- Email: [email protected]
Optimized graphene oxide foam with enhanced performance and high selectivity for mercury removal from water
Accepted Manuscript Title: Optimized graphene oxide foam with enhanced performance and high selectivity for mercury removal from water Author: Bruno Henriques Gil Gonc¸alves Nazanin Emami Eduarda Pereira Mercedes Vila Paula A.A.P. Marques PII: DOI: Reference:
S0304-3894(15)30083-2 http://dx.doi.org/doi:10.1016/j.jhazmat.2015.09.028 HAZMAT 17101
To appear in:
Journal of Hazardous Materials
Received date: Revised date: Accepted date:
1-6-2015 9-9-2015 13-9-2015
Please cite this article as: Bruno Henriques, Gil Gonc¸alves, Nazanin Emami, Eduarda Pereira, Mercedes Vila, Paula A.A.P.Marques, Optimized graphene oxide foam with enhanced performance and high selectivity for mercury removal from water, Journal of Hazardous Materials http://dx.doi.org/10.1016/j.jhazmat.2015.09.028 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Optimized graphene oxide foam with enhanced performance and high selectivity for mercury removal from water
Bruno Henriquesa, Gil Gonçalvesb,c*, Nazanin Emamie, Eduarda Pereiraa, Mercedes Vilab,c,d*, Paula A.A.P. Marquesb,c,
a
b
CESAM & Department of Chemistry, University of Aveiro, 3810-193 Aveiro, Portugal
TEMA-NRD, Mechanical Engineering Department, University of Aveiro, 3810-193 Aveiro, Portugal c
Aveiro Institute of Nanotechnology, AIN, University of Aveiro, 3810-193 Aveiro, Portugal d
e
Centro de Investigación Biomédica en Red. Bioingeniería, Biomateriales y Nanomedicina, CIBER-BBN, Spain
Department of Applied Physics and Mechanical Eng, Lulea University of Technology, Sweden.
*Corresponding authors. E-mail: [email protected] (Gil Gonçalves); [email protected] (Mercedes Vila)
1
Graphical Abstract
2
Highlights
There is a lack of studies on Hg removal from natural waters at realistic levels Three-dimensional graphene oxide were modified to enhance the affinity toward Hg The removal process was optimized in terms of pH, mass and contact time. The process efficiency was assessed in multimetallic solutions and natural matrices Small dose of GO foam (10 mg L-1) allows to remove up to 96% of Hg(II) after 24h
3
Abstract This work explores the preparation of three-dimensional graphene oxide macroscopic structures, shaped by self-assembling single graphene oxide (3DGO) sheets with control of its surface chemistry by combining with nitrogen functional groups (3DGON), or with nitrogen and sulphur functional groups (3DGOSN), and their application in the removal of mercury (Hg(II)) from aqueous solutions. The chemical structure of the materials was assessed by using different characterization techniques: SEM, XPS and BET. Adsorption studies conducted in Hg(II) contaminated ultra-pure water reveal the enhanced ability of 3DGON for the adsorption of this metal, when compared to the other GO foams. A small dose of 3DGON (10 mg L-1) allows to remove up to 96% of Hg(II) after 24h of contact time, leading to a residual concentration in solution close to the guideline value for drinking water (1 µg L-1). The ability of this material to adsorb Hg (II) was evaluated relatively to different experimental parameters such as pH, sorbent dose, time and effect on different competing metal ions. Real application was also evaluated by testing its performance in two different natural matrices, river and sea water, with very promising results.
Keywords: mercury, graphene oxide, sorption, competitive ions, remediation, natural waters
4
1. Introduction Carbon based materials have been, since long time, proved to be one of the best options for water purification [1]. Industrially, the most relevant in our days is activated carbon [2] due to its versatility on removing various organic compounds and also metallic ions, such as Cr3+, Hg2+, Pb2+, Cr6+, Cd2+, Ni2+ and Zn2+ [3]. The removal process by activated carbon includes adsorption, absorption and ion exchange, being the overall process designated as sorption [4]. Recently, other carbon materials with higher specific surface areas have been proposed for purifying water, such as carbon nanotubes [5] and graphene based materials [6]. Among them, graphene oxide (GO) stands out from others because of its high oxidation degree and the ability of self-assembly in 3D macrostructures. GO is usually synthetized by the chemical exfoliation of graphite [8]. This approach has the advantage of large scale and low cost production and high efficiency [7]. GO presents a high degree of oxidation ratio (1.8< C/O <2.5) [9] which grants this material with a high hydrophilic character, allowing highly stable aqueous colloidal suspensions [10]. Furthermore, the presence of oxygen functional groups, which acts as electron donors, allows the easy coordination with metallic ions [11]. The ability of graphene-based materials to adsorb metallic ions from aqueous solutions has been already reported for: Pb2+ [12, 13], Sb3+ [14] Cd2+ [15], Co2+ [15] and U5+ [12]. The capacity of GO for the adsorption of divalent metal ions (Cu2+, Cd2+, Pb2+ and Zn2+) in aqueous solution has been discussed by Stiko et al, as the highest than any reported sorbent [13]. Recently, graphene based three-dimensional structures were synthetized by the selfassembling GO sheets [14] with or without the addition of additives [15-17]. The resulting architectures have been already explored for water purification, typically for the adsorption and degradation of organic contaminants [6]. However, the elimination of metallic ions from polluted waters using these 3D structures is of recent interest. Mi et.
5
al. reported the preparation of GO aerogels by a unidirectional freeze-drying method for adsorption of Cu2+ [18] showing that the GO aerogels reached the equilibrium in a shorter period than oxidized carbon nanotubes (CNT) sheets or activated carbon. Li et. al. showed that 3D graphene macrostructures prepared by Chemical Vapour Deposition growth in a nickel substrate present ultra-high electrical adsorption ability for heavy metals ions (Cd2+, Pb2+, Cu2+, Ni2+) removal from water solutions by using this material as an electrode [17]. Herein, we report for the first time the development of GO macrostructures with a controlled surface chemistry and with high affinity for the adsorption of mercury from aqueous solutions. GO macrostructures were prepared at different experimental conditions allowing the presence of surface modifications with oxygen, nitrogen or sulphur functional groups. The adsorption studies in mercury contaminated waters reveal the superior capacity of GO macrostructures modified with combination of oxygen and nitrogen functional groups (3DGON). Different experimental parameters such as pH, mass, time and effect of different competing metal ions were explored. The potential applicability of this material was also evaluated by testing its performance in river and sea waters. 2. Experimental 2.1. Synthesis of three dimensional GO structures GO was prepared by the chemical exfoliation of graphite (Graphite powder, <45 μm, ≥99.99%, Sigma-Aldrich) following a method described in our previous work [19]. The 3D structures were build up by the simple self-assembly of GO sheets in water (4 mg/mL) inside an autoclave (180ºC; 12h) (3DGO) [18]. The preparation of 3DGON and 3DGOSN structures was performed under the same experimental conditions described before by 6
adding respectively, 1mL of NH4OH (20% v/v) or 10 mg of NH4SCN to the initial GO solution.[20] 2.1.1 Structural characterization The surface chemistry of the 3D macroscopic materials was analysed by XPS (VG Scientific ESCALAB 200 A; UK). The morphological analysis was made using a field emission scanning electron microscope (Model SU70; Hitachi, Tokyo, Japan), operated at 15 kV.
2.2. Hg sorption studies 2.2.1 Sorption kinetics Batch sorption experiments were conducted using a Schott Duran® glass bottle (1 L) as reaction vessel, where sorbent and solution were maintained under constant magnetic stirring (700 rpm), at 21±1ºC. The Hg(II) solution was prepared by adding the desired volume of a certified Hg standard solution (1001±2 mg L−1 of Hg(II) in HNO3 0.5 mol L−1, from Merck) to ultra-pure water. The selected concentration of Hg(II) aimed to simulate a real scenario, since 50 µg L-1 is the old maximum permissible value for Hg discharges from industrial sectors (Directive 84/156/EEC). The rate of Hg(II) uptake by GO foams was followed by determining the concentration of this metal in solution samples (5-10 mL) collected at defined time intervals. All the samples were immediately centrifuged at 5000 rpm for 3 min, the supernatant liquid was collected, acidified to pH ≤ 2 using Suprapur© HNO3 (65% v/v) and stored at 4ºC until analysis. The removal trial was conducted in duplicate with a control (solution of Hg(II) in the absence of sorbent) running in parallel. The initial pH of the solutions was ca. 4.5 and no significant variations were observed during the removal process. The quantification of Hg in all water samples was performed by cold vapour atomic fluorescence spectroscopy (CV-AFS), using a PSA 7
10.025 Millennium Merlin Hg analyser and SnCl2 (2% m/v in HCl 10% v/v) as a reducing agent. 2.2.2 Effect of dosage The effect of the sorbent dosage on Hg(II) sorption was evaluated by placing different amounts of 3DGON (1, 2.5, 5, 10 and 20 mg L−1) in contact with a 50 µg L−1 Hg(II) solution. The experiments and the Hg(II) quantification were carried out as mentioned above. 2.2.3 Effect of pH The effect of pH on the sorption process was investigated for an initial Hg(II) concentration of 50 µg L−1 in ultra-pure water, with different initial pH values: 3, 5, 7 and 9. The pH adjustments were performed using HNO3 (31.5 g L−1) or NaOH (4.0 g L−1) solutions, before the addition of the pre-weighed sorbent (10 mg L-1). The experiments were carried out as described before. 2.2.4 Effect of the presence of co-ions The sorption of Hg(II) by 3DGON under metal competitive conditions was studied for ternary mixtures of Hg(II), Pb(II) and Cd(II) in equal concentration (50 µg L−1). Experiments were performed for an initial pH of ca. 4.5 following the same design described on 2.2.1. The spiked solutions were prepared by adding the desired volume of the each metal standard solution (1000±2 g L-1 of Pb(II) in HNO3 0.5 mol L−1, Panreac; 1000±2 mg L-1 of Cd(II) in HNO3 0.5 mol L−1, Merck) to ultra-pure water. The concentration of Hg(II) in solution was determined by CV-AFS, whereas Pb(II) and Cd(II) quantification was performed by inductively coupled plasma spectroscopy, using a Jobin–Yvon JY70 Plus Spectrometer. The limits of detection for Pb(II) and Cd(II) where found to be 6.7 µg L−1 and 1.3 µg L−1, respectively, with an acceptable relative standard deviation among replicates: <10 %.
8
2.2.4 Assessment of Hg(II) removal from natural waters The efficiency of 3DGON in natural waters was evaluated by conducting sorption tests in river and sea waters spiked with Hg(II), following the experimental design described in section 2.2.1. The river and sea waters were collected in the Vouga river (40°40′42′′ N, 08°22′18′′ W) and on the coast of Aveiro (40°32'58"N 8°46'31"W), respectively, and physicochemical characterization was previously reported [21, 22]. All water was filtered through 0.45 μm pore size filters and the solutions were left to pre-equilibrate during 24h before adding the sorbent. Initial pH of the spiked solutions were, respectively, ca. 6.5 and 7.8 for river water and seawater. 2.3 Analysis of sorption data The amount of Hg(II) bound per unit of sorbent mass, at a given time t (qt, µg g−1), was deduced from the mass balance between the initial Hg concentration in the solution (C0, µg L−1) and the concentration after a particular period of contact time t (Ct, µg L−1): qt
(C 0 C t ) V
(1)
m
where V (L) is the volume of the solution and m (g) is the sorbent mass. When equilibrium was attained: t=te, qt=qe and Ct=Ce (residual Hg(II) concentration in solution). The performance of the removal process was evaluated and compared using the Hg(II) removal percentage (R, %), which at time t is defined by: R t (%)
(C
0
Ct) C
100
(2)
0
2.3.1 Kinetics and Equilibrium models In this study, the kinetics of Hg(II) uptake by 3DGON was studied by applying three widely used reaction models to the experimental data, namely, the Lagergren pseudofirst-order model [23], the Ho’s pseudo-second-order model [24] and the Elovich model
9
[25]. In addition to the reaction models, two widely known diffusion-based models, Boyd’s film-diffusion [26] and Webber’s pore-diffusion [27], were used. (Details at SI) The experimental equilibrium data was fitted using 3 different models in their non-linear forms: the two-parameters Langmuir and Freundlich isotherm models and the threeparameters Sips or Langmuir-Freundlich isotherm model. (Details at SI)
3. Results and discussion 3.1. Characterization of the 3D GO structures The hydrothermal treatment of GO aqueous solutions results in the assembly of 3D macrostructures (Figure 1a). The SEM analyses show an interconnected porous network of partial overlapping flexible GO nanosheets with a foam-like structure, Figure 1b and c, for 3DGON. The chemical mechanism behind the assembly of GO nanosheets into 3D architectures with interconnected porous network is not yet completely understood. Wosley et al. proposed that the GO functional groups are the main responsible for the crosslinking and respective assembly onto 3D macrostructures [28]. Xu et al claim that non-covalent bonds like π-π stacking interactions between GO nanosheets are the main responsible for this assembly. HR-XPS spectra with curve fittings for C1s (Figure 1 d) and e)) shows high restructuration of sp2 carbon, presence of C-O functional groups and an increase of π-π stacking interactions. The curve fittings for N1s (Figure 1f) confirm the doping with different N-type groups of GO: pyridinic (31.9 at.%), amino (42.1 at.%), pyrrolitic (16.2 at.%), graphitic (4.1 at.%) and oxidized N (5.7 at.%) [29]. These results suggest that 3DGON are hybrid systems resulting from the assembly of GO nanosheets with the establishment of covalent bonds between carbon and nitrogen functional groups and π-π noncovalent interactions. 10
Analysis of the specific surface area of 3DGON macrostructure using nitrogen adsorption/desorption isotherms, show that they exhibit type IV isotherm patterns with H3-type hysteresis loop following the BDDT classification, which are characteristics of the mesoporous materials [30] (Figure S1), with a BET surface area of ~400 m2g-1 and a pore volume of 0.27 cm3g-1 (1.6
11
The optimization of the 3DGON sorbent dosage (mass of sorbent per volume of solution) is essential to maximize the efficiency of the sorption system. Figure 3 presents the removal of Hg(II) for different sorbent dosage (1.0, 2.5, 5.0, 10.0, and 20 mg L-1), after 24 and 48 hours of contact time. The results show that higher percentages of Hg(II) removal are obtained with the increase of the sorbent concentration, from 33% to 96% for doses 1 to 20 mg L-1 respectively, and the reaction equilibrium was reached after 24h contact time. It is evident that the increase of sorbent mass leads to a larger number of available coordination sites improving the removal efficiency of sorbent. However it was observed that using 10 mg L-1 of 3DGON allows to obtain removal values (95%) very similar to the maximum reached with 20 mg L-1. The significant increase of sorbent concentration did not lead to an enhancement in the sorption extension, revealing that the process is not only dependent on the number of free sites but also on the equilibrium of charges between sorbent and solution. 3.2.3 Effect solution pH The effect of the solution pH on the sorption of Hg(II) by GO3DN foams, at different periods of contact time (3, 12 and 24h) is presented in Figure 4. The results suggest that the sorption efficiency is almost independent of the solution pH on the interval between 3 and 9. After 3, 12 and 24 hours of contact time the amplitude on the removal percentage, for the different pH values studied, was 3.8 and 8.0% respectively. Actually other sorption studies of metals by GO based materials showed different sorption stages, lower efficiency for pH values lower than 3.9 and higher sorption rates for higher pH values [13, 18]. The decrease of sorption at acidic pH is usually attributed to the preferential protonation of the surface functional groups of the material over the biding to the metal ions. In the case of 3DGON foams the presence of nitrogen simultaneously with oxygen functional groups allows to maintain the same Hg(II) sorption efficiency at lower pH 12
values. From our best knowledge, 3DGON shows an acid basic behaviour similar to that of amino acids, which have the capacity to form zwitterionic species by establishing ionic dipole (-COO- and the adjacent -NH3+) on the pH range of 2 to 9 [33]. 3.3. Sorption kinetics Three kinetic reaction models in their non-linear form were used to examine the sorption of Hg(II) by 3DGON foam, namely pseudo-first-order (PFO), pseudo-second order (PSO) and Elovich model. The fittings of the experimental data and the parameters values obtained from the kinetic modelling are presented in Figure 5a) and Table 1, respectively. Clearly among the models tested and the under the experimental conditions used the PFO model had the poorest performance on describing the experimental data, which is corroborated by the lowest R2 (0.917) and the highest Sy.x (464). Moreover this model was not able to correctly predict the experimental qe, underestimating its value (qe1= 4173 vs qe exp= 4515), contrary to the PSO model which precisely estimated the qe value (relative error lower than 0.3%). Despite the good agreement between the experimental data and the fittings accomplished by both PSO (R2=0.971) and the Elovich (R2=0.981) models, comparison using Akaike's information criteria (AIC) obtained for each model indicates that Elovich model is more likely to be correct (probability of 89.7%). This gives indication that the Hg(II) sorption process has a chemisorption nature as well as the surface of the sorbent is heterogeneous [34].
To obtain more insight into the mechanism of Hg(II) sorption by 3DGON and potential rate-controlling steps two diffusion-based models were explored: Boyd’s film-diffusion and Webber’s intraparticle-diffusion. The results of the piecewise linear regression (PLR) treatment [35] applied to the experimental data are presented in Table 2. From Boyd filmdiffusion fitting it was obtained a first linear segment (R2=0.9941) and the confidence 13
interval of the intercept (95 % confidence level) includes the zero, which strongly suggests that film-diffusion is not the rate determining step in the sorption process [36]. The Webber’s intraparticle-diffusion´s plot is presented in Figure 5b. The existence of multiple linear stages is clear which shows that more than one diffusion step takes place. The Akaike’s information criteria were used to define the numbers of linear segments. Lower AIC value was obtained when considering only two linear segments and the calculated evidence ratio indicates that the two stages approach is 1.13×107 times more likely to be the correct model than the approach considering three stages. The first steeper section (until 7.89 h) does not pass through the origin which points that intra-particle diffusion is not only the rate-controlling step, but is also affected by the boundary layer diffusion process [34]. Another phenomenon could be a surface enhancement associated with highly energetic heterogeneous surface, leading to a high rapid uptake of Hg(II) on the sorbent before the system stabilizes [37]. The second less pronounced portion is attributed to the final equilibrium stage for which the intra-particle diffusion starts to slow down due to low Hg(II) concentration left in the solution.
3.4. Chemical mechanism of Hg(II) sorption XPS survey spectra for 3DGON surfaces before and after Hg(II) sorption were conducted to advance the perception into the sorption mechanism (Figure 6a). As expected, the structural analysis exhibits the peaks matching C1s, O1s and N1s and Hg4f after exposure to the Hg (II) spiked water. High-resolution spectra of Hg4f, in Figure 5b, shows two peaks at 100.9 eV and 104.8 eV, that can be assigned to the presence of Hg 4f7/2 and 4f5/2. Those results indicates that mercury is adsorbed in an oxidized state instead of its metallic form (Hg0, Hg 4f7/2 – 104 eV and 4f5/2 – 99.9 eV) [38] suggesting that the main sorption mechanism is oxidative chemisorption. 14
HRXPS of 3DGON at N1s region can be deconvoluted into five Gaussian peaks (Figure 6c), Pyridinic, Amino, Pyrrolic, Graphitic and Oxidized N. However after exposure to Hg(II), the N1s peak correspondent to graphitic structures disappears. Moreover the exposure of 3DGON to Hg(II) aqueous solution promotes significant changes in the N1s peaks areas, meaning a different relative distribution of nitrogen species or/and the formation of new covalent bonds with metal. The higher decrease on the peaks area corresponds to amine (9.1 at%) and pyridinic groups (4.8 at%) suggests that are the preferential groups for the coordination with the Hg(II). Li et al. described the formation of Hg(II)-NH2 complex on the sorption mechanism of Hg(II) by chitosan beads grafted with polyacrylamide [39]. The soft basic ligand character of the amino functional groups can favors an easy interaction with very soft acid such as Hg(II). HRXPS of C1s spectrum shows a decrease of 3.26 at% for C=O/COOH groups after Hg(II) sorption at equilibrium, which indicates that oxygen functional groups also play an active role on the interaction with Hg (II). The chemical interaction of carboxylic and phenolic groups by proton exchange with metal ions was already reported for oxidized CNTs,[40] as the main sorption mechanism for divalent metal ions in aqueous solution. 3.5. Sorption Isotherms Isotherm models allow the quantitative assessment of sorbent’s sorption capacity, giving indication about its sorption strength. In this study, Langmuir, Freundlich and SIPS isotherm models, in their nonlinear form, were applied to the experimental equilibrium results for the Hg(II)-3DGON system, at 21±1ºC ([Hg(II)] of 50 µg L-1 and 3DGON dosages of 1, 2.5, 5, 10 and 20 mg L-1). Figure 7 shows the fitted equilibrium data, while Table 3 presents the isotherms parameters and the good fit. Under the experimental conditions studied, both Langmuir and Freundlich isotherms models successfully described the equilibrium data, which implies that both monolayer adsorption and 15
heterogeneous surface conditions existed [18, 41]. However, the good fit (Table 3) indicates that Langmuir isotherm model performed slightly better than Freundlich isotherm model. The comparison between Langmuir and SIPS isotherms using AIC (since these two models have different degrees of freedom) suggests that the former is more likely to be adjusted to the experimental data than the latter (AIC Langmuir=85.10 and AICSIPS=112.0). In fact, the heterogeneity index from the SIPS isotherm, 1/n=0.88±0.10, points to a heterogeneous or multilayer system, yet this value is not far away from the unit, case where SIPS isotherm reduces to Langmuir isotherm. The maximum sorption capacity of Hg(II) by 3DGON predicted by Langmuir isotherm model is c.a. 35000 µg g1
, which highlights the potential of this material. Moreover the separation factor, RL,
characteristic of the Langmuir isotherm, indicates that the sorption is a favourable process (RL=0.33), almost irreversible at high initial Hg(II) concentrations (RL→0) [41, 42]. The performance evaluation of 3DGON on mercury removal regarding other sorbent materials is not a simple task, and not always feasible. Experimental conditions, such pH, temperature and sorbate/sorbent ratio are very different among studies, being the results not entirely comparable. It should be emphasized that the majority of these studies have focused on “unrealistic” concentrations of Hg(II), i.e. using initial concentrations extremely larger than those that may be found in effluents or aquatic systems affected to wastewater discharges. Furthermore, the dose of sorbent used in most of these works (up to 50,000 times higher than ours) did not reflect an environmental concern with the final residue generated (Table S1). Using an initial Hg(II) concentration of 0.1 mg L-1 and sorbent/solution ratios from 200 to 800 mg L-1 Tawabini et al. [43] found that the maximum sorption capacity of MWCNTs is 13.16 mg g-1 (less than half of the value found for 3DGON). Similar value (14 mg g-1) was reported by Dong et al. [44] for mesoporous sílica-coated magnetic particles, using 2500 mg L-1 and concentrations of
16
Hg(II) between 10 to 60 mg L-1. Two of the few studies with GO-based materials applied to Hg(II) removal were performed by Kyzas et al. [45] (Magnetic GO, qe=14.02 to 123.10 mg g-1) and by Bao et al. [46] (GO, qe=5 mg g-1; Fe3O4/GO, qe=20 mg g-1; Thiolfunctionalized-Fe3O4/GO, qe=30.94 mg g-1). However, in addition to the very high concentrations of Hg(II) used (200-10000 times higher than the limit for effluent discharges), there is a lack of information in both studies regarding the pH and the amount of sorbent used, which prevents a fair comparison with 3DGON. Overall, these results demonstrate the efficacy of 3DGON foam in the treatment of contaminated waters with Hg(II) levels representative of environmental real situations.
3.6. Effect of co-ions on Hg(II) sorption The selectivity of 3DGON to Hg(II) ions was evaluated by the competitive adsorption with Pb(II) and Cd(II) co-ions. Figure 8 showed the preferential affinity for this material to the Hg(II) > Pb(II) > Cd(II) ions with removal percentage of approximately 91%, 10% and 5% at equilibrium, respectively. That high selectivity for Hg(II) by GO nanocomposites in co-ions solutions was already observed by Chandra et al [32] however they didn’t show any plausible explanation for the selective adsorption of the metals. It is know that divalent metal ions can be present in deionized water in different ionized species M2+, M(OH)+1, M(OH)20, M(OH)3-1… depending the percentage of each species of the form and concentration of negative ions in solution, the pH and electronic structure of metal. The simulations performed in the MinteQ software at constant pH of 4.5 for mixture Hg(II), Pb(II) and Cd(II) showed that the predominant species of each element on aqueous solutions are Hg(OH)2 (98.0%), Pb2+(92.12%) and Cd2+(98.0%), respectively (Table S2).
17
The preferential adsorption of metallic species by 3DGON in aqueous solutions can be explained by Pearson’ theory that postulates hard acids coordinates preferentially to hard bases and soft acids to soft bases [47]. Based on the chemical hardness (7.70 eV (Hg2+), 8.46 eV (Pb2+), 10.29 eV (Cd2+)), the metallic ions involved on the adsorption process, should have very similar adsorptive behaviour[48]. However, on contrary to the other metallic elements, the predominant species for Hg in aqueous solution is a neutral form Hg(OH)2, which is considered a softer base than the metallic cations (Hg2+, Pb2+ and Cd2+). Taking into account that the predominant groups at 3DGON structure are amines, that can be considered as a soft bases, the affinity with Hg(OH)2 will be higher, which is in agreement with our experimental results. 3.7. Sorption of Hg(II) on natural waters To evaluate the real efficiency of 3DGON under real conditions we perform tests of sorption of Hg(II) on river and seawater (Figure 9). The results showed that Hg(II) sorption in ultra-pure water is higher (93%), which can be explained by the fact that natural waters are complex matrices due to the presence of various ions that can compete for the binding sites at sorbent surface (Figure S4). Furthermore the organic matter existent on natural systems can complex with the metallic ions under study, decreasing its mobility and affinity for sorption [49]. Relatively to the natural waters, the results achieved show a higher Hg(II) sorption capacity on river water (82%) then seawater (42%) ecosystems. The larger concentration of positive ions (Ca+, Na+, Mg+2) in seawater leads to an increased competition for the functional groups on the surface of sorbent material. On the other hand, the high levels of Cl- in seawater enables the formation of mercury chloro-complexes (HgCl42−, HgCl3−…), which have less affinity to the sorbent and thus remain in the solution [50].
18
4. Conclusions In summary we observed that GO macrostructures are a very effective material on the sorption of Hg(II) from water solution. The different chemical surface modifications performed showed that the combination of oxygen and nitrogen functional groups (3DGON) increases the removal efficiency of Hg (II). The different acidic character of the nitrogen and oxygen functional groups allows shifting the equilibrium of the reaction towards the increase o Hg(II) sorption. Small dose of 3DGON (10 mg L-1) allows to remove up to 95% of Hg(II) after 24h of contact time, leading to a residual concentration close to the guideline value for drinking water (1 µg L-1). The Hg(II) sorption follows the Elovich model kinetics and the equilibrium data was well fitted by the Langmuir isotherms, with a predicted maximum sorption capacity of 35 mg g-1. The efficiency of this new material was not significantly affected by the pH and presence of co-ions Pb(II) and Cd(II) in solution, which highlights its great selectivity and affinity to Hg(II). Tests conducted in natural waters showed higher potential of 3DGON for real applications.
Ackowledgements: B. Henriques and G. Gonçalves would like to acknowledge the Portuguese Foundation for Science and Technology (FCT) for the financial support of the PostDoc grants (SFRH/BPD/112576/2015 and SFRH/BDP/84419/2012, respectively). M Vila and PAAP Marques thank the FCT for the investigator grant. We would like to thank FCT, the European Union, QREN, FEDER, COMPETE, for funding the TEMA Research Unit (PEst-C/QUI/UI0062/2013)
and
the
CESAM
associated
laboratory
(UID/AMB/50017/2013).
19
References [1]
P. Hadi, M-H. To, C-W. Hui, CSK. Lin, G. McKay, Aqueous mercury
adsorption by activated carbons, Water Res. 73 (2015) 37-55. [2]
R.C Bansal, M. Goyal, Activated Carbon Adsorption, CRC Press 2005, USA,
pp. 297-436. [3]
A. Bhatnagar, A.K. Minocha, Conventional and non-conventional adsorbents for
removal of pollutants from water - A review, Indian J. Chem. Techn. 13 (2006) 203-17. [4]
F. Fu, Q. Wang, Removal of heavy metal ions from wastewaters: A review, J.
Environ. Manage. 92 (2011) 407-18. [5]
X. Ren, C. Chen, M. Nagatsu, X. Wang, Carbon nanotubes as adsorbents in
environmental pollution management: A review, Chem. Eng. J. 170 (2011) 395-410. [6]
K.C. Kemp, H. Seema, M. Saleh, N.H. Le, K. Mahesh, V. Chandra et al.,
Environmental applications using graphene composites: water remediation and gas adsorption, Nanoscale 5 (2013) 3149-71. [7]
D.C. Marcano, D.V. Kosynkin, J.M. Berlin, A. Sinitskii, Z. Sun, A. Slesarev, et
al. Improved Synthesis of Graphene Oxide, ACS Nano 4 (2010) 4806-14. [8]
S. Park S, R.S. Ruoff, Chemical methods for the production of graphenes. Nat.
Nanotechnol. 4 (2009) 217-24. [9]
D.R. Dreyer, S. Park, C.W. Bielawski, R.S. Ruoff, The chemistry of graphene
oxide. Chem. Soc. Rev. 39 (2010) 228-40. [10]
J.I. Paredes, S. Villar-Rodil, A. Martinez-Alonso, J.M.D. Tascon Graphene
oxide dispersions in organic solvents. Langmuir 24 (2008) 10560-4. [11]
I.L. Laure, S.V. Tkachev, E.Y. Buslaeva, E.V. Fatushina, S.P. Gubin The
coordination chemistry of graphene oxide: Interactions with metal ions in water. Russ. J. Coord. Chem. 39 (2013) 487-92.
20
[12]
Z. Li, F. Chen, L. Yuan, Y. Liu, Y. Zhao, Z. Chai, et al. Uranium(VI) adsorption
on graphene oxide nanosheets from aqueous solutions. Chem. Eng. J. 210 (2012) 53946. [13]
R. Sitko, E. Turek, B. Zawisza, E. Malicka, E. Talik, Heimann J, et al.
Adsorption of divalent metal ions from aqueous solutions using graphene oxide. Dalton Trans. 42 (2013) 5682-9. [14]
Y. Xu, K. Sheng, C. Li, G. Shi, Self-Assembled Graphene Hydrogel via a One-
Step Hydrothermal Process. ACS Nano. 4 (2010) 4324-30. [15]
L. Jiang, Z. Fan, Design of advanced porous graphene materials: from graphene
nanomesh to 3D architectures. Nanoscale. 6 (2014) 1922-45. [16]
S. Nardecchia, D. Carriazo, M. Luisa Ferrer, M.C. Gutierrez, F. del Monte,
Three dimensional macroporous architectures and aerogels built of carbon nanotubes and/or graphene: synthesis and applications. Chem Soc. Rev. 42 (2013) 794-830. [17]
C. Cheng, D. Li, Solvated Graphenes: An Emerging Class of Functional Soft
Materials. Adv. Mater. 25 (2013) 13-30. [18]
X. Mi, G.B. Huang, W.S. Xie, W. Wang, Y. Liu, J.P. Gao, Preparation of
graphene oxide aerogel and its adsorption for Cu2+ ions. Carbon 50 (2012) 4856-64. [19]
G. Goncalves, P.A.A.P. Marques, C.M. Granadeiro, H.I.S. Nogueira, M.K.
Singh, J. Gracio, Surface Modification of Graphene Nanosheets with Gold Nanoparticles: The Role of Oxygen Moieties at Graphene Surface on Gold Nucleation and Growth. Chem. Mater. 21 (2009) 4796-802. [20] M.A. Worsley; S.O. Kucheyev, H.E. Mason; M.D. Merrill, B.P. Mayer, J. Lewicki, et. al., Mechanically robust 3D graphene macroassembly with high surface area, Chem. Commun. 48 (2012) 8428-8430.
21
[21] C.B. Lopes, P. Figueira, D.S. Tavares, Z. Lin, A.L. Daniel-da-Silva, A.C. Duarte, J. Rocha, T. Trindade, E. Pereira, Core–shell magnetite-silica dithiocarbamatederivatised particles achieve the Water Framework Directive quality criteria for mercury in surface waters, Environ. Sci. Pollut. Res. 20 (2013) 5963-5974. [22] C. Lopes, J. Oliveira, L. Rocha, D. Tavares, C. Silva, S. Silva, N. Hartog, A. Duarte, E. Pereira, Cork stoppers as an effective sorbent for water treatment: the removal of mercury at environmentally relevant concentrations and conditions, Environ. Sci. Pollut. Res. 21 (2014) 2108-2121. [23]
S. Lagergren About the theory of so-called adsorption of soluble substances,
Kung Sven Veten Hand. 24 (1898) 1-39. [24]
Y.S. Ho, G. McKay, Pseudo-second order model for sorption processes, Process
Biochem. 34 (1999) 451-65. [25]
M.J.D. Low, Kinetics of Chemisorption of Gases on Solids, Chem Rev. 60
(1960) 267-312. [26]
G.E. Boyd, A.W. Adamson, L.S. Myers The Exchange Adsorption of Ions from
Aqueous Solutions by Organic Zeolites, J. Am. Chem. Soc. 69 (1947) 2836-48. [27]
W.J. Weber, J.C. Morris, Kinetics of adsorption on carbon from solution.
Kinetics of Adsorption on Carbon from Solution. 1963. [28]
M.A. Worsley, S.O. Kucheyev, H.E. Mason, M.D. Merrill, B.P. Mayer, J.
Lewicki, et al. Mechanically robust 3D graphene macroassembly with high surface area, Chem. Commun. 48 (2012) 8428-30. [29]
C. Zhang, R. Hao, H. Liao, Y. Hou Synthesis of amino-functionalized graphene
as metal-free catalyst and exploration of the roles of various nitrogen states in oxygen reduction reaction, Nano Energy 2 (2013) 88-97.
22
[30]
S.J. Gregg, K.S.W. Sing, Adsorption, surface area, and porosity. London:
Academic Press; 1982. [31]
Council Directive 98/83/EC of 3 November 1998 on the quality of water
intended for human consumption. Journal of the European Communities. 1998. [32]
V. Chandra, K.S. Kim, Highly selective adsorption of Hg2+ by a polypyrrole-
reduced graphene oxide composite, Chem Commun. 47 (2011) 3942-4. [33]
J. Berg, J. Tymoczko, L. Stryer Protein Composition and Structure.
Biochemistry. 7th ed: W.H. Freeman and Company 2012. [34]
C.W. Cheung, J.F. Porter, G. McKay, Sorption kinetics for the removal of
copper and zinc from effluents using bone char, Sep. Purif. Technol. 19 (2000) 55-64. [35]
G.F. Malash, M.I. El-Khaiary, Piecewise linear regression: A statistical method
for the analysis of experimental adsorption data by the intraparticle-diffusion models, Chem. Eng. J. 163 (2010) 256-63. [36]
L. Rocha, C. Lopes, J.A. Borges, A.C. Duarte, E. Pereira, Valuation of
Unmodified Rice Husk Waste as an Eco-Friendly Sorbent to Remove Mercury: a Study Using Environmental Realistic Concentrations, Water Air Soil Pollut. 224 2013 1-18. [37]
Y.S. Ho, J.C.Y. Ng, G. Mckay, Kinetics of pollutant sorption by biosorbents:
Review, Separ. Purif. Method. 29 (2000) 189-232. [38]
E. Sasmaz, A. Kirchofer, A.D. Jew, A. Saha, D. Abram, T.F. Jaramillo, et al.
Mercury chemistry on brominated activated carbon, Fuel 99 (2012) 188-96. [39]
N. Li, R.B. Bai, C.K. Liu, Enhanced and selective adsorption of mercury ions on
chitosan beads grafted with polyacrylamide via surface-initiated atom transfer radical polymerization. Langmuir. 21 (2005) 11780-7. [40]
G.P. Rao, C. Lu, F. Su, Sorption of divalent metal ions from aqueous solution
by carbon nanotubes: A review, Sep. Purif. Technol. 58 (2007) 224-31.
23
[41]
A. Gunay, E. Arslankaya, I. Tosun, Lead removal from aqueous solution by
natural and pretreated clinoptilolite: Adsorption equilibrium and kinetics, J. Hazard Mater. 146 (2007) 362-71. [42]
B.H. Hameed, M.I. El-Khaiary, Malachite green adsorption by rattan sawdust:
Isotherm, kinetic and mechanism modeling, J. Hazard. Mater. 159 (2008) 574-9. [43]
B. Tawabini, S. Al-Khaldi, M. Atieh, M. Khaled, Removal of mercury from
water by multi-walled carbon nanotubes, Water Sci. Technol. 61 2010 591-8. [44]
J. Dong, Z.H. Xu, F. Wang, Engineering and characterization of mesoporous
silica-coated magnetic particles for mercury removal from industrial effluents, Appl Surf. Sci. 254 (2008) 3522-30. [45]
G.Z. Kyzas, E.A. Deliyanni, K.A. Matis, Graphene oxide and its application as
an adsorbent for wastewater treatment, J. Chem. Technol. Biot. 89 (2014) 196-205. [46]
J. Bao, Y. Fu, Z. Bao, Thiol-functionalized magnetite/graphene oxide hybrid as a
reusable adsorbent for Hg2+ removal, Nanoscale Res. Lett. 8 (2013) 486-92. [47]
R.G. Pearson, Absolute electronegativity and hardness – application to
inorganic-chemistry, Inorg Chem. 27 (1988) 734-40. [48]
V. Gomez-Serrano, A. Macias-Garcia, A. Espinosa-Mansilla, C. Valenzuela-
Calahorro, Adsorption of mercury, cadmium and lead from aqueous solution on heattreated and sulphurized activated carbon, Water Res. 32 (1998) 1-4. [49]
L.S. Rocha, C.B. Lopes, B. Henriques, D.S. Tavares, J. Borges, A.C. Duarte, et
al. Competitive effects on mercury removal by an agricultural waste: application to synthetic and natural spiked waters. Environ. Technol. 35 (2014) 661-73. [50]
J. Wang, B. Deng, H. Chen, X. Wang, J. Zheng Removal of Aqueous Hg(II) by
Polyaniline: Sorption Characteristics and Mechanisms, Environ Sci. Technol. 43 (2009) 5223-8. 24
Figure 1. Digital image of 3DGON foam a); SEM images of 3DGON internal microstructure at different magnifications b) and c). XPS survey spectra of 3DGON d) with high resolution C1s e) and N 1s f). XPS spectra with deconvoluted peaks
25
Figure 2: Variation of the normalized concentration of Hg(II) in solution (Ct/C0) during the contact time (h) with 10 mg L-1 of 3DGO (■), 3DGON (●) and 3DGOSN (∆). Initial concentration of Hg(II) in ultra-pure water: 50 µg L-1. The dotted line correspond to the control (Hg(II) spiked ultra-pure water in the absence of sorbent).
26
Figure 3. Variation in the amount of Hg(II) removed (R (%)) for different 3DGON dosage used (mg L-1), after 24 and 48 hours of contact time. Initial concentration of Hg(II) in ultra-pure water: 50 µg L-1.
27
Figure 4: Variation in the amount of Hg(II) removed (R (%)) for different initial solution pH, after 3, 12 and 24h of contact time. Initial concentration of Hg(II) in ultra-pure water: 50 µg L-1.
28
Figure 5. Kinetic modelling of the experimental data obtained from the sorption process of Hg(II) onto 3DGON. a) using three kinetic reaction models: pseudo-first order (…), pseudo-second order (▬) and Elovich (---). b) using Webber’s intraparticle-diffusion model.
29
Figure 6. XPS spectra for 3DGON surfaces before and after Hg(II) sorption (a). Highresolution spectra of Hg 4f (b) and N 1s (before (c) and after Hg sorption(d)).
30
Figure 7. Equilibrium data (black squares) and sorption isotherms for the Hg(II)-3DGON system (21±1ºC). Initial concentration of Hg(II) in MQ water 50 µg L-1 and 1.0, 2.5, 5.0 10.0 and 20.0 mg L-1 of 3DGON.
31
Figure 8. Variation of the normalized concentrations of Hg(II) (■), Pb(II) (●) and Cd(II) (∆) in solution (Ct/C0) during the contact time (h) with 10 mg L-1 of 3DGON. Initial concentration of each metal in the mixture: 50 µg L-1. The dotted line correspond to the Hg(II) control (mixture in absence of sorbent).
32
Figure 9. Variation of the normalized concentrations (Ct/C0) of Hg(II) in spiked ultrapure water (■), river water (○) and seawater (▲) during the contact time (h) with 10 mg L-1 of 3DGON. Initial concentration of Hg(II): 50 µg L-1. The black and the dotted lines correspond to the controls (spiked solutions in absence of sorbent) of river water and seawater, respectively.
33
Table 1. Kinetic parameters obtained by fitting the experimental data on Hg(II) uptake by 3DGON to the following reaction models: Pseudo-first order, Pseudo-second order and Elovich. Experimental qe and the goodness of fit were also presented in order to assess and compare models. Pseudo-first order
Best fit values qe exp
qe1 ±
Initial [Hg]
± SD
SD
µg L-1
(µg g1 )
(µg g1 )
(h-1)
50
4515± 35
4173 ± 234
0.619 ±0.13 7
Pseudo-second order
Goodness of fit
qe2 ±
k1 ± SD
Best fit values
SD 2
R
0.9 17 6
Sy.x
464
Goodness of fit
(µg g1 )
(h-1)
4528 ± 170
0.000 2±0.0 0004
Best fit values
β ± SD
k2 ± SD
Elovich
2
R
0.9 72 2
Sy.x
270
(g µg1 ) 0.0014 ±0.000 1
Goodnes s of fit
α± SD R2
Sy. x
0.9 81 9
21 7
(µg g1 -1 h ) 16010 ±4489
34
Table 2. Kinetic parameters obtained by fitting the experimental data on Hg(II) uptake by 3DGON (10 mg L-1) to the following diffusion models: Boyd’s film-diffusion and Webber’ s intraparticle-diffusion. Boyd’s film-diffusion
Webber’ s intraparticle-diffusion
Stag e
intercept
R2
Breakpoint (h)
Ki (µg g-1 h-1/2)
R2
0.03478 1st
[0.02865;0.09839 ]
0.9941 (n=6)
7.89
1225.185
0.9936 (n=6)
2nd
-
-
-
79.79848
0.8026 (n=3)
35
Table 3. Isotherm constants obtained from fitting the equilibrium data for Hg(II) sorption onto 3DGON to Langmuir, Freundlich and SIPS isotherm models. Langmuir Isotherm
Best fit values qm ± SD
Goodness of fit
bL ± SD
g-
)
(L µg1 )
35022 ± 1946
0.039 ± 0.004
Best fit values KF ± SD
R2 (µg 1
Freundlich Isotherm
Sy.x (µg1-n Ln/g)
0.9 98 4
328
2067 ± 204
1/nF ± SD
0.65 ± 0.03
Goodness of fit
R
2
0.9 961
SIPS Isotherm
Best fit values
Goodnes s of fit
qm ± Sy.x SD
1/n ± SD
bS ± SD
R2
Sy. x
449 66± 105 78
0.88 ± 0.10
0.02 3± 0.01 1
0.9 99 1
29 3
516
36
S0304-3894(15)30083-2 http://dx.doi.org/doi:10.1016/j.jhazmat.2015.09.028 HAZMAT 17101
To appear in:
Journal of Hazardous Materials
Received date: Revised date: Accepted date:
1-6-2015 9-9-2015 13-9-2015
Please cite this article as: Bruno Henriques, Gil Gonc¸alves, Nazanin Emami, Eduarda Pereira, Mercedes Vila, Paula A.A.P.Marques, Optimized graphene oxide foam with enhanced performance and high selectivity for mercury removal from water, Journal of Hazardous Materials http://dx.doi.org/10.1016/j.jhazmat.2015.09.028 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Optimized graphene oxide foam with enhanced performance and high selectivity for mercury removal from water
Bruno Henriquesa, Gil Gonçalvesb,c*, Nazanin Emamie, Eduarda Pereiraa, Mercedes Vilab,c,d*, Paula A.A.P. Marquesb,c,
a
b
CESAM & Department of Chemistry, University of Aveiro, 3810-193 Aveiro, Portugal
TEMA-NRD, Mechanical Engineering Department, University of Aveiro, 3810-193 Aveiro, Portugal c
Aveiro Institute of Nanotechnology, AIN, University of Aveiro, 3810-193 Aveiro, Portugal d
e
Centro de Investigación Biomédica en Red. Bioingeniería, Biomateriales y Nanomedicina, CIBER-BBN, Spain
Department of Applied Physics and Mechanical Eng, Lulea University of Technology, Sweden.
*Corresponding authors. E-mail: [email protected] (Gil Gonçalves); [email protected] (Mercedes Vila)
1
Graphical Abstract
2
Highlights
There is a lack of studies on Hg removal from natural waters at realistic levels Three-dimensional graphene oxide were modified to enhance the affinity toward Hg The removal process was optimized in terms of pH, mass and contact time. The process efficiency was assessed in multimetallic solutions and natural matrices Small dose of GO foam (10 mg L-1) allows to remove up to 96% of Hg(II) after 24h
3
Abstract This work explores the preparation of three-dimensional graphene oxide macroscopic structures, shaped by self-assembling single graphene oxide (3DGO) sheets with control of its surface chemistry by combining with nitrogen functional groups (3DGON), or with nitrogen and sulphur functional groups (3DGOSN), and their application in the removal of mercury (Hg(II)) from aqueous solutions. The chemical structure of the materials was assessed by using different characterization techniques: SEM, XPS and BET. Adsorption studies conducted in Hg(II) contaminated ultra-pure water reveal the enhanced ability of 3DGON for the adsorption of this metal, when compared to the other GO foams. A small dose of 3DGON (10 mg L-1) allows to remove up to 96% of Hg(II) after 24h of contact time, leading to a residual concentration in solution close to the guideline value for drinking water (1 µg L-1). The ability of this material to adsorb Hg (II) was evaluated relatively to different experimental parameters such as pH, sorbent dose, time and effect on different competing metal ions. Real application was also evaluated by testing its performance in two different natural matrices, river and sea water, with very promising results.
Keywords: mercury, graphene oxide, sorption, competitive ions, remediation, natural waters
4
1. Introduction Carbon based materials have been, since long time, proved to be one of the best options for water purification [1]. Industrially, the most relevant in our days is activated carbon [2] due to its versatility on removing various organic compounds and also metallic ions, such as Cr3+, Hg2+, Pb2+, Cr6+, Cd2+, Ni2+ and Zn2+ [3]. The removal process by activated carbon includes adsorption, absorption and ion exchange, being the overall process designated as sorption [4]. Recently, other carbon materials with higher specific surface areas have been proposed for purifying water, such as carbon nanotubes [5] and graphene based materials [6]. Among them, graphene oxide (GO) stands out from others because of its high oxidation degree and the ability of self-assembly in 3D macrostructures. GO is usually synthetized by the chemical exfoliation of graphite [8]. This approach has the advantage of large scale and low cost production and high efficiency [7]. GO presents a high degree of oxidation ratio (1.8< C/O <2.5) [9] which grants this material with a high hydrophilic character, allowing highly stable aqueous colloidal suspensions [10]. Furthermore, the presence of oxygen functional groups, which acts as electron donors, allows the easy coordination with metallic ions [11]. The ability of graphene-based materials to adsorb metallic ions from aqueous solutions has been already reported for: Pb2+ [12, 13], Sb3+ [14] Cd2+ [15], Co2+ [15] and U5+ [12]. The capacity of GO for the adsorption of divalent metal ions (Cu2+, Cd2+, Pb2+ and Zn2+) in aqueous solution has been discussed by Stiko et al, as the highest than any reported sorbent [13]. Recently, graphene based three-dimensional structures were synthetized by the selfassembling GO sheets [14] with or without the addition of additives [15-17]. The resulting architectures have been already explored for water purification, typically for the adsorption and degradation of organic contaminants [6]. However, the elimination of metallic ions from polluted waters using these 3D structures is of recent interest. Mi et.
5
al. reported the preparation of GO aerogels by a unidirectional freeze-drying method for adsorption of Cu2+ [18] showing that the GO aerogels reached the equilibrium in a shorter period than oxidized carbon nanotubes (CNT) sheets or activated carbon. Li et. al. showed that 3D graphene macrostructures prepared by Chemical Vapour Deposition growth in a nickel substrate present ultra-high electrical adsorption ability for heavy metals ions (Cd2+, Pb2+, Cu2+, Ni2+) removal from water solutions by using this material as an electrode [17]. Herein, we report for the first time the development of GO macrostructures with a controlled surface chemistry and with high affinity for the adsorption of mercury from aqueous solutions. GO macrostructures were prepared at different experimental conditions allowing the presence of surface modifications with oxygen, nitrogen or sulphur functional groups. The adsorption studies in mercury contaminated waters reveal the superior capacity of GO macrostructures modified with combination of oxygen and nitrogen functional groups (3DGON). Different experimental parameters such as pH, mass, time and effect of different competing metal ions were explored. The potential applicability of this material was also evaluated by testing its performance in river and sea waters. 2. Experimental 2.1. Synthesis of three dimensional GO structures GO was prepared by the chemical exfoliation of graphite (Graphite powder, <45 μm, ≥99.99%, Sigma-Aldrich) following a method described in our previous work [19]. The 3D structures were build up by the simple self-assembly of GO sheets in water (4 mg/mL) inside an autoclave (180ºC; 12h) (3DGO) [18]. The preparation of 3DGON and 3DGOSN structures was performed under the same experimental conditions described before by 6
adding respectively, 1mL of NH4OH (20% v/v) or 10 mg of NH4SCN to the initial GO solution.[20] 2.1.1 Structural characterization The surface chemistry of the 3D macroscopic materials was analysed by XPS (VG Scientific ESCALAB 200 A; UK). The morphological analysis was made using a field emission scanning electron microscope (Model SU70; Hitachi, Tokyo, Japan), operated at 15 kV.
2.2. Hg sorption studies 2.2.1 Sorption kinetics Batch sorption experiments were conducted using a Schott Duran® glass bottle (1 L) as reaction vessel, where sorbent and solution were maintained under constant magnetic stirring (700 rpm), at 21±1ºC. The Hg(II) solution was prepared by adding the desired volume of a certified Hg standard solution (1001±2 mg L−1 of Hg(II) in HNO3 0.5 mol L−1, from Merck) to ultra-pure water. The selected concentration of Hg(II) aimed to simulate a real scenario, since 50 µg L-1 is the old maximum permissible value for Hg discharges from industrial sectors (Directive 84/156/EEC). The rate of Hg(II) uptake by GO foams was followed by determining the concentration of this metal in solution samples (5-10 mL) collected at defined time intervals. All the samples were immediately centrifuged at 5000 rpm for 3 min, the supernatant liquid was collected, acidified to pH ≤ 2 using Suprapur© HNO3 (65% v/v) and stored at 4ºC until analysis. The removal trial was conducted in duplicate with a control (solution of Hg(II) in the absence of sorbent) running in parallel. The initial pH of the solutions was ca. 4.5 and no significant variations were observed during the removal process. The quantification of Hg in all water samples was performed by cold vapour atomic fluorescence spectroscopy (CV-AFS), using a PSA 7
10.025 Millennium Merlin Hg analyser and SnCl2 (2% m/v in HCl 10% v/v) as a reducing agent. 2.2.2 Effect of dosage The effect of the sorbent dosage on Hg(II) sorption was evaluated by placing different amounts of 3DGON (1, 2.5, 5, 10 and 20 mg L−1) in contact with a 50 µg L−1 Hg(II) solution. The experiments and the Hg(II) quantification were carried out as mentioned above. 2.2.3 Effect of pH The effect of pH on the sorption process was investigated for an initial Hg(II) concentration of 50 µg L−1 in ultra-pure water, with different initial pH values: 3, 5, 7 and 9. The pH adjustments were performed using HNO3 (31.5 g L−1) or NaOH (4.0 g L−1) solutions, before the addition of the pre-weighed sorbent (10 mg L-1). The experiments were carried out as described before. 2.2.4 Effect of the presence of co-ions The sorption of Hg(II) by 3DGON under metal competitive conditions was studied for ternary mixtures of Hg(II), Pb(II) and Cd(II) in equal concentration (50 µg L−1). Experiments were performed for an initial pH of ca. 4.5 following the same design described on 2.2.1. The spiked solutions were prepared by adding the desired volume of the each metal standard solution (1000±2 g L-1 of Pb(II) in HNO3 0.5 mol L−1, Panreac; 1000±2 mg L-1 of Cd(II) in HNO3 0.5 mol L−1, Merck) to ultra-pure water. The concentration of Hg(II) in solution was determined by CV-AFS, whereas Pb(II) and Cd(II) quantification was performed by inductively coupled plasma spectroscopy, using a Jobin–Yvon JY70 Plus Spectrometer. The limits of detection for Pb(II) and Cd(II) where found to be 6.7 µg L−1 and 1.3 µg L−1, respectively, with an acceptable relative standard deviation among replicates: <10 %.
8
2.2.4 Assessment of Hg(II) removal from natural waters The efficiency of 3DGON in natural waters was evaluated by conducting sorption tests in river and sea waters spiked with Hg(II), following the experimental design described in section 2.2.1. The river and sea waters were collected in the Vouga river (40°40′42′′ N, 08°22′18′′ W) and on the coast of Aveiro (40°32'58"N 8°46'31"W), respectively, and physicochemical characterization was previously reported [21, 22]. All water was filtered through 0.45 μm pore size filters and the solutions were left to pre-equilibrate during 24h before adding the sorbent. Initial pH of the spiked solutions were, respectively, ca. 6.5 and 7.8 for river water and seawater. 2.3 Analysis of sorption data The amount of Hg(II) bound per unit of sorbent mass, at a given time t (qt, µg g−1), was deduced from the mass balance between the initial Hg concentration in the solution (C0, µg L−1) and the concentration after a particular period of contact time t (Ct, µg L−1): qt
(C 0 C t ) V
(1)
m
where V (L) is the volume of the solution and m (g) is the sorbent mass. When equilibrium was attained: t=te, qt=qe and Ct=Ce (residual Hg(II) concentration in solution). The performance of the removal process was evaluated and compared using the Hg(II) removal percentage (R, %), which at time t is defined by: R t (%)
(C
0
Ct) C
100
(2)
0
2.3.1 Kinetics and Equilibrium models In this study, the kinetics of Hg(II) uptake by 3DGON was studied by applying three widely used reaction models to the experimental data, namely, the Lagergren pseudofirst-order model [23], the Ho’s pseudo-second-order model [24] and the Elovich model
9
[25]. In addition to the reaction models, two widely known diffusion-based models, Boyd’s film-diffusion [26] and Webber’s pore-diffusion [27], were used. (Details at SI) The experimental equilibrium data was fitted using 3 different models in their non-linear forms: the two-parameters Langmuir and Freundlich isotherm models and the threeparameters Sips or Langmuir-Freundlich isotherm model. (Details at SI)
3. Results and discussion 3.1. Characterization of the 3D GO structures The hydrothermal treatment of GO aqueous solutions results in the assembly of 3D macrostructures (Figure 1a). The SEM analyses show an interconnected porous network of partial overlapping flexible GO nanosheets with a foam-like structure, Figure 1b and c, for 3DGON. The chemical mechanism behind the assembly of GO nanosheets into 3D architectures with interconnected porous network is not yet completely understood. Wosley et al. proposed that the GO functional groups are the main responsible for the crosslinking and respective assembly onto 3D macrostructures [28]. Xu et al claim that non-covalent bonds like π-π stacking interactions between GO nanosheets are the main responsible for this assembly. HR-XPS spectra with curve fittings for C1s (Figure 1 d) and e)) shows high restructuration of sp2 carbon, presence of C-O functional groups and an increase of π-π stacking interactions. The curve fittings for N1s (Figure 1f) confirm the doping with different N-type groups of GO: pyridinic (31.9 at.%), amino (42.1 at.%), pyrrolitic (16.2 at.%), graphitic (4.1 at.%) and oxidized N (5.7 at.%) [29]. These results suggest that 3DGON are hybrid systems resulting from the assembly of GO nanosheets with the establishment of covalent bonds between carbon and nitrogen functional groups and π-π noncovalent interactions. 10
Analysis of the specific surface area of 3DGON macrostructure using nitrogen adsorption/desorption isotherms, show that they exhibit type IV isotherm patterns with H3-type hysteresis loop following the BDDT classification, which are characteristics of the mesoporous materials [30] (Figure S1), with a BET surface area of ~400 m2g-1 and a pore volume of 0.27 cm3g-1 (1.6
11
The optimization of the 3DGON sorbent dosage (mass of sorbent per volume of solution) is essential to maximize the efficiency of the sorption system. Figure 3 presents the removal of Hg(II) for different sorbent dosage (1.0, 2.5, 5.0, 10.0, and 20 mg L-1), after 24 and 48 hours of contact time. The results show that higher percentages of Hg(II) removal are obtained with the increase of the sorbent concentration, from 33% to 96% for doses 1 to 20 mg L-1 respectively, and the reaction equilibrium was reached after 24h contact time. It is evident that the increase of sorbent mass leads to a larger number of available coordination sites improving the removal efficiency of sorbent. However it was observed that using 10 mg L-1 of 3DGON allows to obtain removal values (95%) very similar to the maximum reached with 20 mg L-1. The significant increase of sorbent concentration did not lead to an enhancement in the sorption extension, revealing that the process is not only dependent on the number of free sites but also on the equilibrium of charges between sorbent and solution. 3.2.3 Effect solution pH The effect of the solution pH on the sorption of Hg(II) by GO3DN foams, at different periods of contact time (3, 12 and 24h) is presented in Figure 4. The results suggest that the sorption efficiency is almost independent of the solution pH on the interval between 3 and 9. After 3, 12 and 24 hours of contact time the amplitude on the removal percentage, for the different pH values studied, was 3.8 and 8.0% respectively. Actually other sorption studies of metals by GO based materials showed different sorption stages, lower efficiency for pH values lower than 3.9 and higher sorption rates for higher pH values [13, 18]. The decrease of sorption at acidic pH is usually attributed to the preferential protonation of the surface functional groups of the material over the biding to the metal ions. In the case of 3DGON foams the presence of nitrogen simultaneously with oxygen functional groups allows to maintain the same Hg(II) sorption efficiency at lower pH 12
values. From our best knowledge, 3DGON shows an acid basic behaviour similar to that of amino acids, which have the capacity to form zwitterionic species by establishing ionic dipole (-COO- and the adjacent -NH3+) on the pH range of 2 to 9 [33]. 3.3. Sorption kinetics Three kinetic reaction models in their non-linear form were used to examine the sorption of Hg(II) by 3DGON foam, namely pseudo-first-order (PFO), pseudo-second order (PSO) and Elovich model. The fittings of the experimental data and the parameters values obtained from the kinetic modelling are presented in Figure 5a) and Table 1, respectively. Clearly among the models tested and the under the experimental conditions used the PFO model had the poorest performance on describing the experimental data, which is corroborated by the lowest R2 (0.917) and the highest Sy.x (464). Moreover this model was not able to correctly predict the experimental qe, underestimating its value (qe1= 4173 vs qe exp= 4515), contrary to the PSO model which precisely estimated the qe value (relative error lower than 0.3%). Despite the good agreement between the experimental data and the fittings accomplished by both PSO (R2=0.971) and the Elovich (R2=0.981) models, comparison using Akaike's information criteria (AIC) obtained for each model indicates that Elovich model is more likely to be correct (probability of 89.7%). This gives indication that the Hg(II) sorption process has a chemisorption nature as well as the surface of the sorbent is heterogeneous [34].
To obtain more insight into the mechanism of Hg(II) sorption by 3DGON and potential rate-controlling steps two diffusion-based models were explored: Boyd’s film-diffusion and Webber’s intraparticle-diffusion. The results of the piecewise linear regression (PLR) treatment [35] applied to the experimental data are presented in Table 2. From Boyd filmdiffusion fitting it was obtained a first linear segment (R2=0.9941) and the confidence 13
interval of the intercept (95 % confidence level) includes the zero, which strongly suggests that film-diffusion is not the rate determining step in the sorption process [36]. The Webber’s intraparticle-diffusion´s plot is presented in Figure 5b. The existence of multiple linear stages is clear which shows that more than one diffusion step takes place. The Akaike’s information criteria were used to define the numbers of linear segments. Lower AIC value was obtained when considering only two linear segments and the calculated evidence ratio indicates that the two stages approach is 1.13×107 times more likely to be the correct model than the approach considering three stages. The first steeper section (until 7.89 h) does not pass through the origin which points that intra-particle diffusion is not only the rate-controlling step, but is also affected by the boundary layer diffusion process [34]. Another phenomenon could be a surface enhancement associated with highly energetic heterogeneous surface, leading to a high rapid uptake of Hg(II) on the sorbent before the system stabilizes [37]. The second less pronounced portion is attributed to the final equilibrium stage for which the intra-particle diffusion starts to slow down due to low Hg(II) concentration left in the solution.
3.4. Chemical mechanism of Hg(II) sorption XPS survey spectra for 3DGON surfaces before and after Hg(II) sorption were conducted to advance the perception into the sorption mechanism (Figure 6a). As expected, the structural analysis exhibits the peaks matching C1s, O1s and N1s and Hg4f after exposure to the Hg (II) spiked water. High-resolution spectra of Hg4f, in Figure 5b, shows two peaks at 100.9 eV and 104.8 eV, that can be assigned to the presence of Hg 4f7/2 and 4f5/2. Those results indicates that mercury is adsorbed in an oxidized state instead of its metallic form (Hg0, Hg 4f7/2 – 104 eV and 4f5/2 – 99.9 eV) [38] suggesting that the main sorption mechanism is oxidative chemisorption. 14
HRXPS of 3DGON at N1s region can be deconvoluted into five Gaussian peaks (Figure 6c), Pyridinic, Amino, Pyrrolic, Graphitic and Oxidized N. However after exposure to Hg(II), the N1s peak correspondent to graphitic structures disappears. Moreover the exposure of 3DGON to Hg(II) aqueous solution promotes significant changes in the N1s peaks areas, meaning a different relative distribution of nitrogen species or/and the formation of new covalent bonds with metal. The higher decrease on the peaks area corresponds to amine (9.1 at%) and pyridinic groups (4.8 at%) suggests that are the preferential groups for the coordination with the Hg(II). Li et al. described the formation of Hg(II)-NH2 complex on the sorption mechanism of Hg(II) by chitosan beads grafted with polyacrylamide [39]. The soft basic ligand character of the amino functional groups can favors an easy interaction with very soft acid such as Hg(II). HRXPS of C1s spectrum shows a decrease of 3.26 at% for C=O/COOH groups after Hg(II) sorption at equilibrium, which indicates that oxygen functional groups also play an active role on the interaction with Hg (II). The chemical interaction of carboxylic and phenolic groups by proton exchange with metal ions was already reported for oxidized CNTs,[40] as the main sorption mechanism for divalent metal ions in aqueous solution. 3.5. Sorption Isotherms Isotherm models allow the quantitative assessment of sorbent’s sorption capacity, giving indication about its sorption strength. In this study, Langmuir, Freundlich and SIPS isotherm models, in their nonlinear form, were applied to the experimental equilibrium results for the Hg(II)-3DGON system, at 21±1ºC ([Hg(II)] of 50 µg L-1 and 3DGON dosages of 1, 2.5, 5, 10 and 20 mg L-1). Figure 7 shows the fitted equilibrium data, while Table 3 presents the isotherms parameters and the good fit. Under the experimental conditions studied, both Langmuir and Freundlich isotherms models successfully described the equilibrium data, which implies that both monolayer adsorption and 15
heterogeneous surface conditions existed [18, 41]. However, the good fit (Table 3) indicates that Langmuir isotherm model performed slightly better than Freundlich isotherm model. The comparison between Langmuir and SIPS isotherms using AIC (since these two models have different degrees of freedom) suggests that the former is more likely to be adjusted to the experimental data than the latter (AIC Langmuir=85.10 and AICSIPS=112.0). In fact, the heterogeneity index from the SIPS isotherm, 1/n=0.88±0.10, points to a heterogeneous or multilayer system, yet this value is not far away from the unit, case where SIPS isotherm reduces to Langmuir isotherm. The maximum sorption capacity of Hg(II) by 3DGON predicted by Langmuir isotherm model is c.a. 35000 µg g1
, which highlights the potential of this material. Moreover the separation factor, RL,
characteristic of the Langmuir isotherm, indicates that the sorption is a favourable process (RL=0.33), almost irreversible at high initial Hg(II) concentrations (RL→0) [41, 42]. The performance evaluation of 3DGON on mercury removal regarding other sorbent materials is not a simple task, and not always feasible. Experimental conditions, such pH, temperature and sorbate/sorbent ratio are very different among studies, being the results not entirely comparable. It should be emphasized that the majority of these studies have focused on “unrealistic” concentrations of Hg(II), i.e. using initial concentrations extremely larger than those that may be found in effluents or aquatic systems affected to wastewater discharges. Furthermore, the dose of sorbent used in most of these works (up to 50,000 times higher than ours) did not reflect an environmental concern with the final residue generated (Table S1). Using an initial Hg(II) concentration of 0.1 mg L-1 and sorbent/solution ratios from 200 to 800 mg L-1 Tawabini et al. [43] found that the maximum sorption capacity of MWCNTs is 13.16 mg g-1 (less than half of the value found for 3DGON). Similar value (14 mg g-1) was reported by Dong et al. [44] for mesoporous sílica-coated magnetic particles, using 2500 mg L-1 and concentrations of
16
Hg(II) between 10 to 60 mg L-1. Two of the few studies with GO-based materials applied to Hg(II) removal were performed by Kyzas et al. [45] (Magnetic GO, qe=14.02 to 123.10 mg g-1) and by Bao et al. [46] (GO, qe=5 mg g-1; Fe3O4/GO, qe=20 mg g-1; Thiolfunctionalized-Fe3O4/GO, qe=30.94 mg g-1). However, in addition to the very high concentrations of Hg(II) used (200-10000 times higher than the limit for effluent discharges), there is a lack of information in both studies regarding the pH and the amount of sorbent used, which prevents a fair comparison with 3DGON. Overall, these results demonstrate the efficacy of 3DGON foam in the treatment of contaminated waters with Hg(II) levels representative of environmental real situations.
3.6. Effect of co-ions on Hg(II) sorption The selectivity of 3DGON to Hg(II) ions was evaluated by the competitive adsorption with Pb(II) and Cd(II) co-ions. Figure 8 showed the preferential affinity for this material to the Hg(II) > Pb(II) > Cd(II) ions with removal percentage of approximately 91%, 10% and 5% at equilibrium, respectively. That high selectivity for Hg(II) by GO nanocomposites in co-ions solutions was already observed by Chandra et al [32] however they didn’t show any plausible explanation for the selective adsorption of the metals. It is know that divalent metal ions can be present in deionized water in different ionized species M2+, M(OH)+1, M(OH)20, M(OH)3-1… depending the percentage of each species of the form and concentration of negative ions in solution, the pH and electronic structure of metal. The simulations performed in the MinteQ software at constant pH of 4.5 for mixture Hg(II), Pb(II) and Cd(II) showed that the predominant species of each element on aqueous solutions are Hg(OH)2 (98.0%), Pb2+(92.12%) and Cd2+(98.0%), respectively (Table S2).
17
The preferential adsorption of metallic species by 3DGON in aqueous solutions can be explained by Pearson’ theory that postulates hard acids coordinates preferentially to hard bases and soft acids to soft bases [47]. Based on the chemical hardness (7.70 eV (Hg2+), 8.46 eV (Pb2+), 10.29 eV (Cd2+)), the metallic ions involved on the adsorption process, should have very similar adsorptive behaviour[48]. However, on contrary to the other metallic elements, the predominant species for Hg in aqueous solution is a neutral form Hg(OH)2, which is considered a softer base than the metallic cations (Hg2+, Pb2+ and Cd2+). Taking into account that the predominant groups at 3DGON structure are amines, that can be considered as a soft bases, the affinity with Hg(OH)2 will be higher, which is in agreement with our experimental results. 3.7. Sorption of Hg(II) on natural waters To evaluate the real efficiency of 3DGON under real conditions we perform tests of sorption of Hg(II) on river and seawater (Figure 9). The results showed that Hg(II) sorption in ultra-pure water is higher (93%), which can be explained by the fact that natural waters are complex matrices due to the presence of various ions that can compete for the binding sites at sorbent surface (Figure S4). Furthermore the organic matter existent on natural systems can complex with the metallic ions under study, decreasing its mobility and affinity for sorption [49]. Relatively to the natural waters, the results achieved show a higher Hg(II) sorption capacity on river water (82%) then seawater (42%) ecosystems. The larger concentration of positive ions (Ca+, Na+, Mg+2) in seawater leads to an increased competition for the functional groups on the surface of sorbent material. On the other hand, the high levels of Cl- in seawater enables the formation of mercury chloro-complexes (HgCl42−, HgCl3−…), which have less affinity to the sorbent and thus remain in the solution [50].
18
4. Conclusions In summary we observed that GO macrostructures are a very effective material on the sorption of Hg(II) from water solution. The different chemical surface modifications performed showed that the combination of oxygen and nitrogen functional groups (3DGON) increases the removal efficiency of Hg (II). The different acidic character of the nitrogen and oxygen functional groups allows shifting the equilibrium of the reaction towards the increase o Hg(II) sorption. Small dose of 3DGON (10 mg L-1) allows to remove up to 95% of Hg(II) after 24h of contact time, leading to a residual concentration close to the guideline value for drinking water (1 µg L-1). The Hg(II) sorption follows the Elovich model kinetics and the equilibrium data was well fitted by the Langmuir isotherms, with a predicted maximum sorption capacity of 35 mg g-1. The efficiency of this new material was not significantly affected by the pH and presence of co-ions Pb(II) and Cd(II) in solution, which highlights its great selectivity and affinity to Hg(II). Tests conducted in natural waters showed higher potential of 3DGON for real applications.
Ackowledgements: B. Henriques and G. Gonçalves would like to acknowledge the Portuguese Foundation for Science and Technology (FCT) for the financial support of the PostDoc grants (SFRH/BPD/112576/2015 and SFRH/BDP/84419/2012, respectively). M Vila and PAAP Marques thank the FCT for the investigator grant. We would like to thank FCT, the European Union, QREN, FEDER, COMPETE, for funding the TEMA Research Unit (PEst-C/QUI/UI0062/2013)
and
the
CESAM
associated
laboratory
(UID/AMB/50017/2013).
19
References [1]
P. Hadi, M-H. To, C-W. Hui, CSK. Lin, G. McKay, Aqueous mercury
adsorption by activated carbons, Water Res. 73 (2015) 37-55. [2]
R.C Bansal, M. Goyal, Activated Carbon Adsorption, CRC Press 2005, USA,
pp. 297-436. [3]
A. Bhatnagar, A.K. Minocha, Conventional and non-conventional adsorbents for
removal of pollutants from water - A review, Indian J. Chem. Techn. 13 (2006) 203-17. [4]
F. Fu, Q. Wang, Removal of heavy metal ions from wastewaters: A review, J.
Environ. Manage. 92 (2011) 407-18. [5]
X. Ren, C. Chen, M. Nagatsu, X. Wang, Carbon nanotubes as adsorbents in
environmental pollution management: A review, Chem. Eng. J. 170 (2011) 395-410. [6]
K.C. Kemp, H. Seema, M. Saleh, N.H. Le, K. Mahesh, V. Chandra et al.,
Environmental applications using graphene composites: water remediation and gas adsorption, Nanoscale 5 (2013) 3149-71. [7]
D.C. Marcano, D.V. Kosynkin, J.M. Berlin, A. Sinitskii, Z. Sun, A. Slesarev, et
al. Improved Synthesis of Graphene Oxide, ACS Nano 4 (2010) 4806-14. [8]
S. Park S, R.S. Ruoff, Chemical methods for the production of graphenes. Nat.
Nanotechnol. 4 (2009) 217-24. [9]
D.R. Dreyer, S. Park, C.W. Bielawski, R.S. Ruoff, The chemistry of graphene
oxide. Chem. Soc. Rev. 39 (2010) 228-40. [10]
J.I. Paredes, S. Villar-Rodil, A. Martinez-Alonso, J.M.D. Tascon Graphene
oxide dispersions in organic solvents. Langmuir 24 (2008) 10560-4. [11]
I.L. Laure, S.V. Tkachev, E.Y. Buslaeva, E.V. Fatushina, S.P. Gubin The
coordination chemistry of graphene oxide: Interactions with metal ions in water. Russ. J. Coord. Chem. 39 (2013) 487-92.
20
[12]
Z. Li, F. Chen, L. Yuan, Y. Liu, Y. Zhao, Z. Chai, et al. Uranium(VI) adsorption
on graphene oxide nanosheets from aqueous solutions. Chem. Eng. J. 210 (2012) 53946. [13]
R. Sitko, E. Turek, B. Zawisza, E. Malicka, E. Talik, Heimann J, et al.
Adsorption of divalent metal ions from aqueous solutions using graphene oxide. Dalton Trans. 42 (2013) 5682-9. [14]
Y. Xu, K. Sheng, C. Li, G. Shi, Self-Assembled Graphene Hydrogel via a One-
Step Hydrothermal Process. ACS Nano. 4 (2010) 4324-30. [15]
L. Jiang, Z. Fan, Design of advanced porous graphene materials: from graphene
nanomesh to 3D architectures. Nanoscale. 6 (2014) 1922-45. [16]
S. Nardecchia, D. Carriazo, M. Luisa Ferrer, M.C. Gutierrez, F. del Monte,
Three dimensional macroporous architectures and aerogels built of carbon nanotubes and/or graphene: synthesis and applications. Chem Soc. Rev. 42 (2013) 794-830. [17]
C. Cheng, D. Li, Solvated Graphenes: An Emerging Class of Functional Soft
Materials. Adv. Mater. 25 (2013) 13-30. [18]
X. Mi, G.B. Huang, W.S. Xie, W. Wang, Y. Liu, J.P. Gao, Preparation of
graphene oxide aerogel and its adsorption for Cu2+ ions. Carbon 50 (2012) 4856-64. [19]
G. Goncalves, P.A.A.P. Marques, C.M. Granadeiro, H.I.S. Nogueira, M.K.
Singh, J. Gracio, Surface Modification of Graphene Nanosheets with Gold Nanoparticles: The Role of Oxygen Moieties at Graphene Surface on Gold Nucleation and Growth. Chem. Mater. 21 (2009) 4796-802. [20] M.A. Worsley; S.O. Kucheyev, H.E. Mason; M.D. Merrill, B.P. Mayer, J. Lewicki, et. al., Mechanically robust 3D graphene macroassembly with high surface area, Chem. Commun. 48 (2012) 8428-8430.
21
[21] C.B. Lopes, P. Figueira, D.S. Tavares, Z. Lin, A.L. Daniel-da-Silva, A.C. Duarte, J. Rocha, T. Trindade, E. Pereira, Core–shell magnetite-silica dithiocarbamatederivatised particles achieve the Water Framework Directive quality criteria for mercury in surface waters, Environ. Sci. Pollut. Res. 20 (2013) 5963-5974. [22] C. Lopes, J. Oliveira, L. Rocha, D. Tavares, C. Silva, S. Silva, N. Hartog, A. Duarte, E. Pereira, Cork stoppers as an effective sorbent for water treatment: the removal of mercury at environmentally relevant concentrations and conditions, Environ. Sci. Pollut. Res. 21 (2014) 2108-2121. [23]
S. Lagergren About the theory of so-called adsorption of soluble substances,
Kung Sven Veten Hand. 24 (1898) 1-39. [24]
Y.S. Ho, G. McKay, Pseudo-second order model for sorption processes, Process
Biochem. 34 (1999) 451-65. [25]
M.J.D. Low, Kinetics of Chemisorption of Gases on Solids, Chem Rev. 60
(1960) 267-312. [26]
G.E. Boyd, A.W. Adamson, L.S. Myers The Exchange Adsorption of Ions from
Aqueous Solutions by Organic Zeolites, J. Am. Chem. Soc. 69 (1947) 2836-48. [27]
W.J. Weber, J.C. Morris, Kinetics of adsorption on carbon from solution.
Kinetics of Adsorption on Carbon from Solution. 1963. [28]
M.A. Worsley, S.O. Kucheyev, H.E. Mason, M.D. Merrill, B.P. Mayer, J.
Lewicki, et al. Mechanically robust 3D graphene macroassembly with high surface area, Chem. Commun. 48 (2012) 8428-30. [29]
C. Zhang, R. Hao, H. Liao, Y. Hou Synthesis of amino-functionalized graphene
as metal-free catalyst and exploration of the roles of various nitrogen states in oxygen reduction reaction, Nano Energy 2 (2013) 88-97.
22
[30]
S.J. Gregg, K.S.W. Sing, Adsorption, surface area, and porosity. London:
Academic Press; 1982. [31]
Council Directive 98/83/EC of 3 November 1998 on the quality of water
intended for human consumption. Journal of the European Communities. 1998. [32]
V. Chandra, K.S. Kim, Highly selective adsorption of Hg2+ by a polypyrrole-
reduced graphene oxide composite, Chem Commun. 47 (2011) 3942-4. [33]
J. Berg, J. Tymoczko, L. Stryer Protein Composition and Structure.
Biochemistry. 7th ed: W.H. Freeman and Company 2012. [34]
C.W. Cheung, J.F. Porter, G. McKay, Sorption kinetics for the removal of
copper and zinc from effluents using bone char, Sep. Purif. Technol. 19 (2000) 55-64. [35]
G.F. Malash, M.I. El-Khaiary, Piecewise linear regression: A statistical method
for the analysis of experimental adsorption data by the intraparticle-diffusion models, Chem. Eng. J. 163 (2010) 256-63. [36]
L. Rocha, C. Lopes, J.A. Borges, A.C. Duarte, E. Pereira, Valuation of
Unmodified Rice Husk Waste as an Eco-Friendly Sorbent to Remove Mercury: a Study Using Environmental Realistic Concentrations, Water Air Soil Pollut. 224 2013 1-18. [37]
Y.S. Ho, J.C.Y. Ng, G. Mckay, Kinetics of pollutant sorption by biosorbents:
Review, Separ. Purif. Method. 29 (2000) 189-232. [38]
E. Sasmaz, A. Kirchofer, A.D. Jew, A. Saha, D. Abram, T.F. Jaramillo, et al.
Mercury chemistry on brominated activated carbon, Fuel 99 (2012) 188-96. [39]
N. Li, R.B. Bai, C.K. Liu, Enhanced and selective adsorption of mercury ions on
chitosan beads grafted with polyacrylamide via surface-initiated atom transfer radical polymerization. Langmuir. 21 (2005) 11780-7. [40]
G.P. Rao, C. Lu, F. Su, Sorption of divalent metal ions from aqueous solution
by carbon nanotubes: A review, Sep. Purif. Technol. 58 (2007) 224-31.
23
[41]
A. Gunay, E. Arslankaya, I. Tosun, Lead removal from aqueous solution by
natural and pretreated clinoptilolite: Adsorption equilibrium and kinetics, J. Hazard Mater. 146 (2007) 362-71. [42]
B.H. Hameed, M.I. El-Khaiary, Malachite green adsorption by rattan sawdust:
Isotherm, kinetic and mechanism modeling, J. Hazard. Mater. 159 (2008) 574-9. [43]
B. Tawabini, S. Al-Khaldi, M. Atieh, M. Khaled, Removal of mercury from
water by multi-walled carbon nanotubes, Water Sci. Technol. 61 2010 591-8. [44]
J. Dong, Z.H. Xu, F. Wang, Engineering and characterization of mesoporous
silica-coated magnetic particles for mercury removal from industrial effluents, Appl Surf. Sci. 254 (2008) 3522-30. [45]
G.Z. Kyzas, E.A. Deliyanni, K.A. Matis, Graphene oxide and its application as
an adsorbent for wastewater treatment, J. Chem. Technol. Biot. 89 (2014) 196-205. [46]
J. Bao, Y. Fu, Z. Bao, Thiol-functionalized magnetite/graphene oxide hybrid as a
reusable adsorbent for Hg2+ removal, Nanoscale Res. Lett. 8 (2013) 486-92. [47]
R.G. Pearson, Absolute electronegativity and hardness – application to
inorganic-chemistry, Inorg Chem. 27 (1988) 734-40. [48]
V. Gomez-Serrano, A. Macias-Garcia, A. Espinosa-Mansilla, C. Valenzuela-
Calahorro, Adsorption of mercury, cadmium and lead from aqueous solution on heattreated and sulphurized activated carbon, Water Res. 32 (1998) 1-4. [49]
L.S. Rocha, C.B. Lopes, B. Henriques, D.S. Tavares, J. Borges, A.C. Duarte, et
al. Competitive effects on mercury removal by an agricultural waste: application to synthetic and natural spiked waters. Environ. Technol. 35 (2014) 661-73. [50]
J. Wang, B. Deng, H. Chen, X. Wang, J. Zheng Removal of Aqueous Hg(II) by
Polyaniline: Sorption Characteristics and Mechanisms, Environ Sci. Technol. 43 (2009) 5223-8. 24
Figure 1. Digital image of 3DGON foam a); SEM images of 3DGON internal microstructure at different magnifications b) and c). XPS survey spectra of 3DGON d) with high resolution C1s e) and N 1s f). XPS spectra with deconvoluted peaks
25
Figure 2: Variation of the normalized concentration of Hg(II) in solution (Ct/C0) during the contact time (h) with 10 mg L-1 of 3DGO (■), 3DGON (●) and 3DGOSN (∆). Initial concentration of Hg(II) in ultra-pure water: 50 µg L-1. The dotted line correspond to the control (Hg(II) spiked ultra-pure water in the absence of sorbent).
26
Figure 3. Variation in the amount of Hg(II) removed (R (%)) for different 3DGON dosage used (mg L-1), after 24 and 48 hours of contact time. Initial concentration of Hg(II) in ultra-pure water: 50 µg L-1.
27
Figure 4: Variation in the amount of Hg(II) removed (R (%)) for different initial solution pH, after 3, 12 and 24h of contact time. Initial concentration of Hg(II) in ultra-pure water: 50 µg L-1.
28
Figure 5. Kinetic modelling of the experimental data obtained from the sorption process of Hg(II) onto 3DGON. a) using three kinetic reaction models: pseudo-first order (…), pseudo-second order (▬) and Elovich (---). b) using Webber’s intraparticle-diffusion model.
29
Figure 6. XPS spectra for 3DGON surfaces before and after Hg(II) sorption (a). Highresolution spectra of Hg 4f (b) and N 1s (before (c) and after Hg sorption(d)).
30
Figure 7. Equilibrium data (black squares) and sorption isotherms for the Hg(II)-3DGON system (21±1ºC). Initial concentration of Hg(II) in MQ water 50 µg L-1 and 1.0, 2.5, 5.0 10.0 and 20.0 mg L-1 of 3DGON.
31
Figure 8. Variation of the normalized concentrations of Hg(II) (■), Pb(II) (●) and Cd(II) (∆) in solution (Ct/C0) during the contact time (h) with 10 mg L-1 of 3DGON. Initial concentration of each metal in the mixture: 50 µg L-1. The dotted line correspond to the Hg(II) control (mixture in absence of sorbent).
32
Figure 9. Variation of the normalized concentrations (Ct/C0) of Hg(II) in spiked ultrapure water (■), river water (○) and seawater (▲) during the contact time (h) with 10 mg L-1 of 3DGON. Initial concentration of Hg(II): 50 µg L-1. The black and the dotted lines correspond to the controls (spiked solutions in absence of sorbent) of river water and seawater, respectively.
33
Table 1. Kinetic parameters obtained by fitting the experimental data on Hg(II) uptake by 3DGON to the following reaction models: Pseudo-first order, Pseudo-second order and Elovich. Experimental qe and the goodness of fit were also presented in order to assess and compare models. Pseudo-first order
Best fit values qe exp
qe1 ±
Initial [Hg]
± SD
SD
µg L-1
(µg g1 )
(µg g1 )
(h-1)
50
4515± 35
4173 ± 234
0.619 ±0.13 7
Pseudo-second order
Goodness of fit
qe2 ±
k1 ± SD
Best fit values
SD 2
R
0.9 17 6
Sy.x
464
Goodness of fit
(µg g1 )
(h-1)
4528 ± 170
0.000 2±0.0 0004
Best fit values
β ± SD
k2 ± SD
Elovich
2
R
0.9 72 2
Sy.x
270
(g µg1 ) 0.0014 ±0.000 1
Goodnes s of fit
α± SD R2
Sy. x
0.9 81 9
21 7
(µg g1 -1 h ) 16010 ±4489
34
Table 2. Kinetic parameters obtained by fitting the experimental data on Hg(II) uptake by 3DGON (10 mg L-1) to the following diffusion models: Boyd’s film-diffusion and Webber’ s intraparticle-diffusion. Boyd’s film-diffusion
Webber’ s intraparticle-diffusion
Stag e
intercept
R2
Breakpoint (h)
Ki (µg g-1 h-1/2)
R2
0.03478 1st
[0.02865;0.09839 ]
0.9941 (n=6)
7.89
1225.185
0.9936 (n=6)
2nd
-
-
-
79.79848
0.8026 (n=3)
35
Table 3. Isotherm constants obtained from fitting the equilibrium data for Hg(II) sorption onto 3DGON to Langmuir, Freundlich and SIPS isotherm models. Langmuir Isotherm
Best fit values qm ± SD
Goodness of fit
bL ± SD
g-
)
(L µg1 )
35022 ± 1946
0.039 ± 0.004
Best fit values KF ± SD
R2 (µg 1
Freundlich Isotherm
Sy.x (µg1-n Ln/g)
0.9 98 4
328
2067 ± 204
1/nF ± SD
0.65 ± 0.03
Goodness of fit
R
2
0.9 961
SIPS Isotherm
Best fit values
Goodnes s of fit
qm ± Sy.x SD
1/n ± SD
bS ± SD
R2
Sy. x
449 66± 105 78
0.88 ± 0.10
0.02 3± 0.01 1
0.9 99 1
29 3
516
36