Semiconductor lasers are uniquely suited to a variety of applications because they have extremely high efficiency compared to other laser media and can be designed to produce a wide range of wavelengths. One of the major disadvantages of semiconductor lasers that has limited their application is their inherently high divergence.
Theoretically, laser output propagates in precisely the same direction over the entire cross section of the beam and remains perfectly parallel until it is scattered. However, the realities of defective materials and diffraction causes the beam to spread out as it propagates; the measure of the angle over which the beam spreads is termed its divergence. In this week's Nature Photonics, researchers from Harvard and Hamamatsu photonics demonstrate a beam-shaping technique that results in a 25 times less divergence from an edge-emitting quantum cascade laser (QCL).
In this work, the output of the quantum cascade laser is coupled to the surface plasmon modes of a metal. The best way to imagine this is that the light output of the laser gets converted into electron oscillations on the surface of the metal that behave a lot like the original laser light. As the plasmons move down the metal surface, they get scattered by patterned grooves in the metal surface. By adjusting the spacing, thickness, and depth of the grooves, the plasmon scattering can be controlled so that all of the energy scattered out of the beam undergoes destructive interference, resulting in a highly collimated beam. Photonic crystals guide light by the same principle.
A good analogy of this process is diffraction in reverse. Diffraction occurs when a wave moving in one direction is scattered by an object and produces a diverging set of waves not unlike the ripples from a stone thrown into water. In almost all scientific applications, diffraction occurs when a collimated or coherent electromagnetic beam interacts with a small scattering object—this effect is put to good use in X-ray diffraction and demonstrations of the wave nature of light like Young's famous two slit experiment. However the time reversal symmetry inherent to all physical laws tells us that this process can just as easily happen in reverse; we just do not typically have the right conditions to diffract diverging radiation into a collimated beam.
While this technique is an excellent demonstration of the power of plasmonics and metamaterials, it is severely limited by the fact that it only works in one dimension and only with edge-emitting semiconductor lasers. Three dimensional metallic grids offer hope of greater versatility, but they may be prohibitively complicated to produce.
Nature Photonics DOI: 10.1038/nphoton.2008.152 10.1038/nphoton.2008.152