waals (mainly dispersion) interaction between adsorbate and adsorbent, whereas electrostatic force plays only a minimal role. Compared with apolar hydrophobic organic chemicals such as polycyclic aromatic hydrocarbons (PAHs), the relatively low adsorption affinity of AC and SWNT for PFOA and PFOS is likely due to the low electronic polarizability of these molecules (Uneyama, 2006), thus lessoning potential van der Waals interaction in spite of the large electronic polarizability of graphitic carbons. Electrostatic interaction is proposed to be the main driving force for adsorption of PFAAs to CTF and Resin. The triazine groups of CTF have a pKa of 6.5, and therefore are partially (76%) protonated at the test pH 6.0. The negatively charged head group in PFAAs can interact strongly with the positively charged protonated triazine groups of CTF or the permanently charged quaternary ammonium groups of Resin. The electrostatic mechanism was explored in previous studies to explain the strong adsorption of anionic dyes and surfactants to CTF materials at neutral or slightly acidic pH conditions (Kuhn et al., 2008b; Bhunia et al., 2015). The role of hydrophobic effect in adsorption to CTF is reflected by the chain length dependency on adsorption (see the n-octanol-water partition coefficients and water solubilities in Table S1). Compared with electrostatic interaction, hydrophobic effect is expected to be less important in adsorption to CTF. This is evidenced by the fact that despite the much lower surface hydrophobicity (as reflected by the water vapor adsorption data, see Fig. S5), CTF exhibits stronger adsorption of PFOA and PFOS than does AC and SWNT. Electrostatic interactions alone, however, cannot fully account for the strongest adsorption observed for PFAAs to CTF, otherwise the sorption affinities and capacities of PFOA and PFOS to Resin would have been comparable with that to CTF. Resin has a higher anion exchange capacity (> 3000 mmol/kg according to the manufacturer) compared to CTF with a theoretical anion exchange capacity of CTF (3800 mmol/kg), which is equivalent to the quantity of protonated N atoms in the triazine groups at pH 6. It was proposed in our previous study that the benzene rings adjacent to the triazine group make the surrounding microenvironment partially dehydrated and somewhat hydrophobic (Liu et al., 2012). This in turn effectively alleviates the “desolvation penalty” encountered in the break of the hydration shells around the protonated N atom in the triazine group and the anionic functional group of the solute when they form electrostatic complexes. In a similar fashion, the electrostatic interaction of PFAAs with CTF could also be facilitated. Bhunia et al. (2015) found that alkylbenzenesulfonate surfactants could penetrate into the interlayer space of CTF, as evidenced by an increase in the interlayer spacing via XRD after adsorption. The XRD patterns of CTF at different loading levels of PFOA/PFOS are compared in Fig. S3, along with that of the non-adsorbed control sample. The peak at 26.2 characteristic of the (001) diffraction was shifted to 18.2 after adsorption of PFOA/PFOS, suggesting an

increase of interlayer distance. The interlayer distance progressively increased from 0.342 nm to 0.352, 0.513 and 0.516 nm upon adsorption of PFOA at 419, 687 and 1031 mmol/kg, and from 0.342 nm to 0.354, 0.497 and 0.525 nm upon adsorption of PFOS at 318, 607 and 1338 mmol/kg. Apparently, the interlayer space of CTF could only be opened up and accessible to adsorbate molecules at a relatively high loading level of PFOA/PFOS, with a threshold of 300e400 mmol/kg. Additionally, the original (001) diffraction peak at 26.2 never disappeared despite the attenuated intensity by adsorption of PFOA/PFOS. Thus, it is derivable that not all interlayer spaces of CTF are opened up upon adsorption of PFOA/PFOS, and adsorbate molecules tend to enter the interlayer spaces that have already been opened by previously adsorbed PFOA/PFOS molecules (conditioning effects). Nonetheless, the abovementioned mechanism for alleviating the “desolvation penalty” in adsorption could be further favored by the narrow hydrophobic interlayer spaces (less than 0.6 nm) of CTF where the PFAA molecules are embedded. Resin would not be able to take advantage of this “desolvation penalty”-relieving mechanism because the quaternary ammonium groups within the polymer matrix are expected to be fully saturated with water molecules stemming from the gel-type cationic nature. 4. Conclusion CTF has unique characteristics of rigid, homogeneous nanopore structures with pH-dependent adsorption activity ascribed to the protonation-deprotonation transition of the triazine groups. CTF has exhibited promising potential for being an effective adsorbent for PFAAs in water and wastewater treatment with its strong adsorption affinity and very high adsorption capacity. The mechanisms of CTF adsorption yields an adsorbent that can be easily regenerated and is not fouled by the presence of humic acids, which far exceeds other currently used sorbent technologies. Further research is warranted to investigate the sorption properties of CTF in fixed-bed mode which is more analogous to practical water treatment. Acknowledgments We thank Mr. Nanyang Yu for the assistance with the HPLC-MS analysis. This work was supported by the National Key Basic Research Program of China (Grant 2014CB441103) and the National Natural Science Foundation of China (Grants 21237002, 21225729 and 21428701). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.envpol.2016.06.062 References Bhunia, A., Dey, S., Bous, M., Zhang, C.Y., von Rybinski, W., Janiak, C., 2015. High adsorptive properties of covalent triazine-based frameworks (CTFs) for surfactants from aqueous solution. Chem. Commun. 51, 484e486. Carmosini, N., Lee, L.S., 2008. Partitioning of fluorotelomer alcohols to octanol and different sources of dissolved organic carbon. Environ. Sci. Technol. 42, 6559e6565. Chan-Thaw, C.E., Villa, A., Katekomol, P., Su, D.S., Thomas, A., Prati, L., 2010. Covalent triazine framework as catalytic support for liquid phase reaction. Nano. Lett. 10, 537e541. Deng, S.B., Yu, Q.A., Huang, J., Yu, G., 2010. Removal of perfluorooctane sulfonate from wastewater by anion exchange resins: effects of resin properties and solution chemistry. Water Res. 44, 5188e5195. Deng, S.B., Zhang, Q.Y., Nie, Y., Wei, H.R., Wang, B., Huang, J., Yu, G., Xing, B.S., 2012. Sorption mechanisms of perfluorinated compounds on carbon nanotubes. Environ. Pollut. 168, 138e144.

Please cite this article in press as: Wang, B., et al., Covalent triazine-based framework: A promising adsorbent for removal of perfluoroalkyl acids from aqueous solution, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.06.062

B. Wang et al. / Environmental Pollution xxx (2016) 1e9 Fujii, S., Polprasert, C., Tanaka, S., Lien, N.P.H., Qiu, Y., 2007. New POPs in the water environment: distribution, bioaccumulation and treatment of perfluorinated compoundsea review paper. J. Water Supply Res. Technol. Aqua 56, 313e326. Hansen, K.J., Johnson, H.O., Eldridge, J.S., Butenhoff, J.L., Dick, L.A., 2002. Quantitative characterization of trace levels of PFOS and PFOA in the tennessee river. Environ. Sci. Technol. 36, 1681e1685. Higgins, C.P., Luthy, R.G., 2006. Sorption of perfluorinated surfactants on sediments. Environ. Sci. Technol. 40, 7251e7256. Hori, H., Hayakawa, E., Einaga, H., Kutsuna, S., Koike, K., Ibusuki, T., Kiatagawa, H., Arakawa, R., 2004. Decomposition of environmentally persistent perfluorooctanoic acid in water by photochemical approaches. Environ. Sci. Technol. 38, 6118e6124. Houde, M., De Silva, A.O., Muir, D.C.G., Letcher, R.J., 2011. Monitoring of perfluorinated compounds in aquatic biota: an updated review. Environ. Sci. Technol. 45, 7962e7973. Hug, S., Mesch, M.B., Oh, H., Popp, N., Hirscher, M., Senker, J., Lotsch, B.V., 2014. A fluorene based covalent triazine framework with high CO2 and H2 capture and storage capacities. J. Mater. Chem. A 2, 5928e5936. Ji, L.L., Liu, F.L., Xu, Z.Y., Zheng, S.R., Zhu, D.Q., 2009. Zeolite-templated microporous carbon as a superior adsorbent for removal of monoaromatic compounds from aqueous solution. Environ. Sci. Technol. 43, 7870e7876. Ji, L.L., Liu, F.L., Xu, Z.Y., Zheng, S.R., Zhu, D.Q., 2010. Adsorption of pharmaceutical antibiotics on template-synthesized ordered micro- and mesoporous carbons. Environ. Sci. Technol. 44, 3116e3122. Johnson, R.L., Anschutz, A.J., Smolen, J.M., Simcik, M.F., Penn, R.L., 2007. The adsorption of perfluorooctane sulfonate onto sand, clay, and iron oxide surfaces. J. Chem. Eng. Data 52, 1165e1170. Karanfil, T., Dastgheib, S.A., 2004. Trichloroethylene adsorption by fibrous and granular activated carbons: aqueous phase, gas phase, and water vapor adsorption studies. Environ. Sci. Technol. 38, 5834e5841. Karoyo, A.H., Wilson, L.D., 2013. Tunable macromolecular-based materials for the adsorption of perfluorooctanoic and octanoic acid anions. J. Colloid. Interface. Sci. 402, 196e203. Kilduff, J.E., Karanfil, T., Chin, Y.P., Weber, W.J., 1996. Adsorption of natural organic polyelectrolytes by activated carbon: a size-exclusion chromatography study. Environ. Sci. Technol. 30, 1336e1343. Konkena, B., Vasudevan, S., 2012. Understanding aqueous dispersibility of graphene oxide and reduced graphene oxide through pKa measurements. J. Phys. Chem. Lett. 3, 867e872. Kuhn, P., Antonietti, M., Thomas, A., 2008a. Porous, covalent triazine-based frameworks prepared by ionothermal synthesis. Angew. Chem. Int. Ed. 47, 3450e3453. Kuhn, P., Krüger, K., Thomas, A., Antonietti, M., 2008b. “Everything is surface”: tunable polymer organic frameworks with ultrahigh dye sorption capacity. Chem. Commun. 5815e5817. Lau, C., Anitole, K., Hodes, C., Lai, D., Pfahles-Hutchens, A., Seed, J., 2007. Perfluoroalkyl acids: a review of monitoring and toxicological findings. Toxicol. Sci. 99, 366e394. Lin, H., Niu, J.F., Xu, J.L., Huang, H.O., Li, D., Yue, Z.H., Feng, C.H., 2013. Highly efficient and mild electrochemical mineralization of long-chain perfluorocarboxylic acids (C9-C10) by Ti/SnO2-Sb-Ce, Ti/SnO2-Sb/Ce-PbO2, and Ti/BDD electrodes. Environ. Sci. Technol. 47, 13039e13046. Lindstrom, A.B., Strynar, M.J., Libelo, E.L., 2011. Polyfluorinated compounds: past, present, and future. Environ. Sci. Technol. 45, 7954e7961. ~ o, S.M., 2013. Microbial degradation of polyfluoroalkyl chemicals Liu, J.X., Avendan in the environment: a review. Environ. Int. 61, 98e114. Liu, J.L., Chen, H., Zheng, S.R., Xu, Z.Y., 2013. Adsorption of 4,4’-(Propane-2,2-diyl) diphenol from aqueous solution by a covalent triazine-based framework. J. Chem. Eng. Data 58, 3557e3562. Liu, J.L., Zong, E.M., Fu, H.Y., Zheng, S.R., Xu, Z.Y., Zhu, D.Q., 2012. Adsorption of aromatic compounds on porous covalent triazine-based framework. J. Colloid. Interf. Sci. 372, 99e107. Ma, R.W., Shih, K., 2010. Perfluorochemicals in wastewater treatment plants and sediments in Hong Kong. Environ. Pollut. 158, 1354e1362. Moody, C.A., Field, J.A., 1999. Determination of perfluorocarboxylates in groundwater impacted by fire-fighting activity. Environ. Sci. Technol. 33, 2800e2806.

9

Moody, C.A., Kwan, W.C., Martin, J.W., Muir, D.C.G., Mabury, S.A., 2001. Determination of perfluorinated surfactants in surface water samples by two independent analytical techniques: liquid chromatography/tandem mass spectrometry and 19F NMR. Anal. Chem. 73, 2200e2206. Moody, C.A., Hebert, G.N., Strauss, S.H., Field, J.A., 2003. Occurrence and persistence of perfluorooctanesulfonate and other perfluorinated surfactants in groundwater at a fire-training area at Wurtsmith Air Force Base, Michigan. USA. J. Environ. Monit. 341e345. Moriwaki, H., Takagi, Y., Tanaka, M., Tsuruho, K., Okitsu, K., Maeda, Y., 2005. Sonochemical decomposition of perfluorooctane sulfonate and perfluorooctanoic acid. Environ. Sci. Technol. 39, 3388e3392. Ochoa-Herrera, V., Sierra-Alvarez, R., 2008. Removal of perfluorinated surfactants by sorption onto granular activated carbon, zeolite and sludge. Chemosphere 72, 1588e1593. Oliaei, F., Kriens, D., Weber, R., Watson, A., 2013. PFOS and PFC releases and associated pollution from a PFC production plant in Minnesota (USA). Environ. Sci. Pollut. Res. 20, 1977e1992. Place, B.J., Field, J.A., 2012. Identification of novel fluorochemicals in aqueous filmforming foams (AFFF) used by the US military. Environ. Sci. Technol. 46, 7120e7127. Post, G.B., Cohn, P.D., Cooper, K.R., 2012. Perfluorooctanoic acid (PFOA), an emerging drinking water contaminant: a critical review of recent literature. Environ. Res. 116, 93e117. Prevedouros, K., Cousins, I.T., Buck, R.C., Korzeniowski, S.H., 2006. Sources, fate and transport of perfluorocarboxylates. Environ. Sci. Technol. 40, 32e44. Rahman, M.F., Peldszus, S., Anderson, W.B., 2014. Behaviour and fate of perfluoroalkyl and polyfluoroalkyl substances (PFASs) in drinking water treatment: a review. Water Res. 50, 318e340. Schroder, H.F., Meesters, R.J.W., 2005. Stability of fluorinated surfactants in advanced oxidation processes-a follow up of degradation products using flow injection-mass spectrometry, liquid chromatography-mass spectrometry and liquid chromatography-multiple stage mass spectrometry. J. Chromatogr. A 1082, 110e119. Sinclair, E., Kannon, K., 2006. Mass loading and fate of perfluoroalkyl surfactants in wastewater treatment plants. Environ. Sci. Technol. 40, 1408e1414. Steenland, K., Fletcher, T., Savitz, D.A., 2010. Epidemiologic evidence on the health effects of perfluorooctanoic acid (PFOA). Environ. Health. Perspect. 118, 1100e1108. Sun, H.Y., Zhu, D.Q., Mao, J.D., 2008. Sorption of polar and nonpolar aromatic compounds to two humic acids with varied structural heterogeneity. Environ. Toxicol. Chem. 27, 2449e2456. Tang, H.Q., Xiang, Q.Q., Lei, M., Yan, J.C., Zhu, L.H., Zou, J., 2012. Efficient degradation of perfluorooctanoic acid by uv-fenton process. Chem. Eng. J. 184, 156e162. th, A., Voitko, K.V., Bakalinska, O., Prykhod’ko, G.P., Berto ti, I., Martínez-Alonso, A., To  n, J.M.D., Gun’ko, V.M., L , K., 2012. Morphology and adsorption Tasco aszlo properties of chemically modified MWCNT probed by nitrogen, n-propane and water vapor. Carbon 50, 577e585. Uneyama, K., 2006. Organofluorine Chemistry. Blackwell, Oxford, U.K. Yu, J., Lv, L., Lan, P., Zhang, S.J., Pan, B.C., Zhang, W.M., 2012. Effect of effluent organic matter on the adsorption of perfluorinated compounds onto activated carbon. J. Hazard. Mater 225, 99e106. Yu, Q., Zhang, R.Q., Deng, S.B., Huang, J., Yu, G., 2009. Sorption of perfluorooctane sulfonate and perfluorooctanoate on activated carbons and resin: kinetic and isotherm study. Water Res. 43, 1150e1158. Zhang, R.M., Yan, W., Jing, C.Y., 2014. Mechanistic study of PFOS adsorption on kaolinite and montmorillonite. Colloid. Surf. A 462, 252e258. Zhao, L.X., Bian, J.N., Zhang, Y.H., Zhu, L.Y., Liu, Z.T., 2014. Comparison of the sorption behaviors and mechanisms of perfluorosulfonates and perfluorocarboxylic acids on three kinds of clay minerals. Chemosphere 114, 51e58. Zhi, Y., Liu, J.X., 2015. Adsorption of perfluoroalkyl acids by carbonaceous adsorbents: effect of carbon surface chemistry. Environ. Pollut. 202, 168e176. Zhou, Y.P., Wen, B., Pei, Z.G., Chen, G.C., Lv, J.T., Fang, J., Shan, X.Q., Zhang, S.Z., 2012. Coadsorption of copper and perfluorooctane sulfonate onto multi-walled carbon nanotubes. Chem. Eng. J. 203, 148e157.

Please cite this article in press as: Wang, B., et al., Covalent triazine-based framework: A promising adsorbent for removal of perfluoroalkyl acids from aqueous solution, Environmental Pollution (2016), http://dx.doi.org/10.1016/j.envpol.2016.06.062