10.7 Properties of Bose-condensed gases 241Fig. 10.14The interference fringes ob-
served when two independent Bose con-
densates are released from nearby po-
tential wells and the clouds of atoms
expand and overlap. This experiment
was carried out with sodium atoms by
the team led by Wolfgang Ketterle at
MIT (Andrewset al.1997). Copyright
1997 by the American Association for
the Advancement of Science.and the usual double-slit experiments. In the MIT experiment there was
no fixed relation between the phases of the two condensates and before
the experiment was carried out it was hotly debated whether interference
would be observed. Clear interference fringes were observed each time
the experiment was carried out. However, the position of these fringes
depended on the difference between the phases of the condensates in
that particular run—the bright and dark fringes appeared at a different
place each time the experiment was repeated; thus the fringe pattern
would ‘wash out’ if averaged over many runs. The observation of the
interference of two condensates relies on the ability to see interference
fringes in a single shot.
The experiment was carried out with an Ioffe trap in which the atoms
form a long, cigar-shaped cloud. At MIT they used a sheet of light
to chop the cloud into two pieces of roughly half the original length.
The light exerted a force that pushed the atoms out of the region of
high intensity (because it had a blue frequency detuning as explained in
Section 9.6). This configuration gave two separate potential wells. In
practice, the following situations both give the same results: (a) when
the two condensates are created independently; and (b) when a single
large condensate is divided into two parts after it has formed. The pro-
cess of turning on a sheet of light in the middle of an already-formed
condensate produces such a strong perturbation that the two resulting
condensates have almost random phases. Only recently has the con-
trolled separation of a condensate into two parts whilst preserving the
phase been demonstrated in a double-well dipole-force trap—a system
that corresponds to a beam splitter for matter waves. (Atom optics
is discussed further in Chapter 11.) However, the intriguing aspect of
the interference of two independent condensates is its dissimilarity to
previous experiments.