Anal. Chem. 1998, 70, 3155-3161
Voltammetry at Micropipet Electrodes Yuanhua Shao and Michael V. Mirkin*
Department of Chemistry and Biochemistry, Queens CollegesCUNY, Flushing, New York 11367
The use of micropipet electrodes for quantitative voltammetric measurements of ion-transfer (IT) and electrontransfer (ET) reactions at the interface between two immiscible electrolyte solutions (ITIES) requires knowledge of geometry of the liquid interface. The shape of the meniscus formed at the pipet tip was studied in situ by video microscopy under controlled pressure. The shape of the interface can be changed from a complete sphere to a concave spherical cap by varying the pressure applied to the pipet, and the diffusion current to the pipet changes accordingly. With no external pressure applied, the water/organic interface turned out to be flat, and the voltammetric response of a pipet must follow the wellknown theory for a microdisk electrode. The large deviations from this theory observed previously can be attributed to a small amount of the filling aqueous solution which escapes from the pipet and forms a thin layer on its outer wall. This effect can be eliminated by making the outer pipet wall hydrophobic. Procedures have been developed for independent silanization of the inner and outer walls of the pipet. Pipets with a silanized inner wall can be filled with an organic solvent (e.g., 1,2-dichloroethane) and be used for voltammetric measurements in aqueous solutions. Another mode of voltammetry is based on trapping of a thin layer of organic solvent in the narrow shaft of a pipet between the filling solution and the aqueous outer phase. This arrangement is potentially useful for electrochemical catalysis and sensor applications. After the pioneering work of Taylor and Girault,1 several research groups employed micrometer-sized pipets for voltammetric studies of charge-transfer processes at the interface between two immiscible electrolyte solutions (ITIES).2-7 Com(1) Taylor, G.; Girault, H. H. J. Electroanal. Chem. 1986, 208, 179. (2) (a) Campell, J. A.; Girault, H. H. J. Electroanal. Chem. 1989, 266, 465. (b) Campell, J. A.; Stewart, A. A.; Girault, H. H. J. Chem. Soc., Faraday Trans. 1989, 85, 843. (c) Stewart, A. A.; Shao, Y.; Pereira, C. M.; Girault, H. H. J. Electroanal. Chem. 1991, 305, 135. (d) Shao, Y.; Osborne, M. D.; Girault, H. H. J. Electroanal. Chem. 1991, 319, 101. (e) Shao, Y.; Girault, H. H. J. Electroanal. Chem. 1992, 334, 203. (3) (a) Senda, M.; Kakutani, T.; Osakai, T.; Ohkouchi, T. In Bioelectroanalysis 1. Proceedings of the 1st Bioelectroanalytical Symposium; Pungor, E., Ed.; Akademiai Kiado: Budapest, 1987; p 353. (b) Ohkouchi, T.; Kakutani, T.; Osakai, T.; Senda, M. Anal. Sci. 1991, 7, 371. (c) Vanysek. P.; Hernandez, I. C. J. Electrochem. Soc. 1990, 137, 2763. (d) Vanysek. P.; Hernandez, I. C.; Xu, J. Microchem. J. 1990, 41, 327. (4) Solomon, T.; Bard, A. J. Anal. Chem. 1995, 67, 2787. (5) Cunnane, V. J.; Schiffrin D. J.; Willams, D. E. Electrochim. Acta 1995, 40, 2943. S0003-2700(98)00244-3 CCC: $15.00 Published on Web 06/26/1998
© 1998 American Chemical Society
pared to a large ITIES, a micropipet has a number of advantages typical of ultramicroelectrodes (UMEs), i.e., the absence of the limitations caused by the charging current and ohmic potential drop, relative simplicity of data analysis, and high mass-transfer rate essential for fast kinetic measurements. Unlike metal UMEs, small pipets are easy to make. Pipets as small as a few nanometers radius have recently been prepared and used for quantitative experiments.7a,8 Both micro- and nanopipets can be used as amperometric probes for scanning electrochemical microscopic studies.4,7b Most electrochemical processes occurring at metal electrodes can also be conducted at the liquid/liquid interface. These include electron transfer (ET) and photoinduced ET, electrochemical catalysis, adsorption, and electrodeposition.9 A large group of facilitated and unassisted ion-transfer (IT) reactions9a can be studied at the ITIES but not at metal electrodes. Voltammetry at a micropipet electrode can be a powerful tool for studies of all these processes. In some cases, it may be advantageous to replace a metal UME with a micropipet because the ITIES is highly uniform, is free of surface features, and can be easily modified, e.g., by self-assembly of a molecular monolayer.9,10 Several obstacles have so far prevented a wide use of micropipets as amperometric sensors. Steady-state voltammograms at micropipets are usually of lower quality than those obtained at metal UMEs. Typically, the curves are not fully retraceable, and the diffusion plateau is not flat because of a considerable background current. The background current at a pipet is larger than that at an equivalent size solid electrode because of the larger double-layer capacitance. It is also hard to completely eliminate parallel charge-transfer processes at the ITIES. It was shown recently that high-quality voltammograms can be obtained even at nanometer-sized pipets using background subtraction.7a The background voltammogram to be subtracted is obtained with the same pipet electrode in a solution of the same composition except for the absence of species participating in the charge-transfer reaction. Another problem is the large deviations of the experimental values of diffusion-limiting current to a micropipet from the theory (6) Tokuda, K.; Kitamura, F.; Liao, Y.; Okuwaki, M.; Ohsaka, T. In Charge Transfer at Liquid/Liquid and Liquid/Membrane Interface; Kyoto, 1996; p 7. (7) (a) Shao, Y.; Mirkin, M. V. J. Am. Chem. Soc. 1997, 119, 8103. (b) Shao, Y.; Mirkin, M. V. J. Electroanal. Chem. 1997, 439, 137. (8) Wei, C.; Bard, A. J.; Feldberg, S. W. Anal. Chem. 1997, 69, 4627. (9) (a) Girault, H. H. In Modern Aspects of Electrochemistry; Bockris, J. O’M., Conway, B. E., White, R. E., Eds.; Plenum Press: New York, 1993; Vol. 25, p 1. (b) Liquid-Liquid Interfaces. Theory and Methods; Volkov, A. G., Deamer, D. W., Eds; CRC Press: Boca Raton, FL, 1996. (10) Tsionsky, M.; Bard, A. J.; Mirkin, M. V. J. Am. Chem. Soc. 1997, 119, 10785.
Analytical Chemistry, Vol. 70, No. 15, August 1, 1998 3155
for a disk-shaped microelectrode. Assuming that the pipet orifice is disk-shaped, one can calculate the steady-state diffusion limiting current as
id ) 4nFaDc
(1)
where D is the diffusion coefficient of species in the outer solution responsible for the interfacial charge-transfer reaction, a is the inner pipet radius, F is the Faraday constant, and n is the transferred charge. Equation 1 should be equally applicable to ion-transfer and electron-transfer processes. However, values about 3 times higher than expected from eq 1 were measured for interfacial IT11 and ET.4 Two explanations for this discrepancy have been given previously:11b (i) the ITIES formed at the pipet tip is convex-shaped and has a larger surface area due to bulging of the aqueous phase out of the pipet and (ii) the diffusion current is increased by additional flux from the back of the pipet due to small thickness of its wall (this may account for ∼40% increase in the diffusion-limiting current12). An attempt was recently made to evaluate the shape of the solution meniscus formed at the tip of a micropipet by scanning electrochemical microscopy (SECM).7b The assumption of a convex-shaped ITIES was found to be in conflict with some experimental results. We report here a video microscopic study of the shape of the pipet/solution interface and show that a quantitative agreement can be achieved between the theory and experimental voltammograms at micropipets. Another limitation of the previous voltammetric experiments with pipets is that the filling solution has to be aqueous. This limits the applicability of pipet electrodes to measurements in organic solvents immiscible with water. To do experiments in aqueous solutions, one has to put an organic solvent inside a pipet. Two problems need to be solved here. Because the inner glass wall of the pipet is hydrophilic, the outer aqueous solution tends to penetrate inside the pipet and displace the organic solvent from its narrow shaft. One can observe microscopically a micrometerthick layer of water inside a pipet, near its tip, which impairs amperometric measurements. We describe below the ways to eliminate this effect by silanization of the inner wall. Another concern here is the high ohmic resistance of an organic-filled pipet. However, with a reasonably high concentration of supporting electrolyte in the organic phase, the resistance of short (patchtype) pipets does not greatly affect steady-state voltammetric measurements. We are aware of only one publication describing voltammetry at micropipets filled with an organic solvent (nitrobenzene).3b However, without surface pretreatment,3b those experiments must have been strongly affected by penetration of water inside the pipet. Unlike solid electrodes, the shape of the ITIES can be varied by application of an external pressure to the pipet.13 The combination of pressure control with video microscopic monitoring allowed us to do steady-state voltammetry at convex and concave ITIES and to explore the possibility of trapping a thin layer of an immiscible solvent in the narrow shaft of a pipet (11) (a) Stewart, A. A.; Taylor, G.; Girault, H. H.; McAleer, J. J. Electroanal. Chem. 1990, 296, 491. (b) Beattie, P. D.; Delay, A.; Girault, H. H. J. Electroanal. Chem. 1995, 380, 167. (12) (a) Shoup, D.; Szabo, A. J. Electroanal. Chem. 1984, 160, 27. (b) Fang, Y.; Leddy, J. Anal. Chem. 1995, 67, 1259. (13) Allen, R. M.; Williams, D. E. Faraday Discuss. Chem. Soc. 1996, 104, 281.
3156 Analytical Chemistry, Vol. 70, No. 15, August 1, 1998
between the filling solution and the outer phase. Such an arrangement can be useful for studies of change-transfer processes in liquid membranes and electrochemical catalysis. EXPERIMENTAL SECTION Chemicals. Tetrabutylammonium tetrakis(4-chlorophenyl)borate (TBATPBCl), tetraethylammonium tetrakis(4-chlorophenyl)borate (TEATPBCl), and bis(triphenylphosphoranylidene)ammonium tetraphenylborate (BTPPATPB) were prepared as described previously14 and served as a supporting electrolyte for organic phase. 1,2-Dichloroethane (DCE, 99.8% HPLC grade), dibenzo-18-crown-6 (DB18C6), tetrabutylammonium chloride (TBACl), potassium tetrakis(4-chlorophenyl)borate (KTPBCl), bis(triphenylphosphoranylidene)ammonium chloride, and KCl from Aldrich (Milwaukee, WI) and trimethylchlorosilane (Hu¨ls America, Inc., Bristol, PA) were used as received. All aqueous solutions were prepared from deionized water (Milli-Q, Millipore Corp.). Nitrobenzene (NB, >99%, Fluka Chemika, Switzerland) solutions were washed with a larger volume of Milli-Q water several times before measurements to remove impurities from the organic phase. Instrumentation. A BAS 100B electrochemical workstation (Bioanalytical Systems, West Lafayette, IN) and an EI-400 fourelectrode potentiostat (Ensman Instruments, Bloomington, IN) were employed to record the cyclic voltammograms. An Olympus BH2 optical microscope (×100 to ×1000 magnification) was used to inspect all prepared pipets before measurements. Video microscopic experiments were carried out using a Zoom 70 optical system (Optem, Fairport, NY) equipped with an IK-TU40A 3 CCD color camera (Toshiba, Japan) and an AIGotcha video capture system (AITech International, Fremont, CA). The resistance of the organic-filled pipets was measured with a lock-in amplifier (SRS 850, Stanford Research Systems, Inc., Sunnyvale, CA). Electrodes and Electrochemical Cells. Voltammetric experiments with micropipets were carried out in a two-electrode U-tube cell inside a Faraday cage. A rectangular spectrophotometric cuvette was used as a cell in experiments involving video microscopy. The voltage was applied between two reference electrodes. Both of them were 0.25- or 0.125-mm-diameter Ag wires coated with AgCl (aqueous reference) or AgTPBCl (organic reference). One reference electrode was inserted into a pipet from the back, and another one was immersed into the external solution. All the experiments were carried out at room temperature (25 ( 2 °C). Pipet Preparation. Micropipets were made from borosilicate capillaries (o.d./i.d. ) 1.0/0.58 or 1.0/0.5 mm) from Sutter Instrument Co. (Novato, CA) using a laser-based pipet puller (model P-2000, Sutter Instrument). The shape of a micropipet and the diameter of its orifice were controlled by proper choice of the puller’s five parameters (heat, filament, velocity, delay, and pull).15 Since the ohmic resistance of a pipet is largely determined by the length of the narrow shaft leading to the orifice,7a,11,16 a pulling program was developed to produce short (patch-type) pipets. Both the orifice radius and the thickness of the pipet wall were measured microscopically. The micropipets were filled with (14) Shao, Y.; Girault, H. H. J. Electroanal. Chem. 1990, 282, 59. (15) Model P-2000 Micropipet Puller, Instruction Manual; Sutter Instrument Co.: Novato, CA, 1993. (16) Wei, C.; Bard, A. J.; Nagy, G.; Toth, K. Anal. Chem. 1995, 67, 1346.
either aqueous or organic solution from the back using a small (10-25 µL) syringe. For experiments with an organic solvent inside a pipet, its inner wall was silanized to render it hydrophobic. This was done by putting the back of the pipet in a solution of trimethylchlorosilane, which penetrated inside the pipet, and then pushing it toward the tip by a syringe from the back. The solution was removed from the pipet in about 30 min with a syringe, and the silanized pipet was allowed to dry in the air overnight. In this procedure, only a minor portion of the outer pipet wall is silanized by the vapor emerging from the tip. In contrast, when silanization is done by dipping the pipet tip into trimethylchlorosilane, both the inner and the outer walls become silanized. However, voltammetric responses of organic-filled pipets silanized by these two approaches were practically identical (see below). To silanize only the outer surface of the pipet, the flow of argon was passed through the pipet from the back while its tip was immersed in a solution of trimethylchlorosilane. It is crucial to avoid silanization of the inner wall of a pipet that is going to be filled with an aqueous solution because the outer organic solvent is drawn inside a pipet if its inner surface is hydrophobic. Video-Microscopic and Controlled-Pressure Experiments. The desired shape of the micro-ITIES (i.e., convex, flat, or concave) was obtained by variation of the external pressure applied to the pipet. Both positive and negative pressures were produced by small movements of the plunger of a 1-mL plastic syringe. The syringe was connected to the back of a micropipet with a thin tube, and Torr Seal epoxy (Varian, Lexington, MA) was used to make a seal between a syringe and a tube. The plunger of the syringe was attached to a piezoelectric inchworm motor (Burleigh Instruments, Fishers, NY) controlled by a CE-6000 micropositioning device (Burleigh). Software generously provided by Prof. David O. Wipf (Mississippi State University) was used to control the motion of the inchworm. The shape of the solution meniscus formed at the tip of a pipet was monitored with an Optem Zoom 70 video microscope, allowing more than ×1000 on-screen magnification at a working distance of about 3 cm. To minimize the effect of vibrations on image quality, the whole system was mounted on a vibration-free table (Newport Corp., Fountain Valley, CA). When required, steady-state voltammograms were obtained at a controlled pressure simultaneously with video monitoring. This arrangement was also used for trapping of a thin organic layer inside the narrow shaft of a pipet between two layers of water. A pipet silanized inside and filled with an aqueous solution from the back was immersed into DCE or NB, which was drawn into its narrow shaft by capillary forces. The thickness of the trapped liquid membrane was varied by changing the applied pressure as described above and monitored video microscopically. The pipet with a micrometer-thick organic layer formed near its tip was placed into an aqueous solution for voltammetric measurements. RESULTS AND DISCUSSION Diffusion Current to a Pipet. We have used two extensively studied reactions, i.e., ion transfer of potassium from water to DCE facilitated by dibenzo-18-crown-6 (DB18C6) and simple IT of tetraethylammonium (TEA+) from DCE to the aqueous phase, to probe the origins of a larger diffusion current to micropipets. The first experimental system can be represented by the following cell:
Ag|AgTPBCl|0.25 mM DB18C6 + 0.01 M TBATPBCl|| outer DCE solution 0.01 M KCl|AgCl|Ag (cell 1) pipet The facilitated IT reaction at the water/DCE interface is
K+ (w) + DB18C6 (DCE) a [K+DB18C6] (DCE) (2) The ITIES formed at the pipet tip is polarizable, and the voltage applied between the micropipet and the reference electrode provides the driving force for the IT process. With a concentration of KCl inside a pipet much higher than the concentration of DB18C6 in DCE, the current is limited by diffusion of DB18C6 to the pipet orifice. Steady-state voltammograms were obtained previously for various concentrations of K+ and DB18C611 and various pipet radii (from 5 nm to 15 µM7a), and the plateau current was always about 2.6 times higher than the value found from eq 1. This was noticed earlier by Beattie et al.,11b who proposed an empirical equation (eq 3) for the limiting current at a pipet electrode.
ipip ) 3.35πnFaDc
(3)
The shape of the micro-ITIES can be changed by application of an external pressure to the pipet. The effect of the applied pressure is shown in Figure 1. When a negative pressure was applied, the ITIES shape was concave (Figure 1A). As expected from the theory,17a the diffusion current to a recessed ITIES (see inset in Figure 1A) was lower than that in the absence of negative external pressure (inset in Figure 1B). When a positive pressure was applied to the pipet, the solution meniscus became convex, and the diffusion current increased (Figure 1C). The diffusionlimiting current increased with increasing height of the spherical segment (up to the complete sphere), as the theory predicts.17b Importantly, with no external pressure applied to the pipet, the micro-ITIES was essentially flat (Figure 1B). This observation was corroborated by numerous experiments performed with different concentrations of dissolved species and different pipet radii. Nevertheless, the limiting current to such a flat ITIES was found to be about 2.6 times higher than expected from microdisk theory in accord with eq 3. A significantly larger current at a micropipet may be explained either by another charge-transfer process occurring in parallel with facilitated IT of potassium or by the existence of an additional pathway for delivery of DB18C6 species to the orifice. The first explanation is not plausible because this experimental system has been well-characterized previously.9a Besides, similar values of the ipip/id ratio, between 2 and 3, were obtained for completely different charge-transfer reactions at micropipets. An additional pathway, which may exist for all different charge-transfer processes, is surface diffusion along the outer pipet wall. This implies the presence of a thin (probably submicrometer-thick) layer of aqueous phase, which flows out of the pipet and covers its (17) (a) Ferrigno, R.; Brevet, P.-F.; Girault, H. H. Electrochim. Acta 1997, 42, 1895. (b) Alfred, L. C. R.; Oldham, K. B. J. Phys. Chem. 1996, 100, 2170.
Analytical Chemistry, Vol. 70, No. 15, August 1, 1998
3157
Figure 2. Schematic diagram of facilitated IT of potassium from the aqueous solution inside a micropipet to the outer DCE solution. The transfer occurs by interfacial complexation with DB18C6 (eq 2). The presence of an aqueous layer on the outer pipet wall increases the effective interfacial area available for the IT reaction. The dimensions of this layer are greatly exaggerated in the figure.
Figure 1. Video micrographs of a 15.5-µm-radius micropipet filled with an aqueous KCl solution and immersed in a DCE solution of DB18C6. The solution meniscus is made concave (A) or convex (C) by application of a negative (A) or positive (C) pressure to the pipet. (B) No external pressure was applied to the pipet, and the microITIES is flat. The insets show corresponding steady-state voltammograms of facilitated transfer of potassium. For parameters, see cell 1 in the text.
hydrophilic glass wall (Figure 2). The effective area of the ITIES gets much larger than the geometrical area of the pipet orifice; hence, a significant increase in the diffusion current is observed. To verify this assumption, we compared steady-state voltammograms obtained before and after the outer pipet wall was silanized. Figure 3 shows a video micrograph of a micropipet with a silanized outer wall. The micro-ITIES in Figure 3 is flat and 3158 Analytical Chemistry, Vol. 70, No. 15, August 1, 1998
Figure 3. Video micrograph of a silanized (outer wall only) micropipet filled with an aqueous KCl solution and immersed in a DCE solution of DB18C6. With no external pressure applied to the pipet, the ITIES is essentially flat. a ) 17 µm. The inset shows a corresponding voltammogram of facilitated transfer of potassium. For parameters, see cell 1 in the text.
similar to that shown in Figure 1B. However, the diffusion current decreased markedly after silanization. While the diffusion-limiting current to nonsilanized pipets was always close to ipip values obtained from eq 3, the currents measured at silanized pipets (iexp in Table 1) are much lower. The ratio, iexp/id ) 1.44 ( 0.25 for eight different pipets in Table 1, is in a good agreement with the theoretical prediction of the iexp/id ratio of about 1.4 for b/a ) 1.1 typical for our pipets (b is the radius of glass at the pipet tip).12 A reasonably small scatter of iexp/id values in Table 1 can be explained by uncertainties in optical determination of both a and b values. One should notice that the inner wall of the pipets was not silanized in the above experiments. This excludes any possibility
Table 1. Effect of Silanization on Steady-State Diffusion-Limiting Current of Facilitated Transfer of Potassium at Micropipets in Cell 1 pipet no.
a (µm)
id (nA) (eq 1)
ipip (nA) (eq 3)
iexp (nA) (silanized)
iexp/id
1 2 3 4 5 6 7 8
14.5 21.8 10.0 14.0 7.5 19.5 11.5 13.0
0.73 1.09 0.50 0.70 0.38 0.98 0.58 0.65
1.91 2.88 1.32 1.85 0.99 2.57 1.52 1.72
1.05 1.88 0.68 1.03 0.48 1.20 0.80 1.10
1.44 1.72 1.35 1.42 1.28 1.23 1.39 1.69 Figure 4. Cyclic voltammograms of TEA+ transfer across the water/ DCE interface formed at nonsilanized (A) and silanized (B) micropipets of the same radius (4.5 µm). The potential sweep rate was 100 mV/s. For other parameters, see cell 2 in the text.
Table 2. Effect of Pipet Silanization on Limiting Current of Facilitated Transfer of Potassium into Nitrobenzene pipet i (nA)a ipip (nA) iexp (nA)b id (nA) no. a (µm) (nonsilanized) (eq 3) (silanized) (eq 1) iexp/id 1 2 3 4 5
4.5 6.5 14.5 8.0 11.5
0.50 0.66 1.73 0.85 1.39
0.53 0.76 1.69 0.93 1.34
0.26 0.33 1.02 0.51 0.78
0.20 0.29 0.64 0.36 0.51
1.28 1.16 1.59 1.43 1.52
a Experimental data for nonsilanized pipets. b Experimental data for silanized pipets.
of a change in the geometrical radius of the pipet orifice caused by silanization. A major decrease in the plateau current can be attributed only to elimination of the diffusion flux of ions along the pipet’s outer wall. The existence of a thin aqueous layer on the surface of glass immersed in organic solvent is consistent with the results of previous studies, which showed that a glass-sealed metal UME penetrating the ITIES comes into the organic phase (i.e., nitrobenzene18a or benzonitrile18b) with a micrometer- or submicrometer-thick layer of aqueous electrolyte. To check that the described effect is not related to a specific solvent, we carried out voltammetric measurements of IT of potassium into nitrobenzene (NB) facilitated by DB18C6. The experimental arrangement was similar to that represented by cell 1, except for DCE being replaced by NB and a higher concentration of DB18C6, i.e., 0.5 mM. The data obtained with five pairs of nearly identical micropipets (each pair represents two halves of the same pulled capillary having the same orifice radius; one of them was silanized before measurements) were similar to the results obtained in DCE. The diffusion currents at nonsilanized pipets were much higher than corresponding id values calculated from eq 1 with D ) 2.3 × 10-6 cm2/s.19 The silanization of a pipet produced a significant decrease in the diffusion current (Table 2; i and iexp are the diffusion-limiting currents before and after silanization). In agreement with the theory, the mean iexp/ id value of 1.4 ( 0.2 was obtained for pipets with silanized outer walls. The current values measured before silanization were close to those expected from eq 3, and the least-squares fit gave the numerical factor of 3.32, which is very close to the 3.35 factor in (18) (a) Wei, C.; Bard, A. J.; Mirkin, M. V. J. Phys. Chem. 1995, 99, 16033. (b) Shao, Y.; Mirkin, M. V.; Rusling, J. F. J. Phys. Chem. B 1997, 101, 3202. (19) Kakutani, T.; Nishiwaki, Y.; Osakai, T.; Senda, M. Bull. Chem. Soc. Jpn. 1986, 59, 781.
Table 3. Effect of Silanization on Steady-State Diffusion-Limiting Current of Simple Transfer of TEA+ from DCE into Water-Filled Micropipetsa pipet i (nA)b ipip (nA) iexp (nA)c id (nA) no. a (µm) (nonsilanized) (eq 3) (silanized) (eq 1) iexp/id 1 2 3 4 5
1.5 4.5 6.5 10.0 17.5
0.64 1.60 1.93 4.04 7.54
0.59 1.78 2.58 3.96 6.94
0.31 1.03 1.37 2.41 4.04
0.23 0.68 0.98 1.51 2.64
1.36 1.51 1.40 1.60 1.53
a The diffusion coefficient of TEA in DCE, D ) 9.76 × 10-6 cm2/s, was obtained from the D value measured in NB20 using Walden’s rule. For other parameters, see cell 2 and Figure 4. b Experimental data for nonsilanized pipets. c Experimental data for silanized pipets.
eq 3. This indicates that the effect of thin-layer formation is similar in different solvents. Diffusion currents significantly exceeding the values predicted by eq 1 were also reported for unassisted IT reactions, e.g., the transfer of tetraethylammonium ion across an oil/water interface.1,11 We used the following cell to study the effect of silanization on the transfer of TEA+ across the water/DCE interface:
Ag|AgCl|0.01 M NaTPB + 0.01 M NaCl|| aqueous reference solution 0.4 mM TEATPBCl + 0.01 M BTPPATPB|| outer DCE solution 0.01 M LiCl|AgCl|Ag (cell 2) pipet
where the large interface between the aqueous Na(TPB) solution and DCE was not polarizable because of the common ion (TPB-) present in both phases. A typical voltammogram at a nonsilanized pipet (Figure 4, curve 1) consists of a steady-state wave, corresponding to the transfer of TEA+ from DCE into the aqueous phase inside a pipet, and a non-steady-state reverse peak. The latter is due to the transfer of TEA+ back to DCE, which is limited by quasi-linear diffusion inside a pipet.1,11 The height of the steadystate wave decreased markedly after the pipet outer wall was silanized (Figure 4, curve 2). The limiting steady-state current values for nonsilanized pipets in Table 3 are fairly close to those Analytical Chemistry, Vol. 70, No. 15, August 1, 1998
3159
Table 4. Effect of Silanization on Maximum Voltammetric Current of Simple Transfer of TEA+ from Aqueous Solutions Inside Micropipets into DCEa pipet no.
a (µm)
ip (pA)b (nonsilanized)
ip (pA)c (silanized)
1 2 3 4 5
3.5 7.5 8.0 10.0 15.0
95 252 270 354 794
83 260 273 345 745
a v ) 100 mV/s. See cell 3 for other parameters. b Experimental data for nonsilanized pipets. c Experimental data for silanized pipets.
calculated from eq 3, although that equation was obtained empirically for a completely different process (i.e., facilitated IT11b). The iexp/id values obtained after silanization yield 1.48 ( 0.12, in a good agreement with the theory. From Figure 4, one may conclude that the reverse peak corresponding to the ejection of TEA+ from the pipet also decreases when the outer wall of the pipet is silanized. However, the reverse peak height in Figure 4 depends on the amount of TEA+ accumulated inside a pipet during the forward potential sweep. To simplify the situation, the following cell was used for investigation of the effect of silanization on IT from the pipet into the surrounding solvent:
Ag|AgCl|0.01 M NaTPB + 0.01 M NaCl|| aqueous reference solution 0.01 M BTPPATPB|| outer DCE solution 0.4 mM TEACl + 0.01 M LiCl|AgCl|Ag (cell 3) micropipet The TEA+ ions initially present in the pipet were transferred outside when the pipet potential was swept in a positive direction. Unlike all experiments described above, the peak current in this case was practically unaffected by the silanization procedure (Table 4). This observation is consistent with our model. Clearly, the ion ejection current should be proportional to the crosssectional area of the pipet inner shaft.11 The presence of an aqueous layer on the outer wall does not change the rate of this process. Voltammetry at the Organic-Filled Pipets. So far, we have discussed voltammetry at pipets containing aqueous phase and immersed in organic media. To do experiments in aqueous solutions, one has to put an organic solvent inside a pipet. This can be done by silanizing the inner pipet wall. A typical voltammogram of TEA+ transfer between the outer aqueous solution and DCE solution inside a micropipet,
Ag|AgCl|0.4 mM TEACl + 0.01 M LiCl|| outer aqueous solution 0.01 M TBATPBCl|AgTPBCl|Ag (cell 4) DCE solution in a micropipet is shown in Figure 5. The obtained voltammograms are well 3160 Analytical Chemistry, Vol. 70, No. 15, August 1, 1998
Figure 5. Cyclic voltammogram of TEA+ transfer between DCE inside an 11.5-µm silanized pipet and the outer aqueous solution. The potential sweep rate was 20 mV/s. For other parameters, see cell 4 in the text.
shaped and are suitable for quantitative applications, e.g., mapping of ion-transfer rate through membranes21a and studying of ion binding to DNA.21b The main concern here is a considerable resistance which may impair kinetic measurements with micropipets. The resistance of a typical 10-µm-radius pipet filled with DCE and containing 0.01 M supporting electrolyte was found to be about 10 MΩ, and it is approximately inversely proportional to the pipet radius. Such a resistance produces ∼15 mV iR drop when 0.5 mM TEA+ is transferred from the outer aqueous phase into a 10-µm-radius pipet (the limiting current is ∼1.5 nA according to eq 1). The resistive potential drop can be minimized by using higher concentrations of supporting electrolyte inside a pipet, but the resistance may not be strictly proportional to 1/celectrolyte (as it was observed in water7a,11b) because of ion pairing. To carry out fast kinetic measurements with nanopipets filled with organic solvents, the pulling program has to be improved to fabricate even shorter, less resistive pipets. For the organic-filled pipets with both inner and outer walls silanized, one may expect a current enhancement due to the formation of a layer of organic solvent on the hydrophobic outer wall of a pipet. However, this was not observed experimentally. The ratio, iexp/id ) 1.48 ( 0.3, was obtained with six different micrometer-sized pipets. This number corresponds to the limiting current increased by diffusion from the back, but it does not point to thin-layer formation on the outer wall. The responses of completely silanized pipets and those with only the inner wall silanized were practically indistinguishable. Apparently, the interactions between hydrophilic glass and water are stronger than those between a hydrophobic (silanized) surface and DCE. Trapping a Liquid Membrane inside a Pipet. We noted above that organic solvents get drawn inside a water-filled pipet if its inner wall is silanized. The thickness of the organic phase inside a pipet can be varied by changing the applied external pressure (Figure 6A). When such a pipet is transferred into an aqueous solution, the trapped organic layer acts as a liquid membrane, separating two aqueous phases. Steady-state mea(20) Wandlowski, T.; Marecek, V.; Holub, K.; Samec, Z. J. Phys. Chem. 1989, 93, 8204. (21) (a) Shao, Y.; Mirkin, M. V., unpublished results. (b) Horrocks, B. R.; Mirkin, M. V. J. Phys. Chem., submitted.
processes in membranes, the thin-layer arrangement can potentially be useful for measurements of concentrations in very small (subpicoliter) samples and electrochemical catalysis with a mediator confined to the liquid membrane.
Figure 6. (A) Video micrograph of a 21-µm-radius silanized micropipet filled with an aqueous KCl solution. A micrometer-thick layer of DCE is trapped near the pipet tip under a negative external pressure applied to a pipet. (B) Cyclic voltammogram of facilitated K+ transfer between the outer aqueous solution and DCE layer trapped inside the pipet. The potential sweep rate was 100 mV/s. The filling aqueous solution contained 0.1 M (TBA)Cl; the DCE solution was 0.25 mM in DB18C6 and 0.01 M in (TBATPB)Cl; and the external aqueous solution contained 0.01 M KCl.
surements with a liquid membrane containing either a redox mediator or an ion-complexing agent (Figure 6B) can be carried out similarly to voltammetric experiments at large-area liquid membranes.22 Although the steady-state current of potassium transfer in Figure 6B is well defined, the overall quality of cyclic voltammograms obtained in this configuration is not high. The main problem is extraction of both the mediator and the supporting electrolyte from the trapped thin layer of a solvent. This process occurs on a minute time scale and results in a decrease in the limiting current. The successive voltammograms become more distorted by the growing pipet resistance and the instability of the potential drop across the interface between the trapped layer and the filling aqueous solution. This is also accompanied by a gradual decrease in the layer thickness which can be monitored by the video microscope. Work is in progress to eliminate the depletion effect by loading the membrane with highly hydrophobic species and to stabilize the organic layer thickness by controlling the pressure inside the pipet. Besides studying charge-transfer (22) Shirai, O.; Kihara, S.; Yoshida, Y.; Matsui, M. J. Electroanal. Chem. 1995, 389, 61.
CONCLUSIONS We have used the combination of steady-state voltammetry and video microscopy to investigate the applicability of micropipet electrodes to quantitative electrochemical measurements. The voltammetric response of a pipet filled with an aqueous solution depends strongly on the extent of hydrophilicity of its outer surface. When a glass pipet is immersed in an organic solvent, its hydrophilic outer wall retains a submicrometer-thick aqueous layer. This layer increases the apparent interfacial area and results in diffusion current values more than twice the theoretical prediction based on the pipet radius. This effect is equally important for any charge-transfer process at a micropipet (i.e., simple and facilitated IT and ET) as long as its rate is controlled by diffusion of species in the external solution to the pipet orifice. The aqueous layer can be eliminated by silanizing the outer pipet wall to render it hydrophobic. This procedure allows one to achieve quantitative agreement between the experimental and theoretical values of diffusion current to the pipet. When diffusion inside a pipet is rate-limiting, the current is independent of the nature of the outer wall. A pipet with a silanized inner wall can be filled with an organic solvent and used for voltammetric measurements in aqueous solutions. This technique was recently employed to measure binding constants of cations to DNA21b and to image pores in artificial membranes.21a The diffusion-limiting current in this case agrees well with the theory, regardless of silanization of the outer pipet wall. Thus, no continuous organic layer is formed on a silanized external surface of a DCE-filled pipet. This indicates that the interactions between hydrophilic glass and water are stronger than those between a hydrophobic (silanized) surface and DCE. Unlike solid UMEs, the shape and size of the micro-ITIES can be easily varied by application of an external pressure to a pipet. Video microscopy showed that the ITIES formed at the micropipet tip is flat when no external pressure is applied. The shape of the liquid interface can be changed from a recessed spherical cap inside the micropipet to a complete sphere, and the voltammetric response changes accordingly. One can also trap a micrometeror submicrometer-thick layer of organic phase near the pipet tip between two aqueous solutions. Such a system may be useful for voltammetric studies of charge-transfer processes in liquid membranes and also for electrochemical catalysis and sensor applications. ACKNOWLEDGMENT The support by the Donors of the Petroleum Research Fund administrated by the American Chemical Society and a grant from PSC-CUNY are gratefully acknowledged.
Received for review March 4, 1998. Accepted May 12, 1998. AC980244Q Analytical Chemistry, Vol. 70, No. 15, August 1, 1998
3161
