This thesis describes production of and experiments with a Bose-Einstein condensate of approximately 2 × 10[superscript 4] [superscript 87]Rb atoms, trapped at the surface of an atom chip. In the first half of this thesis I describe the process of trapping and cooling the atomic vapour close to the surface of an atom chip. This process, which cools the vapour by over 9 orders of magnitude, involves a highly complex sequence of events which I implemented and optimised over the first two years of my PhD. In the early stages of this process, the atomic vapour is laser cooled and magneto-optically trapped. The vapour is then transferred to a highly elongated magnetic trap produced by high field gradients a few hundred microns from the surface of the atom chip. Here the vapour is evaporatively cooled to below the transition temperature where a Bose-Einstein condensate emerges. A simple existing analytic model of evaporative cooling is extended in this work to account for the shape of our highly elongated trap. Predictions of this model are presented here along with experimental observations with which it has good agreement. The second part of my thesis investigates some of the characteristics of the condensate, and dynamics of its low energy collective oscillations in the trap, based on experimental measurements taken in the final 18 months of my PhD. In particular, measurements taken of the centre of mass oscillations of the condensate along the long axis of the trap are presented. In the zero temperature limit the condensate is expected to behave as a perfect superfluid, and these low energy oscillations should go undamped. However, at finite temperature where not all atoms in the gas are condensed, damping is observed. In our experiment significant damping is found with an 1/e decay rate which varies between 2s[superscript -1] and 8s[superscript -1], depending on the fraction of non-condensed atoms in the gas. A finite temperature formalism is then used to describe the likely damping mechanism - Landau damping. We use a simple model of this formalism which estimates the temperature dependence of the damping rate γ(T), but find this gives a significant overestimation of the rates we measure. However, we argue that a straightforward adaptation to this model reduces the predicted damping rate significantly, and suggests a functional form of γ(T) that is in much better agreement with our experimental measurements.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:634120 |
| Date | January 2014 |
| Creators | Yuen, Benjamin |
| Contributors | Hinds, Edward |
| Publisher | Imperial College London |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://hdl.handle.net/10044/1/18833 |
Page generated in 0.0024 seconds