Theoretical studies of perfluorochemicals (PFCs) adsorption mechanism on the carbonaceous surface

Chemosphere 235 (2019) 606e615

Contents lists available at ScienceDirect

Chemosphere journal homepage: www.elsevier.com/locate/chemosphere

Theoretical studies of perfluorochemicals (PFCs) adsorption mechanism on the carbonaceous surface Mu Li, Feiyun Sun, Wentao Shang, Xiaolei Zhang*, Wenyi Dong, Tongzhou Liu, Wen Pang Harbin Institute of Technology (Shenzhen), Shenzhen Key Laboratory of Water Resource Utilization and Environmental Pollution Control, Shenzhen, 518055, China

h i g h l i g h t s
Binding mechanism of PFCs onto PAC was studied by using DFT.
PFCs adsorption involves electrostatic and hydrophobic interactions.
Chain length of PFBS and PFOS impacts on adsorption paths.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 22 May 2019 Received in revised form 21 June 2019 Accepted 25 June 2019 Available online 27 June 2019

Understanding the mechanism by which perfluorochemicals (PFCs) adsorbed on carbonaceous surface is eventually important to the design and process optimization of effective PFCs removal technologies. In this study, the possible binding mechanism of six different PFCs onto carbonaceous surface was investigated by means of first principles quantum mechanical methods based on density functional theory (DFT) calculation and wave function analysis. The adsorption process fitted well with pseudo-secondorder kinetic indicated that chemical bonding could not be underestimated. The results indicate that there were monolayer adsorption, electrostatic and hydrophobic interactions existed in PFCs adsorption process. DFT results suggested that the adsorption of PFCs on carbonaceous surface was one chemisorption process that accompanied by Van der Waals interactions. As there was different head functional groups in PFOS and PFOA, their adsorption capacity mainly controlled by the availability of active sites that was occupied by PFCs. The variation of chain length of PFBS and PFOS also take a certain responsible for different adsorption paths, due mainly to their hydrophobic effect. The obtained results from wave function and DFT analysis give in-depth understanding of PFCs adsorption on carbonaceous surface and help to their effectively removal. © 2019 Elsevier Ltd. All rights reserved.

Handling Editor: Y Yeomin Yoon Keywords: Adsorption Adsorption behaviors Perfluorochemicals (PFCs) Carbonaceous surface Density functional theory (DFT)

1. Introduction Perfluorochemicals (PFCs) have been widely used in many industrial areas, such as surfactants, surface protectors in paper, semiconductor, photograph, and coating materials (Rumsby et al., 2009; Paul et al., 2009). Recently, PFCs catch great attention due to their bio-accumulative, toxic and persistent characteristics (Lau et al., 2007; Newsted et al., 2005; Viberg and Eriksson, 2011). Perfluorooctane sulphonate (PFOS) and Perfluorooctanoic acid (PFOA) are two predominant PFCs. They have been listed in the Stockholm Convention in 2009, and 2015, respectively. However, PFCs could

* Corresponding author. E-mail address: [email protected] (X. Zhang). https://doi.org/10.1016/j.chemosphere.2019.06.191 0045-6535/© 2019 Elsevier Ltd. All rights reserved.

still be detected in wild animals and human breast epithelial cells till to date (Jian et al., 2017). The one major route of PFCs entering to environment is from the wastewater discharge from fire-fighting forms and surface coatings manufacturing (Jian et al., 2017). However, the general wastewater treatment processes are not designed for PFCs removal (Dong et al., 2017; Sun et al., 2011; Naile et al., 2010), which leads to the discharge of PFCs to the receiving water bodies. It has been reported that a high PFCs concentration over 1.41 ng/L was found in Songhuajiang River in the Northeast of China (Yan et al., 2014). Sun et al. (2011) observed an obvious increasing of PFCs concentrations in the Hun River after it passed through two cities (Sun et al., 2011). In fact, efforts were made to remove PFCs from the water and other sources. However, PFCs comprise of a large group of chemicals with a fully fluorinated alkyl chain, mainly including two

M. Li et al. / Chemosphere 235 (2019) 606e615

types, i.e. perfluorinatedsulfonates (PFSAs) and perfluorinated carboxylic acids (PFCAs) (Bao et al., 2010), both which are extremely refractory organics. Many studies performed the occurrences and degradations of PFOS and PFOA by physical, chemical and biological ways (Supreeyasunthorn et al., 2016; Lopez-Antia et al., 2017). Among all, PFCs adsorption by anion-exchange resin (Deng et al., 2010), boehmite (Wang et al., 2012), montmorillonite (Lu et al., 2016), nanosized inorganic oxides (Senevirathna et al., 2010; Zhang et al., 2010) are considered as an economic and effective method for PFCs removal from water. The adsorption of PFCs onto different adsorbents was summarized in Table S1. It was recognized that the adsorption of PFCs onto granular activated carbon (GAC) and resins required over 100 h, while only about 4 h was demanded to reach the sorption equilibrium by using powder activated carbon (PAC) (Yu et al., 2009). It indicates that PAC adsorption to remove PFCs may be a simple and high cost-effective way from water. Currently, studies are focusing on the understanding the kinetic and isotherms performance of PAC. However, the relations between physical and chemical characteristic of PFCs and adsorption properties, the activated site and the mechanism taking responsible for PFCs adsorption on PAC surface have also not been well investigated. In recent years, theoretical chemical calculations attracted increasing attention in metal adsorption (Padak and Wilcox, 2009). Liu has been studied the effects of halogen and oxygen functional groups on the mercury adsorption by the carbonaceous surface using density function theory (DFT) (Liu et al., 2011). These studies opened up the investigations required for elucidating the behaviors of different adsorbate on different adsorbent using DFT calculation method. So far, limited works has been conducted in the theoretical analysis of PFCs adsorption behaviors at the molecular level. The objective of this study was to compare the macroscopic adsorption behavior of six PFCs compounds by traditional adsorption experiments and apply the theoretical-based cluster modeling to examine the possible binding behaviors of different PFCs on PAC. 2. Materials and methods 2.1. Experimental materials and pretreatment Perfluorooctane sulphonate (PFOS), perfluorooctanoic acid (PFOA), perfluorohexanesulfonate (PFHxS), perfluorohexanoic acid (PFHxA), perfluorobutylsulfonate (PFBS) and perfluorobutyrate (PFBA) were purchased from Aladdin, China. PAC was obtained from Sigma Aldrich, USA. UPLC-grade methanol was provided by Fisher Chemical (America). Other chemicals were of reagent grade. The PAC was firstly rinsed with deionized water (DI water, Fisher, America) for several times and then was washed in deionized water to remove any dirt. The PAC was then dried in an oven at 105  C for 48 h, and transferred in a mortar and screened. 2.2. PFCs adsorption experiments Batch adsorption experiments were carried out in an orbital shaker at 120 rpm. Around 0.02g PAC and 2 mM NaH2PO4 as pH buffer were put into 250 mL flask. In the sorption kinetic experiments, 0.5 mg/L of PFCs solution was used and the experiments were conducted at 25  C. The sorption isotherm experiments were conducted at the initial PFCs concentrations ranging from 0.025 to 2 mg/L at 25  C for 24 h. All experiments were conducted twice and the average value was adopted. The Langmuir and Freundlich adsorption isothermal models were used to investigate the sorption equilibrium behavior of PFCs on PAC with the fitting formula of Eq. (1) and Eq. (2), respectively. It was assumed that the adsorbent surface is uniform and the adsorption capacity of all sites is constant.

q ce qe ¼ 1 m þ ce b

607

(1)

where qm is the maximum adsorption capacity of PFCs (mg/g), ce is equilibrium PFCs concentration in the solution (mg/L), and b is adsorption equilibrium constant (L/mg). 1=n

qe ¼ KF C e

(2)

where qe is equilibrium adsorption capacity of PFCs (mg/g), ce is equilibrium concentration of PFCs (mg/L), n is Freundlich index that representing the degree of deviation of linear adsorption, and KF is adsorption capacity constant of PFCs (mg/g). 2.3. Theoretical description and calculation Density functional theory (DFT) provides a satisfactory way to reach relatively high simulation accuracy and to guarantee the computational efficiency. Gaussian09 was employed to perform DFT calculation in this study. Geometry optimizations, vibration frequency analysis and energy calculation were carried out by using Becke three-parameter nonlocal exchange function and the LeeYang-Parr correlation function (B3LYP), while the Van der Waals interactions calculation was evaluated by M06-2X-D3 function, during which the atoms are described based on the 6e31g (d) basis set (Montoya et al., 2000). Carbonaceous surface and inner structure were established to examine its bonding pattern and thereby result adsorption behaviors of PFCs at the molecular and theoretical levels. From solidstate 13C nuclear magnetic resonance (NMR) experiments (Perry et al., 2000), it states that the carbonaceous surface with a mainly carbonaceous surface consisted of three to seven benzene rings. Especially, the cluster model consisting of four to seven fused benzene rings could well examine carbonaceous surface adsorption behavior of different species, by providing accurate description of adsorbate structure and adsorption energies (Radovic, 2005; Zhang and Truong, 2003). Accordingly, finite clusters with five fused benzene rings was chosen to simulate the carbonaceous surface structure (Liu et al., 2013), and the edge atoms on the carbonaceous surface outside of the model were set to be unsaturated to simulate active sites while the other edge carbon atoms were terminated with hydrogen atoms. Multiwfn program (version 3.2), a powerful wave function analysis software, was normally used to simulate the process of chemical bond formation by calculating Mayer bond order, bond energy analysis and Atoms in Molecules (AIM). AIM of real space function mainly refers to the critical point and the AIM path to reach the connection critical point. The AIM analysis of electron density was proposed by Bader and it is the most significant part in AIM theories (Lu and Chen, 2012). To analyze the non-covalent interactions between PFCs and carbonaceous surface, and an elegant way was used to assess its characteristic by analyzing the sign of l_2 and the second derivative of the electronic density in the perpendicular direction of the bond (Kozuch and Martin, 2013), during which the value of reduced density gradient (RDG) function can be determined via Eq. (3).

1 jVrj RDG ¼
1=3 4=3 2 r 2 3p

(3)

Van der Waals interactions are described by low electron density (r) at the critical points where l_2*r is negative, according to the method established by Johnson and Contreras-Garcia defined as NCI (Johnson et al., 2010; Contreras et al., 2011),. The function,

608

M. Li et al. / Chemosphere 235 (2019) 606e615

l_2*r with different colors to define different values (Fig. S1), filled into RDG isosurface, and thus visualization of the size and type of Van der Waals interactions can be observed with the assistance of VMD software (Lu and Chen, 2012). Afterwards, the Mayer bond order was calculated that was between atom A and B and defined by Eq. (4).

IAB ¼

XX

½ðPSÞba ðPSÞab þ P s S


ba

Ps S

 i ab

(4)

a2Ab2B

where Pa and Pb are alpha and beta density matrix, respectively, S is overlap matrix. As for a restricted close-shell circumstance, Eq. (4) can be simplified to Eq. (5), where spin density matrix is zero.

IAB ¼

XX

ðPSÞba ðPSÞab

(5)

a2Ab2B

The bonding and stable structure of PFOS and carbonaceous surface were also analyzed. Moreover, the chemisorption process of PFOS on the carbonaceous surface adsorption was accordingly examined. Based on PFCs adsorption kinetic and isotherms experiments, it was found that the physical adsorption process played an important role in the process of adsorption. Multiwfn combined with VMD software visualization was employed to analyze the Van der Waals interactions in the adsorption process.

2.4. Analytical methods The concentrations of PFCs were measured by UPLC-MS-MS (Waters), which was equipped with a BEH C18 column (2.1  50 mm, 1.7 mm). A mixture of acetonitrile and 2 mM ammonium acetate were used as the mobile phase. During the measurement, the injection flow rate was maintained at 0.3 mL/min. Selected ion monitoring mode was employed to quantify the analyzers, i.e. m/z ¼ 499 for PFOS, m/z ¼ 399 for PFHxS, m/z ¼ 299 for PFBS, m/z ¼ 413 for PFOA, m/z ¼ 313 for PFHxA, m/z ¼ 213 for PFBA. 3. Results and discussion 3.1. Adsorption behaviors of PFCs on PACs PFCAs and PFSAs compounds showed different adsorption behaviors and kinetics onto PAC, which were displayed in Fig. 1(a) and (b), respectively. It was observed that the rates of PFCAs adsorption onto PAC were quite faster than those of PFSAs. PFCAs could reach saturation at 16 h, while PFSAs need more than 24 h. This discrepancy was also observed by Yu (Yu et al., 2009), who though that PFOA reached the saturated adsorption faster than PFOS that may also depend on the PAC materials. Especially, although with the same C-F chain length, the adsorption amount of sulfonic acid head group by PAC, e.g. PFOS, PFHxS and PFBS, was greater than that of carboxylic head group, such as PFOA, PFHxA and PFBA

Fig. 1. Sorption kinetics of PFSAs (a), PFCAs (b), on PAC, and simulation results with pseudo-second-order models, and sorption isotherms of PFSAs (c) and PFCAs (d) on PAC and simulation by using Langmuir and Freundlich models.

M. Li et al. / Chemosphere 235 (2019) 606e615

(Fig. 1(a) and (b)). It indicates that the PAC adsorption sites were easily occupied by sulfonic acid group in contrast to carboxylic acid group, since that the hydrophobic groups in the tail played an important role in the PAC adsorption (Fig. 1(a) and (b)). Meanwhile, as for the same head group, PAC preferentially adsorbed more PFCs with long C-F chain compared to those with the short C-F chain. To further investigate the sorption kinetics, the pseudosecond-order model was used to fit the PAC adsorption of varied PFCs. It was found that the correlation coefficients (R2) between simulated and experimental results for all PFCs were greater than 0.9 (Fig. 1(a) and (b)). This good fitting of the pseudosecond-order model indicates that the chemical interactions are possibly involved in the sorption processes. The pKa of all the PFCs was below 1.0 when he adsorption tests were conducted at pH of 4.0 (Steinledarling and Reinhard, 2008). It reveals that all the molecules were at an anionic form. The isoelectric point of the PAC in this study was 7.5. Hence, it suggests that the electrostatic effect also may play an important role in the adsorption of negative charged PFCs onto positive charged PAC surface. The thermodynamic data for the adsorption of six PFCs were fitted with the Langmuir and Freundlich models, respectively. The adsorption isotherms at pH of 4.0 were shown in Fig. 1(c) and (d). The results demonstrate that the R2 in all the regression were greater than 0.9. According to the t-test results (Table S2 in the Supplementary materials), the p values were all more than 0.01. It indicates that the thermodynamic data are significantly correlated with the two models. These good fitting coefficients of Langmuir and Freundlich models revealed that it was uncertain to distinguish

609

the process of sorption with monolayer or multilayer. As the feeding PFCs concentration was ranged from 0.1 to 2 mg/L in this study, monolayer adsorption could occur for the PAC adsorption of PFCs. However, the good Freundlich fitting demonstrated that there were multilayer or hemi-micelles during PAC adsorption, or chemical adsorption occurred (Schwarzenbach et al., 2005). Hence, it is highly expected to reveal the real paths by conducting the adsorption of various PFCs with PAC. 3.2. PFOS defined in DFT and its adsorption path analysis on carbonaceous surface During the adsorptions of PFOS onto carbonaceous surface, PFOS molecules approached and attacked active sites of carbonaceous surface; sulfonic head-group vertically attacked the edge and center C on carbonaceous surface; the C-F chain as its tail may preferentially attack the edge C and C (Fig. 2(a)). Meanwhile, the possible surface complexes, 1a, 2a and 3a/1b showed in Fig. 2(b). Corresponded adsorption energies of different direction PFOS adsorbed on carbonaceous surface at different adsorption sites were defined in DTF according to Eq. (6).

Eads ¼ EðPFOScarbonaceous surfaceÞ EðPFOSÞ Eðcarbonaceous surfaceÞ (6) where E(PFOS-carbonaceous surface) is the energy of given geometry containing carbonaceous surface and the adsorbing molecule,

Fig. 2. Adsorption of PFOS at different sites on carbonaceous surface (a), and its adsorption complexes on carbonaceous surface (b), where yellow, red, gray, dark gray, blue and dark blue circles denotes S, O, H, C and F element, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

610

M. Li et al. / Chemosphere 235 (2019) 606e615

Table 1 DFT-calculated molecule-chain length, Gibbs free energy of PFOS and PFOScarbonaceous surface. Structures molecule chain length(Å) Gibbs free energy(KJ/mol) DG(KJ/mol) C PFOS 1a 2a 3a/1b

1912936.675 6894460.606 8807531.181 8807554.810 8807355.272

\ 12.8 17.3 19.1 16.2

\ \ 132.789 157.206 43.720

Table 2 Atomic bond population of PFOS and PFOS-carbonaceous surface. Bond

carbonaceous surface

1a

2a

3a/1b

24C-25C 25C-10C 10C-9C 9C-8C 8C-11C 11C-12C C-O

0.47 0.45 0.35 0.35 0.45 0.47

0.45 0.36 0.32 0.43 0.44 0.43 0.16

0.47 0.54 0.25 0.26 0.44 0.47 0.23

0.46 0.46 0.29 0.25 0.39 0.44

E(PFOS) is the energy of PFOS, and E(carbonaceous surface) is the energy of the carbonaceous surface. According to the previous study, the adsorption is considered to be weak when Eads is ranged from 10 to 30 kJ/mol, and adsorption energy (between 50 and 960 kJ/mol) deemed as a strong adsorption that could be considered as chemical adsorption (Liu et al., 2013). Accordingly, the calculated Eads for adsorption energy was listed in Table 1. The stability order of these three adsorption products was 2a(-157.206 kJ/mol)>1a(-132.789 kJ/mol)>3a/1b(43.72 kJ/mol). The chemical adsorption of the process was displayed in Fig. 1(a). The possibility of PFOS adsorbed at the center of C was relatively high, while the possibility at the edge C was low, as the energy of the edge of C was higher than that at the central position. The atomic bond population was calculated with Mulliken population analysis and the results were displayed in Table 2. It could be used as a measure of bond strength between PFOS and carbonaceous surface. High positive value of bond population refers to strong covalent bond, while there is not any interaction occurred if the bond population between the two atoms closed to zero (Hu

et al., 2007). From Table 2, it was observed that the bond population of C(8)-C(9) and C(9)-C(10) decreased while C(10)-C(25) increased during the adsorption progression. The bond population of C(9)-O bonds in 2a was 0.23. It indicates that some of the electrons may be transferred to the adsorbed PFOS molecules with the progress of carbonaceous surface adsorption, induced a rather weak C-C bond strength on the carbonaceous surface. Meanwhile, it was also found that the bond population of C(13)-O bond in 1a was only 0.16, which was smaller than that of 2a (0.23), evidencing that the 2a structure was rather stable. Afterwards, the Mayer bond order of 2a was calculated, and the values of the single, double and triple bond were close to 1.0, 2.0 and 3.0, respectively. In PFOS adsorption onto carbonaceous surface, the Mayer bond order of C(9)-O in 2a is 0.863 (closed to 1.0). It suggests that atom C bonding to atom O seems as a single bond. RDG function and l_2*r scatter graph of 2a structure calculated by Multiwfn function were shown in Fig. 3(a). It was observed that a spike demonstrating the Van der Waals interactions between PFOS molecule and carbonaceous surface was occurred in the range from 0.02 to 0.01, and from 0.01 to 0.02, respectively (Kozuch and Martin, 2013). This visualization model results reveal that the range of l_2*r from 0.02 to 0.01 was steric hindrance among benzene rings, and the range from 0.01 to 0.02 referred to physical attraction (Fig. 3(b)) that may be caused by the sulfonic acid group and adjacent C-F chain to the adsorption sites of carbonaceous surface (Kozuch and Martin, 2013). Thus, as for the PFOS adsorption onto carbonaceous surface, it was deduced that chemical adsorption was the most important adsorption ways during all adsorption processes through Gibbs energies analysis, and C-C single bond was the one major connection bond between carbonaceous surface and PFOS through Mayer bond order calculation. RDG function and l_2*r scatter graphs confirms that physical adsorption accompanied with chemical adsorption was occurred in the adsorption process. 3.3. PFBS defined in DFT and its adsorption path analysis on carbonaceous surface The major adsorption sites during PFBS adsorption onto carbonaceous surface were shown in Fig. 4 (a). Different from PFOS, the chain length of CF2 in PFBS changed from seven to three, thus its

Fig. 3. RDG function and l_2*r function scatter graphs of 2a structure calculated by using Multiwfn (a) and visualized model of 2a structure (b), where yellow, red, while, light blue, and pink circles denotes S, O, H, C and F elements, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

M. Li et al. / Chemosphere 235 (2019) 606e615

whole length was shorter than those of PFOS. The possible surface complexes 1a0 , 2a0 and 3a’/1b’ were shown in Fig. 4 (b). The adsorption energies of the different direction PFBS adsorbed on carbonaceous surface at different adsorption sites were defined in Eq. (7).

Eads ¼ EðPFBScarbonaceous surfaceÞ EðPFBSÞ Eðcarbonaceous surfaceÞ (7) As listed in Table 3, the stability order of three adsorption products was 2a’ (160.785 kJ/mol) >1a’ (143.457 kJ/mol) >3a’/ 1b’ (50.304 kJ/mol). The results implied that the possibility of PFBS adsorption at the central position of C was relatively higher compared to the edge one. Similarly, 2a0 was the most stable structure for all the complexes, which deemed as chemical

611

adsorption. Atomic bond populations for PFBS on carbonaceous surface were given in Table 4. It can be found that the change of the bond strength was almost the same as the bond strength of PFOS adsorbed on carbonaceous surface, as that CF2 chain length had insignificant influence on the chemical adsorption of PFCs. The Van der Waals interactions between the PFBS head group and the carbonaceous surface were also observed in Multiwfn combined with VMD visualization of the Van der Waals interactions process. RDG function and l_2*r function scatter graph calculated by Multiwfn of 2a0 structure was shown in Fig. 5(a), and the visualization model of 2a0 structure was shown in Fig. 5(b). These results were rather similar as those displayed in Fig. 5(a) and (b). It was found that the chain length of C-F had a weak correlation with the interaction between the head sulfonic acid group and the

Fig. 4. Adsorption of PFBS at different sites on carbonaceous surface (a), and its adsorption complexes on carbonaceous surface (b), where yellow, red, gray, dark gray, blue and dark blue circles denotes S, O, H, C and F element, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

612

M. Li et al. / Chemosphere 235 (2019) 606e615

Table 3 DFT-calculated molecule-chain length, Gibbs free energy of PFBS and PFBScarbonaceous surface. Structures molecule chain length(Å) Gibbs free energy(KJ/mol) DG(KJ/mol) C PFBS 1a0 2a0 3a’/1b0

1912936.675 4397300.297 6310383.999 6310399.752 6310189.712

\ 7.6 12.6 14.2 15.6

\ \ 143.457 160.785 50.304

Table 4 Atomic bond population of PFBS and PFBS-carbonaceous surface. Bond

carbonaceous surface

1a0

2a0

3a’/1b0

24C-25C 25C-10C 10C-9C 9C-8C 8C-11C 11C-12C C-O

0.47 0.45 0.35 0.35 0.45 0.47

0.45 0.36 0.34 0.41 0.45 0.42 0.17

0.47 0.55 0.25 0.26 0.44 0.48 0.23

0.45 0.45 0.30 0.28 0.39 0.47

neighboring CF2 chain close to the carbonaceous surface adsorption site. The C-F chain as tail attacking carbonaceous surface adsorption site was also calculated. The l_2*r scatter graph of 3a’/1b0 structure of function and visualization model were shown in Fig. 5(c) and (d), respectively. It was observed that the head sulfonic acid group has great influence on carbonaceous surface adsorption site, although the C-F tail was forced to attack carbonaceous surface adsorption site. It was due to that the C-F chain of PFBS was too short to unfold on the carbonaceous surface compared to the C-F length of PFOS. Therefore, the calculation result of 3a’/1b’ structure cannot fully express the Van der Waals interactions between C-F tail of PFBS and carbonaceous surface. According to the above calculation, as for all the processes of PFBS adsorbed onto carbonaceous surface, chemical adsorption and the head of sulfonic acid group physical adsorption were rather similar to those of PFOS. In the chemical adsorption process, C-C bond was the connection between carbonaceous surface and PFBS. For the difference of equilibrium adsorption process, it is mainly attributed to the different hydrophobic effect of adsorption.

Fig. 5. RDG function and l_2*r function scatter graphs of 1a0 2a0 (a) and 3a’/1b0 (c) structure calculated by Multiwfn, and visualized model of 1a’/2a’ (b) and 3a’/1b0 (d) structure, where yellow, red, while, light blue, and pink circles denotes S, O, H, C and F element, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

M. Li et al. / Chemosphere 235 (2019) 606e615

3.4. PFOA defined in DFT and its adsorption path analysis on carbonaceous surface Different adsorption sites during PFBS adsorption onto carbonaceous surface and the possible surface complexes, 1a’’/2a’’ and 3a’’/1b’’, were shown in Fig. 6(a) and Fig. 6(b). The adsorption energies of different direction PFOA adsorbed on carbonaceous surface at different adsorption sites were defined according to Eq. (8).

Eads ¼ EðPFOAcarbonaceous surfaceÞ EðPFOAÞ Eðcarbonaceous surfaceÞ (7) As listed in Table 5, adsorption energy results indicated the

613

stability order of two adsorption products was 1a’’/2a’’ (525.14 kJ/ mol) >3a’/1b’ (18.49 kJ/mol). From the viewpoint of chemical adsorption, there was a great difference in PFOA adsorption because of its carboxylic head group. In the stable 1a’’/2a’’ structure, its adsorption energy was much smaller than the 2a structure of PFOS. Each PFOA molecule occupied two carbonaceous surface adsorption sites (Fig. 6), which was one of the main reasons leading to the faster adsorption rate of PFOA than PFOS onto carbonaceous surface. Meanwhile, the calculated results were also in good agreement with the experimental results. Atomic bond populations for PFOA on carbonaceous surface were given in Table 6. The bond population of C(8)-C(9), C(9)-C(10), C(10)-C(25) and C(24)-C(25) decreased due to the adsorption. It can

Fig. 6. Adsorption of PFOA at different sites on carbonaceous surface (a), and PFOA adsorption complexes on carbonaceous surface (b), where yellow, red, gray, dark gray, blue and dark blue circles denotes S, O, H, C and F element, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

614

M. Li et al. / Chemosphere 235 (2019) 606e615

Table 5 DFT-calculated molecule-chain length, Gibbs free energy of PFOA and PFOAcarbonaceous surface. Structures molecule chain length(Å) Gibbs free energy(KJ/mol) DG(KJ/mol) C PFOA 1a’’/2a’’ 3a’’/1b’’

1912936.675 5127407.174 7040870.537 7040326.901

\ 11.1 17.1 13.1

\ \ 525.140 18.490

carbonaceous surface adsorption site was weaker than that of PFOS. The visualization model displayed that a part of green region appeared in the space between C-F chain and carbonaceous surface adsorption site, and a part of the green region appeared in carbonyl O and carbonyl O action space. It can be seen that its area was significantly less than PFOS (Fig. 7(b)), which implies that physical interaction between PFOA and carbonaceous surface was weaker than PFOS on carbonaceous surface.

4. Conclusions

Table 6 Atomic bond population of PFOA and PFOA- carbonaceous surface. Bond

carbonaceous surface

1a’’/2a’’

3a’’/1b’’

24C-25C 25C-10C 10C-9C 9C-8C 8C-11C 11C-12C 9C-C(PFOA) C¼O

0.477 0.458 0.357 0.357 0.458 0.477

0.36 0.24 0.31 0.29 0.51 0.47 0.18 0.53

0.43 0.42 0.32 0.30 0.38 0.45

be seen that the effect of PFOA on the intensity of C-C bond of carbonaceous surface was rather significant than those in PFOS adsorption process. The C-C single bond was formed accompanied with formation of C¼O double bond. Its impact became more significant as the bond population of C-C bonds (0.18) was smaller than that of C¼O double bond (0.54). It indicates that much more of the electrons would be transferred to C¼O double bond rather than to C-C single bond. The Mayer bond order of C-C and C¼O of 1a’’/ 1b’’ were 0.84 (closed to 1.0) and 1.676 (close to 2.0), respectively. It further proved that C-C was a single bond and C¼O was a double bond. The Van der Waals interactions between the PFOA head group and the carbonaceous surface was also observed by Multiwfn combined with VMD visualization of the Van der Waals interactions process. RDG function and l_2*r function scatter graph calculated with Multiwfn of 1a’’/2a’’ structure were shown in Fig. 7(a) and the visualization model of 1a’’/2a’’ structure was shown in Fig. 7(b). From Fig. 7(a), the spike occurred mainly in the range of l_2*r from 0.015 to 0.005 It is closer to zero than the spike range of PFOS and PFBS. It indicates that the effect of Van der Waals interactions between the carboxylic acid of PFOA and

PFCs adsorption behaviors on carbonaceous surface were investigated by series of experiments and density functional theory. The kinetic analysis revealed that the sorption of PFCs on carbonaceous surface related significantly with the head group and the C-F chain length. The PFCAs adsorption rate onto carbonaceous surface was slower than that of PFSAs. The pseudo-second-order model could well describe the PFCs sorption, indicating that chemisorption plays a significant role. The energy diagrams were used to compare the energies of various possible carbonaceous surface complexes and PFCs adsorption paths on carbonaceous surface. It was found that PFOA adsorbed on carbonaceous surface was much more stable than PFOS and PFBS. The atoms C on the carbonaceous surface linked to atoms O(PFOS) by a C-O single bond during PFOS adsorption, and there was a strong attraction in the stable adsorption structure. As for the PFBS adsorption, the length of tail C-F chain has insignificant impact on the formation of chemical bonding of C-O single bond and physical interaction strength. It is found that the different hydrophobic interaction between PFOS and PFBS would take important role. In the PFOA adsorption onto carbonaceous surface, the C¼O bond of PFOA molecule was broken to form atom O, which would link to atom C on carbonaceous surface to form C¼O double bond, and to form atom C that linked to atom C of center site with C-C single bond. Meanwhile, it was also revealed that the change of head group affected the Van der Waals interactions interaction strength between tail C-F chain and carbonaceous surface.

Declarations of interest None.

Fig. 7. RDG function and l_2*r function scatter graphs of 1a’’/2a’’(a) structure calculated by Multiwfn, and visualized model of 1a’’/2a’’(b) structure, where yellow, red, while, light blue, and pink circles denotes S, O, H, C and F element, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

M. Li et al. / Chemosphere 235 (2019) 606e615

Acknowledgement This research was supported by the grant Project supported by Guangdong Natural Science Foundation [2017A030313285], the Shenzhen Science and Technology Funding Project [grant number JCYJ20160406162038258], Program of International S&T Cooperation [grant number 2016YFE0123400], National Natural Science Foundation of China [grant numbers 51678183, 51408149], and from Chinese National Water Pollution Control [grant number 2012ZX07313001-008]. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.chemosphere.2019.06.191. References Bao, J., Liu, W., Liu, L., Jin, Y., Ran, X., Zhang, Z., 2010. Perfluorinated compounds in urban river sediments from Guangzhou and Shanghai of China. Chemosphere 80, 123e130. Contreras, J., Johnson, E.,R., Keinan, S., Chaudret, R., Piquemal, J.,P., Beratan, D.,N., Yang, W., 2011. NCIPLOT: a program for plotting noncovalent interaction regions. J. Chem. Theory Comput. 7, 625e632. Deng, S., Yu, Q., Huang, J., Yu, G., 2010. Removal of perfluorooctanesulfonate from wastewater by anion exchange resins: effects of resin properties and solution chemistry. Water Res. 44, 5188e5195. Dong, W., Liu, B., Song, Y., Zhang, H., Li, J., Cui, X., 2017. Occurrence and partition of perfluorinated compounds (PFCs) in water and sediment from the Songhua river. China. Arch. Environ. Contam. Toxicol. 74, 1e10. Hu, Q., Wu, Q., Sun, G., Luo, X., Liu, Z., Xu, B., He, J., Tian, Y., 2007. First-principles study of atomic oxygen adsorption on boron-substituted graphite. Surf. Sci. 602, 37e45. Jian, J.,M., Guo, Y., Zeng, L.,X., Liu, L.,Y., Lu, X.,W., Wang, F., Zeng, E., 2017. Global distribution of perfluorochemicals (PFCs) in potential human exposure sourceA review. Environ. Int. 108, 51e62. Johnson, E.,R., Keinan, S., Mori-Sanchez, P., Contreras-Garcia, J., Cohen, A.,J., Yang, W., 2010. Revealing noncovalent interactions. Am. Chem. Soc. 132, 6498e6506. Kozuch, S., Martin, J.,M.,L., 2013. Halogen bonds: benchmarks and theoretical analysis. J. Chem. Theory Comput. 9, 1918e1931. Lau, C., Anitole, K., Hodes, C., Lai, D., 2007. Perfluoroalkyl acids: a review of monitoring and toxicological findings. Toxicol. Sci. 99, 366e394. Liu, J., Cheney, M.,A., Wu, F., Li, M., 2011. Effects of chemical functional groups on elemental mercury adsorption on carbonaceous surfaces. J. Hazard Mater. 186, 108e113. Liu, J., Qu, W., Zheng, C., 2013. Theoretical studies of mercuryebromine species adsorption mechanism on carbonaceous surface. Proc. Combust. Inst. 34, 2811e2819. Lopez-Antia, A., Dauwe, T., Meyer, J., Maes, K., Bervoets, L., Eens, M., 2017. High levels of PFOS in eggs of three bird species in the neighbourhood of a fluorochemical plant. Ecotoxicol. Environ. Saf. 139, 165e171. Lu, X., Deng, S., Wang, B., Huang, J., Wang, Y., Yu, G., 2016. Adsorption behavior and

615

behaviors of perfluorooctanesulfonate on nanosized inorganic oxides. J. Colloid Interface Sci. 474, 199. Lu, T., Chen, F., 2012. Multiwfn: a multifunctional wavefunction analyzer. J. Comput. Chem. 33, 580e592. Montoya, A., Truong, T.,N., Sarofilm, A.,F., 2000. Application of density functional theory to the study of the reaction of NO with char-bound nitrogen during combustion. J. Phys. Chem. B 104, 8409e8417. Naile, J.,E., Jongseong, K., Wang, T.,Y., Chen, C., Luo, W., Kwon, B., Park, J., Park, C., Jones, P., Lu, Y., Giesy, J.,P., 2010. Perfluorinated compounds in water, sediment, soil and biota from estuarine and coastal areas of Korea. Environ. Pollut. 158, 1237e1244. Newsted, J.,L., Jones, P.,D., Coady, K., Giesy, J.,P., 2005. Avian toxicity reference values for perfluorooctanesulfonate. Environ. Sci. Technol. 39, 9357e9362. Padak, B., Wilcox, J., 2009. Understanding mercury binding on activated carbon. Environ. Prog. 47, 2855e2864. Paul, A.,G., Jones, K.,C., Sweetman, A.,J., 2009. A first global production, emission, and environmental inventory for perfluorooctanesulfonate. Environ. Sci. Technol. 43, 386e392. Perry, S.,T., Hambly, E.,M., Fletcher, T.,H., Solum, M.,S., Pugmire, R.,J., 2000. Solidstate 13C NMR characterization of matched tars and chars from rapid coal devolatilization. Proc. Combust. Inst. 28, 2313e2319. Radovic, L.,R., 2005. The behaviors of CO2 chemisorption on zigzag carbon active sites: a computational chemistry study. Carbon 43, 907e915. Rumsby, P.,C., Mclaughlin, C.,L., Hall, T., 2009. Perfluorooctanesulphonate and perfluorooctanoic acid in drinking and environmental waters. Philos. Trans. R. Soc. A-Math. Phys. Eng. Sci. 367, 4119e4136. Schwarzenbach, R.,P., Gschwen, P.,M., Imboden, D.,M., 2005. Organic acids and bases: acidity constant and partitioning behavior. Environ. Inorg. Chem. Senevirathna, S.,T., Tanaka, S., Fujii, S., Kunacheva, C., Harada, H., Shivakoti, B.,R., Okamoto, R., 2010. A comparative study of adsorption of perfluorooctanesulfonate (PFOS) onto granular activated carbon, ion-exchange polymers and non-ion-exchange polymers. Chemosphere 80, 647e651. Steinledarling, E., Reinhard, M., 2008. Nanofiltration for trace organic contaminant removal: structure, solution, and membrane fouling effects on the rejection of perfluorochemicals. Environ. Sci. Technol. 42, 5292e5297. Sun, H., Li, F., Zhang, T., Zhang, X., He, N., S, Q., Zhao, L., Sun, L., Sun, T., 2011. Perfluorinated compounds in surface waters and WWTPs in Shenyang, China: mass flows and source analysis. Water Res. 45, 4483e4490. Supreeyasunthorn, P., Boontanon, S.,K., Boontanon, N., 2016. Perfluorooctanesulfonate (PFOS) and perfluorooctanoic acid (PFOA) contamination from textiles. Environ. Lett. 51, 1e6. Viberg, H., Eriksson, P., 2011. Perfluorooctanesulfonate (PFOS) and Perfluorooctanoic Acid (PFOA), Reproductive and Developmental Toxicology, pp. 623e635. Wang, F., Liu, C., Shih, K., 2012. Adsorption behavior of perfluorooctanesulfonate (PFOS) and perfluorooctanoate (PFOA) on boehmite. Chemosphere 89, 1009e1014. Yan, A., Zhihong, Z., Ranran, L., 2014. Water quality assessment in the Harbin reach of the Songhuajiang river (China) based on a fuzzy rough set and, an attribute recognition theoretical model. Int. J. Environ. Res. Public Health 11, 3507e3520. Yu, Q., Zhang, R., Deng, S., Huang, J., Yu, G., 2009. Sorption of perfluorooctanesulfonate and perfluorooctanoate on activated carbons and resin: kinetic and isotherm study. Water Res. 43, 1150e1158. Zhang, Q., Deng, S., Yu, G., Huang, J., 2010. Removal of perfluorooctanesulfonate from aqueous solution by crosslinked chitosan beads: sorption kinetics and uptake behaviors. Bioresour. Technol. 102, 2265e2271. Zhang, S., Truong, T.N., 2003. Kinetics of hydrogen abstraction reaction class H þ HC(sp3): first-principles predictions using the reaction class transition state theory. J. Phys. Chem. A 107, 1138e1147.