Organic Solvent-Free Cloud Point Extraction-like Methodology

Because of its unique properties and capability of formation of well-dispersed aqueous colloids in aqueous phase, graphene oxide can be used for the e...
0 downloads 5 Views 3MB Size
Download PDF
Article pubs.acs.org/ac

Organic Solvent-Free Cloud Point Extraction-like Methodology Using Aggregation of Graphene Oxide Dongyan Deng,† Xiaoming Jiang,‡ Lu Yang,§ Xiandeng Hou,†,‡ and Chengbin Zheng*,† †

Key Laboratory of Green Chemistry & Technology, Ministry of Education, College of Chemistry, Sichuan University, Chengdu, Sichuan 610064, China ‡ Analytical & Testing Center, Sichuan University, Chengdu, Sichuan 610064, China § Chemical Metrology, Measurement Science and Standards, National Research Council Canada, Ottawa, Ontario K1A 0R6, Canada S Supporting Information *

ABSTRACT: Because of its unique properties and capability of formation of well-dispersed aqueous colloids in aqueous phase, graphene oxide can be used for the efficient preconcentration of heavy metal ions prior to their determination. The complete collection of graphene oxide colloids from water has generally been considered to be insurmountable. Here, graphene oxide aggregation triggered by introducing NaCl was used to develop a novel organic solvent-free cloud point extraction-like method for the determination of trace toxic metals. The graphene oxide sheets were uniformly dispersed in aqueous samples or standard solutions for a fast and efficient adsorption of Pb(II), Cd(II), Bi(III), and Sb(III) owing to its hydrophilic character and the electrostatic repulsion among the graphene oxide sheets, and its aggregation immediately occurred when the electrostatic repulsion was eliminated via adding NaCl to neutralize the excessive negative charges on the surface of graphene oxide sheets. The aggregates of graphene oxide and analytes ions were separated and treated with hydrochloric acid to form a slurry solution. The slurry solution was pumped to mix with KBH4 solution to generate hydrides, which were subsequently separated from the liquid phase and directed to an atomic fluorescence spectrometer or directly introduced to an inductively coupled plasma optical emission spectrometer for detection. On the basis of a 50 mL sample volume, the limits of detection of 0.01, 0.002, 0.01, and 0.006 ng mL−1 were obtained for Pb, Cd, Bi, and Sb, respectively, when using atomic fluorescence spectrometry, providing 35-, 8-, 36-, and 37-fold improvements over the conventional method. Detection limits of 0.6, 0.15, 0.1, and 1.0 ng mL−1 were obtained with the use of slurry sampling inductively coupled plasma optical emission spectrometry. The method was applied for analysis of two Certified Reference Materials and three water samples for these elements.

H

Graphene, processing two-dimensional carbon nanostructure with large surface area (2630 m2/g) and excellent physicochemical properties, holds great potential in analytical chemistry for applications such as preconcentration and biological and electronic sensors.11,12 Like other nanomaterials, graphene is usually coated on a substrate or loaded into a column/cartridge for solid phase extraction/microextraction (SPE/SPME) considering its toxicity and the environmental risk of nanomaterials.13−15 However, these immobilized techniques may result in higher back-pressure in the packing process (ca. 52 MPa), lower extraction efficiency, and longer extraction time to achieve adsorption equilibrium.16 A dispersive solid liquid extraction method by direct addition of graphene to an aqueous phase has been applied to solve these problems.17 Direct dispersion of hydrophobic graphene sheets in water has generally been considered to be an insurmountable

eavy metals such as Pb, Cd, Bi, and Sb can have serious toxic effects on human and animals even at trace levels due to biomagnification of hundreds and thousands of times in human and animal bodies via the food chain, leading to damage of the brain and central nervous system.1 Despite modern analytical instrumentation with adequate sensitivities is available, the determination of these elements in a complex matrix can be complicated due to their low concentrations and potential concomitant matrix effects. Thus, numerous protocols for matrix separation and analyte preconcentration are desired to circumvent these obstacles.2−5 Among these, cloud point extraction based on nonionic surfactants aggregation via an increase in temperature or by adding a “salting out” agent is considered to be an efficient and environmentally benign extraction technique for separation and preconcentration of analyte from bulky matrix.6−10 However, consumption of organic reagents and a tedious operation process still remain in this technique, which could result in potential pollution to the environment and low sample throughput. © XXXX American Chemical Society

Received: October 8, 2013 Accepted: November 20, 2013

A

dx.doi.org/10.1021/ac403345s | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

emission spectrometer (ICP-OES) (ARCOS FHS12, SPECTRO Analytical Instruments GmbH, Germany) was also used to evaluate this preconcentration method. The operating conditions for AFS and ICP-OES were summarized in Table S1 (see Section 1 of the Supporting Information). GO and its aggregates were characterized by a transmission electron microscope (TEM, Tecnai G2 F20 S-TWIN, FEI Company, USA), a Fourier transform infrared spectrometer (FT-IR, IS10, Thermo Nicolet, USA), and a UV−vis absorption spectrometer (UV−vis, UV-1750, Shimadzu, Japan), respectively. The Zeta potential measurement was carried out on a Zeta Nano Series (Malvern Instruments Led). A vortex mixer (XH-B, Jiangsu Kangjian Medical Apparatus Co., Ltd., China) was used to disperse GO evenly in standard or sample solutions. The aggregate of GO and analyte ions was easily deposited to the bottom of the centrifuge tube using a high speed centrifuge (HC-3018, Anhui USTC Zokia Scientific Instruments Co., Ltd., China). A benchtop ultrasonic cleaner (KQ-300DE, Kunshan Ultrasonic Instruments Co., Ltd., China) was applied to prepare GO dispersion. Reagents and Materials. All reagents used in this work were analytical grade. All solutions were prepared using 18.2 MΩ-cm deionized water (DIW) produced by a water purification system (Chengdu Ultrapure Technology Co., Ltd., China). Standard solutions were prepared daily by serial dilutions of the stock solutions (1000 mg L−1) of Pb(II), Cd(II), Bi(III), and Sb(III) purchased from National Research Center of China (NRCC, Beijing, China). Hydrochloric acid and sodium chloride were purchased from Kelong Chemical Factory (Chengdu, China). Graphite powder (99.85%, Kelong Chemical Factory, Chengdu, China) was used to synthesize graphene oxide. Two Certified Reference Materials (CRMs, rice powder GBW08510 and simulated natural water GBW08608) from National Research Center for Standard Materials (Beijing, China), a high salt water sample (containing 3.5% NaCl), and two real water samples collected from a local river and a lake (Chengdu, China) were analyzed for these elements by this method. GO was synthesized according to the modified Hummers method.29,30 Briefly, 3 g of graphite powder was added into a 1500 mL round-bottom flask and reacted with a mixture of concentrated H2SO4 (12 mL), K2S2O8 (2.5 g), and P2O5 (2.5 g) at 80 °C for 5 h under stirring, and then, 500 mL of DIW was added and allowed to stay overnight. Preoxidized gaphite was obtained after centrifugation of the mixture at 5000 rpm for 15 min and then added into cold concentrated H2SO4 (120 mL). Meanwhile, KMnO4 (18 g) was gradually added with stirring. Afterward, H2O2 (30 mL, 30%) and DIW were added in the suspension to eliminate the excess MnO4− and terminate the reaction. The resultant solution was filtered and dialyzed in DIW for about 1 month to remove the remaining impurities of inorganic acid and metals. Finally, the product was dried at 60 °C under vacuum. The obtained GO was used to prepare a stable GO suspension by ultrasonication for 2 h in DIW prior to use. In order to avoid the tedious and time-consuming GO synthesis and simplify routine analysis, a commercial GO powder purchased from Chengdu Organic Chemicals Co. Ltd. (Chengdu, China) was used as an alternative to the synthesized one, and similar results were obtained. Sample Preparation. Water samples were obtained from the shore by sampling aliquots into precleaned quartz bottles and directly transported to the laboratory for immediate study after filtration with 0.22 μm membrane filters. The simulated

challenge because of its high specific surface area and very strong van der Waals interactions of layer to layer. This could lead to irreversible agglomerates or even restack to form graphite and thus interrupt the further evolution of graphene for analytes preconcentration. Over the past decade, many efforts have been undertaken in order to obtain stable and uniform dispersions of graphene in water by means of chemical oxidation,18−20 using surface-active agents21 or polar solvents22 as solubilizing or dispersing agents. Chemical oxidation of graphite with Hummers method to produce graphene oxide (GO) has been proved to be a relatively mild, simple, and efficient strategy.23,24 In contrast to graphene, GO can form well-dispersed aqueous colloids in aqueous phase because of its polar and hydrophilic character arising from a number of reactive oxygen functional groups including epoxide, hydroxyl, and carboxylic acid present on the basal planes and edges of GO. Moreover, the large amounts of negative charges from ionization of these reactive oxygen functional groups make GO more favorable to adsorb cations through the electrostatic attraction as it is dispersed in water. Therefore, GO has been applied as a superior sorbent to preconcentrate heavy metal ions.25−27 Unfortunately, it is difficult to completely collect the miniscule GO sheets from a well-dispersed solution even by high-speed centrifugation.14 Instead, 0.22 μm membrane filters are usually used to separate solid phase of GO from the solution, which is still incomplete, relatively cumbersome, and time-consuming.27 Fortunately, Wallace et al.24 found that chemically converted graphene (CCG) from reduction of GO with N2H4 could form stable aqueous colloids and then aggregate when an electrolyte solution such as sodium chloride was added or the pH of solution was decreased. On the basis of our preliminary colloid experiments, we believed that GO could form immediate aggregation by changing pH or adding electrolyte solution, similar to that conventional CPE method and, thus, it can be used for efficient preconcentration of metal ions. This hypothesis is supported through the findings by Wang et al.,28 that a high concentration of Cu2+ caused GO to be folded to form a large aggregate, and our feasibility study. Therefore, the aggregation of graphene oxide (GO) from well-dispersed GO aqueous colloids triggered by introducing NaCl was investigated in the current work. To our knowledge, this is the first report that utilized GO aggregation as an effective cloud point extraction (CPE)-like but organic reagentfree technique for the preconcentration of toxic elemental ions. The isolated aggregates of GO and the ions were treated with hydrochloric acid to form slurry solution and subsequently pumped to mix with KBH4 solution for the generation of hydrides prior to AFS or ICP-OES detection. Taking advantage of the superior properties of GO and the slurry hydride generation/slurry sample introduction, the proposed method not only provides the advantages of CPE but also has excellent extraction efficiency and high sample throughput, is free of organic chemicals, and eliminates tedious elution steps in the conventional methods.

■

EXPERIMENTAL SECTION Instrumentation. A commercial four channel hydride generation nondispersive atomic fluorescence spectrometer (Model AFS-9600, Beijing Haiguang Instrument Co., Beijing, China) equipped with a quartz argon−hydrogen flame atomizer, a quartz gas−liquid separator (GLS), and coded high intensity hollow cathode lamps of Pb, Cd, Bi, and Sb were used for AFS detection. An inductively coupled plasma optical B

dx.doi.org/10.1021/ac403345s | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

nm and a shoulder peak at around 300 nm were the absorption bands corresponding to the π→π* electron transitions of polyaromatic CC bonds and the n→π* electron transitions of CO bonds, respectively.31 The results obtained agree well with those reported in previous studies.32 Clearly, both FT-IR and UV−vis spectra provided the evidence of the presence of large amounts of oxygen functional groups (hydroxyl, carboxyl, carbonyl, and epoxy groups) on the surface of the prepared GO, which resulted in the formation of well-dispersed GO aqueous suspensions and the adsorption of analyte ions. Proposed Mechanism. Wallace24 found that CCG sheets were highly negatively charged, arising from the ionization of its carboxylic acid and phenolic hydroxyl groups. They suggested that the formation of stable CCG colloids should be attributed to electrostatic repulsion, rather than just the hydrophilicity of GO as previously presumed.33 On the basis of this information, we speculate that adsorption of analyte ions and dispersion and aggregation of GO may also depend on the degree of the ionization of its carboxylic acid and phenolic hydroxyl groups. It is well-known that aqueous colloids/dispersions are stabilized via electrostatic repulsions, which are pH dependent.24,34 Therefore, the effect of pH on the Zeta potential of 0.1 mg mL−1 GO was investigated, as shown in Figure 2a. The results indicate that the GO nanosheets are highly negatively charged at pH 2−10, which is consistent with the expected fact that the ionization of carboxylic acid and phenolic hydroxyl groups exist on the GO surface. These highly negative charges at this pH range resulted in well-dispersed GO through the strong electrostatic repulsion. Although the GO aggregation occurred and phase separation was observed as the pH decreased to lower than 2, analytes cannot be efficiently adsorbed under this condition. This will be further addressed later. Therefore, NaCl as an alternative for HC1 was chosen to neutralize the excessive negative charges and decrease the electrostatic repulsion to subsequently accomplish the purpose of GO aggregation. The effect of the concentration of NaCl on the Zeta potential was thus evaluated. The results showed that the Zeta potential of 0.1 mg mL−1 GO gradually decreased from −44.2 to −20.6 mV when the concentration of NaCl increased to 5.75 mg mL−1. A previous study found that the instable dispersion of colloid caused by insufficient mutual repulsion would appear when the absolute Zeta potential values were less than 20 mV.35 Figure 2b shows the Zeta potentials of 0.1 mg mL−1 GO measured at different conditions. The results confirm that the change of the Zeta potentials due to the introduction of trace level metal ions (1−1000 μg L−1) can be negligible and thus fulfill the demand of routine analysis by adding sufficient quantities of NaCl to decrease the Zeta potential values and thus to achieve GO aggregation and phase separation. The TEM images of GO sheets summarized in Figure 3 further illuminated the change of morphologies of GO sheets as a result of addition of trace level metal ions (e.g., Pb2+) and/or NaCl. As shown in Figure 3a,b, a large GO sheet with intrinsic wrinkles was observed, similar to those prepared by the chemical exfoliation method. Furthermore, Figure 3c shows that GO sheets were aggregated together, confirming that neutralization of GO negative charges leads to the GO aggregation in the aqueous GO suspension by introducing NaCl. In order to visually demonstrate the phenomenon of GO aggregation, the stabilities of GO aqueous suspensions under different conditions were studied, shown in Figure 4. The GO sheets can be stably and homogeneously dispersed even after

water sample GBW08608 was simply diluted before analysis. For determination of Cd in rice powder GBW08510, 0.1 g of the rice powder was weighed into precleaned Teflon vessels. Eight mL of HNO3 and 2 mL of H2O2 were then added. Sample blanks were processed along with the sample. The sealed vessels were heated in a microwave oven (Master 40, Shanghai Sineo Microwave Chemistry Technology Co., China) operated under the following conditions: 15 min at 130 °C and 2200 W; 20 min at 150 °C and 2200 W; and 25 min at 180 °C and 2200 W. After cooling, the caps were removed and the digests were transferred to precleaned 50 mL volume Teflon tubes and evaporated on an electrical hot plate to near dryness; Then, 50 mL of DIW was added to redissolve the analytes. Finally, the samples were transferred into a 50 mL polyethylene tube and stored in the dark at 4 °C prior to use. Extraction and Analysis Procedure. For standard solution and lake and river water sample solutions, 47.5 mL of these solutions with different amounts of Pb(II), Cd(II), Bi(III), and Sb(III) was added to 50 mL volume polyethylene tubes and adjusted to pH 6.0 using NaOH and HCl prior to extraction. For CRMs, 5 mL of the rice sample solution or 10 mL of the diluted natural water was diluted to 47.5 mL using DIW, NaOH, and HCl to adjust the pH to 6.0 prior to extraction. After that, 2.5 mL of a high concentration of GO dispersion solution (2 mg mL−1) was then added to the tubes, and the tubes were allowed to stand for 5 min to achieve complete adsorption of the analyte ions. GO was quickly aggregated and gradually deposited onto the bottom of the tubes when NaCl was mixed with the solution to a final concentration of 5.75 mg mL−1. The complete phase separation was accomplished by either standing for 48 h or centrifuging for 5 min at 10 000 rpm. The whole preconcentration process is schematically described in Figure 1. The supernatant was

Figure 1. Schematic diagram of the whole preconcentration process.

discarded, and 3% HCl (v/v) was added to dilute and acidify the residual GO aggregates to 1 mL of slurry. The slurry was directly transported to ICP-OES for detection or pumped to mix with KBH4 solution for the generation of the hydride, which was swept into AFS for detection.

■

RESULTS AND DISCUSSION Characterization of Graphene Oxide. The chemical composition of the synthesized GO was initially characterized by FT-IR and UV−vis. The FT-IR spectra (see the Supporting Information for Figure S1a) revealed that the characteristic bands of CC at 1623 cm−1 and the broad and intense band of O−H in GO appeared at 3408 cm−1; the CO band stretching vibrations from carbonyl and carboxylic groups of GO appeared at around 1730 cm−1; and the bands around 1280 and 1090 cm−1 are attributed to C−OH and C−O stretching vibrations, respectively. According to the UV−vis spectra of the stable GO dispersion (Figure S1b in the Supporting Information), the maximum absorption peak at about 230 C

dx.doi.org/10.1021/ac403345s | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

Figure 2. Zeta potential measurement of GO suspensions: (a) Zeta potential of GO at 0.1 mg mL−1 as a function of pH and (b) Zeta potential values after addition of NaCl (5.75 mg mL−1) into GO suspensions (0.1 mg mL−1) and suspensions with Pb ions (Pb′ refers to 1 ng mL−1 and Pb″ refers to 1000 ng mL−1), at pH 6.

Figure 3. TEM images of GO suspensions (a), GO+Pb(II) (b), and GO+Pb(II)+NaCl (c).

Figure 4. Aggregation of GO (left), GO+Pb(II) (middle), and GO+Pb(II)+NaCl (right) after 0 h (a), 2 h (b), 5 h (c), 17 h (d), 24 h (e), and 48 h (f).

groups is significantly affected by the pH. Moreover, the initial experiment showed that the stable GO suspensions were aggregated when the pH of solution was lower than 2. Since tested analyte ions form hydroxides at pH > 9.0, the effect of pH on the extraction efficiencies of analyte ions was

standing for 48 h without NaCl. The GO aggregation can be immediately observed and completely deposited to the bottom of bottles when NaCl was added. Optimization of Extraction Conditions. As mentioned above, the ionization of carboxylic acid and phenolic hydroxyl D

dx.doi.org/10.1021/ac403345s | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

when the pH is lower than the IEP of GO, thus making it difficult to absorb positive metal ions. It is interesting that the extraction efficiencies of Bi(III) and Sb(III) are still stable when pH value is lower than IEP of GO, whereas the surface of GO is positively charged. This phenomenon is inconsistent with electrostatic interaction, which is presumably due to the restoration of a graphitic network of sp2 bonds.25 There are enough functional groups on GO surfaces that can form surface complexes with Bi(III) and Sb(III) and result in these stable extraction efficiencies, which agree well with the subsequent experiment that shows Bi(III) and Sb(III) retaining high saturation of adsorptive capacities on GO. To eliminate tedious pH adjustments to maintain simplicity and maximum extraction efficiencies, pH at 6.0 (pH value of normal natural water) was thus selected for all subsequent experiments. The adsorption kinetics of Pb(II), Cd(II), Bi(III), and Sb(III) onto GO were investigated at pH 6.0. The obtained results reveal that adsorption equilibriums were achieved very quickly as fast as 5 min, which is much faster than conventional extraction techniques. The fast adsorption kinetics is likely due to the highly attainable surface area of GO and the good dispersion of GO in aqueous solution. The fast kinetics is important for practical applications of GO for adsorption of Pb(II), Cd(II), Bi(III), and Sb(III) prior to their determination, as this can remarkably reduce the analysis time. On the basis of these observations, an adsorption time of 5 min was chosen for subsequent experiments in order to have high sample throughput. The adsorption capacities of the prepared GO for Pb(II), Cd(II), Bi(III), and Sb(III) were evaluated via adding 2.5 mL of concentrated GO dispersion solution (2 mg mL−1) to a series of 47.5 mL solutions containing various concentration of these tested metal ions, respectively. Saturation of adsorptive capacities of GO was observed at 149, 42, 1170, and 78.3 mg g−1 for Pb(II), Cd(II), Bi(III), and Sb(III), respectively, suitable for routine analysis. The effect of NaCl concentration on extraction efficiency was investigated using 1 μg mL−1 of Pb(II) as a test element under optimized experimental conditions (also see the Supporting Information for Figure S2). Extraction efficiency dramatically increased when NaCl concentration increased from 4.7 to 5.75 mg mL−1 and decreased at higher concentrations of NaCl.

investigated in a range of pH 2.0−9.0. The extraction efficiency was estimated from a comparison of the relative concentrations of analytes in the tested standard solution and its supernatant after the GO adsorption. As shown in Figure 5, the extraction

Figure 5. Sorption of Pb(II), Cd(II), Bi(III), and Sb(III) on GO nanosheets as a function of pH. Experimental conditions: 1 μg mL−1, Pb(II), Cd(II), Bi(III), and Sb(III); T = 318 K; t = 5 min; and 0.1 mg mL−1 GO.

efficiencies of Cd(II) and Pb(II) increased significantly with increasing pH of sample solution from 2.0 to 4.0, followed by a plateau at the pH range of 4.0−9.0. No significant variation on the extraction efficiencies of Bi(III) and Sb(III) was observed over the pH range studied (pH 2.0−9.0). Although the exact mechanism of adsorption of analyte ions onto GO is not clear currently, this phenomenon may be realized via electrostatic attraction between analyte ions and surface charges of GO, which is similar to that demonstrated earlier using inorganic solid nanosorbent including TiO2, SiO2, or other nanomaterials to adsorb metal ions. As the pH of sample solution becomes higher than the isoelectric point (IEP) value of GO (3.9),27 the surface of GO appears negatively charged and is favorable to adsorb metal ions. However, the surface of GO is protonated because the ionization of the functional groups is prohibited

Figure 6. Effects of different concentrations of HCl and KBH4 on the signal of HG-AFS: (a) HCl and (b) KBH4. Experimental conditions: 2.5 μg L−1, Pb(II), Bi(III), and Sb(III); 0.25 μg L−1, Cd(II); T = 318 K; t = 5 min; 0.1 mg mL−1, GO; and 5.75 mg mL−1, NaCl. E

dx.doi.org/10.1021/ac403345s | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

Table 1. Analytical Results of Pb, Cd, Bi, and Sb in Spiked Mimic Sea Water and Two Natural Water Samples by Using the CPElike Extraction Method added (μg L−1) sample

Pb

Cd

Bi

Sb

mimic seawater

0 0.5 1.0 0 0.5 1.0 0 0.5 1.0

0 0.01 0.05 0 0.01 0.05 0 0.01 0.05

0 0.5 1.0 0 0.5 1.0 0 0.5 1.0

0 0.5 1.0 0 0.5 1.0 0 0.5 1.0

lake water

river water

a

found (μg L−1)a Pb 0.10 0.62 1.13 0.068 0.590 0.960 1.97 2.49 2.94

± ± ± ± ± ± ± ± ±

Cd 0.03 0.05 0.12 0.006 0.006 0.023 0.09 0.02 0.08

0.049 0.059 0.105 0.028 0.038 0.083 0.046 0.057 0.095

± ± ± ± ± ± ± ± ±

recovery (%) Bi

0.004 0.008 0.010 0.006 0.001 0.009 0.003 0.003 0.007

Sb

0.53 ± 0.02 1.00 ± 0.03 0.47 1.06 0.07 0.61 1.05

± ± ± ± ±

0.01 0.02 0.02 0.03 0.04

0.52 0.94 0.12 0.61 1.02 0.083 0.507 0.943

± ± ± ± ± ± ± ±

0.04 0.05 0.01 0.01 0.04 0.009 0.011 0.011

Pb

Cd

Bi

Sb

104 103

100 112

106 94

104 94

104 89

100 90

98 106

98 90

104 97

110 98

108 98

85 86

Mean and standard deviation (n = 3).

analyte hydrides before they can be separated from the liquid phase.38 Moreover, the coexisting ions may affect the adsorption efficiencies of analytes because of the competitive adsorption between the coexisting ions and tested ions. The effects of 12 concomitant ions on the AFS responses from 0.001 μg mL−1 Pb(II), Cd(II), Bi(III), and Sb(III) standard solutions were tested. With tolerance limits defined as the largest concentrations of coexisting ions resulting in less than 10% signal variation, no significant interferences from 0.1 mg L−1 Fe3+, Ag+, and Hg2+; 0.5 mg L−1 Cu2+, Mg2+, and Cr3+; 1 mg L−1 Ca2+ and Co2+; 2 mg L−1 Zn2+ and Ni2+ were observed. Moreover, no significant interferences from 4 mg L−1 of Na+ or K+ were observed, suggesting the proposed method is capable of analyzing samples with high content of salts. Analytical Performance. Analytical figures of merit obtained using slurry sampling HG-AFS or ICP-OES under optimal experimental conditions were evaluated. The linear correlation coefficient for each calibration curve was better than 0.998. The limit of detection (LOD) was defined as the analyte concentration equivalent to three times of the standard deviation of 11 measurements of the blank solution. For the slurry HG-AFS system, the LODs were 0.01, 0.002, 0.01, and 0.006 ng mL−1 for Pb, Cd, Bi, and Sb using a 50 mL sampling volume, respectively, with significant improvements of 35-, 8-, 36-, and 37-fold compared to those obtained from the conventional HG system without preconcentration. Precisions of replicate measurements, expressed as relative standard deviations (RSDs, n = 11), are better than 4.2% with analyte ions at a concentration of 1 ng mL−1. Detection of limits of 0.6, 0.15, 0.1, and 1.0 ng mL−1 were obtained for Pb, Cd, Bi, and Sb with use of slurry sampling inductively coupled plasma optical emission spectrometry and successively improved 30-, 10-, 33-, and 31-fold, compared to those arising from conventional solution nebulization. It should be noted that the concentrations of standard solutions used to establish the calibration curve of Cd is 10-fold lower than those of other elements because of the much higher sensitivity of Cd over other tested elemental ions using HG-AFS or ICP-OES. Therefore, such low concentration in aqueous solution might influence its extraction efficiency. Meanwhile, the relative high blank to such low content of Cd also resulted in high deviation and limited its LOD improvement. Preliminary Analytical Application. The accuracy of the proposed method was tested by analysis of several water samples including river water and lake water. The water samples were collected from the shore and immediately filtered through a 0.22 μm membrane filter prior to their preconcentra-

Lower NaCl concentration resulted in inefficient GO aggregation and low efficiency of extraction of analyte from aqueous phase. Whereas higher NaCl concentrations resulted in significant elution of analyte from GO because of competition between positive sodium ions and analyte ions. A NaCl concentration of 5.75 mg mL−1 was selected for all subsequent experiments. In addition, this concentration of NaCl was found to be suitable for the extraction of Cd(II), Bi(III), and Sb(III) with good phase separation and quantitative extraction efficiencies. Optimization of the Slurry HG-AFS Determination. The overall efficiency of HG defined as the convolution of the efficiency of species formation, gas−liquid separation, and transportation to atomization was affected by many factors including the nature and concentration of acid, the concentration of KBH4, and carrier gas flow rate. According to previous works,36,37 1.0% (m/v) potassium ferricyanide (K3Fe(CN)6) and a mixture containing 2.0% (m/v) thiourea and ascorbic acid were selected as enhancing reagents for HG of Pb(II) and Sb(III), respectively. An eluent was required to release the analytes from GO aggregation prior to determination by slurry HG-AFS or ICPOES. In this work, HCl solution was chosen as an eluent to decrease the pH of slurry over the IEP of GO for elution of analytes since it facilitates both desorption of analytes and their subsequent HG. A series of HCl concentrations were used to investigate the effect on the AFS responses of Pb(II), Cd(II), Bi(III), and Sb(III), shown in Figure 6a. The responses from all the tested elements increased significantly in the range of 0−3% (v/v) HCl, reached a plateau at the range of 3−5% (v/v), and decreased at higher concentrations. Low HCl concentration led to inefficient elution and low HG efficiencies; high concentration resulted in acute reaction and produced copious bubbles, leading to difficulty in mixing of the reactant solutions. A concentration of 3% (v/v) was thus selected for the subsequent measurements. The effect of the concentration of KBH4 on responses was also investigated, as shown in Figure 6b. High concentration of KBH4 resulted in dilution of hydrides because of the generation of a large amount of hydrogen gas, and 2% (m/v) KBH4 was thus selected for HG of Pb(II), Cd(II), Bi(III), and Sb(III) to have high HG efficiencies for all analytes. Effect of Coexisting Ions. The major shortcoming associated with conventional hydride generation systems is the serious interferences arising from transition and noble metal ions due to these ions being very easily reduced to their metallic states or colloidal forms which then scavenge or decompose F

dx.doi.org/10.1021/ac403345s | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

■

Table 2. Determination of Cd and Pb in Certified Reference Materials determineda (μg L−1) sample rice powder GBW08510b simulated natural water GBW08608 a

Pb

Pb

1.976 ± 0.105 52 ± 4

10.3 ± 0.9

*E-mail: [email protected] (C. B. Zheng). Author Contributions

Cd

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

2.062 ± 0.052 50 ± 1

10.0 ± 0.4

Notes

The authors declare no competing financial interest.

■

−1

Mean and standard deviation (n = 3). μg g . b

ACKNOWLEDGMENTS The authors gratefully acknowledge the National Nature Science Foundation of China (Nos. 21075085 and 21128006) for financial support. C. B. Zheng is grateful for the financial support by Ministry of Education of China through the Grant NCET-11-0361. The authors also acknowledge parts of this work were reported in the 5th Asia Pacific Winter Conference on Plasma Spectrochemistry, Jeju Island, South Korea, August, 2012.

tion. A water containing 3.5% of NaCl (m/v) was utilized to evaluate the usefulness of this method for analysis of samples containing high salts. However, direct determination of analytes in this high salt containing sample was impossible because the aggregation of GO was immediately observed and decreased the extraction efficiency when GO was added. Moreover, a high content of salt also resulted in low extraction efficiencies of analytes because of the competition between positive sodium ions and analyte ions. Fortunately, simple dilution can be used to overcome this problem due to its ultrahigh sensitivity. It is worthwhile to note that no additional NaCl was needed to trigger the GO aggregate in this case. The results are summarized in Table 1. Good spike recoveries of analytes (85−112%) for these samples were achieved, confirming accuracy of the proposed method for high salt containing water. The utility of the proposed technique was further demonstrated by the determination of two representative elements, Cd and Pb, in rice powder GBW08510 and simulated natural water GBW08608. Analytical results are summarized in Table 2. The t test showed that the analytical results obtained by the proposed method were not significantly different from the certified values at the confidence level of 95%.

■

REFERENCES

(1) Potocki, S.; Rowinska-Zyrek, M.; Witkowska, D.; Pyrkosz, M.; Szebesczyk, A.; Krzywoszynska, K.; Kozlowski, H. Curr. Med. Chem. 2012, 19, 2738−2759. (2) Yan, X. P.; Kerrich, R.; Hendry, M. J. Anal. Chem. 1998, 70, 4736−4742. (3) Rao, T. P.; Daniel, S.; Gladis, J. M. TrAC, Trends Anal. Chem. 2004, 23, 28−35. (4) Sohrin, Y.; Urushihara, S.; Nakatsuka, S.; Kono, T.; Higo, E.; Minami, T.; Norisuye, K.; Umetani, S. Anal. Chem. 2008, 80, 6267− 6273. (5) Yousefi, S. R.; Shemirani, F. Anal. Chim. Acta 2010, 669, 25−31. (6) Casero, I.; Sicilia, D.; Rubio, S.; Perez-Bendito, D. Anal. Chem. 1999, 71, 4519−4526. (7) Quina, F. H.; Hinze, W. L. Ind. Eng. Chem. Res. 1999, 38, 4150− 4168. (8) Stalikas, C. D. TrAC, Trends Anal. Chem. 2002, 21, 343−355. (9) Bosch Ojeda, C.; Sanchez Rojas, F. Anal. Bioanal. Chem. 2009, 394, 759−782. (10) Majedi, S. M.; Lee, H. K.; Kelly, B. C. Anal. Chem. 2012, 84, 6546−6552. (11) Stoller, M. D.; Park, S.; Zhu, Y.; An, J.; Ruoff, R. S. Nano Lett. 2008, 8, 3498−3502. (12) Guo, S. J.; Dong, S. J. Chem. Soc. Rev. 2011, 40, 2644−2672. (13) Liu, Q.; Shi, J.; Zeng, L.; Wang, T.; Cai, Y.; Jiang, G. J. Chromatogr., A 2011, 1218, 197−204. (14) Liu, Q.; Shi, J.; Sun, J.; Thanh, W.; Zeng, L.; Jiang, G. Angew. Chem., Int. Ed. 2011, 50, 5913−5917. (15) Wang, Y.; Gao, S.; Zang, X.; Li, J.; Ma, J. Anal. Chim. Acta 2012, 716, 112−118. (16) Jiang, X.; Huang, K.; Deng, D.; Xia, H.; Hou, X.; Zheng, C. TrAC, Trends Anal. Chem. 2012, 39, 38−59. (17) Huang, Z.-H.; Zheng, X.; Lv, W.; Wang, M.; Yang, Q.-H.; Kang, F. Langmuir 2011, 27, 7558−7562. (18) Zhao, G.; Jiang, L.; He, Y.; Li, J.; Dong, H.; Wang, X.; Hu, W. Adv. Mater. 2011, 23, 3959−3963. (19) Liu, L.; Ryu, S.; Tomasik, M. R.; Stolyarova, E.; Jung, N.; Hybertsen, M. S.; Steigerwald, M. L.; Brus, L. E.; Flynn, G. W. Nano Lett. 2008, 8, 1965−1970. (20) McAllister, M. J.; Li, J. L.; Adamson, D. H.; Schniepp, H. C.; Abdala, A. A.; Liu, J.; Herrera-Alonso, M.; Milius, D. L.; Car, R.; Prud’homme, R. K.; Aksay, I. A. Chem. Mater. 2007, 19, 4396−4404. (21) Lin, S. C.; Shih, C. J.; Strano, M. S.; Blankschtein, D. J. Am. Chem. Soc. 2011, 133, 12810−12823. (22) Shih, C. J.; Lin, S. C.; Strano, M. S.; Blankschtein, D. J. Am. Chem. Soc. 2010, 132, 14638−14648.

■

CONCLUSION A novel, CPE-like but organic reagent-free method was developed using GO aggregation with addition of NaCl for preconcetration/separation of Pb(II), Cd(II), Bi(III), and Sb(III) from seawater and river water prior to their determination by slurry HG-AFS and ICP-OES. Owing to the benefit of well-dispersed aqueous colloids of GO and the abundant oxygen-containing functional groups on GO nanosheets, the analyte ions can be fast and efficiently adsorbed onto the GO surface. This method not only retains the principle advantages of cloud point extraction but also provides for simpler operation, greener analytical chemistry, an organic reagents free method, enhancements of sample throughput, and cost-effectiveness. Moreover, sufficiently low LODs obtained using the proposed method are well-suitable for the quantification of these analyte ions in river water samples or even the samples containing high concentration salts such as seawater samples. The proposed method has potential for the preconcentration of other metals or proteins that can be preconcentrated/separated by the CPE technique.

■

AUTHOR INFORMATION

Corresponding Author

certified (μg L−1)

Cd

Article

ASSOCIATED CONTENT

S Supporting Information *

Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org. G

dx.doi.org/10.1021/ac403345s | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

(23) William, S.; Hummers, R. E. O., Jr. J. Am. Chem. Soc. 1958, 80, 1339−1339. (24) Li, D.; Muller, M. B.; Gilje, S.; Kaner, R. B.; Wallace, G. G. Nat. Nanotechnol. 2008, 3, 101−105. (25) Sun, Y. B.; Wang, Q.; Chen, C. L.; Tan, X. L.; Wang, X. K. Environ. Sci. Technol. 2012, 46, 6020−6027. (26) Chandra, V.; Kim, K. S. Chem. Commun. 2011, 47, 3942−3944. (27) Zhao, G. X.; Li, J. X.; Ren, X. M.; Chen, C. L.; Wang, X. K. Environ. Sci. Technol. 2011, 45, 10454−10462. (28) Yang, S. T.; Chang, Y. L.; Wang, H. F.; Liu, G. B.; Chen, S.; Wang, Y. W.; Liu, Y. F.; Cao, A. N. J. Colloid Interface Sci. 2010, 351, 122−127. (29) Hirata, M.; Gotou, T.; Horiuchi, S.; Fujiwara, M.; Ohba, M. Carbon 2004, 42, 2929−2937. (30) Song, H.; Zhang, L.; He, C.; Qu, Y.; Tian, Y.; Lv, Y. J. Mater. Chem. 2011, 21, 5972−5977. (31) Skoog, D. A.; Holler, F. J.; Nieman, T. A. Principles of Instrumental Analysis; Hartcourt Brace & Company: Philadelphia, 1998; Chapter 13. (32) Paredes, J. I.; Villar-Rodil, S.; Martinez-Alonso, A.; Tascon, J. M. D. Langmuir 2008, 24, 10560−10564. (33) Stankovich, S.; Dikin, D. A.; Piner, R. D.; Kohlhaas, K. A.; Kleinhammes, A.; Jia, Y.; Wu, Y.; Nguyen, S. T.; Ruoff, R. S. Carbon 2007, 45, 1558−1565. (34) Li, D.; Kaner, R. B. Chem. Commun. 2005, 3286−3288. (35) Vallar, S.; Houivet, D.; El Fallah, J.; Kervadec, D.; Haussonne, J. M. J. Eur. Ceram. Soc. 1999, 19, 1017−1021. (36) Karadjova, I. B.; Lampugnani, L.; D’Ulivo, A.; Onor, M.; Tsalev, D. L. Anal. Bioanal. Chem. 2007, 388, 801−807. (37) Liu, R.; Wu, P.; Xi, M.; Xu, K.; Lv, Y. Talanta 2009, 78, 885− 890. (38) Kumar, A. R.; Riyazuddin, P. TrAC, Trends Anal. Chem. 2010, 29, 166−176.

H

dx.doi.org/10.1021/ac403345s | Anal. Chem. XXXX, XXX, XXX−XXX