Excess quantum noise fluctuations in unstable-resonator lasers

Experiments completed during the past year confirm the existence of a sizable excess quantum noise factor in lasers using unstable optical resonators or, more generally, resonators with nonorthogonal oscillation modes. Schawlow and Townes predicted in 1958, before the first laser was built, that even an ideal laser should exhibit a finite linewidth resulting from spontaneous emission by the laser atoms.1 Lamb and others, using standard techniques of cavity mode expansion and second quantization, then showed that the spontaneous emission in any laser should have a magnitude equal to the downward stimulated emission due to one additional quantum of signal energy acting on the inverted laser medium.2 This "one extra noise photon" approach to quantum noise has since become conventional wisdom in the field. Petermann noted in 1979, however, that spontaneous emission into the oscillating mode of a gain-guided semiconductor laser could be significantly larger than one photon per mode, leading to potentially measurable consequences for such lasers.3 In 1985, Haus and Kawakami showed that there would be partial coherence between the excess noise emission into different cavity modes, thus avoiding apparent conflicts between this excess emission and basic concepts of thermal equilibrium.4 One of us then showed in 1989 that this excess quantum noise was actually associated with the nonhermitian or biorthogonal character of the modes in certain laser structures rather than with gain guiding per se, so that large excess quantum noise effects should be observed in unstable resonator lasers in particular.5