Thiocholesterol on Gold: A Nanoporous Molecular Assembly

Apr 3, 1996 - The formation of thiocholesterol (TC) monolayers on gold has been studied by ellipsometry, contact angle measurements, infrared spectros...
19 downloads 7 Views 178KB Size
+

+

1704

Langmuir 1996, 12, 1704-1707

Thiocholesterol on Gold: A Nanoporous Molecular Assembly Z. P. Yang,†,‡ I. Engquist,‡ J.-M. Kauffmann,† and B. Liedberg*,‡ Institute of Pharmacy, Universite Libre de Bruxelles, Campus Plaine CP205/6, B-1050, Brussels, Belgium, and Laboratory of Applied Physics, Linko¨ ping University, S-581 83 Linko¨ ping, Sweden Received December 11, 1995. In Final Form: February 9, 1996X The formation of thiocholesterol (TC) monolayers on gold has been studied by ellipsometry, contact angle measurements, infrared spectroscopy, and cyclic voltammetry. Subsequent treatment of the TC assembly with 11-mercaptodeuterioundecanoic acid (MDUA) shows that the average surface coverage is about 65% of that of a self-assembled alkanethiolate monolayer and that it has a large number of molecular defects. These defects exist because of a mismatch between the size and shape of the TC molecule and the pinning distance at the Au(111) crystal lattice. Potential uses of these defect-rich structures are microelectrode arrays for electroanalytical and biosensor applications.

Introduction Self-assembled monolayers (SAMs) of ω-substituted alkanethiols on gold1-8 are dense structures with a high degree of order. Extensive characterizations2,4,9-12 have revealed that these monolayers form a (x3 × x3) R30° overlayer structure on Au(111) surfaces. Various superlattice structures have also been observed, depending on annealing history and sample temperature, with the c(4 × 2) structure being the most common one.13-16 The monolayers serve as ideal model systems in surface chemical and biological research, and they have during the last 10 years contributed to a deeper understanding of interfacial phenomena in areas like long range electron transfer,17-19 wetting,20,21 protein adsorption,22 and mo* To whom correspondence should be addressed. † Universite Libre de Bruxelles. ‡ Linko ¨ ping University. X Abstract published in Advance ACS Abstracts, March 15, 1996. (1) Nuzzo, R. G.; Allara, D. L. J. Am. Chem. Soc. 1983, 105, 44814483. (2) Porter, M. D.; Bright, T. B.; Allara, D. L.; Chidsey, C. E. D. J. Am. Chem. Soc. 1987, 109, 3559-3568. (3) Troughton, E. B.; Bain, C. D.; Whitesides, G. M.; Nuzzo, R. G.; Allara, D. L.; Porter, M. D. Langmuir 1988, 4, 365-385. (4) Bain, C. D.; Troughton, E. B.; Tao, Y.-T.; Evall, J.; Whitesides, G. M.; Nuzzo, R. G. J. Am. Chem. Soc. 1989, 111, 321-335. (5) Bain, C. D.; Evall, J.; Whitesides, G. M. J. Am. Chem. Soc. 1989, 111, 7155-7164. (6) Bain, C. D.; Whitesides, G. M. J. Am. Chem. Soc. 1989, 111, 71647175. (7) Bain, C. D.; Whitesides, G. M. Angew. Chem., Int. Ed. Engl. 1989, 28, 506-512. (8) Dubois, L. H.; Nuzzo, R. G. Annu. Rev. Phys. Chem. 1992, 437463. (9) Widrig, C. A.; Alves, C. A.; Porter, M. D. J. Am. Chem. Soc. 1991, 113, 2805-2810. (10) Samant, M. G.; Brown, C. A.; Gordon, J. G. I. Langmuir 1991, 7, 437-439. (11) Bain, C. D.; Davies, P. B.; Ong, T. H.; Ward, R. N.; Brown, M. A. Langmuir 1991, 7, 1563-1566. (12) Alves, C. A.; Smith, E. L.; Porter, M. D. J. Am. Chem. Soc. 1992, 114, 1222-1227. (13) Mozaffary, H. Thin Solid Films 1994, 244, 874-877. (14) Small, D. M. In The Physical Chemistry of Lipids: From Alkanes to Phospholipids; Plenum: New York, 1986; p 395. (15) Mozaffary, H. Thin solid Films 1994, 244, 847-877. (16) Fenter, P.; Eisenberger, P.; Liang, K. S. Phys. Rev. Lett. 1993, 70, 2447-2450. (17) Finklea, H. O.; Avery, S.; Lynch, M.; Furtsch, T. Langmuir 1987, 3, 409-413. (18) Collinson, M.; Bowden, E. F.; Tarlov, M. J. Langmuir 1992, 8, 1247-1250. (19) Becka, A. M.; Miller, C. J. J. Phys. Chem. 1992, 96, 2657-2688. (20) Nuzzo, R. G.; Dubois, L. H.; Allara, D. L. J. Am. Chem. Soc. 1990, 112, 558-569. (21) Ulman, A.; Evans, S. D.; Shnidman, Y.; Sharma, R.; Eilers, J. E.; Chand, J. C. J. Am. Chem. Soc. 1991, 113, 1499-1506.

Chart 1. Molecular Structure of Thiocholesterol

lecular recognition.23 However, the SAMs are seldom perfect.24 Structural defects exist in most self-assembled monolayers, due to surface roughness of the gold substrates,25 mismatches in the chain orientation of adjacent molecular domains,24,26,27 and etching processes during the chemisorption of the thiols onto the gold substrate.28 The formation of defects is hard to control, both with respect to their type and density. However, once characterized, they can be used to address a series of critical issues concerning charge and mass transport phenomena on and across membrane-like structures.2,19,29 In this letter, we investigate the structure of and defect formation in self-assembled monolayers of thiocholesterol (TC, Chart 1) on gold. TC is a cholesterol molecule bearing a thiol moiety instead of a hydroxyl group at the 3βposition. The molecular structure is a nearly planar polycyclic steroid ring with a branched aliphatic chain. Our research indicates that the asymmetric TC molecule forms a self-assembled monolayer on gold. This novel cholesterol interface should be very useful in fundamental studies both of biomembranes and of a broad range of molecular recognition processes at surfaces, since cholesterol occurs at a quite high concentration in the membranes of many animal cells. A large number of defects are expected to exist in the TC monolayer because of the size and irregular shape of the molecule, which will prevent the formation of structures as well ordered and densely packed as the (x3 × x3) R30° structure normally found for alkanethiols on Au(111). (22) Prime, K. L.; Whitesides, G. M. Science 1991, 252, 1164-1167. (23) Ha¨ussling, L.; Michel, B.; Ringsdorf, H.; Rorher, H. Angew. Chem., Int. Ed. Engl. 1991, 30, 569-572. (24) Delamarche, E.; Michel, B.; Gerber, C.; Anselmetti, D.; Guntherodt, H.-J.; Wolf, H.; Ringsdorf, H. Langmuir 1994, 10, 28692871. (25) Creager, S. E.; Hockett, L. A.; Rowe, G. K. Langmuir 1992, 8, 854-861. (26) Camillone, N., III; Chidsey, C. E. D.; Eisenberger, P.; Fenter, P.; Li, J.; Liang, K. S.; Liu, G.-Y.; Scoles, G. J. Chem. Phys. 1993, 99, 744747. (27) Butt, H. J.; Seifert, K.; Bamberg, E. J. Phys. Chem. 1993, 97, 7316-7320. (28) Sondag-Huethorst, J. A. M.; Scho¨nenberger, C.; Fokkink, L. G. J. J. Phys. Chem. 1994, 98, 6826-6834. (29) Wang, J.; Wu, H.; Angnes, L. Anal. Chem. 1993, 65, 1893-1896.

+

+

Letters

Langmuir, Vol. 12, No. 7, 1996 1705

Figure 1. (a) IR transmission spectrum of thiocholesterol in an isotropic sample (KBr pellet, 0.005 mg of TC/mg of KBr, pellet thickness ) 0.4 mm). (b) IRAS spectrum of thiocholesterol in the SAMs on gold. Table 1. Peak Assignments for TC in KBr (Isotropic Sample) and on Gold wavenumber (cm-1) TC in KBr

TC on gold

peak assignment

3032 2961 2932 2903 2867 2848 1465 1378

2963 2937 2903 2865, 2874 2847 1469 1389

dCHsCsH str CH3 CsH asym str CH2 CsH asym str CH CsH str (tentative) CH3 CsH sym str CH2 CsH sym str CH2 scissors def CH3 sym def

Experimental Section Self-assembled monolayers of TC were prepared by immersing cleaned substrates (2000 Å of Au on Si), which exhibit a very strong (111) texture,30 in a 1 mM ethanol solution of TC (Aldrich) overnight. Characterization by contact angle goniometry and ellipsometry confirmed formation of a TC monolayer on the gold substrates. Values routinely obtained for the advancing and receding contact angles were 107 ( 1° and 93 ( 2° with water and 15 ( 1° and 0° with hexadecane, respectively. If the liquid drop spread to an irregular shape, the contact angle value was assumed to be zero. The monolayer thickness was found to be 16 ( 1 Å by ellipsometry. The film thickness calculation was based on a three-phase model assuming a nonabsorbing, isotropic TC film with a refractive index of 1.50. Infrared spectroscopy and electrochemical measurements have been used to further determine the structural characteristics of the TC monolayers. Infrared reflection-absorption spectroscopy (IRAS) measurements were performed in an ultrahigh vacuum system,31 where a base pressure of 10-9 mbar or better was maintained throughout the experiments. Spectra of the monolayers were obtained by using a Ne ion sputter-cleaned sample as reference. Electrochemical investigations were made in an electrochemical cell,25 allowing measurement at working electrodes with a reproducible geometrical electrode area of 0.38 cm2.

Results and Discussion An IRAS spectrum of TC on gold is presented in Figure 1, together with the transmission spectrum of an isotropic sample (KBr pellet). The peak frequencies along with suggested mode assignments are shown in Table 1. To the authors’ knowledge, although cholesterol is an important component in biomembranes, there are no available IR data on cholesterol or TC on solid surfaces. (30) Bertilsson, L.; Liedberg, B. Langmuir 1993, 9, 141-149. (31) Engquist, I.; Lundstro¨m, I.; Liedberg, B. J. Phys. Chem. 1995, 99, 12257-12267.

A considerable part of the infrared work on SAMs has been devoted to the CsH stretching vibrations in the 2700-3100 cm-1 region, since they provide useful information about molecular conformation and orientation. In the case of TC, we observe substantial differences in the relative peak intensities between the monolayer and bulk spectra. This is illustrated by the CH2 νas vibration, whose intensity changes from 1.4 times the CH3 νas peak in the bulk spectrum to 0.4 times the CH3 νas peak in the monolayer spectrum. The large dissimilarity indicates that the TC monolayer is not isotropic, i.e., that it must be ordered. A further sign of ordering is the intensity of the asymmetric CH3 stretching vibration at 2963 cm-1, which is by far the strongest peak in the spectrum of TC on gold. The transition dipole moment of this vibration must be aligned essentially normal to the gold surface for this to happen, according to the surface selection rule.32 Since the majority (4 of 5) of the TC methyl groups have transition dipole moments that lie in a plane roughly parallel to the molecular axis of TC, we suggest that the TC molecular axis is aligned parallel to or at a fairly small angle with respect to the surface normal. The orientational ordering referred to here should not be confused with positional (lateral) ordering, which we are unable to address with IRAS. From space-filling models of the TC molecule, we conclude that the maximum expected thickness of a TC monolayer is about 20 Å, assuming a close-packed, fully extended molecular structure normal to the surface. Since the measured ellipsometric thickness of a SAM of TC is only 16 Å, this fully extended molecular structure cannot be the correct one. The TC molecules must either be inclined with respect to the surface normal or be less densely packed. If we assume that tilted TC molecules are the sole cause for measuring 16 Å in thickness, the tilt angle may be calculated to be 37°. On the other hand, if the packing density is reduced significantly, the refractive index of the TC monolayer will undoubtedly be lower than 1.5, perhaps 1.45, a value more representative for liquid-like polymethylenes.4 When the refractive index is lowered to 1.45 in the monolayer, the thickness calculation gives a value of 17 Å instead of 16 Å, which in turn means that the molecular tilt angle must be smaller than the previously calculated 37°. The present set of data does not allow us to separate these two effects from each other. However, as the IRAS data clearly indicate that the assembly consists of highly oriented TC molecules, the possibility of a completely isotropic (liquid-like) monolayer is ruled out. Given the molecular structure of TC, it is likely that TC monolayers cannot be as densely packed as alkanethiolate SAMs. We have examined this hypothesis by taking a supposedly fully assembled monolayer of TC and immersing it in an ethanol solution of 11-mercaptodeuterioundecanoic acid (MDUA), HS(CD2)10COOH, overnight. For comparison, a fully assembled monolayer of MDUA was put in an ethanol solution of TC for the same period of time. Similar experiments have been described by Allara and Nuzzo33 and Bain and co-workers.5 IRAS spectra of the two monolayers before and after immersion in the second solution can be seen in Figure 2. If we compare spectrum 2a (TC) and 2b (TC and then MDUA), we see that peaks characteristic of MDUA have appeared in spectrum 2b, indicating that a significant amount of MDUA has been incorporated into the TC monolayer. We can determine the postadsorbed amount of MDUA by integrating the CD stretching peak at 2196 cm-1 in spectra (32) Francis, S. A.; Ellison, A. H. J. Opt. Soc. Am. 1959, 49, 131-138. (33) Allara, D. L.; Nuzzo, R. G. Langmuir 1985, 1, 45-52.

+

1706

+

Langmuir, Vol. 12, No. 7, 1996

Figure 2. IRAS spectra of the SAMs of thiocholesterol (TC) and 11-mercaptodeuterioundecanoic acid (MDUA) in the exchange experiments (see text): (a) TC/Au; (b) TC/Au exchanged with MDUA; (c) MDUA/Au; (d) MDUA/Au exchanged with TC.

Letters

layers is suppressed in the MDUA-filled TC monolayer, due to an even distribution of the postadsorbed MDUA molecules. The advancing contact angle of water on the SAM changes from 108° to 70° as a result of the filling process, but the ellipsometric thickness stays the same. For comparison, the same kind of posttreatment experiment was performed using bis(16-hydroxyhexadecyl) disulfide, (S(CH2)15OH)2, instead of MDUA. Interestingly, no incorporation of these molecules into the TC monolayer occurs, as evidenced by contact angle measurements (the advancing contact angle of water only changes from 108° to 106°) and IRAS (no change after posttreatment). We conclude that the defects in the TC monolayer are smaller than the bis(16-hydroxyhexadecyl) disulfide molecule, i.e., smaller than two hydrocarbon chains side by side. When, inversely, a SAM of MDUA is immersed in a TC solution, the IRAS spectra show only small changes in the 2700-3100 cm-1 region (spectra 2c and 2d) and the contact angle of water and the ellipsometric thickness stay the same (32 ( 2° and 17 ( 1Å, respectively). This result suggests that the SAMs of MDUA are highly dense. Judging from the small peaks that appear in the 27003100 cm-1 region, a small amount of TC however seems to fill in defects in the MDUA monolayer or contaminate the MDUA monolayer surface, but the amount is less than 16% of the full TC monolayer coverage. From the CsH and CD2 vibrational modes in Figure 2, the molecular surface coverage, θ, of the TC monolayer can be calculated. We assume that the molecular surface coverage of an MDUA SAM on gold equals 1, and by making a number of assumptions,34 we arrive at the following expression for θ:

θ)

Figure 3. Schematic diagram illustrating the formation of a mixed TC/MDUA monolayer by immersion of the TC monolayer into an MDUA solution.

2b and 2c (pure MDUA). This shows that around 50% of a full MDUA monolayer is postadsorbed into the TC monolayer. Furthermore, the CsH stretching peaks in the 3000 cm-1 region change in intensity by a factor of ∼0.75 from spectrum 2a to 2b, meaning that around 25% of the TC molecules have desorbed. The possibility that the intensity change is due to a conformational rearrangement of TC as MDUA is postadsorbed has been considered but found unlikely, since the relative peak intensities in the CH stretching region are virtually the same in spectra 2a and 2b. A conformational rearrangement normally would change the different peak intensities by different factors, according to the surface selection rule.32 Spectra 2a, 2b, and 2c together indicate that the MDUA molecules fill the defects and displace a small amount of TC, forming a true mixed monolayer. The process is schematically outlined in Figure 3. Further evidence is gained from the double CdO stretching peak, which is slightly shifted in spectrum 2b compared to 2c (1713 and 1741 cm-1 compared to 1717 and 1738 cm-1) and which also displays a significantly different intensity distribution in the two spectra. This behavior suggests that the lateral hydrogen bonding in pure MDUA mono-

(

(ATC)a (ATC)b

1-

)

(AMDUA)b (AMDUA)c

where (ATC)a and (ATC)b are the integrated intensities of the CH3 νas vibrational mode at 2963 cm-1 in the SAMs of pure TC and TC posttreated with MDUA, respectively. (AMDUA)c and (AMDUA)b are the integrated intensities of the CD2 νas vibrational mode at 2196 cm-1 in the SAMs of pure MDUA and TC posttreated with MDUA, respectively. A molecular surface coverage θ ) 0.65 ( 0.04 (average of three measurements) is obtained, meaning that a TC monolayer covers 65% of the gold surface area, leaving 35% uncovered. Thus, significantly more defects exist in the SAMs of TC, as compared to the highly dense SAMs (34) We assume that the fraction of the gold surface area that is covered by TC molecules in a SAM is proportional to the integrated intensity of the CH3 νas peak at 2963 cm-1, with the proportionality factor fTC. We assume a similar case for the MDUA molecules and the C-D peak at 2196 cm-1 with the proportionality factor fMDUA. Further, we denote the integrated intensities of the 2963 and 2196 cm-1 peaks in Figure 3a, b, and c by (ATC)a, (ATC)b, (AMDUA)b, and (AMDUA)c. Assuming that the coverage for a pure TC monolayer (Figure 3a) is θ and that the coverages for MDUA (Figure 3c) and TC filled with MDUA (Figure 3b) both are 100% (as shown by the cyclic voltammograms in Figure 4b and c, which indicate densely packed monolayers in both cases), we can set up three equations:

fTC(ATC)a ) θ fTC(ATC)b + fMDUA(AMDUA)b ) 1 fMDUA(AMDUA)c ) 1 This system of equations can easily be solved to yield θ as a function of the four peak areas. A slight uncertainty might arise from using the same fMDUA and fTC for both pure and mixed monolayers, but the variations (if any) should be small, considering the similarity of the IRAS spectra, and we expect the resulting error in θ to be minor.

+

+

Letters

Figure 4. Cyclic voltammograms at (a) a thiocholesterol-coated gold electrode, (b) a thiocholesterol-coated gold electrode which was further modified with 11-mercaptodeuterioundecanoic acid, and (c) an 11-mercaptodeuterioundecanoic acid coated gold electrode in a 0.1 M KNO3 solution containing 1 mM K3Fe(CN)6 at pH 7. Scan rate ) 100 mV/s.

of ω-substituted alkanethiols. As suggested before, this is probably due to the irregular shape of the TC molecule and to a mismatch between the TC molecular size and the distance between pinning sites at the Au(111) crystal lattice. The lower packing density is also in agreement with the conclusions drawn from the ellipsometric data discussed above. Electrochemical measurements, i.e., cyclic voltammetry, using ferricyanide as the electroactive compound, have been applied to the SAMs of TC on gold electrodes (Figure 4) in order to further identify the presence of defects. Compact monolayers, prepared by using alkanethiol SAMs on gold electrodes, have been reported to inhibit diffusion of electroactive molecules to the gold surface, giving a very small response in cyclic voltammetry.2 Curve 4a shows a cyclic voltammogram of a gold electrode, modified with a SAM of TC, in a 0.1 M KNO3 solution containing 1 mM K3Fe(CN)6 at pH 7. The voltammogram displays a plateau current, independent of the scan rate in the

Langmuir, Vol. 12, No. 7, 1996 1707

range from 50 to 1000 mV/s. The electrochemical behavior resembles that at a microelectrode array,35 which again indicates the existence of structurally induced defects in the SAMs of TC. After posttreatment of a TC-modified gold electrode by immersing it in a 1 mM MDUA solution, the plateau current disappears (Figure 4b). The electrochemical behavior is close to that on gold electrodes modified with a pure MDUA SAM (Figure 4c), which is known to be ordered and densely packed.36 These results further confirm that the defects in TC SAMs can be filled by MDUA, resulting in densely packed mixed SAMs. In conclusion, a combination of IRAS and electrochemical measurements confirms that TC assembles into oriented SAMs on gold substrates. The surface coverage is around 65% of that of alkanethiolate SAMs, which results in the formation of a large number of defects with a size of about 5-8 Å. It is suggested that the defects are formed as a result of the size and shape of the thiocholesterol molecule itself and thus are of significantly different nature from the defects observed with ordinary SAMs of alkanethiols, which mainly result from defects in the underlying metal substrate. Posttreatment experiments have shown that the defects offer a good template for further modification of this membrane-like monolayer. Microscopic defects are also desirable in electrochemical applications, where these substrates may function as a microelectrode array.35,37 Acknowledgment. Thanks are expressed to the European Science Foundation (ESF) for providing an ABI (Artificial Biosensing Interfaces) Fellowship to Z.Y. (Free University of BrusselssULBs Pharmaceutical Institute) in 1994 and 1995. This work was supported by a grant from the Swedish Research Council for Engineering Sciences. LA951524T (35) Finklea, H. O.; Snider, D. A.; Fedyk, J.; Sabatani, E.; Gafni, Y.; Rubinstein, I. Langmuir 1993, 9, 3660-3667. (36) The peak position of the CD2 νas vibration (2196 cm-1) clearly indicates a well-ordered, densely packed SAM. For comparison, the same peak for a Au-S(CD2)19CD3 SAM is found at 2195 cm-1. (37) Turyan, I.; Mandler, D. Anal. Chem. 1994, 66, 58-61.