The implementation of a fast actuating valve to remove the buffer gas and thermally isolate the trapped atoms extends buffer gas loading to atoms whose magnetic moments are 2 Bohr magneton or greater. This triples the number of atoms accessible to buffer gas loading and bridges the gap between atoms which can be buffer gas loaded and those which have been Bose-condensed, most of which have less than or equal to 2 Bohr magneton. Of these, metastable helium He* (2 Bohr magneton ), has the unique property of having 20 eV of internal energy.

Experiments to produce Bose-condensates of He* have thus far used laser cooling as the initial loading stage. However, the number of 4He* capable of being loaded into a magnetic trap is limited by both the low efficiency for exciting helium to the metastable state and the lower cooling rate of the 1083 nm cooling transition as compared to the transitions used in cooling the alkali metals.

Since buffer gas loading already uses a cryogenic helium gas to load atoms into a trap, we felt it was ideally suited for producing and trapping He*. The advantage over laser cooling is the capability of loading orders of magnitude larger numbers of He* into a trap. The second advantage is the ability to load 4He* or 3He* or even mixtures of 4He* and 3He*. The advantage of larger trapped samples afforded to us by buffer gas loading should lead to larger condensates and facilitate novel experiments in quantum atom optics and is a promising candidate to sympathetically cool polar molecules into the ultracold regime

Production and Magnetic Trapping
He* is produced via an rf-discharge. The copper rf-coil surrounds the trapping region of the cell and couples in a 10 W, 100 microsecond pulse. We also use a 5 ns, 1 mJ pulse from a frequency-doubled Nd:YAG laser to reliably ignite the discharge.

With the cell at 400 mK and a trap depth of 4.9 K, we are able to trap upwards of 10 11 He* atoms. We are able to individually trap both 4He* and 3He*. Figure 1 shows the magnetically broadening spectra of trapped 4He* and 3He*. Helium atoms that were not magnetically trapped were pumped away to either the walls or charcoal sorb.

Figure 1. Spectra of trapped He*. 10 11 He* atoms are trapped at an initial temperature of 400 mK. Both isotopes can be trapped.

Evaporative Cooling
After magnetically trapping, we evaporatively cool to try to reach quantum degeneracy. However rather than uniformly lowering the trap depth, we evaporate by lowering the trap sample towards to the cell window. Atoms which hit the window are lost through adsorption. As the atoms at the edge of the cloud have higher energy than the average energy of an atom in the cloud, evaporative cooling occurs. In this method, the trap depth is decoupled from the confinement. Therefore we are able to maintain tight confinement and high collision rates throughout the evaporation process.

We start with the atoms 5 cm from the window. Over 180 s, the trap center is gradually ramped towards the window to a final separation of 580 micron. The 400 mK clouds has been evaporatively cooled to 1.4 mK with 2x10 9 atoms remaining (Fig. 2).

Figure 2. Trapping and evaporative cooling of 4He*.
A) The magnetic field contour lines at the initial trap depth of 4.9 K.
Inset: Spectrum of trapped 4He* initially loaded at 0.4~K.
B) The contour lines at the end of the evaporation at a final trap depth of 2.7 mK Inset: Spectrum of trapped 4He* after evaporation. The fit (solid line) gives T= 1.4 mK, n=2x10 12 cm 3 and N=2x10 9 .

This result marks three significant feats. First this is comparable or larger in number than that attained via laser cooling and with much room for improvement. Second, evaporation is performed well in the multi-partial-wave regime at 400 mK and was efficient all the way into the ultracold regime. And third, this is an increase of 5 orders of magnitude in phase space density from the initial loading conditions, the first time a significant increase in phase space density for a buffer gas loaded sample has been achieved.

Reaching Quantum Degeneracy
Further cooling right now is limited by Majorana flops from the magnetic field zero of the quadruple trap. Our lifetimes range from hundreds of seconds to just a few seconds at our coldest temperatures. Therefore to reach quantum degeneracy, we are currently implementing a new cloverleaf Ioffe trap in which the atoms can be transferred into after the initial evaporation stage (Fig. 3).

Figure 3. Implementation of the clover leaf Ioffe trap.
A) Schematic diagram of the modified cell with the clover leaf trap.
B) Photo of modified cell.
C) Photo of the clover leaf trap.

Once transferred into the Ioffe trap, the final evaporation to degeneracy will be further enhanced by using an rf knife. An evaporation model predicts that for our initial number and density at 1.4 mK, a Bose-Einstein condensate of greater 10 7 should be easily achievable (Fig. 4).

Figure 4. Comparison of laser cooling and buffer gas loading evaporation trajectories. The dashed line is the projected path based on an evaporation model using the known cross sections.