Anal. Chem. 1997, 69, 471-477
Nanoscale Characterization of Gold Colloid Monolayers: A Comparison of Four Techniques Katherine C. Grabar,†,§ Kenneth R. Brown,† Christine D. Keating,† Stephan J. Stranick,‡,⊥ Sau-Lan Tang,‡ and Michael J. Natan*,†
Department of Chemistry, The Pennsylvania State University, 152 Davey Laboratory, University Park, Pennsylvania 16802-3600, and DuPont Central Research and Development, Experimental Station, P.O. Box 80356, Wilmington, Delaware 19880-0356
Atomic force microscopy (AFM), field emission scanning electron microscopy (FE-SEM), transmission electron microscopy (TEM), and near-field scanning optical microscopy (NSOM) have been used to characterize the nanostructure of Au colloid-based surfaces. Because these substrates are composed of particles whose dimensions are known prior to assembly, they are well-suited for a critical comparison of the capabilities and limitations of each nanoscale imaging technique. The three criteria for this comparison, which are relevant to the field of nanoparticle assemblies in general, are (i) accuracy in establishing particle size, particle coverage, and interparticle spacing; (ii) accuracy in delineating surface topography; and (iii) ease of sample preparation, data acquisition, and image analysis. For colloidal Au arrays, TEM gives the most reliable size and spacing information but exhibits the greatest constraints with regard to sample preparation; in contrast, AFM is widely applicable but yields data that are the least straightforward to interpret. For accurate information regarding nanometer-scale architecture of particle-based surfaces, a combination of at least one scanning probe method (AFM, NSOM) and one accelerated-electron method (TEM, FE-SEM) is required. Controlling the nanostructure of solids and thin films is among the most important problems in materials chemistry today. Catalysis, information storage, nonlinear optics, sensors, semiconductor photoelectrochemistry, and a host of other fields depend on production of materials that exhibit a well-defined spatial relationship between features in both two and three dimensions. Accordingly, nanostructure characterization is becoming an increasingly common aspect of materials chemistry.1-15 A particularly good example is the nanometer-scale analysis of noble metal surfaces, where there exist well-defined relationships between surface morphology and bulk physical properties. For instance, surface-enhanced Raman scattering (SERS)16 depends critically on the nanometer-level details of Ag, Au, or Cu surface topography, and there have been numerous studies describing the nanostructure of thin-film substrates designed for SERS.17-20 More generally, continuous and discontinuous Au or †
The Pennsylvania State University. DuPont Central Research and Development. Current address: Union Carbide Corp., P.O. Box 8361, South Charleston, WV 25303. ⊥ Current address: National Institute of Standards and Technology, B248 Chemistry Building, Chemical Science and Technology Laboratory, Gaithersburg, MD 20899. (1) Chang, S.-Y.; Liu, L.; Asher, S. A. J. Am. Chem. Soc. 1994, 116, 67396744. ‡ §
S0003-2700(96)00596-3 CCC: $14.00
© 1997 American Chemical Society
Ag films have been characterized by field emission scanning electron microscopy (FE-SEM),21 transmission electron microscopy (TEM),22-25 scanning tunneling microscopy (STM)26 and photon STM,27 atomic force microscopy (AFM),28 and near-field scanning optical microscopy (NSOM).29 These studies have afforded a morphological description of continuous Au and Ag thin films prepared by a variety of methods. Thin films composed of discrete, nanometer-sized Au or Ag particles present opportunities for characterization30-40 beyond the rudimentary topographical details that can be determined for continuous or discontinuous films. For a single layer of nanoparticles, determination of particle dimensions (i.e., diameter for spherical particles, major and minor axes for nonspherical particles), interparticle spacing (i.e., most frequent distance to nearest neighbor), the planarity of particles (i.e., vertical position relative to an underlying substrate), and overall particle number density (i.e., number/unit area) yields a thorough description of two-dimensional (2-D) nanometer-scale architecture.41-44 These parameters are important for understanding bulk properties of (2) Kondo, M.; Shinozaki, K.; Bergstrom, L.; Mizutani, N. Langmuir 1995, 11, 394-397. (3) Kamenetzky, E. A.; Magliocco, L. G.; Panzer, H. P. Science 1994, 207210. (4) Coury, J. E.; Pitts, E. C.; Felton, R. H.; Bottomley, L. A. J. Vac. Sci. Technol. B 1995, 13, 1167-1171. (5) Duteil, A.; Schmid, G.; Meyer-Zaika, W. J. Chem. Soc., Chem. Commun. 1995, 31-32. (6) Golan, Y.; Hodes, G.; Rubinstein, I. J. Phys. Chem. 1996, 100, 2220-2228. (7) Motte, L.; Billoudet, F.; Pileni, M. P. J. Phys. Chem. 1995, 99, 16425-16429. (8) Alivisatos, A. P. Science 1996, 271, 933-937. (9) Gao, M.; Zhang, X.; Yang, B.; Shen, J. J. Chem. Soc., Chem. Commun. 1994, 2229-2230. (10) Heath, J. R.; Williams, R. S.; Shiang, J. J.; Wind, S. J.; Chu, J. D’Emic, D.; Chen, W.; Stanis, C. L.; Bucchignano, J. J. J. Phys. Chem. 1996, 100, 31443149. (11) Ogawa, S.; Fan, F.-R. F.; Bard, A. J. J. Phys. Chem. 1995, 99, 11182-11189. (12) Yang, J.; Fendler, J. J. Phys. Chem. 1995, 99, 5505-5511. (13) Huck, W. T. S.; van Veggel, F. C. J. M.; Kropman, B. L.; Blank, D. H. A.; Keim, E. G.; Smithers, M. M. A.; Reinhoudt, D. N. J. Am. Chem. Soc. 1995, 117, 8293-8294. (14) Murray, C. B.; Kagan, C. R.; Bawendi, M. G. Science 1995, 270, 13351338. (15) Wang, D.; Liu, Y.; Xu, G.; Bai, C. J. Vac. Sci. Technol. B 1994, 12, 18981900. (16) Brandt, E. S.; Cotton, T. M. In Investigations of Surfaces and Interfaces, Part B, 2nd ed.; Rossiter, B. W., Baetzold, R. C., Eds.; John Wiley & Sons: New York, 1993; Vol. IXB, Chapter 8, pp 633-718. (17) Van Duyne, R. P.; Hulteen, J. C.; Treichel, D. A. J. Chem. Phys. 1993, 99, 2101-2115. (18) Roark, S. E.; Semin, D. J.; Rowlen, K. L. Anal. Chem. 1996, 68, 473-480. (19) Lacy, W. B.; Williams, J. M.; Wenzler, L. A.; Beebe, T. P., Jr.; Harris, J. M. Anal. Chem. 1996, 68, 1003-1011. (20) Alak, A. M.; Vo-Dinh, T. Anal. Chem. 1989, 61, 656-660. (21) Naitoh, M.; Shoji, F.; Oura, K. Jpn. J. Appl. Phys. 1992, 31, 4018-4019. (22) Krakow, W. J. Appl. Phys. 1990, 69, 2206-2210. (23) Davis, C. A.; McKenzie, D. R.; McPhedran, R. C. Opt. Commun. 1991, 85, 70-82.
Analytical Chemistry, Vol. 69, No. 3, February 1, 1997 471
nanoparticulate noble metal films; importantly, they are relevant to characterization of any nanoparticulate film. What is the best way to acquire this information? Immobilized Au and Ag nanoparticles have previously been studied using one of a variety of techniques (SEM, TEM, STM, and AFM).30-40 This approach can be problematic, particularly when the topographic information is convoluted with instrument response.45,46 Moreover, when two approaches have been employed, they have involved different samples. This paper describes the nanoscale characterization of self-assembled, covalently bound 2-D arrays of colloidal Au particles41-44 using AFM, FE-SEM, TEM, and NSOM. The use of four different methods to study exceedingly similar or, in several cases, identical samples allows the relative advantages/disadvantages of each technique to be elaborated. EXPERIMENTAL SECTION Materials. (3-Aminopropyl)methyldiethoxysilane (APMDES) and (3-mercaptopropyl)methyldimethoxysilane (MPMDMS) were obtained from United Chemical Technologies. CH3OH (spectrophotometric grade) was obtained from EM Science. HOOC(CH2)2SH and HAuCl4 were purchased from Aldrich. All H2O was distilled through a Barnstead Nanopure H2O purification system and had a resistance of 18 MΩ. Substrates were obtained as follows: glass microscope slides from Fisher Scientific, Sb-doped SnO2 (100 Ω/square) from Delta Technologies, formvar- and SiOxcoated TEM grids from Ted Pella, Inc., In-doped SnO2 (30 Ω/square) from PPG Industries, Inc., and Au/Cr/Si samples from the Allara group at Penn State. Colloid Preparation. Small Au particles (“Sol A”) were prepared by citrate reduction of HAuCl4 in aqueous solution.41-44 (24) Arai, M.; Mitsui, M.; Ozaki, J.-I.; Nishiyama, Y. J. Colloid Interface Sci. 1994, 168, 473-477. (25) Blacher, S.; Brouers, F.; Gadenne, P.; Lafait, J. J. Appl. Phys. 1993, 74, 207213. (26) Rucker, M.; Knoll, W.; Rabe, J. P. J. Appl. Phys. 1992, 72, 5027-5031. (27) Dawson, P.; de Fornel, F.; Goudonnet, J.-P. Phys. Rev. Lett. 1994, 72, 29272930. (28) de Hollander, R. B. G.; van Hulst, N. F.; Kooyman, R. P. H. Ultramicroscopy 1995, 57, 263-269. (29) Fischer, U. C. J. Opt. Soc. Am. B 1986, 3, 1239-1244. (30) Zoval, J. V.; Stiger, R. M.; Biernacki, P. R.; Penner, R. M. J. Phys. Chem. 1996, 100, 837-844. (31) Masuda, H.; Fukuda, K. Science 1995, 268, 1466-1468. (32) Foss, J. C. A.; Hornyak, G. L.; Stockert, J. A.; Martin, C. R. J. Phys. Chem. 1992, 96, 7497-7499. (33) Casella, I. G.; Destradis, A.; Desimoni, E. Analyst 1996, 121, 249-254. (34) Chumanov, G.; Sokolov, K.; Gregory, B. W.; Cotton, T. M. J. Phys. Chem. 1995, 99, 9466-9471. (35) Doron, A.; Katz, E.; Willner, I. Langmuir 1995, 11, 1313-1317. (36) Giersig, M.; Mulvaney, P. Langmuir 1993, 9, 3408-3413. (37) Shaiu, W.-L.; Vesenka, J.; Jondle, D.; Henderson, E.; Larson, D. D. J. Vac. Sci. Technol. A 1993, 11, 820-823. (38) Dolan, C.; Yuan, Y.; Jao, T.-C.; Fendler, J. H. Chem. Mater. 1991, 3, 215218. (39) Peschel, S.; Schmid, G. Angew. Chem., Int. Ed. Engl. 1995, 34, 1442-1443. (40) Terrill, R. H.; Postlethwaite, T. A.; Chen, C.-h.; Poon, C.-D.; Terzis, A.; Chen, A.; Hutchison, J. E.; Clark, M. R.; Wignall, G.; Londono, J. D.; Superfine, R.; Falvo, M.; Johnson, C. S., Jr.; Samulski, E. T.; Murray, R. W. J. Am. Chem. Soc. 1995, 117, 12537-12548. (41) Grabar, K. C.; Smith, P. C.; Musick, M. D.; Davis, J. A.; Walter, D. G.; Jackson, M. A.; Guthrie, A. P.; Natan, M. J. J. Am. Chem. Soc. 1996, 118, 1148-1153. (42) Freeman, R. G.; Grabar, K. C.; Allison, K. J.; Bright, R. M.; Davis, J. A.; Guthrie, A. P.; Hommer, M. B.; Jackson, M. A.; Smith, P. C.; Walter, D. G.; Natan, M. J. Science 1995, 267, 1629-1632. (43) Grabar, K. C.; Freeman, R. G.; Hommer, M. B.; Natan, M. J. Anal. Chem. 1995, 67, 735-743. (44) Grabar, K. C.; Allison, K. J.; Baker, B. E.; Bright, R. M.; Brown, K. R.; Freeman, R. G.; Fox, A. P.; Keating, C. D.; Musick, M. D.; Natan, M. J. Langmuir 1996, 12, 2353-2361. (45) Griffith, J. E.; Grigg, D. A. J. Appl. Phys. 1993, 9, 83-109. (46) Keller, D. Surf. Sci. 1991, 253, 353-364.
472
Analytical Chemistry, Vol. 69, No. 3, February 1, 1997
For this procedure, average particle diameters (as measured by TEM) were consistently in the range of 12-14 nm, with standard deviations of 1-2 nm. One set of larger Au particles (“Sol B”) was prepared through a slight modification of a previously described procedure,43 resulting in elliptical particles of major axis 53 ( 5 nm and minor axis 44 ( 4 nm. Au particles with 31- and 45-nm diameters (major axes) were prepared by adding 12-nmdiameter particles and then sodium citrate to refluxing solutions of 0.01% HAuCl4, yielding particles with a narrower size dispersity47 than those obtained by conventional citrate reduction.48 Substrate Derivatization. Sample preparation followed previously described protocols.41-44 Briefly, cleaned substrates were derivatized with dilute organosilane or alkanethiol solutions. After exhaustive rinsing, the substrates were then immersed in aqueous solutions of dilute Au colloid (between 1 and 20 nM in particles) for time periods ranging from a few minutes to several days, thoroughly rinsed, and then stored in H2O. Prior to imaging, samples were rinsed sequentially with CH3OH/H2O mixtures with increasing CH3OH percentages (0-100%) and allowed to dry in air. Instrumentation. AFM measurements were performed with either a Digital Instruments Nanoscope III or a Burleigh Personal AFM. Imaging was performed under ambient conditions using standard Si cantilevers. FE-SEM images were acquired with a JEOL JSM 6320F or with a JEOL 6300 FV instrument. TEM analysis was performed on a JEOL Model 1200 EXII instrument operating at 80-kV accelerating voltage. NSOM analysis was performed with an in-house instrument that utilized shear force measurements as a feedback mechanism to maintain constant probe height above the surface.49 Other instrumentation has been described previously.41-44 Sample Analysis. Particle sizes obtained from TEM or FESEM data represent an average of >100 particles and were calculated using previously described methods.41-44 RESULTS AND DISCUSSION Atomic Force Microscopy. Contact mode AFM has been carried out on silanized glass slides prior to and after deposition of colloidal Au (Supporting Information). Except for occasional protrusions of ∼5 nm in height, the silanized surface is flat, exhibiting an rms roughness of 0.8 nm over a 4-mm2 area. The line scan for the colloidal Au-derivatized surface indicates height variations of less than 2 nmsmuch lower than expected for 12nm Au particlessyet a strong absorbance in the optical spectrum at 520 nm (not shown) confirms the presence of a large number of Au particles on the surface. Repeated attempts at imaging in contact mode yielded similar results and suggest that the force of the probe tip is sufficient to move particles across the substrate surface. The presence of streaks in some of the contact mode images and the observed instability of these colloid films to mechanical forces (e.g., scratching with tweezers) support this conclusion. The problems associated with tip-induced particle mobility are improved considerably through the use of intermittent contact mode (“tapping mode”). In this configuration, the cantilever is oscillated near its resonant frequency so that the probe tip intermittently makes contact with the sample surface. Changes in the amplitude of the vibrating cantilever are related to the (47) Brown, K B.; Walter, D. G.; Natan, M. J., manuscript in preparation. (48) Hayat, M. A., Ed. Colloidal Gold: Principles, Methods, and Applications; Academic Press: San Diego, CA, 1989; Vol. 1. (49) Betzig, E.; Finn, P.; Weiner, J. Appl. Phys. Lett. 1992, 60, 2484-2486.
Figure 1. AFM images (500 nm × 500 nm) of MPMDMS-coated glass (a, b), of SAMs of HOOC(CH2)2SH on Au/Cr/Si (c, d), and of MPMDMScoated In-doped SnO2 (e, f) surfaces, before (top set) and after (bottom set) immobilization of 12-nm-diameter colloidal Au particles.
attractive van der Waals forces acting between the tip and surface. Frictional forces that can damage the substrate surface are typically eliminated. Tapping mode AFM images of three different substrates before and after Au colloid monolayer derivatization are shown in Figure 1. Because the contrast mechanism in AFM derives from sample-tip forces, polymer-coated samples are as easily imaged as Au-treated samples; more importantly, AFM data yield direct evidence regarding the planarity of the substrate. Two of the three surfaces imaged, an MPMDMS-coated glass slide (panels a, b) and a self-assembled monolayer (SAM) film of HOOC(CH2)2SH (panels c, d), exhibit minimal nanoscale roughness (30-nm diameter) colloidal Au particles is well-established,56 (56) Goodman, S. L.; Hodges, G. M.; Trejdosiewicz, L. K.; Livingston, D. C. J. Microsc. 1981, 123, 201-213.
476
Analytical Chemistry, Vol. 69, No. 3, February 1, 1997
but even small colloidal particles (i.e., 30-nm-diameter particles. The weaknesses of FE-SEM are similar to those discussed for TEM, the principal problems being the lack of topographic information about the substrate or the colloidal film. The utility of AFM for colloidal Au array nanostructure determination is amply demonstrated by the data in Figures 1-3. Generation of topographic information is central to the instrument’s operational mechanism and is available for macroscopic substrates before and after deposition of organosilane, as well as after Au colloid immobilization. Comparison of the image for a Au colloid multilayer (Figure 3) with those of monolayers indicates that AFM is suitable for distinguishing between single layers and multilayers of particles on the surface. Approximate particle number densities are available by feature counting. Sample preparationsmounting a sample onto a sample holdersis trivial, and AFM experiments can be run rapidly on nearly any sample type. All these benefits make AFM nearly mandatory for characterization of nanoparticle arrays. Nevertheless, there are obvious shortcomings with the technique as well. The topographic information furnished by AFM is difficult to interpret when the substrate roughness and particle diameter are similar in size. This underscores the lack of chemical information obtained by routine AFM: an organosilane protrusion is indistinguishable from a Au colloid bump in tapping mode AFM, a phenomenon we have observed on many occasions. Moreover, accurate size and spacing information is not directly available from AFM images due to tip convolution. Software that deconvolutes the tip profile from the image have been described, but their value in analyzing nanoparticle arrays is limited, for two reasons: (i) They usually require calibration with incompressible, spherical objectssoften colloidal Au particles.37 Unfortunately, as mentioned above, colloidal Au particles are actually not spherical. (ii) Center-to-center interparticle spacing does not require actual knowledge of particle shape. On the other hand, edge-to-edge particle spacing, which is critical to understanding bulk optical behavior, does require knowledge of particle size. For arrays of particles that are not truly monodisperse, accurate knowledge of particle shapes is essential, and it is not clear how deconvolution algorithms will perform on irregularly shaped tips probing irregularly sized and shaped particles. NSOM overcomes the tip convolution difficulties encountered in AFM, at least in the particle size range of ∼15 nm and larger. The shear force topographical imaging mode operates at up to a few nanometers from the surface, minimizing intrusive interactions of the tip with the sample. As mentioned above, the chemical identification inherent in optical microscopy adds a desirable dimension to nanoscale particle characterization. Applicability to General Issues in Nanoscale Imaging. Reports of nanostructured materials are prevalent in the recent literature.1-15,17-40 Some of these materials derive their nanoscale features from metal or semiconductor particles; others rely upon lithographic techniques or metal deposition into porous membranes. In all of these studies, nanoscale imaging provides important information for understanding the material’s performance and characteristics. Collectively, these studies include (58) Mulvaney, P.; Giersig, M. J. Chem. Soc., Faraday Trans. 1996, 3137-3143.
characterization by AFM, SEM, TEM, and NSOM. The application of these techniques to the characterization of Au colloid arrays in this work and others58 has made clear some principles regarding the best use of imaging capabilities for nanostructure elucidation. The following principles apply generally to characterization requirements for a variety of nanostructures. (i) TEM characterization of macroscopic surfaces is limited to the fragile films formed by sectioning of larger samples and to thin films that can be transferred from a solid support to a copper mesh grid.50 Many samples are not amenable to either of these treatments; thus, the small size requirements and the necessity of the support to be transparent to the electron beam limit TEM’s use in nanostructure characterization of large surface areas. (ii) SEM analysis, which in the past was limited to conducting samples, has been further developed through the field emission sources and new in-lens detection methods to permit the imaging of nonconducting samples. At low accelerating voltages, these samples can be imaged without a conductive coating of C or Cr. Without the inlens detection apparatus, insulating substrates may still require a conductive coating. For most applications, this coating does not obscure the features of interest. However, the best resolution available (∼1 nm) is still nearly a factor of 10 worse than that possible with state-of-the art TEM instrumentation. Moreover, the coating does prevent subsequent use of the sample. (iii) AFM has proven to be a very versatile imaging technique; it has no limitations regarding the composition or size of the sample surface. For relatively densely packed colloidal arrays, quantitative analysis of particle coverage and spacing is obscured, although qualitative measurements are still informative. On rougher substrates like SnO2, the most reliable information available is a yes-no answer to the presence of small (12 nm) colloidal Au particles. Like TEM, AFM is useful in distinguishing relative differences in nanostructure. (iv) For particles with strong extinction or emission in the ultraviolet-visible region of the spectrum, the combination of NSOM and shear force AFM is a powerful method for nanostructure characterization. (v) Most importantly, even when the particle size is independently known prior to immobilization, no single technique yields all the information needed to fully characterize a 2-D nanoparticle array: a combination of AFM and TEM/FE-SEM is needed. The increased availability of this instrumentation should ultimately make this level of analysis routine. Acknowledgment is made to the National Science Foundation (CHE 92-08614, CHE 92-56692, and CHE-9307485) and to the Beckman Foundation for partial support of this research, as well as to the Eastman Kodak Co. for a graduate fellowship to K.C.G. Acknowledgment is also made to the Electron Microscopy Facility for the Life Sciences in the Biotechnology Institute at The Pennsylvania State University. We thank JEOL USA, Inc., for FESEM analyses and the Allara group for a gift of Au-coated Si samples. SUPPORTING INFORMATION AVAILABLE Contact mode AFM images, triangular AFM artifacts, AFM tip convolution of large and small colloidal particles, and slope mode AFM images of SnO2 as well as colloidal Au/SnO2 (4 pages). Ordering information is given on any current masthead page. Received for review June 14, 1996. Accepted October 27, 1996.X AC9605962 X
Abstract published in Advance ACS Abstracts, January 1, 1997.
Analytical Chemistry, Vol. 69, No. 3, February 1, 1997
477
