Extraction of Copper (II) Ions from Aqueous Solutions with a

Aug 25, 2010 - Designated volumes of water and ionic liquid were used to extract copper(II) from aqueous solutions. After contact was made between the...
0 downloads 0 Views 2MB Size
Download PDF
Anal. Chem. 2010, 82, 7691–7698

Extraction of Copper(II) Ions from Aqueous Solutions with a Methimazole-Based Ionic Liquid Juan M. Reyna-Gonza´lez,† Angel A. J. Torriero,*,†,‡ Amal I. Siriwardana,†,‡ Iko M. Burgar,§ and Alan M. Bond*,†,‡ School of Chemistry and ARC Special Research Centre for Green Chemistry, Monash University, Clayton, Victoria 38000, Australia, and CMSE Division, CSIRO, Clayton, Victoria 3800, Australia The recently synthesized ionic liquid (IL) 2-butylthiolonium bis(trifluoromethanesulfonyl)amide, [mimSBu][NTf2], has been used for the extraction of copper(II) from aqueous solution. The pH of the aqueous phase decreases upon addition of [mimSBu]+, which is attributed to partial release of the hydrogen attached to the N(3) nitrogen atom of the imidazolium ring. The presence of sparingly soluble water in [mimSBu][NTf2] also is required in solvent extraction studies to promote the incorporation of Cu(II) into the [mimSBu][NTf2] ionic liquid phase. The labile copper(II) system formed by interacting with both the water and the IL cation component has been characterized by cyclic voltammetry as well as UV-vis, Raman, and 1H, 13C, and 15N NMR spectroscopies. The extraction process does not require the addition of a complexing agent or pH control of the aqueous phase. [mimSBu][NTf2] can be recovered from the labile copper-water-IL interacting system by washing with a strong acid. High selectivity of copper(II) extraction is achieved relative to that of other divalent cobalt(II), iron(II), and nickel(II) transition-metal cations. The course of microextraction of Cu2+ from aqueous media into the [mimSBu][NTf2] IL phase was monitored in situ by cyclic voltammetry using a well-defined process in which specific interaction with copper is believed to switch from the ionic liquid cation component, [mimSBu], to the [NTf2] anion during the course of electrochemical reduction from Cu(II) to Cu(I). The microextraction-voltammetry technique provides a fast and convenient method to determine whether an IL is able to extract electroactive metal ions from an aqueous solution. Over the past decade, ionic liquids (ILs) have attracted considerable attention for metal extraction from aqueous solutions, since they can replace the use of volatile organic solvents that have been widely used for this purpose.1-7 However, a major * Corresponding authors. E-mail: [email protected] (A.A.J.T.); [email protected] (A.M.B.). Fax: +61(3)99054597. Phone: +61(3)99051177. † School of Chemistry, Monash University. ‡ ARC Special Research Centre for Green Chemistry, Monash University. § CSIRO. (1) Chun, S.; Dzyuba, S. V.; Bartsch, R. A. Anal. Chem. 2001, 73, 3737–3741. (2) Dietz, M. L.; Dzielawa, J. Chem. Commun. 2001, 2124–2125. 10.1021/ac101321a  2010 American Chemical Society Published on Web 08/25/2010

limitation has been the loss of the IL components during the course of extraction.8-15 To improve the extraction of the metal ion from the aqueous solution without IL loss, hydrophobic ILs, with the addition of organic ligands in biphasic solutions, have been introduced.8-15 Task-specific ionic liquids have been also synthesized and employed for the extraction of metal ions from aqueous solutions.16,17 Nevertheless, the recovery of the ionic liquid from the metal complex for reuse is an important factor to be considered for the extraction system. Electrochemical deposition of the metal ion from the complex is one option to recover the IL, although this procedure is not as simple as in aqueous media.14 Due to the industrial and environmental importance of copper and its Cu(I) and Cu(II) redox states, the electrochemical behavior of Cu-based redox couples has been studied in a range of ILs as a function of temperature.18-20 Awakura et al. have investigated copper electrodeposition from the ionic liquid trimethyl-n-hexylammonium bis(trifluoromethanesulfonyl)amide.18,19 They found that monovalent copper ion was stable in this IL and that it could be reduced to metallic copper on a Pt electrode at 50 °C. Endres and co-workers studied the electrodeposition of copper using (3) Visser, A. E.; Holbrey, J. D.; Rogers, R. D. Chem. Commun. 2001, 2484– 2485. (4) Visser, A.; Swatloski, R. P.; Reichert, W. M.; Griffin, S. T.; Rogers, R. D. Ind. Eng. Chem. Res. 2000, 39, 3596–3604. (5) Visser, A.; Rogers, R. D. J. Solid State Chem. 2003, 171, 109–113. (6) Cocalia, V. A.; Holbrey, J. D.; Gutowski, K. E.; Bridges, N. J.; Rogers, R. D. Tsinghua Sci. Technol. 2006, 11, 188–193. (7) Heitzman, H.; Young, B. A.; Rausch, D. J.; Rickert, P.; Stepinski, D. C.; Dietz, M. L. Talanta 2006, 69, 527–531. (8) Wei, G. T.; Yang, Z.; Chen, C. J. Anal. Chim. Acta 2003, 488, 183–192. (9) Sun, X.; Peng, B.; Ji, Y.; Chen, J.; Li, D. Sep. Purif. Technol. 2008, 63, 61–68. (10) Sun, X. Q.; Peng, B.; Chen, J.; Li, D. Q.; Luo, F. Talanta 2008, 74, 1071– 1074. (11) Shimojo, K.; Goto, M. Anal. Chem. 2004, 76, 5039–5044. (12) Baghdadi, M.; Shemirani, F. Anal. Chim. Acta 2008, 613, 56–63. (13) Baghdadi, M.; Shemirani, F. Anal. Chim. Acta 2009, 634, 186–191. (14) Hsu, S. C. N.; Su, C. J.; Yu, F. L.; Chen, W. J.; Zhuang, D. X.; Deng, M. J.; Sun, I. W.; Chen, P. Y. Electrochim. Acta 2009, 54, 1744–1751. (15) Luo, H.; Dai, S.; Bonnesen, P. V. Anal. Chem. 2004, 76, 2773–2779. (16) Visser, A.; Swatloski, R. P.; Reichert, W. M.; Mayton, R.; Sheff, S.; Wierzbicki, A.; Davis, J. H.; Rogers, R. D. Environ. Sci. Technol. 2002, 36, 2523–2529. (17) Papaiconomou, N.; Lee, J. M.; Salminen, J.; Stosch, M. V.; Prausnitz, J. M. Ind. Eng. Chem. Res. 2008, 47, 5080–5086. (18) Murase, K.; Nitta, K.; Hirato, T.; Awakura, Y. J. Appl. Electrochem. 2001, 31, 1089–1094. (19) Katase, T.; Murase, K.; Hirato, T.; Awakura, Y. J. Appl. Electrochem. 2007, 37, 339–344. (20) Zein-El-Abedin, S.; Saad, A. Y.; Farag, H. K.; Borisenko, N.; Liu, Q. X.; Endres, F. Electrochim. Acta 2007, 52, 2746–2754.

Analytical Chemistry, Vol. 82, No. 18, September 15, 2010

7691

1-butyl-1-methylpyrrolidinium bis(trifluoromethanesulfonyl)amide, [bmpyr][NTf2], as the solvent.20 Because of the poor solubility of copper compounds in this IL, copper ions were introduced by anodic dissolution of a copper electrode. The oxidation state of copper was again found to be 1, with a [NTf2]- complex being formed. This study also established that higher temperatures promoted a progressive shift of the copper deposition potential and enhanced the efficiency of the stripping voltammetry. These effects were ascribed to both an increase in the mobility of the electroactive species toward the electrode surface and a reduced nucleation overvoltage at elevated temperatures.20 In this paper, we report the extraction of Cu(II) from aqueous systems using the IL 2-butylthiolonium bis(trifluoromethanesulfonyl)amide, [mimSBu][NTf2] (see the structural details in Scheme S1, Supporting Information),21 without addition of a complexing agent (IL acts in this role) or control of the pH. Voltammetric and spectroscopic techniques are used to establish the mechanism, which has several novel features. Importantly, the IL can be recovered for reuse by simple addition of a strong acid at room temperature. The microextraction of Cu2+ by an IL film adhered to an electrode was followed by in situ cyclic voltammetry. This technique is proposed as a rapid method to determine whether an IL is able to extract electroactive metallic ions from an aqueous solution. In addition, the electrochemical behavior of copper(II) in aqueous solutions containing [mimSBu]+ or [NTf2]- ions in the supporting electrolyte also is reported. EXPERIMENTAL SECTION Reagents. Copper(II) acetate monohydrate (Sigma-Aldrich), copper(II) sulfate pentahydrate, ferrous sulfate, and nickel chloride (May&Baker), cobaltous bromide (BDH), potassium chloride (Merck), anhydrous sodium acetate and glacial acetic acid (Ajax), lithium bis(trifluoromethanesulfonyl)amide (LiNTf2) and silver trifluoromethanesulfonate (AgOTf; Aldrich), sulfuric acid (Sigma), and 1-butyl-1-methylpyrrolidinium bis(trifluoromethanesulfonyl)amide ([bmpyr][NTf2]; Merck) were used as received from the manufacturer. The synthesis and characterization of 2-butylthiolonium chloride ([mimSBu][Cl]), 2-butylthiolonium bis(trifluoromethanesulfonyl)amide ([mimSBu][NTf2]), and 2-butylthiolonium hexafluorophosphate ([mimSBu][PF6]) were as reported in the literature.21,22 The structures of all ILs used in this study are shown in Scheme S1, Supporting Information. Deionized water from a Milli-Q-MilliRho purification system was used to prepare all aqueous solutions. Apparatus and Procedures. Designated volumes of water and ionic liquid were used to extract copper(II) from aqueous solutions. After contact was made between the aqueous and IL phases, extraction into the IL phase was facilitated by N2 bubbling for 10 min. The ionic liquid phase changes from colorless to green-blue as copper(II) is extracted from the aqueous phase (Figure 1). Centrifugation was not required to enhance the separation of the two phases. (21) Siriwardana, A.; Crossley, I. R.; Torriero, A. A. J.; Burgar, I. M.; Dunlop, N. F.; Bond, A. M.; Deacon, G. B.; MacFarlane, D. R. J. Org. Chem. 2008, 73, 4676–4679. (22) Torriero, A. A. J.; Siriwardana, A. I.; Bond, A. M.; Burgar, I. M.; Dunlop, N. F.; Deacon, G. B.; MacFarlane, D. R. J. Phys. Chem. B 2009, 113, 11222– 11231.

7692

Analytical Chemistry, Vol. 82, No. 18, September 15, 2010

Figure 1. (A) Photograph of the aqueous solution/IL system after extraction of Cu2+ with 200 µL of [mimSBu][NTf2]. (B) Control experiment under the same conditions as (A) but in the absence of Cu2+.

The loss of the IL cation into the aqueous phase during extraction was investigated by UV-vis spectroscopy. In these studies, 4 mL of aqueous solutions containing known concentrations of Cu2+-LiNTf2 were mixed with 50 µL of IL. After extraction, 0.5 mL of the aqueous phase was transferred to a new vial, and the solution was diluted by addition of 30 mL of deionized water. UV-vis spectra on the diluted solution were obtained with a Varian Cary 5000 UV-vis-NIR spectrophotometer. Raman spectra of [mimSBu][NTf2] and the [mimSBu][NTf2]-copper(II) system were recorded with a Renishaw RM 2000 Raman spectrograph using a 50× objective lens. 1H NMR, 13C NMR, and 15N NMR spectra on the IL phase before and after extraction of copper(II) were recorded at 21 ± 1 °C with a Varian 300 MHz VNMRS spectrometer. 1H and 13 C chemical shifts (ppm) are reported relative to the peak for an external tetramethylsilane (TMS) standard. Chemical shifts for 15N NMR spectra are reported relative to the peak for CH3NO3. In these NMR experiments, 4 mL of aqueous solution containing 20 mM Cu2+ were mixed with 300 µL of IL. The copper(II) concentration in the IL phase was estimated to be ∼16 mM by using UV-vis spectroscopy and application of Beer’s law. The IL phase, after separation from the aqueous phase, was used directly for the NMR experiments. Voltammetric measurements were undertaken at 21 ± 1 °C with a BAS 100B/W (Bioanalytical Systems, West Lafayette, IN) electrochemical workstation. A three-electrode cell was employed, with 1 mm diameter glassy carbon (GC) or platinum (Pt) (Cypress Systems, Inc., Lawrence, KS) working electrodes and a Pt wire as the counter electrode. For measurements in purely aqueous electrolyte solutions or when a thin layer of ionic liquid is present on the working electrode, which is in contact with aqueous electrolyte solution, a Ag|AgCl (3 M KCl) reference electrode was used. A Ag|10 mM AgOTf reference electrode containing the IL of interest in a separate compartment was used for voltammetric studies in the ionic liquid phase after extraction of copper(II). Prior to each experiment, the working electrodes were polished with 0.30 µm alumina (Buehler, Lake Bluff, IL) on a clean polishing cloth (Buehler), sequentially rinsed with distilled water and acetone, and then dried with lint-free tissue paper. Effective electrode areas of 0.85 mm2 for the GC electrode and 0.73 mm2

for the Pt electrode were determined using previously established procedures.22 A modified electrode for in situ voltammetric monitoring of extraction experiments was fabricated by pipetting microliter amounts of [mimSBu][NTf2] onto the electrode surface. The microdroplet-covered electrode was then directly placed in contact with 1.5 mL of aqueous solution containing known concentrations of copper(II). Dissolution of the ionic liquid droplet layer could be prevented by presaturation of the bulk aqueous electrolyte solution with the ionic liquid. The IL used to prepare the chemically modified electrode also was presaturated with water to maintain the viscosity of the IL phase constant. Scheme S2 (Supporting Information) provides a representation of the working electrode and interfaces present in this configuration. Prior to the attachment of the IL film, the working electrode was polished as described above. The pH of the aqueous solutions was determined by use of a Methrom model 744 pH meter with a glass pH electrode.

Figure 2. Cyclic voltammograms obtained at a GC electrode with v ) 0.1 V s-1 for reduction of 1.0 mM Cu2+ in aqueous 0.1 M LiNTf2 solution using switching potentials of (a) -0.1, (b) -0.2, (c) -0.3, (d) -0.4, (e) -0.6, and (f) -0.9 V.

RESULTS AND DISCUSSION Cyclic Voltammetry in Aqueous Solution Containing [mimSBu]+ and [NTf2]- and Their Mixtures. The influences of the potential (E), scan rate (v), pH, counteranion, background electrolyte, and temperature (T) on copper electrodeposition have been studied extensively in a wide range of media.23-28 The electrochemical reduction of copper(II) to copper(I) and sometimes to the metal state was studied initially in aqueous solutions containing electrolyte derived from LiNTf2 and water-soluble [mimSBu][Cl] and prior to studies in [mimSBu][NTf2] ionic liquid after extraction of copper(II) from the aqueous phase being undertaken. To probe the interaction between copper(II) ions and the [mimSBu][NTf2] IL cation and anion components in aqueous media, voltammetric experiments were undertaken with LiNTf2 as the electrolyte. Water-miscible [mimSBu][Cl] was then added to aqueous solutions containing 0.1 M LiNTf2 or KCl. Figure 2 displays voltammograms obtained for reduction of Cu2+ at a GC electrode in aqueous solutions containing 0.1 M LiNTf2 as a function of the switching potential with a scan rate of 0.1 V s-1. Two reduction processes with peak potentials of 0.02 V (process Ic) and -0.3 V (process IIc) were found when the potential was scanned in the negative direction over the range of +0.65 to -0.9 V. In the reverse scan, a major oxidation peak at 0.085 V (process Ia) with a shoulder at about 0.038 V (process IIa) was detected. Process Ic was assigned to the reduction of Cu(II) to Cu(I). The second process (IIc) was assigned to the reduction of Cu(I) or Cu(II) to metallic copper, although disproportionation of Cu+ to metallic copper and Cu2+ may also occur26 in competition with its electrochemical (23) Oskam, G.; Vereecken, P. M.; Searson, P. C. J. Electrochem. Soc. 1999, 146, 1436–1441. (24) Danilov, A. I.; Molodkina, E. B.; Polukarov, Y. M. Elektrokhimiya 2002, 38, 825–835. (25) Ramos, A.; Miranda-Herna´ndez, M.; Gonza´lez, I. J. Electrochem. Soc. 2001, 148, C315–C321. (26) Grujicic, D.; Pesic, B. Electrochim. Acta 2002, 47, 2901–2912. (27) Grujicic, D.; Pesic, B. Electrochim. Acta 2005, 50, 4426–4443. (28) Majidi, M. R.; Asadpour-Zeynali, K.; Hafezi, B. Electrochim. Acta 2009, 54, 1119–1126.

Figure 3. (A) Cyclic voltammograms obtained at v ) 0.1 V s-1 with a GC electrode for reduction of 1.0 mM Cu2+ in aqueous solution (0.1 M LiNTf2) after addition of (a) 0.66 mM (pH 5.1), (b) 1.33 mM (pH 4.9), and (c) 3.33 mM [mimSBu][Cl] (pH 4.6). (B) Repetitive cycles of potential for solution c, T ) 21 ( 1 °C.

reduction. As shown in Figure S1A (Supporting Information), the peak potential for process IIc, but not process Ic, is highly dependent on the scan rate. Process IIc has characteristics usually associated with an electrodeposition reaction.26 The shoulder IIa (Figure 2) is probably due to the oxidation of metallic copper to Cu2+, which can in turn react with the residual copper to form Cu+.18,25,27 As the potential becomes more positive, Cu+ and residual copper are oxidized to Cu2+, giving rise to process Ia. When the Analytical Chemistry, Vol. 82, No. 18, September 15, 2010

7693

potential is switched at less negative values, the oxidation current associated with process IIa decreases, which indicates less copper is deposited and hence stripped. When a switching potential of -0.1 V is used, process IIa does not occur, leaving only processes Ic and Ia (Figure 2), which now constitute a chemically reversible Cu2+/Cu+ couple without any evidence of stripping of metallic copper. Figure S1B (Supporting Information) shows the effect of repetitive scanning of the potential on copper electrodeposition. The copper deposition potential shifts from -0.3 V to less negative values with successive scans, making the electrodeposition easier. Clearly, the electrode surface activity is modified by cycling the potential. In summary, the basic reduction processes at a GC electrode are defined in aqueous solution containing 0.1 M LiNTf2 as the supporting electrolyte as Cu2+ + e- h Cu+

(1)

Cu+ + e- h Cumetal

(2)

The IL of interest in this study is [mimSBu][NTf2]. The above results reflect the role of [NTf2]- as the anion. The role of the [mimSBu] cation was now probed by addition of water-soluble [mimSBu][Cl] as the electrolyte. Figure S2 (Supporting Information) shows the effect of the [mimSBu][Cl] concentration on the pH of an aqueous solution. Clearly, the acidity increases as [mimSBu][Cl] is added. The probable source of H+ is the hydrogen attached to the N(3) nitrogen atom of the cation of the ionic liquid (Scheme S1, Supporting Information). On the basis of analysis of the data obtained in Figure S2, an estimated pKa value of 6.0 was obtained for the reaction given in the following equation:

Figures 3 and S3 (Supporting Information) show the effect of the [mimSBu][Cl] concentration on the voltammetry of copper(II) in aqueous solutions containing 0.1 M LiNTf2. As [mimSBu]+ can interact with [NTf2]- to form [mimSBu][NTf2], which is only sparingly soluble (see below), concentrations of [mimSBu][Cl] e 3.33 mM were added to avoid precipitation, at least in the bulk solution. Figure 3A-a shows voltammograms obtained for 1.0 mM Cu2+ in the presence of 0.66 mM [mimSBu][Cl]. Many of the features found in voltammograms for Cu2+ in 0.1 M LiNTf2 aqueous solutions are retained (Figure 2), but new processes are now evident at about 0.25 V (I′c) and 0.3 V (I′a) vs Ag|AgCl (3 M NaCl). These two new processes (I′c and I′a) are associated with the electrochemistry of a system formed by interaction of copper(II) and the [mimSBu]+ component now introduced. Further addition of [mimSBu][Cl] leads to a progressive shift of the potential needed for deposition of copper to more negative values and a decrease in both the reduction and oxidation peak currents for processes IIc and IIa (Figure 3A). Apparently, interaction of copper(II) with the [mimSBu] component in aqueous media increases the potential 7694

Analytical Chemistry, Vol. 82, No. 18, September 15, 2010

Figure 4. Cyclic voltammograms obtained at a GC electrode at v ) 0.1 V s-1 for 1 mM Cu2+ aqueous solution (0.1 M KCl) after addition of 0.0 mM (a) and 3.33 mM (b) [mimSBu][Cl].

needed for deposition of metallic copper and also provides a new pathway for reduction of Cu(II) to Cu(I). Other effects also may be operative29 that give rise to inhibition of reduction and appearance of the new reduction pathway. In particular, formation of a film or an array of microdroplets of the water-immiscible IL [mimSBu][NTf2] will be shown to be important. The inhibition of copper deposition was not simply a result of a pH change, as when H2SO4 was added instead of [mimSBu][Cl], the electrodeposition was enhanced (data not shown). Thus, the electrodeposition voltammetry is dependent on the concentration of [mimSBu]+ and [NTf2]- present in the aqueous solution. Apparently, the presence of [mimSBu]+ inhibits copper electrodeposition (Figures 3 and S3C), whereas [NTf2]- enhances the deposition process (Figure 2 and S1). Figure 3B provides a comparison of the initial and second cycles of potential for reduction of 1.0 mM Cu2+ in aqueous 0.1 M LiNTf2 solution containing 3.33 mM [mimSBu][Cl]. At this [mimSBu][Cl] concentration, process I′c becomes even better defined in the second cycle. This current increase may arise from the presence of a thin layer or an array of microdroplets of [mimSBu][NTf2] on the electrode surface, which allows preconcentration of Cu(II), which is then reduced to Cu(I) in the liquid film, in a process that is complementary to solvent extraction of Cu(II). To provide supporting evidence for processes assigned under conditions related to those present in Figure 3, the electrochemistry of Cu2+ also was examined in KCl-[mimSBu][Cl] aqueous solutions, without any LiNTf2 being present. Cyclic voltammograms obtained in aqueous media containing 0.1 M KCl as the supporting elecrolyte (Figure 4a) exhibit four well-defined processes (two reduction and two oxidation), which have their origin as assigned in eqs 1 and 2. The effect of addition of [mimSBu][Cl] on the electrodeposition of copper is shown in Figure 4b. It can be readily appreciated that the potential to achieve deposition of copper (process IIc) from copper(II) shifts to more negative values. Thus, electrodeposition of metallic copper is inhibited by addition of [mimSBu][Cl] as occurred when LiNTf2 was present in the supporting electrolyte (Figures 3 and S3, (29) Zhang, Q. B.; Hua, Y. X.; Wang, Y. T.; Lu, H. J.; Zhang, X. Y. Hydrometallurgy 2009, 98, 291–297.

Supporting Information). However, the blocking effect depends on the Cu2+-[NTf2]--[mimSBu]+ concentration relationship. This was confirmed by examination of Figure S4 (Supporting Information), which displays cyclic voltammograms obtained as a function of the Cu2+ concentration in an aqueous 0.4 M LiNTf2 electrolyte solution saturated with [mimSBu][NTf2]. When 1.0 mM Cu2+ is present (Figure S4a), one broad reduction process is observed in the initial negative potential direction scan, with one oxidation process located at 0.35 V (process I′a) detected upon reversal of the scan direction. This voltammogram exhibits features similar to those seen in Figure S3C. Therefore, the broad reduction step is assigned to reduction of a labile Cu(II) complex, accumulated in a film or array of microdroplets of [mimSBu][NTf2], to Cu(I), and process I′a is a result of the oxidation of copper(I) to copper(II), which can form a labile coordinated system with [mimSBu] in aqueous media. At higher Cu(II) concentrations, additional reduction occurs at more negative potential, giving rise to current crossover (Figure S4c,d, peak IIc), in a process which is assigned to the electrodeposition of copper, with the anodic stripping peak located at 0.17 V being assigned to the redissolution of the deposited metallic copper. Significantly, these results support the process assignments made under conditions relevant to Figure 3. Cyclic Voltammetry of Cu2+ in [mimSBu][NTf2]. The studies in aqueous media imply that Cu2+ interaction with [mimSBu][NTf2] should facilitate extraction of Cu2+ into the IL phase and that this should lead to detection of a Cu(II)/ Cu(I) couple if voltammograms are obtained in the watersaturated [mimSBu][NTf2] phase. Monitoring of in situ Microextraction of Cu2+ into the IL. Microextraction monitored by in situ cyclic voltammetry confirms that the ionic liquid [mimSBu][NTf2] is able to extract Cu2+ from aqueous media. Figures 5A and S5 (Supporting Information) provide examples of in situ cyclic voltammograms obtained when microextraction is achieved using a modified electrode fabricated according to Scheme S2 (Supporting Information) and where 0.6 µL of initially colorless [mimSBu][NTf2] was adhered to a GC electrode surface and placed in contact with a Cu2+ aqueous solution. This experiment produces a thick film of watersaturated [mimSBu][NTf2], rather than the thin film or array of microdroplets achieved in the studies described above. As the copper(II) aqueous solution was not stirred, the extraction occurs solely by diffusion processes. The color of the ionic liquid thick film adhered to the GC electrode was green-blue at the end of the experiment. When 0.1 M KCl is present in the aqueous phase, one broad reduction peak and its corresponding oxidation component characterize the cyclic voltammetry at the chemically modified electrode. Early in the extraction process, Cu2+ is mainly in the aqueous phase, and the reduction and oxidation peaks are detected at about 0.0 and 0.52 V, respectively. At longer times, and when the extent of extraction has increased, the reduction peak potential shifts to more negative values and the oxidation one to more positive values, probably due to an increase in the IRu (Ohmic) drop (where I ) current and Ru ) uncompensated resistance).

Figure 5. (A) In situ cyclic voltammograms obtained at v ) 0.1 V s-1 during the course of microextraction from an aqueous solution (0.1 M KCl) containing 1.0 mM copper(II) acetate into 0.6 µL of [mimSBu][NTf2] adhered onto a GC electrode. (B) Dependence of the cathodic peak current for reduction of copper ion with time. The aqueous phase contains (a) 1.0 mM copper acetate and 0.1 M acetate-acetic acid buffer, pH 4, (b) 1.0 mM copper acetate and 0.1 M KCl, (c) 1.0 mM copper acetate, 0.1 M KCl, and 0.1 M acetate-acetic acid buffer, pH 4, (d) 1.0 mM copper sulfate, 0.1 M KCl, and 0.1 M acetate-acetic acid buffer, pH 4, and (e) 1.0 mM copper acetate and 0.1 M LiNTf2.

A plot of the peak current for the reduction current versus time is shown in Figure 5B. The peak currents increase with time because of the transferral of copper ions from the aqueous phase into the ionic liquid film. In addition, data show that the extraction occurs at similar rates, irrespective of whether the aqueous phase contains buffer (Figure 5B-a), KCl (Figure 5B-b), or buffer-KCl (Figure 5B-c). Thus, controlling the pH of the aqueous solution phase with a buffer does not significantly modify the efficiency of the copper(II) extraction. At very long times, a decrease in peak current is detected, which may be a result of a change in film thickness or viscosity of the IL associated with the incorporation of very high Cu2+ concentrations. The extraction process was found to be slower when the copper(II) counteranion was changed to SO42- (Figure 5B-d) and even slower when LiNTf2 was present in the aqueous phase (Figure 5B-e). Thus, the presence of different anions can affect the extraction rate of Cu2+. Significantly, the presence of Co2+, Fe2+, and Ni2+ does not affect the Cu2+ extraction rate, nor are these ions extracted, indicating selectivity for Cu(II) is Analytical Chemistry, Vol. 82, No. 18, September 15, 2010

7695

Figure 6. (a) Cyclic voltammograms obtained at a scan rate of 0.1 V s-1 on a GC electrode in the IL phase after extraction of 25 mM Cu2+ aqueous solution with 400 µL of [mimSBu][NTf2]. (b) Voltammogram obtained after the potential was scanned to and then held at -1.4 V for 15 min.

achieved relative to these other transition-metal dications. A different situation was observed when silver ions were present in the aqueous solution, where competition with the copper extraction was detected. Cyclic Voltammetry of Cu2+ Extracted into [mimSBu][NTf2]. Copper(II) acetate is insoluble in dry [mimSBu][NTf2], and the presence of water (at a saturated level) is required to promote the incorporation of Cu2+ into the ionic liquid. In one set of experiments, 4.0 mL of a 25 mM Cu2+ aqueous solution was extracted with 400 µL of [mimSBu][NTf2]. The IL phase was separated from the aqueous phase, and the electrochemical behavior of the extracted copper system formed in [mimSBu][NTf2] was investigated by cyclic voltammetry (Figure 6). Figure 6a contains two well-defined processes (Ic and Ia) with a peak-to-peak separation of about 0.675 V that exhibit most of the features contained in Figure 5A. The reduction peak at -0.525 V (I′c) is assigned to the one-electron reduction of a labile complex formed by Cu2+, [mimSBu] (in its neutral or cationic form, eq 3), and water to Cu+, which could interact with the [NTf2] anion.20 The oxidation peak at about 0.15 V (I′a) corresponds to the oxidation of the Cu+-[NTf2]- complex back to Cu2+-[mimSBu]-H2O. Therefore, it is postulated that the ability of copper to coordinate with different ligands in each of the oxidation states I and II gives rise to the large peak-topeak potential separation. No copper electrodeposition was detected when the potential was switched at -1.5 V under these conditions. However, as shown in Figure 6b, when the potential was scanned to and then held at -1.4 V for 15 min and then scanned in the positive direction, two additional oxidation processes having peak potentials at about -0.237 V (IIa) and 0.15 V (Ia) were detected. These two processes are assumed to be related to the oxidation of electrodeposited metallic copper, as noted in the earlier discussion. The cyclic voltammogram obtained for reduction of copper(II) extracted into the IL phase at a Pt electrode (Figure S6A, Supporting Information) exhibits an even larger separation in reduction and oxidation peak potentials than seen in Figure 6 at a GC electrode. However, as was the case at a GC electrode, 7696

Analytical Chemistry, Vol. 82, No. 18, September 15, 2010

copper electrodeposition was not detected under short time scale conditions of cyclic voltammetry. Nevertheless, when the potential was held at -1.2 V for 30 min and the electrode was removed from the IL phase, washed with acetone, and dried with N2 prior to a cyclic voltammogram being recorded in 0.1 M KCl aqueous solution (Figure S6B), essentially the same four well-defined processes shown in Figure 4 were evident. Importantly, the ionic liquid can be recovered by washing with a concentrated solution of sulfuric acid, which back-extracts Cu2+ into the acid phase. The complete absence of copper ions in the IL after it was washed with sulfuric acid was confirmed by cyclic voltammetry. Details of the Extraction Process. Cation Exchange and IL Extraction Capacity. To study the chemistry associated with Cu2+ extraction from the aqueous to the IL phase, UV-vis measurements were undertaken. UV-vis spectra in the aqueous solution as a function of the Cu(II) concentration are shown in Figure S7A-a,b (Supporting Information). At a low 0.45 mM concentration of Cu2+ (Figure S7A-a) an absorption band is found in the UV region between 200 and 330 nm. At a high, 30.0 mM, concentration of Cu2+ (Figure S7A-b) the weak absorption band, with λmax ) 763 nm, is attributed to a Cu2+ d-d transition.14 The UV-vis absorption spectrum derived from 0.17 mM [mimSBu][Cl] is shown in Figure S7A-c. The IL cation exhibits absorption bands at λmax ) 222 and 255 nm. The absorption spectrum of [mimSBu]+ in the aqueous phase when 50 µL of IL and 4 mL of a 0.1 M LiNTf2 aqueous solution are mixed is shown in Figure S7B-a (Supporting Information). On addition and then extraction of Cu2+ into the IL phase, the IL cation bands at λmax ) 222 and 255 nm decrease as [mimSBu]+ is lost from the aqueous phase (Figure S7B-b1,b2). To quantify the [mimSBu]+ concentration change in the aqueous phase, a calibration plot was prepared using [mimSBu][Cl] and absorbance at λmax ) 222 nm. Initially, the concentration of the IL cation in the aqueous phase was 12 mM (Figure S7B-a). After 0.8 and 3.2 mM concentrations of Cu2+ were added to the water phase and extracted with 50 µL of IL, the concentration of the IL cation in the aqueous phase decreased to 10.2 and 6 mM, respectively. Thus, the solubility of the IL cation in the aqueous phase decreases as a result of Cu2+ extraction. The UV-vis absorption spectra of 34 mM Cu2+, in the aqueous phase containing an acetate-acetic acid buffer, before and after addition of different [mimSBu][NTf2] volumes are depicted in Figure S7C (Supporting Information), which shows that the characteristic Cu2+ absorption band at λmax ) 763 nm decreases as Cu2+ is extracted into the IL phase. Using copper acetate as a standard and via use of Beer’s law, it was estimated that 50 µL of [mimSBu][NTf2] can extract about 7 mM from a 34 mM Cu2+ aqueous solution, whereas 400 µL of IL can extract about 19 mM Cu2+, so the extraction ratio is not linear with the volume of the IL at high concentrations of IL. This is likely to be a consequence of the decrease in the effective contact area between phases as the IL volume increase and also to changes in viscosity and hence diffusion characteristics of the copper-IL system within the IL phase. The inset to Figure S7C represents the dependence of the pH of unbuffered aqueous solution on the volume of [mimSBu][NTf2]. Again, eq 3 is operative under these conditions.

Study of the Extracted Copper(II) System. The nature of the green-blue copper complex extracted into the IL phase was investigated by Raman, 1H, 13C, and 15N NMR, and UV-vis spectrometry. There are several possibilities for the Cu2+-IL interaction. In principle, copper ions can interact with the sulfur and the nitrogen atoms of the cation as well as with the [NTf2]of the IL. To establish which component of the [mimSBu][NTf2] IL system plays the dominant role in the extraction, related ionic liquids were studied by in situ cyclic voltammetry for Cu(II) extraction. The ionic liquid [bmpyr][NTf2] did not extract Cu2+ from aqueous solution in the pH range examined (3-7). In contrast, Cu2+ was extracted by the ionic liquid [mimSBu][PF6], implying that Cu2+ is extracted predominantly by [mimSBu]+ rather than [NTf2]-. Figure S8 (Supporting Information) shows the Raman spectra of [mimSBu][NTf2] before and after extraction of Cu2+ over the wavenumber range of 700-800 cm-1. The band at 742 cm-1, assigned to the νs(SNS) band for the “free” [NTf2]- in [mimSBu][NTf2],30 is unaltered by extraction of Cu2+, again implying that complexation of Cu2+ with [NTf2]- is not significant. Since the UV-vis absorption bands of Cu2+ and [mimSBu]+ overlap in the 200-300 nm wavelength range (Figure S7, Supporting Information), complexation of Cu2+ and the [mimSBu] component was investigated over the wavelength range from 400 to 800 nm, where [mimSBu] does not absorb light. The shift in the characteristic absorption band of Cu2+ at λmax ) 763 nm to λmax ) 565 nm (data not shown) is attributed to interaction of Cu2+ ions with the cation component of the ionic liquid. The 1H NMR resonance associated with the NH(im) of the neat (dry) ionic liquid was previously assigned and appears at an unexpectedly high frequency (δH 12.77 ppm),21 which implies it participates in strong intermolecular association and/ or intramolecular site exchange. In contrast, when the IL is saturated with water, the NH(im) resonance intensity is 3 times larger than expected and has shifted to 7.11 ppm (Figure S9Aa, Supporting Information). [mimSBu][NTf2] has been shown to be a source of H+ for the aqueous phase; this implies that NH(im) hydrogen ions are released from the IL cation to form H3O+ and rapidly replaced by hydrogen atoms from other water molecules, which explains the enhanced intensity observed during the 1H NMR integration for the NH(im) group. The addition of copper(II) ions by extraction into watersaturated [mimSBu][NTf2] affects several of the 1H, 13C, and 15 N NMR resonances (Figures S9 and S10, Supporting Information) due to their paramagnetism. Figure S9A reveals that the 1H resonance of NH(im) at 7.11 ppm is more extensively broadened than any other 1H NMR resonance after copper extraction, which suggests rapid exchange and interaction between copper(II) and the N(3) nitrogen occur. In the 13C NMR spectra (Figure S9B), all imidazolium ring carbons exhibit broader lines after copper extraction, except for the butyl carbons attached to sulfur atom. Importantly, 13C NMR resonances for the CF3 carbons of the IL [NTf2] anion are not affected by the presence of paramagnetic Cu2+ (Figure S9B-b). This finding supports the conclusions derived from Raman spectroscopy that copper(II) ions do not significantly interact with the [NTf2](30) Babai, A.; Mudring, A. V. Inorg. Chem. 2006, 45, 3249–3255.

Scheme 1. Schematic Representation of the Principal Processes That Occur at the IL/Water Interface When a [mimSBu][NTf2]-Modified Electrode Is Placed in Contact with an Aqueous Electrolyte Solution Containing Cu2+ ionsa

a A ) [NTf2]- or other anions. [mimSBu]+ and [mimSBu] are defined in eq 3.

component of the IL. The 13C spectra also show that the presence of copper ions does not produce chemical shift changes in the butyl carbons attached to sulfur atom (Table S1, Supporting Information). The 15N resonance of the IL cation in the presence of water (Figure S10) shows a small difference of about 1 ppm between the two imadizolium nitrogens. The presence of copper(II) ions broadens both resonances so that they are no longer resolved (line-broadening is at least 5 ppm, Figure S10). In contrast, the 15N resonance from the [NTf2]anion is not affected, consistent with the absence of any significant interaction with copper(II). Integration of the relative intensities of the [mimSBu]+ and [NTf2]- 15N resonances remains at exactly 2:1 in the presence of copper(II). Thus, 15N NMR spectra provide strong evidence for direct interaction and rapid exchange between the copper(II) ions and the N(3) atoms of the ionic liquid [mimSBu] component, to give a labile copper(II)-[mimSBu] coordinated system, aided by interaction with water molecules. Spectroscopic data lead to the postulation of the copper(II) extraction scheme depicted in Scheme 1. Copper(II) salts are insoluble in dry [mimSBu][NTf2], but the introduction of water, at a saturated concentration level, induces interaction with the IL cation system and hence promotes incorporation of Cu2+ into the ionic liquid. The addition of [mimSBu][Cl], [mimSBu][PF6], or [mimSBu][NTf2] to an aqueous solution produces a decrease in the pH value (Figure S2, Supporting Information). However, the addition of copper(II) into the aqueous phase of the biphasic extraction system does not induce further pH modification. This pH decrease in the aqueous phase implies that the hydrogen ions released by the IL cations (NH(im)) to form H3O+ and [mimSBu] (eq 3) are rapidly replaced by hydrogen atoms from other water molecules, which interact with [mimSBu] through hydrogen bond formation. This explains the enhanced intensity in the 1H NMR integration for the NH(im) group (Figure S9, Supporting Information) in water-saturated IL. The moiety formed between [mimSBu] and water facilitates extraction of copper(II) through interaction with the N(3) atom to give a labile Cu(II)-[mimSBu]-H2O complex system (Scheme 1), which is responsible for the green-blue color detected in the IL phase, as copper ions are extracted from the aqueous phase (Figure 1). Analytical Chemistry, Vol. 82, No. 18, September 15, 2010

7697

CONCLUSIONS The microextraction of Cu2+ by a thick IL film adhered to an electrode can be conveniently followed by in situ cyclic voltammetry. This technique provides a fast method to ascertain whether an ionic liquid is able to extract electroactive species from an aqueous solution. Cu2+ ions can be readily extracted from aqueous media into the new ionic liquid [mimSBu][NTf2]. High selectivity in the extraction is achieved against cobalt(II), iron(II), and nickel(II) transition metals. The extraction process does not require the addition of a complexing agent or pH control. The presence of water assists release of the hydrogen atom attached to the N(3) atom of the imidazolium ring and promotes incorporation of copper(II) ions into the ionic liquid. Copper(II) ions extracted into the IL phase form a labile complex system with the IL cation component and water. The ionic liquid can be recovered and reused after being washed with a strong acid. ACKNOWLEDGMENT This research was funded by Australian Research CouncilOrica Linkage Grant LP0668123. The Monash University ARC Special Research Centre for Green Chemistry and Monash University are gratefully acknowledged for provision of the

7698

Analytical Chemistry, Vol. 82, No. 18, September 15, 2010

facilities that supported this project. J.M.R.-G. is grateful to CONACyT for a postdoctoral fellowship. SUPPORTING INFORMATION AVAILABLE Structure and nomenclature of the ILs employed, schematic representation of a modified electrode/solution interface used in microextraction experiments, cyclic voltammograms for Cu2+ in aqueous solutions containing 0.1 M LiNTf2 as the supporting electrolyte as a function of the scan rate and switching potential, effect of the [mimSBu][Cl] concentration on the pH of the aqueous solution, voltammograms of Cu2+ in aqueous (0.1 M LiNTf2) solutions after addition of [mimSBu][Cl], voltammograms of Cu2+ in 0.1 M LiNTf2 and 0.1 M acetate-acetic acid buffer solution, in situ cyclic voltammograms in the microextraction configuration, cyclic voltammograms in the IL phase after copper(II) extraction, UV-vis absorption spectra, Raman spectra, 1H, 13C, and 15N NMR spectra, and a table with NMR data. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review June 9, 2010. Accepted August 6, 2010. AC101321A