In an old vaudeville joke a pursued individual is told: "Head to the roundhouse, they'll never corner you there." And so it is with the so-called whispering-gallery disc laser. With no corners, the radiation goes round and round, eventually escaping evenly via quantum-mechanical tunneling. While the design makes for extremely small lasers, they suffer from low power and a lack of directionality. The circular laser, said Douglas Stone, chairman of Yale University's applied physics department, reflects too well, and the power that does escape is negligible.The intensity pattern of the bow-tie laser shows up clearly in false color representation. Now, as reported in the July issue of Photonics Spectra, researchers at Lucent Technologies' Bell Laboratories, Yale and the Max Planck Institute of Physics have deformed the cylindrical disc laser into a stadiumlike shape (see figure) and produced a multidirectional, high-power, low-threshold device. Rather than traveling in an infinite loop constrained by similar angles of incidence and total internal reflection, radiation now travels in a bow-tie pattern and emits in four controllable beams at 10 mW each in the mid-IR. The device evolved from work done at Yale on asymmetric resonator cavities and chaotic motion and from experiments conducted at Bell Labs, where Federico Capasso, head of Semiconductor Physics Research, and researcher Claire Gmachl were experimenting with disc lasers in an attempt to observe chaotic activity. Capasso was not aware of the Yale work when he appeared there for a meeting that, he said, later proved to be quite a "serendipitous" event. Meanwhile, Stone had suggested to Yale physicist Richard Chang that chaos theory might offer a solution to the circular laser problem, since classical optical ray tracing or even massive computer modeling was not providing one. According to Stone, chaos theory holds that a deformed circular resonator would allow the beam to escape by refraction at certain points. However, the bow-tie form was not anticipated by either group. Seeking two beamsNow both groups are exploring ways to move the device from the lab to the marketplace. Stone's group is working on changing the laser's shape to get to two beams rather than four, which is theoretically possible. Can they get to just one? Stone replies that this is easy to do in conventional lasers but not so in microlasers. Capasso and Gmachl's goals are slightly different. Beam steering via electronic rather than optical means is one goal. Getting the wavelength in the 1-µm telecommunication region is another. Bell Labs envisions a variety of roles for the device as an optical interconnect, citing free-space communications and a variety of short network applications including laser/detector arrays on a chip. Capasso said he believes that the impact of optics on optical computing is more likely to evolve through interconnects than through processors, a role that the new laser eventually might play. Technology aside, the development illustrates the beauty and effectiveness of interdisciplinary science. The project successfully combined semiconductor technology, optics, quantum mechanics, chaos theory, and academic and corporate research. In Capasso's words, "A strange but successful marriage."