Optical cooling using the dipole force

Abstract

The term ‘laser cooling’ is applied to the use of optical means to cool the motional energies of either atoms and molecules, or micromirrors. In the literature, these two strands are kept largely separate; both, however suffer from severe limitations. Laser cooling of atoms and molecules largely relies on the internal level structure of the species being cooled. As a result, only a small number of elements and a tiny number of molecules can be cooled this way. In the case of micromirrors, the problem lies in the engineering of micromirrors that need to satisfy a large number of constraints—these include a high mechanical Q-factor, high reflectivity and very good optical quality, weak coupling to the substrate, etc.—in order to enable efficient cooling. During the course of this thesis, I will draw these two sides of laser cooling closer together by means of a single, generically applicable scattering theory that can be used to explain the interaction between light and matter at a very general level. I use this ‘transfer matrix’ formalism to explore the use of the retarded dipole–dipole interaction as a means of both enhancing the efficiency of micromirror cooling systems and rendering the laser cooling of atoms and molecules less species selective. In particular, I identify the ‘external cavity cooling’ mechanism, whereby the use of an optical memory in the form of a resonant element (such as a cavity), outside which the object to be cooled sits, can potentially lead to the construction of fully integrated optomechanical systems and even two-dimensional arrays of translationally cold atoms, molecules or even micromirrors. The concept of an optical memory is a very general one, and I use it to link together mechanisms that would otherwise appear disparate, including the cavity cooling of atoms and cooling mechanisms based on the non-adiabatic following of atomic populations. A fully vectorial three-dimensional scattering theory including the effects of such a memory is also presented and used to explore several different experimentally-realisable cooling configurations.