Anal. Chem. 1999, 71, 2564-2570
Quantitative SERS Measurements on Dielectric-Overcoated Silver-Island Films by Solution-Deposition Control of Surface Concentrations William B. Lacy, Lydia G. Olson, and Joel M. Harris*
Department of Chemistry, University of Utah, Salt Lake City, Utah 84112
A simple method to control the dosing of small adsorbate molecules onto solid surfaces from liquid solution is applied to quantitative surface-enhanced Raman scattering measurements on dielectric-overcoated silver-island films. The deposition method, based on substrate withdrawal from solution, is evaluated by measuring fluorescence (ex situ) and optical absorption (in situ) of dye molecules deposited onto glass surfaces. Control of adsorbate surface concentrations was accomplished by varying the withdrawal rate and the concentration of the dye in solution. The dosing method was used to study the dependence of the electromagnetic contribution to SERS enhancement on surface coverage of scatterer. The sensitivity enhancement was found to be constant for adsorbate coverages up to 60-80% of a monolayer. Beyond a full monolayer, SERS enhancement for additional molecules deposited onto the surface was found to drop significantly, by as much as 1 order of magnitude.
Considerable research has been directed toward development of surface-sensitive and surface-selective optical spectroscopies1-3 to characterize thin films and monolayers of molecules on surfaces. These spectroscopies include infrared internal- and external-reflectance measurements, total-internal-reflection excitation of fluorescence, surface-enhanced Raman scattering, and surface-plasmon resonance spectroscopy. These methods do not generally produce data that can be interpreted quantitatively without calibration. In many studies, flattening of the spectroscopic response versus adsorbate exposure is interpreted as the development of a full monolayer, and a fractional response as a proportional reduction in surface coverage. This interpretation neglects whether the spectroscopic sensitivity is constant or coveragedependent. Optical absorption and fluorescence, for example, is influenced by dipole-dipole interactions and state mixing (exciton coupling) between resonant chromophores; these interactions depend sensitively on the distance between the adsorbate (1) Bohn, P. W.; Walls, D. J. Microchim. Acta 1991, 1, 3-35. (2) Garrell, R. L. Anal. Chem. 1989, 61, 401A-411A. (3) Suetaka, W.; Yates, J. T. Surface Infrared and Raman Spectroscopy; Plenum Press: New York, 1995.
2564 Analytical Chemistry, Vol. 71, No. 13, July 1, 1999
molecules,4-6 which can lead to a vastly different response as a surface layer accumulates. Calibration of surface concentrations is a challenging issue for the chemical analysis of interfaces. Radiochemical methods can yield absolute in situ adsorbate concentrations.7 A quartz microbalance can be used to measure the mass of an adsorbate at gassolid interfaces. For measurements at a liquid-solid interface, however, the QCM can respond to changes in the properties of the interfacial solvent,8,9 which can vary with adsorption of molecules to the solid surface and thus prevent a simple interpretation of the frequency change. An alternative approach to quantitative standardization of a surface-spectroscopic measurement is to vary the surface concentration of an adsorbate in a known way. One approach to quantitatively transfer molecules to a solid surface is LangmuirBlodgett film transfer, where concentrations of molecules on the surface are controlled by the adsorbate film pressure at the airwater interface from which the LB film is cast.10 While this technique allows accurate control of surface concentrations, it can only be used for amphiphilic molecules with low water solubility that accumulate at an air-water interface. Polymer materials are routinely deposited from liquid solution onto solid substrates by spin-coating. With this method,11 the centripetal acceleration of the underlying substrate leaves a film of polymer on the surface, the thickness of which depends on the polymer concentration, density, and viscosity of the solution and the rate of rotation of the substrate. A similar approach that is applicable to low-viscosity solutions is film deposition by substrate withdrawal from solution. In this method, a gravitational field replaces the centripetal acceleration of the substrate and allows solutions of lower viscosity to be deposited as uniform films. This method was first applied to deposition of very thin (10 nm (4) Song, Q.; Evans, C. E.; Bohn, P. W. J. Phys. Chem. 1993, 97, 13736-13741. (5) Ovchinnikov, M.; Wight, C. A. J. Chem. Phys. 1994, 100, 972-977. (6) Yamamoto, K.; Ishida, H. Appl. Spectrosc. 1994, 48, 775-787. (7) Li, M.; Schlenoff, J. B. Anal. Chem. 1994, 66, 824-829. (8) Duncan-Hewitt, W. C.; Thompson, M. Anal. Chem. 1992, 64, 94-99. (9) Yang, M.; Thompson, M.; Duncan-Hewitt, W. C. Langmuir 1993, 9, 802811. (10) Ulman, A. Ultrathin Organic Films; Academic Press: Boston, 1991; Chapter 2. (11) Emslie, A. G.; Bonner, F. T.; Peck, L. G. J. Appl. Phys. 1958, 29, 858-862. 10.1021/ac981024f CCC: $18.00
© 1999 American Chemical Society Published on Web 05/21/1999
to 1 µm) polymer films on glass and Mylar substrates.12 The substrate withdrawal method has also been used to vary the surface concentration of dye molecules on silica surfaces to observe coverage effects on optical absorption and fluorescence yields of adsorbed dye13 and, at a fixed concentration of dye, to study molecular orientation on quartz surfaces using reflection and transmission UV-visible spectroscopy.14 The method has been applied to dosing of fixed concentrations of an adsorbate onto metal-island films to investigate the enhancement of SERS, surface-enhanced resonance Raman scattering (SERRS), and surface-enhanced fluorescence.15,16 In the present paper, the quantitative control of the substratewithdrawal method is evaluated for dosing of adsorbate molecules onto glass surfaces from solution. Quantitative control of adsorbate surface concentrations is then applied to determining the surfacecoverage dependence of SERS intensity from SiO2-overcoated silver-island films.17-19 Since the SiO2-dielectric overlayer prevents contact between the adsorbate and metal-island surface,19 the results address the coverage dependence of the electromagnetic contribution to the SERS enhancement2,20 in the presence of a dielectric layer over the metal. Study of the surface-coverage dependence of SERS enhancement has been a topic of interest over the past decade.21 In early theoretical work, Kerker and co-workers22 studied dipole interactions between scattering molecules within a classical electromagnetic model for SERS in an effort to predict surface coverage effects; the calculation was limited to “two-body” interactions and coverages well below a monolayer and predicted a fractional loss in scattering cross section as the surface density is increased. Murray and Bodoff23 produced experimental evidence of the nonlinearity of SERS by dosing cyanide onto silver from the gas phase; they observed maximum scattering at ∼0.25 monolayer followed by a slight decay with increases in coverage. They modeled these results using electromagnetic theory to account for increases in surface polarizability and depolarization of surface fields as adsorbate accumulates.24 Several studies of the coverage dependence for SERRS have been reported.25-27 Zeman et al.25 evaporated cobalt phthalocyanine onto roughened silver films and observed a sharp maximum in SERRS intensity at coverages of less than 0.1 monolayer; on the basis of an electromagnetic model, they attributed this behavior to strong damping of the plasmon (12) Yang, C.-C.; Josefowicz, J. Y.; Alexandru, L. Thin Solid Films 1980, 74, 117-127. (13) Garoff, S.; Stephens, R. B.; Hanson, C. D.; Sorenson, G. K. Opt. Commun. 1982, 41, 257-262. (14) Elking, M. D.; He, G.; Xu, Z. J. Chem. Phys. 1996, 105, 6565-6573. (15) Weitz, D. A.; Garoff, S.; Gersten, J. I.; Nitzan, A. J. Chem. Phys. 1983, 78, 5324-5338. (16) Schlegel, V.; Cotton, T. M. Anal. Chem. 1991, 63, 241-247. (17) Walls, D. J.; Bohn, P. W. J. Phys. Chem. 1989, 93, 2976-2982. (18) Hill; W.; Rogalla, D.; Klockow, D. Anal. Methods Instrum. 1993, 1, 89-96. (19) Lacy, W. B.; Williams, J. M.; Wenzler, L. A.; Beebe, T. P., Jr.; Harris, J. M. Anal. Chem. 1996, 68, 1003-1011. (20) Weitz, D. A.; Moskovits, M.; Creighton, J. A. Surface-Enhanced Raman Spectroscopy; VCH Publishers: New York, 1986. (21) Aroca, R.; Kovacs, G. J. In Vibrational Spectra and Structure; Durig, J. R., Ed.; Elsevier: New York, 1991; Vol. 19, Chapter 2 and references therein. (22) Chew, H.; Wang, D-. S.; Kerker, M. Phys. Rev. B 1983, 28, 4169-4178. (23) Murray, C. A.; Bodoff, S. Phys. Rev. Lett. 1984, 52, 2273-2276. (24) Murray, C. A.; Bodoff, S. Phys. Rev. B 1985, 32, 671-688. (25) Zeman, E. J.; Carron, K. T.; Schatz, G. C.; Van Duyne, R. P. J. Chem. Phys. 1987, 87, 4189-4200. (26) Kim, J.; Cotton, T. M.; Uphaus, R. A. J. Phys. Chem. 1989, 93, 3713-3720. (27) Aroca, R.; Battisti, D. Langmuir 1990, 6, 250-254.
resonance of the metal by adsorbed dye molecules. Similar results were reported for a cynanine dye26 and a vanadylphthalocyanine27 dispersed into Langmuir-Blodgett monolayers deposited onto silver-island films; interestingly, deposition of multiple LB layers containing dye showed increases in SERRS scattering for up to several layers.27 Several recent studies of the coverage dependence of nonresonant SERS on roughened electrodes have shown a linear response to much higher surface coverages. Stolberg et al.28 controlled the surface concentration of pyridine adsorbed from solution to a roughened gold electrode by varying the electrochemical potential and while measuring the surface coverage by coulometry; the SERS intensity was found to be linear with surface coverage to ∼0.5 monolayer. In a similar study of the adsorption of pyrazine to a roughened gold electrode,29 the SERS intensity was proportional to surface concentration up to ∼0.7 monolayer. Similar results were reported for in situ SERS monitoring of oxidative adsorption of ethanethiolate onto roughened silver electrodes, where coverage was determined by both voltammetry and QCM measurements.30 The variety of results for the dependence of SERS intensity on coverage is not surprising when one considers that chemical, electromagnetic, and often resonance effects all play a role in determining the scattering enhancement. Since each of these effects is influenced in a different way by accumulation of adsorbate on the surface, the coverage dependence of scattering varies from one system to the next. In the present investigation, a silica layer deposited over the silver-island film blocks direct contact between the metal and the adsorbate; this eliminates “chemical” enhancement of scattering and thereby avoids changes in charge-transfer interactions that would accompany the accumulation of dipolar or charged adsorbates at a surface.31 Electromagnetic enhancement can be influenced by changes in the local dielectric properties of the metal interface, which are changed by accumulation of a polarizable adsorbate;22-24 in the present study, the local dielectric environment near the metal a dominated by the silica overlayer and should not be strongly affected by accumulation of molecular adsorbate. Strong optical absorption by resonant scatterers can dampen the metal plasmon resonance and lower the electromagnetic enhancement for higher coverages in SERRS;25-27 in the present study, resonant scattering (and plasmon resonance damping) should not contribute to enhancement since the lowest-lying electronic transition of the adsorbate is 15 000 cm-1 higher in frequency than the incident laser excitation. With direct adsorption to a metal-island surface, there can also be a correlation between site-specific adsorption and the corresponding enhancement of that site. Murray and Bodoff23 reported this behavior for cyanide adsorption onto silver islands, where a change in enhancement was observed with coverage as a stronger site filled and adsorption at a lower-energy site began to dominate the scattering. The differences in metal surface structures used in coverage-dependent studies could affect the results one obtains since the electromagnetic field strength (28) Stolberg, L.; Lipkowski, J.; Irish, D. E. J. Electroanal. Chem. 1991, 300, 563-584. (29) Brolo, A. G.; Irish, D. E.; Szymanski, G.; Lipkowski, J. Langmuir 1998, 14, 517-527. (30) Hatchett, D. W.; Stevenson, K. J.; Lacy, W. B.; Harris, J. M.; White, H. S. J. Am. Chem. Soc. 1997, 119, 6596-6606. (31) Frumkin, A. N. Z. Phys. Chem. 1925, 116, 466-484.
Analytical Chemistry, Vol. 71, No. 13, July 1, 1999
2565
should vary strongly with the location of the molecule on an island or particular surface structure.32,33 In the present study, the SiO2 overlayer both covers the Ag islands and fills the spaces between them,19 which should provide a more uniform enhancement and remove any correlation between adsorption-site energetics on the metal and local enhancement characteristics. The SiO2-overcoated Ag-island SERS substrate should therefore provide a simpler coverage-dependent response than is expected from direct adsorption of molecules onto a metal surface. A study of the SERS coverage dependence of this substrate is also important for understanding the response of SERS-based sensors that block direct, nonspecific adsorption of species to the metal surface and that develop a selective response.34-36 The application of selectively coated SERS substrates to analyte quantitation can benefit from a study of their coverage dependence of SERS enhancement. EXPERIMENTAL SECTION Substrate Preparation and Surface Dosing. Substrates for the surface dosing experiments were glass microscope slides. For the surface-enhanced Raman experiments, the glass slides were coated on one side with a thermally evaporated silver-island film grown to the silver mass thickness between 4.5 and 4.7 nm, producing a densely packed film of 40-nm islands as measured by AFM.19 The silver-island film was overcoated with a thin (6nm) layer of SiO2. The thermally evaporated SiO2 greatly reduces the measured surface roughness of the island film, from σ ) 1.8 nm to
