Langmuir 1998, 14, 5139-5146
5139
Dynamics of the Formation of Mixed Alkanethiol Monolayers: Applications in Structuring Biointerfacial Arrangements Christopher Cotton,† Andrew Glidle,† Graham Beamson,‡ and Jonathan M. Cooper*,† Bioelectronics Research Centre, Department of Electronic & Electrical Engineering, Glasgow University, Glasgow, U.K., and Research Unit for Surfaces, Transforms and Interfaces, Daresbury Laboratory, Warrington, U.K. Received March 23, 1998. In Final Form: June 15, 1998 The control of molecular recognition at interfaces is of importance in many areas of bioelectronics and biocatalysis, particularly in the design of biosensors, where it may be desirable to be able to manipulate the degree of interaction between a protein and a metal surface. To develop appropriate models to study such phenomena, we have produced mixed monolayers on gold by the displacement of molecules from a previously formed, homogeneous self-assembled monolayer (SAM) of an alkanethiol (mercaptoethylamine) using a solution containing a thiol with a different functional headgroup (mercaptopropanol). The dynamics of the formation of the resultant surfaces were investigated using high-resolution X-ray photoelectron spectroscopy (XPS) to probe both the C(1s) and N(1s) environments, corresponding to the different species in the SAMs. To illustrate the application of such interfacial arrangements and to further understand their role in controlling protein-surface interactions, two systems were investigated: in the first instance, the electron transfer between the small redox protein, cytochrome c, and the mixed monolayers was investigated using cyclic voltammetry; second, the potential for producing surfaces with different amounts of immobilized antibodies was demonstrated using fluorescence microscopy.
Introduction Over the past decade, chemical modification of sensor surfaces has been recognized as a means of introducing both specificity and functionality.1-4 For example, as long ago as 1983, Nuzzo and Allara showed that alkanethiols will chemisorb spontaneously onto gold surfaces, to form organized self-assembled monolayers (SAMs),5 thereby providing a means of introducing a degree of molecular functionality.6-10 In the context of understanding biointerfacial interactions, the tail group of such a selfassembled alkanethiol can be considered to tether the molecular structure to the gold surface, while the headgroup has the potential to act as a site for molecular recognition, e.g. for the selective immobilization of a ligandbinding protein (such as an antibody, or streptavidin) as may be appropriate for a biosensor.2,11-20 * To whom correspondence should be addressed. † Glasgow University. ‡ Daresbury Laboratory. (1) Williams, R. A.; Blanch, H. W. Biosens. Bioelectron. 1994, 9, 159. (2) Zhong, C. J.; Porter, M. D. Anal. Chem. 1995, 367, 709A. (3) Byfield, M. P.; Abuknesha, R. A. GEC J. Res. 1991, 9, 97. (4) Scouten, W. H.; Luong, J. H. T.; Brown, R. S. Trends Biotechnol. 1995, 13, 178. (5) Nuzzo, R. G.; Allara, D. L. J. Am. Chem. Soc. 1983, 105, 4481. (6) Ulman, A. Chem. Rev. 1996, 96, 1533. (7) Xu, J.; Li, H. L. J. Colloid Interface Sci. 1995, 176, 138. (8) Allara, D. L. Biosens. Bioelectron. 1995, 10, 771. (9) Bain, C. D.; Evans, S. D. Chem. Br. 1995, Jan, 46. (10) Haussling, L.; Ringsdorf, H.; Schmitt, F.-J.; Knoll, W. Langmuir 1991, 7, 1837. (11) Piehler, J.; Brecht, A.; Geckeler, K. E.; Gauglitz, G. Biosens. Bioelectron. 1996, 11, 579. (12) Lotzbeyer, T.; Schuhmann, W.; Katz, E.; Falter, J.; Schmidt, H. J. Electroanal. Chem. 1994, 377, 291. (13) Allen, P. M.; Allen, H.; Hill, O.; Walton, N. J. J. Electroanal. Chem. 1984, 178, 69. (14) Rickert, J.; Gopel, W.; Beck, W.; Jung, G.; Heiduschka, P. Biosens. Bioelectron. 1996, 11, 757. (15) Mrksich, M.; Chen, C.; Xia, Y.; Dike, L. E.; Ingber, D. E.; Whitesides, G. M. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 10775. (16) Delamarche, E.; Sundarababu, G.; Biebuyck, H.; et al. Langmuir 1996, 12, 1997. (17) Schramm, W.; Paek, S.-H. Anal. Biochem. 1992, 205, 47.
The abundance and the spatial distribution of such headgroup recognition sites has important implications on both the micro- and macroscopic properties of a bioelectronic interface, as shall be illustrated in this paper. In this context, it has previously been shown that antibodies, immobilized at a mixed monolayer, show a surface-composition dependent variation in their ability to bind an antigen, a fact which was attributed to steric hindrance.17-19 Although similar results have been observed for the binding of streptavidin using mixed monolayers of biotinylated thiols and alcohol-terminated thiols,20 in none of these examples has the extent of the immobilization process been used as a probe to measure the monolayer structure. The use of long chain alkanethiols to produce mixed monolayers on gold surfaces has received attention from a wide number of groups working in physics, biochemistry, and the analytical sciences.3 For example, Takami et al. have used scanning tunneling microscopy analysis to distinguish between different headgroups in a mixed monolayer consisting of alkanethiols of chain lengths of >12 carbon atoms, demonstrating that the resulting (modified) surface reflected the molar ratio of headgroups in the (loading) solution.21 Other research, producing mixed monolayers of CH3- and OH-terminated long chain alkanethiols (of the same chain length) through coadsorption has shown that self-assembly results in true molecular mixing (with no single-component domains).22-25 (18) Wimalasena, R. L.; Wilson, G. S. J. Chromatogr. 1991, 572, 85. (19) Rickert, J.; Gopel, W.; Beck, W.; Jung, G.; Heiduschka, P. Biosens. Bioelectron. 1996, 11, 757. (20) Haussling, L.; Ringsdorf, H.; Schmitt, F. J.; Knoll, W. Langmuir 1991, 7, 1837. (21) Takami, T.; Delmarche, E.; Michel, B.; Gerber, C.; Wolf, H.; Ringsdorf, H. Langmuir 1995, 11, 3876. (22) Bain, C. D.; Whitesides, G. M. J. Am. Chem. Soc. 1989, 111, 7164. (23) Bain, C. D.; Evall, J.; Whitesides, G. M. J. Am. Chem. Soc. 1989, 111, 7155. (24) Bain, C. D.; Biebuyck, H. A.; Whitesides, G. M. Langmuir 1989, 5, 723. (25) Bertilsson, L.; Liedberg, B. Langmuir 1993, 9, 141.
S0743-7463(98)00321-7 CCC: $15.00 © 1998 American Chemical Society Published on Web 08/12/1998
5140 Langmuir, Vol. 14, No. 18, 1998
While long chain SAMs have the advantage of forming highly ordered structures,20-26 they do have the potential disadvantage that they serve to distance the solution phase from the underlying (sensor) surface, which may result in the reduction of the rate of electron transfer between a (solution phase) redox center and the electrode.27 In addition, long chain monolayers, by their nature, often form a close-packed, ordered barrier, which limits the diffusion of a solution phase electroactive species, an arrangement which has practical limitations if such an interface were to be used for electrochemical sensing. In common with other authors,28 we have prepared mixed monolayers of short chain alkanethiols by the time dependent displacement of an initial modifier. In our work, we have found that the use of short chain thiols not only helps to maximize the rate of interfacial electron transfer between proteins, but also provides surfaces which are more amenable to quantification using XPS. This latter aspect of the study is a consequence of higher ratios of headgroup specific signals (e.g. NH2, COH) to the alkyl backbone signals, than is the case for long chain thiols. Thus we not only are able to probe the dynamics of monolayer formation using spectroscopic studies but also, in addition, have related these results to the study of the redox interaction of cytochrome c with the resultant monolayers. We are able to show that, by successive immersions of a gold substrate into single component thiol solutions, it is possible to control the ratio of two alkanethiols, containing two different functional headgroups, within a mixed monolayer. Finally, we demonstrate one potential application of such a system, by showing differing (controlled) extents of immobilization of antibodies at a suitably modified gold surface. Experimental Section Materials. All chemicals and solvents were used as supplied by Sigma (Poole U.K.). Metals were obtained from Goodfellows (Cambridge, U.K.). In addition, a gold disk working electrode (diameter 2 mm) and an Ag|AgCl reference electrode were both obtained from BioAnalytical Systems (Luton, U.K.). Formation of Mixed Monolayers. Polycrystalline gold working electrodes were cleaned by polishing with 0.3 a µm alumina slurry, sonication in reverse osmosis (RO) water for 15 min, and then treatment for 2 min in a solution of freshly prepared 1:4 piranha solution (30% H2O2:98% H2SO4) prior to finally rinsing thoroughly with RO water. Alternatively, gold substrates were prepared for SAM formation by electron-beam evaporation of a Ti/Pd/Au (10/10/100 nm) multilayer structure onto glass. Immediately prior to deposition of the thiol monolayer, the gold surface was cleaned by a reactive ion etch for 5 min (50 mT, 20 sccm, and 10 W) first in an oxygen plasma and then in an argon plasma using a PlasmaFab ET340. In either case, after cleaning, the gold surface was immediately immersed in a 1 mM solution of an initial thiol modifier (i.e. mercaptoethylamine) for 2 h and then thoroughly rinsed using RO water. Mixed monolayers were subsequently prepared by drying the surface in a stream of nitrogen and then immersing them in a second thiol solution (i.e. 1 mM mercaptopropanol) for between 10 s and 180 min, before rinsing. Characterization of the Mixed Monolayers by XPS. All XPS spectra were collected using a Scienta ESCA300 spectrometer, with a rotating anode source providing Al KR (1486.7 eV) radiation. After monochromation the radiation is focused on the sample, positioned such that the emitted electron takeoff angle (TOA) is 10 ( 3° relative to the substrate surface. Selfassembled monolayer films were prepared as described above, and S(2p), C(1s), N(1s), and O(1s) spectra were acquired using (26) Ulman, A. In An Introduction to Ultrathin Organic Films; Academic Press: Boston, MA, 1991. (27) Bartlett, P. N.; Tebbut, P.; Whitaker, R. G. Prog. React. Kinet. 1991, 16, 55. (28) Allen, H.; Hill, O.; Page, D. J.; Walton, N. J. J. Electroanal. Chem. 1987, 217, 141.
Cotton et al. the Scienta Software. The slit width (0.8 mm) and the TOA were kept constant for each of the samples measured, to probe each sample to the same depth. Prior to data fitting (SigmaPlot, Jandel Scientific, Poole, U.K.), the binding energies of the spectra were adjusted to compensate for sample charging (note: the insulating glass substrates underneath the evaporated Au surface necessitated the use of a flood gun). The Au(4f7/2) spectrum was chosen as the reference binding energy (84.00 eV). Electrochemical Characterization of Mixed Monolayers. Electrochemical measurements were performed using a BAS-CV37 potentiostat, and recorded on a Goerz Servogor 780 X-Y chart recorder. Experiments were performed in an all-glass cell which had a working volume of 1 mL, incorporating a conventional three-electrode configuration with a 2 mm diameter polycrystalline gold working electrode, a fixed counter electrode of platinum wire, and an Ag|AgCl reference (3.0 M NaCl, +190 mV vs NHE). Cyclic voltammetry was carried out at potentials between -200 and +200 mV using scan rates (v) of between 5 and 100 mV s-1 in a solution of 700 µM horse heart cytochrome c (Sigma type VI) (supporting electrolyte of 20 mM sodium phosphate buffer, pH 7.0, with 100 mM sodium perchlorate), at a temperature of 22 ( 2 °C. Immobilization of Fluorescent Antibodies. Gold slides were prepared and modified as for XPS analysis. Immobilization of 20 µg/mL fluorescein isothiocyanate (FITC) labeled goat antirabbit IgG was performed following an optimized glutaraldehyde cross-linking procedure as described by Williams and Blanch.1 Following immobilization, the samples were examined using a Nikon Microphot fluorescent microscope (FITC: λex) 494 nm, λem) 520 nm). Relative fluorescence was estimated as the “in focus” exposure time at a fixed aperture when photographing the surface (thus using the camera to integrate the average fluorescent signal over an area of 100 × 100 µm). Exposure times for unmodified mercaptoethlyamine SAMs were used to correct for background light levels in the above fluorescence measurements.
Results and Discussion A. XPS of Homogeneous and Mixed SAMs. To determine any changes in molecular composition as a consequence of immersing a mercaptoethylamine/Au SAM in a solution of mercaptopropanol, we performed XPS measurements at a TOA of 10° to record the C(1s), N(1s), S(2p), and O(1s) spectra. This low TOA had the advantage of reducing the background signal due to emission of Au electrons and maximizing the surface area of the SAM that was sampled. As a consequence, it was possible to collect spectra with a good signal-to-noise ratio before significant X-ray damage to the SAM occurs. Typical collection times for the spectra presented here were 40 min, and examination of XPS spectra obtained over varying lengths of time, up to 4 h after irradiation, showed little change. However, it should be noted that collection of spectra at a low TOA angle highlights atomic centers closest to the sample/air(vacuum) interface; two consequences of this are, first, the requirement for stringent precautions to minimize the inevitable surface contamination, and, second, that the use of elemental quantification (based on bulk sensitivity factors) does not give the stoichometric composition of the SAM (discussed further below). Two methods of quantification of the mixed SAM composition were employed, namely integration of N(1s) signals and spectral decomposition of C(1s) spectra. While the N(1s) spectra are simple Gaussian peaks, the complex features in the C(1s) spectra of homogeneous SAMs are discussed in detail before presenting the spectra arising from mixed SAMs. Influence of Low TOA in Elemental Quantification of SAMs. A consequence of the ordered nature of SAM interfaces is that particular parts of the immobilized molecule will always be nearer the immobilizing surface (here Au), and other parts nearer the air (or solution)
Formation of Mixed Alkanethiol Monolayers
Langmuir, Vol. 14, No. 18, 1998 5141
interface. When XPS measurements are performed at low TOAs, this well-defined compositional variation in the plane normal to the SAM surface means that significantly less photoelectrons escape from the atomic centers which anchor the SAM compared to those from the SAMs headgroup. A simple model which describes this variation in measured photoelectrons intensity with SAM thickness and TOA is based on the Beer-Lambert law: denoting the photoelectron intensity at the surface of a film as I, and the intensity of the source a distance (d) from the film’s interface as Io, then
I/Io ) exp (-d/(λ sin θ))
(1)
where λ is the mean free path of the emitted photoelectron in the material being examined (ca. 25 Å for the transitions of this study) and θ is the TOA, or escape angle, of the photoelectrons (with respect to the sample surface).29 In the SAMs under consideration here, the sulfur tail group is approximately 4.5 Å further from the photoelectron escaping interface than is the headgroup. If, for simplicity, we assume the part of the SAM above the sulfur tail group can be represented by a complete, homogeneous layer 4.5 Å thick, then photoelectrons measured with a TOA of 10° will pass through 4.5/sin(10°) Å of material before escaping. Using the above formula (eq 1), this predicts that the signals from the underlying sulfur tail groups will be attenuated by ca. 65%, when compared to the magnitude they would have if they were at the sample/ vacuum interface. This attenuation of photoelectron intensities is considered below in both the deconvolution of homogeneous SAM C(1s) spectra and the ratio of N(1s) to S(2p) photoelectron counts in appropriate SAM systems. Electronic Effects in Homogeneous SAM C(1s) Spectra. Curve fitting to the C(1s) spectra for the singlecomponent SAMs (parts a and b of Figure 1) and thin film mercaptoethylamine (Figure 1c) show the distinct shifts for the carbon centers bonded to either the tail group (SAu, S-H), or the headgroup (NH2 or OH), Table 1. Note that mercaptopropanol, being a liquid, was not amenable to XPS measurements of the bulk phase in this study. The binding energy of the C(1s) electrons associated with the carbon center adjacent to the sulfur tail group (peaks 1 in Figure 1) is influenced by the electronegativities of the headgroup (O or N) and the adjacent S species, together with (where appropriate) the significant electron donation effect of the electropositive Au surface. The former effect raises the relevant C(1s) binding energy, whereas the latter effect lowers it. Thus, in these short chain SAMs, the domination of the Au electron-donating effect causes the C(1s) binding energy of these tail group carbon centers to be shifted more negative compared to that corresponding to general aliphatic carbon centers (ca. 285 eV). This effect is clearly illustrated in the case of mercaptoethylamine by comparison of peak 1 positions in Figure 1b and Figure 1c, which show a 1.1 eV reduction in this C(1s) binding energy for the SAM species (the large peak in Figure 1b at ca. 285.1 eV corresponds to a carboncontaining contaminating overlap always present to some degree in mercaptoethylamine SAMs). The small low binding energy peak at 284.6 eV in Figure 1c corresponds to Au-bound mercaptoethylamine observed through “thin” regions of the bulk film when a 90° TOA is used. As expected the proximity of the electron-withdrawing headgroups in mercaptoethylamine and mercaptopropanol (29) Hofmann, S. In Practical Surface Analysis; Briggs, D., Seah, M. P., Eds.; J.Wiley and Sons: Chichester, U.K., 1993.
Figure 1. Curve fits to (a) mercaptopropanol/Au, (b) mercaptoethylamine/Au, and (c) a thin multilayer film of mercaptoethylamine on Au surface C(1s) spectra to a minimal number of Gaussian peaks. Numbered peaks refer to carbon centers as detailed in insert; the large peak in part b corresponds to a contaminating overlayer, and a small peak at 284.6 eV in part c corresponds to the region of Au/mercaptoethylamine SAM. (TOA 10° for spectra a and b; TOA 90° for spectrum c).
5142 Langmuir, Vol. 14, No. 18, 1998
Cotton et al.
Table 1. C(1s) Binding Energies for Different Carbon Centers in Homogeneous Mercaptopropanol and Mercaptoethylamine Systemsa
a Peak positions were determined by deconvolution of spectra in Figure 1 using Gaussian basis curves. Estimates of distances between carbon centers and head/tail groups correspond to simple summation of the appropriate “standard” bond lengths; i.e., no account is taken of the bond angles.
Figure 2. N(1s) spectra for mercaptoethylamine/Au SAMs after immersion in mercaptopropanol solution for varying lengths of time (1, 10, 20, and 90 min, a-d), together with a mercaptopropanol/Au SAM which has been left in RO water for 50 min (e) (TOA 10°).
to the tail group C further influences the latter’s measured binding energy (Table 1): mercaptoethylamine SAMs (N two bonds away from the “tail” carbon) have a tail C(1s) binding energy of 284.6 eV, whereas mercaptopropanol (with an oxygen atom three bonds away from the tail carbon) has a lower tail C(1s) binding energy of 284.3 eV, despite the more electronegative character of oxygen compared to nitrogen. In mercaptoethylamine SAMs, a similar interplay of proximity effects and the electropositive character of the Au surface causes the binding energy of the headgroup C (peak 2 in Figure 1b) to be fractionally lower (-0.1 eV) than in the bulk species (peak 2 in Figure 1c); see Table 1. The influence of the electron donating Au surface on the electron distribution in these SAMs is however further manifested in a more significant lowering of the N(1s) binding energy in the SAM (399.1 eV), Figure 2, from that found in the bulk (400.0 eV) (not shown). The increase in electron density around the head N group as a consequence of anchoring to the Au surface implied by this measurement suggests an increase in basicity of the SAM species when compared to the same species in solution.
Significant nitrogen basicity of the mercaptoethylamine SAM is also evidenced by the finding that the N(1s) spectra frequently have a high binding energy shoulder, indicating the presence of a number of N centers bearing a positive charge. In general, associated with the observations of charged N centers, is a higher than average amount of surface contamination (corresponding to the peak at ca. 285.1 eV noted above in Figure 1b). Thus far, it has proved impossible to prepare homogeneous mercaptoethylamine SAMs which contain very low levels of C contamination, in complete contrast to, e.g., SAMs of mercaptopropanol, mercaptopropionic acid, mercaptopropane, etc. handled and prepared in a similar manner. In fitting these spectra using Scienta software, the Gaussian peaks were constrained to be of increasing areas for centers closer to the non-gold interface of the SAM in accord with the discussion of low TOA quantification above. Additionally, the peak area corresponding to the carbon center adjacent to sulfur was constrained to be less than that derived from quantification of the sulfur species on the Au surface (by integration of the S(2p) spectrum). While the instrumental resolution and closeness of C(1s) binding energies leads to uncertainties in the fitting of Gaussian peaks, it was noted that sets of curves with similar goodnesses of fit vary most in the relative peak heights, rather than peak positions and halfwidths. The narrowness of the fitted peaks for the carbon centers adjacent to the sulfur center (half-width ca. 0.7 eV) when compared to that of the headgroup carbon centers (>1.0 eV) may be attributable to their reduced vibrational freedom (similarly the half-width of sulfur centers bonded to the gold surface is markedly less than that found for physisorbed species). Evaluation of SAM Composition Based on N(1s) Spectra. Figure 2 shows N(1s) spectra for mercaptoethylamine SAMs following immersion for increasing lengths of time in a mercaptopropanol solution. Likewise, the N(1s) spectrum for a mercaptopropanol SAM which had been left soaking in water for an extended length of time is given (Figure 2), illustrating the absence of contamination by nitrogen containing species during a typical SAM treatment. These spectra illustrate that immersion in mercaptopropanol solution leads to a loss of nitrogen functionalities from the surface. The time course of reduction of the nitrogen spectra (as a consequence of displacement by mercaptopropanol) is given in Figure 3 which shows the change in proportion
Formation of Mixed Alkanethiol Monolayers
Figure 3. Ratio of N(1s) to S(2p) photoelectron counts for mercaptoethylamine/Au SAMs after different lengths of immersion time in a mercaptopropanol solution. Error bars estimated from variations in fitting the underlying baseline to the N(1s) and S(2p) spectra. (TOA 10°).
Figure 4. Evolution of C(1s) spectra for mercaptoethylamine/ Au SAMs on immersion in mercaptopropanol solution for varying lengths of time (b, 2 min; c, 20 min; d, 90 min), together with the C(1s) spectra for homogeneous mercaptoethylamine/ Au (a) and mercaptopropanol/Au SAMs (e) (TOA 10°). The spectra are offset for clarity.
of N(1s) ejected photoelectrons to those originating from S(2p) centers corresponding to the SAM tail group. This normalization of N(1s) counts by S(2p) counts corrects for small variations between samples in total SAM coverage, surface contamination on top of SAMs, and instrumental parameters (e.g. TOA and X-ray flux). This estimation of the time variation in the SAM’s surface functionality compares well with the additional methods of SAM composition evaluation described below (Figure 6), and the measured biomolecular recognition responses (Figures 9 and 10). The attenuation of photoelectron intensity (at low TOA) with film thickness discussed above is illustrated by the ratio of N(1s) to S(2p) photoelectron counts for a homogeneous mercaptoethylamine SAM (Figure 3). This ratio was measured to be ∼1.36, which, when bulk sensitivity factors (1.73 and 2.08, respectively) are used to calculate
Langmuir, Vol. 14, No. 18, 1998 5143
Figure 5. Spectral decomposition of mercaptoethylamine/Au SAM which has been immersed in a mercaptopropanol solution for 5 min. Curve a corresponds to the measured spectrum. The component spectra corresponding to proportioned library spectra are as follows: mercaptopropanol/Au (b) and mercaptoethylamine/Au (c) (derived from the peak fitting of Figure 1b). The composite spectrum is labeled d with the residual spectrum labeled e (TOA 10°).
Figure 6. Percent mercaptoethylamine remaining in mercaptoethylamine/Au SAMs after their immersion in mercaptopropanol for varying lengths of time as evaluated by spectral decomposition of C(1s) spectra (TOA 10°).
the elemental composition gives a N:S atomic ratio of 1.6:1 (rather than 1:1). This “high” N:S ratio confirms that the interfaces we are measuring are ordered with the S group further from the outer interface than is the N group. It is interesting to note that we have observed that the N:S ratio more closely approaches 1:1 when examining nitrogen-containing SAMs which have been either poorly rinsed or participate in strong electrostatic interactions with deposition solution species leading to additional physisorbed layers (e.g. cysteine, N-acetylcysteine). This reduced N:S ratio corrolates with a significant amount of physisorbed thiol on top of an underlying SAM (judged by the multiple values of S(2p) binding energies) and is likely to reflect the more random orientation of molecules in the interface. Examination of C(1s) Spectra for Variously Prepared SAMs. The above method of analysis (using the
5144 Langmuir, Vol. 14, No. 18, 1998
Cotton et al.
N(1s):S(2p) photoelectron ratio) does not directly indicate whether the loss of nitrogen functionality from the surface on immersion of a mercaptoethylamine SAM in a mercaptopropanol solution is concomitant with the latter’s adsorption. Thus we collected the corresponding C(1s) spectra to compare with spectra from SAMs prepared using either mercaptopropanol or mercaptoethylamine alone (Figure 4). The changes in shape of these C(1s) envelopes indicate a progression of the monolayer composition from that of pure mercaptoethylamine to one containing a significant proportion of mercaptopropanol. Evaluation of Mixed SAM Composition Using C(1s) Spectra. Two approaches to the deconvolution of C(1s) spectra for the “mixed” monolayer SAMs involve the fitting of the C(1s) envelopes either to linear combinations of Gaussian peaks or linear combinations of the twocomponent spectra (derived from the homogeneous mercaptoethylamine and mercaptopropanol SAMs). While, in the former approach, the half-widths, their positions, and their relative heights can be partially constrained to be close to those found in the homogeneous mercaptoethylamine and mercaptopropanol SAM, the large number of parameters (18 in this case) requires complex fitting procedures to determine which of the “best” fits most closely represent the “true” solution. The second approach, in which library spectra (corresponding to homogeneous mercaptopropanol and mercaptoethylamine SAMs) are used in the spectral decomposition process, is a widely used technique and one we successfully exploited in the determination of enzyme concentrations in polymer films via their C(1s) spectra.30 When an attempt is made to decompose the mixed monolayer SAM spectra using combinations of the “asmeasured” two library spectra in an unconstrained manner, it was found that the presence of an additional component spectrum due to surface contamination led to a degenerate set of solutions. If compared to our previous XPS examinations of enzyme/polymer systems,30 the additional surface contamination spectral component is significantly larger. Thus, we sought to apply a constraint to the proportions of SAM library spectra using the magnitude of the N(1s) signal for a particular mixed monolayer SAM. Consequently, the C(1s) spectral decomposition (Figure 5) was performed using the following algorithm: a difference spectrum was created by subtracting a library mercaptoethylamine SAM spectrum (which had been proportioned according to the magnitude of the N(1s) signal in the mixed monolayer) from the spectrum of the mixed monolayer SAM. This difference spectrum was then fitted using a mercaptopropanol SAM C(1s) library spectrum such that the residual spectrum contained a minimal intensity in the regions 286-287 and 283-284.5 eV. Thus the final residual spectrum corresponded to a Gaussian peak centered on 285 eV (which arose from the (variable) surface contamination). Two formulations of the mercaptoethylamine library spectrum were employed in separate fitting attempts: either the “as measured” spectrum of Figure 1b, or a spectrum synthesized from the curve fitting of Figure 1b (i.e. corresponding to a library spectrum lacking in the contamination component at ca. 285 eV). Quantifications of the mixed SAM composition estimated using either of these mercaptoethylamine library spectra were generally within 5%, and the results presented in Figures 5 and 6 correspond to the usage of the synthetically derived mercaptoethylamine library spectrum.
The validity of this constrained spectral decomposition approach relies on the low level of nitrogen in any surface contamination resulting from the experimental procedure (as evidenced by trace e in Figure 2). Thus, after the determination of the relative mercaptoethylamine and mercaptopropanol components, as outlined above, the proportion of mercaptoethylamine was allowed to vary by 5% and the mercaptopropanol spectrum was refitted; this leads to the error bars shown in Figure 6 for the variation with immersion time of the mercaptoethylamine fraction of the total SAM content (i.e., mercaptoethylamine plus mercaptopropanol). Comparison of Figures 3 and 6 show that both methods of XPS analysis of the SAM composition give similar estimates of the mercaptoethylamine content (the data of Figure 3 can be normalized to the N(1s):S(2p) photoelectron count ratio for the homogeneous mercaptoethylamine SAM). However, the requirement to use a mercaptopropanol spectrum in calculating the data of Figure 6 supports the assertion that a mercaptopropanol/mercaptoethylamine mixed SAM is being formed. An error in the interpretation of the above quantifications and deconvolutions would arise if either of the two thiol species was physisorbed rather than chemisorbed onto the underlying SAM interface. Thus it should be noted that the S(2p) signals for all these monolayers indicate >95% of the sulfur species have a binding energy appropriate for Au-S bonding i.e., ∼162 eV (as compared to 163 eV for physisorbed species). This lack of physisorbed thiol is in contrast to our (and others31) observations for carboxylate terminated or other very polar SAMs. In this respect it is noticeable that the amount of surface contamination on the mercaptopropanol SAMs is markedly less than for other short chain SAM systems we have examined, which may suggest its use as an antifouling layer on metal electrode surfaces. Reaction Mechanism in the Formation of Mixed SAMs. In seeking a mechanistic explanation for the displacement of a mercaptoethylamine SAM by solution phase mercaptopropanol (the reverse displacement is not observed; see also electrochemical observations, later), we found that the S(2p) binding energy in mercaptopropanol SAMs was 0.1 eV lower (at 162.0 eV) than that in mercaptoethylamine SAMs. These values compare with the 161.3 eV binding energy that we have measured for Na2S derived SAMs and 163 eV for physisorbed species (data not shown). The higher binding energy for the S(2p) electrons in mercaptoethylamine SAMs indicates that the outer S(3s) and S(3p) electron densities are further away from the sulfur center (a consequence of the combined electron-donating and -withdrawing effects of atoms bonded to the sulfur species). When this observation is combined with the higher binding energy measured for the tail group C(1s) electrons in the mercaptoethylamine SAM (compared to that for the mercaptopropanol SAM), this indicates that the electrons in the S-Au bond are centered closer to the Au species, perhaps reflecting a longer, and consequently weaker, S-Au bond in the mercaptoethylamine SAM, as compared to that for the mercaptopropanol SAM. Thus, if (in the thiol-containing modification solution) there is a transition state in which a disulfide bond exists between the solution species and the thiol of the SAM, the following dissociation of the transition state would lead to a SAM composed of the thiol species with the stronger S-Au bond; i.e., in the present case, a mercaptopropanol SAM. Further
(30) Griffith, A.; Glidle, A.; Beamson, G.; Cooper, J. M. J. Phys. Chem. B 1997, 101, 2092.
(31) Uvdal, K.; Bodo, P.; Liedberg, B. J. Colloid Interface Sci. 1992, 149, 162.
Formation of Mixed Alkanethiol Monolayers
investigation of this proposed mechanism requires examination of far infrared reflectance, or Raman, spectra in the region corresponding to Au-S vibrations. An indication that the changing composition of the SAM from mercaptoethylamine to mercaptopropanol is not via an initial dissociative step comes from our observations that the N(1s) signal does not diminish appreciably on immersion of the mercaptoethylamine SAM in (thiol-free) water of various pHs, for similar extended periods of time. In summary, the XPS measurements on mercaptoethylamine SAMs immersed in mercaptopropanol solutions indicate that by using readily manageable solution concentrations (millimolar) and immersion times (minutes) it is possible to control the proportions of these two species in the resultant mixed SAM. Since prolonged immersion leads to larger C contamination levels, which hamper C(1s) deconvolution and obscure N(1s) signals, it is not possible to state if the level of mercaptoethylamine still present in the SAM after 90 min represents an equilibrium or a kinetic state. Confidence in the determination of the fractional mercaptoethylamine or mercaptopropanol content in the SAMs comes from the similarity of the results obtained from both C(1s) deconvolution and N(1s) quantification. A further point to note is the absence of an increase in the high binding energy region of the C(1s) spectra for the treated mercaptoethylamine SAMs, indicating the nitrogen functionality has not been lost by oxidation to a carboxylate species. B. Electrochemistry of Cytochrome c at Mixed Monolayers. The ability of alkanethiols bearing certain headgroups to promote electron transfer at gold electrodes is well documented. For example, over a decade ago, Allen et al. produced a comprehensive survey of over 50 thiolcontaining molecules studied for their ability to promote the electrochemistry of cytochrome c at gold electrodes.13 Their studies suggest that for a compound to be an effective electron-transfer promoter, it is essential that it has the ability to form hydrogen bonds with the terminal amine groups on the lysine residues which surround the heme crevice on cytochrome c. In this respect, it is clear that the mercaptopropanol has an ability to promote the protein redox chemistry (see Figure 7), while the protonated NH3+ group of mercaptoethylamine that exists at pH 7.0 does not. Generally the rate of electron transfer between an electrode and many metalloproteins is extremely slow and as expected the distinctive electron-transfer peaks of horse heart cytochrome c electrochemistry are not observed at an unmodified gold electrode, Figure 7a.13 Figure 7b, however, shows the voltammetry which is observed at a gold electrode which has been modified with mercaptopropanol, with well-defined, diffusion-controlled, reversible cytochrome c electrochemistry (the peak to peak separation (∆E) ) 61 mV and the ratio of the anodic and cathodic currents (Ipa/Ipc) ) 1, both indicating a high degree of reversibility). In common with other studies,13 the size of both the cathodic and anodic peaks were proportional to v1/2 in the range 10-100 mV s-1, indicating that the electrochemical reaction is diffusion controlled. In contrast, electrodes modified with mercaptoethylamine gave poor voltammetric responses, Figure 7c, ∆E ) 95 mV, suggesting slow electron transfer. As a consequence we sought to use cytochrome c electrochemistry to probe the molecular composition of the mixed monolayer. The change in the peak anodic current (Ipa) following immersion of the mercaptoethylamine SAM in a mercaptopropanol solution for differing lengths of time was sought to corroborate information obtained from the above XPS measurements.
Langmuir, Vol. 14, No. 18, 1998 5145
Figure 7. Representative cyclic voltammograms of a 700 µM solution of horse heart cytochrome c, recorded at a gold electrode (surface area ) 3 mm2), v ) 50 mV s-1. Key: (‚‚‚) the scan of a clean unmodified electrode, a; (s) the voltammetry observed at an electrode modified by incubation in 1 mM aqueous solution of mercaptopropanol for 30 min, b; (- - -) the response of an electrode modified in an aqueous solution of mercaptoethylamine for 30 min, c. Electrolyte conditions: 20 mM sodium phosphate, adjusted to pH 7.0; 100 mM sodium perchlorate.
Figure 8. Cyclic voltammograms of a cytochrome c solution obtained when a gold electrode modified with mercaptoethylamine (surface area ) 3 mm2), was placed in a 1 mM aqueous solution of mercaptopropanol for repeated periods of time (0120 min). Increasing peak heights occur for increasing lengths of immersion time; see Figure 9 for precise time values. Solution and instrumental parameters were as for Figure 7.
In this context, Figure 8 shows the consequence of immersion of a mercaptoethylamine modified electrode in a 1 mM mercaptopropanol solution, over a period of time between 10 s and 180 min. Two main effects are observed: first the peak currents (Ipa or Ipc) increase in size as a function of the time of exposure to mercaptopropanol, Figure 9; second, the separation of these peaks decreases with time. Both results are indicative of an increase in the rate of biological electron transfer, an observation which can be attributed to the XPS findings that mercaptopropanol (a better electron-transfer promoter) displaces mercaptoethylamine.
5146 Langmuir, Vol. 14, No. 18, 1998
Cotton et al.
mobilized on a homogeneous mercaptoethylamine SAM. Figure 10 shows that the fluorescence observed, after the immobilization of antibody, decreases with the time that the mercaptoethylamine-modified gold sample is immersed in mercaptopropanol. This decrease in fluorescence is consistent with a fall in the number of available surface amine groups as demonstrated by the XPS measurements above. Conclusion
Figure 9. Variation with immersion time of the anodic peak current for electrodes prepared and treated as in Figure 8.
Figure 10. Change in fluorescence of mercaptoethylaminemodified gold samples after immersion in 1 mM mercaptopropanol for a specified time period, followed by the glutaraldehyde immobilization of FITC-labeled goat anti-rabbit IgG. (λex 494 nm - λem 520 nm)
Interestingly, an attempt at the “reverse” displacement, involving the immersion of a mercaptopropanol-modified electrode in a solution of mercaptoethylamine, had no effect on the size or shape of the resulting voltammogram. This suggests that the mechanism of displacement is not reversible and is consistent with our XPS measurements for similarly prepared and treated SAMs (data not shown). C. Immobilization of Fluorescent Antibodies. The displacement of mercaptoethylamine by mercaptopropanol was used to manipulate the number of potential sites (NH2 headgroups) for protein binding to the surface, and thereby control the amount of fluorescently labeled antibody (IgG), immobilized through via glutaraldehyde coupling. Measurement of the amount of fluorescence present was determined by recording the intensity of emitted fluorescence (from the exposure time for a Nikon camera) at a fixed aperture with appropriate filters. Results were corrected using background fluorescence values for the native gold modified with mercaptoethylamine alone followed by scaling to the fluorescence from IgG im-
We have shown that the degree to which a surface can be shown to interact with biological molecules can be controlled by the manipulation of the ratio of functional groups on that surface. The responses in the electrochemical and fluorescence systems above are favored by different (OH or NH2) headgroups in the “mixed” SAM. The SAMs were probed with high-resolution XPS to establish that they were of “mixed” composition, with the second component coming from the solution thiol species (mercaptopropanol). Conversly, this paper illustrates that not only is it possible to use displacement reactions to tailor an appropriate biointerfacial arrangement but also that highly sensitive protein-surface interactions can, themselves, be used to monitor the changes in the molecular structure of a self-assembled monolayer. It is interesting to note, however, that in order to transform the interfacial properties of a surface with respect to certain bioelectrochemical processes, it is not necessary to produce a homogeneous surface coating; i.e., under the electrochemical conditions employed, cyt-c redox currents similar to those on suitable homogeneous SAMs are obtainable even when the mixed SAM surface contains a significant proportion (40%) of “nonpromoter” functionalities. This latter observation may be analogous to the macroscopic electrochemical response found for interconnected arrays of microelectrodes randomly distributed on a surface, when slow scan rates are used. Finally, it is interesting to note that the displacement reactions by solution phase thiols are not necessarily reversible on the moderately long time scales explored here. Thus exposure of a gold electrode surface to a binary thiol solution may not lead to a mixed monolayer SAM of the same relative composition as the solution. Of further importance to the results presented here, our preliminary measurements on other SAMs used in bioelectrochemical interfaces suggest that in addition mercaptopropionic acid will readily displace mercaptoethylamine from an Au surface, whereas mercaptopropanol does not displace mercaptopropionic acid or vica versa.32 Consequently, through the determination of a hierarchy corresponding to the ability of a solution thiol to displace a surface species, it may be possible to readily create SAMs containing controlled quantities of multiple thiols. Acknowledgment. The authors wish to thank both Glaxo-Wellcome and The Leverhulme Trust for supporting of this work. The XPS, carried out at RUSTI, was supported by the EPSRC, through their funding of the CCLRC. LA980321C (32) Cotton, C.; Glidle, A.; Jiang, L.; Beamson, G.; Cooper, J. M. Manuscript in preparation.
