Preparation, Structure, and Optical Properties of Nanoporous Gold

Jan 24, 2007 - The thin film samples were prepared directly on a QCM crystal face and then were mounted in the measurement head, located in a vacuum c...
1 downloads 10 Views 516KB Size
2414

Langmuir 2007, 23, 2414-2422

Preparation, Structure, and Optical Properties of Nanoporous Gold Thin Films Matthew C. Dixon,† Thomas A. Daniel,† Mitsunori Hieda,‡,§ Detlef M. Smilgies,| Moses H. W. Chan,‡ and David L. Allara*,† Department of Chemistry and The Materials Research Institute, PennsylVania State UniVersity, UniVersity Park, PennsylVania 16802, Department of Physics, PennsylVania State UniVersity, UniVersity Park, PennsylVania 16802, and CHESS, Wilson Laboratory, Cornell UniVersity, Ithaca, New York 14853 ReceiVed August 5, 2006. In Final Form: NoVember 16, 2006 Thin nanoporous gold (np-Au) films, ranging in thickness from ∼40 to 1600 nm, have been prepared by selective chemical etching of Ag from Ag/Au alloy films supported on planar substrates. A combination of scanning electron microscopy (SEM) imaging, synchrotron grazing incidence small angle X-ray scattering, and N2 adsorption surface area measurements shows the films to exhibit a porous structure with intertwined gold fibrils exhibiting a spectrum of feature sizes and spacings ranging from several to hundreds of nanometers. Spectroscopic ellipsometry measurements (300-800 nm) reveal the onset of surface plasmon types of features with increase of film thicknesses into the ∼200 nm film thickness range. Raman scattering measurements for films functionalized with a self-assembled monolayer formed from 4-fluorobenzenethiol show significant enhancements which vary sharply with film thickness and etching times. The maximum enhancement factors reach ∼104 for 632.8 nm excitation, peak sharply in the ∼200 nm thickness range for films prepared at optimum etching times, and show high spot to spot reproducibility with ∼1 µm laser spot sizes, an indication that these films could be useful as durable, highly reproducible surface-enhanced Raman substrates.

1. Introduction Selectively leaching one component from a bimetallic alloy to leave behind a porous structure has been practiced for centuries. Specifically, refinement of Au from its alloys by selective dissolution of the less noble metal is a well-known industrial process and dates as far back as pre-Colombian Indians in South America, where it may have been the basis of the art of depletion gilding.1 More recently, the dealloying mechanism of gold alloys has dominated work in this field.2 The AgX/Au1-X alloy is of particular fundamental interest since the two metals have same unit cell structure (fcc) with nearly identical lattice constants (within 0.2%) across the full compositional range.3 It has been reported that gold samples prepared by chemical etching of the silver from AgX/Au1-X alloys in the composition range of X ) 0.5-0.8 exhibit porous bulk nanostructures with variations in pore diameters, similar to porous Vycor.4,5 Chan and coworkers have reported the use of these nanoporous gold (np-Au) structures5-10 for He superfluidity studies. Supported thin film forms of np-Au are of particular interest since they provide * To whom correspondence should be addressed. E-mail: dla3@ psu.edu. † Department of Chemistry and The Materials Research Institute, Pennsylvania State University. ‡ Department of Physics, Pennsylvania State University. § Present address: Department of Physics, Nagoya University, Furo-cho, Chikusa 464-8602, Japan. | Cornell University. (1) Forty, A. J. In Sir Charles Frank: An 80th Birthday Tribute; Chamber, R. B., Ed.; Adam Hilger: Bristol, England, 1991. (2) Erlebacher, J.; Aziz, M. J.; Karma, A.; Dimitrov, N.; Sieradzki, K. Nature 2001, 410, 450-453 and references therein. (3) Forty, A. J.; Durkin, P. Philos. Mag. A 1980, 42, 295-318. (4) Li, R.; Sieradzki, K. Phys. ReV. Lett. 1992, 68, 1168-1171. (5) Yoon, J.; Chan, M. H. W. Phys. ReV. Lett. 1997, 78, 4801-4804. (6) Csathy, G. A.; Tulimieri, D.; Yoon, J.; Chan, M. H. W. Phys. ReV. Lett. 1998, 80, 4482-4485. (7) Tulimieri, D. J.; Yoon, J.; Chan, M. H. W. Phys. ReV. Lett. 1999, 82, 121-124. (8) Csathy, G. A.; Chan, M. H. W. J. Low Temp. Phys. 2000, 121, 451-458. (9) Csathy, G. A.; Chan, M. H. W. Phys. ReV. Lett. 2001, 87, 045301. (10) Kim, E.; Chan, M. H. W. J. Low Temp. Phys. 2005, 138, 859-864.

confined geometry media in a convenient, well-defined planar form. Thin films appear to have been first developed and utilized in studies of the dealloying mechanism in which transmission electron microscopy (TEM) or scanning tunneling microscopy (STM) characterization was used to follow the structure evolution during etching of supported AgX/Au1-X films formed by thermal3,11-13 or sputter14 co-deposition. Subsequently, Chan and co-workers reported the preparation of np-Au films in the ∼1 µm thickness range and characterized their utility for quartz crystal microbalance measurements, including He superfluidity and gas adsorption measurements.15,16 Preliminary physical characterization of the fully dealloyed film showed internal surface areas ∼40 times the geometric area of the planar substrate surfaces and a wormlike structure with ∼30 nm diameter interconnected fibrils similar to the bulk structure.15 Given the rich electromagnetic (EM) response of gold nanostructures,17 including planar thin films,18 periodic gratings,19 roughened surfaces,19 and nanoparticle arrays,17 and the wide ranging ability of gold surfaces to support self-assembled monolayer (SAM) functionalization chemistry,20 it was of interest to us to pursue the physical characterization and chemical behavior of our films in greater detail and to correlate the behaviors with the detailed nanostructures. In this first study, we focus on the (11) Oppenheim, I. C.; Trevor, D. J.; Chidsey, C. E. D.; Trevor, P. L.; Sieradzki, K. Science 1991, 254, 687-689. (12) Durkin, P.; Forty, A. J. Philos. Mag. A 1982, 45, 95-105. (13) Forty, A. J. Gold Bull. 1981, 14, 25-35. (14) Wagner, K.; Brankovic, S. R.; Dimitrov, N.; Sieradzki, K. J. Electrochem. Soc. 1997, 144, 3545-3555. (15) Hieda, M.; Garcia, R.; Dixon, M.; Daniel, T.; Allara, D.; Chan, M. H. W. Appl. Phys. Lett. 2004, 84, 628-630. (16) Hieda, M.; Clark, A. C.; Chan, M. H. W. J. Low Temp. Phys. 2004, 134, 91-96. (17) Burda, C.; Chen, X.; Narayanan, R.; El-Sayed, M. A. Chem. ReV. 2005, 105, 1025-1102. (18) Kretschmann, E.; Raether, H. Z. Naturforsch. 1968, 23, 2135. (19) Raether, H. Surface Plasmons on Smooth and Rough Surfaces and on Gratings; Springer-Verlag: Berlin, 1988. (20) For example, see: Ulman, A. Chem. ReV. 1996, 96, 1533-1554 or Love, J. C.; Estroff, L. A.; Kriebel, J. K.; Nuzzo, R. G.; Whitesides, G. M. Chem. ReV. 2005, 105, 1103-1169.

10.1021/la062313z CCC: $37.00 © 2007 American Chemical Society Published on Web 01/24/2007

Properties of Nanoporous Gold Thin Films

Figure 1. Schematic of the ion beam sputter-deposited thin film alloy structures utilized to prepare the nanoporous gold film samples. Layer thicknesses are given in parentheses. In our experiments, X ∼ 0.68.

Langmuir, Vol. 23, No. 5, 2007 2415

(SEM), grazing incidence small angle X-ray scattering (GISAXS), adsorption isotherm surface area measurements, and electrochemistry. These data show that the films consist of highly porous structures with a wide range of internal feature sizes from a few to tens of nanometers. Dielectric function spectra, determined from spectroscopic ellipsometry (300-800 nm) measurements, show a complex dependence on film thickness and, consistent with the range of physical feature sizes in any given sample, this behavior cannot be quantitatively explained in terms of simple effective medium models of gold-void composites. Preliminary experiments on the SERS responses (632.8 nm excitation) were determined for a self-assembled monolayer of test aromatic thiol molecules. Intensity enhancements of ∼104, as estimated by standard methods, were observed with relatively high spot to spot and sample to sample reproducibility, an indication that these nanoporous Au films could be useful as SERS substrates for applications such as trace chemical sensing. 2. Experimental

Figure 2. DPXPS analysis of 500 nm thick np-Au films which underwent etch times of 6, 12, and 24 h. After the initial sputtering/ analysis cycle, the Ag concentration drops to ∼1% until the adhesion Au layer is reached. The Ag and Au compositions were determined from the Ag 3d and Au 4f peaks (pass energy of 160 eV and dwell times of 0.6 s for Ag and 0.3 s for Au). For details, see text.

Figure 3. Ag compositions from XPS for various thickness np-Au films formed at different etching times. Inset shows the expanded region from 0 to 22% Ag. The analysis details are the same as given for the DPXPS data in Figure 2.

physical characterization and selected aspects of the EM responses, including the ability of these films to serve as useful substrates for surface-enhanced Raman spectroscopy (SERS).21-25 In subsequent publications, we will report on the surface functionalization and related properties. The films were prepared by sputter depositing AgX/Au1-X alloy films, with X in the 0.6-0.7 range, at thicknesses ranging from 40 to 1600 nm onto planar substrates and selectively leaching out the Ag using nitric acid. The physical characteristics of the resulting films were characterized by a combination of X-ray photoelectron spectroscopy (XPS), scanning electron microscopy (21) Kneipp, K.; Kneipp, H.; Itzkan, I.; Dasari, R. R.; Feld, M. S. Chem. ReV. 1999, 99, 2957-2975. (22) Otto, A.; Mrozek, I.; Grabhorn, H.; Akemann, W. J. Phys.: Condens. Matter 1992, 4, 1143-1212. (23) Moskovits, M. ReV. Mod. Phys. 1985, 57, 783-826. (24) Otto, A. In Light Scattering in Solids IV; Cardona, M., Guntherodt, G., Eds.; Springer-Verlag: Berlin, 1984. (25) Van Duyne, R. P. In Chemical and Biochemical Applications of Lasers; Moore, C. B., Ed.; Academic Press: New York, 1979; Vol. 4, pp 101-184.

2.1. Materials. Hydrogen peroxide and ammonium hydroxide were obtained from Brand-Nu Laboratories, sulfuric acid and nitric acid were from VWR, cerium ammonium nitrate (A.C.S. grade) was from J. T. Baker, ethanol (pure) was from Pharmco, and anhydrous hexadecane (99% purity) and potassium nitrate (A.C.S. grade) were from Aldrich. Water was purified using a Millipore purification system to deionize (>18 mΩ cm resistivity) and remove organic impurities. Silicon wafer substrates [2 in. round, ∼0.05 cm thick; either (100) or (111) orientation] were cut to fit the specific analytical technique needed. Quartz crystal microbalance (QCM) crystals were AT-cut with 4 MHz or 5 MHz fundamental frequency and were obtained from Lap-Tech Inc. or Maxtek Inc. All metals for substrate deposition were >99.99% pure including Au, Cr, and a customprepared Ag68Au32 alloy sputter target. High-purity 4-fluorobenzenethiol was purchased from Aldrich and was recrystallized in ethanol prior to use. 2.2. Sample Preparation. Before metal deposition, the silicon or bare QCM substrates were cleaned using a H2SO4(conc)/H2O2(30%) 3:1 (V:V) mixture at ∼80 °C (caution: reacts violently with organics). The cleaned samples were loaded in the sputtering chamber and ∼5-10 nm of Cr followed by ∼100-200 nm of Au were deposited as an adhesion layer.26 Without this adhesion layer, the AgX/Au1-X alloys would often delaminate during etching. Various thicknesses of alloy were subsequently sputter deposited and the mass/unit area deposited was tracked by QCM (Sigma SQM-160). The mass coverages were converted to the corresponding thicknesses of planar films using 13.13 g/cm3 for the mean density of the AgX/ Au1-X alloy.27 The maximum alloy film thicknesses in our study were limited to ∼2 µm but thicker samples can readily be made by longer deposition times. The depositions were done by accelerating 8 kV Ar atoms (3 mA current) from two ion guns (IBS/e system guns; South Bay Tech, San Clemente, CA) focused on the selected target material (background pressure ∼1 × 10-5 Torr). Sample substrates were kept titled ∼70° from the surface normal and were rotated at ∼600 Hz to ensure uniform deposition.28 Typical sputtering rates were ∼0.003-0.009 nm/s, with the highest rate used for the AgX/Au1-X (26) By trial and error, it was found that surface attachment of the AgX/Au1-X alloy required a clean gold adhesion layer. Otherwise, the alloy would delaminate from the substrate during HNO3 etching. Freshly sputtered gold surfaces adhered AgX/Au1-X the best (i.e., it was best to use freshly made samples as opposed to making a bunch of gold samples and then periodically cleaning one when necessary with either H2SO4:H2O2 or H2O2:NH4OH:H2O), although in preliminary trials thermally deposited gold when used immediately usually adhered alloy upon etching. (27) These QCM thicknesses will be used for convenience for the sample designation throughout the text. Cross-sectional SEM images were taken for a variety of the different thicknesses to determine a direct physical thickness. Typically, the SEM thickness values were ∼0.9 of the QCM values. (28) This particular configuration was recommended by the manufacturer to ensure an even coating.

2416 Langmuir, Vol. 23, No. 5, 2007

Dixon et al.

Figure 4. SEM images showing the time evolution of structure due to HNO3 etching of AgX/Au1-X films. Left: Top down 150 000× images of 400 nm alloy films etched at different times (marked on each image in hours). Right: Cross-sectional 50 000× images of 1600 nm alloy films etched at different times (marked on each image in hours). alloy. These conditions produced extremely smooth surfaces (