Scanning Electrochemical Microscopy of Membrane Transport in the

Dec 21, 2000 - Prospects for the Application of Scanning Electrochemical Microscopy in Ionic Liquids. Maurizio Carano , Alan M. Bond. Australian Journ...
0 downloads 7 Views 160KB Size
Anal. Chem. 2001, 73, 533-539

Scanning Electrochemical Microscopy of Membrane Transport in the Reverse Imaging Mode Olivia D. Uitto and Henry S. White*

Department of Chemistry, University of Utah, Salt Lake City, Utah 84112

Scanning electrochemical microscopy (SECM), operated in reverse imaging mode (RIM), has been used to visualize the steady-state transport of molecules entering into porous membranes. RIM imaging is advantageous for investigating transport across biological membranes in situations where the SECM tip can access only the exterior membrane surface. Examples of RIM images of a synthetic membrane (mica with pores filled with the ionselective polymer Nafion) and a biological membrane (hairless mouse skin) recorded during diffusive and iontophoretic transport, are reported. RIM imaging during diffusive transport allows visualization of the depletion of solute molecules in the solution adjacent to the pore openings. However, an accumulation of solute molecules above the pore opening is observed during iontophoresis, which is a consequence of the separation of the solute from the solvent (i.e., ultrafiltration). The separation results from differences in the rates of molecule transfer across the pore/solution interface when electroosmotic flow is operative. The results suggest that RIM imaging may be useful for measuring the kinetics of interfacial molecule transfer at biological membranes. Spatially resolved measurements of molecular transport across porous membranes using scanning electrochemical microscopy1 (SECM) provides a means to identify microscopic structures in a membrane that are associated with molecular transport paths and mechanisms.2 SECM has been used to measure molecular transport rates through individual pores,3-7 and to quantify the individual contributions of diffusion, migration, and convective to the overall flux.4-6,8,9 “Images of molecular flux” across a porous * Phone: 801-585-6256. Fax: 801-585-3207. E-mail: [email protected]. (1) (a) Bard, A. J.; Fan, F.-R. F.; Kwak, J.; Lev, O. Anal. Chem. 1989, 61, 132. (b) Bard, A. J.; Fan, F.-R. F.; Mirkin, M. V. In Electroanalytical Chemistry; Bard, A. J., Ed.; Marcel Dekker: New York, 1994; Vol. 18, p 243. (c) Engstrom, R. C.; Weber, M.; Wunder, D. J.; Burgess, R.; Winquist, S. Anal. Chem. 1986, 58, 844. (d) Engstrom, R. C.; Meaney, T.; Tople, R.; Wightman, R. M. Anal. Chem. 1987, 59, 2005. (e) Engstrom R. C. Anal. Chem. 1984, 56, 890. (2) Bath, B. D.; White, H. S.; Scott, E. R. In Scanning Electrochemical Microscopy; Bard, A. J., Mirken, M., Eds.; John Wiley: New York, in press. (3) Scott, E. R.; White, H. S.; Phipps, J. B. Anal. Chem. 1993, 65, 1537. (4) Bath, B. D.; Lee, R. D.; Scott, E. R.; White, H. S. Anal. Chem. 1998, 70, 1047. (5) Bath, B. D.; White, H. S.; Scott, E. R. Anal. Chem. 2000, 72, 433. (6) Bath, B. D.; White, H. S.; Scott, E. R. J. Pharm. Sci., in press. (7) Bath, B. D.; White, H. S.; Scott, E. R. Pharm. Res. 2000, 17, 471. (8) Scott, E. R.; Laplaza, A. I.; White, H. S.; Phipps, J. B. Pharm. Res. 1993, 10, 1699-1709. 10.1021/ac0009301 CCC: $20.00 Published on Web 12/21/2000

© 2001 American Chemical Society

membrane are now being routinely obtained with micrometer and submicrometer resolution.3,10 This capability has been especially useful in characterizing transport across biological membranes,6,7,11-13 where the molecular flux varies significantly as a function of spatial position. In related studies, SECM has also been successfully applied in investigations of fast ion and electron transfer across air/liquid,14 liquid/liquid15-18 and liquid/lipid bilayer/liquid19-22 interfaces. In investigations of membrane transport, the SECM tip is typically rastered across the membrane surface at a tip-surface separation of a few micrometers, Figure 1. The membrane separates a donor solution, which contains an electroactive species at millimolar or submillimolar concentrations, from a receptor solution in which the redox species is absent. In the conventional mode of SECM operation,2-9 which we refer hereafter to as forward imaging mode or FIM, the SECM tip is placed in the receptor compartment and is poised at a potential such that the electroactive molecule is oxidized or reduced as it emerges from pores in the membrane, Figure 1a. The magnitude of the tip current measured by the SECM tip is proportional to the local concentration of electroactive species above the pore opening. Because the local concentration of the redox species above the pore opening is proportional to the species flux within the membrane,4 a plot of the SECM tip current versus spatial position provides a direct means to visualize transport paths in the membrane. Imaging free-standing membranes using FIM is particularly attractive because this mode of operation has a zero background signal; the redox molecule is not initially present in (9) Scott, E. R.; Phipps, J. B.; White, H. S J. Invest. Dermatol. 1995, 104, 142. (10) Macpherson, J. V.; Unwin, P. R. Anal. Chem. 2000, 72, 276. (11) (a) Unwin, P. R.; Macpherson, J. V.; Beeston, M. A.; Evans, N. J.; Littlewood, D.; Hughes, N. P. Adv. Dent. Res. 1997, 11, 548. (b) Macpherson, J. V.; Beeston, M. A.; Unwin, P. R.; Hughes, N. P.; Littlewood, D. Langmuir 1995, 11, 3959. (c) Macpherson, J. V.; Beeston, M. A.; Unwin, P. R.; Hughes, N. P.; Littlewood, D. J. Chem. Soc., Faraday Trans. 1995, 91, 1407. (12) Macpherson, J.; Unwin, P. R.; Winlove, C. P. Biophys. J. 1997, 73, 2771. (13) Nugues, S.; Denuault, G. J. Electroanal. Chem. 1996, 408, 125. (14) (a) Slevin, C. J.; Ryley, S.; Walton, D. J.; Unwin, P. R. Langmuir 1998, 14, 5331. (b) Barker, A. L.; Macpherson, J. V.; Slevin, C. J.; Unwin, P. R. J. Phys. Chem. B 1998, 102, 1586. (15) Slevin, C. J.; Macpherson, J. V.; Unwin, P. R. J. Phys. Chem. B 1997, 101, 10851. (16) (a) Liu, B.; Mirkin, M. V. J. Am. Chem. Soc. 1999, 121, 8352. (b) Shao, Y.; Mirkin, M. J. Phys. Chem. B 1998, 102, 9915. (17) Tsionsky, M.; Bard, A. J.; Mirkin, M. J. Phys. Chem. 1996, 100, 17881. (18) Wei, C.; Bard, A. J.; Mirkin, M. J. Phys. Chem. 1995, 99, 16033. (19) Amemiya, S.; Ding, Z.; Zhou, J.; Bard, A. J. J. Electroanal. Chem. 2000, 483, 7. (20) Pohl, P.; Saparov, S. M.; Antonenko, Y. N. Biophys. J. 1997, 72, 1711. (21) Antonenko, Y. N.; Pohl, P.; Denisov, G. A. Biophys. J. 1997, 72, 2187. (22) Pohl, P.; Saparov, S. M.; Antonenko, Y. N. Biophys. J. 1998, 75, 1403.

Analytical Chemistry, Vol. 73, No. 3, February 1, 2001 533

Figure 2. Schematic diagram of the iontophoresis cell and scanning electrochemical microscope (SECM) (see text).

membrane/solution interface. Figure 1. Schematic presentation of the two basic imaging modes of SECM (see text): forward imaging mode (FIM) and reverse imaging mode (RIM). The arrows represent the direction of transport of an electroactive molecule, Az+, through a porous membrane. The graph corresponds to the current measured by the tip as it is scanned above the pore in the x direction.

the receptor compartment solution where the signal arises, nor does it accumulate to any appreciable extent during the course of a typical measurement. It is, thus, relatively easy to quantify very small transport rates (∼10-10 mol/s) in individual microscopic pores.5,6 There are many problems in biological chemistry where FIM cannot be used to image membrane transport. For instance, it is not feasible to place the solute permeant and SECM tip on opposite sides of the membrane (as depicted in Figure 1a) to investigate transport of molecules into single cells or to perform transport measurements in vivo. In these situations, it would be clearly advantageous to have both the solute and SECM tip positioned in the solution that is contacting the exterior surface of the membrane or tissue being imaged. We refer to this operation of SECM as reverse imaging mode (RIM): the tip is scanned on the donor side of the membrane in order to observe molecules entering into the membrane, Figure 1b. In RIM, the diffusion of molecules into the pore depletes solute in the donor solution immediately adjacent to the pore entrance, resulting in a decrease in the tip current as the tip is rastered across the area above the pore. Analogous to the increase in tip current measured above a pore in FIM, the decrease in current in RIM signifies a local region in the membrane where the flux is large. In principle, RIM has an inherently lower sensitivity than FIM, because the signal is measured relative to a large background. However, RIM provides a means to image molecular transport into biological membranes, which is not possible using FIM. This is a key advantage for monitoring the uptake of molecular species into single cells and biological tissues. In this report, we describe preliminary studies using RIM to study molecular transport across biological and synthetic membranes. In addition to demonstrating the potential utility of this method for investigating transport rates, we also demonstrate that RIM may be used to directly observe ultrafiltration effects at the 534 Analytical Chemistry, Vol. 73, No. 3, February 1, 2001

EXPERIMENTAL SECTION Chemicals. Acetaminophen (99.9%, Aldrich) was used as received. Ferrocenylmethyltrimethylammonium hexafluorophosphate was prepared by metathesis of the corresponding iodide salt with ammonium hexafluorophosphate (Aldrich, 99.9%) in H2O. The precipitate was dissolved in acetone, recrystallized by the addition of ether, vacuum filtered, and dried in air. Hairless Mouse Skin (HMS). Skin samples were obtained from 7-week-old, male, hairless mice (Charles River, SKH-1) euthanized by CO2 asphyxiation. A thin layer of subcutaneous fat was left on the skin tissue to ensure that no damage was done to the hair follicle structure. The samples were kept under refrigeration between phosphate-buffered saline-filled (Dulbecco’s) gauze pads. The skin was used within 8 h of sacrifice. The research adhered to the “Principles of Laboratory Animal Care” (NIH publication #85-23, revised 1985). Mica/Nafion Membrane. Membranes were prepared as previously described from a 100-µm-thick sheet of mica.5 A circular-shaped pore was generated in the center of a 0.75 × 0.75 in. square sample of mica by laser ablation using a 355-nm line of a Nd:YAG pulsed laser (Spectra Physics). Five 90-mJ pulses of 10 ns duration were used to create a 14.6-µm-radius pore. The pore was then filled with a solution of the cation-selective polymer, Nafion. Two drops (∼40 µL) of a 5 wt % solution of Nafion 117 in alcohol/water (Aldrich, 1100 equivalent weight) were placed over the pore and drawn into the pore by capillary forces. The membrane was allowed to dry and the process was repeated on the reverse side. Scanning Electrochemical Microscopy and the Iontophoresis Cell. The SECM and iontophoresis cell that were employed in these studies have been described in detail.3,4 As depicted in Figure 2, the membrane being investigated separated the upper and lower compartments of the cell, both of which contained an aqueous electrolyte (0.2 M NaCl). In both FIM and RIM, the SECM tip was rastered across the membrane surface that faces the upper compartment. The molecular permeant may be placed in either the lower or upper solution. In the RIM measurements described here, the permeant was placed in the upper compartment and was transported through the pore to the lower compartment. Thus, the upper compartment was defined as the donor compartment.

The SECM tip position was controlled by a piezoelectric inchworm microtranslation stage (model TSE-75, Burleigh Instruments) having a precision of 1 µm. A custom-built potentiostat controlled the tip potential with respect to a Ag/AgCl reference electrode. Tip current was measured with a precision of 2 pA. In iontophoretic transport measurements, a constant current (100, 50, 0, and -100 µA) is applied across the membrane using a galvanostat (model RDE-4, Pine Instruments). A positive iontophoretic current corresponds to migration of cations from the donor to the receptor compartment. Carbon Fiber Tip Fabrication. SECM tips were constructed by inserting a 4-µm-radius carbon fiber into a 5-mm-diameter glass capillary and sealing the junction with epoxy. A tungsten wire was then inserted at the other end of the glass capillary. Contact between the tungsten rod and the carbon fiber was achieved using silver composition paint. Flame etching reduced the radius of the exposed end of the carbon fiber.23 The SECM tips were insulated by electropolymerization of a thin film of poly(oxyphenylene) oxide following the procedure of Kamloth et al.23 The tip was inserted in a 1:1 methanol:water solution containing 2-butoxyethanol (2 wt %), 60 mM 2-allylphenol, and 90 mM phenol and poised at 4 V for 14 min. The film was then cured at 120 °C for 30 min. The carbon fiber tip was cut at the end with a sharp razor blade to expose a disk-shaped area. The radius of the tip, rt, was determined by measuring the voltammetric limiting current in an unstirred aqueous solution containing 5 mM acetaminophen and 0.2 M NaCl. The limiting current is related to rt by eq 1,24

it ) 4nFDC*rt

(1)

where n is the number of electrons transferred, F is Faraday’s constant, and D and C* are the diffusivity (9.1 × 10-6 cm2/s) and bulk concentration of acetaminophen. The average tip radius of the exposed carbon fibers used in these experiments was 1.1 µm, with an insulating layer thickness of 250 nm. RESULTS AND DISCUSSION The Influence of Electroosmotic Flow and Slow Interfacial Molecule Transfer on SECM Images. As will be demonstrated by experimental SECM images, solution flow and slow molecule transfer at the membrane/solution interface can alter the depletion or accumulation of solute molecules at the pore openings, giving rise to a myriad of qualitatively different SECM image shapes. An overview of these factors, and how they impact SECM images, is presented in the following paragraphs. Figure 3 depicts the two imaging modes of SECM (FIM and RIM) and a qualitative representation of the possible solute concentration profiles that are established above a pore opening during diffusive or mixed diffusive/convective transport. We consider only the concentration profile above the membrane surface that is being imaged, which corresponds to the top surfaces of the membranes that are depicted in Figure 3. In FIM (panels a-c on the left-hand side of Figure 3), the solute molecule of interest, Az+, is initially placed in the lower compartment and (23) Potje-Kamloth, K.; Janata J.; Josowicz, M. Phys. Chem. 1989, 93, 1480. (24) Saito, Y. Rev. Polarogr. 1968, 15, 177.

Figure 3. Schematic drawing of six SECM imaging modes (see text) employed to investigate transport of an electroactive molecule, Az+, through a pore. Solid arrows represent diffusion of Az+ due to a concentration gradient. Dashed arrows represent the direction of solution flow due to electroosmotic or mechanical pressure. The predicted concentration profiles, measured by the SECM tip, are shown for each imaging mode. The solid line corresponds to the profile expected for fast molecule partitioning of Az+ between the pore and the donor solution. The dashed lines correspond to slow molecule transfer from the donor solution into the pore.

is transported across the membrane toward the SECM tip, where it is detected. In RIM (panels d-f on the right-hand side), Az+ is initially placed in the upper compartment and is transported across the membrane away from the SECM tip. In both cases, the direction of diffusion that results from the concentration gradient is shown as two solid arrows within the pore. In the absence of convective flow, Az+ is transported across the pore solely by diffusion. As described earlier, the accumulation of Az+ at the pore exit in FIM gives rise to a peak in the SECM tip current, but the depletion of Az+ at the pore entrance in RIM yields a dip in the SECM current. The relative magnitudes of the peaks and valleys, respectively, in FIM and RIM depend largely on the rate of diffusion of Az+ in the pore and the separation distance between the tip and sample. Our laboratory and others have reported peak shape images of membrane transport using FIM in imaging biological and synthetic membranes.2-13 To our knowledge, no reports of SECM imaging using RIM have been previously described. Analytical Chemistry, Vol. 73, No. 3, February 1, 2001

535

Although SECM images may be readily obtained under conditions of passive diffusive transport, the application of forced convective flow across the membrane can greatly enhance the rate of solute transport, thereby increasing the tip current and the overall signal-to-noise ratio of the SECM image. For example, Unwin and co-workers employed the simple method of applying hydraulic pressure to increase the rate of solute transport across thin slices of dentine in order to enhance SECM images of pore transport.12 In our investigations of iontophoretic transdermal drug delivery,6,7 we have been interested in the use of applied currents to control the rate of transport of both ionic and neutral molecules across skin. Because skin is negatively charged,25,26 the applied iontophoretic current results in the electroosmotic flow of solution across the skin, the velocity being determined by the direction and magnitude of the applied current. As with hydraulic-pressure driven flow, the electroosmotic flow can enhance the S/N of the SECM images. Figure 3b,c shows the effect of solvent flow on the concentration profiles above the pore opening during SECM imaging in FIM. Solvent flow in the same direction as spontaneous diffusion, that is, positive solvent flow, denoted by the dashed arrow in the pore, Figure 3b, increases the rate at which Az+ is transported across the membrane, which results in an increase in the concentration of Az+ above the pore opening (solid curve in Figure 3b), and an increased SECM signal. Negative solvent flow, Figure 3c, acts to counter passive diffusion, resulting in a decrease in SECM signal. If the rate of negative solvent flow is sufficiently large, the amount of solute transport across the pore is negligibly small, resulting in the disappearance of the SECM signal altogether. A quantitative treatment of the SECM response in FIM, with both positive and negative electroosmotic flows, as well as inclusion of migrational transport, has been reported by our laboratory.5 In RIM, convective flow has a significantly different effect on the image. When flow is in the direction of diffusion, Figure 3e, the normal depletion of Az+ above the pore opening is reduced, because the flow brings fresh solution containing Az+ at its bulk concentration to the pore opening. In principle, positive flow should result in the concentration of Az+ adjacent to the pore being equal to that of the bulk solution in the donor compartment (solid concentration profile in Figure 3e), resulting in a negligible change in the SECM current above the pore. The pore would be invisible in SECM images under these conditions. Negative solvent flow, Figure 3f, pushes solution from the receptor compartment through the pore. Because Az+ is not present in the receptor compartment solution, the concentration of Az+ at the pore opening is greatly reduced. Thus, negative flow in RIM enhances the relative decrease in the SECM tip current as it is rastered across the pore opening, resulting in a larger signal in the SECM image. In addition to the influence of the solution flow, slow molecule transfer at the interfaces between the pore openings and the solutions can result in qualitative changes in the SECM images. Here, we consider only the effect of slow kinetics of molecule transfer at the interface between the donor solution and the pore (i.e., at the upper surface of the membranes, depicted in Figure 3). A more elaborate treatment would include both interfaces; (25) Burnette, R. R.; Marrero, D. Pharm. Sci. 1986, 75, 738. (26) Burnette, R. R.; Ongpipattanakul, B. J. Pharm. Sci. 1987, 76, 765.

536

Analytical Chemistry, Vol. 73, No. 3, February 1, 2001

however, this is not necessary for the treatment of the experimental images presented below. For the case of passive diffusion during imaging in FIM and RIM, Figure 3a and d, respectively, slow interfacial transfer reduces the rate of solute transport, resulting in a decrease in the SECM tip current. The resulting concentration profiles, shown as dashed lines in Figure 3, would exhibit a smaller peak or valley above the pore opening. The effect of slow interfacial kinetics would be difficult to casually observe in SECM images of passive diffusional transport, because interfacial transfer and transport in the pore are in series.27 The resistance associated with the latter transport process would necessarily have to be known a priori in order to discern the effect of an interfacial resistance. The situation is more interesting in the case in which both convective flow and slow interfacial transfer effect the overall transport. In both FIM and RIM, if the solvent and solute cross the interface at similar rates, then a decrease in the interfacial transfer rates would simply result in a decrease in the tip current. As in the case of diffusion, this kinetic limitation would be difficult to identify in the SECM image and would require a more quantitative study to separate the interfacial and pore resistances. However, if a relative difference exists between the rates of interfacial transfer of the solvent and solute, then this difference may be apparent as a qualitative change in the shape of the SECM image. The situation is shown in Figure 3e for RIM, in which both diffusion and convection occur in the direction away from the tip. As before, for fast interfacial transfer, the flow into the pore results in the diminishment of the concentration profile above the pore opening. However, if solvent transfer is fast and solute transport is slow, then the solute will begin to accumulate above the pore opening. The resulting SECM image will display a peak above the pore in this case (dashed line in Figure 3e), as compared to the normal dip or baseline current expected in the RIM in the absence of the kinetic limitations. Thus, the appearance of the peak shape current above a pore when imaging in RIM is a clear indication that the rate of transfer Az+ across the interface is relative slow in comparison to that of the solvent. The preferential rejection of Az+ at the membrane interface is commonly referred to in the field of chemical separations as ultrafiltration.28 In biological applications, this separation of solvent and solute is important for the understanding of the barrier function of membranes. The ability to observe the ultrafiltration effect in SECM images should depend significantly on the convective flow rate, becoming easier to observe at higher flow rates in which Az+ is brought to the surface at a higher rate. In addition, the ultrafiltration efficiency will depend on the chemical and physical properties of the solute molecules. We show that both predictions are borne out in the SECM experiments presented below. Slow interfacial solute transfer will decrease the SECM signal in FIM when positive solvent flow is operative (Figure 3b), but as in the case of passive diffusion, the effect will be difficult to qualitatively discern. SECM images in either FIM and RIM under (27) The overall mass-transport resitance of the membrane may be modeled as Rm ) Rp + Rentrance, where Rp and Rentrance are the resistances of the pore interior and pore entrance, respectively. Knowledge of Rm and Rp would allow Rentrance to be computed. (28) Giddings, J. C. Unified Separation Science; John Wiley and Sons: New York, 1991.

Figure 4. SECM images (concentration vs position) of acetaminophen above a 14.6-µm-radius pore filled with 1100 equiv wt Nafion as a function of the iontophoretic current, iapp. The tip was scanned 5 µm above the membrane. The donor compartment contained 2 mM acetaminophen; both donor and receptor compartments contained 0.05 M NaCl.

conditions of negative solvent flow are not influenced by slow interfacial kinetics at the pore entrances, (Figure 3c,f), because ultrafiltration cannot occur in these cases. We leave it to the reader to consider these cases in more detail. SECM Images in RIM of Mica/Nafion Membranes. Figure 4 shows RIM images (concentrated vs position) obtained during transport of acetaminophen above a 14.6-µm-radius pore in mica filled with 1100 equiv wt Nafion. The image was obtained by measuring the tip current, it, as a function of position and computing the local concentration, C, using eq 1 (after substituting C for C*). The SECM tip was scanned 5 µm above the membrane and poised at 1.0 V vs Ag/AgCl (3 M NaCl) to detect acetaminophen. The donor compartment contained 2 mM acetaminophen; both donor and receptor compartments contained 0.05 M NaCl. The tip current arises from the irreversible two-electron oxidation of acetaminophen, E° ′ ) 0.6 V vs Ag/AgCl.29 The positive iontophoretic currents (iapp ) 50 and 100 µA in Figure 4) correspond to transport of Na+ from the donor to receptor compartment, in the same direction as the spontaneous diffusion of acetaminophen (corresponding to Figure 3e)). Because the interior of Nafion contains immobilized sulfonate groups, the predominant charge carrier in the pore is Na+. The Na+ flux results in electroosmotic flow through the pore from the donor to the receptor compartment. The images in Figure 4 clearly show an increase in the concentration of acetaminophen above the pore opening under conditions of positive flow at iapp ) 50 and 100 µA, which indicates an accumulation of the solute resulting from ultrafiltration, that is, as the solvent is forced through the pore, the solute is separated from the solvent at the pore/solution (29) Lau, O. W.; Luk, S. F.; Cheung, Y. M. Analyst 1989, 114, 1047.

Figure 5. SECM images (concentration vs position) of FeCp2TMA+ above the same pore as in Figure 4 as a function of the iontophoretic current, iapp. The tip was scanned 5 µm above the membrane. The donor compartment contained 2 mM FeCp2TMA+; both donor and receptor compartments contained 0.05 M NaCl.

interface. Previous SECM studies of acetaminophen transport using FIM clearly demonstrate that this neutral molecule is transported across Nafion.5 Thus, partial rejection of the molecule occurs at the interface. No attempt has been made to quantify the efficiency of this process. However, the separation of solvent and solute in this case is sufficiently large that this phenomenon is clearly observable in the RIM images. The accumulation of acetaminophen above the pore entrance increases when iapp is increased (compare images at iapp ) 50 and 100 µA), which is consistent with a larger flow rate at the larger applied current. The shape of the concentration profile above the pore is determined by the accumulation of the solute, balanced at steady state by back diffusion into the bulk solution. At iapp ) 0 µA, the image is featureless, a consequence of the low current sensitivity that was employed in acquiring the images in Figure 4. As shown below, the same image replotted at higher sensitivity exhibits a decrease in acetaminophen concentration above the pore opening, which is consistent with formation of a depletion layer. The direction of convective flow is reversed in the image obtained at iapp ) -50 µA, resulting in convection opposing the diffusion of acetaminophen. As the solution from the receptor compartment (which does not contain acetaminophen) enters the donor compartment, it forces the donor solution (which contains acetaminophen) away from the pore opening, thereby lowering the local concentration of acetaminophen. Thus, the SECM tip Analytical Chemistry, Vol. 73, No. 3, February 1, 2001

537

Figure 6. Enhanced SECM images of (a) acetaminophen and (b) FeCp2TMA+ (from data in Figures 4 and 5) corresponding to diffusional transport (iapp ) 0 µA).

senses a lower concentration above the pore opening, which is consistent with the prediction depicted in Figure 3f. The RIM images in Figure 4 and all following figures represent steady-state conditions. In previous studies using FIM,4 we have observed that the time that is necessary to obtain a new steadystate image following a step change in iapp is ca. 10 min. Because iapp is controlled by the galvanostatic circuitry, the iontophoretic current in the pore and, thus, the electroosmotic solvent flow, must be established very rapidly following any change in iapp. It follows that the rather slow transient response reflects the establishment of new steady-state concentration profiles within the Nafion-filled pore following the change in iapp. Qualitatively similar RIM images, Figure 5, were obtained of the mica/Nafion pore during the transport of 2 mM FeCp2TMA+. This molecule undergoes a reversible 1 -e- oxidation at E° ′ ) 0.40 V vs Ag/AgCl and is readily detected at the SECM tip. As apparent in the images in Figure 5, a larger amount of FeCp2TMA+ is accumulated at the pore opening during iontophoresis at 50 and 100 µA than was previously observed for acetaminophen. Interesting, the equilibrium partition coefficient of FeCp2TMA+ between Nafion and the aqueous solution is approximately 2000 larger than that of acetaminophen,5 which is due to the strong electrostatic binding between the sulfonate groups in Nafion and the positively charged FeCp2TMA+. The SECM images indicate that this electrostatic force is considerably less important in determining the rate of partitioning, at least under the iontophoretic conditions employed here. The transport of FeCp2TMA+ across the pore also contains a migrational component that tends to force FeCp2TMA+ through the pore at a larger rate than the uncharged acetaminophen. This enhancement in driving force is responsible, in part, for the larger accumulation of FeCp2TMA+. RIM images under conditions of diffusive transport, that is, iapp ) 0 µA, are replotted in Figure 6 at a higher sensitivity in order to see the depletion layer formed at the pore opening. The 538 Analytical Chemistry, Vol. 73, No. 3, February 1, 2001

Figure 7. SECM images (concentration vs position) of acetaminophen above a 16.2-µm-radius hair follicle in hairless mouse skin as a function of the iontophoretic current, iapp. The tip was scanned 9 µm above the skin surface. The donor compartment contained 5 mM acetaminophen; both donor and receptor compartments contained 0.05 M NaCl.

general shape of the images is in agreement with expectations of RIM imaging, as discussed above. Qualitatively, the dip in the concentration profile for acetaminophen is ∼2 larger that that for FeCp2TMA+, which suggests that the net diffusive flux of the acetaminophen in the Nafion-filled pore is larger than that of FeCp2TMA+. This difference is consistent with previous quantitative SECM measurements of the diffusive fluxes of these molecules in Nafion using FIM.5 The concentration profiles above the circular-shaped pore are expected to be circularly symmetric. The asymmetry apparent in the images presented in Figure 6 is due to fluid convection induced by rastering the tip over the pore. RIM Images of Hairless Mouse Skin. In recent investigations, our laboratory has used SECM to quantify iontophoretic transport in individual hair follicles of hairless mouse skin (HMS). These previous investigations, based on FIM images, have demonstrated that electroosmotic transport of small molecules and ions occurs predominantly through hair follicles in HMS. Skin carries a net negative charge at physiological pH,25,26 and thus, analogous to the mica/Nafion membrane described above, electroosmotic transport occurs in the same direction as cation migration. The negative charge on skin has been suggested to be the result of a larger number of carboxylate than ammonium groups associated with amino acid residues. Changes in pH, thus, affect the net charge on skin and, subsequently, the electroosmotic

noisier tip.) Analogous to the mica/Nafion pore, the RIM images demonstrate an accumulation of acetaminophen above the pore opening during positive solvent flow, indicating that acetaminophen is selectively rejected, relative to the solvent, by the hair follicle. As with the mica/Nafion pore, rejection of acetaminophen is only partial, because FIM images clearly indicate that acetaminophen is transported through the hair follicle.6

Figure 8. Enhanced SECM image of acetaminophen (from data in Figure 7) corresponding to diffusional transport (iapp ) 0 µA).

flow characteristics of the tissue, as shown recently for electroosmotic transport of a neutral molecule within individual hair follicles.7 RIM images of a 16.2-µm-radius30 hair follicle in HMS were measured as a function of the iontophoretic current, iapp. The tip was scanned ∼9 µm above the skin surface. The donor compartment contained 5 mM acetaminophen; both donor and receptor compartments contained 0.05 M NaCl. Figure 7 shows the resulting images acquired at iapp ) 100, 50, 0, and -100 µA, and Figure 8 shows an enhanced image of the pore at iapp ) 0 µA. (These images are noisier than those in Figures 4-6 due to a (30) The radius of the hair follicle was measured by tip retraction experiments in FIM using the methodology described in refs 5 and 6.

CONCLUSION The RIM images presented here demonstrate that SECM can be employed to investigate the transport of molecules into individual pores of porous membranes. RIM imaging will likely be most useful for investigating transport across biological membranes in situations where the SECM tip can access only the exterior membrane surface. Although RIM has an inherently lower sensitivity than FIM, our preliminary studies indicate that RIM images can be obtained without significant loss of signal. In addition, the shape of RIM images is qualitatively sensitive to the rate of interfacial molecule transfer. This feature should be useful in identifying and quantifying kinetically slow steps in membrane transport. ACKNOWLEDGMENT The authors gratefully acknowledge the assistance of Dr. B. D. Bath and Ms. Melissa Mitchell in the preliminary stages of this work. We also appreciate stimulating discussions with Dr. E. R. Scott and Dr. Bath about potential applications of SECM operated in RIM. This research was supported by ALZA, Inc. (Palo Alto, CA). Received for review August 7, 2000. Accepted November 1, 2000. AC0009301

Analytical Chemistry, Vol. 73, No. 3, February 1, 2001

539