Lasers Disturbed by Vacuum?

Posted on 2006-08-01. Permanent link: http://www.rp-photonics.com/spotlight_2006_08_01.html

Ref.: S. Reynaud and A. Heidmann, Opt. Commun. 71 (3-4), 209 (1989); encyclopedia articles on laser noise, quantum noise, shot noise, and others

Lasers are known to be rather sensitive devices. However, most people will be surprised to learn that even nothing – the vacuum! – introduces noise into lasers.

This is no joke, but rather a crucial element of a quantum-optical description of laser noise, and also in semiclassical theories as developed e.g. by researchers at the Ecole Normale Supérieure in Paris in the 1980s (see the reference above). In this picture, noise is introduced into the optical field circulating in a laser cavity not only by pump fluctuations and the like, but also results from the zero-point fluctuations of the electromagnetic field, which leak into the laser cavity. This happens particularly at the output coupler mirror, but also at any other location where optical losses occur. (There is no way to let radiation disappear without opening a path for quantum fluctuations coming in.) These "vacuum fluctuations" thus affect the intracavity field, and interestingly they again mix with it in the laser output: the output field is described as a superposition of the intracavity field partly leaking out of the cavity and the vacuum fluctuations coming from outside the cavity and being partly reflected at the output coupler. How these two fields are correlated with each other depends on the optical frequency. Somewhat aside from a cavity resonance there are phase shifts associated with off-resonant excitation. For frequency further away from the resonance (i.e., for large noise frequencies), one obtains little intensity coming from inside the cavity, and the output is dominated by the reflected vacuum fluctuations, which turn out to be responsible for the shot noise observed at these frequencies.

Skeptical readers may find it disturbing that one cannot test such claims by blocking the "vacuum input" and checking the difference in outcome: even a totally absorbing medium will remove any incident photons, but can not kill the vacuum fluctuations. However, the mentioned theories are experimentally well tested, surely accurate, and particularly their semiclassical forms are very convenient tools to calculate the noise properties of all sorts of devices – not only lasers, but also laser amplifiers as well as parametric devices such as optical parametric oscillators and parametric amplifiers, and frequency doublers.