Langmuir 2003, 19, 9781-9791
9781
Molecular Recognition at Model Organic Interfaces: Electrochemical Discrimination Using Self-Assembled Monolayers (SAMs) Modified via the Fusion of Phospholipid Vesicles Mariusz Twardowski and Ralph G. Nuzzo* Department of Chemistry and the Frederick Seitz Materials Research Laboratory, University of Illinois at Urbana-Champaign, Urbana, Illinois 61801 Received May 25, 2003. In Final Form: August 7, 2003 Supported lipid layers were formed via the fusion of large unilamellar vesicles of 1,2-dimyristoyl-snglycero-3-phosphocholine (DMPC) to mixed self-assembled monolayers (SAMs) on gold comprised of ferrocene-functionalized hexadecanethiol chains (FcCO2C16SH) diluted in either hexadecanethiol (C16SH) or 1-hydroxylhexadecanethiol (HOC16SH). For the former case, the DMPC adsorbs predominantly as a single layer to form a hybrid bilayer membrane (HBM). The structures obtained in this way were characterized by methods that include electrochemical measurements, ellipsometry, and surface plasmon resonance (SPR). Cyclic voltammetry (CV) reveals that the electrochemistry of the ferrocene groups present in the SAM is strongly perturbed by the adlayer structure. The electrochemical behaviors of the ferrocene groups incorporated into a mixed SAM prepared using the more polar hydroxyl terminated thiol are quite different. The adsorption of DMPC via vesicle fusion in this case leads to the adsorption of bilayer assemblies of the lipid on top of the SAM. The coverages of the DMPC suggested by the SPR data lie between the values expected for fusion processes depositing either one or two bilayers of the lipid on top of the SAM. The electrochemical properties of the ferrocene moieties present in this structure were found to be largely unperturbed following the DMPC adsorption. Subsequent studies revealed that the adsorbed DMPC strongly influences the interactions of the tethered ferrocene groups with secondary aqueous molecular redox probes present in the electrolyte solution; permselective properties are seen in this adlayer structure. The varying degrees of electrochemical rectification seen in CV surveys demonstrated that probes such as K4Fe(CN)6, C5H5Fe(C5H4CH2N(CH3)3)PF6, and Ru(NH3)6Cl3 appear to penetrate the DMPC layer while species such as C5H5Fe[C5H4CH2N+H(CH3)2], C5H5Fe(C5H4CH2OH), and C5H5Fe(C5H4COO-) do not. We believe that molecular scale defect structures present in the adsorbed DMPC layers confer the molecular discrimination properties seen. A qualitative structural model is proposed.
Introduction An important theme of research in surface and interface science is one directed toward the development of molecular level understandings of interactions and processes occurring at solid-liquid interfaces.1 Interfacial interactions of this sort play crucial roles in such areas as adhesion,2 wetting,3,4 electrochemistry,5 and catalysis6,7 and critically underpin the functioning of numerous sensor technologies8 and the biological systems that support life.9,10 SAMs11,12 of alkanethiols are a valuable tool for * Corresponding author. Phone: (217) 244-0809. Fax: (217) 2442278. E-mail:
[email protected]. (1) Swalen, J. D.; Allara, D. L.; Andrade, J. D.; Chandross, E. A.; Garoff, S.; Israelachvil, J.; McCarthy, T. J.; Murry, R.; Pease, R. F.; Rabolt, J. F.; Wynne, K. J.; Yu, H. Langmuir 1987, 3, 932-950. (2) Kaelble, D. H. Physical Chemistry of Adhesion; Wiley-Interscience: New York, 1971. (3) Kloubek, J. Adv. Colloid Interface Sci. 1992, 38, 99-142. (4) Bracke, M.; DeBisschop, F.; Joos, P. Prog. Colloid Polym. Sci. 1988, 76, 251. (5) Murray, R. W. In Molecular Design of Electrode Surfaces; Saunders, J., William, H., Ed.; John Wiley & Sons: New York, 1992; Vol. 22. (6) Zak, J.; Yuan, H.; Ho, M.; Woo, L. K.; Porter, M. D. Langmuir 1993, 9, 2772-2774. (7) Willner, I.; Blonder, R.; Katz, E.; Stocker, A.; Bueckmann, A. F. J. Am. Chem. Soc. 1996, 118, 4717-4718. (8) Janata, J. Principles of Chemical Sensors; Plenum Press: New York, 1989. (9) Maget-Dana, R. Biochim. Biophys. Acta 1999, 1462, 109-140. (10) Kornberg, R. D.; McConnell, H. M. Biochemistry 1971, 10, 1111. (11) Nuzzo, R. G.; Allara, D. L. J. Am. Chem. Soc. 1983, 105, 4481. (12) Dubois, L. H.; Nuzzo, R. G. Annu. Rev. Phys. Chem. 1992, 43, 437-63.
modeling such phenomena because of their ease of preparation, stability, highly ordered structure,12-15 and the diversity of the compatible chemistries that can be utilized to covalently incorporate a wide range of functionalities into them.16,17 The study of molecular recognition (which concerns the selective interaction or noncovalent binding of a probe molecule with a receptor) is well suited for modeling with SAMs. The receptors in many natural systems frequently are large, flexible, structurally complex molecules that present many potential binding sites for a ligand or other chemical species.18-20 For this reason, the systematic study of mechanisms underlying their selective binding of a target molecular species is very challenging. By covalently incorporating simpler models of receptors into well-ordered thiol SAMs, the study of these interactions is simplified and a large host of highly sensitive analytical techniques have been developed for studies directed toward this end. (13) Finklea, H. O. Electroanal. Chem. 1996, 19, 109-335. (14) Ulman, A. Chem. Rev. 1996, 96, 1533-1554. (15) Schreiber, F. Prog. Surf. Sci. 2000, 65, 151-256. (16) Chechik, V.; Crooks, R. M.; Stirling, C. J. M. Adv. Mater. 2000, 12, 1161-71. (17) Laibinis, P. E.; Palmer, B. J.; Lee, S.-K.; Jennings, G. K. In The Synthesis of Organothiols and Their Assembly into Monolayers on Gold; Ulman, A., Ed.; Academic Press: New York, 1998; Vol. 24, pp 2-36. (18) Mammen, M.; Choi, S. K.; Whitesides, G. M. Angew. Chem., Int. Ed. Engl. 1998, 37, 2755-2794. (19) Payne, J. W.; Marshall, N. J.; Grail, B. M.; Gupta, S. Curr. Org. Chem. 2002, 6, 1221-1246. (20) Fulton, D. A.; Cantrill, S. J.; Stoddart, J. F. J. Org. Chem. 2002, 67, 7968-7981.
10.1021/la0349018 CCC: $25.00 © 2003 American Chemical Society Published on Web 10/09/2003
9782
Langmuir, Vol. 19, No. 23, 2003
These include such methods as SPR,21 electrochemistry,13 reflection absorbance infrared spectroscopy (RAIRS),22 and binding measurements made using mass-sensitive methods such as the quartz-crystal microbalance.23 Thiols functionalized with DNA,24 barbituric acid receptors,25 cyclodextrins,26 cyclophanes,27 and calixerenes28,29 also have been synthesized and applied in SAM-based forms to study very specific “host-guest” interactions. When taken together, these representative examples illustrate the vital progress being made in this area of research. An alternative approach to the development of surfaces capable of effecting molecular discrimination involves the intentional incorporation of nanoporous defects into SAMs. This may be done simply, for example, through the coadsorption of short-chain, defect-forming thiols along with longer chain thiol adsorbates that are known to form well-packed, highly ordered monolayers.30,31 Such SAMs have shown some selectivity in their interactions with a variety of redox probes in solution, discriminating between them on the basis of both the dimensions of their hydrated radii and charge. In an interesting extension of this work, polyamidoamine (PAMAM) dendrimers coadsorbed with an alkanethiol SAM were found to act as pH sensitive molecular gates for charged solution redox probes.32 In this report, we describe a novel SAM-based electrochemical system that exhibits selectivity in its interactions with a variety of aqueous redox probes. The mechanisms that support this functional behavior involve aspects of both of the aforementioned strategies for modifying electrode surfaces. We specifically show that a mixed SAM composed of a ferrocene-functionalized thiol diluted in either alkane- or hydroxythiol can have substantially altered electrochemical properties upon adsorption of 1,2dimyristoyl-sn-glycero-3-phosphocholine (DMPC) vesicles. By varying the diluent thiol to form either a hydrophobic or hydrophilic interface in solution, the DMPC adsorption can occur through the binding of a single lipid leaflet either in the form of a hybrid bilayer membrane (HBM) or in the form of supported bilayers, respectively.33,34 The effect of the DMPC in the HBM is one of largely blocking the reversible electrochemistry of the ferrocene moiety immobilized in the nonpolar hydrocarbon environment of this structure. The ferrocene units immobilized in the polar environment of the multi-bilayer structure, however, retain their reversible electrochemistry. This observation clearly establishes that the DMPC layers of the latter assembly do not block (or at least not catastrophically so) the correlated motions of the anions that are needed to (21) Ostuni, E.; Yan, L.; Whitesides, G. M. Colloids Surf., B 1999, 15, 3-30. (22) Laibinis, P. E.; Bain, C. D.; Nuzzo, R. G.; Whitesides, G. M. J. Phys. Chem. 1995, 99, 7663-76. (23) Patolsky, F.; Lichtenstin, A.; Willner, I. Nat. Biotechnol. 2001, 19, 253-257. (24) Brockman, J. M.; Frutos, A. G.; Corn, R. M. J. Am. Chem. Soc. 1999, 121, 8044-8051. (25) Motesharei, K.; Myles, D. C. J. Am. Chem. Soc. 1998, 120, 73287336. (26) Rojas, M. T.; Ko¨niger, R.; Stoddart, J. F.; Kaifer, A. E. J. Am. Chem. Soc. 1995, 117, 336. (27) Kaifer, A. E. Acc. Chem. Res. 1999, 32, 62-71. (28) Thoden van Velzen, E. U.; Engbersen, J. F. J.; de Lange, P. J.; Mahy, J. W. G.; Reinhoudt, D. N. J. Am. Chem. Soc. 1995, 117, 68536862. (29) Faull, J. D.; Gupta, V. K. Langmuir 2001, 17, 1470-1476. (30) Chailapakul, O.; Crooks, R. M. Langmuir 1993, 9, 884-888. (31) Chailapakul, O.; Crooks, R. M. Langmuir 1995, 11, 1329-1340. (32) Zhao, M.; Tokuhisa, H.; Crooks, R. M. Angew. Chem., Int. Ed. Engl. 1997, 36, 2596-2598. (33) Lingler, S.; Rubinstein, I.; Knoll, W.; Offenhauser, A. Langmuir 1997, 13, 7085-7091. (34) Jenkins, A. T. A.; Bushby, R. J.; Evans, S. D.; Knoll, W.; Offenhauser, A.; Ogier, S. D. Langmuir 2002, 18, 3176-3180.
Twardowski and Nuzzo
support the one electron oxidation and reduction processes of the chain-end probes. In contrast, though, other redox species present in the electrolyte solution show varying abilities to permeate the DMPC adlayer. These effects allow a number of new behaviors not seen with the SAM alone, ones that include chemical rectification that is innately coupled to the permselective properties afforded by the DMPC. In either case, the tethered ferrocene’s electrochemical activity is significantly impacted. This ability to tailor the nanoscale environment of a SAMtethered redox functional group appears to directly complement the results of research carried out using redox molecules encapsulated within the cores of dendrimers.35,36 Unlike these systems, however, the modification of the SAM via DMPC adsorption is completely reversible, and the original properties of the supporting SAM can be restored by simply washing the adsorbed phospholipid away with ethanol. The origins of the novel properties we observe appear to relate to structural features of the adsorbed lipid assembly that are quite subtle in nature. Inferences about possible structural models of these “permselective defects” can be gleaned from the known structural features of supported bilayers described in the literature. Experimental Section Chemicals. Hexadecanethiol (C16SH), hydroxymethylferrocene (HMFc), (dimethylaminomethyl)ferrocene (FcNH+), ferrocene carboxylic acid (FcCOO-), tris(2,2′-bipyridyl)dichlororuthenium(II) hexahydrate, hexaamineruthenium(III) chloride, sodium perchlorate, and potassium ferrocyanide were obtained from Aldrich Chemical (Milwaukee, WI) and used without further purification. The compound N,N-dimethylaminomethylferrocene methiodide was obtained from Alfa Aesar. It was used either as obtained or in the form of its hexafluorophosphate salt. The latter compound was prepared according to literature procedures by mixing an aqueous solution of it with a saturated solution of NH4PF6.37,38 The precipitate, dimethylaminomethylferrocene hexafluorophosphate (FcNCH3+), was filtered, washed, and dried in vacuo. DMPC was obtained as a solution (10 mg/mL) in chloroform from Avanti Polar Lipids (Alabaster, AL). Reagent grade HCl, HNO3, and HClO4 were obtained from Fisher Scientific (Pittsburgh, PA). The thiol 16-mercapto-1-hexadecanol (HOC16SH) was obtained as a generous gift from Professor D. L. Allara (Pennsylvania State University, University Park, PA). The C16 oligomethylene chain tethered ferrocene, FcCO2C16SH [Fc ) (η5-C5H5)Fe(η5-C5H4)], was synthesized by the method of Chidsey et al.39 and characterized by 1H NMR and mass spectrometry. Unless otherwise noted, all solvents utilized were of commercial grade. Deionized water was further purified with a Milli-Q (Waters Associates) purification system. Gold Substrate Preparation. Glass microscope slides (Fisher Scientific) were cleaned in a hot piranha solution (3:1 concentrated H2SO4:30% H2O2). Caution: piranha solution reacts violently with organic matter and should be handled with extreme care! The slides were rinsed copiously with water and then methanol, subsequently dried briefly in a 130 °C oven, and used immediately. Gold films were grown on the clean glass slides by DC sputtering using an AJA International A320 MultiGun Sputtering System. The sputtering was done under an Ar atmosphere at an operating pressure of 2 mTorr. A 1-5 Å layer of Ti was sputtered initially to promote adhesion, followed by 1800-2000 Å of Au. For substrates used in electrochemical studies, the sputtering was carried out using a Teflon shadow mask to define circular gold electrodes with an area of 0.283 cm2 and a 1 mm thin line extending from their perimeter to provide electrical contact in the cell. This design (35) Cardona, C. M.; Mendoza, S.; Kaifer, A. E. Chem. Soc. Rev. 2000, 29, 37. (36) Hecht, S.; Frechet, J. M. J. Angew. Chem., Int. Ed. 2001, 40, 74. (37) Hauser, C.; Lindsay, J. J. Org. Chem. 1956, 21, 382. (38) Lindsay, J.; Hauser, C. J. Org. Chem. 1957, 22, 355. (39) Chidsey, C. E. D.; Bertozzi, C. R.; Putvinski, T. M.; Mujsce, A. M. J. Am. Chem. Soc. 1990, 112, 4301-4306.
Molecular Recognition at Model Organic Interfaces minimized mechanical deformation of the gold electrode by the O-rings used to seal it in the cell. We had previously found that a specific pretreatment of the gold films dramatically decreased the density of electrochemically active defects present in the monolayers grown on them, and this procedure was used in all the studies reported here.40 In a typical procedure, the Au substrates were dipped for 2 min in piranha solution held at 35 °C, rinsed copiously with water, dried with filtered air or N2, and then immersed for 2 min in a 3:1:16 mixture of HCl/HNO3/H2O. Following this, the slide was rinsed with water and ethanol, dried with filtered N2, and immediately immersed in a thiol solution to form a SAM. Monolayer and Hybrid Multilayer Preparation. The treated gold substrates were exposed to 2 mM solutions of FcCO2C16SH/C16SH or FcCO2C16SH/HOC16SH of varying ratios in ethanol for times of 12-24 h. The SAM-covered substrates were then cleaned with hexanes, ethanol, and water. For electrochemical measurements, the gold electrode was in the cell prior to the SAM formation step. Large unilamellar vesicles of DMPC were prepared as described below. A volume of DMPC/chloroform solution containing 1 mg of DMPC was roughly dried using a flowing stream of N2, and then it was subsequently dried in vacuo for an additional 8-16 h. The dried film was hydrated and resuspended in 2 mL of 175 mM NaClO4 (pH ) 7, 10 mM phosphate buffer) for 30 min with periodic vortexing. The lipid solution was then subjected to five freeze/thaw cycles using liquid nitrogen and a warm water bath. This acts to disrupt the vesicles and improve the homogeneity of the size distribution present in the final solution. The vesicle solution was extruded using a bench top syringe extruder (Avanti Polar Lipids, Alabaster, AL) with 100 nm Nucleopore Track-Etch membranes (Corning Separations Div., Cambridge, MA) for a total of 11 times. In each case the extruded vesicle solutions were used within 24 h of their preparation. For the vesicle adsorption studies, the SAM-covered gold substrates were exposed to the vesicle solution and heated to 30-35 °C, a value chosen such that the DMPC was above its gel transition temperature of 23 °C. The typical time scale for exposure was 8-24 h. For supported lipid monolayers, the substrates were then soaked in water or a dilute saline solution to remove excess vesicles. Substrates for electrochemical studies were rinsed in place, taking care not to expose the treated substrates to air at any point. Ellipsometry. Ellipsometric measurements were made on emersed samples using a Gaertner Scientific (Chicago, IL) model L116C ellipsometer equipped with a He-Ne laser (λ ) 6328 Å) set at an incidence angle of 70°. Measurements were taken at four or five spots on the substrate surface and averaged. A twolayer transparent film model was used for the thickness calculations based on pseudosubstrate constants measured on the clean substrate. The refractive index of the organic film was fixed at 1.5. Surface Plasmon Resonance (SPR). SPR measurements were made utilizing an apparatus based on the Kretchman configuration41 whose design has been described elsewhere.42 A small-volume flow cell (approximately 0.25 mL in volume) was used for the adsorption measurements. A system of three-way valves was used to inject buffer or vesicle solution to the substrate. Substrates for SPR were prepared as described previously, but using thicknesses of 15 Å for a Cr adhesion layer and 150 Å for the Au film. Electrochemistry. The electrochemistry cell consisted of a hollow Teflon cone pressed against the glass bearing the gold electrode. The orifice defined a working cell area of 0.71 cm2. Capacitance and CV data were measured in 175 mM NaClO4 or KCl electrolyte. The CV measurements were typically performed immediately after adding the solution containing a soluble probe, and subsequent scans showed that the electrochemical properties of the mobilized SAMs reported here remained stable for a period of 10 min or longer. Extended incubation times in electrolyte (40) Twardowski, M.; Nuzzo, R. G. Langmuir 2002, 18, 5529-5538. (41) Davies, J. Surface Plasmon ResonancesTheory and Experimental Considerations; CRC Press: New York, 1996. (42) Lavrik, N.; Leckband, D. Langmuir 2000, 16, 1842-1851.
Langmuir, Vol. 19, No. 23, 2003 9783 were not studied due to the noted deleterious effects of the perchlorate anion on gold substrates. A Ag/AgCl/saturated KCl reference electrode and platinum wire counter electrode were used. A RDE4 potentiostat (Pine Instruments Co., Grove City, PA) interfaced to a personal computer with LabVIEW software (National Instruments, Austin, TX) was used to perform and record cyclic voltammograms. Capacitance (C) values were calculated using the formula43
C ) I/AS where I is half of the difference in current between the forward and reverse scans at a given potential (200 mV in this case), A is the electrode surface area, and S is the scan rate. Atomic Force Microscopy (AFM). AFM micrographs were obtained using a PicoSPM 300 (Molecular Imaging, Tempe, AZ) instrument controlled with a Nanoscope E controller (Digital Instruments, Santa Barbara, CA). Images were acquired in the so-called MAC mode, which has been described in detail elsewhere.44 This noncontact measurement method has proven to be well-suited for imaging the fragile structures of interest in this work.45 The spring constant of the Si cantilever was 2.8 N/m. For the AFM experiments, atomically flat Au on mica substrates (SPI Supplies, West Chester, PA) was used to improve the visualization of the lipid layer.
Results Figure 1a presents a typical cyclic voltammogram measured for a mixed thiol monolayer of FcCO2C16SH/ C16SH on gold having a ferrocene surface coverage (as calculated using a literature protocol)39 of 1.06 × 1014 cm-2. As has been noted previously,39,46 the voltammograms measured at low ferrocene surface coverages had peak widths that were similar in magnitude to the ideal 90 mV width expected for identical and independent electroactive groups tethered to a surface.47 For example, the anodic wave in Figure 1a has a peak width of about 105 mV, while that of the cathodic peak is about 130 mV. The larger width in the latter peak suggests some minor perturbations from ion-pairing48,49 or double-layer effects50,51 related to the development of the positive charge at the SAM interface. At a slower scan rate of 10 mV/s (not shown), the cathodic peak width reduces to 109 mV. The peaks measured here follow those reported in the literature and thus reflect the expectations for a well-organized monolayer with tethered ferrocene functionalities effectively localized at the outer periphery of the SAM.13,39,46 A more detailed comparison with the leading work of Chidsey et al.39,46 is warranted here. The peak splitting he reports for the identical SAM system (FcCO2C16SH/ C16SH) with similar ferrocene surface coverage at 100 mV/s is 66 mV, a value comparable to the 72 mV measured in this work. It should be noted, though, that the SAMs used by Chidsey et al. were prepared by a somewhat different procedure; his SAMs were ones obtained via the exchange of a FcCO2C16SH/C16SH SAM with a solution of pure C16SH thiol. This exchange progressively increased the peak splitting and evolved the cyclic voltammograms toward the more idealized quantitative metrics described above. (43) Lang, H.; Duschl, C.; Vogel, H. Langmuir 1994, 10, 197-210. (44) Han, W.; Lindsay, S. M.; Jing, T. Appl. Phys. Lett. 1996, 69, 4111-4113. (45) Xie, A. F.; Yamada, R.; Gewirth, A. A.; Granick, S. Unpublished results. (46) Chidsey, C. E. D. Science 1991, 251, 919-22. (47) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications, 2nd ed.; John Wiley & Sons: New York, 2001. (48) Rowe, G. K.; Creager, S. E. Langmuir 1991, 7, 2307-2312. (49) Andreu, R.; Calvente, J. J.; Fawcett, W. R.; Molero, M. J. Phys. Chem. B 1997, 101, 2884-2894. (50) Fawcett, W. R. J. Electroanal. Chem. 1994, 378, 117. (51) Smith, C. P.; White, S. H. Anal. Chem. 1992, 64, 12753.
9784
Langmuir, Vol. 19, No. 23, 2003
Twardowski and Nuzzo
It is known that the exchange process (where a SAM is equilibrated against a solution of different composition)52,53 is kinetically quite heterogeneous,52 and it is believed that it leads to the preferential exchange of adsorbates located near gross defects in the SAM.39,54 The removal of electroactive groups from these regions (an exchange that took ∼10 days to effect) generates the more idealized behaviors noted.39 Given the close similarity between our data and that of Chidsey’s exhaustively exchanged samples (see above), it is striking to note that our SAM was grown directly in 24 h (or less) from a mixed thiol solution and used without any subsequent exchange incubation. At present we do not fully understand why these very different preparative procedures generate the close correspondences that they do. We simply note them as being reflective of the complex, and as yet incompletely understood, assembly kinetics that characterize the formation of SAMs.12,14,15,17,55 Upon DMPC vesicle fusion, a dramatic change is seen in the voltammetric behavior of the SAM. As shown in Figure 1b, the well-defined redox peaks of the original SAM are extensively broadened and shift markedly toward more anodic potentials. The anodic peak is broadened such that a clear maximum is no longer visible and beyond 1000 mV is obscured by a rapid rise in the measured current. A very broad cathodic peak centered at approximately 880 mV, a positive shift of approximately 375 mV with respect to the data shown in Figure 1a, is evident on the reverse scan. Figure 1c shows the expected decrease in the capacitive currents measured for a pure C16SH SAM before and after DMPC fusion. We use these data to calculate a change in capacitance but defer further comment on it until later in this manuscript. For now, we simply note that it is consistent with the deposition of a lipid thin film on the SAM in the form shown in Chart 1.56,57 Ellipsometry measurements gave an average thickness of 22 Å for the SAM and 45-48 Å for the HBM, in agreement with previous reports.39,57 We also studied by SPR the change in the apparent mass coverage (a quantity deduced from the models used to interpret SPR data) that results due to the vesicle fusion on the SAM (Figure 2a). From these data we calculated a thickness for the HBM of 47 Å. The electrochemical irreversibility observed in the cyclic voltammogram shown in Figure 1b stems from the very hydrophobic environment created around the tethered ferrocene moiety as the DMPC vesicles fuse to the SAM. In related earlier experiments,48,58,59 the cyclic voltammograms of mixed ferrocene/alkanelthiol SAMs, where the alkyl chains were systematically made longer than the ferrocene-tethered chains, were shown to shift toward positive potentials and broaden markedly with increasing alkyl chain length (although peaks were still clearly discernible for the systems examined in that work). It has also been previously demonstrated that the ferrocene redox peaks show similar shifts when the SAM is exposed to aliphatic alcohols such as 1-decanol.60 This also was attributed to a surface segregation of the alcohols, as shown experimentally by the measured decrease in Figure 1. CV scans that illustrate the effect of DMPC adsorption on the electrochemistry of binary and single-phase SAMs on gold: (a) a cyclic voltammogram on a SAM composed of FcC16SH/C16SH; (b) same as part a but after DMPC adsorption; (c) an overlay of cyclic voltammograms of a C16SH SAM before (black) and after DMPC adsorption (gray); (d) a cyclic voltammogram of a FcC16SH/HOC16SH SAM before (black) and after DMPC adsorption (gray); (e) a cyclic voltammogram of a pure HOC16SH SAM before (black) and after DMPC adsorption (gray). All CV measurements were performed in 175 mM NaClO4 at a scan rate of 100 mV/s. pH ) 7.
(52) Bain, C. D.; Evall, J.; Whitesides, G. M. J. Am. Chem. Soc. 1989, 111, 7155-7164. (53) Allara, D. L.; Nuzzo, R. G. Langmuir 1985, 1, 45-52. (54) Collard, D. M.; Fox, M. A. Langmuir 1991, 7, 1192-1197. (55) Laibinis, P. E.; Nuzzo, R. G.; Whitesides, G. M. J. Phys. Chem. 1992, 96, 5097-5105. (56) Plant, A. L. Langmuir 1993, 9, 2764-67. (57) Ha, J.; Henry, C. S.; Fritsch, I. Langmuir 1998, 14, 5850-5857. (58) Creager, S. E.; Rowe, G. K. J. Electroanal. Chem. 1997, 420, 291-299. (59) Creager, S. E.; Rowe, G. K. J. Electroanal. Chem. 1994, 370, 203-211. (60) Creager, S. E.; Rowe, G. K. Langmuir 1993, 9, 2330-2336.
Molecular Recognition at Model Organic Interfaces
Langmuir, Vol. 19, No. 23, 2003 9785
Chart 1. Schematics of (a) a Hybrid Bilayer Membrane on a Hydrophobic SAM Support and (b) a Bilayer Assembly on a Hydrophilic SAM Support
the SAM’s capacitance in a solution containing these aliphatic alcohols. In each case, the electrochemical irreversibility derives both from the stability of the ferrocene relative to its oxidized ionic form in increasingly hydrophobic environments and from the steric impediment to counterion motion toward the redox sites that is obtained from the presence of long chains and/or other nonpolar adlayer species. With this in mind, we also selfassembled HBMs where we added tetrabutlyammonium salts to the vesicle mixture to see if these more hydrophobic species would partially partition into the HBM and affect the tethered ferrocene redox chemistry; the measured cyclic voltammograms were similar to that shown in Figure 1b. One of the useful properties of SAMs on gold is the ability they afford to readily tailor the surface properties of the substrate by simply changing the end group of the thiol chain.22,55,61-65 By replacing only one hydrogen of the end methyl group in an alkanethiol SAM with a hydroxyl (OH) functionality, one can change the surface from a highly hydrophobic interface to a hydrophilic one, corresponding with the static contact angle for water dropping from 113° to 10 MΩ cm2).125 Water molecules and electrolyte ions residing in a layer between the SAM and the lipid (a phase that is believed to be necessary to support the lipid)126 must contribute to some degree to the differences noted here.56,57,123 Still, the larger body of data suggests that other issues may also need to be considered, namely the field-induced motions of mobile ions that penetrate the lipid adlayer structure, either via permeation or in a defect-based transport process.96 The CV data strongly implicate the occurrence of some degree of ion transport through the supported lipid assemblies derived from DMPC. Even so, the degree of permselection noted in this transport cannot be easily understood in the context of a mechanism based solely on the gross structural defects (e.g. such as pinholes). One possible model of the transport mediated by the DMPC lipid assemblies follows from the results of recent studies reported in the literature.127 This work illustrates that small unilamellar vesicles can adsorb on a surface such as was used here without rupture.127 These isolated structures should be unstable, though, with respect to fusion and incorporation into the layered structure derived from them, although it is possible that some could still remain as larger vesicles embedded in the bilayer assembly.122 The phase boundaries created in this way might then possibly serve to mediate the transport seen. The AFM data we have obtained, while not conclusive, do not tend to fully support this structural model. The lipid adlayer we obtained appears to be one that simply presents multiple bilayers of the DMPC. As a result, the overlayer growth tends to form multilayer structures. The AFM data of others further suggest that related adsorption processes of DMPC on a planar substrate such as mica occur so as to give bilayers exhibiting mixed phase statessa coexistence of both liquid and gel phases.45 Our data are consistent with this notion (the lesser ∼1 nm corrugations seen in the AFM data reflecting a boundary crossing for the domains). This complex structure presents many other topographical features as well, with those of other bilayer sheaths being prominent. Our data also did not follow the patterns observed in a recent study by Jenkins et al.,34 (120) Ogier, S. D.; Bushby, R. J.; Cheng, Y.; Evans, S. D.; Evans, S. W.; Jenkins, A. T. A.; Knowles, P. F.; Miles, R. E. Langmuir 2000, 16, 5696-5701. (121) Horswell, S. L.; Zamlynny, V.; Li, H.-Q.; Merill, R.; Lipkowski, J. Faraday Discuss. 2002, 121, 405-422. (122) Radler, J.; Strey, H.; Sackmann, E. Langmuir 1995, 11, 45394548. (123) Porter, M. D.; Bright, T. B.; Allara, D. L.; Chidsey, C. E. D. J. Am. Chem. Soc. 1987, 109, 3559-68. (124) Steinem, C.; Janshoff, A.; Sieber, M. Bioelectrochem. Bioenerg. 1997, 42, 213. (125) Muller, P.; Rudin, D. O. Nature 1968, 217, 713-719. (126) Groves, J. T.; Ulman, N.; Boxer, S. G. Science 1997, 275, 651653. (127) Reviakine, I.; Brisson, A. Langmuir 2000, 16, 1806-1815.
Molecular Recognition at Model Organic Interfaces
who noted that the fusion of DMPC on a mercaptoethanol SAM led to adlayers decorated by vesicle-sized objects. In qualitative terms, though, the ideas that emerge in this work are that the effective mass coverage of the DMPC on the SAM is larger than that of a simple bilayer. While the SAM does appear to be covered by a full bilayer, excess material in a variety of structural forms is present (albeit in degrees that may depend on the exact experimental conditions used). The inferences made above on the basis of capacitance measurements are strongly supported by the results of the CV studies. First, and most notable, the ferrocene groups tethered to the supporting 1-hydroxylhexadecanethiol mixed thiol SAM remain electroactive after the deposition of the DMPC. These later waveforms are only weakly perturbed relative to those seen for the unmodified SAM. This illustrates that the transport of electrolyte ions to this active region of the SAM interface is possible and that the rate of this process is, in fact, relatively fast.48 If this were not the case, the reversible potentials for the ferrocene oxidation and reduction should be very strongly perturbed in ways that are not seen experimentally. This latter argument, then, describes a metric for ion transport kinetics, one that is referenced to the rate discriminated by the CV measurement. The data also show that other redoxactive ions will readily pass through the DMPC adlayer. It is intriguing to note, however, that the DMPC layers are not uniformly leaky in the sense that all soluble redox active ions will cross through easily. Perhaps the most striking example of this selectivity is that the soluble ferrocene derivative, FcNH+, a molecule bearing a similar charge and possessing a structure much like that of the DMPC permeable FcNCH3+, is strongly excluded by the DMPC adlayer (Figure 5a). These data thus suggest a qualitative limit for the length scale at which the SAM-supported DMPC adlayers discriminate against the transport of soluble ionic species, be it in the form of electrolyte ions or redox active species. This limit is not one expected for a gross defect (such as are seen in lipid phases exhibiting gel-liquid coexistence)45 but rather is one appropriate for discriminating between small molecular species, albeit in a fairly rudimentary way. The question remains then as to the nature of the dynamics or structural features that could act in this way. First, with the coverage of lipid in the lower bilayer being less than that needed to form a full layer of the gel
Langmuir, Vol. 19, No. 23, 2003 9791
state, coexistent phases are both expected and seen. Closely related AFM studies suggest that these phases embed discontinuities that are of much lower lipid density than those of either the gel or liquid phases and thus might be regions that are permeable to redox probes and other ions.45 An alternative model described in the literature that serves to rationalize our experimental findings128 is one based on intrinsic “ion-like” channels. Their relevance for mass-transfer processes such as are of interest here has yet to be established. Finally, one must consider the possibility that probe partitioning (even at the low levels established by the SPR data) might also be contributing. This might occur by, for example, providing a low-resistance path for electron hopping. Whatever the case, one must bear in mind that the behaviors seen for the tethered ferrocene chains in the presence of the lipid are ones that require fairly responsive motions of counterions through the mesophase to the support. A strategy for developing a better understanding of the various aspects of the mechanism(s) might be found in studies that exploit phase behaviors of the lipid assembly more directly. Fusion of unimolecular vesicles comprised of multiple lipid types may be a promising method for doing this. These studies are described in a future paper. Acknowledgment. The support of the National Science Foundation (CHE 0097096) and the Department of Energy (DEFG02-91ER45439) is gratefully acknowledged. We wish to thank Marco Bayas and Professor D. Leckband of the Department of Chemical Engineering at UIUC for their invaluable help in running and interpreting the SPR experiments and their generosity in donating their time to these matters. We thank Z. Vivian Feng of Professor A. Gewirth’s group in the Department of Chemistry at UIUC for helping to image the supported assemblies. We would also like to thank Professor A. Gewirth for his numerous helpful discussions that aided in developing our interpretations of the data. Supporting Information Available: AFM images of both the bare SAM and lipid/SAM hybrid structures on gold substrates. This information is available free of charge via the Internet at http://pubs.acs.org. LA0349018 (128) Nagle, J. F.; Scott, H. L. Biochim. Biophys. Acta 1978, 513, 236-243.
