Article pubs.acs.org/est
Effects of Solution Chemistry on Adsorption of Selected Pharmaceuticals and Personal Care Products (PPCPs) by Graphenes and Carbon Nanotubes Fei-fei Liu,†,‡ Jian Zhao,§,‡ Shuguang Wang,† Peng Du,‡ and Baoshan Xing*,‡ †
Shandong Provincial Key Laboratory of Water Pollution Control and Resource Reuse, School of Environmental Science and Engineering, Shandong University, Jinan, Shandong 250100, China ‡ Stockbridge School of Agriculture, University of Massachusetts, Amherst, Massachusetts 01003, United States § College of Environmental Science and Engineering, Ocean University of China, Qingdao, Shandong 266100, China S Supporting Information *
ABSTRACT: Adsorption of three selected pharmaceuticals and personal care products (PPCPs) (ketoprofen (KEP), carbamazepine (CBZ), and bisphenol A (BPA)) by two reduced graphene oxides (rGO1 and rGO2) and one commercial graphene was examined under different solution conditions. Single-walled carbon nanotubes (SWCNTs), multiwalled carbon nanotubes (MWCNTs), and powdered graphite were also investigated for comparison. All adsorption isotherms followed the order of SWCNTs > rGO1 > rGO2 > MWCNTs > graphene > graphite, consistent with the orders of their surface areas and micropore volumes. After surface area normalization, adsorption affinities of the three PPCPs onto graphenes were lower than onto graphite, suggesting incomplete occupation for adsorption sites because of the aggregation of graphene sheets and the presence of oxygen-containing functional groups. The observed pH effects on adsorption correlated well with the pH-regulated distribution of the protonated neutral species of the three PPCPs. Increasing ionic strength from 0 to 20 mM increased KEP adsorption due to the electrostatic screening by Na+ and Ca2+. Both humic acid (HA) and sodium dodecylbenzenesulfonate (SDBS) suppressed PPCPs adsorption to all adsorbents, but their impacts onto graphenes were lower than those onto CNTs because of their lower adsorption by graphenes. More severe HA (or SDBS) effect was found on negatively charged KEP at the tested solution pH 6.50 due to the electrostatic repulsion between the same charged KEP and HA (or SDBS). The findings of the present study may have significant implications for the environmental fate assessment of PPCPs and graphene.
■
INTRODUCTION Graphene, a two-dimensional sheet consisting of a single layer of sp2 carbon atoms, has attracted increasing attention and has become a novel star material since it was first isolated in 2004.1 Potential applications of graphene have been demonstrated, such as electronic sensors, transparent conductors, polymer composites, and energy storage, due to its remarkable electrical, optical, mechanical, and thermal properties.2 Development of chemical and thermal synthesis techniques has opened up facile ways for large-scale productions of graphene.3,4 It has been reported that the global market for graphene-based materials will reach $67 million in 2015 and $675 million in 2020 at annual growth rate of 58.7%.5 Such large-quantity production and wide application will inevitably lead to the release of graphene into the environment via discharge from manufacturing facilities and attrition from composite products. The environmental fate and transport of graphene are attracting more and more concern as its toxicity evidence is accumulating.6 Because of its highly hydrophobic surface and © 2014 American Chemical Society
large theoretical specific surface area, graphene exhibits strong adsorption affinities to organic pollutants, such as polycyclic aromatic hydrocarbons (PAHs), chlorobenzenes, and phenols.7−9 Therefore, the toxicity of graphene may be further enhanced through adsorption of these toxic chemicals. As emerging organic contaminants, pharmaceuticals and personal care products (PPCPs) have gained growing attention in recent years because of their frequent detection in the environment and their long-term potential risks to drinking water safety and aquatic organisms.10,11 From the engineering and environmental perspectives, carbon nanotubes (CNTs) have been investigated extensively for removal of PPCPs, such as tetracyclines, sulfamethoxazole, and tylosin,12 ibuprofen and triclosan,13 bisphenol A and 17α-ethinyl estradiol from aqueous Received: Revised: Accepted: Published: 13197
July 17, 2014 October 26, 2014 October 29, 2014 October 29, 2014 dx.doi.org/10.1021/es5034684 | Environ. Sci. Technol. 2014, 48, 13197−13206
Environmental Science & Technology
Article
solutions.14 These studies demonstrated that CNTs−PPCPs interaction is of great importance to the real-world application of CNTs in wastewater treatment and the evaluation of the environmental fate and transport of PPCPs. Compared with CNTs, only a few studies have investigated the interaction between graphene and PPCPs currently because graphene is a more recent discovery.15,16 The structural and chemical differences of graphene from CNTs may cause distinctive interactions of graphene with PPCPs. First, graphene is composed of several two-dimensional single graphene sheets, possessing open layered structure for organic molecules adsorption. Second, graphene tends to aggregate owing to the strong van der Waals forces, and thus may form interconnected pores within the aggregates. Third, graphene may have some remaining oxygen-containing functional groups because of the incomplete reduction through synthesis methods. Therefore, coupling the characteristics between CNTs and graphene is expected to provide in-depth information for graphene−PPCPs interactions. Adsorption can be greatly affected by the aqueous solution chemistry parameters.13,14,17−22 The existing forms of chemicals, especially ionizable organic chemicals, are highly pHdependent.21 Ions (e.g., Na+ or Ca2+) may have a “salting-out” or “squeezing-out” effect on adsorption of hydrophobic compounds.22 The ubiquitous natural organic matter (NOM) may also have severe impacts on adsorption because of its two opposite effects: decreasing adsorption sites through competition and pore blockage or increasing adsorption sites owing to its better dispersibility.13,14,17−20 In addition, surfactant can be viewed as another important factor influencing PPCPs adsorption because of its wide application in PPCPs emulsions stabilization, drug delivery, and its frequent presence in wastewaters.23,24 However, to our best knowledge, little work has been conducted to examine these aqueous chemistry conditions on PPCPs adsorption by graphene. Overall, we hypothesize that graphene could have different adsorption performance toward PPCPs in comparison with CNTs because of their distinct structural conformation and surface area; and the graphene−PPCPs interaction is highly influenced by solution chemistry such as pH, ionic strength, NOM, and surfactant, thus altering the adsorption capacity of graphene. Therefore, in the present study, we systematically investigated the adsorption mechanisms of three selected PPCPs (ketoprofen, carbamazepine, and bisphenol A) to three graphenes (two reduced graphene oxides and one commercial graphene). Single-walled CNTs (SWCNTs), multiwalled CNTs (MWCNTs), and graphite were used to serve as comparisons to better evaluate their adsorption behavior. Moreover, we further studied the solution chemistry conditions (pH, ionic strength, NOM, and surfactant) on the adsorption of the three PPCPs by adsorbents.
(HA) extracted from Amherst peat soil were used as models of surfactant and NOM, respectively (SI Tables S2 and S3). Two synthetically reduced graphene oxides (rGO1 and rGO2), one commercial graphene (Graphene Supermarket, USA), one SWCNTs (outer diameter 1−2 nm, length 5−30 μm, purity >90%, Shenzhen Nanotech Port Co., Ltd., China), one MWCNTs (inner diameter 3−5 nm, outer diameter 8−15 nm, length 10−50 μm, purity >95%, Chengdu Organic Chemicals Co., Ltd., China), and one powdered graphite (∼300 mesh, Fisher Scientific, USA) were used in this study. rGOs were produced by reducing graphene oxides (GOs) which were prepared by improved hummers’ method from graphite powder and flakes. Detailed procedures for preparation of rGO1 and rGO2 are described in the SI. Characterization. UV−visible spectra of rGOs were recorded on an Agilent 8453 UV−visible spectrophotometer (Agilent Technologies, USA). Transmission electron microscopy (TEM) images were observed with a microscope (JEOL 2000FX, Japan). X-ray diffraction (XRD) patterns were recorded with a powder diffractometer (MAC Science M18XHF, UK) using Cu Kα radiation at a scanning ratio of 8°/min ranging from 5° to 40°. Nitrogen adsorption− desorption isotherms were obtained at 77 K with Quantachrome Autosorb-1 (Quantachrome, USA). Multipoint Brunauer−Emmett−Teller (BET) equation and density functional theory (DFT) were used to calculate their BET surface area and pore size distribution. Elemental composition (C, O, H, N, and S) was determined in triplicates using an elemental analyzer (MicroCube, Elementar, Germany). X-ray photoelectron spectroscopy (XPS) measurements were performed with a Physical Electronics PHI 5500 XPS system (PerkinElmer, USA) using Mg Kα radiation. Point of zero charge (pHpzc) was obtained by the zeta potentials−pH curves using a 90Plus particle size analyzer equipped with ZetaPALS (Brookhaven Instruments, USA). Batch Adsorption Experiments. Adsorption isotherms were obtained using a batch equilibration technique. Target PPCPs solutions were prepared by background solution containing 20 mM NaCl (to maintain constant ionic strength) and 200 mg/L NaN3 (to prevent biodegradation), and adjusted to pH 6.50 ± 0.05 using 0.1 M HCl or 0.1 M NaOH. The test solutions with varying initial PPCPs concentrations (KEP 2.5− 50 mg/L, CBZ 6−100 mg/L, BPA 25−300 mg/L) were added to Teflon-lined screw cap glass vials with certain amount of adsorbents until a minimum headspace was achieved. The solid to solution ratio was adjusted to obtain 20−80% PPCPs removal. All vials were mixed on a rotary shaker (150 rpm, 25 °C) for 7 days, which was long enough to reach adsorption equilibrium through our preliminary experiments (SI Figure S1). After mixing, the vials were centrifuged for 20 min at 3000 rpm for further measurements. To investigate pH and ionic strength effects on adsorption, single-point adsorption experiments were carried out. The pH experiments were conducted in pH 2−12 using 20 mM NaCl as background solution. For the ionic strength experiments, adsorption was conducted in 0−100 mM NaCl or CaCl2 solutions (pH 6.50). Another set of experiments was performed to examine the effect of HA or SDBS on PPCPs adsorption. Test PPCPs solutions were prepared with HA or SDBS solutions (50 mg/L, to mimic the NOM-rich natural water or highly surfactant-polluted water). The pH and ionic strength were 6.50 and 20 mM NaCl, respectively.
■
MATERIALS AND METHODS Materials. PPCPs used in this study were selected according to their representative properties and environmental relevance.25,26 These compounds were ketoprofen (KEP), a carboxylic anti-inflammatory; carbamazepine (CBZ), an amide antiepileptic; and bisphenol A (BPA), a plasticizer and phenolic estrone. They were purchased from Sigma-Aldrich with minimum purities of 98% and used as received. The selected physical−chemical properties of the three compounds are listed in Supporting Information (SI) Table S1. Sodium dodecylbenzenesulfonate (SDBS) (Fisher Scientific, USA) and humic acid 13198
dx.doi.org/10.1021/es5034684 | Environ. Sci. Technol. 2014, 48, 13197−13206
Environmental Science & Technology
Article
Figure 1. Photographs of the aqueous suspensions (A), UV−vis spectra (B), XRD patterns (C), and TEM images (D) of GOs and rGOs.
Determination of PPCPs. Supernatant of each sample was analyzed by high-performance liquid chromatography (HPLC, Shimadzu, Japan) with an UV detector using a 4.6 mm × 150 mm SB-C18 column (SI Figure S2). The mobile phase was used under the following conditions: acetonitrile, 0.02 M KH2PO4 adjusted to an apparent pH 3.0 with H3PO4 (45:55, v/ v) for KEP; methanol, ammonium acetate (10 mM) and acetic acid (5 mM) (60:40, v/v) for CBZ; acetonitrile, deionized water with 1% acetic acid (50:50, v/v) for BPA. The determination wavelengths were 256, 285, and 278 nm for KEP, CBZ, and BPA, respectively.27−29 Data Analysis. All adsorption data were obtained in duplicate experiments. Three different isotherm models were employed to fit the adsorption data: Langmuir model (LM): qe =
rGO1 and rGO2 (Figure 1A). The UV−vis absorption peaks (∼226 nm) of GOs are red-shifted to ∼269 nm for rGOs (Figure 1B), suggesting that the conjugated structures within the graphene sheets were partly restored upon hydrazine reduction. XRD analysis revealed that, after reduction, (001) diffraction peak (2θ ∼ 10.8°) of GOs disappeared and a broad (002) peak (2θ ∼ 24.2°) appeared for rGOs (Figure 1C), corresponding to the decrease of d spacing from 8.19 to 3.68 Å. TEM images revealed that compared with GOs, rGOs had compacted and highly aggregated sheets because of the elimination of some functional groups on rGOs (Figure 1D). The physical−chemical properties of the six adsorbents are presented in SI Table S4. The synthesized rGO1 and rGO2 had similar surface areas (331 and 325 m2/g, respectively) although they were produced from different raw graphite materials. The commercial graphene had much lower surface area value of 132 m2/g, indicating thicker sheets than rGOs. The two rGOs were hybrid in micropores and mesopores (and macropores) with an approximate ratio of 1:1, while the commercial graphene was dominated by mesopores and macropores. Our previous study demonstrated that pores of graphene may originate from the folding/aggregation of graphene sheets,30 which was also observed from TEM images of rGOs (Figure 1D). Moreover, structural defects of graphene contribute pores too.31 The physical characteristics of graphenes were compared with SWCNTs, MWCNTs, and graphite, showing that the surface areas and micropore volumes of the six adsorbents followed the order of SWCNTs > rGO1 ∼ rGO2 > graphene > MWCNTs > graphite. It should be noted that structural characteristics of GOs were not determined because GO sheets were welldispersed and their pore distribution could be largely changed. More importantly, GOs had very much lower adsorption capacities toward hydrophobic organic compounds than rGOs.32 Therefore, GOs were not investigated as adsorbents in this study. Elemental analysis showed that rGO1 and rGO2 had the highest remaining oxygen contents (16.5−17.3%)
qmKLCe 1 + KLCe
Freundlich model (FM): qe = KFCen
Dubinin−Ashtakhov model (DAM): log qe = log Q 0 − (ε /E)b
where qe (mg/g) is the solid-phase concentration, Ce (mg/L) is the solution concentration, qm and Q0 (mg/g) are the maximum adsorption capacity for LM and DAM, and KL (L/ mg) is the Langmuir adsorption affinity parameter. KF and n are the Freundlich adsorption constants. ε (kJ/mol) = −RTln (Ce/ Cs) is the effective adsorption potential, where Cs (mg/L) is the water solubility of adsorbate, R is the universal gas constant (8.314 × 10−3 kJ/mol·K), and T is the absolute temperature (K). E and b are the DAM fitting parameters.
■
RESULTS AND DISCUSSION Characterization of Adsorbents. Brown GOs solutions were reduced by hydrazine to yield stable black suspensions of 13199
dx.doi.org/10.1021/es5034684 | Environ. Sci. Technol. 2014, 48, 13197−13206
Environmental Science & Technology
Article
because of the incomplete chemical reduction of −OH, CO and C−O−C functional groups on GOs (oxygen contents in the range of 54.1−54.7%, SI Table S5).33 The commercial graphene had relatively lower oxygen content (∼9.4%) than the two chemically reduced graphene oxides, followed by SWCNTs, MWCNTs, and graphite. Similar results were also obtained by XPS measurements (SI Table S4). The relative higher N content (3.86−4.25%) of rGOs than that of GOs might come from hydrazine in the reduction reaction. Adsorption Isotherms. Adsorption data of KEP, CBZ, and BPA onto the six adsorbents are presented in Figure 2; all
> rGO2 > MWCNTs > graphene > graphite. Maximum adsorption capacities (Q0) of KEP, CBZ, and BPA were positively correlated with the surface areas of the six adsorbents (SI Figure S3A). To account for the role of surface area, the adsorption isotherms were normalized by surface area (SI Figure S4). The isotherms still scattered after the normalization, indicating the surface area was not the only contributing factor in the adsorption process. The lower normalized Q0 of the three PPCPs onto graphenes and CNTs than graphite might be due to (i) weaker hydrophobic interaction as a result of the higher oxygen contents on graphenes and CNTs and (ii) the impediment for the accessible adsorption sites because of the presence of water clusters formed on the oxygen-containing functional groups of graphenes and CNTs.34 The largest normalized Q0 of PPCPs onto graphite also indicated that the adsorption was mainly driven by the direct interaction with the graphene surface.35,36 It has been reported that micropores also accounted for the adsorption of small molecular organic compounds by microporous adsorbents.37 Q0 values of the three PPCPs and the micropore volumes of the six adsorbents had a linear relationship (SI Figure S3B), but the micropore volume normalized isotherms scattered too (SI Figure S4), suggesting that micropore-filling was also an important but not the only factor influencing PPCPs adsorption. Single point adsorption coefficients (Kd) were calculated at Ce = 0.05 Cs and Ce = 0.5 Cs based on DAM fitting results (Table 1). Apparently, Kd values of the three PPCPs adsorption onto graphenes followed the order of KEP > CBZ > BPA, however, Kd was in the order of CBZ > KEP > BPA for SWCNTs, MWCNTs, and graphite. The intensity of hydrophobic interaction can be predicted by the octanol−water distribution coefficient (Kow) of organic chemicals. If hydrophobic interaction was the dominant factor for the three PPCPs adsorption, the adsorption isotherms onto CNTs and graphite should also follow the order of KEP > CBZ > BPA. Although the order of Kd values onto graphenes was in accordance with the PPCPs hydrophobicity, the hydrophobic interaction cannot play significant roles because the surfaces of the three graphenes were more hydrophilic (due to their higher oxygen contents, SI Table S4) compared with CNTs and graphite. Therefore, there must be additional mechanisms controlling the adsorption. To single out nonhydrophobic interaction mechanisms, adsorption isotherms and Kd values were normalized by Kow values of the three PPCPs (SI Figure S5, Table 1). After Kow normalization, the adsorption isotherms of the three PPCPs onto a given adsorbent followed the order of BPA > CBZ > KEP, which followed the reverse order of their hydrophobicity, implying the importance of other specific interactions in the adsorption process. Considering the characteristics of graphenes and the three PPCPs, these nonhydrophobic mechanisms can be evaluated as follows:9,35,38−40 (i) hydrogen bonding between the adsorbates −COOH, −NH2, −OH groups and the oxygen-containing functional groups of graphenes; (ii) electrostatic interaction between negatively or positively charged graphenes and PPCPs; and (iii) π−π interactions between πelectrons of graphenes surfaces and the adsorbates aromatic rings. Hydrogen bond formed between functional groups of aromatic compounds and carbon nanotubes has been reported to be one, but not the most important, adsorption mechanism. In this study, hydrogen bond was also unlikely to be a
Figure 2. Adsorption of KEP (A), CBZ (B), and BPA (C) onto SWCNTs (○), rGO1 (△), rGO2 (▽), MWCNTs (□), graphene (◊), and graphite (☆) at pH 6.50. Dashed lines are the curve fitting using DAM.
isotherms were nonlinear. Thus, three common nonlinear adsorption isotherm models were used to fit the adsorption data. Compared with LM and FM (SI Table S6), DAM is much better to describe the adsorption isotherms with relative higher r2adj and lower mean weighted square error (MWSE) (Table 1). Therefore, the following discussion is based on the adsorption parameters calculated from the DAM fitting results. Figure 2 compares the adsorption of the three PPCPs among different adsorbents on a unit mass basis, showing that all adsorption isotherms followed the order of SWCNTs > rGO1 13200
dx.doi.org/10.1021/es5034684 | Environ. Sci. Technol. 2014, 48, 13197−13206
Environmental Science & Technology
Article
Table 1. Fitting Results of KEP, CBZ, and BPA Adsorption Isotherms on SWCNTs, rGO1, rGO2, MWCNTs, Graphene, and Graphite by DAM Q0 (mg/g)
b
E (kJ/mol)
radj2a
SWCNTs rGO1 rGO2 MWCNTs graphene graphite
91.5 62.5 54.1 25.3 15.3 1.97
± ± ± ± ± ±
0.5 0.7 0.6 0.2 0.1 0.02
1.14 1.05 1.14 1.10 1.43 1.17
± ± ± ± ± ±
0.02 0.04 0.03 0.03 0.04 0.03
23.2 22.6 19.1 23.5 19.6 22.1
± ± ± ± ± ±
0.4 0.6 0.4 0.5 0.4 0.4
0.998 0.996 0.996 0.998 0.996 0.997
SWCNTs rGO1 rGO2 MWCNTs graphene graphite
185 115 105 55.0 22.8 3.65
± ± ± ± ± ±
1.0 0.5 0.7 0.4 0.5 0.05
1.99 1.35 1.36 6.46 1.72 8.79
± ± ± ± ± ±
0.04 0.02 0.03 0.32 0.15 1.18
16.8 18.4 17.5 12.9 19.3 13.2
± ± ± ± ± ±
0.2 0.1 0.2 0.1 1.0 0.2
0.998 0.999 0.998 0.991 0.969 0.957
SWCNTs rGO1 rGO2 MWCNTs graphene graphite
199 152 128 59.4 26.0 4.14
± ± ± ± ± ±
2.0 1.5 1.0 0.3 0.3 0.03
1.76 1.36 1.45 2.20 1.57 1.68
± ± ± ± ± ±
0.08 0.04 0.04 0.06 0.08 0.06
22.0 24.0 24.7 21.4 19.7 27.4
± ± ± ± ± ±
0.7 0.5 0.5 0.3 0.7 0.7
0.991 0.997 0.997 0.996 0.990 0.994
MWSEb KEP 0.000170 0.000486 0.000598 0.000337 0.000685 0.000363 CBZ 0.000481 0.000119 0.000216 0.00107 0.00948 0.00784 BPA 0.000839 0.000318 0.000205 0.000206 0.00100 0.000434
Kd1c (L/kg)
Kd1/KOWd (L/kg)
Kd2c (L/kg)
Kd2/KOWd (L/kg)
19200 12000 9710 5210 3370 406
14.5 9.10 7.36 3.95 2.56 0.308
3190 2100 1830 873 559 69.0
2.42 1.60 1.39 0.662 0.424 0.0520
20900 10500 9140 9210 2600 643
74.3 37.2 32.4 32.7 9.24 2.28
3220 1870 1700 983 392 65.0
11.4 6.65 6.02 3.49 1.39 0.231
7470 5020 4490 2490 833 168
47.1 31.7 28.3 15.7 5.26 1.06
1020 751 641 310 130 21.0
6.46 4.74 4.04 1.96 0.822 0.132
a Adjusted square of correlation coefficient. bMean weighted square error, equal to 1/ν[(Q0measured − Q0model)2/Q0measured2], where ν is the amount of freedom, equal to N − 3 for DAM, where N is the number of observations. cDistribution coefficient calculated from DAM at Ce = 0.05 Cs and Ce = 0.5 Cs, respectively. dDistribution coefficients normalized by KOW.
adsorption of aromatic chemicals onto CNTs.29,36,38,40 Surfaces of graphenes also contained both the π-electron-depleted regions (because of the presence of defects or functional groups) and π-electron-rich regions (because of the delocalization of π electrons), showing the π-electron acceptor or donor properties.44 As an electron-donating group, −OH makes the benzene rings of BPA electron-rich.39 The unshared pair of electrons of −N(NH2)O can also result in electron conjugation with the π electrons of the benzene rings,45 making CBZ to be electron-donor. KEP can function as electron-acceptor resulting from the strong electron-withdrawing ability of the ketone group.35 Therefore, π−π EDA interaction is expected to affect the adsorption of the three PPCPs onto graphenes. Graphite is free of functional groups and can only involve the π−π EDA interaction in adsorption. The Kow normalized PPCPs adsorption onto graphite followed the order of BPA > CBZ > KEP, indicating the π−π EDA interaction between graphite and the three PPCPs was also in the same order. Therefore, π−π EDA interaction also contributed to PPCPs adsorption onto graphenes and CNTs. Effect of pH. Figure 3 shows the effects of pH on adsorption of the three PPCPs. Clearly, for a given adsorbate, all the six adsorbents exhibited almost similar pH-dependent patterns, indicating the predominant role of graphene surfaces in adsorption. On the other hand, for a given adsorbent, the three PPCPs had very different pH-dependent adsorption, suggesting that pH effects were somewhat adsorbate-specific. Changing pH from 2.0 to 12.0 decreased adsorption of KEP on all adsorbents by more than 4 times. Adsorption of CBZ was almost pH-independent over the pH range of 2.0−12.0. BPA adsorption changed slightly as the pH increased from 2.0 to 8.0, followed by a significant decrease of adsorption over its pKa. It should be pointed out that the observed pH effects on adsorption correlated well with the pH-regulated distribution of
predominant factor controlling PPCPs adsorption onto graphenes for the following reasons. First, graphite contained no or few oxygen-containing functional groups and could not form hydrogen bond, but the Kow normalized adsorption of the three PPCPs still followed the order of BPA > CBZ > KEP, indicating the presence of other adsorption interactions. Second, compared with the hydrogen bond between the three PPCPs and functional groups of graphenes, the interaction between water molecules and the polar sites of graphenes to form water clusters was much stronger.39,41,42 Third, oxygen-containing functional groups of graphenes would be ionized because the tested solution pH (6.50) exceeded their pHpzc (2.0−3.0 and CBZ > BPA (Table 2, SI Figure S8), which positively related to the decrease of their hydrophobicities. Higher suppression of KEP adsorption than CBZ and BPA might be partly attributed to the highest hydrophobicity of KEP, which led to higher adsorption KEP to HA in the aqueous phase. At the tested solution pH (6.50), both HA and SDBS molecules were negatively charged. KEP was also deprotonated because of its low pKa. Thus, besides the direct competition for the adsorption sites or pore blockage effect of HA and SDBS, KEP adsorption was further suppressed by the electrostatic repulsion between the adsorbed HA (SDBS) and negatively charged KEP. Whereas for CBZ and BPA, the electrostatic attraction between oppositely charged CBZ (or BPA) and adsorbed HA (or SDBS) may compensate the suppression by HA and SDBS to some extent. It should be noted that HA could highly solubilize CBZ molecules as indicated from the solubility enhancement of CBZ by HA (SI Figure S10). Therefore, besides competition
■
ENVIRONMENTAL IMPLICATION The interaction between graphenes and three selected typical PPCPs was investigated in comparison with CNTs and graphite under various water chemistry conditions. As exhibited in this present work, graphenes, especially rGOs, possessed high adsorption capacity for the three PPCPs. The adsorption capacities were further compared with activated carbon (SI Table S7), which shows Q0/Asurf values of rGOs were much higher than that of activated carbon even though activated carbon has the largest surface area. All these findings suggested that graphenes could be potentially used as superior adsorbents. However, the current energy usage of graphenes production is estimated to be as high as 1100 MJ/kg,50 which is about 50-fold greater than that of activated carbon. Thus, more efforts are still needed to produce high-surface-area graphenes in large scale at relative low cost. In addition, the strong interaction between graphenes and PPCPs may potentially affect their environmental fate and transport and possibly toxicity and risks. Both graphenes and PPCPs have toxic effects on aquatic organisms,11,32 but their synergistic/antagonistic effects have received little attention to date. Therefore, it is important to consider the coexistence of graphenes and PPCPs when their potential environmental risks are evaluated. Also, the presence of HA and SDBS exhibited different influences on graphenes− PPCPs interaction. Because of the ubiquitous presence of dissolved organic matter (DOM) in water environments, the graphenes−PPCPs interaction should also be investigated in the presence of other DOM, such as proteins and polysaccharides. Furthermore, because of the various kinds and classifications of PPCPs, a predictive model for PPCPs adsorption by graphenes should be established in the future after collecting enough adsorption data (at least 20 compounds).32
■
ASSOCIATED CONTENT
S Supporting Information *
Ten figures and seven tables. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*Phone: +1 413 545 5212; e-mail:
[email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was supported by USDA-NIFA Hatch program (MAS 00475) and NSFC (51178254). F.F.L. thanks the China Scholarship Council to support her study in the U.S. 13204
dx.doi.org/10.1021/es5034684 | Environ. Sci. Technol. 2014, 48, 13197−13206
Environmental Science & Technology
■
Article
(22) Zhang, S.; Shao, T.; Bekaroglu, S. S. K.; Karanfil, T. Adsorption of synthetic organic chemicals by carbon nanotubes: Effects of background solution chemistry. Water Res. 2010, 44 (6), 2067−2074. (23) Gupta, S.; Moulik, S. P. Biocompatible microemulsions and their prospective uses in drug delivery. J. Pharm. Sci. 2008, 97 (1), 22−45. (24) Ying, G. G. Fate, behavior and effects of surfactants and their degradation products in the environment. Environ. Int. 2006, 32 (3), 417−431. (25) Klečka, G. M.; Staples, C. A.; Clark, K. E.; van der Hoeven, N.; Thomas, D. E.; Hentges, S. G. Exposure analysis of bisphenol A in surface water systems in north America and Europe. Environ. Sci. Technol. 2009, 43 (16), 6145−6150. (26) Tixier, C.; Singer, H. P.; Oellers, S.; Müller, S. R. Occurrence and fate of carbamazepine, clofibric acid, diclofenac, ibuprofen, ketoprofen, and naproxen in surface waters. Environ. Sci. Technol. 2003, 37 (6), 1061−1068. (27) Gallagher, S. J.; Trottet, L.; Heard, C. M. Ketoprofen: Release from, permeation across and rheology of simple gel formulations that simulate increasing dryness. Int. J. Pharm. 2003, 268 (1−2), 37−45. (28) Chong, M. N.; Jin, B.; Laera, G.; Saint, C. P. Evaluating the photodegradation of carbamazepine in a sequential batch photoreactor system: Impacts of effluent organic matter and inorganic ions. Chem. Eng. J. 2011, 174 (2−3), 595−602. (29) Pan, B.; Lin, D.; Mashayekhi, H.; Xing, B. Adsorption and hysteresis of bisphenol A and 17α-ethinyl estradiol on carbon nanomaterials. Environ. Sci. Technol. 2008, 42 (15), 5480−5485. (30) Zhao, J.; Wang, Z.; Zhao, Q.; Xing, B. Adsorption of phenanthrene on multilayer graphene as affected by surfactant and exfoliation. Environ. Sci. Technol. 2013, 48 (1), 331−339. (31) Xu, B.; Yue, S.; Sui, Z.; Zhang, X.; Hou, S.; Cao, G.; Yang, Y. What is the choice for supercapacitors: Graphene or graphene oxide? Energy Environ. Sci. 2011, 4 (8), 2826−2830. (32) Zhao, J.; Wang, Z.; White, J. C.; Xing, B. S. Graphene in the aquatic environment: Adsorption, dispersion, toxicity and transformation. Environ. Sci. Technol. 2014, 48 (17), 9995−10009. (33) Stankovich, S.; Dikin, D. A.; Piner, R. D.; Kohlhaas, K. A.; Kleinhammes, A.; Jia, Y.; Wu, Y.; Nguyen, S. T.; Ruoff, R. S. Synthesis of graphene-based nanosheets via chemical reduction of exfoliated graphite oxide. Carbon 2007, 45 (7), 1558−1565. (34) Cho, H. H.; Smith, B. A.; Wnuk, J. D.; Fairbrother, D. H.; Ball, W. P. Influence of surface oxides on the adsorption of naphthalene onto multiwalled carbon nanotubes. Environ. Sci. Technol. 2008, 42 (8), 2899−2905. (35) Ji, L.; Chen, W.; Duan, L.; Zhu, D. Mechanisms for strong adsorption of tetracycline to carbon nanotubes: A comparative study using activated carbon and graphite as adsorbents. Environ. Sci. Technol. 2009, 43 (7), 2322−2327. (36) Ji, L.; Chen, W.; Zheng, S.; Xu, Z.; Zhu, D. Adsorption of sulfonamide antibiotics to multiwalled carbon nanotubes. Langmuir 2009, 25 (19), 11608−11613. (37) Zhu, D.; Pignatello, J. J. Characterization of aromatic compound sorptive interactions with black carbon (charcoal) assisted by graphite as a model. Environ. Sci. Technol. 2005, 39 (7), 2033−2041. (38) Wang, F.; Yao, J.; Sun, K.; Xing, B. Adsorption of dialkyl phthalate esters on carbon nanotubes. Environ. Sci. Technol. 2010, 44 (18), 6985−6991. (39) Chen, W.; Duan, L.; Wang, L.; Zhu, D. Adsorption of hydroxyland amino-substituted aromatics to carbon nanotubes. Environ. Sci. Technol. 2008, 42 (18), 6862−6868. (40) Wang, Z.; Yu, X.; Pan, B.; Xing, B. Norfloxacin sorption and its thermodynamics on surface-modified carbon nanotubes. Environ. Sci. Technol. 2009, 44 (3), 978−984. (41) Zhang, S.; Shao, T.; Bekaroglu, S. S. K.; Karanfil, T. The impacts of aggregation and surface chemistry of carbon nanotubes on the adsorption of synthetic organic compounds. Environ. Sci. Technol. 2009, 43 (15), 5719−5725. (42) Efremenko, I.; Sheintuch, M. Predicting solute adsorption on activated carbon: Phenol. Langmuir 2006, 22 (8), 3614−3621.
REFERENCES
(1) Novoselov, K. S.; Geim, A. K.; Morozov, S. V.; Jiang, D.; Zhang, Y.; Dubonos, S. V.; Grigorieva, I. V.; Firsov, A. A. Electric field effect in atomically thin carbon films. Science 2004, 306 (5696), 666−669. (2) Wassei, J. K.; Kaner, R. B. Oh, the places you’ll go with graphene. Acc. Chem. Res. 2013, 46 (10), 2244−2253. (3) Choucair, M.; Thordarson, P.; Stride, J. A. Gram-scale production of graphene based on solvothermal synthesis and sonication. Nat. Nanotechnol. 2009, 4 (1), 30−33. (4) Obraztsov, A. N. Chemical vapour deposition: Making graphene on a large scale. Nat. Nanotechnol. 2009, 4 (4), 212−213. (5) Gagliardi, M.; McWilliams, A.; Moran, R. Graphene: Technologies, Applications, and Markets; BCC Research Market Forecasting: Wellesley, MA, 2011. (6) Hu, X.; Zhou, Q. Health and ecosystem risks of graphene. Chem. Rev. 2013, 113 (5), 3815−3835. (7) Zhao, G.; Jiang, L.; He, Y.; Li, J.; Dong, H.; Wang, X.; Hu, W. Sulfonated graphene for persistent aromatic pollutant management. Adv. Mater. 2011, 23 (34), 3959−3963. (8) Apul, O. G.; Wang, Q.; Zhou, Y.; Karanfil, T. Adsorption of aromatic organic contaminants by graphene nanosheets: Comparison with carbon nanotubes and activated carbon. Water Res. 2013, 47 (4), 1648−1654. (9) Pei, Z.; Li, L.; Sun, L.; Zhang, S.; Shan, X.; Yang, S.; Wen, B. Adsorption characteristics of 1,2,4-trichlorobenzene, 2,4,6-trichlorophenol, 2-naphthol, and naphthalene on graphene and graphene oxide. Carbon 2013, 51 (0), 156−163. (10) Shen, R.; Andrews, S. A. Demonstration of 20 pharmaceuticals and personal care products (PPCPs) as nitrosamine precursors during chloramine disinfection. Water Res. 2011, 45 (2), 944−952. (11) Liu, J. L.; Wong, M. H. Pharmaceuticals and personal care products (PPCPs): A review on environmental contamination in China. Environ. Int. 2013, 59 (0), 208−224. (12) Ji, L.; Shao, Y.; Xu, Z.; Zheng, S.; Zhu, D. Adsorption of monoaromatic compounds and pharmaceutical antibiotics on carbon nanotubes activated by KOH etching. Environ. Sci. Technol. 2010, 44 (16), 6429−6436. (13) Cho, H. H.; Huang, H.; Schwab, K. Effects of solution chemistry on the adsorption of ibuprofen and triclosan onto carbon nanotubes. Langmuir 2011, 27 (21), 12960−12967. (14) Joseph, L.; Zaib, Q.; Khan, I. A.; Berge, N. D.; Park, Y. G.; Saleh, N. B.; Yoon, Y. Removal of bisphenol A and 17α-ethinyl estradiol from landfill leachate using single-walled carbon nanotubes. Water Res. 2011, 45 (13), 4056−4068. (15) Ji, L.; Chen, W.; Xu, Z.; Zheng, S.; Zhu, D. Graphene nanosheets and graphite oxide as promising adsorbents for removal of organic contaminants from aqueous solution. J. Environ. Qual. 2013, 42 (1), 191−198. (16) Xu, J.; Wang, L.; Zhu, Y. Decontamination of bisphenol A from aqueous solution by graphene adsorption. Langmuir 2012, 28 (22), 8418−8425. (17) Chen, J.; Chen, W.; Zhu, D. Adsorption of nonionic aromatic compounds to single-walled carbon nanotubes: Effects of aqueous solution chemistry. Environ. Sci. Technol. 2008, 42 (19), 7225−7230. (18) Pan, B.; Zhang, D.; Li, H.; Wu, M.; Wang, Z.; Xing, B. Increased adsorption of sulfamethoxazole on suspended carbon nanotubes by dissolved humic acid. Environ. Sci. Technol. 2013, 47 (14), 7722−7728. (19) Wang, X.; Lu, J.; Xing, B. Sorption of organic contaminants by carbon nanotubes: Influence of adsorbed organic matter. Environ. Sci. Technol. 2008, 42 (9), 3207−3212. (20) Zhang, S.; Shao, T.; Karanfil, T. The effects of dissolved natural organic matter on the adsorption of synthetic organic chemicals by activated carbons and carbon nanotubes. Water Res. 2011, 45 (3), 1378−1386. (21) Zhang, D.; Pan, B.; Zhang, H.; Ning, P.; Xing, B. Contribution of different sulfamethoxazole species to their overall adsorption on functionalized carbon nanotubes. Environ. Sci. Technol. 2010, 44 (10), 3806−3811. 13205
dx.doi.org/10.1021/es5034684 | Environ. Sci. Technol. 2014, 48, 13197−13206
Environmental Science & Technology
Article
(43) Li, X.; Pignatello, J. J.; Wang, Y.; Xing, B. New insight into adsorption mechanism of ionizable compounds on carbon nanotubes. Environ. Sci. Technol. 2013, 47 (15), 8334−8341. (44) McDermott, M. T.; McCreery, R. L. Scanning tunneling microscopy of ordered graphite and glassy carbon surfaces: Electronic control of quinone adsorption. Langmuir 1994, 10 (11), 4307−4314. (45) Oleszczuk, P.; Pan, B.; Xing, B. Adsorption and desorption of oxytetracycline and carbamazepine by multiwalled carbon nanotubes. Environ. Sci. Technol. 2009, 43 (24), 9167−9173. (46) Fontecha-Cámara, M. A.; López-Ramón, M. V.; Á lvarez-Merino, M. A.; Moreno-Castilla, C. Effect of surface chemistry, solution pH, and ionic strength on the removal of herbicides diuron and amitrole from water by an activated carbon fiber. Langmuir 2006, 23 (3), 1242−1247. (47) Ni, N.; Yalkowsky, S. H. Prediction of Setschenow constants. Int. J. Pharm. 2003, 254 (2), 167−172. (48) Zhao, J.; Wang, Z.; Ghosh, S.; Xing, B. Phenanthrene binding by humic acid−protein complexes as studied by passive dosing technique. Environ. Pollut. 2014, 184 (0), 145−153. (49) Wang, X.; Shu, L.; Wang, Y.; Xu, B.; Bai, Y.; Tao, S.; Xing, B. Sorption of peat humic acids to multi-walled carbon nanotubes. Environ. Sci. Technol. 2011, 45 (21), 9276−9283. (50) Arvidsson, R.; Kushnir, D.; Sandén, B. A.; Molander, S. Prospective life cycle assessment of graphene production by ultrasonication and chemical reduction. Environ. Sci. Technol. 2014, 48 (8), 4529−4536.
13206
dx.doi.org/10.1021/es5034684 | Environ. Sci. Technol. 2014, 48, 13197−13206
