pubs.acs.org/Langmuir © 2010 American Chemical Society
Nanorods versus Nanospheres: A Bifurcation Mechanism Revealed by Principal Component TEM Analysis Fabien Hubert,† Fabienne Testard,† Giancarlo Rizza,‡ and Olivier Spalla*,† †
IRAMIS-SIS2M-LIONS, CEA Saclay, 9191 Gif-sur-Yvette, France, and ‡IRAMIS-LSI Ecole Polytechnique, 91128 Palaiseau, France Received February 27, 2010. Revised Manuscript Received April 10, 2010
A quantitative analysis of object populations obtained by TEM images is performed for the classical scheme of aqueous seedless synthesis of nanorods. Using an effective way to represent nanoparticle size distributions, we unravel that spheres, usually considered to be a side-product, are in fact coming from a competing route during nanorod formation. The differentiation between spheres and rods appears above a critical size of 5 nm and is due to different growth rates between faces. The initial repartition of faces on nuclei or on the nanoparticles at the critical size can be the source for the final differentiation between globules and rods. The efficiency of the selection is strongly influenced by the production of the initial seeds and, in particular, by the amount of borohydride added in the present scheme.
Introduction The synthesis of anisotropic particles and, more particularly, nanorods is a topic of active research because of the importance of these materials’applications in electronic devices (conduction properties) and biological sensors (because of the versatility of their longitudinal plasmon).1,2 In the case of gold nanorods, many routes of synthesis have been proposed over the years1-6 with the continuous goal of improving the reaction yield, shape selection (minimizing the so-called side products), and anisotropic ratio. TEM is generally used to characterize the final nanoparticles and to improve the merits of the synthesis scheme. A common way is to obtain TEM images of the final particles and then to present some selected images with the most typical nanorod content. Even when the synthesis produces different shapes, there is nearly no statistical analysis of the TEM images. Generally, only the nanoparticles of interest are described after the appropriate extraction/ washing procedure. The statistical analysis of the different final shape of nanoparticles can bring about interesting features in the mechanism. In the seed-mediated synthesis of gold nanorods, this approach has shown that a high yield of nanorods is obtained only from the smallest seeds.7 In the ultraviolet-driven photochemical synthesis of gold nanorods in aqueous surfactants, the description of the size and shape distribution of all of the final nanoparticles clearly illustrated the key role of silver in controlling the anisotropy.8 Jana et al.9 have also shown from TEM measurement on seed solutions that the production of a high yield of nanorods requires either a nonseeding growth method or the use of the smallest seeds (∼1.5 nm) in a seeding growth method. *Corresponding author. E-mail:
[email protected].
(1) Perez-Juste, J.; et al. Coord. Chem. Rev. 2005, 249, 1870–1901. (2) Sharma, V. K.; Srinivasarao, M.; et al. Mater. Sci. Eng. R 2009, 65, 1–38. (3) Jana, N. R.; Gearheart, L.; Murphy, C. J. Langmuir 2001, 17, 6782–6786. (4) Jana, N. R.; Gearheart, L.; Murphy, C. J. J. Phys. Chem. B 2001, 105, 4065– 4067. (5) Nikoobakht, B.; El-Sayed, M. A. Chem. Mater. 2003, 15, 1957–1962. (6) Zijlstra, P.; et al. J. Phys. Chem. B 2006, 110, 19315–19318. (7) Gole, A.; Murphy, C. J. Chem. Mater. 2004, 16, 3633–3640. (8) Placido, T.; et al. Chem. Mater. 2009, 21, 4192–4202. (9) Jana, N. R. Small 2005, 1, 875–882. (10) Lofton, C.; Sigmund, W. Adv. Funct. Mater. 2005, 15, 1197–1208. (11) Elechiguerra, J. L.; Reyes-Gasgab, J.; Yacama, M. J. J. Mater. Chem. 2006, 16, 3906–3919.
Langmuir 2010, 26(10), 6887–6891
HRTEM also enables us to determine that the size and shape of the initial seeds strongly influence the final structure of the nanoparticles.10-13 As an example, the presence of twin planes in the seeds strongly controls the final structure. Finally, the symmetry breaking from isotropic seeds toward anisotropic nanoparticles is still debated,14 leaving the origin of the anisotropic growth incompletely understood. Regarding the distribution of the dimensions of nanorods, decoupled Gaussian distributions in length and diameter are usually assumed.15 Such distributions have been used to quantify in situ SAXS patterns of nanorods in formation (with SAXS being a powerful technique for analyzing a large number of particles).15 We will see that in our study the distribution is different. In the present work, using a statistical counting of the objects from TEM images of the final state and a simple representation, we unravel the mechanism of growth and the formation of gold nanorods and update the status of what people use to call side products. Furthermore, at a more experimental level it represents a very effective way to follow the influence of the chemical conditions and additives.
Results and Discussion Synthesis and TEM Grid Preparation. Gold nanorods were prepared via a seedless synthesis inspired by Jana.4 A 10 mL solution containing 4 mM aqueous tetrachloroauric acid, 0.2 M CTAB, and 1 mM silver nitrate was assembled. Next, 0.12 mL of 0.4 M ascorbic acid was added under stirring. Finally, 32 μL of an ice-cold aqueous 1.6 mM sodium borohydride solution was added (corresponding to a final concentration of 5 μM). The reaction, followed using UV-visible spectroscopy, was complete in less than 20 min. This synthesis is the reference and is used for comparison in the following text to describe the influence of the chemical conditions and additives. The solutions contain at the end of the synthesis a large concentration of surfactants that have to be eliminated before TEM observation. To do so, the nanoparticles are separated by (12) (13) (14) (15)
Johnson, C. J.; et al. J. Mater. Chem. 2002, 12, 1765–1770. Orendorff, C. J.; Murphy, C. J. J. Phys. Chem. B 2006, 110, 3990–3994. Viswanath, B.; et al. J. Phys. Chem. C 2009, 113, 16866–16883. Henkel, A.; et al. J. Phys. Chem. C 2009, 113, 10390–10394.
Published on Web 04/20/2010
DOI: 10.1021/la100843k
6887
Letter
three cycles of centrifugation to extract the majority of them (9000 rpm for 1 h) and redispersed in pure water. The remaining supernatant was nearly colorless and translucent at the end of the procedure, indicating that nearly all nanoparticles were extracted. Finally, one unique small drop of the surfactant-free suspension of nanoparticles was deposited onto the TEM grid (carbon film grid of 200 mesh Du (50) from Agar Scientific), which was further dried in ambient air. Principles of MET Analysis. The TEM measurements were performed using a Philips CM30 300 keV transmission electron microscope (TEM). TEM micrographs were processed with a slow-scan CCD camera. Several images of each sample were made in order to analyze more than 500 particles per each tested chemical set of conditions. The length and diameter were measured for every particle using a semiautomated numerical treatment based on Image J software. First, the average background defined outside the particles was subtracted from all of these pixels. Then an ad hoc threshold was applied to the gray scale, and every pixel above that threshold was set equal to 1 and every pixel below the threshold was set equal to zero. Using this resulting binary image, the algorithm “Measure_ROI.class” by Dougherty (http://www. optinav.com/Measure-Roi.htm) was applied to measure the size and area of these objects and analyze their shape. The algorithm first determines the maximum distance between two points in the particles; then it measures the maximum size in the perpendicular direction. For nanorods having a rounded end shape in the image (Figure 1), the procedure determined the length (L) and diameter (D) correctly. This would not have been the case for an apparent rectangular shape. In other words, this process determines the two first principal components of the objects. The results are simply reported in 2D space as a set of coordinates (L, D), as shown in Figure 1 for the reference synthesis. One remarkable point is that two branches exist in this representation, one for the nanorods and one for the nanospheres. Very few nanoparticles are located between these two branches. The rods are distributed along a line that is not crossing the origin; this line crosses the sphere’s line at a point called BP (bifurcation point) with coordinates of L = 5 nm and D = 5 nm. Below these sizes, only spherical particles are present. Nuclei appear at a much smaller size (typically 1 nm) and then grow either by adsorbing new monomers or by coagulating with other nuclei. However, whatever the mechanism of growth, the present observation means that there is a critical size beyond which the original globular nuclei split into two populations. For nanorods, this corresponds to a size of bifurcation from spherical symmetry. Using this two-component analysis, it clearly appears that small nanoparticles (
