Macroscopic and Spectroscopic Investigations of the Adsorption of

Apr 16, 2015 - carbon nanotube (CNT), and graphite materials.1,2 In addition, graphene has a wide .... mg/L), and p-nitrotoluene (0−200 mg/L) was ca...
1 downloads 0 Views 2MB Size
Article pubs.acs.org/est

Macroscopic and Spectroscopic Investigations of the Adsorption of Nitroaromatic Compounds on Graphene Oxide, Reduced Graphene Oxide, and Graphene Nanosheets Xiaoxiao Chen†,‡ and Baoliang Chen*,†,‡ †

Department of Environmental Science, Zhejiang University, Hangzhou 310058, China Zhejiang Provincial Key Laboratory of Organic Pollution Process and Control, Hangzhou 310058, China



S Supporting Information *

ABSTRACT: The surface properties and adsorption mechanisms of graphene materials are important for potential environmental applications. The adsorption of m-dinitrobenzene, nitrobenzene, and p-nitrotoluene onto graphene oxide (GO), reduced graphene oxide (RGO), and graphene (G) nanosheets was investigated using IR spectroscopy to probe the molecular interactions of graphene materials with nitroaromatic compounds (NACs). The hydrophilic GO displayed the weakest adsorption capability. The adsorption of RGO and G was significantly increased due to the recovery of hydrophobic π-conjugation carbon atoms as active sites. RGO nanosheets, which had more defect sites than did GO or G nanosheets, resulted in the highest adsorption of NACs which was 10−50 times greater than the reported adsorption of carbon nanotubes. Superior adsorption was dominated by various interaction modes including π−π electron donor−acceptor interactions between the π-electron-deficient phenyls of the NACs and the π-electronrich matrix of the graphene nanosheets, and the charge electrostatic and polar interactions between the defect sites of graphene nanosheets and the −NO2 of the NAC. The charge transfer was initially proved by FTIR that a blue shift of asymmetric −NO2 stretching was observed with a concomitant red shift of symmetric −NO2 stretching after m-dinitrobenzene was adsorbed. The multiple interaction mechanisms of the adsorption of NAC molecule onto flat graphene nanosheets favor the adsorption, detection, and transformation of explosives.



INTRODUCTION Graphene is a two-dimensional (2D) monolayer of sp2 hybridized carbon atoms that are packed in a hexagonal honeycomb lattice,1 and is the basic building block of fullerene, carbon nanotube (CNT), and graphite materials.1,2 In addition, graphene has a wide range of potential applications due to its unique structure and outstanding mechanical, optical, and electronic properties.1−3 Because of their hydrophobicity and large surface area,3 graphene nanosheets are expected to serve as a good adsorbent for organic compounds.4 Recent studies have indicated that graphene is a superior adsorption material for nonpolar organic compounds,5−11 such as polycyclic aromatic hydrocarbons (PAHs)5−7,11 and chlorinated aromatic hydrocarbons.8−10 In contrast with CNTs,12−14 studies regarding the interactions of graphene materials with organic pollutants have only recently begun.4 The surface properties and adsorption mechanisms of graphene are vital for its environmental application and for designing novel graphenebased materials.4,11,15 In addition, graphene nanosheets generally exhibit nanometer-scale corrugations.1,5 Furthermore, the associated strain and curvature can markedly influence local reactivity and adsorption.5,16−20 The adsorptive sites of graphene surfaces are various and complex,4,5 and they include flat π networks, © XXXX American Chemical Society

wrinkles, defects, and functional groups attached to the surfaces and edges of the graphene nanosheets.11 The adsorption availability of these sites is regulated by surface properties (such as functional groups) and graphene aggregation5,7,11 and can be distinctly probed by different sorbates with various sizes and polarities.5,7 The high affinity of PAHs for graphene nanosheets is dominated by π−π interactions with the flat surface and the sieving effects of the powerful groove regions that are formed by wrinkles.5 Similarly, Wang and co-workers7 suggested that pore filling and flat surface adsorption both contribute to the adsorption of PAHs onto reduced graphene oxide (RGO). Dramatic changes in the conformation and aggregation of the surface of RGO during PAH adsorption alter subsequent adsorption.5 After graphene is loaded with silica particles, PAH adsorption by graphene nanosheets dramatically increases because the aggregation between the graphene layers via the strong π−π interactions is overcome.11 In general, graphene nanosheets concomitantly contain polarized electron-rich sites and electron-depleted sites, and Received: November 10, 2014 Revised: April 5, 2015 Accepted: April 16, 2015

A

DOI: 10.1021/es5054946 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

Figure 1. FTIR spectra (a) and Raman spectra (b) of graphene oxide (GO), reduced graphene oxide (RGO), and graphene (G).

surface defects can result in significant delocalization of π electrons.9,21 Therefore, the electronic properties and functional groups of adsorbate molecules play important roles in the adsorption mechanisms of graphene. For example, adsorption to graphene improves as the π-electron donor ability of the solute increases,9 and it is dominated by π−π electron donor− acceptor (EDA) interactions with the graphitic surface of the adsorbent. Similarly, Chen et al. reported that the strong adsorptive interactions of nitroaromatics onto CNTs resulted from π−π EDA interactions between the nitroaromatic molecules (electron acceptors) and graphene nanosheets (electron donors) of CNTs.12 However, few studies have investigated the molecular-level mechanisms of nitroaromatic compound (NAC) adsorption onto graphene nanosheets.18,22,23 In addition, spectroscopic evidence is required to prove the occurrence of site-specific EDA interactions or charge transfer process. NACs are used in pesticides, explosives, and as intermediates in the synthesis of dyes and other chemicals24,25 and are common organic pollutants. Because of their molecular and electronic structure, aromatic explosives are strongly adsorbed onto graphene surfaces. The main objectives of the present study were to examine the molecular mechanisms controlling the adsorption of NACs onto graphene nanosheets and to obtain spectroscopic evidence regarding this adsorption. The adsorption processes of m-dinitrobenzene, nitrobenzene, and pnitrotoluene onto graphene oxide (GO), reduced graphene oxide (RGO), and graphene (G) nanosheets were compared. Spectroscopic methods were used to observe the specific electron transfer between the NACs and graphene-based materials. In addition, the influences of pH on the adsorption of NACs onto graphene-based materials and on spectroscopic changes were monitored.

Characterization of GO, RGO, and G. The properties of the GO, RGO, and G nanosheets were measured using elemental (C, H, N) analyses, BET-N2 specific surface area and Raman spectroscopy. The surface elemental composition and functional groups were observed using X-ray photoelectron spectroscopy (XPS) and Fourier transform infrared spectroscopy (FTIR). The detailed methods used in this study are presented in the Supporting Information (SI). Adsorption Experiments. m-Dinitrobenzene, nitrobenzene, and p-nitrotoluene were selected as model solutes with different electron-withdrawing groups. Their selected physicochemical properties are listed in Table S-1 of the SI. Adsorption experiments were conducted using a batch approach reported in our previous studies.5,29 The adsorption of m-dinitrobenzene (0−400 mg/L), nitrobenzene (0−500 mg/L), and p-nitrotoluene (0−200 mg/L) was carried out by mixing a certain amount of GO, RGO, or G in 8-mL vials that were sealed with Teflon-lined screw caps at 25 ± 1 °C. The vials were placed on a shaker and agitated in the dark at 120 rpm for 3 d to reach an apparent equilibrium. The adsorption kinetics (SI Figure S-1) indicated that apparent equilibrium was achieved after 24 h. After separation, the supernatant was immediately analyzed using high-performance liquid chromatography (HPLC). Detailed information is presented in the SI. After adsorption, the graphene material was collected and freeze-dried for 24 h prior to FTIR analysis. Effects of pH Values on Adsorption of m-DNB onto GO and RGO. The effects of pH on the adsorption of mdinitrobenzene were investigated by adjusting the pH from 2 to 12 using 0.01 mol/L HCl and 0.1 mol/L NaOH. The initial concentration of m-dinitrobenzene was 400 mg/L, and the solid-to-solution ratios were 40 mg/8 mL and 2.5 mg/8mL for GO and RGO, respectively. The equilibrium pH value of the solution was recorded by a pH meter (Mettler Toledo) at room temperature. After adsorption, the GO was collected and freeze-dried for 24 h prior to FTIR analysis.



EXPERIMENTAL SECTION Preparation of GO, RGO, and G. GO nanosheets were synthesized from natural graphite flakes (325 mesh, 99.8%, Alfa Aesar) using the modified Hummer method.26 G nanosheets were obtained by thermal reduction of GO nanosheets,27 and RGO nanosheets were obtained by reducing GO nanosheets with NaBH4 at 80 °C for 2 h.28



RESULTS AND DISCUSSION

Structural Characterization. The elemental composition and BET surface areas of the GO, RGO, and G nanosheets are listed in SI Table S-2. The C content of the GO nanosheets (49.14%) was low, indicating that the π-system of the graphene B

DOI: 10.1021/es5054946 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology material was damaged. The C content reached 61.95% in the RGO nanosheets and 85.07% in the G nanosheets, which suggested that thermal reduction can reduce GO thoroughly and that the surface of RGO has more oxygen-containing functional groups than does G. The FTIR spectra of GO, RGO, and G are shown in Figure 1a. For GO, a broad peak at 3357 cm−1 was assigned to the −OH stretching vibration. The FTIR spectrum of GO showed sharp CO (1721 cm−1), aromatic CC (1619 cm−1), carboxyl OCO (1358 cm−1), epoxy CO (1228 cm−1), and alkoxy CO (1048 cm−1) stretching vibrations.30 For G, a peak at 1574 cm−1 corresponded to the benzene ring CC stretching vibration.10 In addition, weak peaks at 1739 and 1240 cm−1 were assigned to carboxyl CO and epoxy CO stretching vibrations, respectively. The FTIR spectrum of RGO showed relatively weak stretching vibrations of CO (1732 cm−1), aromatic CC (1582 cm−1), and alkoxy CO (1048 cm−1). The abundance of O-containing functional groups was significantly higher on GO than on RGO or G, which is consistent with their elemental composition. The FTIR spectra indicated that not only GO but also RGO and G had some Ocontaining functional groups and defect sites on their surfaces. Raman spectroscopy is the most direct and nondestructive technique for characterizing the structure and quality of carbon materials, particularly for identifying defects and ordered and disordered structures.5,31 The Raman spectra of GO, RGO, and G exhibited two significant peaks at approximately 1350 and 1580 cm−1 (Figure 1b), which corresponded to the D band and G band. The G band, which is related to the E2g vibration mode of sp2 carbon domains, can be used to determine the degree of graphitization. In contrast, the D band is associated with structural defects and the partially disordered structures of the sp2 domains.5 The ID/IG value of G (0.87) was less than that of the RGO (1.13). This phenomenon indicated that the graphitic degree of the G sample was improved due to the superior reduction and the self-repair of the graphene layer annealing at high temperature in comparison with the RGO obtained by chemical reduction. The ID/IG ratio of RGO was the highest among the graphene materials, which suggested that the newly formed sp2 domains RGO were smaller but more prevalent than those in G.31 Thus, more defects/edges were created in RGO. The distribution of the various C (C1−C4) and O (O1−O3) components in the GO, RGO, and G materials were analyzed using XPS (Figure 2 for C 1s and SI Figure S-2 for O 1s). The C1 (284.4 eV), C2 (285.9 eV), C3 (287.9 eV), and C4 (289.9 eV) components were assigned to CC, CO, CO, and C(O)O groups, respectively. Regarding the O 1s spectra, the three peaks at (i) 530.5, (ii) 533.2, and (iii) 535.8 eV corresponded to (i) quinone-type [C(O)O, CO], (ii) −C−O and −OH groups, and (iii) chemisorbed oxygen (carboxylic groups) and/or water,32−38 respectively. The atomic percentage of O heteroatoms and the O/C atomic ratio were calculated and are listed in SI Table S-2. The content of the oxygen functionalities sharply decreased after the reduction process, which indicated that most of the epoxide and hydroxyl functional groups on the GO surface were successfully removed and that the aromatic or conjugated systems were recovered.39 The fitting results indicated that RGO had more quinone-type C(O)O and CO (530.46 eV) on its edges than the synthesized G and more −C−O and −OH (533.2 eV) on its surface.40 Consequently, the RGO lamellas had relatively low hydrophobicity and could act as

Figure 2. C 1s XPS fine scan spectra and the deconvolution of the C 1s for graphene oxide (GO), reduced graphene oxide (RGO), and graphene (G).

spacers that keep the layers from coagulating due to hydrophobicity and improve the dispersibility.40 The N2 adsorption−desorption isotherms of graphene materials are depicted in SI Figure S-3. According to the IUPAC classification, the N2 isotherms for G are of type II, suggesting that G is of nonporous material. Whereas for GO and RGO, the N2 isotherms are of type I, which is reflective of a microporous structure. The distinct structures of GO, RGO, and G impact their adsorption capabilities and interaction mechanisms toward NACs. Adsorption Isotherms for NACs on GO, RGO, and G. The adsorption isotherms of m-dinitrobenzene, nitrobenzene, and p-nitrotoluene onto GO, RGO, and G are shown in Figure 3. The isotherms were fit using the Freundlich and Langmuir models (SI Table S-3). On the basis of Figure 3, it is found that the adsorption isotherms of GO, RGO, and G are fitted well by the Freundlich model over the Langmuir model. Interestingly, the deviation of experimental data points from Langmuir model follows the order of GO < G < RGO, which indicates that the heterogeneity of the adsorption sites is greatest for RGO among the tested graphene materials because Langmuir model describes the one-type-sorption set isotherm without any heterogeneity of sorption sites. Furthermore, the Freundlich N value indicates the heterogeneity of the surface of the graphene C

DOI: 10.1021/es5054946 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

Figure 3. Adsorption isotherms of m-dinitrobenzene (a), nitrobenzene (b), and p-nitrotoluene (c) onto graphene oxide (GO), reduced graphene oxide (RGO), and graphene (G).

Figure 4. Hexadecane−water partition coefficient (KHW) normalized adsorption coefficient (Kd/KHW) at different concentrations (Ce) of mdinitrobenzene (DNB), nitrobenzene (NB), and p-nitrotoluene (p-NT) by graphene oxide (a), graphene (b), and reduced graphene oxide (c).

materials, and a higher N value indicates a homogeneous surface with a narrow adsorption site distribution.8,41 Among the graphene materials, the Freundlich N values (0.693−0.843) of GO were the highest, which was attributed to the relatively homogeneous surface of GO due to surface hydration in the aqueous solution. The lowest Freundlich N values (0.253− 0.306) were those of RGO, which displayed strong nonlinear adsorption. This finding indicates that the RGO surface was the most heterogeneous, which is consistent with the structural characterization according to the Raman and XPS spectra. The residual O-containing functional groups result in a more heterogeneous RGO surface with more defects. For G, the Freundlich N values were 0.313−0.422. Similarly, the Freundlich N value approached 0.3 for the adsorption of 2naphthol, 1-naphthamine, and tylosin onto G9 and for the adsorption of PAHs onto reduced GO.7 The Freundlich N values for a given graphene material were similar for mdinitrobenzene, nitrobenzene, and p-nitrotoluene, indicating that the surface properties of the graphene materials dominated the adsorption process. As shown in Figure 3, RGO exhibited the strongest adsorption capability for NACs, and GO displayed the lowest adsorption. The saturated adsorption amounts (Qo) of RGO calculated from Langmuir model are 265.7 mg/g for mdinitrobenzene, 260.9 mg/g for nitrobenzene, and 238.8 mg/g for p-nitrotoluene, which are higher than the Qo values of naphthalene (126 mg/g) and phenanthrene (142 mg/g) onto graphene.5 Obviously, RGO is a powerful adsorbent for polar organic pollutant control. Although G had the most hydrophobic surface and the largest surface area, it adsorbed significantly less NAC than RGO did. Therefore, in addition to hydrophobic and π−π stacking interactions, other specific mechanisms should be involved in the interactions of graphene materials with NACs. To evaluate the adsorption capability of

graphene materials toward different organic pollutants, the hexadecane−water partition coefficient (KHW) normalized adsorption coefficient (Kd/KHW) of NACs at different equilibrium concentrations (Ce) were calculated to rule out the effects of NAC hydrophobicity. The relationships between Kd/KHW and Ce for GO, RGO, and G are presented in Figure 4, and the relationships between Kd/KHW and different sorbed concentrations (Qe) are presented in SI Figure S-4. The adsorption capability of NACs onto the graphene materials decreased as the sorbate equilibrium concentration increased (Figure 4) or as the sorbate sorbed amount increased (SI Figure S-4). This trend is distinct from the adsorption behavior of PAHs,5 where an unexpected peak was found in the plot of the adsorption coefficient (Kd) versus the PAH equilibrium concentration. For GO, the Kd/KHW ratio approached 10 for nitrobenzene and p-nitrotoluene and increased to 20−100 for m-dinitrobenzene. For G, the Kd/ K HW ratio reached up to 10 4 at low NAC sorbate concentrations (Figure 4b), which was greater than the adsorption of PAHs onto G nanosheets.5 For RGO, the Kd/ KHW ratio of NACs was dramatically higher, reaching up to 105 at low sorbate concentrations (Figure 4c). The large Kd/KHW ratio indicates the strong interaction of a nanomaterial with an organic pollutant. On the basis of Figure 4, the adsorption capability of the graphene materials increased with the πacceptor strength of sorbate (p-nitrotoluene < nitrobenzene < m-dinitrobenzene), which corresponded with the electron deficiency potential of the sorbate rather than the hydrophobicity (KHW). The electron deficiency potential is largest for m-dinitrobenzene due to the presence of two electron density withdrawing groups (−NO2). The electron deficiency potential of p-nitrobolutene is less than that of nitrobenzene due to the presence of the methyl group in p-nitrobolutene. At the same sorbed concentrations, the Kd/KHW ratios of NACs onto D

DOI: 10.1021/es5054946 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

Figure 5. FTIR spectra of (a) graphene oxide (GO), (b) graphene (G), and (c) reduced graphene oxide (RGO) before and after adsorption of mdinitrobenzene (DNB).

Table 1. Peak Wavenumbers and Intensities of the Asymmetric (Iasym) and Symmetric (Isym) NO2 Stretching Bands of mDinitrobenzene (DNB) Adsorbed by Graphene Oxide (GO), Reduced Graphene Oxide (RGO), or Graphene (G), and the Iasym/Isym Ratios of DNB Adsorbed onto GO under Different Solution pH Values DNB G−DNB RGO−DNB GO−DNB GO−DNB GO−DNB GO−DNB GO−DNB GO−DNB GO−DNB GO−DNB GO−DNB GO−DNB GO−DNB

pH

Iasym

vasym(NO)

Isym

vsym(NO)

Iasym/Isym

6.52 9.00 2.65 3.01 4.73 5.47 6.00 6.92 7.69 8.63 8.85 10.98 11.92

1.190 0.093 0.062 0.062 0.066 0.077 0.085 0.085 0.121 0.129 0.132 0.165 0.197 0.180

1522 1533 1530 1535 1535 1534 1534 1535 1534 1533 1535 1533 1537 1536

1.455 0.077 0.041 0.165 0.144 0.143 0.146 0.145 0.175 0.183 0.178 0.206 0.225 0.204

1344 1341 1342 1345 1346 1346 1345 1346 1346 1346 1346 1346 1346 1346

0.82 1.20 1.52 0.37 0.46 0.54 0.58 0.59 0.70 0.71 0.75 0.80 0.87 0.88

graphene materials were larger than those of PAHs.5 These observations suggest that the adsorption mechanism of NACs onto graphene materials is related to the nitro-functional groups in NACs and that the hydrophobic interactions are not important, both of which are consistent with previous studies on adsorption of NACs onto graphite and charcoal.42 Interestingly, the adsorption capability of NACs onto RGO and G in this study was 10−50 times greater than that reported on CNTs.12 Next, the mechanisms underlying the superior adsorption of NACs by graphene materials will be elucidated using spectroscopic investigations. FTIR of m-Dinitrobenzene Adsorbed to Graphene Materials. Overall, FTIR is a useful tool to examine the nature of the dehydrated complexes, which provides a reference for probing the interactions of NACs with graphene surfaces. For m-dinitrobenzene, the peaks at 1522 and 1344 cm−1 correspond to the asymmetric stretching vibrations (vasym) and symmetric stretching vibrations (vsym) of the N−O bonds in the −NO2 groups of m-DNB (SI Figure S-5). The FTIR spectra of GO, RGO, and G before and after m-DNB adsorption are shown in Figure 5. The adsorption of m-DNB onto graphene materials is evident from the FTIR spectra in Figure 5, which clearly show the vasym and vsym absorbance peaks that are characteristic of an N−O bond. And it is reported that the bands at 711 and 915 cm−1 were assigned to N−O and C−H out-of-plane wagging deformations.24 Interestingly, the spectrum of m-dinitroben-

zene adsorbed onto the graphene material was rather different from the spectrum of the solid m-dinitrobenzene, which suggested a strong interaction between m-dinitrobenzene and the graphene surfaces. The peak wavenumber of the NO2 stretching models and the relative intensity of the asymmetric NO2 stretching band versus the symmetric NO2 stretching band of adsorbed m-dinitrobenzene (Iasym/Isym) are presented in Table 1. After adsorption onto the graphene material, the asymmetric NO2 stretching peak of m-dinitrobenzene was blue-shifted from ∼1522 cm−1 (pure sorbate) to ∼1534 cm−1 for GO−DNB, ∼1530 cm−1 for RGO−DNB, and ∼1533 cm−1 for G-DNB. Simultaneously, the symmetric NO2 stretching peak at ∼1344 cm−1 for m-DNB remained unchanged for GO−DNB and was red-shifted to ∼1341 cm−1 in the RGO−DNB and G−DNB samples. As shown in Table 1, the Iasym/Isym ratio was 0.82 for pure m-DNB, and after adsorption onto GO the Iasym/Isym ratio decreased to 0.37. This decrease indicated that the intensity of the asymmetric NO2 stretching band at 1534 cm−1 was significantly lower than that of the vsym (NO) band at 1340 cm−1. When m-dinitrobenzene was adsorbed onto G or RGO, the relative intensities of the vasym (NO) and vsym (NO) bands (Iasym/Isym) changed to 1.20 and 1.52 (Table 1), respectively. Thus, the adsorbed state of m-DNB on G and RGO was completely distinct from that on GO. Mechanisms of Nitroaromatic Compound Adsorption onto GO, RGO, and G. The adsorption mechanisms of E

DOI: 10.1021/es5054946 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

Figure 6. Schematic of the interactions of graphene oxide (a), reduced graphene oxide (b), and graphene (c) with m-dinitrobenzene. (The used color schemes indicate as follows: gray for carbon, red for oxygen, white for hydrogen, and blue for nitrogen).

nitroaromatics onto graphene materials. We hypothesized that additional interactions (such as polar interaction and charge electrostatic interaction) between the electron-withdrawing −NO2 groups of NACs and the defect sites/edges or dopant water of graphene nanosheets contributed to the NAC adsorption. This hypothesis was tested using FTIR. The vasym and vsym of the N−O groups in the NACs and their relative intensities (Iasym/Isym) are very sensitive to different surface interaction modes, which could be used to probe the interactions between −NO2 groups of NACs with graphene materials and their charge transfer direction. After NACs were adsorbed onto the graphene material, the vasym (NO) band was blue-shifted and the vsym (NO) band was red-shifted. Correspondingly, the split between the asymmetric and symmetric stretching bands of the −NO2 groups was larger after adsorption. The blue-shifting of the vasym (NO) band resulted from the strengthening of the N−O bonds, whereas the red-shifting of the vsym (NO) band was attributed to the weakening of the C−N bonds. It is understandable that the strengthening of the N−O bond is the result of the increase of charge density in orbitals because of electron transfer process from graphene surface to the electron-withdrawing −NO2 groups of the NAC. The charge transfer results from a mix of various interactions modes which may involve π−π donor− acceptor interaction, electrostatic interactions, and polarization. In contrast, when NACs are adsorbed onto K+-hectorite or K+smectite, the charge transfer of the −NO2 groups to K+ is suggested as the dominant mechanism, which is evidenced by a red-shift of the vasym (NO) band accompanying the blue-shift of the vsym (NO) band.24,25,46 Furthermore, molecular simulations showed that electron donating or withdrawing substituents of nitrobenzene tended to either add charge to, or remove charge from, the nitro groups, leaving the electrodensity of the aromatic ring itself relatively unchanged.46 Therefore, specific interactions involving the nitro group over the aromatic ring dominated the adsorption of nitrobenzenes onto K+-smectite.46 On the basis of the IR spectroscopic evidence in the current study, the interactions of the −NO2 groups of NACs with graphene materials originate from strong electron charge transfer (EDA) rather than from proton charge transfer (Hbonds). After m-DNB interacted with GO (Iasym/Isym = 0.37), the relative intensity of the asymmetric NO2 stretching band at 1534 cm−1 obviously decreased relative to the vsym (NO) band at 1340 cm−1. Thus, it is reasonable to assume that the NO2 groups of m-DNB adsorbed onto GO mainly exist in a chelating bidentate configuration with O-containing groups/edges/ defects/and the hydrogen of H2O (Figure 6a). When mdinitrobenzene was adsorbed onto G, the Iasym/Isym ratio increased to 1.20, which indicated that the nitro groups interacted with the G surface with a bigger O−N−O bond angle and a transfer from the electron rich sites to the −NO2 groups via electrostatic interactions, and polarization (Figure

graphene materials at the molecular level are highly dependent on the structure of the sorbate as a probe and the surface properties of the graphene nanosheets because graphene surfaces are more heterogeneous. Adsorption by the graphene materials occurred in the following order: m-dinitrobenzene > nitrobenzene > p-nitrotoluene. The adsorption of NACs onto the surface of graphene is governed by electronic properties rather than by hydrophobic effects of the sorbate. Furthermore, NACs are π-electron-deficient molecules that act as π-electron acceptors. However, graphene acts as a π-electron-rich matrix because of its delocalized π-bond structure. Thus, NACs can be adsorbed onto the surface of graphene through π-electron donor−acceptor interactions,43 which are contributed by dispersion and perhaps polar interactions of the whole aromatic molecule with graphene. Good electron acceptors (i.e., mDNB) exhibit an electron-deficient π-system due to two electron-withdrawing substituents, which strongly adsorb onto graphene nanosheets. In addition, m-DNB adsorbs onto graphene and interacts with graphene through its aromatic structure, which corresponds with previously reported results.42−45 The presence of the methyl group (electron donor group) in p-nitrotoluene makes the nitro group less electron-withdrawing, thus p-NT exhibits relatively low affinity. After ruling out the effects of sorbate hydrophobicity, the adsorption of NACs (electron-deficient) onto graphene materials was significantly higher than that of PAHs5 and the electron-rich aromatic pollutants (e.g., 2-naphthol).10 These results are consistent with a study that showed that the adsorption of NACs onto CNTs is much stronger than other aromatic compounds of similar hydrophobicity or less hydrophobicity.12 The availability of adsorptive sites on graphene materials for NACs is regulated by surface properties and defects. According to FTIR and XPS analyses, the edges and surfaces of GO have abundant O-containing groups, such as CO, C(O)O, and C−O/O−H, which display hydrophilic properties and hinder π−π EDA interactions. The oxygen contents of G and RGO are lower, and the π-conjugation of the nanosheets is significantly increased, compared to GO. GO had a relatively low surface area (7.707 m2/g) compared to RGO (139.7 m2/g), whereas G had the highest surface area (298.8 m2/g). On the basis of the hydrophobicity and surface area of the graphene materials, it is reasonable to speculate that G should be the best adsorbent for all of the tested nitroaromatics. However, the adsorption capability of RGO for NACs was surprisingly much greater than that of G, with the following order: RGO > G > GO. The basal plane near edges and defects are electron-rich regions.42 A certain amount of O-containing groups/defects/edges on the surface of RGO, which increased electron donating ability of graphene materials, improved the adsorption performance of the NACs. In addition to hydrophobic effects and π−π EDA interactions, these results clearly indicate that other mechanisms significantly contribute to the superior adsorption of F

DOI: 10.1021/es5054946 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

GO gradually dissociated, and the Iasym/Isym ratio of the NO2 stretching bands gradually increased from 0.37 at pH = 2.65 to 0.88 at pH = 11.92, approaching the value of pure mdinitrobenzene (0.82) (SI Figure S-7b). This result indicated that the compression of the O−N−O angle on the surface of GO became weak and finally disappeared as the surface functional groups of GO dissociated. At low pH values, the −COOH and −OH or the adsorbed water molecules onto GO interacted with the two oxygens of the −NO2 groups of mDNB through various interaction modes including charge electrostatic interactions, polar interactions to reduce the O− N−O angle. When the −COOH and −OH began to dissociate at relatively high pH values, the reduction of the O−N−O angle became smaller. Next, the −COO− or −O− (electron donor) functioned as adsorptive sites for m-dinitrobenzene (electron acceptor) through various interaction modes including polar interaction and electron donor−acceptor interactions. Thus, the amount of m-DNB adsorbed onto GO remained constant under different solution pH values may be due to the presence of the various compensation mechanisms. Environmental Implications. The high adsorption of NACs and the multiple interaction mechanisms of graphene nanosheets enable many environmental applications, including pollutant removal, detection, and transformation. Graphene has been widely used as a sensing substrate for detecting various molecules/biomolecules by taking advantage of its adsorption properties toward analytes. The adsorbate molecules either release electrons to graphene (n-type doping) or remove them from graphene (p-type doping),17 which induces a large charge transfer between graphene and the molecular dopant.47 The increase in the graphene charge carrier concentration can be used to detect environmental pollutants and individual molecules.47,48 For example, a graphene-like film with a thickness of 1.5 nm displays high chemical sensitivity toward NO2 gas in the range of tens of ppb due to the enriched holelike carrier concentration through the electron transfer from graphene to a NO2 acceptor adsorbed onto graphene (p-type doping).47,49,50 The strong accumulation of nitroaromatic explosives on the surface of RGO through π−π stacking interactions improves the detection sensitivity for explosives.51 In addition, trinitrotoluene can be detected with high sensitivity and at a low detection limit by using an electrode modified with graphene film, and a uniform and richly wrinkled graphene film can be used as the prominent analytical platform for sensitive electrochemical determination of nitroaromatic explosive compounds.39 Gao et al.16 first reported that RGOs exhibit high catalytic activity and stability for the hydrogenation of nitrobenzene at room temperature and that the unsaturated carbon atoms at the edges of RGOs and the defects on RGOs are catalytically active centers for nitrobenzene. In summary, RGO is a powerful adsorbent for explosive pollutants that is dominated by multiple molecular mechanisms. The nitro-functional groups of NACs not only enhance the π−π EDA between the π-electron-deficient phenyls of NACs and the π-electron-rich matrix of graphene nanosheets, but they also promote the charge electrostatic interaction with the O-containing groups/edges/defects of graphene materials. We speculatively separate adsorption through π-systems from that through NO2 of NACs onto graphene materials, but the proposed mechanism warrants further study.

6c). Regarding RGO, the value of Iasym/Isym reached 1.52, which suggested that the EDA interactions with the −OH/edges/ defects on RGO were stronger (Figure 6b). Vacancies or small holes were ubiquitous in RGO, and these defects (electron-rich regions) could serve as additional adsorption sites for NAC molecules.42 In addition to the π−π EDA interactions, the greater number of −OH/−COOH/defects/edges on RGO relative to G could act as extra adsorption sites for charge electrostatic interaction and polar interaction. In addition, the −OH/−COOH groups could further improve the dispersion ability of RGO, and then favor a better availability of surfaces in hydrated RGO as compared with hydrophobic surfaces of less polar graphene undergoing a strong aggregation. Finally, the band at 711 cm−1 (N−O) increased significantly, which means enhanced out-of-plane wagging deformations, especially for RGO (Figure 5). Therefore, RGO could be the best adsorbent for removing NACs in water. Site-specific interactions between the nitro- or phenylgroups of NACs and the surface of graphene materials were shown using macroscopic adsorption and spectroscopic data. Therefore, a multiple molecular mechanism was initially proposed in this study to explain the distinct NAC adsorption, including the π−π EDA interactions between the electrondeficient phenyl groups of the NACs and the electron-rich matrix of RGO or G and the charge electrostatic interactions between the electron-withdrawing nitro groups of the NACs and defect sites of graphene nanosheets. Furthermore, the adsorption of NACs onto graphene nanosheets in the current study was much higher (10−50-fold) than the adsorption onto CNTs in a previous study.12 This difference could be attributed to the flat surface of graphene nanosheets, which favors multiple molecular interactions in comparison with the cylindrical nanostructures of the CNTs. In comparison with graphene, RGO had more defects that were favorable for the multiple interaction mechanism. The NAC adsorption and spectroscopic data in the current study are consistent with the theoretical calculations in a previous report.17 On the basis of first principle density functional theory, the tilting of the nitro group of nitrobenzene toward the graphene surface indicates that the nitro group−surface interaction is relatively strong.17 Therefore, electron transfer from the π-electron-rich RGO or G nanosheets to the adsorbed NACs results in enriched hole concentration and enhances the interactions of the NAC molecules with G or RGO nanosheets. To assess the contributions of hydrogen bonding to NAC adsorption, the effects of solution pH on adsorption by GO or RGO were investigated along with the FTIR spectra. The adsorption amount (Qe) remained nearly unchanged as the solution pH was altered (SI Figure S-6); thus, the H-bonding mechanisms can be ruled out. Similarly, Chen et al. ruled out H-bonding mechanisms in the adsorption of 2,4-dinitrotoluene onto CNTs because the value of logKd remained constant when the pH changed.12 To further examine the mechanism of mdinitrobenzene’s interactions with the surface of GO, the influences of solution pH on the intensity of the NO2 stretching bands of m-dinitrobenzene adsorbed onto GO are shown in SI Figure S-7. In addition, the calculated Iasym/Isym ratios are listed in Table 1. Under different pH conditions, the vasym (NO) of mDNB adsorbed onto GO were all blue-shifted, which indicated that an electron was removed from the surface of GO and transferred to the m-DNB via various interaction modes, which further excludes the role of H-bonding. When the pH values changed, the surface functional groups (−COOH and −OH) of G

DOI: 10.1021/es5054946 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology



(13) Yang, K.; Xing, B. Adsorption of organic compounds by carbon nanomaterials in aqueous phase: Polanyi theory and its application. Chem. Rev. 2010, 110, 5989−6008. (14) Fang, Q.; Chen, B. Adsorption of perchlorate onto raw and oxidized carbon nanotubes in aqueous solution. Carbon 2012, 50, 2209−2219. (15) Fang, Q.; Chen, B. Self-assembly of graphene oxide aerogels by layered double hydroxides cross-linking and their application in water purification. J. Mater. Chem. A 2014, 2 (23), 8941−8951. (16) Gao, Y.; Ma, D.; Wang, C.; Guan, J.; Bao, X. Reduced graphene oxide as a catalyst for hydrogenation of nitrobenzene at room temperature. Chem. Commun. 2011, 47, 2432−2434. (17) Saha, S.; Chandrakanth, R.; Krishnamurthy, H.; Waghmare, U. Mechanisms of molecular doping of graphene: A first-principles study. Phys. Rev. B 2009, 80, 155414. (18) Dai, Z.; Zhao, Y. First-principles study of nitrobenzene adsorption on graphene. Appl. Surf. Sci. 2014, 305, 382−385. (19) Fu, H.; Zhu, D. Graphene oxide-facilitated reduction of nitrobenzene in sulfide-containing aqueous solutions. Environ. Sci. Technol. 2013, 47, 4204−4210. (20) Kovtyukhova, N. I.; Ollivier, P. J.; Martin, B. R.; Mallouk, T. E.; Chizhik, S. A.; Buzaneva, E. V.; Gorchinskiy, A. D. Layer-by-layer assembly of ultrathin composite films from micron-sized graphite oxide sheets and polycation. Chem. Mater. 1999, 11, 771−778. (21) McDermott, M. T.; McCreery, R. L. Scanning tunneling microscopy of ordered graphite and glassy carbon surfaces: Electronic control of quinone adsorption. Langmuir 1994, 10, 4307−4314. (22) Wang, F.; Haftka, J. J. H.; Sinnige, T. L.; Hermens, J. L. M.; Chen, W. Adsorption of polar, nonpolar, and substituted aromatics to colloidal graphene oxide nanoparticles. Environ. Pollut. 2014, 186, 223−233. (23) Zhang, B.; Li, F.; Wu, T.; Sun, D.; Li, Y. Adsorption of pnitrophenol from aqueous solutions using nanographite oxide. Colloids Surf., A: Physicochem. Eng. Aspects 2015, 464, 78−88. (24) Weissmahr, K. W.; Haderlein, S. B.; Schwarzenbach, R. P. In situ spectroscopic investigations of adsorption mechanisms of nitroaromatic compounds at clay minerals. Environ. Sci. Technol. 1997, 31, 240−247. (25) Johnston, C. T.; De Oliveira, M. F.; Teppen, B. J.; Sheng, G. Y.; Boyd, S. A. Spectroscopic study of nitroaromatic-smectite sorption mechanisms. Environ. Sci. Technol. 2001, 35, 4767−4772. (26) Hummer, W. S.; Offeman, R. E. Preparation of graphite oxide. J. Am. Chem. Soc. 1958, 80, 1339. (27) Sheng, Z.; Shao, L.; Chen, J.; Bao, W.; Wang, F.; Xia, X. Catalyst-free synthesis of nitrogen-doped graphene via thermal annealing graphite oxides with melamine and its excellent electrocatalysis. ACS Nano 2011, 5, 4350−4358. (28) Shin, H. J.; Kim, K. K.; Benayad, A.; Yoon, S. M.; Park, H. K.; Jung, I. S.; Jin, M. H.; Jeong, H. K.; Kim, J. M.; Choi, J. Y.; Lee, Y. H. Efficient reduction of graphite oxide by sodium borohydride and its effect on electrical conductance. Adv. Funct. Mater. 2009, 19, 1987− 1992. (29) Chen, Z.; Chen, B.; Zhou, D.; Chen, W. Bisolute sorption and thermodynamic behavior of organic pollutants to biomass-derived biochars at two pyrolytic temperature. Environ. Sci. Technol. 2012, 46, 12476−12483. (30) Chandra, V.; Park, J.; Chun, Y.; Lee, J. W.; Hwang, I.; Kim, K. S. Water-dispersible magnetite-reduced graphene oxide composites for arsenic removal. ACS Nano 2010, 4, 3979−3986. (31) Stankovich, S.; Dikin, D.; Piner, R.; Kohlhaas, K.; Kleinhammes, A.; Jia, Y.; Wu, Y.; Nguyen, S.; Ruoff, R. Synthesis of graphene-based nanosheets via chemical reduction of exfoliated graphite oxide. Carbon 2007, 45, 1558−1565. (32) Chen, C.; Zhang, Q.; Yang, M.; Huang, C.; Yang, Y.; Wang, M. Structural evolution during annealing of thermally reduced graphene nanosheets for application in supercapacitors. Carbon 2013, 50, 3572− 3584. (33) Yan, J.; Wei, T.; Shao, B.; Ma, F.; Fan, Z.; Zhang, M.; Zheng, C.; Shang, Y.; Qian, W.; Wei, F. Electrochemical properties of graphene

ASSOCIATED CONTENT

S Supporting Information *

Selected properties of NACs (Table S-1), structural characteristics (Table S-2), regression parameters of isotherms (Table S3), adsorption kinetics (Figure S-1), O 1s XPS spectra and deconvolution (Figure S-2), nitrogen adsorption/desorption isotherms (Figure S-3), relationship between Kd/KHW and Qe (Figure S-4), FTIR spectra of the solid m-dinitrobenzene (Figure S-5), effects of solution pH on the adsorption of GO and RGO (Figure S-6), and FTIR spectra of m-dinitrobenzeneGO under different pH (Figure S-7). This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

* Phone: 0086-571-88982587; fax: 0086-571-88982587; e-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This project was supported by the National Science Foundation for Distinguished Young Scholars of China (Grant 21425730), the National Basic Research Program of China (Grant 2014CB441106), and the National Natural Science Foundation of China (Grants 21277120 and 41071210).



REFERENCES

(1) Geim, A. K. Graphene: Status and prospects. Science 2009, 324, 1530−1534. (2) Geim, A. K.; Novoselov, K. S. The rise of graphene. Nat. Mater. 2007, 6, 183−191. (3) Stoller, M. D.; Park, S.; Zhu, Y.; An, J.; Ruoff, R. S. Graphenebased ultracapacitors. Nano Lett. 2009, 8, 3498−3502. (4) Shen, Y.; Fang, Q.; Chen, B. Environmental applications of threedimensional graphene-based macrostructures: Adsorption, transformation and detection. Environ. Sci. Technol. 2015, 49, 67−84. (5) Wang, J.; Chen, Z.; Chen, B. Adsorption of polycyclic aromatic hydrocarbons by graphene and graphene oxide nanosheets. Environ. Sci. Technol. 2014, 48, 4817−4825. (6) Zhao, J.; Wang, Z.; Zhao, Q.; Xing, B. Adsorption of phenanthrene on multilayer graphene as affected by surfactant and exfoliation. Environ. Sci. Technol. 2014, 48, 331−339. (7) Sun, Y.; Yang, S.; Zhao, G.; Wang, Q.; Wang, X. Adsorption of polycyclic aromatic hydrocarbons on graphene oxides and reduced graphene oxides. Chem.Asian J. 2013, 8, 2755−2761. (8) Apul, O. G.; Wang, Q.; Zhou, Y.; Karanfil, T. Adsorption of aromatic organic contaminants by graphene nanosheets: Comparison with carbon naontubes and activated carbon. Water Res. 2013, 47, 1648−1654. (9) 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, 191− 198. (10) Pei, Z.; Li, L.; Sun, L.; Zhang, S.; Shan, X.; Yang, S.; Wen, B. Adsorption characteristics of 1,2,4-trichloobenzene, 2,4,6-trichlorophenol, 2-naphthol and naphthalene on graphene and graphene oxide. Carbon 2013, 51, 156−163. (11) Yang, K.; Chen, B. Facile fabrication of stable monolayer and few-layer graphene nanosheets as superior sorbents for persistent aromatic pollutant management in water. J. Mater. Chem. A 2014, 2 (43), 18219−18224. (12) Chen, W.; Duan, L.; Zhu, D. Adsorption of polar and nonpolar organic chemicals to carbon nanotubes. Environ. Sci. Technol. 2007, 41, 8295−8300. H

DOI: 10.1021/es5054946 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology nanosheet/carbon black composites as electrodes for supercapacitors. Carbon 2010, 48, 1731−1737. (34) Arrigo, R.; Havecker, M.; Wrabetz, S.; Blume, R.; Lerch, M.; McGregor, J.; Parrott, E. P. J.; Zeitler, J. A.; Gladden, L. F.; Gericke, A. K.; Schlog, R.; Su, D. Tuning the acid/base properties of nanocarbons by functionalization via amination. J. Am. Chem. Soc. 2010, 132, 9616− 9630. (35) Chen, C.; Huang, J.; Zhang, Q.; Gong, W.; Yang, Q.; Wang, M.; Yang, Y. Annealing graphene oxide film to produce a free standing high conductive graphene film. Carbon 2012, 48, 4466−4474. (36) Knipe, S. W.; Mycroft, J. R.; Pratt, A. R.; Nesbitt, H. W.; Bancroft, G. M. X-ray photoelectron spectroscopic study of water adsorption on iron sulphide minerals. Geochim. Cosmochim. Acta 1995, 59, 1079−1090. (37) Fan, L.; Liu, J.; Din, R. U.; Yan, X.; Qu, X. The effect of reduction time on the surface functional groups and supercapacitive performance of graphene nanosheets. Carbon 2012, 50, 3724−3730. (38) Sasha, S.; Dmitriy, A.; Richard, D.; Kevin, A.; Alfred, K.; Jia, Y.; Wu, Y. Synthesis of graphene-based nanosheets via chemical reduction of exfoliated graphite oxide. Carbon 2007, 45, 1558−1565. (39) Tang, L. H.; Feng, H. B.; Cheng, J. S.; Li, J. H. Uniform and rich-wrinkled electrophoretic deposited graphene film: A robust electrochemical platform for TNT sensing. Chem. Commun. 2010, 46, 5882−5884. (40) Kim, J.; Cote, L. J.; Kim, F.; Yuan, W.; Shull, K. R.; Huang, J. Graphene oxide sheets at interfaces. J. Am. Chem. Soc. 2010, 132, 8180−8186. (41) Carter, M. C.; Kilduff, J. E.; Weber, W. J., Jr. Site energy distribution analysis of preloaded adsorbents. Environ. Sci. Technol. 1995, 29, 1773−1780. (42) Zhu, D.; Pignatello, J. Characterization of aromatic compound sorption interactions with black carbon (charcoal) by graphite as a model. Environ. Sci. Technol. 2005, 39, 2033−2041. (43) Riskin, M.; Tel-Vered, R.; Bourenko, T.; Granot, E.; Willner, I. Imprinting of molecular recognition sites through electropolymerization of functionalized nanoparticles: Development of an electrochemical TNT sensor based on π-donor-acceptor interactions. J. Am. Chem. Soc. 2008, 130, 9726−9733. (44) Das, B.; Voggu, R.; Rout, C. S.; Rao, C. N. R. Changes in the electronic structure and properties of graphene induced by molecular charge-transfer. Chem. Commun. 2008, 44, 5155−5157. (45) Voggu, R.; Rout, C. S.; Franklin, A. D.; Fisher, T. S.; Rao, C. N. R. Extraordinary sensitivity of the electronic structure and properties of single-walled carbon nanotubes to molecular charge-transfer. J. Phys. Chem. C 2008, 112, 13053−13056. (46) Body, S. A.; Sheng, G. Y.; Teppen, B. J.; Johnston, C. T. Mechanisms for the adsorption of substituted nitrobenzenes by smectite clays. Environ. Sci. Technol. 2001, 35, 4227−4234. (47) Leenaerts, O.; Partoens, B.; Peeters, F. M. Adsorption of H2O, NH3, CO, NO2, and NO on graphene: A first-principles study. Phys. Rev. B 2008, 77, 125416. (48) Schedin, F.; Geim, A. K.; Morozov, S. V.; Hill, E. W.; Blake, P.; Katsnelson, M. I.; Novoselov, K. S. Detection of individual gas molecules adsorbed on graphene. Nat. Mater. 2007, 6, 652−655. (49) Samsonau, S. V.; Shvarkov, S. D.; Meinerzhagen, F.; Wieck, A. D.; Zaitsev, A. M. Growth of graphene-like films for NO2 detection. Sens. Actuators B 2013, 182, 66−70. (50) Lu, G.; Ocola, L. E.; Chen, J. Gas detection using lowtemperature reduced graphene oxide sheets. Appl. Phys. Lett. 2009, 94, 083111. (51) Chen, T.-W.; Sheng, Z.-H.; Wang, K.; Wang, F.-B.; Xia, X.-H. Determination of explosives using electrochemically reduced graphene. Chem.Asian J. 2011, 6, 1210−1216.

I

DOI: 10.1021/es5054946 Environ. Sci. Technol. XXXX, XXX, XXX−XXX