Latex molds spherical-micromirror arrays

Arrays of spherical micromirrors created by researchers at the University of Southampton (Southampton, England) may one day result in arrayed devices containing lasers or optically active material-filled microcavities as the elements.¹ The researchers have been optically characterizing the arrays, which make use of gold (Au), platinum (Pt), or other materials as the reflective surface.

To fabricate the arrays, latex microspheres in a colloidal solution are allowed to settle on a gold film. Under the right conditions, the spheres settle into a close-packed monolayer. Next, a metal is electroplated onto the spheres, settling into the spaces between them. Dissolving the microspheres with a solvent leaves an array of concave spherical mirrors.

These cavities can be made down to 200 nm in diameter; the smallest diffract light substantially as well as reflect it, and are thus difficult to examine optically. Larger mirrors of a few microns in diameter can be examined under an optical microscope. The researchers compared cavities fabricated from Au and Pt using a scanning-electron microscope and an atomic-force microscope (AFM) as well as optically. The AFM showed that, while the Au cavities were smooth across the bottom and up the sides, the Pt cavities had rough sides. Optical micrographs confirmed this result (see figure).

Click here to enlarge image

A spherical-micromirror array with a 10-µm period was fabricated from platinum (Pt; left); a similar array with a 5-µm period was made from gold (Au; right). In this case, the mirrors are almost, but not quite, hemispherical. These micrographs were taken at 200X with a 0.9-numerical-aperture (NA) microscope objective. The Pt mirrors have rough sides as well as rough spaces between the mirrors and therefore appear smaller than they are, even at 0.9 NA; in contrast, the Au mirrors have smooth sides and appear more closely packed. The light ring at the outer edges of the Au mirrors are light that is doubly reflected within the mirrors, while similar-looking rings around the Pt mirrors are actually light singly reflected from rough plateaus within the mirrors (the plateaus were also seen in atomic-force-microscope measurements).

Why use micromirrors of Pt or other materials when Au versions appear to be so well shaped? "Gold is much better optically, but there are good reasons to combine different metals in multilayered structures," says Jeremy Baumberg, one of the researchers. "For instance, Pt has different chemical attachment properties (for optical biosensors), cobalt has magnetic properties, and so on." Baumberg notes that many materials, including silver, cobalt, copper, polymers (for example, polyanaline), cadmium telluride, cuprous oxide, zinc oxide, and other semiconductors can be formed into spherical-micromirror arrays.

The mirrors are large enough to be modeled by geometric ray tracing. When polarized light is reflected from an Au micromirror array and sent back through a crossed polarizer (which blocks all light that has not undergone a change in polarization), some light gets through; this transmitted light has a complicated polarization state that can only be obtained from multiple cavity bounces—thus confirming the ray-tracing model.

Arrays covering an area of several square centimeters have been created, says Baumberg. "The technique is easy to scale up even further," he notes. "We could probably do a square meter without too much problem." Filling the cavities with a laser gain material and placing a flat mirror on top of a microcavity array (at a distance closer than the radius of curvature) could result in an array of lasers.

The researchers have also created three-dimensional photonic crystals using the same technique, but with periods of less than a micron.^{2, 3} While these show grating effects, "they are much more complicated than we thought, since there is an interplay between plasmons and photonic crystal/diffraction effects," says Baumberg.