An intense source of rubidium condensates

A major goal of the rubidium project has been the development of an intense source of rubidium condensates. Drawing from our expertise with sodium condensates, we adopted important components from our latest-generation BEC apparatus, including the vacuum system, the magnetic trap and the electronic control. However, the use of rubidium also led to several new developments: A new oven design with large cold plates prevents the build-up of background vapor due to the much higher vapor pressure of rubidium. The design of the Zeeman slower for rubidium addresses potential complications from the large hyperfine coupling of rubidium and, finally, the experiment employs stable, user-friendly diode lasers instead of dye lasers.

LOADING A MAGNETIC TRAP
Bose-Einstein condensates of rubidium-87 are produced in a multistage process involving laser cooling and trapping, as well as evaporative cooling in a magnetic trap. In our apparatus, a sample of rubidium is first heated to 120°C in an oven chamber. A thermal atomic beam escapes from the oven and enters a Zeeman slower tube where it is decelerated by photon scattering from a red-detuned, counterpropagating laser beam. As the atoms slow down and travel down the tube, they are kept on resonance using an increasing, longitudinal magnetic field. The maximum of the magnetic field profile of the Zeeman slower is at 270 G, corresponding to a capture velocity of 325 m/s (taking the final velocity of slowed atoms to be 30 m/s). In our design an additional, homogeneous bias field of 180 G prevents optical pumping out of the cycling transition. With this Zeeman slower, we have achieved a beam flux exceeding 1x1011 atoms/s, a value that had so far only been achieved with sodium. After leaving the slower tube, the slow atoms enter the main chamber with a background pressure in the sub-10-11 torr range, where they are accumulated in a magneto-optical trap (MOT). We use the standard geometry in which 6 beams intersect in the zero of a three-dimensional quadrupole magnetic field (no retroreflected beams) .With 60 mW of total laser power (beam diameters about 1 inch) we are able to load >4x1010 atoms in the MOT within 2 seconds. After a short compression of the MOT we switch to a dark molasses and then transfer the atoms into a spatially overlapped magnetic trap for evaporative cooling.

Laser cooling and trapping is performed using external-grating cavity diode lasers, injection-locked slave lasers and a tapered amplifier. All lasers are locked to atomic transitions using rubidium vapor cells. The laser system is located on a separate optical table. The light is delivered to the apparatus using single-mode optical fibers which ensures high stability and reliability of the beam geometry.

EVAPORATIVE COOLING
Our Ioffe-Pritchard magnetic trap is of the cloverleaf type. In the apparatus the magnetic coils are located outside the main chamber in a recessed, stable “bucket window” geometry that minimizes shot-to-shot fluctuations in BEC production. To load the magnetic trap, atoms from the molasses are first pumped from F=2 to the F=1, mF=-1 ground state. In the subsequent evaporation process, resonant RF radiation from a small coil inside the vacuum is used to couple the hottest atoms to untrapped sublevels that are expelled from the trap, while the remaining atoms rethermalize at a lower temperature. In a rampdown of the RF frequency we typically reach quantum degeneracy in 20 seconds Over the last year we have optimized the performance of the rubidium BEC apparatus for large atom numbers. This involves lowering the magnetic confinement in the final phase of the evaporation, which reduces the loss of atoms due to inelastic collisions.

RESULT: RECORD PERFORMANCE
We are now able to create large F=1 condensates with more than 10 million atoms which to date are the largest numbers for rubidium anywhere (we also have achieved efficient production of F=2 condensates). The goal of the Center to transfer our experience of producing large condensates of sodium to rubidium has thus been accomplished. We are in the process of documenting in detail the technological steps untertaken to share them with the community.

A BEAM LINE FOR RUBIDIUM BEC
Many experiment require large experimental access to the condensate from the outside. Our approach is modular by combining our high-performance production apparatus (with its inherent limitations in experimental access) with a bare science chamber/glass cell. Over the last year we have taken important experimental steps toward this goal by adding a science-chamber vacuum system into which condensates will be delivered with an optical tweezer. Similar to what is done in one of our sodium-BEC machines, condensates are trapped in the focus of a laser beam which is translated using a linear translation stage. It is thus possible to decouple experiments with condensates from their production, in analogy to a beam line in high-energy physics. This way a variety of new experimental directions, including surface atom optics, can be explored.

As a first application for this scheme we have demonstrated trapping of neutral atoms using magnetic fields produced by a thin magnetic film mounted in the science chamber. The film is magnetized in alternating north/south stripes with a 10 µm period. The magnetizable surface used was a hard disk platter provided by Hitachi Global Storage Technologies. Tube shaped traps were created with an additional radial bias field, and the traps were loaded with atoms from a condensate held in a wire trap. Radial trap frequencies of up to 20 kHz have been observed.

OPTICAL TRAPPING AND TRANSPORT
We have successfully transported rubidium atoms using the optical tweezers. The geometry of the apparatus requires the laser beam to be perpendicular to the long axis of the magnetically trapped condensate. Loading into this geometry is complicated by the large three-body loss and high mass of rubidium as compared to sodium. To overcome this restriction we load cold thermal atoms into the tweezer at low densities. After translation we then perform all-optical evaporation in the science chamber by ramping down the power in the tweezer beam. Using this method we create condensates of up to 1 million atoms in the science chamber. Both atom number and transport time are comparable to those achieved in the sodium experiment.