The Effect of Trap Geometry on Phase Separation in a Polarized Fermi Gas*

We have observed phase separation in a polarized $^6$Li atomic Fermi gas in an elongated (aspect ratio of 30) single-beam optical trap.\footnote{G. B. Partridge {\it et al.}, {\it Science} {\bf 311}, 503 (2006); {\it PRL} {\bf 97}, 190407 (2006).} The phase-separated phase consists of a paired superfluid core surrounded on each end of the trap by a completely polarized normal gas. The observed density distributions represent a violation of the local density approximation, and have been explained by surface tension. In addition, we find that the superfluid core survives until the system is nearly completely polarized, with $P\geq 0.95$. These results contrast with a similar experiment, done at MIT, that observes the superfluid core to survive up to $P\sim 0.74$, in agreement with the Clogston limit. In this case, no surface tension is observed.\footnote{C. H. Schunck {\it et al.}, {\it Science} {\bf 316}, 867 (2007).} Though the the explanation of these discrepancies is not yet clear, one major difference is trap geometry, since the MIT trap has an aspect ratio 6 times smaller than ours. We have implemented a crossed beam trap with an aspect ratio $\sim$3 in order to study the effects of trap geometry. Our latest experimental results will be presented.