Adsorption of Cd2+ on Carboxyl-Terminated Superparamagnetic Iron

Mar 18, 2012 - Magnetic particle imaging: advancements and perspectives for real-time in vivo monitoring and image-guided therapy. Michele H. Pablico-...
0 downloads 4 Views 5MB Size
Article pubs.acs.org/ac

Adsorption of Cd2+ on Carboxyl-Terminated Superparamagnetic Iron Oxide Nanoparticles Zhange Feng, Shun Zhu, Denis Ricardo Martins de Godoi, Anna Cristina S. Samia, and Daniel Scherson* Department of Chemistry, Case Western Reserve University, Cleveland, Ohio 44106, United States S Supporting Information *

ABSTRACT: The affinity of Cd2+ toward carboxyl-terminated species covalently bound to monodisperse superparamagnetic iron oxide nanoparticles, Fe3O4(np)COOH, was investigated in situ in aqueous electrolytes using rotating disk electrode techniques. Strong evidence that the presence of dispersed Fe3O4(np)COOH does not affect the diffusion limiting currents was obtained using negatively and positively charged redox active species in buffered aqueous media (pH = 7) devoid of Cd2+. This finding made it possible to determine the concentration of unbound Cd2+ in solutions containing dispersed Fe3O4(np)COOH, 8 and 17 nm in diameter, directly from the Levich equation. The results obtained yielded Cd2+ adsorption efficiencies of ∼20 μg of Cd/mg of Fe3O4(np)-COOH, which are among the highest reported in the literature employing ex situ methods. Desorption of Cd2+ from Fe3O4(np)-COOH, as monitored by the same forced convection method, could be accomplished by lowering the pH, a process found to be highly reversible.

T

This contribution describes a simple and versatile electrochemical method based on the use of a rotating disk electrode (RDE) to achieve the same goal. The merits of this strategy were illustrated by utilizing iron oxide nanoparticles of narrow size distributions functionalized with a surfactant incorporating siloxane surface anchoring groups and terminal carboxylic moieties, denoted hereafter as Fe3O4(np)-COOH, as sorbents for Cd2+. As will be shown, this technique can be applied in situ, i.e., in the presence of dispersed particles in the solution, thereby eliminating the need of the isolation step, yielding results in quantitative agreement with those reported by other research groups. It may be noted that most attention in the literature involving electrochemistry in the presence of nanoparticles has involved their incorporation into metal films during electrodeposition under a variety of conditions including forced convection.24−26

he declining quality of drinking water in vast regions of our planet due primarily to human activity has become a matter of grave concern.1 This alarming state of affairs has prompted the development and implementation of economic and effective methods for the removal of toxic organic and inorganic contaminants, particularly heavy metals.2−6 A uniquely intriguing strategy to mitigate some of these problems involves the use of iron oxide particles functionalized with species bearing terminal groups, e.g., −SH, −COOH, and −NH2, which display high affinity for such ions as Hg2+, Pb2+, and Cd2+.7−22 This core material is not only cheap and abundant but exhibits superparamagnetic properties when in the form of particles of nanometric dimensions,23 turning magnetic only in the presence of an external magnetic field. Such attributes would make it in principle possible to disperse the particles in water, enable them to bind the heavy metal ions and, subsequently, by using an external field collect them for further processing and reactivation. This approach, originally introduced by our research group using carboxylic groups as the metal binding moieties and Cd2+ as a model heavy metal ion,9,10 has been extended by a number of laboratories to other ligands and metal ions.13,15,17,18,20,22 Particularly noteworthy are the reports of Warner et al., who anchored five kinds of ligands onto iron oxide nanoparticles and compared their capacity as sorbents for a variety of metal ions.22 The ultimate aim of these efforts, neglecting financial aspects, is to maximize the adsorption efficiency expressed, most often, in terms of the weight of bound metal ions over the weight of the particles. With only a few exceptions, this figure of merit is determined by first removing the particles by filtration or other means and then assaying the liquid phase with conventional analytical techniques, such as inductively coupled plasma (ICP) or atomic absorption spectroscopy (AAS). © 2012 American Chemical Society



EXPERIMENTAL SECTION Chemicals. 1-Octadecene (90%), oleic acid (90%), FeCl3·6H2O (97%), triethylamine (≥99%), 4-morpholinepropane sulfonic acid sodium salt (≥99.5%), NaClO4 (≥98%), and ferrocenecarboxylic acid (97%) were purchased from SigmaAldrich Co. Sodium oleate (97%), 3-(triethoxysilyl)propyl succinic anhydride (>95%), HClO 4 (60%), potassium hexacyanoferrate (III) (>99%), KCl (>99%), sodium phosphate monobasic monohydrate (>98%), ammonium hydroxide (28−30 wt %), sodium phosphate dibasic heptahydrate (>98%), HCl (36.5 wt %), and NaOH (>97%) were obtained from Fisher Scientific. Hexaammineruthenium(III) chloride Received: February 9, 2012 Accepted: March 17, 2012 Published: March 18, 2012 3764

dx.doi.org/10.1021/ac300392k | Anal. Chem. 2012, 84, 3764−3770

Analytical Chemistry

Article

Figure 1. Transmission electron micrographs (top panels) for Fe3O4(np)-COOH of 8 nm (left) and 17 nm average diameter (right) and their corresponding size distributions (bottom panels).

of 1 M ammonium hydroxide in 1-butanol, 0.5 mL of water, 1.4 mL of triethylamine, and 0.1 mL of 3-(triethoxysilyl)propyl succinic anhydride. After shaking for 5 min, the suspension was allowed to settle for 1 h. The suspension was then centrifuged and the Fe3O4(np)-COOH transferred to water. Subsequently, 30 mL of ethanol was added and the functionalized nanoparticles were isolated by centrifugation. As evident from the results obtained, Fe3O4(np)-COOH could be easily dispersed in water over the pH range 5 to 12. Determination of Cd2+ in Solution. All electrochemical experiments were performed using a glassy carbon (GC) rotating disk electrode (RDE, Pine Instruments; disk area, 0.164 cm2) mounted on a Pine analytical rotator, a carbon rod counter, and an Ag/AgCl (3 M KCl) reference electrode, respectively. Measurements were conducted in ∼100 mL of N2purged 10 mM 4-morpholinepropane sulfonic acid, sodium salt (MOPS) buffer (pH 7), in 0.1 M NaClO4 aqueous solutions prepared with ultrapure water, to be referred to hereafter as the base electrolyte, containing Cd(ClO4)2 at specified concentrations. Potential control was achieved with an Autolab potentiostat (AUT83194). Dynamic polarization curves were recorded under forced convection at a scan rate ν = 10 mV/s to select a potential at which Cd deposition would proceed under diffusion limited conditions. The concentration of Cd2+ both in the absence and in the presence of well dispersed Fe3O4(np)COOH was measured chronopotentiometrically by monitoring the current as a function of time following application of a potential step from −0.4 V vs Ag/AgCl, i.e., positive to the onset of Cd deposition, to −1.0 V vs Ag/AgCl, i.e., negative enough for Cd deposition to occur under diffusion limited conditions, as determined from the polarization curves, allowing sufficient time was allowed for the current to achieve constant values. Also examined by the same techniques was the desorption of adsorbed Cd2+ induced by decreasing the pH.

(32.5% Ru) and Cd(ClO4)2 (99%) were purchased from Alfa Products. Ethanol (Decon Laboratories. Inc.) and 1-butanol (99.5%, Acros Organics) were used without further purification. Ultrapure water (18.3 Ω cm) was obtained from a Barnstead UV pure system. Synthesis and Characterization of Superparamagnetic Iron Oxide Nanoparticles.27 In total, 20 mmol of iron(III) chloride and 60 mmol of oleate were dissolved in a solution containing 40 mL of ethanol, 30 mL of distilled water, and 70 mL of hexane. The solution was heated to 60 °C for 4 h, cooled down to room temperature, and then thoroughly washed with water yielding, after drying overnight under vacuum, iron oleate. In total, 4 mmol of iron oleate and 16 mmol of oleic acid were dissolved in 15 mL of 1-octadecene at room temperature, ∼25 °C. The mixture was stirred until homogeneous and then heated to 320 °C at a rate of ∼2.5 °C min−1 and kept at that temperature until the mixture turned first clear brown and then black indicating formation of the iron oxide nanoparticles, Fe3O4(np). After about 1 h of subsequent heating under the specified temperature, the heat source was removed and the suspension allowed to cool down to room temperature. Particles were washed by adding 25 mL of ethanol to an aliquot of the suspension and then collected by centrifugation at 7000 rpm for 20 min. Nanoparticles were always kept in a toluene suspension before functionalization (see below). Different sizes of Fe3O4(np) were prepared by varying the relative amounts of oleic acid and the iron precursor, yielding nanoparticles with diameters of 8 and 17 nm. The size distribution of the Fe3O4(np) was determined from the analysis of transmission electron micrographs recorded with a JEOL 1200CX microscope (80 kV). Carboxyl Modified Superparamagnetic Iron Oxide Nanoparticles.28 In total, 4 mL of a 25 mg/mL Fe3O4(np) dispersion in toluene was added to a solution containing 4 mL 3765

dx.doi.org/10.1021/ac300392k | Anal. Chem. 2012, 84, 3764−3770

Analytical Chemistry

Article

Figure 2. (left panel) Dynamic polarization curves recorded with a GC RDE at ν = 10 mV/s in a buffered (pH 7) solution containing 2 mM FcCOOH in 0.1 M KCl for ω = 100, 400, 900, and 1600 rpm in the sequence defined by the arrow before (red lines) and after (green lines) adding 20 mg of Fe3O4(np)-COOH to the solution. Insert: Plots of the limiting currents, ilim vs ω1/2 recorded in the absence (red) and in the presence (green) of 20 mg of dispersed Fe3O4(np)-COOH based on the data in the left panel. (right panel) Chronoamperometric curves recorded under otherwise the same experimental conditions following a potential step from 0 to 0.5 V at t = 0 for ω = 100, 400, 900 and 1600, waiting 60 s before changing ω. The red and green lines represent data collected before and after adding 20 mg of dispersed Fe3O4(np)-COOH. Insert: Same as the insert in the left panel, where the red, green, and blue symbols correspond to data collected in the absence and the presence of 20 mg, about 10 mg of dispersed Fe3O4(np)-COOH, respectively.

Figure 3. (left panel) Chronoamperometric curves recorded with a GC RDE electrode following a potential step from 0 to −0.5 V at t = 0, in a 1 mM Ru(NH3)6Cl3 solution in 0.1 M KCl for ω = 100, 400, 900, and 1600 rpm in the sequence defined by the arrow. (right panel) Plot of ilim vs ω1/2 recorded in the absence (red symbols) and in the presence of 20 mg (green, ΔS = 0.7% and ΔI = 19%) and about 10 mg (blue, ΔS = 4.3% and ΔI = 10%).

rather narrow-sized distributions (see lower panels in the same figure). Effect of Dispersed Nanoparticles on the Diffusion Limited Currents. As stated in the introduction, the main objective of this work is to develop an electrochemical RDE method to measure in situ the efficiency of Fe3O4(np)-COOH for the adsorption of Cd2+ from solution phase. In order to validate this approach, it becomes essential to show that the presence of the dispersed Fe3O4(np)-COOH has a negligible effect on the limiting currents, ilim. To this end, we selected one positively charged, [Ru(NH3)6]3+, and two negatively charged, ferrocene carboxylate (FcCOO)− and [Fe(CN)6]3− redox active species.

The influence of dispersed Fe3O4(np)-COOH on both dynamic polarization and chronoamperometric curves was examined with the same GC RDE in N2-purged aqueous 0.1 M KCl containing either ferrocenecarboxylic acid (FcCOOH) (pH 7), Ru(NH3)6Cl3 and K3Fe(CN)6 before and after adding Fe3O4(np)-COOH.



RESULTS AND DISCUSSION

Structural Characterization. Representative TEM images of the three types of Fe3O4(np)-COOH deposited from a water dispersion and dried under ambient conditions (see upper panels, Figure 1) revealed the presence of clearly defined features of an average size of ∼8 and 17 nm, respectively, with 3766

dx.doi.org/10.1021/ac300392k | Anal. Chem. 2012, 84, 3764−3770

Analytical Chemistry

Article

Figure 4. Dynamic polarization curves (ν = 10 mV/s) recorded with a GC RDE in a solution containing 1 mM K3Fe(CN)6 in 0.1 M KCl at ω 900 rpm in the absence (red) and after in the presence of 20 mg (green) and about 10 mg of dispersed Fe3O4(np)-COOH (blue).

Figure 6. Chronoamperometric curves following a potential step from −0.4 to −1.0 V vs Ag/AgCl at t = 0 s recorded with a GC RDE in the same solution specified in the caption Figure 5 at ω = 100, 400, 900, and 1600 rpm in the sequence defined by the arrow.

Table 1. Slopes and Average Efficiencies for Cd2+ Adsorption on Fe3O4(np)-COOH slope (mA cm−2 rpm−1/2 ) specimen blank 8 nm, 20 mg 8 nm, 30 mg blank 17 nm, 20 mg 17 nm, 30 mg

trial A trial B 375 227 181 375 240 185

average efficiencya (μmol Cd/mg Fe3O4(np)-COOH)

376 225

0.199

376 254

0.189

a

Calculated from three batches based on the amount of 20 mg particles.

zero intercept 0.073 mA cm−2 for data collected before (red) and after adding the particles (green). In order to avoid uncertainties derived from these transient effects, values of ilim were determined by chronoamperometry by stepping the potential from 0 to 0.5 V at t = 0 (see right panel, Figure 2), while monitoring the current allowing sufficient time for steady state values to be achieved under otherwise the same conditions specified for the left panel, Figure 2. A Levich plot, i.e., ilim vs ω1/2, yielded, in this case, a straight line with approximately the same slope (S) intercept (I, see red symbols, insert, left panel, Figure 2). Once this set of measurements was completed, 20 mg of Fe3O4(np)-COOH was introduced into the media and the same experiments repeated. The data in this case (see green symbols) was found to be virtually identical to that observed in the pristine solution, i.e., ΔS = 1.06% and ΔI = 3.6%. Subsequently, the volume of the dispersion was reduced to one-half, an equal volume of the same solution without nanoparticles added to the cell, and the experiment was once again repeated. The results obtained are shown in blue symbols in the insert yielding very little difference compared to the other two sets (ΔS = 0.04% and ΔI = 5.7%). Hardly any differences were also observed in the case of solutions of 1 mM Ru(NH3)6Cl3 solution in 0.1 M KCl (see

Figure 5. Dynamic polarization curves (ν = 10 mV/s) recorded with a GC RDE electrode in a neat 0.1 mM Cd(ClO4)2 solution in the base electrolyte (see the Experimental Section) for ω = 100, 400, 900, and 1600 rpm in the sequence defined by the arrow.

Shown in the left panel of Figure 2 are dynamic polarization curves for the oxidation of 2 mM FcCOO− in 0.1 M KCl (PBS buffer, pH 7) recorded at a scan rate, ν = 10 mV/s, with a GC RDE at rotation rates ω = 100, 400, 900, and 1600 rpm in the sequence defined by the arrow before (red lines) and after adding 20 mg of Fe3O4(np)-COOH (green lines). These curves display classical behavior characterized by well-defined values of ilim for sufficiently positive potentials associated with the one-electron oxidation of the ferrous to yield the corresponding ferric derivative. The differences in the magnitudes of the current for the scans toward positive and negative potentials are due to the high interfacial capacitance of the GC electrode used in these experiments. As indicated, the effects induced by the presence of the dispersed particles are indeed very small. The Levich plots, i.e., ilim vs ω1/2, based on these results (see insert, left panel, Figure 2) yielded straight lines with a slope of 0.026 mA cm−2 rpm−1/2 with a close to 3767

dx.doi.org/10.1021/ac300392k | Anal. Chem. 2012, 84, 3764−3770

Analytical Chemistry

Article

Table 2. Comparison of Different Metal Ions Adsorption average diameter (nm) 8 17 20−30 30 18.4 140 5.8 6, 10 30 8

surface area (m2/g) a

144.8 68.1a b b 216c 62c 114c b b 100a,c

quality of the particles

efficiency (μmol Cd2+/mg adsorbent)

functional group

monodisperse monodisperse non-monodisperse non-monodisperse aggregated, irregular shape not dispersed, irregular shape non-monodisperse monodisperse non-monodisperse non-monodisperse

0.199 0.189 0.08 0.07 0.195 0.09 d 0.34 0.16 d

−COOH −COOH −COOH −COOH −NH2 −COOH −SH −COOH, −NH2, −SH −COOH −COOH, −SH, −NH2

ref this work 9 10 13 14 17 19 21 22

a Surface area calculated from the hard sphere model based on the diameter obtained from TEM images. bSurface area not mentioned in the literature. cSurface area determined by BET through N2 adsorption/desorption. dNo data available in equivalent units.

recorded with a GC RDE at ν = 10 mV/s, in the potential range −0.4 ≤ E ≤ −1.05 V vs Ag/AgCl in a solution containing 0.1 mM Cd(ClO4)2 in the base electrolyte at four different rotation rates, ω = 100, 400, 900, and 1600 rpm, in the sequence defined by the arrow. These curves display well-defined, equally spaced limiting currents, ilim, for sufficiently negative potentials associated with the reduction of Cd2+ to yield metallic Cd. The prominent feature centered at ∼−0.67 V vs Ag/AgCl found upon reversing the scan at −1.05 V is ascribed to stripping of metallic Cd. As described in detail in the previous sections, values of ilim were determined by chronoamperometry by stepping the potential in this case from −0.4 to −1.0 V at t = 0 (see Figure 6). Following each transient, the Cd layer was removed by polarizing the electrode at −0.4 V for ∼30 s. A Levich plot based on the chronoamperometric data was linear with a slope consistent with DCd2+ = 9.2 × 10−6 cm2/s, which compares well with the value of the infinite dilution of 7.2 × 10−6 cm2/s reported in the literature.30 It should be emphasized, however, that our method does not rely on the actual value of DCd2+, as the concentration of Cd2+ in solution will be determined from the ratio of the Levich slopes. The chronoamperometric data collected with the GC RDE in the same solution in the presence of Fe3O4(np)-COOH of different diameters made it possible to determine the concentration of Cd2+ remaining in solution and thus the corresponding amount adsorbed on the particles. For these experiments, the potential was stepped from −0.4 V, i.e., positive to the onset of Cd deposition, to −1.0 V, i.e., negative enough for Cd deposition to proceed under pure diffusion control, and the current monitored until steady state values were achieved. Shown in Table 1 are the values of the slopes of the Levich plots obtained for the three types of particles from a single batch in two different amounts, i.e., 20 and 30 mg, in two

Table 3. Percent of Cd2+ Release As a Function of pH for 8 nm Fe3O4(np)-COOH pH

release efficiency (%)

7 6.7 6.4 5.5

59.8 67.2 70.7 94.7

Figure 3) for which the redox active species is positively charged and could, in principle, interact with the negatively charged carboxylic groups. It can thus be concluded that the presence of the Fe3O4(np)-COOH in this case does not influence the kinetics nor ilim. The third and last redox species examined was the anion Fe(CN)63‑. In agreement with the behavior observed for the other two redox active ions, the effect of dispersed Fe3O4(np)COOH on the slope of ilim vs ω1/2 was minimal (not shown here). However, dynamic polarization curves recorded in 1 mM K3Fe(CN)6 solution containing either 10 or 20 mg of Fe3O4(np)-COOH, at ω = 900 rpm, displayed much lower currents in the mixed kinetics regime than those observed in the neat solution (see Figure 4). Although no satisfactory explanation for this effect can be offered at this time, the ability of this complex to form the highly insoluble Prussian blue could indeed play a role. Sirajuddin and Talbot investigated the effects of 50 nm alumina nanoparticles on the limiting currents for the reduction of Fe(CN)63− using a jet impinging electrode configuration and also found minor differences compared to particle-free solutions; however, distortions were observed in the rising part of their polarization curves.29 Efficiency of Fe3O4(np)-COOH toward Cd2+ Adsorption. Shown in Figure 5 are the dynamic polarization curves

Figure 7. (A) Photograph of dispersed Fe3O4 (np)-COOH of an average diameter of 17 nm in a buffer solution (pH = 7); (B) same as in part A with magnet after 10 min; (C) same as in part B after acidification with HCl (pH = 4.2) with a magnet for 1 min; (D) same as in part C after neutralization with NaOH (pH = 7) with a magnet for 10 min. 3768

dx.doi.org/10.1021/ac300392k | Anal. Chem. 2012, 84, 3764−3770

Analytical Chemistry



independent runs. The results obtained for other batches were found to be very similar. These data can be easily converted into micromoles of Cd2+ per mg of Fe3O4(np)-COOH, and the resulting average values are also given in Table 1. As indicated therein, the adsorption efficiency of the 8 nm particles was a bit higher than that for the 17 nm analogues, which is in line with their corresponding surface areas, and the actual values were found to be among the highest reported in the literature (see Table 2). Reactivation of Fe3O4(np)-COO -Cd2+. The release of adsorbed Cd2+ from Fe3O4(np)-COOH could be easily induced by simply lowering the pH of the solution. Some of the data collected using 20 mg of 8 nm Fe3O4(np)-COOH in 0.1 mM Cd2+ in pH 7 buffer are listed in Table 3, which provides values of the percent of Cd2+ left in the solution as a function of pH. As clearly seen, an increase in the acidity of the media led to the desorption of Cd2+. No data were acquired for pH < 4, as the particles underwent incipient dissolution. In a separate experiment, a small external FeNdB magnet was used to drive the functionalized particles (average diameter 17 nm) dispersed in the pH 7 solution toward the wall of a small container with no success (see parts A and B in Figure 7; however, once the pH was lowered, the particles agglomerated and the magnet was able to collect the particles on the wall leaving a close to clear solution (C). Subsequent addition of a small amount of base led to redispersion of the particles, indicating the process was fully reversible (D). We intend to exploit this property in future work to enable capture of Cd2+ in a cyclic fashion. The same method described earlier was used to test the efficiency of regenerated nanoparticles. Specifically, 20 mg of freshly prepared 8 nm diameter Fe3O4(np)-COOH was first used to adsorb Cd2+, a process that captured 0.45 mg of Cd2+ or, equivalently, 40.1% of the initial amount of metal ions. The particles were then collected by centrifugation, the adsorbed Cd2+ released by lowering the pH to 5, and subsequently washed with water and ethanol and later dried. The adsorptive properties of these reactivated particles were assayed following the same procedure previously described and found to upload 35% of the initial amount Cd2+, i.e., a retention of ∼87% of their original adsorbing capacity.



Article

ASSOCIATED CONTENT

S Supporting Information *

Cyclic voltammetric curves recorded with the GC carbon electrode. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



REFERENCES

(1) Cao, G.; Fryxell, E. G. Synthesis of nanostructured hybrid sorbent materials using organosilane self-assembly on mesoporous ceramic oxide. In Environmental Applications of Nanomaterials: Synthesis, Sorbents, and Sensors; Imperial College Press, London, U.K., 2007; pp 159−178. (2) Kwon, K. C.; Jermyn, H.; Mayfield, H. ACS Symp. Ser. 1992, 509, 161−178. (3) Hayashita, T. ACS Symp. Ser. 1996, 642, 303−318. (4) Battistoni, P.; Cola, E.; Fatone, F.; Bolzonella, D.; Eusebi, A. L. Ind. Eng. Chem. Res. 2007, 46, 6716−6723. (5) Huang, Z. H.; Zheng, X.; Lv, W.; Wang, M.; Yang, Q.-H.; Kang, F. Langmuir 2011, 27, 7558−7562. (6) Gao, W.; Majumder, M.; Alemany, L. B.; Narayanan, T. N.; Ibarra, M. A.; Pradhan, B. K.; Ajayan, P. M. ACS Appl. Mater. Interfaces 2011, 3, 1821−1826. (7) Ferrando, R.; Jellinek, J.; Johnston, R. L. Chem. Rev. 2008, 108, 845−910. (8) Takafuji, M.; Ide, S.; Ihara, H.; Xu, Z. H. Chem. Mater. 2004, 16, 1977−1983. (9) Hai, B.; Wu, J.; Chen, X. F.; Protasiewicz, J. D.; Scherson, D. A. Langmuir 2005, 21, 3104−3105. (10) Wu, J.; Sun, J.; Stefan, L. C.; Scherson, D. A. J. Phys. Chem. B 2004, 108, 12276−12279. (11) Chandra, V.; Park, J.; Chun, Y.; Lee, J. W.; Hwang, I. C.; Kim, K. S. ACS Nano 2010, 4, 3979−3986. (12) Chowdhury, S. R.; Yanful, E. K. Water Environ. J. 2011, 25, 429−437. (13) Wang, J.; Zheng, S.; Shao, Y.; Liu, J.; Xu, Z.; Zhu, D. J. Colloid Interface Sci. 2010, 349, 293−299. (14) Liu, J. F.; Zhao, Z. S.; Jiang, G.-B. Environ. Sci. Technol. 2008, 42, 6949−6954. (15) Siemeling, U.; Rittinghaus, S.; Weidner, T.; Brison, J.; Castner, D. Appl. Surf. Sci. 2010, 256, 1832−1836. (16) Zhou, Y.; Nie, H.; Branford, W. C.; He, Z; Zhu, L. J. Colloid Interface Sci. 2009, 330, 29−37. (17) Yantasee, W.; Warner, C. L.; Sangvanich, T.; Addleman, R. S.; Carter, T. G.; Wiacek, R. J.; Fryxell, G. E.; Timchalk, C.; Warner, M. G. Environ. Sci. Technol. 2007, 41, 5114−5119. (18) Afkhami, A; Asl, R. N. Colloids Surf., A 2009, 52−57. (19) Singh, S.; Barick, K. C.; Bahadur, D. J. Hazard. Mater. 2011, 192, 1539−1547. (20) Sadeghi, M.; Sarabadani, P.; Karami, H. J. Radioanal. Nucl. Chem. 2010, 283, 297−303. (21) Koehler, F. M.; Rossier, M.; Waelle, M.; Athanassiou, E. K.; Limbach, L. K.; Grass, R. N.; Guenther, D.; Stark, W. J. Chem. Commun. 2009, 4862−4864. (22) Warner, C. L.; Addleman, R. S.; Cinson, A. D.; Droubay, T. C.; Engelhard, M. H.; Nash, M. A.; Yantasee, W.; Warner, M. G. ChemSusChem 2010, 3, 749−757. (23) Lin, X. M.; Samia, A. C. S. J. Magn. Magn. Mater. 2006, 305, 100−109. (24) Kaisheva, M.; Fransaer, J. J. Electrochem. Soc. 2004, 151, C89− C96.

CONCLUSIONS

A rotating disk method has been developed and implemented for examining in situ the affinity of Cd2+ toward monodisperse carboxyl functionalized superparamagnetic iron oxide nanoparticles, Fe3O4(np)-COOH, of different sizes in aqueous solutions over a narrow pH range. No changes in the values of the limiting currents were observed for both positively and negatively charged redox species in solution before and after adding nanoparticles to the media. The Cd2+ adsorption efficiency of the particles as determined using this method was on the order of 0.17−0.20 μmol Cd/mg Fe3O4(np)-COOH, which ranks among the highest reported in the literature. Desorption of Cd2+ could be achieved by lowering the pH of the solution leading to agglomeration making it possible to easily capture the particles by an external magnet. Upon redispersion, by raising the pH, the Fe3O4(np)-COOH regained their ability to adsorb Cd2+. 3769

dx.doi.org/10.1021/ac300392k | Anal. Chem. 2012, 84, 3764−3770

Analytical Chemistry

Article

(25) Stappers, L.; Fransaer, J. J. Electrochem. Soc. 2005, 152, C392− C398. (26) Terzieva, V.; Fransaer, J.; Celis, J. P. J. Electrochem. Soc. 2000, 147, 198−202. (27) Park, J.; An, K. J.; Hwang, Y. S.; Park, J. G.; Noh, H. J.; Kim, J. Y.; Park, J. H.; Hwang, N. M.; Hyeon, T. Nat. Mater. 2004, 3, 891− 895. (28) Kohler, N.; Fryxell, G. E.; Zhang, M. Q. J. Am. Chem. Soc. 2004, 126, 7206−7211. (29) Sirajuddin, S.; Talbot, J. B. J. Electrochem. Soc. 2011, 158, D557− D560. (30) Macero, D. J.; Rulfs, C. L. J. Electrochem. Soc. 1964, 328−331.

3770

dx.doi.org/10.1021/ac300392k | Anal. Chem. 2012, 84, 3764−3770