We have created vortices in two-component Bose-Einstein condensates. The vortex state was created through a coherent process involving the spatial and temporal control of interconversion between the two components. Using an interference technique, we map the phase of the vortex state to confirm that it possesses angular momentum. We can create vortices in either of the two components and have observed differences in the dynamics and stability.

We have created spatial dark solitons in two-component Bose-Einstein condensates in which the soliton exists in one of the condensate components and the soliton nodal plane is filled with the second component. The filled solitons are stable for hundreds of milliseconds. The filling can be selectively removed, making the soliton more susceptible to dynamical instabilities. For a condensate in a spherically symmetric potential, these instabilities cause the dark soliton to decay into stable vortex rings. We have imaged the resulting vortex rings.

We have developed an evaporative cooling technique that accelerates the circulation of an ultra-cold 87 Rb gas, confined in a static harmonic potential. As a normal gas is evaporatively spun up and cooled below quantum degeneracy, it is found to nucleate vorticity in a Bose-Einstein condensate. Measurements of the condensate's aspect ratio and surface-wave excitations are consistent with effective rigid-body rotation. Rotation rates of up to 94% of the centrifugal limit are inferred. A threshold in the normal cloud's rotation is observed for the intrinsic nucleation of the first vortex. The threshold value lies below the prediction for a nucleation mechanism involving the excitation of surface-waves of the ground-state condensate. 03.75.Fi,67.90.+z,67.57.Fg,32.80.Pj To paraphrase an ancient riddle, what happens when an irresistible torque meets an irrotational fluid? The answer has been known for more than 50 years: a quantized vortex is nucleated. Vortices alone contribute to a superfluid's rotation, so that the bulk of the fluid may remain curl-free. The nucleation of vortices in bulk superfluid Helium has been the topic of extensive study (for a review see [1]). In the archetypical experiment, a rotatable pot filled with a mixture of superfluid and normal liquid Helium undergoes gradual angular acceleration. The normal fluid and the walls of the pot rotate together as a rigid body, defining a rotating environment. At some threshold angular velocity, a vortex line is nucleated at the circumference of the pot, and then quickly migrates inward until it is collinear with the axis of rotation. Further angular acceleration results in the nucleation of more vortices; eventually the fluid is filled with an array of vortex lines [2]. A central theme [3] of this research is the question: to what extent is the nucleation process "extrinsic," i.e. dependent on such details as the roughness of the surface of the walls, and to what extent is it "intrinsic" [4,5], i.e. driven (in the limit of microscopically smooth walls) by the flow of normal fluid along the boundary of the superfluid? In the analogous rotating-potential experiments with a dilutegas Bose-Einstein condensate (BEC), the confining potential and the normal fluid typically rotate at different rates [6]. In this context, the extrinsic-intrinsic question can be restated as: is it the confining potential or the normal fluid that defines the rotating environment?Vortices in a BEC have been created with wavefunction engineering [7], through the decay of solitons [8,9], and in the wake of moving objects [10,11]. The first rotatingpotential experiment to detect vortices in a BEC was performed by the Paris group [12]; results have also been obtained by the MIT [13] and Oxford [14] groups. In these experiments the role of the normal fluid was secondary to that of the rotating potential; it is conceivable the normal fluid was not rotating at all. This paper presents vortex nucleation experiments performed in the opposite limit, namely in the environment of a rotating...