Disordered 2D gases
PhDs: Guillaume Salomon, Lauriane Fouché
Undergraduates: Jan Lowinski
We focus our research on the physics in two-dimensions in the presence of disorder. In 2D compared to 3D, the roles of interactions as well as quantum and thermal fluctuations are enhanced. For bosons, a specific superfluid phase transition occurs, called the Berezinskii-Kosterlitz-Thouless transition. In our system, the disorder is introduced in a controlled way through a speckle light field. In a sense, we thus simulate intrinsically disordered condensed matter systems such as thin metallic films or Hi-Tc superconductors for which the physics is only partly understood.
We can first study the effect of disorder at the one particle level (i.e. without interaction). We can then observe classical trapping and diffusion, but also quantum phenomena such as weak and strong Anderson localization, which is the absence of diffusion due to interferences between scattering paths. 2D is the critical dimension for Anderson localization.
We also study the effect of disorder on the physics of interacting Bose gases. In such a many-body system, the competition between the interaction and the disorder leads to modification of the physics of the superfluid phase transition. Experimentally, we detect the phase coherence properties of the gas, which are modified at the phase transition. Moreover, for strong disorder and low temperatures, a transition to the elusive insulating Bose glass phase is expected.
Experimentally, we have recently modify our experiment in order to cool bosonic potassium 39, an element that allows us to control the interatomic interactions. As in our previous experiment with Rubidium, we rely on an all optical cooling scheme, which permits to reduce the duration of the experimental cycle to few seconds.
Schematic of the experimental setup. The atoms are directly cool to degeneracy in an optical dipole trap in the middle of the science chamber. They are then transfered in a blue-detuned 2D trap, The momentum distribution can be imaged after 3.4 cm of free fall using florescent imaging on an amplified CCD camera. The large viewports permit to add a speckle potential with a high numerical aperture.
Temperature of the D1 gray molasses as a function of the D1 cooling intensity per beam.
We have developed new techniques in cooling 39K atoms using laser light close to the D1 transition. First, a new compressed-MOT configuration is taking advantage of gray molasses type cooling induced by blue-detuned D1 light. It yields an optimized density of atoms. Then, we use pure D1 gray molasses to further cool the atoms to an ultra-low temperature of 6 µK. The resulting phase-space density is 2x10-4 and will ease future experiments with ultracold potassium. As an example, we use it to directly load up to 3x107 atoms in a far detuned optical trap, a result that opens the way to the all-optical production of potassium degenerate gases.
Typical example of a speckle disorder. The superfluid phase transition can be studied as a function of the disorder strength.
We experimentally study the effect of disorder on trapped quasi two-dimensional (2D) 87Rb clouds in the vicinity of the Berezinskii-Kosterlitz-Thouless (BKT) phase transition. The disorder correlation length is of the order of the Bose gas characteristic length scales (thermal de Broglie wavelength, healing length) and disorder thus modifies the physics at a microscopic level. We analyze the coherence properties of the cloud through measurements of the momentum distributions, for two disorder strengths, as a function of its degeneracy. For moderate disorder, the emergence of coherence remains steep but is shifted to a lower entropy. In contrast, for strong disorder, the growth of coherence is hindered. Our study is an experimental realization of the dirty boson problem in a well controlled atomic system suitable for quantitative analysis.
2D Momentum distribution for different atom number. The left image corresponds to a normal gas, the middle one to a gas at the superfluid transition and the right one to a gas deep in the superfluid phase. The shrinking of the momentum distribution is related to the increasing coherence in the 2D gas when crossing the BKT superfluid transition.
We measure the momentum distribution of a 2D trapped Bose gas and observe the increase of the range of coherence around the Berezinskii-Kosterlitz-Thouless (BKT) transition. We quantitatively compare our observed profiles to both a Hartree-Fock mean-field theory and to quantum Monte-Carlo simulations. In the normal phase, the momentum distribution is observed to sharpen well before the phase transition. This behavior is partially captured in a mean-field approach, in contrast to the physics of the BKT transition.
If you are interested in our research, please do not hesitate to contact Thomas Bourdel. We have open positions for PhD or post-docs. We also offer to some motivated undergraduate students at various levels to work with us every year. A link to a more detail thesis/intership proposition follows.
Group meetings take place every monday at 11:00 in Salle du conseil (2nd floor)
18.11.13: L. Orozco (Maryland University)
25.11.13: News from team Theory
03.12.13: S. Stringari (Trento University)