Simple, Readily Controllable Palladium Nanoparticle Formation on

Department of Chemical and Biological Engineering, Tufts University, Medford Massachusetts 02144. ‡ X-ray Science Division, Argonne National Laborat...
0 downloads 0 Views 4MB Size
Download PDF
pubs.acs.org/Langmuir © 2009 American Chemical Society

Simple, Readily Controllable Palladium Nanoparticle Formation on SurfaceAssembled Viral Nanotemplates Amy K. Manocchi,† Nicholas E. Horelik,† Byeongdu Lee,‡ and Hyunmin Yi*,† †

Department of Chemical and Biological Engineering, Tufts University, Medford, Massachusetts 02144, and ‡ X-ray Science Division, Argonne National Laboratory, Argonne, Illinois 60439 Received August 24, 2009. Revised Manuscript Received October 30, 2009

Transition-metal nanoparticles possess unique size-dependent optical, electronic, and catalytic properties on the nanoscale, which differ significantly from their bulk properties. In particular, palladium (Pd) nanoparticles have properties applicable to a wide range of applications in catalysis and electronics. However, predictable and controllable nanoparticle synthesis remains challenging because of harsh reaction conditions, artifacts from capping agents, and unpredictable growth. Biological supramolecules offer attractive templates for nanoparticle synthesis because of their precise structure and size. In this article, we demonstrate simple, controllable Pd nanoparticle synthesis on surfaceassembled viral nanotemplates. Specifically, we exploit precisely spaced thiol functionalities of genetically modified tobacco mosaic virus (TMV1cys) for facile surface assembly and readily controllable Pd nanoparticle synthesis via simple electroless deposition under mild aqueous conditions. Atomic force microscopy (AFM) studies clearly show tunable surface assembly and Pd nanoparticle formation preferentially on the TMV1cys templates. Grazing incidence small-angle X-ray scattering (GISAXS) further provided an accurate and statistically meaningful route by which to investigate the broad size ranges and uniformity of the Pd nanoparticles formed on TMV templates by simply tuning the reducer concentration. We believe that our viral-templated bottom-up approach to tunable Pd nanoparticle formation combined with the first in-depth characterization via GISAXS represents a major advancement toward exploiting viral templates for facile nanomaterials/device fabrication. We envision that our strategy can be extended to a wide range of applications, including uniform nanostructure and nanocatalyst synthesis.

Introduction Transition-metal-based nanoparticles offer enhanced, unique size-dependent catalytic, electronic, and optical properties arising on the nanoscale. An important example is palladium (Pd) nanoparticles, which play a vital role in catalysis for a wide range of applications in energy,1-5 chemical synthesis,6-8 and environmental cleanup.9,10 Although catalysis is clearly the dominant application of Pd nanoparticles, their electronic properties have also been exploited for nanoelectronic device fabrication11,12 and electrochemical sensing applications.13 Despite these numerous and significant uses of Pd nanoparticles, their controllable and reproducible synthesis remains challenging because of harsh reaction conditions, artifacts from surfactants and capping agents, and often unpredictable particle growth. *To whom correspondence should be addressed. Tel: (617) 627-2195. Fax: (617) 627-3991. E-mail: [email protected]. (1) Durand, J.; Teuma, E.; Gomez, M. Eur. J. Inorg. Chem. 2008, 23, 3577–3586. (2) Forzatti, P.; Groppi, G. Catal. Today 1999, 54, 165–180. (3) Kishore, S.; Nelson, J. A.; Adair, J. H.; Eklund, P. C. J. Alloys Compd. 2005, 389, 234–242. (4) Lampert, J. K.; Kazi, M. S.; Farrauto, R. J. Appl. Catal., B 1997, 14, 211– 223. (5) Pham-Huu, C.; Keller, N.; Ehret, G.; Charbonniere, L. J.; Ziessel, R.; Ledoux, M. J. J. Mol. Catal. A: Chem. 2001, 170, 155–163. (6) Navarro, O.; Kaur, H.; Mahjoor, P.; Nolan, S. P. J. Org. Chem. 2004, 69, 3173–3180. (7) Vittorio, F. Adv. Synth. Catal. 2004, 346, 1553–1582. (8) Xu, Y.-H.; Lu, J.; Loh, T.-P. J. Am. Chem. Soc. 2009, 131, 1372–1373. (9) Guibal, E.; Vincent, T. J. Environ. Manage. 2004, 71, 15–23. (10) Omole, M. A.; K’Owino, I. O.; Sadik, O. A. Appl. Catal., B 2007, 76, 158– 167. (11) Nguyen, K.; Monteverde, M.; Filoramo, A.; Goux-Capes, L.; Lyonnais, S.; Jegou, P.; Viel, P.; Goffman, M.; Bourgoin, J.-P. Adv. Mater. 2008, 20, 1099–1104. (12) Richter, J.; Seidel, R.; Kirsch, R.; Mertig, M.; Pompe, W. Adv. Mater. 2000, 12, 507–510. (13) Guo, S.; Dong, S. Trends Anal. Chem. 2009, 28, 96–109.

3670 DOI: 10.1021/la9031514

Biologically derived macro/supramolecules (e.g., proteins, DNA, and viruses) have attracted significant attention as templates for nanoparticle synthesis because of their precise size, structure, and shape. In addition, their readily manipulated surface properties offer enhanced functionality.14-21 Notably, spherical protein cages have been studied for the templated synthesis of metallic nanoparticles within the cage cores.22,23 The filamentous M13 bacteriophage has also been used extensively as a template for the formation of quantum dot or metallic nanowires.24-26 Another significant example of biological nanotemplates is the tobacco mosaic virus (TMV), a rigid, rod-shaped plant virus. TMV has been utilized as a nanotemplate in a wide (14) Balci, S.; Bittner, A. M.; Hahn, K.; Scheu, C.; Knez, M.; Kadri, A.; Wege, C.; Jeske, H.; Kern, K. Electrochim. Acta 2006, 51, 6251–6257. (15) Bigall, N. C.; Reitzig, M.; Naumann, W.; Simon, P.; Pee, K.-H. v.; Eychmuller, A. Angew. Chem., Int. Ed. 2008, 47, 7876–7879. (16) Deplanche, K.; Woods, R. D.; Mikheenko, I. P.; Sockett, R. E.; Macaskie, L. E. Biotechnol. Bioeng. 2008, 101, 873–880. (17) Mertig, M.; Ciacchi, L. C.; Seidel, R.; Pompe, W. Nano Lett. 2002, 2, 841– 844. (18) Gu, Q.; Cheng, C.; Haynie, D. T. Nanotechnology 2005, 8, 1358. (19) Seeman, N. C.; Belcher, A. M. Proc. Nat. Acad. Sci. U.S.A. 2002, 99(Suppl 2), 6451–6455. (20) Seidel, R.; Ciacchi, L. C.; Weigel, M.; Pompe, W.; Mertig, M. J. Phys. Chem. B 2004, 108, 10801–10811. (21) Zhou, J. C.; Gao, Y.; Martinez-Molares, A. A.; Jing, X.; Yan, D.; Lau, J.; Hamasaki, T.; Ozkan, C. S.; Ozkan, M.; Hu, E.; Dunn, B. Small 2008, 4, 1507– 1515. (22) Douglas, T.; Young, M. Nature 1998, 393, 152–155. (23) Douglas, T.; Young, M. Science 2006, 312, 873–875. (24) Nam, K. T.; Kim, D.-W.; Yoo, P. J.; Chiang, C.-Y.; Meethong, N.; Hammond, P. T.; Chiang, Y.-M.; Belcher, A. M. Science 2006, 312, 885–888. (25) Mao, C.; Solis, D. J.; Reiss, B. D.; Kottmann, S. T.; Sweeney, R. Y.; Hayhurst, A.; Georgiou, G.; Iverson, B.; Belcher, A. M. Science 2004, 303, 213– 217. (26) Lee, S.-W.; Mao, C.; Flynn, C. E.; Belcher, A. M. Science 2002, 296, 892– 895.

Published on Web 11/17/2009

Langmuir 2010, 26(5), 3670–3677

Manocchi et al.

Figure 1. (a) Chimera drawing of a portion (400 proteins) of a TMV1cys nanotemplate. Red dots represent cysteine groups genetically displayed on the outer surface of each coat protein. This drawing was produced using the UCSF Chimera package (http://www.cgl.ucsf.edu/chimera) from the Resource for Biocomputing, Visualization, and Informatics at the University of California, San Francisco.32 The TMV structure (PDB id 2tmv)33 was obtained from the Research Collaboratory for Structural Bioinformatics Protein Data Bank (RCSB PDB, http://www.pdb.org).34 (b) Schematic diagram depicting TMV1cys assembly onto gold surfaces, followed by reductive metallization of the Pd precursor by sodium hypophosphite to form Pd nanoparticles on the TMV1cys templates.

range of applications, such as conductive nanowires, battery electrodes, and digital memory devices.27-30 Despite these advances, studies have been limited to mere demonstration, and an in-depth examination of particle size and attempts at predictable size control are still lacking. Wild-type TMV (wtTMV) is a rigid, tubular plant virus with a length of 300 nm and a diameter of 18 nm and is composed of 2130 identical coat proteins helically assembled around a single genomic mRNA strand.31 Our approach exploits genetically modified TMV (TMV1cys), which possesses one cysteine residue displayed on the outer surface of each coat protein, as shown in the Chimera model drawing of TMV1cys in Figure 1a. (27) Royston, E.; Ghosh, A.; Kofinas, P.; Harris, M. T.; Culver, J. N. Langmuir 2008, 24, 906–912. (28) Royston, E.; Lee, S.-Y.; Culver, J. N.; Harris, M. T. J. Colloid Interface Sci. 2006, 298, 706–712. (29) Tseng, R. J.; Tsai, C.; Ma, L.; Ouyang, J.; Ozkan, C. S.; Yang, Y. Nat. Nano. 2006, 1, 72–77. (30) Wang, X.; Niu, Z.; Li, S.; Wang, Q.; Li, X. J. Biomed. Mater. Res. A 2008, 87A, 8–14. (31) Culver, J. N. Ann. Rev. Phytophathol. 2002, 40, 287–308. (32) Pettersen, E. F.; Goddard, T. D.; Huang, C. C.; Couch, G. S.; Greenblatt, D. M.; Meng, E. C.; Ferrin, T. E. J. Comput. Chem. 2004, 25, 1605–1612.

Langmuir 2010, 26(5), 3670–3677

Article

Previous studies reported the formation of metal nanoparticles on wt-TMV through electrostatic interactions.35-37 However, TMV1cys provides a dense and precisely spaced array of thiol functionalities,38 enabling enhanced and readily controllable Pd nanoparticle formation rising from the sulfur atom’s high affinity for Pd.39,40 In other studies, the high-capacity thiol functionalities of TMV1cys and TMV2cys (displaying two surface-accessible cysteine residues) have been exploited for the attachment of fluorescent markers and metal nanoparticles.38,41-44 Despite these earlier demonstrations of nanoparticle formation on genetically modified TMV’s, facile metal nanoparticle size control and any in-depth size examination are lacking in all studies. Therefore, there is a critical need for a simple route to control and examine the particle size and its uniformity in order for the viral-templated nanoparticle synthesis approaches to be readily employed in practical applications such as catalysis or nanodevice fabrication where such control of particle size and distribution is highly desired. In this article, we demonstrate readily controllable, tunable Pd nanoparticle formation on surface-assembled TMV1cys nanotemplates. As shown in the schematic diagram of Figure 1b, we first assembled TMV1cys onto clean gold surfaces by simply dipping a clean gold chip into TMV1cys solution. Following this surface assembly, Pd nanoparticles were synthesized on the TMV1cys surface via reduction of the Pd precursor with a mild reducing agent (sodium hypophosphite). Initial examination with atomic force microscopy (AFM) showed that the surface-assembly density of TMV1cys was controllable and tunable and that Pd nanoparticles were synthesized at high density preferentially on the TMV1cys surface. Further examination with grazing incidence small-angle X-ray scattering (GISAXS) indicated that the Pd nanoparticle size varied over a broad range and was readily controllable via a simple manipulation of the sodium hypophosphite reducer concentration. Finally, an investigation of other commonly enlisted reducing agents showed a lack of nanoparticle size control and batch-to-batch inconsistency. These results indicate that sodium hypophosphite is an effective reducer for the controllable synthesis of Pd nanoparticles on TMV1cys templates. This study shows, for the first time, an in-depth examination of Pd nanoparticle formation on TMV1cys nanotemplates. Furthermore, the results presented in this study indicate facile, broad Pd nanoparticle size manipulation under mild aqueous conditions. Importantly, analysis with GISAXS facilitated the measurement of a statistically meaningful, accurate Pd nanoparticle size and size distribution. We envision that our (33) Namba, K.; Pattanayek, R.; Stubbs, G. J. Mol. Biol. 1989, 208, 307–325. (34) Bhat, T. N.; Bourne, P.; Feng, Z.; Gilliland, G.; Jain, S.; Ravichandran, V.; Schneider, B.; Schneider, K.; Thanki, N.; Weissig, H.; Westbrook, J.; Berman, H. M. Nucleic Acids Res. 2001, 29, 214–218. (35) Knez, M.; Bittner, A. M.; Boes, F.; Wege, C.; Jeske, H.; Mai, E.; Kern, K. Nano Lett. 2003, 3, 1079–1082. (36) Knez, M; Sumser, M; Bittner, A. M.; Wege, C.; Jeske, H.; Martin, T. P.; Kern, K. Adv. Funct. Mater. 2004, 14, 116–124. (37) Bromley, K. M.; Patil, A. J.; Perriman, A. W.; Stubbs, G.; Mann, S. J. Mater. Chem. 2008, 18, 4796–4801. (38) Yi, H.; Nisar, S.; Lee, S.-Y.; Powers, M. A.; Bentley, W. E.; Payne, G. F.; Ghodssi, R.; Rubloff, G. W.; Harris, M. T.; Culver, J. N. Nano Lett. 2005, 5, 1931– 1936. (39) Okeya, S.; Kameda, K.; Kawashima, H.; Shimomura, H.; Nishioka, T.; Isobe, K. Chem. Lett. 1995, 24, 501–502. (40) Ebner, M.; Otto, H.; Werner, H. Angew. Chem., Int. Ed. 1985, 24, 518–519. (41) Yi, H.; Rubloff, G. W.; Culver, J. N. Langmuir 2007, 23, 2663–2667. (42) Tan, W. S.; Lewis, C. L.; Horelik, N. E.; Pregibon, D. C.; Doyle, P. S.; Yi, H. Langmuir 2008, 24, 12483–12488. (43) Lee, S.-Y.; Royston, E.; Culver, J. N.; Harris, M. T. Nanotechnology 2005, 16, S435–S441. (44) Lee, S.-Y.; Choi, J.; Royston, E.; Janes, D. B.; Culver, J. N.; Harris, M. T. J. Nanosci. Nanotechnol. 2006, 6, 974–981.

DOI: 10.1021/la9031514

3671

Article

Manocchi et al.

viral-templated approach to Pd nanoparticle synthesis can be expanded to a broad range of applications such as nanocatalysis or the fabrication of nanoscale structures where facile size manipulation is highly desired.

Materials and Methods Materials. Acetone (HPLC grade), isopropanol, and methanol were used as received (all from Fisher Scientific, Waltham, MA). Sodium tetrachloropalladate (II) (Na2PdCl4) was used as the Pd precursor for Pd nanoparticle formation (Sigma-Aldrich, St. Louis, MO). Precursor reduction was conducted using sodium hypophosphite (Sigma-Aldrich), sodium borohydride (Fisher Scientific), sodium cyanoborohydride (MP Biomedicals, Santa Ana, CA), or borane dimethylamine complex (DMAB) (97%, Sigma-Aldrich). Ethanol (200 proof, 99.5%) was also used (Fisher Scientific). TMV1cys Surface Assembly on Gold Chips. TMV1cys was generously provided by Dr. James Culver at the University of Maryland Biotechnology Institute, Center for Biosystems Research. Gold-coated silicon wafers (Platypus, Madison, WI) were cut into small chips (about 1 cm  2 cm) and then cleaned sequentially in acetone, isopropanol, and methanol each for 20 min, with thorough rinsing with deionized water between the steps. After this organic solvent cleaning, the chips were dried in a stream of ultrapure nitrogen gas and then etched with plasma (Ernest F. Fullam Inc., Clifton Park, NY). Immediately after being etched, the chips were incubated in 100 μg/mL TMV1cys in 0.01 M sodium phosphate buffer (pH 7) overnight at room temperature. Finally, after TMV1cys binding, the chips were thoroughly rinsed with deionized water, dried in a stream of ultrapure nitrogen gas, and stored at room temperature until AFM and GISAXS analyses were conducted. Palladium Nanoparticle Formation on TMV1cys Templates. Palladium nanoparticles were formed on the TMV1cys templates through the reductive metallization of the palladium precursor, Na2PdCl4, in aqueous sodium hypophosphite solutions. TMV1cys-bound gold chips (TMV chip) were incubated in 0.5 mM Na2PdCl4 and sodium hypophosphite solution for 20 min vertically in a microcentrifuge tube. The metallized TMV chips (TMV-Pd chips) were then thoroughly rinsed with deionized water for 5 min after metallization and dried in a stream of nitrogen gas. Pd nanoparticles were also formed on the TMV1cys surface through the reduction of palladium precursor using sodium borohydride, sodium cyanoborohydride, and dimethylamine borane (DMAB). This reduction was completed in the same manner as with sodium hypophosphite; however, in some cases of sodium borohydride and sodium cyanoborohydride, solutions were prepared in both water and 25% aqueous ethanol solutions. Atomic Force Microscopy. Atomic force microscopy (AFM) images were acquired using a Dimension 3100 series scanning probe microscope (SPM) (Veeco, Woodbury, NY). Images were analyzed using NanoScope software. All AFM measurements were conducted in tapping mode with TAP-Al-50 AFM tips (Budget Sensors, Sofia, Bulgaria).

Grazing Incidence Small-Angle X-ray Scattering (GISAXS). GISAXS measurements were conducted at the Advanced Photon Source (Argonne National Laboratory, Argonne, IL) BESSRC/XOR 12 ID-C beamline. Samples were mounted on a goniometer, and the beam irradiated the sample at an incident angle (Ri) of 0.1°, as shown in Figure 3a. The scattered X-rays were collected on a CCD detector (Rayonix; Mar165) with a sample-to-detector distance of approximately 2 m. Strong scattering and incident beam reflections in the Rf direction were blocked using a vertically mounted beamstop between the sample and detector. The beam energy was 8 keV. For all GISAXS measurements, the scattering pattern was analyzed in terms of the scattering vector, qxy. In this series of 3672 DOI: 10.1021/la9031514

experiments, it was assumed that Pd nanoparticles are spherical and dilute. The nanoparticles were assumed to be spherical on the basis of the isotropic scattering pattern in Figure 4c, which is characteristic of spherical particles. For dilute particle systems, it is assumed that the positions of individual nanoparticles are separated enough for scattering to be considered the sum of individual particle scattering. Importantly, the nanoparticles are assumed to be dilute because of the absence of structure factor features in the scattering curves. The relationship among the scattering intensity, I(qxy), for dilute spherical particles of radius R, particle volume ν, and density F is given in eq.49 Iðqxy Þ ¼ F2 ν2

9ðsinðqxy RÞ -qxy R cosðqxy RÞÞ2 ðqxy RÞ6

ð1Þ

For dilute particle systems where the interparticle distance peak in SAXS is not observed, the intensity data may be analyzed according to the Guinier law and the radius of gyration, Rg, of the particle calculated as shown in eq , which is derived from eq with a small qxy approximation. "

-ðqxy Rg Þ2 Iðqxy Þ ¼ F v exp 3 2 2

# ð2Þ

Therefore, a plot of ln[I(qxy)] versus qxy2 is linear with the slope equaling -Rg2/3. A complete analysis of the radius of the particle, R, requires a knowledge of the density distribution in the particle; however, for simplicity the density is considered to be constant and Rg is expressed as Rg=(3/5)1/2R.50 Only the linear portions of the Guinier plot corresponding to the region where qxy is less than the Guinier knee locations for each sample were used in the Guinier analysis calculations. The standard deviations of average particle size reported in Figure 5b were calculated by conducting a Guinier analysis in 10 portions of the linear region of the Guinier plot. Particle size distributions were calculated using Irena datafitting software, where the scattering curves (I(qxy) vs qxy) were fit using the maximum entropy method with a 10% error allowance.53 Data fitting was conducted in the region where the scattering from Pd nanoparticles appears.

Results and Discussion Simple Surface Assembly of TMV Templates. As shown in the AFM images in Figure 2, we first demonstrate the simple, tunable surface assembly of TMV1cys on gold substrates. For this, clean gold-coated silicon chips were incubated in a TMV1cys solution with varying concentrations overnight at room temperature, thoroughly rinsed with deionized water, dried in stream of ultrapure nitrogen gas, and probed via tapping-mode AFM. As clearly shown in Figure 2a-d, the surface assembly density of TMV1cys increases with increasing TMV1cys concentration, from sparse to dense assembly in the concentration ranges shown. Importantly, the binding of TMV1cys on gold surfaces was very (45) Markiewicz, P.; Goh, M. C. Langmuir 1994, 10, 5–7. (46) Glatter, O.; Kratky, O. Small-Angle X-ray Scattering; Academic Press: London, 1982. (47) Biswas, K.; Varghese, N.; Rao, C. N. R. Small 2008, 4, 649–655. (48) Lee, B.; Lo, C.-T.; Thiyagarajan, P.; Winans, R. E.; Li, X.; Niu, Z.; Wang, Q. Langmuir 2007, 23, 11157–11163. (49) Guinier, A; Fournet, G. Small Angle Scattering of X-rays; John Wiley and Sons: New York, 1955. (50) Roe, R.-J. Methods of X-ray and Neutron Scattering in Polymer Science; Oxford University Press: New York, 2000. (51) Britton, D. T.; Odo, E. A.; GoroGonfa, G.; Jonah, E. O.; Harting, M. J. Appl. Crystallogr. 2009, 42, 448–456. (52) Knez, M.; Sumser, M.; Bittner, A. M.; Wege, C.; Jeske, H.; Martin, T. P.; Kern, K. Adv. Funct. Mater. 2004, 14, 116–124. (53) Jemian, P. R.; Weertman, J. R.; Long, G. G.; Spal, R. D. Acta Metall. Mater. 1991, 39, 2477–2487.

Langmuir 2010, 26(5), 3670–3677

Manocchi et al.

Article

Figure 2. AFM images of surface-assembled TMV1cys at varying surface density controlled using TMV1cys incubation concentrations of (a) 10, (b) 30, (c) 50, and (d) 100 μg/mL.

consistent and stable throughout the assembly and probing procedures, including rigorous rinsing, drying with nitrogen gas, and physical contact with AFM tips, suggesting the utility of the genetically displayed thiol functionality for simple, stable surface assembly. Also, the higher surface density of TMV1cys can be consistently obtained at higher TMV1cys concentrations, whereas wt-TMV (no cysteine on the surface) shows inconsistent binding (results not shown). Notably, Royston et al. recently reported a similar trend in TMV1cys binding proportional to its concentration,27 but at high concentration (1 mg/mL), TMVs were reported to be standing up upon full coating with nickel. Meanwhile, our dense coverage of laterally assembled TMV1cys offers uniform templating conditions for Pd nanoparticle synthesis, provides clear AFM evidence, and allows surface characterization techniques such as GISAXS to be readily employed. As such, TMV-assembled chips (TMV chips) with a dense surface coverage (100 μg/mL) were employed for Pd metallization throughout the remainder of this study. In summary, these results show the simple, readily tunable surface assembly of TMV1cys templates on solid substrates. Palladium Nanoparticle Formation on TMV1cys Templates. As shown in the AFM images of Figure 3, we next demonstrate preferential palladium (Pd) nanoparticle formation on surface-assembled TMV1cys templates. For this, we incubated TMV chips in a sodium hypophosphite (NaPH2O2) solution containing Pd precursors (Na2PdCl4). These chips were then thoroughly rinsed, dried with ultrapure nitrogen gas, and probed via tapping-mode AFM as described above. First, Figure 3a shows a high-resolution AFM image of surface-assembled TMV1cys before exposure to Pd. Notably, TMV1cys appears to have a smooth surface, but the underlying gold surface shows mildly rough granular curvature. It is clear that all of the TMV1cys’s are assembled laterally on the gold chip, as evident by the visible TMV tubular shape and AFM height profiles (data not shown). Next, Figure 3b clearly shows Pd nanoparticles formed on the TMV1cys templates at high density upon exposure to the Pd precursor in reducing buffer. The presence of Pd was further confirmed using X-ray photoelectron Langmuir 2010, 26(5), 3670–3677

Figure 3. Pd nanoparticle formation on TMV1cys templates. (a) AFM image of surface-assembled TMV1cys, bound in 100 μg/mL TMV1cys. (b, c) AFM images of Pd nanoparticles formed on surface-assembled TMV1cys templates. Pd nanoparticles were formed using 0.5 mM Na2PdCl4 in 10 mM sodium hypophosphite for 20 min.

spectroscopy (XPS), as shown in Figure S1( Supporting Information 1). Importantly, the number of Pd nanoparticles on bare gold surfaces was consistently minimal in the presence of TMV1cys templates, clearly demonstrating the preferential templating of Pd nanoparticles on TMV1cys templates. In the absence of this preferred TMV1cys template, minimal Pd nanoparticle formation is observed on the bare gold surface (data not shown). Also, the DOI: 10.1021/la9031514

3673

Article

Figure 4. (a) Schematic diagram of the GISAXS setup. The sample is irradiated at an incident angle of Ri, and the scattered X-rays are measured as a function of 2θ and Rf. (b) Typical GISAXS scattering pattern of surface-assembled TMV1cys before metallization (TMV chips). (c) Typical GISAXS scattering pattern of Pd nanoparticles formed on TMV1cys templates (TMVPd chips). The red line indicates the location of the horizontal line cut.

TMV-Pd complex is very stable and remains strongly bound throughout extensive rinsing, drying, extended storage under ambient conditions, and physical contact during AFM, unlike previously reported results conducted in aqueous solutions.43 Furthermore, another AFM scan of a larger area shown in Figure 3c clearly shows the consistency of Pd nanoparticle formation on TMV1cys templates over a larger area. Combined, these AFM studies illustrate the preferential and high-density Pd nanoparticle formation on the surface-assembled TMV1cys templates. Grazing Incidence Small-Angle X-ray Scattering (GISAXS) Analysis of Palladium Nanoparticles on TMV Templates. Although providing a simple means to probe surface topology on the nanoscale, AFM is not suitable for an accurate examination of particle size or interparticle distance because of the inherent physical contact between the tip and the sample, resulting in a gross overestimation of particle size.45 Grazing incidence small-angle X-ray scattering (GISAXS) provides a simple nondestructive method by which to calculate nanoscale particle size and size distribution, film thickness, crystal structures, and so forth.46 GISAXS measurements of nanoparticle size have been shown to be very accurate when compared to other common analysis methods, such as high-resolution transmission electron microscopy (TEM).47 Importantly, GISAXS measurements result in statistically meaningful measurements of the average properties of the surface by utilizing a sizable sampling 3674 DOI: 10.1021/la9031514

Manocchi et al.

area with a large beam size (∼0.5  5 mm2), in contrast to the small sampling size of TEM. As shown in the schematic diagram of the typical GISAXS setup of Figure 4a, the sample is irradiated with the incident X-rays at a low angle of Ri (0.1°) and the scattering pattern is recorded at small angle ranges on a 2-D CCD detector. Figure 4b,c shows typical GISAXS scattering patterns of a TMV chip and a TMV-Pd chip, respectively. As shown in Figure 4b, surface-assembled TMV1cys exhibits strong scattering in the out-of-plane (at 2θ=0°) direction, characteristic of the rod shape of TMV and inter-TMV scattering.48 In stark contrast, the pattern shown in Figure 4c shows strong scattering rising from spherical Pd nanoparticles in a wider area on the CCD detector. The overall scattering intensity emerging from the TMV-Pd chip is significantly increased as compared to that of the TMV chip as a result of a substantial number of highly scattering Pd nanoparticles. The out-of-plane scattering from TMV1cys remains strong in Figure 4c, indicating that the Pd nanoparticles are formed right on the TMV1cys templates and the TMV templates maintain their overall tubular structure upon Pd nanoparticle formation. Importantly, the isotropic feature in the scattering pattern of Figure 4c clearly signals that the Pd particles that are formed are spherical (above Rf > Rf,c (∼0.5°)). Concurrently, the gold substrate has moderate surface roughness, as shown in the AFM image in Figure 3a, and generates significant scattering (at around 0 < Rf < 1 and -0.5 < 2θ < 0.5). Despite this potential interference, the TMV chips that we examined clearly show oscillating scattering patterns characteristic of TMV shown on other substrates.48 Furthermore, the clear difference between the scattering patterns in Figure 4b,c shows that this potential interference from background scattering from gold is minimal. In summary, these clear distinctions in the GISAXS scattering patterns between the TMV chip and the TMV-Pd chip confirm substantial Pd nanoparticle formation on TMV1cys templates and further suggest the utility of GISAXS in the investigation of parameters that govern Pd nanoparticle formation as described below. Pd Nanoparticle Size Control: Sodium Hypophosphite Concentration. Next, to interpret the GISAXS scattering patterns for particle size estimation, a horizontal line cut is made as shown by the red line in Figure 4c and the scattering intensity is plotted as a function of qxy (qxy= (4π sin θ)/λ), where θ is the scattering angle (Figure 4a) and λ is the X-ray wavelength, as shown in Figure 5a. The features of these scattering curves are then analyzed using the Guinier law,49 a widely accepted method46 by which to calculate particles sizes from X-ray scattering data that relates the scattering profile to the particle diameter (Materials and Methods), as shown in Figure 5b. On the basis of the preliminary examinations via AFM shown in Figure 4, we employed GISAXS to examine several critical reaction parameters that could affect the Pd nanoparticle size, such as Pd precursor concentration, reducer type and concentration, and metallization time. We observed clear distinctions between the scattering curves of samples prepared under various sodium hypophosphite concentrations, indicating that this parameter had the strongest potential as the particle-size-controlling parameter. For this, densely covered TMV chips were incubated in Pd precursor and sodium hypophosphite solutions for 20 min, thoroughly washed with deionized water and dried under ultrapure nitrogen gas, and then examined with AFM and GISAXS. Figure 5a shows scattering curves of TMV-Pd chips prepared in varying concentrations of sodium hypophosphite. First, the scattering curve of an unmetallized TMV chip (black line) at the bottom of Figure 5a shows several oscillations at high Langmuir 2010, 26(5), 3670–3677

Manocchi et al.

Figure 5. (a) GISAXS scattering curves (I(qxy) vs qxy) of samples prepared from 0.5 mM Na2PdCl4 and varying concentrations of sodium hypophosphite. (b) Average Pd nanoparticle size as a function of sodium hypophosphite concentration, calculated using Guinier analysis. The error bars represent standard deviations rising from Guinier analyses conducted in 10 different linear regions of the Guinier plot for each sample. (c) Normalized Pd nanoparticle volume distributions under various sodium hypophosphite concentration conditions.

qxy values, characteristic of TMV nanotubes48 whose frequency is observed only when the object is monodisperse in size, and is inversely propotional to the diameter of an object measured (in this case, TMV). Second, the 60 mM sodium hypophosphite curve (magenta, above the black TMV chip line) maintains some of these oscillations that are shifted to lower qxy locations with higher frequency. This indicates the presence of small particles (with a diameter of less than 4 nm) that evenly coat the TMV1cys surface while maintaining its monodipserse tubular shape with increased diameter. (See the simulation results in Supporting Information 2.) The presence of nanoparticles is signaled not Langmuir 2010, 26(5), 3670–3677

Article

only by the increase in the frequency of TMV scattering but also by Guinier-type scattering (or the form factor scattering that could be described by a Gaussian function centered at qxy=0).50 The width of the Guinier-type form factor scattering is inversely proportional to nanoparticle size and is parametrized with the qxy value of the Guinier knee, qxy*.51 The Gunier knee appears as a bump in a log-log plot of intensity versus qxy, where the slopes of intensity are different before and after qxy*. The size of the nanoparticle is approximated from qxy* using the Bragg equation such that D=2π/qxy*. For instance, the location of the Guinier knee for the 60 mM curve is at about 0.15 A˚-1, indicating that the diameter of the Pd nanoparticle is about 4 nm. Next, the TMV-Pd chips with lower sodium hypophosphite concentrations (5-50 mM) show Guinier knees shifting to lower qxy as the sodium hypophosphite concentration decreases, indicating an increase in the particle size. The TMV oscillation features are not visible for these samples, indicating that the scattering from the Pd particles is much stronger than that of the TMV and that the Pd nanoparticles are becoming polydisperse. In the case of the 60, 55, and 50 mM sodium hypophosphite samples, the locations of the Guinier knees shown in Figure 5a are approximately the same, indicating similar nanoparticle size, which is further confirmed via Guinier analysis49 (approximately 4 nm in diameter, Figure 5b). At 45 mM sodium hypophosphite (dark-blue line in Figure 5a), the Guinier knee is slightly shifted to lower qxy, indicating slightly larger particles (approximately 7 nm in diameter). Importantly, this 45 mM sodium hypophosphite sample exhibits a sharper Guinier knee peak because of scattering between closely packed nanoparticles, indicating that the nanoparticles are more closely packed than the lower-sodium-concentration samples. Notably, the particle diameter plot in Figure 5b shows a sharp increase in nanoparticle size from the 45 to 40 mM sodium hypophosphite samples, followed by a steady increase to a maximum diameter of approximately 16 nm for the 5 mM sodium hypophosphite sample. Importantly, this trend is consistent for several samples prepared on different occasions and batches illustrating the reproducible and robust nature of our TMV1cys assembly, Pd nanoparticle formation, and GISAXS-based particle examination strategy. Further reduction of the Pd nanoparticles with the dimethylamine borane complex (DMAB) did not change the nanoparticle size, as shown in the GISAXS scattering curves in Figure SI4 (Supporting Information 3), where five samples prepared under identical conditions were reduced in varied concentrations of DMAB. More importantly, the five identical samples that show identical scattering curves confirm the reproducibility and consistency of our nanoparticle-formation method and GISAXS measurements. Additionally, Pd nanoparticle size calculated via GISAXS correlated well with results from AFM and TEM, as shown in Figure SI5 (Supporting Information 4). Meanwhile, other parameters that were studied did not show significant potential as particle-size-controlling parameters. Specifically, the effect of incubation time was investigated, where it was observed that the growth of Pd nanoparticles occurred quickly and the particles reached their final size after 1 min of incubation (data not shown). Notably, the size range shown here (4-16 nm) is significantly broader than in previously reported studies on TMV templates under aqueous conditions, such as in Lee et al., where the Pd nanoparticles bound to TMV2cys were not uniform or wellcontrolled.43 Also, in this previous study, some of the large (∼70 nm) particles were not strongly bound to the TMV surface and could be easily washed away, in contrast to nanoparticles presented in this study that were strongly bound to TMV templates. Additionally, Knez et al. previously reported DOI: 10.1021/la9031514

3675

Article

significantly smaller Pd nanoparticles formed on wtTMV in aqueous solutions, yet specific particle sizes were not reported.52 Similarly, in other studies, nanoparticles have been formed on viral surfaces for a variety of purposes, including nanowire fabrication.27,29,35-37 However, none of these previous studies report an in-depth particle-size examination, and particle size and uniformity were estimated only from electron microscopy images. In contrast, the Pd nanoparticles that we report in this study show a significantly different and broad range of sizes (4-16 nm) on surface-assembled TMV1cys templates, further suggesting the uniqueness and utility of our TMV1cys surface assembly strategy for facile size control via the modulation of simple parameters such as the reducing agent concentration. Next, to understand further the size distribution of the Pd nanoparticles shown in Figure 5a,b, we employed Irena datafitting software53 and plotted normalized Pd volume distributions as shown in Figure 5c. The Irena software utilized for sizedistribution fittings uses the maximum entropy method and 10% error allowance (Materials and Methods). As shown in Figure 5c, the Pd nanoparticles are more uniform at small sizes (thus high sodium hypophosphite concentrations) and become polydisperse as the average size increases. A possible explanation could be that nanoparticle growth follows a mechanism similar to the Finke-Watzky two-step mechanism, where nanoparticle size is related to the initial number of nanoclusters formed in the initial nucleation step.54 We hypothesize that the number of initial clusters may be greater at high reducer concentration because of a lower Pd precursor-to-reducer ratio, which would result in the growth of smaller nanoparticles. In summary, the results in Figure 5 show a wide range of Pd nanoparticle sizes (4-16 nm) formed on surface-assembled TMV1cys templates in mild aqueous environments with a clear relationship between the size and the reducer concentration as a key size-controlling parameter. Particle Size and Distribution in Other Reducing Environments. Building on the observation that sodium hypophosphite concentration plays a significant role in Pd nanoparticle size control, we further investigated the effects of other commonly enlisted reducing agents: dimethylamine borane complex (DMAB), sodium cyanoborohydride, and sodium borohydride.55,56 For this, TMV chips were incubated in aqueous or 25% ethanol solutions containing the Pd precursor and reducing agents and were then examined with AFM and GISAXS. Their normalized volume distributions were calculated via Irena software in the same manner as above. Several representative size-distribution curves for each set of reducing environments that produced meaningful nanoparticles are plotted in Figure 6. First, in Figure 6a, an aqueous DMAB solution resulted in particles ranging from 5 to 15 nm average diameter, with some larger particles (15-20 nm) present. Importantly, all size distributions are broad throughout the concentration ranges studied, with no apparent concentration-size relationship. Second, Figure 6b shows that reduction with sodium cyanoborohydride under aqueous conditions results in narrow size distributions (4-6 nm diameter) for the two highest reducer concentrations tested (1 and 0.5 mM). Similarly, the lowest reducer concentration tested (0.01 mM) also resulted in small nanoparticles with a narrow size distribution. In contrast, at 0.1 mM reducer concentration, larger and more polydisperse nanoparticles were observed (8-20 nm). Third, Figure 6c shows nanoparticle sizes (4-6 nm) at the higher (54) Watzky, M. A.; Finney, E. E.; Finke, R. G. J. Am. Chem. Soc. 2008, 130, 11959–11969. (55) Coronado, E.; Ribera, A.; Garcia-Martinez, J.; Linares, N.; Liz-Marzan, L. M. J. Mater. Chem. 2008, 18, 5682–5688. (56) Rao, C. R. K.; Trivedi, D. C. Coord. Chem. Rev. 2005, 249, 613–631.

3676 DOI: 10.1021/la9031514

Manocchi et al.

Figure 6. Normalized Pd nanoparticle volume distributions for Pd nanoparticles formed through the reduction of 0.5 mM Na2PdCl4 using (a) DMAB, (b) sodium cyanoborohydride in water, (c) sodium cyanoborohydride in 25% ethanol, (d) sodium borohydride in water, and (e) sodium borohydride in 25% ethanol.

concentrations of sodium cyanoborohydride in 25% ethanol similar to those under the aqueous condition. Next, Figure 6d shows that aqueous sodium borohydride produces broad nanoparticle size ranges. The size distribution is quite narrow in the case of 1 and 0.01 mM sodium borohydride; however, the nanoparticle size becomes highly polydisperse at 0.1 mM, with no apparent size-concentration relationship. Finally, Figure 6e shows that sodium borohydride in 25% ethanol results in very narrow size distributions at 0.1 mM, with a 4 nm average nanoparticle diameter. At higher concentrations, Pd nanoparticles become Langmuir 2010, 26(5), 3670–3677

Manocchi et al.

Article

highly polydisperse and quite large (12-30 nm). Again, there is no clear size-concentration relationship. In general, the three reducing agents studied here often exhibited polydisperse nanoparticle sizes with no apparent concentration-size relationship. In stark contrast to the reproducible, broad (4-16 nm) size ranges for sodium hypophosphite, these reducing agents did not produce meaningful size ranges and there was high batch-to-bach inconsistency. Furthermore, other reducer concentrations studied (e.g., greater than 4 mM) did not show any meaningful Pd nanoparticle formation via either AFM or GISAXS (data not shown). Nevertheless, all of the reducing conditions studied here (cyanoborohydride, borohydride, and DMAB) show several important aspects consistent with the sodium hypophosphite case. First, both AFM images and GISAXS scattering patterns clearly show preferential Pd nanoparticle formation on TMV templates. Second, GISAXS results show that TMV templates retain their tubular structures upon metallization under all conditions examined. Finally, in all cases, Pd nanoparticles formed on TMV templates (and therefore the TMV-Pd complexes on the surface) retain their overall structures through extensive rinsing, drying with nitrogen gas, extended storage, tapping-mode AFM, and GISAXS, indicating the stability of the TMV-Pd complexes and TMV templates under various reducing and physical evironments. In summary, the five buffer conditions studied here (Figure 6) produced inconsistent, nonreproducible, polydisperse nanoparticles, whereas sodium hypophosphite (Figure 5) yielded easily controlled, consistent, broad size ranges.

Conclusions This report demonstrates the readily controllable formation of uniform Pd nanoparticles on the surface of genetically modified TMV1cys templates via the electroless deposition of the Pd precursor simply by modulating the concentration of a mild reducing agent, sodium hypophosphite. Both AFM- and GISAXS-based examinations clearly showed spherical Pd nanoparticle formation preferentially on TMV1cys templates at high density. In-depth Guinier analyses further indicated that the Pd nanoparticles were consistently and controllably formed over a broad size range (4-16 nm diameter). We observed that the Pd nanoparticles became smaller and more uniform at higher sodium hypophosphite concentration as opposed to larger and more polydisperse at low sodium hypophosphite concentrations. In contrast, the use of other commonly enlisted reducing agents

Langmuir 2010, 26(5), 3670–3677

resulted in polydisperse Pd nanoparticles with no size/reducer concentration relationship and inconsistent particle formation. Importantly, the GISAXS measurements correlate well to the TEM measurements of Pd nanoparticle size, attesting to the accuracy of the GISAXS analysis of nanoparticles. We believe that our viral-templated nanoparticle formation approach employing simple dipping procedures throughout, coupled with the first in-depth nanoparticle size examination via GISAXS, represents a significant advancement toward controlled metal nanoparticle formation. Further studies on particle size control and growth kinetics over a wider range of conditions (e.g., in a bulk aqueous environment) and in-depth examinations of the effect of the Pd nanoparticle size on catalytic activity are currently underway. We envision that our viral-templated approach to readily controllable Pd nanoparticle synthesis can be expanded to a wide range of applications, such as uniform Pd nanocatalyst synthesis57,58 or nanodevice fabrication.27,29 Acknowledgment. We gratefully acknowledge Dr. James Culver at the University of Maryland Biotechnology Institute, Center for Biosystems Research for his generous gift of TMV1cys. The work at Argonne National Laboratory was supported by the U.S. Department of Energy, BES-Chemical Sciences and BESScientific User Facilities under contract DE-AC-02-06CH11357 with UChicago Argonne, LLC, operator of Argonne National Laboratory. Also, we thank Dr. Sonke Seifert at APS Sector 12, Argonne National Laboratory for his valuable assistance with the GISAXS measurements. Finally, we acknowledge the Harvard University Center for Nanoscale Systems (CNS) for the use of TEM for nanoparticle analysis and specifically Dr. David Lange at Harvard CNS for his assistance with the XPS analysis. Partial funding for this work was provided by the Tufts Summer Scholars award (N.E.H.) and the Wittich Family Fund for Energy Sustainability. Supporting Information Available: XPS analysis of TMV and Pd-TMV chips, GISAXS simulation of Pd nanoparticles on the surface of TMV1cys, GISAXS analysis of Pd nanoparticles further reduced with DMAB, and a TEM analysis confirming the nanoparticle size measurements. This material is available free of charge via the Internet at http://pubs.acs.org. (57) Yang, C.; Manocchi, A. K.; Lee, B.; Yi, H. Appl. Catal., B 2009, in press. doi:10.1016/j.apcatb.2009.10.001 (58) Avery, K. N.; Schaak, J. E.; Schaak, R. E. Chem. Mater. 2009, 21, 2176– 2178.

DOI: 10.1021/la9031514

3677