A class of porous metallic nanostructures
Nature, 401, 548 (1999)
Materials and methods. The materials and chemicals used were as follows: sulfate, surfactant-free latex microspheres (IDC, USA); 25 mm diameter polycarbonate membranes (Poretics Inc., USA); inorganic oxidizer Nochromix (Godax, USA). The gold suspensions used were prepared by the established procedure (Slot, J. W. & Geuze, H. J., Eur. J. Cell Biol. 38, 87, 1985; Slot, J. W. & Geuze, H. J., J. Cell Biol. 90, 533, 1981) at an Au concentration of 0.01 wt%. The mean size of the particles obtained was determined by dynamic light scattering and TEM. The observations were carried out on an Olympus BH2 optical microscope, a JEOL JXA-840 scanning electron microscope at 30 kV and a Philips EM400T transmission electron microscope at 100 kV. The conductivity of the flakes was measured by touching the flakes, resting on a gold surface, with a multimeter probe.
Procedures. The colloidal crystal templates are assembled from monodisperse, negatively charged polystyrene latex microspheres ranging in diameter from 300 to 1000 nm. The latex particles used are diluted into 10 ml of deionized water and then deposited into densely packed layers by filtering for about 2 hours at 20 kPa. To improve the quality of the arrays, 0.15 wt% of a nonionic surfactant, Tween 20, is added to the latex suspension. We found polycarbonate membranes of 50 nm pore size to be optimal in retaining both the latex and the gold particles, but still large enough to allow a reasonably high flux of water. The dilute microspheres slowly accumulate on the membrane surface into closely packed 3D ordered layers of thickness of about 35 microns, which typically demonstrate a few reflections of a single bright color when viewed at different angles.
The next step in the process is the deposition of colloidal gold particles in the cavities of the latex crystals. The size of the gold particles used, 15-25 nm, was chosen to be much smaller than the openings between the latex microspheres in the crystal. Although the nominal pore size of the membrane is larger than the typical gold particle diameter, the particles are retained, possibly due to their charge exclusion from the pores or plugging of the pores by larger particles. The deposition of the gold particles on the filter is carried out at a pressure of about 95 kPa with 200-300 ml of gold suspension filtered through the deposited latex layer in 24-72 hours. The growth of the gold structure within the latex layer can be seen by the change in the color of the deposited film from white to gray to black. After the desired amount of gold is deposited, the filter with the layer of colloidal material is detached and dried in air for 2 hours at 50oC.
Two alternative procedures are used to remove the latex beads from within the composite, and they yield porous metal with slightly different properties. The first procedure is calcination in which the temperature is ramped at 0.2 oC/min to 300 oC, where it is maintained for 30 minutes before cooling. The final products are flakes with the distinct yellow color of metallic gold. As an alternative to calcination, the latex templates can be removed by chemical oxidation or solvent dissolution at room temperature. In the oxidation route the latex/gold composites are soaked in a mixture of concentrated sulfuric acid and inorganic oxidizer. The flakes initially float on the surface of the mixture, but sink to the bottom within a few minutes and are left there for 24 hours before being extensively washed with deionized water and dried. In the dissolution scheme the flakes are multiply washed with trichloromethane for 2-3 hours and air-dried. The resulting materials appear as brownish flakes with a yellow metallic appearance in strong reflected light. Brightly colored reflections can again be observed in low-magnification microscopy at appropriate angles of incident illumination.
Size control of the pores via the latex templates. Shown below are micrographs from samples prepared from latex beads of sizes 270 nm and 1000 nm. The lattice dimensions of the pore arrays embedded in the metal correspond to those of the original latex crystals, indicating negligible shrinkage of the metallic structure during template removal.
Wall thickness. The thickness of the walls around the pores varies depending on the gold size and deposition procedure and the type of treatment to remove the latex. Thicker bridges are obtained with smaller gold particles, more complete filling and chemical oxidation. A very delicate, 3D lace-like structure is obtained in some of the samples prepared by dissolution. One interesting structure found in the calcined samples after partial melting of the bridges is a mutation of the lattice into an array of droplets connected by thin nanosized "wires".
Crystalline modifications. Shown below is a porous gold domain with cubic packing obtained from 630 nm latex templates.
Relevance to photonic materials. The lace-like structure shown in Figure 2c of the paper is an almost exact replica of the 3D wire mesh photonic crystals described and studied by Sievenpiper et al. (Phys. Rev. Lett. 76, 2480, 1996), but scaled down by a factor of 20000 to the sub-micrometer region. This wire mesh photonic structure has been demonstrated to have an interesting combination of forbidden bands in the gigahertz region. As the photonic properties scale with array size, miniaturizing the structure should shift these properties to the IR/visible region, which is of significant interest for optoelectronic devices. Our flakes demonstrate a combination of optical features that appear to make these nanostructures unique among materials with periodicity on a colloidal scale. Each ordered domain on the surface shows three different principal modes of reflection at different orientations: (1) no special coloring or reflection; (2) bright pure colors, originating from diffraction at specific angles relative to the incident beam; and (3) strong polychromatic reflection at angles close to the specular angle. The close–packed latex arrays such as those used as templates display only the first two modes of reflection and lack the strong polychromatic reflection. The surface of non-porous gold exhibits strong reflection, but no colors at any angle. It remains to be seen to what extent the strong, non-wavelength dependent reflection of the porous gold can be attributed to the photonic properties of the nanostructured metal mesh, or to trivial reflection from the metal surface.
It has been argued that because of large losses in 3D metallic nanostructures they likely have photonic properties only in reflectance mode, and not in the practically important transmission mode. This is true for thick porous metals, but not so for very thin porous layers. Indeed the series of papers by Ebbesen et al. (e.g. Nature, 391, 667, 1998) have demonstrated the very interesting transmission properties of thin holey arrays, arising from the surface plasmons in the metallic layers. Since our Nature paper was submitted for publication, we have been able to modify the method to the formation of structured metallic layers only 1-3 pores thick deposited on a glass substrate. These layers are both semitransparent and show colors by diffraction. Informal comments, inquiries and discussion via personal communication are welcome.
Related developments. Two papers on the formation of microstructured porous metals by alternative procedures appeared while this material was in press: H. Yan et al., Adv. Mater., 11, 1003 (1999) and P. Jiang et al., J. Am. Chem. Soc., 121, 7957 (1999).
Acknowledgments. We are grateful to Prof. G. H. Watson for useful discussions.
Correspondence to Orlin Velev: email@example.com
© Copyright 1999, Velev, Tessier, Lenhoff & Kaler
Back to colloidal assembly index
Colloidal crystallization in 2D and 3D: The quest for novel materials
Thank you for visiting this page. Please note that this material is copyrighted.