Frequency stability and the optical linewidth of semiconductor laser are critical parameters, affecting a wide range of applications such as coherent and optical communication system performance, wavelength division multiplexing (VMM) based photonic switching and communication systems, room-temperature spectral-hole-burning optical memory, interferometric sensing, path length measurement, and high-resolution spectroscopy. For many applications the operating frequency of a single longitudinal mode (SLM) laser such as the distributed feedback laser (DFB),.or the distributed Bragg reflector laser (DBR), is not stable enough. It is susceptible to thermal fluctuations, and intensity as well as spatial variations of carrier concentration cause chirping. In addition, a typical DFB laser, operating at 1.5 microns has a linewidth of the order of 3- 1 0 MHz, while most coherent optical communication systems require sub MHz linewidth. Several techniques have been used in the past to lock the operating frequency and reduce the optical linewidth, among which are electronic or optical feedback. Reduction of the optical linewidth to the kHz range has successfully been demonstrated using bulk external cavity semiconductor laser. Another approach is to couple the laser cavity to a properly detuned external high finesse resonator. Frequency locking with dramatic linewidth reduction by a factor of ~I 000, using coupling of a semiconductor laser to high-finesse Fabry-Perot etalon, has been demonstrated. However, in spite of impressive reduction of the laser noise, most of these techniques are cumbersome, require a bulky, laboratory-Lype optical set-up, and are sensitive to cantankerous parameters, such as the phase of the reflected signal, which is strongly affected by niicrophonic effects. In addition, due to the length of the cavity required for the set-ups, the longitudinal mode spacing is small and SLM operation is difficult to realize. As a result, these techniques were primarily used only for experimental set-ups.
In this project, we propose to develop and carry out theoretical and experimental studies of a novel scheme for realizing frequency locking and linewidth reduction of semiconductor lasers in a small scale geometry. The approach involves the use of a new type of optical resonator system, based on a very high Q microsphere cavity. It has been known for some time that micrometer sized dielectric spheres act as high-Q resonators, with electromagnetic energy stored in form of spherical cavity modes confined around the sphere, near its surface. Employment of such modes in an axially-symmetric dielectric body results in sharp resonance peaks named "morphology dependent resonances" or MDR'S. The Q-factor of these resonances is projected to approach 10 8 at T = 77=degrees K, and exceed 10 9=at T=4 K. For comparison, a Fabry-Perot etalon type resonator, comprised o= f two mirrors of 98% reflectivity separated 10 cm apart would result in a Q factor of 3X10 7.
These resonance peaks of the micro-spheres occur at specific values of the sphere size parameter X, where X = 2*Pi*a/lambda; a is the radius of the sphere, and lambda is the optical wavelength. Using this phenomenon, cavity effects such as optical bistability, and lasing have been reported. In our case, the micro-sphere can be used as a detuned loading for a semiconductor laser, causing the oscillating field to lock on to one of the MDR's and quench its linewidth at the same time. This approach may achieve extremely narrow linewidth in a system that is stable, cost effective, and not larger than the laser itself.
In addition, the ability to lock into and measure MDR's of single stationary micro-spheres opens up the prospect of performing extremely sensitive adsorption and reaction measurements between species bonded to the micro-sphere surface and reagents in a surrounding solution. With modest Q's of 10 6, a layer having a subatomic thickness (~O.1 Angstrom) may be detected on a micro-sphere with r=3D 10 microns. Since this thicknes= s is considerably less than the molecular size of a typical antigen molecule, the possibility of observing small fractions of a monolayer is reasonable and thus paves the way to new immunological testing techniques.
