Adsorption of Alkylthiol-Capped Gold Nanoparticles onto Alkylthiol

1048

Langmuir 2006, 22, 1048-1054

Adsorption of Alkylthiol-Capped Gold Nanoparticles onto Alkylthiol Self-Assembled Monolayers: An SPR Study Michal Goren, Natalie Galley, and R. Bruce Lennox* Department of Chemistry and the FQRNT Center for Self-Assembled Chemical Structures (CSACS), McGill UniVersity, 801 Sherbrooke St. West, Montreal, Quebec, Canada H3A 2K6 ReceiVed September 10, 2003. In Final Form: June 25, 2004 The kinetics of alkylthiol-capped gold nanoparticle (RS/Au-NP) adsorption to alkylthiol/Au self-assembled monolayers (RS/Au-SAMs) has been studied using SPR (surface plasmon resonance) spectroscopy. Variation of the alkylthiol chain terminus (CH3, COOH) and solvent (H2O, hexane) provides insight into the relative importance of solvation energies (in the context of surface energies) and RS/Au-NP structure on adsorption kinetics. The kinetics, fitted to the Langmuir kinetic model, yield adsorption and desorption rate constants. ∆Gads derived from kinetic data are compared to calculated values of work of adhesion (Wadh), derived from the corresponding surface energies. The absence of a ∆Gads - Wadh correlation arises because the measured kinetics are not reporting the approach to equilibrium and/or because there are factors (i.e., local chain packing defects, surface hydration differences, or particle-particle interactions) not accounted for in calculated Wadh values.

Introduction Most self-assembly phenomena, templating phenomena, and nonspecific adsorption processes of particles or molecules onto surfaces are mediated either by specific chemical bonds or by nonspecific electrostatic interactions. Examples include selfassembled monolayers (SAMs) of thiols on gold,1 SAMs of silanes on silicon,2 and nanoparticles on SAMs.3,4 Templating of ions or charged particles on both charged surfaces5,6 and on surfaces displaying chemical contrast (i.e., spatially distinguished regions of two or more functional groups) have also been documented.7-9 There are, however, fewer examples where selective templating occurs on surfaces whose contrast originates from surface energy contrast rather than chemical contrast.10,11 Such energy differences usually originate from van der Waals interaction differences that are often small in magnitude in comparison to chemical bonding differences.12 How differences in surface energy lead to differences in adsorption kinetics are particularly interesting. This study addresses the issues of surface energy contrast by monitoring the in situ adsorption kinetics of RSH/Au-NP onto a 2D SAM surface using SPR (surface plasmon resonance) spectroscopy. The adsorption and desorption rate constants, and thus ∆Gads, are experimentally determined. Experimental Section Chemicals. 1-Hexadecanethiol and 16-mercaptohexadecanoic acid were obtained from Dr. Antonella Badia (University of Montreal). * To whom correspondence [email protected].

should

be

addressed.

E-mail:

(1) Nuzzo, R. G.; Fusco, F. A.; Allara, D. L. J. Am. Chem. Soc. 1987, 109, 2358-2368. (2) Sagiv, J. J. Am. Chem. Soc. 1980, 102, 92-98. (3) Shipway, A. N.; Lahav, M.; Willner, I. AdV. Mater. 2000, 12, 993-998. (4) Auer, F.; Scotti, M.; Ulman, A.; Jordan, R.; Sellergren, B.; Garno, J.; Liu, G. y. Langmuir 2000, 16, 7554-7557. (5) Dujardin, E.; Peet, C.; Stubbs, G.; Culver, J. M.; Mann, S. Nano. Lett. 2003, 3, 413-417. (6) Monson, C. F.; Wooley, A. T. Nano. Lett 2003, 3, 359-363. (7) Burkett, S. L.; Mann, S. J. Chem. Soc. Commun. 1996, 3, 321. (8) Goren, M.; Qi, Z.; Lennox, R., B. Chem. Mater 2000, 12, 1222-1228. (9) Goren, M.; Lennox, R. B. Nano. Lett. 2001, 1, 735-738. (10) Moraille, P.; Badia, A. Angew. Chem. Int. Ed. 2002, 41, 4303-4306. (11) Noll, J. D.; Nicholson, M. A.; van Patten, P. G.; Chung, C.-W.; Myrick, M. L. J. Electrochem. Soc. 1998, 145, 3320. (12) Israelachvili, J. Intermolecular & Surface Forces, 2nd ed.; Academic Press: New York, 1997.

1-Octanethiol, 1-iodonaphthalene, hexane (Spectrograde), and LiClO4 were purchased from Aldrich and used without further purification. RS/Au-NP were synthesized in our lab by Lawrence Lee using a two phase synthesis procedure.13,14 The CH3-terminated NP were C14S/Au, and the COOH-terminated NP were either HOOCC15S/ Au-NP or HOOCC10S/Au-NP. C14S/Au-NP and HOOCC15S/ Au-NP were soluble in hexane and HOOCC10S/Au-NP was soluble in water. The pH was adjusted using either dilute nitric acid and dilute NaOH. Instrumentation and Procedures. Surface Plasmon Resonance Spectroscopy (SPR). SPR measurements were carried out with a computer-controlled, scanning angle apparatus from Resonant Probes GmbH (Mainz, Germany). Surface plasmons are excited using the Kretschmann configuration where a right angle prism (LaSFN9 glass, Hellma Optik) is index-matched (1-iodonaphthalene, n ) 1.701) to a gold-coated LaSFN9 substrate (n632.8nm ) 1.845, Hellma Optik). The prism/Au substrate assembly is then fixed against one side of a custom-built Teflon liquid cell (1 mL capacity) fitted with Kalrez O-rings so that the Au film is exposed to the solution. The other side of the liquid cell is pressed against a microscope glass slide. The SPR excitation source is a linearly polarized HeNe laser (output power ) 5 mW, λ ) 632.8 nm) from JDS Uniphase (model 1125P). The laser intensity is attenuated using a sheet polarizer before passing through a Glan-Thompson polarizing prism (Halle) which is set for p-polarized incident light with respect to the plane of incidence on the gold surface. An optical chopper (Perkin-Elmer model 197) is used to modulate the optical signal at a frequency of 1 kHz, which is then correlated with detection through a lock-in-amplifier (EG&G PAR model 5210). Light reflected from the metal/prism interface is focused onto a silicon photodiode detector. The photodiode signal is measured with the lock-in-amplifier that is in-phase with the excitation source. A stepper-motor driven θ/2θ goniometer (Huber model 414a) and motor control unit allows the incident angle to be varied in increments g0.01°. The prism/Au substrate/liquid cell assembly is mounted onto the goniometer stage so that the center of the Au/glass substrate is at the axis of rotation. Alignment was performed at 45°, ensuring that the back reflected beam overlaps with the incident beam. The goniometer and data acquisition were controlled through an IEEE interface board (Keithley model KPC488.2) and software provided by Resonant Probes. (13) Brust, M.; Walker, M.; Bethell, M.; Schiffrin, D. J.; Whyman, R. J. Chem. Soc., Chem. Commun. 1994, 7, 801-802. (14) Brust, M.; Fink, J.; Bethell, D.; Schiffrin, D. J.; Kiely, C. J. Chem. Soc., Chem. Commun. 1995, 8, 1655-1656.

10.1021/la035688i CCC: $33.50 © 2006 American Chemical Society Published on Web 12/24/2005

Adsorption of Functional Gold Nanoparticles SPR curves were obtained by measuring the reflected light intensity as a function of the incident angle, θ. These angular reflectance curves were analyzed using a multilayer Fresnel optical model generated via the Resonant Probes software. The refractive index of the solvent was fitted according to the angle of total internal reflection.15 Kinetic adsorption data were obtained by tracking the angle of minimum reflectivity over time with a temporal resolution of 20 s (this is the time required for the SPR instrument to obtain the minimum angle). Tracking the minimum angle is achieved by measuring the reflectivity at three points close to the overall reflectance minimum and fitting these to a parabola. The apex of the parabola is considered to be the coupling angle. However, the shape of the SPR curve minimum is more complex than a simple parabola, so the derived coupling angle depends on the angle increment between the three points used for tracking. We have found that the coupling angle observed at saturation coverage in angle tracking mode is very close to that obtained from a full SPR curve acquired at the end of a kinetic run. There is however a limitation to this procedure when the initial minimum angle shift is large. In this case the instrument is unable to determine a minimum, and another cycle of 3 reflectivity points measurement is performed. This doubles the time interval between measurements in these cases. To verify that the observed shift in the coupling angle is due to RSH/Au-NP adsorption rather than a refractive index change, the refractive indices of pure hexane and a hexane solution of C14S/ Au-NP (0.8 mg/mL) were determined from critical angle measurements by SPR.15 The nanoparticles do not measurably alter the refractive index of hexane. Gold-coated LaSFN9 slides were prepared by evaporation of 46.5-50.0 nm gold (99.99% Plasmaterials. Inc) at a rate of 1.2 Å/s, and an initial pressure of 2 × 10-6 Torr. These slides were cleaned prior to the gold coating by 10min sonication in 2% Helmanex solution (Hellma Optik GmbH Jena, Germany), followed by 5 min sonication in Millipore water, and then 5 min sonication in absolute ethanol. The slides were stored in absolute ethanol both before and after the gold deposition. Self-Assembled Monolayer (SAM) Deposition. Electrodeposition of thiols was carried using a computer-controlled potentiostat (BAS 100B Electrochemical Analyzer, Bioanalytical Systems, West Lafayette, IN) apparatus, according to a procedure that was developed previously in our lab.16 Briefly, the gold-coated LaSFN9 slides were first exposed to a 1 mM LiClO4 solution (ethanol) and then treated with 4 cyclic potential sweeps (-500 to +500 mV, scan rate 100 mV/s). 25% of the solution was then removed, a potential of 200 mV was applied for ca. 5-10 s, and a 4 mM thiol solution in 1 mM LiClO4/ethanol was added. After 20 min, the slides were removed from the solution and washed with ethanol and then Millipore. This was repeated three times. The slides were stored in Millipore water overnight. Preparation of Gold Nanoparticle Monolayers. A monolayer of C14S/Au-NP was formed on a water surface using 600 µL of a C14S/Au-NP solution (0.8 mg/mL in hexane, Spectrograde) spread via ca. 10 µL drops. 15 min elapsed before compression of the spread film (KSV Langmuir film balance, model 3000) was initiated. The monolayer was transferred horizontally (Π ) 11mN/m) onto a LaSFN9 glass slide that had been coated with gold and 1-octadecanethiol. (Figure 1) Data Analysis. Conversion of the SPR signal (∆θ) into surface coverage yields kinetic curves in terms of surface coverage of RSH/ Au-NP vs time. The adsorption rate was determined by numerically differentiating the coverage-time curves. Because numerical differentiation amplifies noise,17 the raw data (coverage-time) requires smoothing. A smoothing procedure was therefore applied to the first 100 data points (corresponding to the first 30 min of each experiment). The smoothing procedure involves (i) performing a “running mean” smoother of 5 points, (ii) fitting a seventh order polynomial to the smoothed raw data, and (iii) assessing the validity of the resulting (15) Grassi, J. H.; Georgiadis, R. M. Anal. Chem. 1999, 71, 4392-4396. (16) Ma, F.; Lennox, R. B. Langmuir 2000, 16, 6188-6190. (17) Jung, L. S.; Campbell, C. T. J. Phys. Chem. B 2000, 104, 11168-11178.

Langmuir, Vol. 22, No. 3, 2006 1049

Figure 1. Surface pressure-area isotherm (T ) 26.5 °C) of C14S/ Au-NP on a water surface. A monolayer of C14S/Au-NP was removed via the LB technique at a surface pressure of 11 mN/m2 (area ) ca. 2760 Å2/particle). This area corresponds to a NP diameter of 5.9 nm and a surface density of 3.6 × 1012 NP/cm2. data by plotting the resulting polynomial values vs the first 10 points of the raw data. All data presented lie within 95% confidence limits for a slope )1 and intercept ) 0 for this plot. In a few cases, the smoothing procedure had been applied to a lesser number of data points so that this confidence level will be achieved.

Results and Discussion Evaluation of Gold RSH/Au-NP Parameters and RSH/ Au-NP Surface Coverage via SPR. Determination of the equilibrium adsorption constant for a series of chemically distinct RSH/Au-NP on a series of SAMs requires knowledge of the RSH/Au-NP surface coverage throughout the course of the experiment. The experimental SPR signal thus has to be converted into values of surface coverage. A brief overview of the SPR experiment as used here helps describe how surface coverage data is accessed from SPR data. SPR measures changes in the refractive index near metals, using metal surface plasmons as a sensor.18-20 An incoming monochromatic p-polarized light is directed at the back of the metal through a prism at variable angle range, and its reflection intensity is detected. The intensity of the reflected light is equal to the intensity of the incoming light except when resonance conditions are fulfilled. Under resonance conditions, the reflection intensity decreases, and a minimum is observed in the SPR spectrum (Figure 2). The angle at which resonance conditions are achieved is the coupling angle, and its position depends on the refractive index at the metalsolution interface. In all measurements represented here, solvent was used as a blank, and in each experiment, the difference between the experimental coupling angle and its blank, (∆θ), was determined. ∆θ thus corresponds to the change in the system arising from adsorption of material onto the surface. To estimate the surface coverage of the RSH/Au-NP (NP/ cm2), the ∆θ values of both an RS/Au-SAM (blank) and an RS/Au-SAM covered with a full monolayer of C14S/Au-NP (Figure 2) were determined. The RS/Au-NP monolayer (obtained as an LB film, Figure 1) serves as the standard for the SPR shift for a known coverage of gold nanoparticles. To prevent dissolution (18) Rather, H. Phys. Thin Films 1977, 9, 145-261. (19) Jung, L. S.; Campbell, C. T.; Chinowsky, T. M.; Mar, M. N.; Yee, S. S. Langmuir 1998, 14, 5636-5648. (20) Frutos, A. G.; Corn, R. M. Anal. Chem. 1998, 449, 9A-455A.

1050 Langmuir, Vol. 22, No. 3, 2006

Figure 2. Effect of RS/Au-NP monolayer deposition on SPR curves: The measured reflectivity in air for C18S/Au-SAM (]) and a monolayer of C14S/Au-NP deposited onto the C18S/Au-SAM (i.e., C14S/Au-NP + C18S/Au-SAM) (2) is plotted as f(θ). ∆θ is the shift in the minimum angle and corresponds to the formation of the C14S/Au-NP overlayer. To obtain greater resolution in the fitting process, a large number of data points were collected in the regions of the total internal reflection angle (dashed line) and the minimum angle (θmin).

of RSH/Au-NP from the LB film, the SPR measurements were performed in air and not in solvent (i.e., hexane). The experimental ∆θ values thus obtained were then converted to hexane conditions.21 This conversion involves fitting the entire SPR curve in air (using Fresnel analysis) yielding a ∆θ value. Setting the dielectric constant of the last layer in the model to that of hexane ( ) 1.884) yields a “corrected ∆θ”. Fitting the entire SPR curve requires the determination of RS/Au-NP parameters. This was accomplished by using an iterative fit procedure.22 The values obtained for the RS/Au-NP from this fit are nR ) 1.88 and nI ) 0.23, where nR is the real part of the refractive index and nI is the imaginary part of the refractive index. Because literature values for the refractive index of gold nanoparticles span a fairly large range,23-25 it is difficult to compare these to the values determined in this study because nanoparticle size, interparticle interactions, and particle coalescence can all affect the dielectric constant. For example, at a wavelength of ca. 630 nm, the reported nI values of bulk gold range from 3.15 to 3.5.26,27 However, nI values reported for isolated nanoparticles or nanoparticles in a transparent matrix range from 0.012524 to 1.55.25 A study by Palpant et al.23 illustrates this problem. When the gold parameters of alumina-embedded (6.5% volume fraction) gold clusters (2-4 nm) are measured as a function of particle size, a blue shift is observed in the nanoparticle adsorption spectrum as the particle size decreases. This leads to decreased nI values. Since nI is proportional to the absorption coefficient, it is expected that a reduction in a particle size will cause a decrease in the nI value at this wavelength. The nR values for bulk gold at λ ) 630 nm are