A Novel Isotherm, Modeling Self-Assembled Monolayer Adsorption

Dec 17, 2008 - The formation of mercaptopropionic acid (MPA) SAMs on gold has been ... on the rate of formation and the degree of monolayer disorder...
8 downloads 3 Views 674KB Size
Download PDF
Langmuir 2009, 25, 931-938

931

A Novel Isotherm, Modeling Self-Assembled Monolayer Adsorption and Structural Changes Andrew P. Henderson,*,† Lalitesh N. Seetohul,† Andrew K. Dean,‡ Paul Russell,† Stela Pruneanu,§,| and Zulfiqur Ali† School of Science and Technology, and Spartan Nano, UniVersity of Teesside, Middlesbrough, United Kingdom, National Institute for R&D of Isotopic and Molecular Technologies, Cluj-Napoca 3400, Romania, and School of Natural Sciences, Newcastle UniVersity, Newcastle upon Tyne, United Kingdom ReceiVed August 15, 2008. ReVised Manuscript ReceiVed NoVember 10, 2008 Self-assembled monolayers (SAMs) have numerous applications, for example, engine wear inhibitors, surface profiling signal enhancement, nanostructure production, sensor production, and catalysis. The adsorbed SAM structure has a major impact on the properties of the outer monolayer surface which dictates the performance and viability of the SAM for individual applications. Substrate growth phases of SAMs have been extensively studied, and two structures have been identified. Initially, a lying down SAM structure is formed that evolves into a standing up structure. It is often critical to know how both structures form as a function of substrate immersion time to be able to design the properties of this structure. The formation of mercaptopropionic acid (MPA) SAMs on gold has been studied. Electrochemical impedance spectroscopy (EIS) was used to measure the adsorption isotherms at five temperatures in the range 4-40 °C. Infrared reflectance absorption spectroscopy (IRRAS) was also used, and the results show close agreement. A new monolayer adsorption isotherm is proposed which models SAM structure formation as a function of immersion time, representing all phases of SAM adsorption. This model represents a significant improvement on previous models based on Langmuir and Kisliuk adsorption isotherms that only model the fractional coverage of a surface with a SAM. The new model predicts the optimum immersion time taken for an MPA monolayer on gold to attain a surface saturated with MPA. It accounts for temperature effects on the rate of formation and the degree of monolayer disorder. It has potential for use in other SAM systems and may become the method of choice for modeling many instances of sequential substrate adsorption of two different structures, each of which exhibits different properties, as a function of immersion time.

Introduction Self-assembled monolayers (SAMs) have a host of applications. For example; they have been used in engine wear inhibitors,1 in surface profiling signal enhancement,2 in nanostructure production,3 in catalysis,4 and in sensor production.5 In particular, the immunosensor community commonly uses SAMs in the attachment of antibody Fc fragments to an inert metal surface.6 Of the several million SAMs known to exist,7 SAMs employing short hydrocarbon chains are most commonly used with Au(111) substrates and are grown in an ethanol or hexane solution. An explanation for this is that Au films are relatively easy to prepare, the substrate is relatively easy to clean, and the short hydrocarbon chain SAMs are simple molecules that exhibit all the properties and degrees of freedom necessary for complete monolayer formation. In this instance, the sulfur terminal group of the SAM automatically forms strong metal-ligand interactions with the Au(111) crystals in the substrate. * To whom correspondence should be addressed. Telephone: +44 (0)1642 342-428.Fax:+44(0)1642342-401.E-mail:[email protected]. † School of Science and Technology, University of Teesside. ‡ Spartan Nano, University of Teesside. § National Institute for R&D of Isotopic and Molecular Technologies. | Newcastle University. (1) Zhou, Y.; Jiang, S.; C¸agin, T.; Yamaguchi, E. S.; Frazier, R.; Ho, A.; Tang, Y.; Goddard, W. A. J. Phys. Chem. A 2000, 104, 2508. (2) Flores, S. M.; Shaporenko, A.; Vavilala, C.; Butt, H. J.; Schmittel, M.; Zharnikov, M.; Berger, R. Surf. Sci. 2006, 600, 151–2847. (3) Choi, S.; Lee, N.; Kim, Y-J; Kang, S. Nanotechnology 2004, 3, 350. (4) Cortial, G.; Goettmann, F.; Mercier, F.; Le Floch, P.; Sanchez, C. Catal. Commun. 2007, 8, 215. (5) Chen, D.; Li, J. Surf. Sci. Rep. 2006, 61, 445. (6) Luppa, P. B.; Sokoll, L. J.; Chan, D. W. Clin. Chim. Acta 2001, 314, 1. (7) Schreiber, F. Prog. Surf. Sci. 2000, 65, 151.

The structure and growth of thioalkane SAM molecules on a Au(111) substrate is dictated by a number of factors including substrate structural defects, temperature during monolayer formation, immobilization technique, solvent used, electrode pretreatment, chain length, monolayer density, and headgroup structure.5,7,8 Although there are many types of thioalkane headgroups, this report focuses specifically on developing a model to describe the self-assembly of a mercaptoalkane carboxylic acid monolayer on a Au(111) substrate (which provides a model case for the purposes of understanding thioalkane behavior). Three Stages of Monolayer Formation. Although Langmuir adsorption isotherm behavior has been observed over the first few minutes of SAM solution immersion, symmetrical deviations from the Langmuir isotherm occurs in measured experimental values that have been reported in the literature.11,12,14,21 An in(8) Gyepi-Garbrah, S. H.; Sˇilerova´, R. Phys. Chem. Chem. Phys. 2001, 3, 2117. (9) Ishida, T.; Tsuneda, S.; Hara, N. N. M.; Sasabe, H.; Knoll, W. Langmuir 1997, 13, 4638. (10) Poirier, G. E.; Pylant, E. D. Science 1996, 272, 1145. (11) Dannenberger, O.; Buck, M.; Grunze, M. J. Phys. Chem. B 1999, 103, 2202. (12) Peterlinz, K. A.; Georgiadis, R. J. Phys. Chem. B 1997, 101, 8041. (13) Poirier, G. E.; Tarlov, M. J.; Rushmeier, H. E. Langmuir 1994, 10, 3383. (14) Bain, C. D.; Troughton, E. B.; Tao, Y.-T.; Evall, J.; Whitesides, G. M.; Nuzzo, R. G. J. Am. Chem. Soc. 1989, 111, 321. (15) Camillone, N.; Eisenberger, P.; Leung, T. Y. B.; Schwartz, P.; Scoles, G.; Poirier, G. E.; Tarlov, M. J. J. Chem. Phys. 1994, 101, 11031. (16) Anderson, M. R.; Evaniak, M. N.; Zhang, M. Langmuir 1996, 12, 2327. (17) Castner, D. G.; Hinds, K.; Grainger, D. W. Langmuir 1996, 12, 5083. (18) Pan, W.; Durning, C. J.; Turro, N. J. Langmuir 1996, 12, 4469. (19) Bensebaa, F.; Voicu, R.; Huron, L.; Ellis, T. H.; Kruus, E. Langmuir 1997, 13, 5335. (20) Karpovich, D. S.; Blanchard, G. J. Langmuir 1994, 10, 3315. (21) Xu, S.; Cruchon-Dupeyrat, S.; Garno, J. C.; Liu, G.-Y.; Jennings, G. K.; Yong, T.-H.; Laibinis, P. E. J. Chem. Phys. 1998, 108, 5002.

10.1021/la802677n CCC: $40.75  2009 American Chemical Society Published on Web 12/17/2008

932 Langmuir, Vol. 25, No. 2, 2009

Henderson et al.

Figure 1. Diagrammatic representation of a low coverage alkanethiol molecular arrangement on a Au(111) surface.

depth investigation was conducted into these occurrences by Himmelhaus et al.,24 using sum frequency generation (SFG), who reported three distinct stages of monolayer formation. Three Stages of Monolayer Formation: (i) The First Stage of Formation. The first stage consists of the chemisorption of the thiol headgroup to form a “lying phase” on the surface of the substrate. Camillone et al.15 established that this stage happens very rapidly, in the order of seconds or minutes, and that alkanethiol hydrocarbon chains in this state were laying flat on the Au surface and some sulfur groups were spaced at a distance of 31.7 Å, which is roughly twice the length of the decanethiol SAM molecule immobilized (L) (see Figure 1). This study was later confirmed in another paper26 which reported that the spacing between sulfur groups increased as a function of alkanethiol chain length and that adjacent sulfur groups were spaced at 2.2 Å which is also the length of a disulfide bond. This hypothesis was supported by two studies of alkanethiols on Au surfaces via X-ray standing wave (XSW) data.27,28 Three Stages of Monolayer Formation: (ii) The Second Stage of Formation. Second stage formation corresponds to reorientation of hydrocarbon chains to a 30° angle with respect to the surface normal and occurs over a much longer immersion period as a result of substrate coverage and monolayer density. This indicates that the effect of hydrocarbon chain reorientation is brought about by interactions between adjacent SAM molecules on the substrate.26,29-31 Moreover, the tilt angle of alkanethiols varies between 90° and approximately 30° as a function of SAM density on the Au surface.32-34 As a further result of the reorientation of hydrocarbon chains, the structure of the SAM monolayer on the Au surface changes from that of the low coverage phase to (22) Yamada, R.; Uosaki, K. Langmuir 1997, 13, 5218. (23) Yamada, R.; Uosaki, K. Langmuir 1998, 14, 855. (24) Himmelhaus, M.; Eisert, F.; Buck, M.; Grunze, M. J. Phys. Chem. B 2000, 104, 576. (25) Ji, Z.; Li, J.; Zhu, B.; Yuan, G.; Cai, S. Electrochem. Commun. 2007, 9, 1667. (26) Camillone, N.; Leung, T. Y. B.; Schwartz, P.; Eisenberger, P.; Scoles, G. Langmuir 1996, 12, 2737. (27) Fenter, P.; Schreiber, F.; Berman, L.; Scoles, G.; Eisenberger, P.; Bedzyk, M. Surf. Sci. 1998, 412/413, 213. (28) Fenter, P.; Schreiber, F.; Berman, L.; Scoles, G.; Eisenberger, P.; Bedzyk, M. Surf. Sci. 1999, 425, 138. (29) Poirier, G. E. Langmuir 1999, 15, 1167. (30) Dubois, L. H.; Zegarski, B. R.; Nuzzo, R. G. J. Chem. Phys. 1993, 98, 678. (31) Gerlach, R.; Polanski, G.; Rubahn, H.-G. Appl. Phys. A: Mater. Sci. Process. 1997, 65, 375. (32) Chidsey, C. E. D.; Liu, G.-Y.; Rowntree, P.; Scoles, G. J. Chem. Phys. 1989, 91, 4421. (33) Strong, L.; Whitesides, G. M. Langmuir 1988, 4, 546. (34) Chidsey, C. E. D.; Loiacono, D. N. Langmuir 1990, 6, 682.

a transitional structural phase and, finally, to the saturated phase structure. Three Stages of Monolayer Formation: Saturated Phase. The third stage was shown to occur over the longest period of immersion time and involved the ordering of the SAM terminal functional groups.24 The reported structure of the saturated phase of an alkanethiol SAM on a Au(111) surface can be seen in Figure 2, and it has been described as having the hexagonal notation: (3 × 3) R30°.32-34 Each sulfur group that has been chemisorbed onto the Au substrate is spaced at 5 Å from adjacent sulfur groups, and molecular interactions between adjacent hydrocarbon chains give rise to a 30° hydrocarbon chain tilt angle (approximately) from the normal35,36 (see Figure 2). However, instances have also been reported where saturated alkanethiol layers have differed in structure from the conventional (3 × 3) R30° structure.37-39 It was suggested by Schrieber7 that a possible reason for this occurrence was that sulfur group interactions played a more dominant role in SAM molecule positioning in very short hydrocarbon chain alkanethiols than in larger hydrocarbon chain alkanethiols. The behavior of mercaptopropionic acid (MPA) is an example of such behavior. Carboxyl Terminated Thioalkanes. In a more recent study, the structure of 3-mercaptopropionic acid (HS-(CH2)2-COOH) on a Au(111) surface was studied.44 This short chain thioalkane exhibited an ordered monolayer with disordering only occurring near etched areas of the substrate. As expected, the structure of the short chain mercaptoalkane carboxylic acid was found to exhibit a different structure from that of the conventional (3 × 3) R30° structure, as a result of the strong intermolecular hydrogen bonds brought about by the SAM molecule COOH terminal group35 (see Figure 3). In this instance, triangular structures, each consisting of three adjacent SAM molecules, are (35) Nuzzo, R. G.; Dubois, L. H.; Allara, D. L. J. Am. Chem. Soc. 1990, 112, 558. (36) Nuzzo, R. G.; Korenic, E. M.; Dubois, L. H. J. Chem. Phys. 1990, 93, 767. (37) Pockels, A. Nature 1891, 43, 437. (38) Fenter, P.; Eberhardt, A.; Eisenberger, P. Science 1994, 266, 1216. (39) Als-Nielsen, J.; Mo¨hwald, H. In Handbook of Synchrotron Radiation; Ebashi, S., Koch, M., Rubenstein, E., Eds.; Elsevier: New York, 1991; Vol. 4. (40) Porter, M. D.; Bright, T. B.; Allara, D. L.; Chidsey, C. E. D. J. Am. Chem. Soc. 1987, 109, 3559. (41) Alves, C. A.; Smith, E. L.; Porter, M. D. J. Am. Chem. Soc. 1992, 114, 1222. (42) Kang, J.; Rowntree, P. A. Langmuir 1996, 12, 2813. (43) Barrena, E.; Ocal, C.; Salmeron, M. J. Chem. Phys. 1999, 111, 9797. (44) Sawaguchi, T.; Sato, Y.; Mizutani, F. Phys. Chem. Chem. Phys. 2001, 3, 3399. (45) Kreuzer, H. J. Surf. Sci. Lett. 1995, 344, L1264. (46) Lewis, P. A.; Donhauser, Z. J.; Mantooth, B. A.; Smith, R. K.; Bumm, L. A.; Kelly, K. F.; Weiss, P. S. Nanotechnology 2001, 12, 231.

Modeling SAM Adsorption and Structural Changes

Langmuir, Vol. 25, No. 2, 2009 933

Figure 2. Diagrammatic representation of a saturated monolayer structure on a Au(111) substrate, with chemisorbed sulfur groups (S) and hydrocarbon chains (solid lines) at a tilt angle of θ° from the normal (dotted line).

of the increasing number of SAM molecules present on the Au surface during the period of monolayer formation. This was accomplished through the use of “sticking coefficients” (kE) which vary depending on hydrocarbon chain length; functional groups that may be present on the SAM’s constituent molecules; SAM molecule solution concentration, and experimental conditions that can influence monolayer growth (examples of these include the type of solvent used, the immobilization temperature, and the degree of solution contamination).11,45

dΘ ) R ′ (1 - Θ)(1 + kEΘ) dt Figure 3. Diagrammatic representation of the structure of a mercaptopropionic acid monolayer. MPA molecules are shown as circles, with an MPA structure highlighted in gray; substrate molecules are shown in white, and the black rhombus shows the interstructural spacing.

arranged in a rhombus shape, with the center of each triangle at a distance of 1 nm between adjacent structures. Each individual SAM molecule on the surface of the Au(111) monolayer is spaced at a distance of 4.53 Å within a triangle. Implications of Molecular Interactions on the Langmuir Growth Curve Model. The Langmuir adsorption isotherm (eq 1), derived in eq 2, is conventionally used to model most instances of molecular adsorption, and it has been found to provide an adequate approximation to SAM coverage observed over time.11 In this instance, the growth rate is proportional to the number of available binding sites on the Au surface, where Θ is the fraction of the Au surface covered, R is the rate constant, and t is immobilization time. This model, however, fails to take into account molecular interactions, which have a significant effect on the rate of formation and the geometry of the monolayer. This encompasses electrostatic interactions of adjacent molecules, in addition to the effects of immobilized SAM molecules and immobilized SAM molecule orientations on aqueous SAM molecule diffusion from the bulk solution to the Au(111) surface.

dΘ ) R(1 - Θ) dt

(1)

Θ ) 1 - e-Rt

(2)

In an attempt to overcome this problem, a modified Kisliuk model (eq 3) was introduced to take into account the interactions (47) Scoles, G., Ed. Atomic and Molecular Beam Methods; Oxford University Press: Oxford, 1992; Vols. 1 and 2. (48) Fenter, P.; Eberhardt, A.; Liang, K. S.; Eisenberger, P. J. Chem. Phys. 1997, 106, 1600.

Θ)

1 - e-R(1+kE)t 1 + kE e-R(1+kE)t

(3) (4)

In the instance of eqs 3 and 4, R′ is a coefficient representing the impact of diffusion on monolayer formation and is proportional to the square root of the system’s diffusion coefficient. Although the Kisliuk and Langmuir adsorption isotherms both give an effective approximation to substrate coverage with SAM as a function of immersion time, neither isotherm models second and third stage growth. This is critical, as, in many instances, functional group orientation affects the monolayer properties as well as the rate and reproducibility of subsequent reactions involving the immobilized SAM layer. Thus, there is a need to develop a model which is capable of predicting SAM growth and degree of ordering as a function of immersion time.

Experimental Section MPA was selected to form a SAM on Au electrodes to provide sample data upon which a model could be based. This short chain mercaptoalkane carboxylic acid was chosen because it results in fewer intermolecular interactions than its intermediate and long chain counterparts. This acted to reduce gauche defects and ultimately resulted in a more ordered monolayer.7,40-42 Electrochemical impedance spectroscopy (EIS) was selected as the principal method for characterization of SAMs. It was selected because it is particularly suited to the analysis of the changes in the (49) Bard, A. J.; Faulkner, L. R. Electrochemical Methods Fundamentals and Applications, 2nd ed.; John Wiley and Sons Inc.: New York, 2001. (50) Limbut, W.; Kanatharana, P.; Mattiasson, B.; Asawatreratanakul, P.; Thavarungkul, P. Biosens. Bioelectron. 2006, 22, 233. (51) Shervedani, R. K.; Mehrjardi, A. H.; Zamiri, N. Bioelectrochemistry 2006, 69, 201. (52) Feng, K.; Li, J.; Jiang, J. H.; Shen, G. L.; Yu, R. Q. Biosens. Bioelectron. 2007, 22, 1651.

934 Langmuir, Vol. 25, No. 2, 2009 interfacial properties of modified electrodes upon monolayer formation occurring at the modified electrode surface. This is due to the ability of EIS to provide detailed information on capacitance and resistance changes occurring at the electrode surface, as well as the kinetics and mechanisms of electron transfer processes (corresponding to the cleanliness of the electrode surface in addition to monolayer formation and ordering occurring at the modified electrode; see the Supporting Information). Infrared reflection adsorption spectroscopy (IRRAS) was used for the characterization of the binding of SAMs to a Au surface as a means of confirming EIS results. IRRAS was chosen because it shows some advantages over conventional IR spectroscopy. For instance, it can be used for samples that are not transparent over an appreciable wavelength range in the IR region. In IRRAS, the IR beam is reflected from the front face of a highly reflective sample. Using this method, typical spectra produce outputs such as transmittance versus wavelength and absorbance versus wavelength or wavenumbers. The data in this research are represented as absorbance versus wavenumber. Experimental Methods. Chemicals and Solutions. MPA (Sigma) at a concentration of 0.05 M was used in ethanol (Fisher Scientific). Phosphate buffer solution (PBS; 0.1 M, pH ) 7.4) was prepared, which contained 0.1 M KCl (BDH) as supporting electrolyte and 0.05 M K3Fe(CN)6 (Fisher Scientific) as redox mediator. Deionized water (DI) (15 MΩ) was used, unless otherwise stated. Cleaning. Au(111) coated borosilicate glass slides for IRRAS studies were immersed in piranha solution [75% v/v H2SO4 (Fluka) and 25% v/v H2O2 (Fisher Scientific)] before being rinsed first in DI water and then subsequently in ethanol to remove inorganic and organic contaminants from the slides prior to MPA immobilization. All glassware was sequentially rinsed with aqua regia [25% v/v HNO3 (Fisher Scientific) and 75% v/v HCl (Fisher Scientific)], concentrated NaOH (Fluka) solution, and DI water to remove sources of organic and ionic contamination. Working electrodes were polished using 0.05 µm alumina powder, followed by ultrasonic cleaning for 10 min in DI water. The working electrodes were then ultrasonicated in piranha solution before being rinsed in DI water and subsequently in ethanol to remove inorganic and organic contaminants from the electrodes prior to MPA immobilization. To ensure effective electrode cleaning, five cyclic voltammogram sweeps were performed between potentials of -0.5 and 0.5 V. Overlapping scans inferred that nonatomically bound matter was no longer transferring into the bulk PBS solution from the electrode, and thus, the working electrode was assumed to be clean. Electrochemical Measurements. Each immersion period test was performed in triplicate using a conventional three electrode system with an Autolab potentiostat/galvanostat (Eco Chemie B.V.) in conjunction with Frequency Response Analyzer (FRA) software and General Purpose Electrochemical System (GPES) software (Eco Chemie B.V.) for EIS and cyclic voltammetry (CV) measurements, respectively. The Ag/AgCl reference electrode, platinum wire counter electrode, and Au working electrode (MF-2014) were all purchased from BASi instruments Inc., and all CV and EIS measurements were performed in PBS. For EIS, a 0.1 mV amplitude sine wave was applied to the working electrode in the tested frequency range of 0.1 Hz to 1 MHz. An AC frequency of 0.1 Hz was used to obtain the data. This frequency was selected because the mass transfer effects of redox ions are more apparent at this frequency, causing a greater number of monolayer properties to influence the impedance data. The working electrode was polarized at a potential of +0.25 V/Ag(AgCl) during measurements, which was shown via CV to be the oxidation potential of the redox indicator. EIS measurements were performed after the working electrode had been cleaned and again after it had been immersed in MPA solution. Spectroscopic Measurements. For surface analysis (mainly used for evidence of 3-mercaptopropionic acid (MPA) terminated SAM formation), infrared spectroscopy was used. IRRAS spectra were (53) Zheng, H.; Hirose, Y.; Kimura, T.; Suye, S.; Hori, T.; Katayama, H.; Arai, J.; Kawakami, R.; Ohshima, T. Sci. Technol. AdV. Mater. 2006, 7, 243.

Henderson et al.

Figure 4. Normalized impedance plot comparing the fit of the proposed adsorption isotherm model (H 40 °C, H 30 °C, H 25 °C, H 20 °C, and H 4 °C) to the normalized experimental impedance data obtained at 0.1 Hz (D 40 °C, D 30 °C, D 25 °C, D 20 °C, and D 4 °C).

acquired using a Varian 7000 Fourier transform infrared (FTIR) spectrometer equipped with a step scan interferometer and a liquid nitrogen cooled narrow band detector. The spectra were collected in a vacuum at a resolution of 2 cm-1 from 600 to 6000 cm-1. A reference Au slide was used as background. SAM Growth Procedure. Working electrodes and Au coated glass slides were rinsed in ethanol after the initial measurement, to avoid MPA solution contamination with PBS, before Au coated glass slides were immersed in MPA solution at 20 °C. The MPA solutions under investigation were held at a range of specific temperatures in either a water bath or refrigerator. Several working electrodes were initially immersed in each MPA solution, and each electrode was sequentially removed for EIS measurement after a certain immersion period. Upon removal from the MPA solution, electrodes were immediately ultrasonicated for 2 min 30 s in ethanol to remove nonchemisorbed MPA molecules prior to EIS measurement. Similarly, prior to IRRAS measurement, Au coated glass slides were soaked in ethanol over a 12 h period before being rinsed with water and then dried to remove nonchemisorbed MPA molecules.

Results and Discussion Monolayer Formation. Figure 4 shows the data obtained from the EIS measurements for five different temperatures. All five temperatures show two clear ramps followed by plateaus in the record that correspond to two distinct phases of monolayer growth. For the data sets collected for temperatures between 20 and 40 °C, the first ramp and plateau correspond to the formation of a monolayer structure (lying down) over an immersion time of between 5 s (0.00139 h) and 5 min (0.0833 h) and the formation of a second structure (standing up) over 5-12 h. This is illustrated in Figure 4 by the two plateaus occurring on each temperature trace. As expected, the rate of formation of the first structure increases from 20 to 40 °C. However, the hydrophobicity and structure of the first fully formed monolayers differ greatly between formation temperatures, which is reflected in the impedance magnitudes of the first structure plateaus.57 The data sets are divided by the Gyepi-Garbrah critical temperature8 (Tc) into two sets. This temperature was defined as being the point at which a SAM structure undergoes a twodimensional phase transition, resulting in a loss of order in the monolayer. For a MPA monolayer, the Tc is shown to lie in the range of 20-25 °C in Figure 4. Two data sets were measured (54) Uzawa, H.; Ito, H.; Izumi, M.; Tokuhisa, H.; Taguchi, K.; Minoura, N. Tetrahedron 2005, 61, 5895. (55) Pruneanu, S.; Boughriet, A.; Henderson, A.; Malins, C.; Ali, Z.; Olenic, L. Part. Sci. Technol. 2008, 26, 136. (56) Fan, H.; Lu, Y.; Stump, A.; Reed, S. T.; Baer, T.; Schunk, R.; Perez-Luna, V.; Lo´pez, G. P.; Brinker, C. J. Nature 2000, 405, 56. (57) Huang, F.; Qu, S.; Zhang, S.; Liu, B.; Kong, J. Talanta 2007, 72, 457.

Modeling SAM Adsorption and Structural Changes

Langmuir, Vol. 25, No. 2, 2009 935

below Tc at 4 and 20 °C. Three experiments at 25, 30, and 40 °C were carried out above this temperature, and the formation of the monolayer structures was observed to be different. Immersion Data Collected at 4 °C. During adsorption of the first structure at 4 °C, the rate of adsorption of MPA was higher than the rate of adsorption of MPA at 20 °C. This was due to the decreased solubility of the MPA in ethanol. In accordance with the principles of the Kisliuk adsorption isotherm, as applied to SAMs,11 this considerably weakened the force that ethanol molecules exerted on the MPA molecules in the precursor state, which act to drive the molecules back into solution. Thus, MPA molecules are more readily adsorbed to the surface of the substrate than at a temperature of 20 °C. It can also be seen from Figure 4 that the first plateau at 4 °C occurs at a much lower impedance than the 20 °C data set. This indicates that the monolayer structure is different between the two cases (see the Supporting Information). Moreover, it can be deduced that a porous disordered monolayer is initially formed and that the gradual ramp preceding the formation of the second plateau corresponds to the reduction in monolayer pinhole defects until the monolayer becomes saturated and the second plateau is attained. In this respect, the 4 °C data set is unique as no phase transitions occur in the monolayer over the immersion times investigated. Immersion Data Collected at 20 °C. Over the first 5 min, at 20 °C, MPA molecules form an ordered fully formed lying down structure (shown in Figure 1). The structure forms as MPA molecules adsorbed onto the surface of the electrode as insulating “islands” that grow with immersion time.7,43 This can be represented by the Gibbs free energy equation, shown in eq 5, where ∆G ads1 is the Gibbs free energy of MPA adsorption (dictated mainly by the S-Au interaction); ∆Hads1 is the enthalpy evolved from adsorption; T is absolute temperature; ∆S ads1 is the entropy introduced as a result of MPA molecule adsorption.

∆Gads1 ) ∆Hads1 - T∆Sads1

(5)

The electrically insulating islands grow as a function of immersion time until the substrate surface is covered and the monolayer enters a transitional state structure, whereby hydrocarbon chains of MPA immobilized on the substrate are displaced by thiol groups of other MPA molecules being adsorbed onto the substrate.7 This is shown in Figure 4 by the shift in impedance levels between plateaus 1 and 2 and can be represented in the energy balance shown in eq 6. ∆Gads2 is the Gibbs free energy of MPA adsorption over the transitional structural formation stage, and ∆GR-Au is the Gibbs free energy evolved from the MPA chain-Au van der Waals interactions.

∆Gads2 ) ∆Gads1 - ∆GR-Au

(6)

Upon comparing eqs 5 and 6, it can be seen that ∆Gads2 < ∆Gads1; thus, the rate of adsorption in the transitional structure phase is slower than in the case of initial adsorption (see Figure 4). Structure 1 (the lying down structure) has high impedance due to the hydrophobic layer formed by the exposure of the carbon chains to the MPA-electrochemical solution boundary. Structure 2 (standing up structure) corresponds to an increase in the number of oxygen and hydrogen atoms being orientated toward the solution-monolayer boundary, which in turn has low impedance and is an indication of increased monolayer disorder. Also, during MPA molecule adsorption in the transition phase, the densely packed (standing up) MPA structure (see Figure 3) gradually forms.44 In this instance, sulfur molecules are spaced at a constant distance of 4.53 Å from each other as a result of a force equilibrium of strong attractive and repulsive

Figure 5. Plot showing how final monolayer normalized impedance measured at 0.1 Hz varies with immersion temperature as a result of monolayer structural differences.

intermolecular forces established between adjacent MPA molecules. This standing up structure does not allow any further MPA molecules to be adsorbed onto the substrate from the solution to form an even more densely packed structure; therefore, the second plateau, of low impedance, is formed. Thus, it can be deduced that the rate of formation of the standing up structure in the transitional phase is proportional to the amount of MPA in the lying down structure present on the substrate. Immersion Data Collected above the Gyepi-Garbrah Critical Temperature: Temperature Effects on the Formation of the First Adsorbed Structure at 20 °C and above Tc. The initial impedance gradients leading to plateau 1 (see Figure 4) increase with temperature. This suggests that the rate of formation of the first adsorbed structure increases with temperature above Tc. For the temperature increase from 20 to 25 °C, the impedance level of plateau 1 is observed to drop significantly. Gyepi-Garbrah and Sˇilerova´8 suggested that this may be result of an increase in disorder of the first monolayer with a structure exhibiting decreased impedance. Our hypothesis, commensurate with GyepiGarbrah suggestion, is that structure 2 (standing up structure) forms at the same time as the normal growth of structure 1 across the surface. The first plateau in the impedance trace (see Figure 4) is the result of a pseudo-steady-state situation where a balance is struck between the formation rates of the two structures: high impedance structure 1 and low impedance structure 2. The relative impedances cancel each other out to give constant impedance. When the surface is covered fully, no more structure 1 can form, the balance is broken, and the impedance starts to rise as structure 1 is replaced with structure 2 Disorder continues to increase with increased temperature from 25 to 40 °C because the relative rates of formation change between the two structures, favoring structure 2 formation at higher temperatures. If the impedance is only a function of the structure of the monolayer, one would expect the impedance to fall. However, as the rate of formation of structure 2 increases, then more imperfections/faults in structure 2 occur.8 These imperfections act to increase the hydrophobicity of the layer and cause the impedance to rise. Immersion Data Collected above the Gyepi-Garbrah Critical Temperature: Temperature Effects on the Formation of the Second Adsorbed Structure at 20 °C and above Tc. Figure 5 shows a plot of final impedance value versus temperature of the data measured above Tc. The impedance exhibits a steady rise, which is caused by the increase in imperfections in the structure 2 monolayer for the temperatures 20-30 °C. The trace then shows a drop in impedance for the 40 °C result. GyepiGarbrah and Sˇilerova´8 observed an increase in monolayer porosity as immersion temperature increased above Tc that would give a drop in impedance. It is clear that the imperfections due to increased porosity now dominate the impedance of the monolayer and explain the drop in impedance observed.

936 Langmuir, Vol. 25, No. 2, 2009

Figure 6. ATR-FTIR spectra of MPA deposited on a Au substrate.

Figure 7. IRRAS spectra of MPA monolayers immobilized at 5 min, 15 min, and 16 h.

The Changes in Monolayer Absorbance as a Function of Immersion Time. To verify the EIS results obtained in the model case, at 20 °C, an IRRAS study was conducted, which was used to quantify the amount of MPA present on the surface of a Au coated glass slide with immersion time. To enable the identification of MPA, the assignment of the observed vibrational mode (obtained via IRRAS) was determined by spectral comparison with Fourier transform infrared (FTIR) spectra. A special sample containing a large amount of MPA was prepared by doping the surface of an Au-plated glass slide with a 0.05 M solution of MPA in ethanol. The ethanol was allowed to evaporate, leaving a large amount of MPA deposited on the surface. The amount of MPA on this sample was quantified using an attenuated total reflectance (ATR) element to measure the FTIR spectrum. Figure 6 shows the characteristic modes including the CdO stretch (1705-1720 (acids) CdO (H-bonded)) that was used to characterize the MPA monolayer on Au by IRRAS. IRRAS was then used to analyze the deposition of MPA on to the Au surfaces of the immersed samples used to monitor the self-assembly process for the monolayer at 20 °C. The IRRAS spectrum exhibited a weak but apparent mode (1717-1750 (acids) CdO (H-bonded)) that revealed the composition of the MPA monolayer on the Au surface as shown in Figure 7. The spectra response is much weaker than that of the doped sample because much less MPA is present on the surface for the immersion sample. The results indicate that a complete monolayer had formed after an immersion time of 15 min. However, due to low signal strength, it was difficult to determine the formation of the standing up structure between an immersion time of 15 min and 16 h. The IRRAS data (shown in Figure 7) and the EIS measurements

Henderson et al.

Figure 8. Area A: bare Au working electrode substrate surface available for binding to MPA to form the first Au-MPA structure. Area B: area occupied by the first MPA-Au structure which is available for binding to more MPA in solution, causing the formation of the second structure. Area C: coverage of the second structure, on which no further adsorption or reordering takes place.

(shown in Figure 4 at 20 °C) differ in that the immersion time required to achieve saturation of the electrode surface is much faster occurring after an immersion time of 5 min with EIS. This can be attributed to a greater degree of contamination of the IRRAS substrate surface than with the EIS working electrode after cleaning. A gentler cleaning regime was employed to clean the IRRAS substrates to avoid damaging them. As a result, IRRAS substrate MPA immobilization kinetics were hampered as collisions with the Au surface had to have enough energy to not only bind with the substrate surface but also displace the remaining surface contaminants. The IRRAS CdO peak exhibits a shift in wavenumber in comparison to the CdO peak yielded from the FTIR spectra, thereby demonstrating that the MPA forms ordered structures when incubated in ethanol solution at 20 °C and that MPA molecules are merely deposited in random orientations when MPA-ethanol solutions are simply doped onto the Au surface and left to evaporate. The spectrum shows a weak but apparent mode (1717-1750 (acids) CdO (H-bonded)) [see region of interest] that revealed the composition of MPA monolayer. The New Model. The EIS results obtained for the impedance are a function of both surface coverage and observed structure. Two MPA monolayer structures form on the Au working electrode substrates at different immersion times and at all temperatures. Moreover, it was found that disordered lying down and standing up structures do form, but the degree of disorder means that these terms, as defined by Himmelhaus et al.24 and Camillone et al.,26 cannot be used to describe the monolayer structures formed at temperatures above TC. Conventional adsorption isotherms that rely on fractional surface coverage with a single species, for example, the Kisliuk and Langmuir isotherms, are therefore rendered ineffective. Thus, a new model which can account for both quantities adsorbed and the structure of the monolayer is required. Development of a New Model. Although Kisliuk and Langmuir models cannot be used to model the impedance and hence adsorption data as a whole for this work, they can be used to model initial growth of the first structure on the substrate. Here the adsorption rate is dictated by the availability of binding sites. The number of possible binding sites for MPA occupying the first structure is solely a function of the number of MPA molecules bound to the Au surface. Modeling the formation of the second structure, however, is more complicated, as formation of the first structure on the substrate surface is a prerequisite to the formation of the second structure. The complexity increases for temperatures above TC where the second structure starts to form in small patches of structure one (see Figure 8) deposited on the surface before a

Modeling SAM Adsorption and Structural Changes

Langmuir, Vol. 25, No. 2, 2009 937

Table 1. Proposed Model Constant Values Used to Model the Experimental EIS Data at Different Immersion Temperatures and the Fit That the Model Function Provided to These Data chi-squared test (χ2) adsorption temperature (°C)

φ1

kE1

R1′

φ2

kE2

R2′

χ

40 30 25 20 4

0.30 0.25 0.20 0.79 0.20

0 0.60 0.60 0.64 0.45

2500 2000 2000 362.2 1500

0.60 1.00 0.48 0.13 0.29

0 0 0.5 0.7 0.3

0.5 0.7 0.7 0.8 0.3

0.04674 0.09681 0.01496 0.5944 0.01209

complete monolayer of the first structure is established. The number of possible binding sites for the second structure is dependent on two factors: Factor 1 is the availability of binding sites, which initially increase with immersion time, as the area of MPA occupying the first structure on the substrate surface also increases with immersion time as the monolayer grows. For factor 2, as structure 2 forms, structure 1 is consumed and eventually the rate of formation of structure 2 exceeds structure 1 and the number of available binding sites for structure 2 formation decreases, as indicated in Figure 8. As both the first and second structure consist of MPA molecules, total fractional coverage of the Au surface with MPA (ΘMPA) can be expressed in eq 7.

ΘMPA ) Θ1 + Θ2

(7)

where Θ1 and Θ2 are the fractional coverage of the first structure and second structure, respectively, on the Au surface at a given time and the fractional coverage of MPA orientated in the first structure decreases as a function of the second structure forming. This is illustrated in eqs 8 and 9, where ΘAu0 is the initial fraction of possible binding sites on the Au working electrode substrate surface and is equal to 1 at time t ) 0. Θ1(t) is a function describing the Langmuir or Kisliuk growth of the first MPA structure as a function of immersion time; Θ2(t) is a function describing the Langmuir or Kisliuk growth of the second MPA structure as a function of immersion time.

Θ1 ) ΘAu0Θ1(t) - Θ2

(8)

Θ2 ) Θ1(t)Θ2(t)

(9)

The immersion time functions in eqs 8 and 9 at 4 °C follow the mechanism used in the derivation of the Kisliuk model; however, molecular enthalpy is sufficiently large at temperatures higher than 20 °C for Langmuir behavior to result (as molecular sticking coefficients become negligible). It was decided that the immersion time functions Θ1(t) and Θ2(t) would be calculated using the Kisliuk adsorption model, as the Kisliuk model could be used to model Langmuir behavior by setting kE to 0 (see eq 4). In this instance, values of R′1 and kE1 are used in place of R′ and kE to obtain Θ1(t), and R′2 and kE2 are used to obtain Θ2(t). So far, only surface coverage has been accounted for. Next the dielectric properties must be taken into account. For impedance measurements, this can be compensated for by introducing a weighting constant φ which is the normalized signal value that would result from a substrate covered entirely with a particular type of MPA structure. In this case, the impedance φ1 was set at the measured impedance for plateau 1, obtained from the experimental data for each temperature data set. Similarly, φ2 is set as the measured impedance for plateau 2. It should be possible to define similar weighting constants based on measured variables for other detection techniques that measure monolayer structure and adsorption. Incorporating φ1 and φ2 into eqs 7-9 allows a model to be derived that can determine the total normalized impedance of an MPA monolayer, zt, after any immersion period, within the time

2

no. of data points 10 9 9 14 9

it takes to form the two different MPA monolayer structures. This model is shown in eq 10

zt ) φ1(ΘAu0Θ1(t) - Θ1(t)Θ2(t)) + φ2(Θ1(t)Θ2(t))

(10)

and is simplified in eq 11.

zt ) Θ1(t)[φ1(1 - Θ2(t)) + φ2Θ2(t)]

(11)

Results of Fitting the Model to the Measured Data. The model has been fitted to the experimental data to obtain the Kisliuk parameters for each temperature set. This was done using the least-squares method as defined in eq 12 and is expanded in eq 13.

∑ ([1 +1 -k e e

∑ √(zt(calc)-zt(exp t))2

-R1(1+kE1)t

E1

-R1(1+kE1)t

{(

φ1 1 -

)

1 - e-R2(1+kE2)t 1 - e-R2(1+kE2)t +φ2 -R2(1+kE21)t 1 + kE2 e 1 + kE2 e-R2(1+kE21)t

(12)

}] ) - zt(exp t)

2

(13)

The objective function was minimized using the Newton method contained within the Microsoft Solver routine and the fitted parameters set as R1′, R2′, kE1, and kE2. The fitted parameters and the goodness of fit to the experimental data are shown in Table 1 for each temperature data set. The rate of adsorption of the first structure shows a positive correlation with temperature, which is in accordance to the principles of the Kisliuk model and is revealed by the trend in the variable R′1 between temperatures of 20 and 40 °C. Increased dissociation with increased temperature influences the formation of the second structure between immersion temperatures of 20 and 40 °C. This causes the rate constant determining R′2 to decrease11 with increased temperature. Moreover, it can be seen that the presence of the first structure of MPA significantly reduces the rate of binding in each case, which is shown in the constant R′2 being significantly smaller than R′1 (see Table 1). For 25-40 °C, the values of both kE1 and kE2 decrease with increased temperature. This is as expected; the MPA molecules possess increased enthalpy at high temperatures, reducing the significance of molecular interactions which cause Θ1(t) and Θ2(t) to deviate from the Kisliuk adsorption model. Indeed, the zero values returned for the 40 °C suggest that above this temperature adsorption follows Langmuir adsorption model behavior. For the 4 °C data, values of R′1, R′2, kE1, and kE2 do not follow the trends observed for the model curves seen to occur between 20 and 40 °C. The adsorption kinetics are different from those observed from 20 to 40 °C because decreased solubility of MPA starts to affect the adsorption. The fitted isotherm models calculated by application of the new expression (eq 11) are displayed with the normalized experimental impedance data in Figure 4. The error between the normalized experimental data and the values predicted by the proposed model at 4 °C are shown in Table 2. It can be seen from Table 1 that a good fit is obtained using the model over all adsorption temperatures investigated, as every temperature data set provided a χ2 value significantly less than the number of data

938 Langmuir, Vol. 25, No. 2, 2009

Henderson et al.

Table 2. Error between the Values Predicted by the Proposed Model and the Experimental Data at a Temperature of 4°C normalized impedance immersion period (h)

predicted by proposed model

experimental data

error (%)

0 0.001389 0.004167 0.016667 0.166667 0.333333 2 8 12

0 0.119346 0.190869 0.200242 0.202400 0.204755 0.225783 0.269349 0.280808

0 0.135058 0.170344 0.203731 0.237921 0.210885 0.226506 0.279999 0.264786

0 11.63344 12.04897 1.712618 14.92957 2.907165 0.318769 3.800433 6.050958

average

5.933547

points obtained at a particular temperature. The majority of the data are scattered randomly about the model curves and can be attributed to microscopic surface irregularities, brought about during cleaning of the surface. In the typical distribution of error shown in Table 2, there is no trend between the amount of error observed and immersion period. Data provided by Himmlehaus et al. showed that vibration mode intensity data yielded by docosanethiol growth on gold24 displays a similar shaped curve to the adsorption curve displayed for MPA on Au at 20 °C, shown in Figure 4. Therefore, it is reasonable to conclude that the modeling equation may also be applicable to other SAM systems. However, it must be noted that, as the nature of the molecular interactions are different, the equation constants will be a function of the molecular interactions specific to the SAM system under consideration.

Conclusions The stages of MPA self-assembled monolayer growth on a Au working electrode have been characterized and quantified using electrochemical impedance spectroscopy (EIS) at temperatures ranging from 4 to 40 °C. Infrared reflection absorption

spectroscopy (IRRAS) was used to confirm EIS data at an immersion temperature of 20 °C. Changes in the resistance and capacitance of the monolayer over immersion time allowed the recognition of two distinct structures in monolayer self-assembly. Initially, MPA occupying the lying down configuration was formed and evolved into a structure consisting of entirely MPA in the standing up configuration. Both structures displayed an increase in disorder with increased immersion temperature above 20 °C. Also, a disordered structure at an immersion temperature of 4 °C was displayed as a result of decreased solubility of MPA in ethanol. This allowed the elucidation of a new monolayer model, which was applied to model normalized impedance data and allowed the deduction of the optimum immersion time for formation of a complete MPA lying down structure and standing up structure. The model was found to provide a good fit to all experimental data. It is thought that the proposed model is suitable for the purposes of modeling the different phases of monolayer growth, as measured via EIS. It is also expected that this model could simulate many two component systems, where sequential structures, that exhibit different properties, are adsorbed onto a substrate surface. Acknowledgment. The authors wish to thank Mr. Timothy Hawlkins of Newcastle University for performing the necessary IRRAS measurements; Dr. David Pritchard and Dr. Christine Peel of the University of Teesside for their advice on model justification; and The University of Teesside School of Science and Technology technical staff. The authors would also like to thank Immunodiagnostic Systems (IDS Ltd.) and the EPSRC for CASE studentship financial support. Supporting Information Available: An expanded discussion of impedance plateau formation at 20 and 4 °C, including Nyquist plot analysis using equivalent circuits. This material is available free of charge via the Internet at http://pubs.acs.org. LA802677N