AgCl Electrodes Based on Hydrophobic Ionic Liquid

Aug 21, 2007 - Application of a pH Electrode Incorporating an Ionic Liquid Salt Bridge to the Measurement of Rainwater Samples. Yota KUDO ... Properti...
8 downloads 19 Views 172KB Size
Download PDF
Anal. Chem. 2007, 79, 7187-7191

New Class of Ag/AgCl Electrodes Based on Hydrophobic Ionic Liquid Saturated with AgCl Takashi Kakiuchi,* Takahiro Yoshimatsu, and Naoya Nishi

Department of Energy and Hydrocarbon Chemistry, Graduate School of Engineering, Kyoto University, Kyoto 615-8510, Japan

A new type of Ag/AgCl electrodes based on a hydrophobic ionic liquid has been proposed. The electrode consists of a Ag/AgCl electrode immersed in or coated with a AgClsaturated ionic liquid, 1-methy-3-octylimidazolium bis(trifluoromethylsulfonyl)imide ([C8mim+][C1C1N-]), instead of the internal aqueous solution. The [C8mim+][C1C1N-] phase plays dual roles, that is, as a medium dissolving AgCl and an ionic-liquid-type salt bridge upon contact with an aqueous solution. The gelation of the [C8mim+][C1C1N-] phase allows us to prepare coatedwire-type solid-state reference electrodes with a welldefined thermodynamic basis for the electrode potential. Both gelled and nongelled types show stable electrode potentials against the change in the concentration of KCl between 0.05 mmol dm-3 and 2 mol dm-3. This new class of reference electrodes opens the way for a variety of miniaturized and solid-state reference electrodes. A salt bridge we proposed recently consisting of a hydrophobic room-temperature ionic liquid (RTIL), 1-methyl-3-octylimidazolium bis(trifluoromethylsulfonyl)imide ([C8mim+][C1C1N-]),1 maintains the phase-boundary potential between the ionic liquid and an aqueous solution (W) constant over a four-orders-of-magnitude change in the KCl concentration in W from 50 µmol dm-3 to 1 mol dm-3.1,2 In this case, the internal aqueous 1-methyl-3octylimidazolium chloride (C8mimCl) solution was an integral part of the reference electrode. For many practical applications of reference electrodes, however, the presence of the internal aqueous phase is unwelcome. For example, for miniaturization, a number of works have been dedicated to obtain solid-state reference electrodes3-19 and miniature liquid-junction reference * To whom correspondence should be addressed. Phone: (81)-75-383-2489. Fax: (81)-75-383-2490. E-mail: [email protected]. (1) Kakiuchi, T.; Yoshimatsu, T. Bull. Chem. Soc. Jpn. 2006, 79, 1017-1024. (2) Yoshimatstu, T.; Kakiuchi, T.; Anal. Sci., submitted for publication on April 16, 2007. (3) Collins, S. D. Sens. Actuators, B 1993, 10, 169-178. (4) Cosofret, V. V.; Ernosy, M.; Johnson, T. A.; Buck, R. P.; Ash, R. B.; Neuman, M. R. Anal. Chem. 1995, 67, 1647. (5) Nagy, K.; Eine, K.; Syverud, K.; Aune, O. J. Elecrochem. Soc. 1997, 144, L1. (6) Lee, H. J.; Hong, U. S.; Lee, D. K.; Shin, J. H.; Nam, H.; Cha, G. S. Anal. Chem. 1998, 70, 3377-3383. (7) Kaden, H.; Vonau, W. J. Prakt. Chem.-Chem. Ztg. 1998, 340, 710-721. (8) Mi, Y.; Mathison, S.; Bakker, E. Electrochem. Solid-State Lett. 1999, 2, 198. (9) Bakker, E. Electroanalysis 1999, 11, 788-792. (10) Yoon, H. J.; Shin, J. H.; Lee, S. D.; Nam, H.; Cha, G. S.; Strong, T. D.; Brown, R. B. Sens. Actuators, B 2000, 64, 8-14. 10.1021/ac070820v CCC: $37.00 Published on Web 08/21/2007

© 2007 American Chemical Society

electrodes.20-25 However, these attempts do not seem to be successful with regard to the stability and reliability in comparison with traditional reference electrodes having an internal solution, such as Ag/AgCl/aq.KCl and Hg/Hg2Cl2/aq.KCl. We herein describe a new type of reference electrodes with no internal aqueous phase. By directly immersing a Ag/AgCl electrode in a AgCl-saturated RTIL, [C8mim+][C1C1N-], or coating the gelled [C8mim+][C1C1N-] on a Ag/AgCl electrode, new reference electrodes with no internal aqueous phase have been prepared. Because the RTIL phase works as an RTIL salt bridge26 upon contact with an aqueous solution, this new type of reference electrodes has a built-in RTIL salt bridge, which outperforms KClbased salt bridges in many respects.1 The potential of the electrode in contact with an aqueous solution is stable over the change in the composition of the aqueous solution, as is the case of the RTIL salt bridge combined with a Ag/AgCl/aq.KCl electrode.1,2 Recently, RTILs have been used for reference electrodes.27-30 Ag/AgCl electrodes in an RTIL containing either Ag+ or Cl- have been shown to work satisfactorily as reference electrodes in ionic liquids.27-29 An important difference between these electrodes and the one we present in this study is that, in the latter, the RTIL (11) Mangold, K. M.; Scha¨fer, S.; Ju ¨ ttner, K. Synth. Met. 2001, 119, 345-346. (12) Ciobanu, M.; Wilburn, J. R.; Buss, N. L.; Ditavong, P.; Lowy, D. A. Electroanalysis 2002, 14, 989-997. (13) Jahn, H.; Kaden, H. Microchim. Acta 2004, 146, 173-180. (14) Tymecki, L.; Zwierkowska, Z.; Koncki, R. Anal. Chim. Acta 2004, 526, 3-11. (15) Blaz, T.; Migdalski, J.; Lewenstam, A. Analyst 2005, 130, 637-643. (16) Kisiel, A.; Marcisz, H.; Michalska, A.; Maksymiuk, K. Analyst 2005, 130, 1655-1662. (17) Ha, J.; Martin, S. M.; Jeon, Y.; Yoon, I. J.; Brown, R. B.; Nam, H.; Cha, G. S. Anal. Chim. Acta 2005, 549, 59-66. (18) Kim, S. K.; Lim, H.; Chung, T. D.; Kim, H. C. Sens. Actuators, B 2006, 115, 212-219. (19) Polk, B. J.; Stelzenmuller, A.; Mijares, G.; MacCrehan, W.; Gaitan, M. Sens. Actuators B 2006, 114, 239-247. (20) Yee, S. H. J.; Lam, L. K. C. Sens. Actuators 1988, 15, 337-345. (21) van den Berg, A.; Grisel, A. Sens. Actuators 1990, B1, 425-432. (22) Jermann, R.; Tercier, M. L.; Buffle, J. Anal. Chim. Acta 1992, 269, 49. (23) Diamond, D.; Mcenroe, E.; Mccarrick, M.; Lewenstam, A. Electroanalysis 1994, 6, 962. (24) Nolan, M. A.; Tan, S. H.; Kounaves, S. P. Anal. Chem. 1997, 69, 1244. (25) Suzuki, H.; Hirakawa, T.; Sasaki, S.; Karube, I. Anal. Chim. Acta 1999, 387, 103-112. (26) Kakiuchi, T.; Tsujioka, N.; Kurita, S.; Iwami, Y. Electrochem. Commun. 2003, 5, 159-164. (27) Fukui, R.; Katayama, Y.; Miura, T. Electrochemistry 2005, 73, 567-569. (28) Saheb, A.; Janata, J.; Josowicz, M. Electroanalysis 2006, 18, 405-409. (29) Snook, G. A.; Best, A. S.; Pandolfo, A. G.; Hollenkamp, A. F. Electrochem. Commun. 2006, 8, 1405-1411. (30) Maminska, R.; Dybko, A.; Wroblewski, W. Sens. Actuators, B 2006, 115, 552-557.

Analytical Chemistry, Vol. 79, No. 18, September 15, 2007 7187

Methods. The electrochemical cell we employed for potentiometry is similar to that we previously reported,26 except for the absence of the internal aqueous solution phase in the present case. The cell voltage, referred to by the left-hand-side terminal in the cell (1), E, was measured with an electrometer with a GPIB

I

II

III

IV

V

VI

Ag AgCl [C8mim+][C1C1N-] a mmol dm-3 AgCl Ag KCl or LiCl satd. AgCl (RTIL) (W) (1)

Figure 1. Picture of gelled-[C8mim+][C1C1N-]-coated Ag/AgCl electrode: 1, silver wire; 2, silicon rubber stopper; 3, Ag/AgCl; 4, [C8mim+][C1C1N-] gel saturated with AgCl.

phase acts not only as a medium dissolving AgCl but also as an RTIL salt bridge whose properties are ditinct from traditional KCltype salt bridges. EXPERIMENTAL SECTION Reagents. C8mimCl was synthesized from 1-methyl-imidazole (Wako Pure Chem., 98+%) and 1-chlorooctane (Aldrich 99%).31 [C8mim+][C1C1N-] was prepared from C8mimCl and bis(trifluoromethylsulfonyl)imide acid (Central Glass Co. Ltd. (Japan)) as described elsewhere.1 [C8mim+][C1C1N-] was saturated with water in the process of washing crude [C8mim+][C1C1N-] with water repeatedly. AgCl-saturated [C8mim+][C1C1N-] was prepared by stirring [C8mim+][C1C1N-] solution with AgCl powder at 25 °C. Poly(vinylidene fluoride-co-hexafluoropropylene) (P(VdFHFP), average MW 400 000, Aldrich) was used for gelation of C8mimC1C1N. Other chemicals were of reagent grade. The gelation of [C8mim+][C1C1N-] was made by adding P(VdF-HFP) in an acetone solution of C8mimC1C1N.32 A typical composition of the solution was 1:1:5 by weight for AgCl-saturated [C8mim+][C1C1N-], P(VdF-HFP), and acetone, respectively. RTIL-coated electrodes were made by painting the acetone solution of P(VdF-HFP) and [C8mim+][C1C1N-] on a 0.5 mm diameter Ag/AgCl electrode and then drying acetone in air. This process was repeated at least 15 times to obtain a thick membrane of [C8mim+][C1C1N-] gel covering the Ag/AgCl electrode. Figure 1 shows a photograph of a typical [C8mim+][C1C1N-] gel-coated Ag/AgCl electrode. The length of the coated portion was about 1.5 cm, and the thickness of the gel was about 1.5 mm. The electrodes were stored in water saturated with [C8mim+][C1C1N-] when not in use. Ag/AgCl electrodes were prepared by anodizing 0.5 mm diameter silver wires in an aqueous KCl solution. (31) Fordon, C. M.; Holbrey, J. D.; Kennedy, A. R.; Seddon, K. R. J. Mater. Chem. 1998, 8, 2627-2636. (32) Fuller, J.; Breda, A. C.; Carlin, R. T. J. Electroanal. Chem. 1998, 459, 2934.

7188

Analytical Chemistry, Vol. 79, No. 18, September 15, 2007

interface, as described elsewhere.26 The electrometer reading was made every 10 s for usually 15 min. The glass cells we used for potentiometric measurements for nongelled RTIL and gelled RTIL cases are similar to that we reported elsewhere.1 In the former, AgCl-saturated [C8mim+][C1C1N-] was filled in the cell to the upper surface of the glass frit in the middle of the cell. AgCl-free [C8mim+][C1C1N-] of a few millimeters thickness was layered on the glass frit. For measuring E in the case of the gel-coated Ag/ AgCl electrode, a Ag/AgCl electrode coated with gelled [C8mim+][C1C1N-] (Figure 1) was immersed in the aqueous solution. Ag/ AgCl electrodes were positioned in the glass cell using silicon rubber stoppers. The temperature of the cell was maintained at 25 ( 0.2 °C by circulating water through the outer jacket of the cell. The solubility of AgCl in [C8mim+][C1C1N-] was measured with an atomic absorption spectrometer (Z-2700, Hitachi) using the standard addition method. RESULTS AND DISCUSSION Time Course of E at Different Concentrations of KCl for Nongelled and Gelled [C8mim+][C1C1N-]. Parts a and b of Figures 2 are time dependences of E at 16 different concentrations -3 to 2 mol dm-3 in aqueous of KCl, cW KCl, from 0.02 mmol dm phase IV for nongelled (Part a of Figure 2) and gel-coated (part b of Figure 2) cases. The aqueous KCl solutions were presaturated with [C8mim+][C1C1N-]. Both parts a and b of Figure 2 show stable readings over 15 min in the wide concentration range of KCl, as was the case of [C8mim+][C1C1N-] salt bridges of nongelled and gelled types.1 At the beginning of the measurements, there was a slight difference in response time. In the gel-coated case, the reading was stable from the beginning of the measurements with a variation of E within (0.2 mV (part b of Figure 2), whereas in the nongelled case, it took ca. 1 min to have a stable reading within (0.2 mV (part a of Figure 2). This is similar to the gelled RTIL salt bridges that gave a more stable reading of E in the [C8mim+][C1C1N-] salt bridges, in the presence of an internal aqueous phase.1 There was no significant difference in response time when the aqueous solutions were not presaturated with [C8mim+][C1C1N-] (data not shown). The dissolution of [C8mim+][C1C1N-] in W is therefor not responsible for this initial change over a few minutes. It may require some time to equilibrate the response of the Ag/AgCl electrode, which was kept in water before use and washed with the solution having the same composition with that in phase IV, prior to the insertion in the cell. Constancy of the Electrode Potential of Ag/AgCl/[C8mim+][C1C1N-] Against the Change in KCl Concentration. Figure 3

Figure 3. Dependence of the cell voltage on the logarithm of the mean ionic activity of KCl for nongelled (b) and gelled (O) [C8mim+][C1C1N-]. Solid lines have a slope of 59.2 mV par decade, which is expected when Ag/AgCl ideally responds to the activity of Cl-, and the phase-boundary potential between phases II and III remains constant. Data for aqueous LiCl solutions are plotted as ([).

Figure 2. Time dependence of the cell voltage with Ag/AgCl in contact with nongelled (a) and gelled (b) [C8mim+][C1C1N-] saturated with AgCl at different concentrations of KCl. Each point represents a value obtained from measurements for a single run. The sampling interval was 10 s. Concentrations of KCl: 2 × 10-5 (1), 5 × 10-5 (2), 1 × 10-4 (3), 2 × 10-4 (4), 5 × 10-4 (5), 1 × 10-3 (6), 2 × 10-3 (7), 5 × 10-3 (8), 1 × 10-2 (9), 2 × 10-2 (10), 5 × 10-2 (11), 1 × 10-1 (12), 2 × 10-1 (13), 5 × 10-1 (14), 1 (15), and 2 (16) mol dm-3.

shows the response of E with the mean ionic activity of KCl, a( KCl, for nongelled (b) and gel-coated (O) cases, respectively. The activity coefficients were from literature values, in addition -3 to those interpolated when cW KCl was above 10 mmol dm . For

-3 cW KCl values less than 10 mmol dm , they were calculated using the Debye-Hu¨ckel’s limiting law. In Figure 3, it is seen that the data points fall on the straight line expressing the Nernst slope, 59.2 mV per decade at 25 °C over the entire concentration range of KCl studied. Because the Ag/AgCl electrode in phase IV responds to a( KCl in a Nernstian manner, this variation of E with a( can be ascribed to the change in the potential at the Ag/ KCL AgCl electrode and provides evidence that the phase-boundary potential between phases III and IV in cell (1) stays constant over the 4 orders of magnitude change in cW KCl. The lowest concentration, 20 µmol dm-3, is close to the solubility of AgCl in water, 13 µmol dm-3;33 hence, the deviation by a few millivolts in the nongelled case may be due to the deviation of the Cl- concentration from that as prepared. When + cW KCl is lower than the solubility of [C8mim ][C1C1N ] in water, 1.6 mmol dm-3 at 25 °C,34 the diffusion potential due to the different mobilities of C8mim+ and C1C1N- ions in W may also contribute to the deviation.2 The deviation in the middle of the plot is likely to be due to the change in the properties of the Ag/ AgCl electrode in phase IV. The potentiometric measurements were made consecutively from a dilute solution to a concentrated solution. Each time when the KCl solution was renewed, the Ag/ AgCl electrode in phase IV was washed repeatedly with water, first, and then washed with a sample solution at least three times. The deviation of E values from the Nernst slope was much smaller in the case of gelled [C8mim+][C1C1N-] ((O) in Figure 3). In this case, washing of the gel-coated Ag/AgCl every time

(33) Waghorne, W. E. Monatsh. Chem. 2003, 134, 655-667. (34) Our most recent solubility measurements show that the solubility of [C8mim+][C1C1N-] in water at 25 °C is 1.6 mmol dm-3, which agrees with the average of our two previously reported values, 1.426 and 1.8 mmol dm-3.1 These values are considerably lower than the value, 4.4 mmol dm-3.55,56

Analytical Chemistry, Vol. 79, No. 18, September 15, 2007

7189

we changed the sample solution apparently did not appreciably alter the properties of the Ag/AgCl electrode. The deviation at -3 downward from the Nernst slope is similar cW KCL ) 20 µmol dm to the nongelled case. Given the constancy of the phase-boundary potential between phases III and IV even at 50 µmol dm-3 KCl (Figure 3), together with the stable time course at this concentration in Figure 2, the RTIL-coated Ag/AgCl reference electrode is particularly useful in potentiometry of low ionic strength solutions, such as rainwater, for which the KCl-type salt bridge has been known to be unreliable.35-44 Potential of [C8mim+][C1C1N-]-Coated Ag/AgCl Electrodes. For a reference electrode equipped with a salt bridge, the constancy of the electrode potential is of crucial imortance, whereas the value of E is of minor importance. Nevertheless, it is sometimes necessary to refer the value of the electrode potential with respect to those of existent reference electrodes. The evaluation of the electrode potential of [C8mim+][C1C1N-]-coated Ag/AgCl electrodes can be easily made through the results in Figure 3 using a value of E at a certain concentration of KCl at which a [C8mim+][C1C1N-]-coated Ag/AgCl electrode works satisfactorily. For example, at the concentration of KCl at 0.1 mol dm-3, E values are 0.02 and 0.07 V for nongelled and gel-coated cases, respectively. Because the standard electrode potential of Ag/AgCl/aq.KCl is 0.222 V at 25 °C,45 the potential of the [C8mim+][C1C1N-]-coated Ag/AgCl electrode with respect to the standard hydrogen electrode is estimated by subtracting 0.02 or 0.07 V from (0.222-0.02569 ln a( KCl) V. Then, the values of the electrode potentials of nongelled and gel-coated [C8mim+][C1C1N-]coated Ag/AgCl electrodes referred to as the standard hydrogen electrode are estimated to be 0.27 and 0.22 V, respectively, at 25 °C. Further detailed studies are required for more precise estimates. Solubility of AgCl in [C8mim+][C1C1N-]. We examined the significance of the saturation of [C8mim+][C1C1N-] with AgCl by dipping a Ag/AgCl electrode in [C8mim+][C1C1N-] without prior saturation with AgCl and found that E values were stable as well. Probably, a small amount of AgCl dissolved in the RTIL phase after the dipping. For better stability of the RTIL, the saturation with AgCl is nonetheless preferable. The solubility of AgCl in [C8mim+][C1C1N-] was (4.8 ( 0.2) × 10-5 mol kg-1, which corresponds to (6.3 ( 0.3) × 10-5 mol dm-3 at 25 °C. AgCl is almost 5 times more soluble in [C8mim+][C1C1N-] than in water. Phase-Boundary Potential between AgCl and [C8mim+][C1C1N-]. The saturation of [C8mim+][C1C1N-] with AgCl is important not only to prevent the dissolution of AgCl covering (35) (36) (37) (38) (39) (40) (41) (42) (43) (44) (45)

Galloway, J. N.; Cosby, B. J. Limnol. Oceanogr. 1979, 24, 1161. Midgrey, D.; Torrance, K. Analyst 1979, 104, 63-72. Brezinski, D. P. Analyst 1983, 108, 425-442. Covington, A. K.; Whalley, P. D.; Davison, W. Pure Appl. Chem. 1985, 57, 877-886. Davison, W.; Woof, C. Anal. Chem. 1985, 57, 2567-2570. Koch, W. F.; Marinenko, G.; Paule, R. C. J. Res. Nat. Bur. Stand. 1986, 91, 23. Davison, W.; Gardner, M. J. Anal. Chim. Acta 1986, 182, 17-31. Franklin, S.; Miller, G. M. Am. Lab. 1989, 40. Durst, R. A.; Davison, W.; Koch, W. F. Pure Appl. Chem. 1995, 66, 649658. Ozeki, T.; Tsubosaka, Y.; Nakayama, S.; Ogawa, N.; Kimoto, T. Anal. Sci. 1998, 14, 749-756. Janz, G. J. In Reference Electrodes; Ives, D. J. G., Janz, G. J., Eds.; Academic Press: New York, 1961; Chapter 4.

7190

Analytical Chemistry, Vol. 79, No. 18, September 15, 2007

Ag but also to establish the phase-boundary potential between the solid AgCl and [C8mim+][C1C1N-]. Unlike conventional Ag/ AgCl/KCl(aq.) electrodes or a Ag/AgCl in contact with an RTIL containing Cl-,27,30 there appears to be no common ion between phases II and III in cell (1). Actually, the phase-boundary potential is through the partition equilibrium of relevant ionic species between phase II and III because of the finite solubility of AgCl in [C8mim+][C1C1N-]. It would therefor be better understood as a distribution potential similar to the one across the interface between two immiscible electrolyte solutions46 or between a hydrophobic RTIL and water.26 In other words, the distribution of ionic species, primarily Ag+ and Cl-, assures the thermodynamic equilibrium between the two phases.47-49 The phaseboundary potential between phase II and III in cell (1), φIL - φAgCl, may then be expressed formally in terms of the transfer Gibbs energies of Ag+ and Cl- ions between the phases II and III, ∆ RTILfAgCl RTILfAgCl Gtr,Ag and ∆Gtr,Cl : + -

φIL - φAgCl )

1 RTILfAgCl RTILfAgCl [∆Gtr,Ag - ∆Gtr,Cl ] + 2F

This form is the same as the expression of the phase-boundary potential between an RTIL and water saturated with the RTIL,26 which is a variant of the well-known expression for the distribution potential due to the partition of a single 1:1 electrolyte between two immiscible electrolyte solutions.46 The electrochemical polarizability of the interface between phases II and III is then determined by the solubility of AgCl in [C8mim+][C1C1N-], just like the electrochemical polarizability of an RTIL/W interface, where the polarization resistance is inversely proportional to the square root of the solubility product.50 For this reason, the saturation of [C8mim+][C1C1N-] with AgCl is crucial. Because the solubility of [C8mim+][C1C1N-] in water is higher than that of AgCl in [C8mim+][C1C1N-], the electrochemical polarizability of this reference electrode would be determined by the latter. There might be a concern that AgCl in [C8mim+][C1C1N-] eventually contaminates sample solutions upon contact, although the solubility of AgCl in W is limited. To avoid this, we tested the effect of layering a AgCl-free [C8mim+][C1C1N-] on top of AgClsaturated [C8mim+][C1C1N-] and found no significant difference. For practical applications, it is probably better to use this approach of layering AgCl-free [C8mim+][C1C1N-]. Response of Ag/AgCl/[C8mim+][C1C1N-] to Aqueous LiCl Solutions. Previously, we showed that the constancy of the phaseboundary potential was disturbed at higher concentrations of HCl and LiCl, and to a lesser degree, of NaCl. This deviation is presumably due to the dissolution of hydrophilic cations and Clions in the RTIL phase. The mechanism of maintaining the phaseboundary potential of Ag/AgCl/[C8mim+][C1C1N-] should be the same as that for the RTIL salt bridge. We expect a similar deviation at higher concentrations of these electrolytes. Figure 3 ([) shows the response of a [C8mim+][C1C1N-]-gel coated Ag/AgCl elec(46) Luther, R. Z. Phys. Chem. 1896, 19, 529-571. (47) Haber, F. Ann. Phys 1908, 26, 927-973. (48) Dole, M. In Principles of Experimental and Theoretical Electrochemistry; Dover: New York, 1961; Chapter 22. (49) Kolthoff, I. M.; Sanders, H. L. J. Am. Chem. Soc. 1937, 59, 416-420. (50) Kakiuchi, T.; Tsujioka, N. J. Electroanal. Chem. 2007, 599, 209-212.

trode against different concentrations of LiCl in W between 1 mmol dm-3 and 2 mol dm-3. A significant deviation from the Nernst slope at higher concentrations started from 0.2 mol dm-3, as was the case for a [C8mim+][C1C1N-] salt bridge in contact with aqueous LiCl solutions.1 The limitations of the use of the RTILcoated electrodes are similar to those of the corresponding salt bridge. CONCLUSIONS A new class of reference Ag/AgCl electrodes has been proposed. The use of a gelled RTIL for both the internal electrolyte and also the RTIL salt bridge opens the way to materialize miniaturized reference electrodes equipped with a solid-state RTIL salt bridge, which have been sought after for a long time. This reference electrode is given all of the advantages, and disadvantages, of RTIL salt bridges, we recently proposed.1 The former includes the stability of the phase-boundary potential against the composition of aqueous solutions over conventional KCl-type reference electrodes, in particular, dilute aqueous solutions, longterm stability over 3 months, less demanding maintenance, and the thermodynamic nature of the phase-boundary potential. Besides, RTILs, such as [C8mim+][C1C1N-], are redox inactive because the potential window is wide enough so that water (51) Schro ¨der, U.; Wadhawan, J.; Compton, R. G.; Marken, F.; Suarez, P. A. Z.; Consorti, C. S.; de Souza, R. F.; Dupont, J. New J. Chem. 2000, 24, 10091015. (52) Buzzeo, M. C.; Klymenko, O. V.; Wadhawan, J. D.; Hardacre, C.; Seddon, K. R.; Compton, R. G. J. Phys. Chem. A 2003, 107, 8872-8878. (53) Evans, R. G.; Klymenko, O. V.; Saddoughi, S. A.; Hardacre, C.; Compton, R. G. J. Phys. Chem. B 2004, 108, 7878-7886. (54) Nishi, N.; Imakura, S.; Kakiuchi, T. Anal. Chem. 2006, 78, 2726-2731. (55) Luo, H. M.; Dai, S.; Bonnesen, P. V. Anal. Chem. 2004, 76, 2773-2779. (56) Toh, S. L. I.; McFarlane, J.; Tsouris, C.; DePaoli, D. W.; Luo, H.; Dai, S. Solvent Extr. Ion Exch. 2006, 24, 33-56.

electrolysis takes place within the window in wet RTILs.51 Furthermore, [C8mim+][C1C1N-] and other homologous RTILs are not oxidized by oxygen, either.52,53 Disadvantages of RTIL salt bridges include the interference by hydrophobic ions in a sample solution and the finite solubility of the RTIL in the sample solution, for example 1.6 mmol dm-3 in the case of [C8mim+][C1C1N-], which may not be negligible for long-term use even after the optimization of the gelation, particularly when the size of the reference electrode is small. The leaching-out problem would be alleviated by using more hydrophobic RTILs, such as trioctylmethylammonium tetrakis[3,5-bis(trifluoromethyl)phenyl] borate,54 but a trade off is a greater degree of interference by hydrophilic ions in W. As [C8mim+][C1C1N-] and other hydrophobic RTILs are soluble in polar organic solvents, like acetonitrile and methanol, the RTIL-coated Ag/AgCl electrodes in such solvents are of limited use. Aside from the optimization of gelation conditions, the selection of a proper RTIL depending on particular purposes is required for practical applications. ACKNOWLEDGMENT This work was supported by Japan Science and Technology Agency under the program, “Development of Systems and Technology for Advanced Measurement and Analysis”. This work was also partially supported by a Grant-in-Aid for Scientific Research (No. 18350006) and a Grant-in-Aid for Exploratory Research (No. 17655031) from the Ministry of Education, Culture, Sports, Science, and Technology, Japan.

Received for review April 24, 2007. Accepted July 17, 2007. AC070820V

Analytical Chemistry, Vol. 79, No. 18, September 15, 2007

7191