Anal. Chem. 1881, 63,1501-1504
A
[~OO~AI
-
C
1
I
TIME Figure 7. Flow injection peaks for glucose at the glucose oxidase walCjet/thin-layer detector: (A) 6 X lo4 M; (B) increasing concentrations (2-6) X lo4 M (a-c). Sample volume, 10 wL; operating potential, +0.95 V; fbw rate, 0.9 mLlmin; carrier and electrolyte, 0.05 M phosphate buffer (pH 6.5).
(between 5 and 20 rL), after which it starts to level off (B). In addition, the peak decreases slowly as the inlet-electrode distance increases between 250 and 500 Wm (C). A unique situation occurs for a zero carrier flow rate (Figure 6). In accordance with the flow rate independence (reported in Figure 5),such a flow rate yields a peak current similar to that observed at higher flow rates (A versus B and C). However, in the absence of the carrier pushing action, the fast rise in the current is followed by a very slow decay (in a manner analogous to stopped-flow FIA experiments). Repetitive injections of ferrocyanide solutions of increasing concentrations ((1- 6) X 10" M) were used to estimate the linearity and detectability. The response was linear over the entire concentration range (slope, 2.61 pAlmM; intercept, 14 nA; correlation coefficient, 0.999 (not shown)). The signalto-noise characteristics ( S I N = 3 ) of the 1 X M ferrocyanide peak indicated a detection limit of 7 X lo-' M. Analogous measurements of caffeic acid (e.g., Figure 4A(L)) yielded a detection limit of 3 X lo-' M. The placement of the sample injector onto the detector is particularly attractive in connection with active detector surfaces. Highly specific enzyme electrodes, resulting from the judicious immobilization of biocatalysts onto the glassy carbon surface, represent such an opportunity. Figure 7 illustrates the performance of a glucose oxidase based walljet/thin-layer detector. The enzyme electrode responds rapidly and reproducibly to the dynamic changes in the glucose concentration. For the 24 repetitive injections (shown in B) the relative standard deviation is 1.2%. At least 240 glucose measurements can be made every hour at physiolog-
1501
ically relevant concentrations. The enzyme immobilization was accompplished by incorporation within a Kodak-AQ poly(ester-sulfonic acid) film. Because of the additional (protection and germselective) functions offered by AQ polymers (8),such incorporation is extremely attractive for integrated injectorlreactorldetector systems for FIA (which lack solution handling capability). In conclusion, the on-detector injector has been demonstrated to be a very practical device, which addresses the trends toward integration and miniaturization of FIA systems. In addition to its ultrafast (flow-rate independent) response and high sensitivity, such a wall-jet/thin-layer configuration holds promise for performing other FIA procedures. In particular, the versatility and flexibility obtained by introducing the carrier and sample through different inlets (on the detector body) allows additional schemes. These may include on-detector reagent introduction (with minimum dispersion of the product, Figure 1C)or reversed FIA (with on-detector injection of the carrier into the flowing sample solution, Figure 1D). Such schemes should further minimize reagent consumption and enhance the sensitivity, in comparison to analogous FIA systems with separated injector and detector components (1,9). Stopped-flow, flow-reversal, and titrimetric procedures, performed a t the surface, can also be envisioned. The separate control of the sample and carrier introduction should facilitate such schemes and may offer new FIA procedures. As such, the on-detector injection is much more versatile than on-valve detection. In addition, such modification of commercial flow cells offers greater convenience, simplicity, and reliability than the on-valve microelectrode detector ( 4 ) . The concept of on-detector injector should be suitable for other (nonelectrochemical)detection schemes. A direct injection into a large optical (absorption) flow cell has already been reported (10). Registry No. Eastman AQ-55D, 116326-12-6; glucose, 50-99-7.
LITERATURE CITED (1) Ruzicka. J.; Hansen, H. E. Flow In/ection Analysis; Wiiey: New York,
1981. (2) Christian, G.: Krull, I.; Tyson, J. Anal. Chem. 1990, 62, 455A. (3) Canete. F.; Rios. A.; Luque de Castro, M. D.; Valcarcei, M. Anal. Chem. 1988. 60. 2354. (4) Wang, J.; Li, R. Anal. Chem. 1990, 62, 2414. (5) Valcarcel, M.; Luque de Castro, M. D. Anal. Roc. 1989, 26, 313. (6) Kissinger. P. R. J. Chem. Ed. lB83, 60, 308. (7) Fleet, 0.; Little, C. J. J . Chromatogr. Sci. 1974, 12, 747. (8) Wang, J.; Golden, T. Anal. Chem. 1989, 61, 1397. (9) Johnson, K. S.; Petty, R. L. Ami. Chem. 1982, 54, 1185. (10) Eswara Dutt. V. V. S.; Mottola, H. A. Anal. Chem. 1975, 47, 357.
RECEIVED for review December 18,1990. Accepted April 8, 1991.
Application of Scanning Electrochemical Microscopy to GenerationKollection Experiments wlth High Collection Efficiency Chongmok Lee, Juhyoun Kwak, and Fred C. Anson* Arthur Amos Noyes Laboratories, Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, California 91125 The technique of scanning electrochemical microscopy (SECM) was introduced by Bard and co-workers (1)who have utilized i t in a variety of experiments (2). The use of a microelectrode placed close to a substrate electrode to detect
* To whom correspondence should be addressed. 0003-2700/9 110383-150 1$02.50/0
electroactive products generated at the latter, as described by Engstrom and ceworkers (3,4),is one attractive application of the SECM technique (I). In previous studies, this type of generationldetection experiment has usually been carried out with the microelectrode used as the detector electrode (1-5). In the present study, a larger (100 pm diameter) substrate 0 1991 American Chemical Society
1502
ANALYTICAL CHEMISTRY, VOL. 63, NO. 14, JULY 15, 1991 Tip ( Generator Electrode )
ox
A
.
B
Red
4
X INCHWORN LOCATION
/
um
0 50 100 150 X I N C N O R M LOCATION
200 / um
Disk ( Collector Electrode )
Figure 1. Schematic depiction of the electrode and mass-transfer processes occurring in generator/collector experiments with a microtip electrode positioned very close to a macrodisk electrode in a solution containing a redox couple, Ox/Red. I n the experiments of this study, the tip and substrate electrode diameters were 11 and 100 pm, respectively. Both were mounted in glass shrouds.
electrode is used to detect and collect electrode reaction products generated at a microtip electrode positioned at various distances above the substrate electrode. The collection efficiency (in the sense employed for rotating ring-disk electrodes (6))of such an arrangement is virtually 100% at easily achievable separation distances. The advantages of carrying out generation/collection experiments with the present apparatus are exemplified with some simple experimental systems. Comparisons with similar recent experiments in which arrays of microband electrodes are utilized (7,8) are also provided.
EXPERIMENTAL SECTION A carbon microtip electrode with a diameter of 11pm and a platinum macrodisk electrode with a diameter of 100 pm were
fabricated, mounted in glass, and polished according to previously described procedures ( I ) . The carbon and platinum electrodes were employed as the tip and substrate, respectively, in SECM experiments. The miniature electrochemical cell and cell mounting were similar to those described in previous studies (9). The computer-controlled instrument employed in the SECM measurements was based on a previous design (9)with substitution of low-noiseoperational amplifiers in the bipotentiostatto increase the signal-to-noise ratio in the current measurements and a modification in the programs which control the x-y scanning of the tip electrode so that tip current-tip position data could be acquired and stored separatelyfor scans in the forward or reverse direction across the substrate (IO). The separation between the tip and substrate electrodes was determined by comparing the measured and calculated (II ) tip current-separationdistance data ~ ~Fe(CN)63-. + in solutions of R u ( N H ~ )or
RESULTS AND DISCUSSION For experiments directed at maximizing the collection efficiency, i.e., the fraction of a reactant generated at a microtip electrode which diffuses to a nearby substrate electrode to be detected by electrochemical oxidation or reduction, one might expect the largest possible substrate electrode to be desirable. However, residual background currents increase with the area of the substrate electrode and substantial increases in signal-to-noise are obtained by limiting the size of the substrate electrode. We found a ratio of the diameters of the substrate to tip electrode of about 10 to provide both good collection efficiencies and acceptable signal-to-noise levels. To carry out high collection efficiency experiments with substrate electrodes having diameters in the range of 100 pm, it is important to position the microtip electrode above the center of the substrate electrode, as depicted in Figure 1. The necessary positioning of the tip electrode is readily accomplished by employing the SECM instrument to scan the tip electrode across the substrate electrode in solutions containing a reversible, electroactive redox couple according to the pro-
Figure 2. Images of the 100-pm platinum substrate electrode obtained by scanning the 11-pm carbon tip electrode at 10 pm s-I over the substrate in a solution of 5 mM Ru(NH,),~+ in 0.1 M K,SO,. The distance between the substrate and the tip when it was centered over the substrate was 11 pm. The tip and substrate potentials were -0.5 and 0.2 V, respectively. The images in (A) and (C) were obtained from the tip current; those in (B)and (D), from the substrate current. (A) and (B) are cross-sectional views; (C) and (D) are threedimensional views. The tip currents in (A) range from 5.4 to 10.0 nA. The substrate currents in (B)range from -0.7 to -9.4 nA. The tic marks on the vertical axis are separated by 2 nA. When the tip was retracted to a position far (>300 pm) from the substrate, the tip and substrate currents were 6.8 and -0 nA, respectively. (Evaporation of the solution during the course of the experiments (5 h) increased the final tip current to 8.1 nA.)
tocols described by Kwak and Bard (9). In Figure 2 are shown results from some typical SECM scans in a solution of Ru(NH3)63+.The tip was held 11pm above the plane containing the substrate electrode and scanned across a 200 pm by 200 pm region containing the substrate electrode. The potential of the tip electrode was held at -0.5 V vs SCE where the Ru(NH3):+ was continuously reduced to Ru(NH3)2+,and the substrate electrode was maintained at 0.2 V so that any Ru(NH3)2+reaching it was oxidized. Very similar behavior resulted when Fe(CN)$- was employed as the reactant with the tip and substrate potentials set at -0.2 and 0.5 V vs SCE, respectively. In Figure 2A the current at the tip electrode at each value of x is shown for various values of y with R u ( N H ~ ) ~ +as/ ~ + the active redox couple. The same data are displayed in a three-dimensional format in Figure 2C. The position of the substrate electrode is clearly defined, and the values of x and y needed to position the tip above the center of the substrate are easily located. The image benefits from the combination of positive and negative feedback (11,12),which enhances the tip currents when the tip is above the conducting substrate electrode and suppresses the tip current when the tip is above the insulating glass shroud (11). The images of the substrate electrode can also be obtained by measuring the current at the substrate as the tip electrode is scanned. The results of this mode of operation are shown in Figure 2B,D. The substrate current is essentially zero until the tip approaches the substrate electrode, which means that the relative changes in the current at the substrate electrode as the tip passes over it are larger than those at the tip electrode where the reduction of Ru(NH,):+ proceeds at all tip positions. It proved important to limit the scanning rates employed in acquiring the images from substrate currents (Figure 2B,D) to values low enough for the currents measured to approach their steady-state values. For example, images obtained by monitoring substrate currents while the tip was scanned at
ANALYTICAL CHEMISTRY, VOL. 63, NO. 14, JULY 15, 1991 f A Y["
A
\
'. L
C
I
s M
0
58
Y -1
188 160 LQATIW
Tip (Generator)
2-
-- -_ - _ - _-
1-
_
g
288 /
um
-3 0
0
t
1503
20
60
40
Distance I
B
80
100
pm
lm0
i z $ U $ .-a!
0 0
E0 E0
186 186
158 158
288 288
$ 0.8
0
58
188
158
288
Y INZ-WXtl LOCATION / um
Flgure 3. Current contour maps resulting from scans of the 100 pm diameter platinumdisk electrode by SECM. The experimental conditions were the same as those in Figure 2 except that the tip was positioned to be 30 pm above the center of the platinum disk and was scanned at 50 pm s-'. (A, 6) Scan in "forward" and "reverse" direction while recording the tip currents, which vary between ca. 6.8 and 7.0 nA, for the six contours shown; (C, D) Repeat of A and B while recording the disk currents, which vary between -4.5 and -5.7 nA, for
the six contours shown.
50 pm s-* portrayed an apparently inclined surface with the direction of the inclination depending on the direction of the scan. By contrast, the images in Figure 2A,C, obtained with a tip scanning rate of 10 pm s-l, were independent of the scan direction and closely resembled the images obtained from tip current measurements at both scanning rates. The current sensitivity and low noise level of the SECM instrument employed in this study permitted some subtle effects to be observed. For example, in Figure 3 are shown a set of isocurrent contours obtained for the tip and substrate electrodes during scans with a separation distance (30 pm) greater than that employed in Figure 2. The asymmetric shape of the pattern obtained for the tip currents (Figure 3A,B) allowed the direction in which the tip was scanned to be established by inspection. The increased separation between the isocurrent lines that results as the tip is moved from positions over the conducting substrate to positions over the insulating glass shroud results from the interplay between the rate of relaxation of the concentration gradients provoked by the crossing of the conductor/insulation interface and the rate a t which the tip is scanned. The asymmetry is diminished as the scan rate or the separation distance is decreased, as w8s observed in a previous study (12). The absence of corresponding asymmetry in the isocurrent contours for the substrate electrode (Figure 3C,D) shows that the substrate currents relax more rapidly than do those at the tip electrode at the separation distance employed. The difference in relaxation times reflects the way in which the layer between the tip and substrate, which is depleted of reactant while the tip is above the insulating glass shroud, is repopulated with reactant as the tip crosses the glass/platinum boundary. Reactant generated at the substrate must diffuse to the vicinity of the tip to establish a new steady-state concentration profile, and this process proceeds less rapidly than the steady-state profile is generated at the substrate electrode. In attempting to position a microtip electrode over the center of a small substrate electrode it is desirable to avoid conditions (i.e., large separation distances and high scan rates) that give rise to patterns such as those in Figure 3A,B in order to avoid misjudgementsabout the true center of the substrate.
u
E W
0.6
c
-
0.4
0
5
0.2 0.0
0
20
40
Distance
60
80
100
I Fm
Flgure 4. Generationkollection measurements wlth the 11 pm diam
eter carbon-fiber tip electrode positioned above the center of the 100-pm piatinumdisk collector electrode. (A) Tip and disk currents as a function of the distance between the two electrodes. Both currents are normalized with respect to the tip current when the separation distance is large (>300Fm). Tip and disk potentials were -0.2 and 0.5 V v s SCE, respectively. The test solution was 5 mM Fe(CN),> in 0.1 M KCi-0.005 M HCI. (e). Collection efficiencies (Le., disk currentltip current) as a function of separation distance: (1) the system from (A); single (2) or a pair (3)of microband collector electrodes flanking a microband generator electrode (data taken from ref 8). For (2) and (3)the distance is the gap separating the band eiectrodes. Positioning by means of the pattern of substrate current contours appears to be preferable in such circumstances. However, we found it difficult to obtain such clear images if much larger substrate electrodes (e.g., 1mm diameter) were employed because of larger contributions to the substrate current from background or "leakage" currents. Fortunately, it was not necessary to employ larger substrate electrodes in order to obtain high collection efficiencies. Collection Efficiency Measurements. Once the tip (generator) electrode was positioned over the center of the substrate (collector) electrode, it was of interest to determine how the collection efficiency (as measured by the ratio of substrate current to tip current) varied with the distance separating the tip and substrate. In Figure 4A are plotted the tip and substrate currents (normalizedwith respect to the tip current obtained when the separation distance is large enough to render the presence of the substrate electrode negligible) vs the distance between the two electrodes. The solution contained Fe(CN)63-,and the potential of the tip was set at a value where the reduction to Fe(CN)64-proceeded a t a diffusion-controlled rate while the substrate was held at a potential where the Fe(CN)64-was oxidized back to Fe(CN)B. The steady-state current at the tip increased significantly above the constant value obtained at large distances when the separation became less than ca. 10 pm. This behavior is the result of the electrochemical feedback process that has been elucidated in earlier reports (11). The current at the substrate electrode was essentially zero at sufficiently large separations
1504
ANALYTICAL CHEMISTRY, VOL. 63, NO. 14, JULY 15, 1991
(not shown) and increased slowly as the separation decreased until the sharp rise resulting from the electrochemicalfeedback set in and the tip and substrate currents became nearly equal. In Figure 4B are plotted collection efficiencies as a function of the tip-to-substrate separation for a solution of Fe(CN)6”. Attractively high collection efficiencies are obtained. They exceed 0.9 for separations below 20 pm, and even at 100 pm the efficiency is 0.3. Thus, the closely spaced microtip-macrodisk system allows very efficient capture by the disk of species generated at the tip. Geometrical differences between microband arrays and tip-substrate systems will, of course, produce behavioral differences, but it is of interest to compare these two methods of conducting generation/collection experiments with separation distances in the micron range. Curves 2 and 3 in Figure 4B give the collection efficiencies obtained in a recent study with interdigitated microband electrodes (8). Larger efficiencies were obtained from generator band electrodes that were flanked by two collector electrodes (curve 3) than from a side-by-side pair of microband electrodes (curve 2). The microtip-macrodisk system provides not only larger collection efficiencies but also a reasonable range of separations where the collection efficiency is close to unity and insensitive to small changes in the separation. Preliminary results of a digital simulationstudy of the effect of tip geometry and substrate size on collection efficiencies (10) show that collection efficiencies greater than 0.9 are to be anticipated at separation distances comparable to the radius of the tip, even when the substrate and tip have similar radii, so long as the tip is perfectly disk-shaped.
CONCLUSIONS The experimentalsystem described in this study has several attractive features: Unusually large collection efficiencies are easily achievable, which would be valuable in the detection of short-lived intermediates generated a t the tip or in measurements of the kinetics of reactions between species generated at the tip and other reactants present in the solution. Compared with rotating ring-disk or interdigitated microband electrode systems that can be employed for the same purposes, the present system allows a wider range of generator-to-col-
lector distances to be examined by means of the continuously variable micropositioning apparatus. The other techniques would require that a series of electrodes with varying separations be constructed. The fabrication of a macrodisk electrode opposed to adjustable microtip electrode is somewhat simpler than is the construction of ring-disk or interdigitated microband electrode assemblies. The ability to use substrate currents instead of tip currents to obtain images by means of the SECM technique is another advantage of the present system because it permits successful imaging at large separation distances where the changes in tip current are considerably smaller than the changes in substrate current. The system described in this study was constructed with the intention of applying it to studies of ionic transport at the surfaces of electrodes coated with polymers and polyelectrolytes. The high collection efficiencies that have been demonstrated should prove useful in quantitative measurements of electroactive species ejected from such coatings when current is passed through them. The results of these experiments will be described in a forthcoming report.
LITERATURE CITED Bard, A. J.; Fan, FA.; Kwak, J.; Lev, 0.Anal. Chem. 1989, 6 1 , 132. Bard, A. J.; Denault, 0.;Lee, C.; Mandler, D.; Wipf, D. 0. Acc. &em. Res. 1990, 23, 357 and references therein. Engstrom. R. C.; Weaver, M.; Wunder, D. J.; Burgess, R.; Winquist, S. Anal. Chem. 1986. 58, 844. Engstrom, R. C.; Meaney, T.; Tople, R.: Wightman. R. M. Anal. Chem. 1987, 5 9 , 2005. Engstrom, R. C.; Pharr, C. M. Anal. Chem. 198% 6 1 , 1099A. Albely, W. J.; HRchman, M. L. Ring-Disc Electrodes; Clarencon Press: Oxford, U.K., 1971. Bard, A. J.; Crayston, J. A.; Kittlesen, G. P.; Varco Shea, T.; Wrighton, M. S. Anal. Chem. W86, 58, 2321. Bartek. J. E.; Deakin, M. R.; Amatore, C.; Wightman, R. M. Anal. Chem. 1988, 6 0 , 2167. Kwak, J.; Bard, A. J. Anal. Chem. 1989. 61. 1794. Kwak, J.; Anson. F. C. Unpublished work. Kwak, J.; Bard, A. J. Anal. Chem. 1989. 6 1 , 1221. Kwak, J.; Lee, C.; Bard, A. J. J . ElecVochem. SOC.1990, 137, 1481.
RECEIVED for review January 14, 1991. Accepted April 18, 1991. This work was supported by the National Science Foundation. This is Contribution No. 8379 from the Division of Chemistry and Chemical Engineering.
