Atom trapping with a thin magnetic film

We have created a ${}^{87}\mathrm{Rb}$ Bose-Einstein condensate in a magnetic trapping potential produced by a hard disk platter written with a periodic pattern. Cold atoms were loaded from an optical dipole trap and then cooled to Bose-Einstein condensation on the surface with radio-frequency evaporation. Fragmentation of the atomic cloud due to imperfections in the magnetic structure was observed at distances closer than $40\mathrm{\mu }\mathrm{m}$ from the surface. Attempts to use the disk as an atom mirror showed dispersive effects after reflection.

Figure 1

(Color online) Geometry of the magnetic trap formed by a magnetic film. Magnetic field lines resulting from the addition of a radial $\left(\widehat{x}\right)$ bias field to the magnetized surface. The tubes represent the locations of the field minima where atoms are trapped. The addition of an axial bias field of about $1.0\mathrm{G}$ along $\widehat{z}$ prevents atom loss from Majorana spin flips.

Figure 2

Magnetic force microscopy image of the disk used in this experiment. This scan is over the region with $\lambda =1.0\mathrm{\mu }\mathrm{m}$. The track edges are of the same uniformity in the $\lambda =100.0\mathrm{\mu }\mathrm{m}$ region. The characteristic size of the roughness of the domain boundaries is about $30\mathrm{nm}$.

Figure 3

Absorption image of the BEC in trap near the surface. (a) Top imaging. The field of view is $0.58\mathrm{mm}\times 0.58\mathrm{mm}$. (b) Side imaging. The field of view is $0.56\mathrm{mm}\times 0.56\mathrm{mm}$. The double image comes from the reflection of the imaging beam on the platter surface. The trap frequencies were $2\pi \times (\mathrm{400,\; 400},10)\mathrm{Hz}$. The configuration of the experiment prevented ballistic expansion, but the bimodal distribution in trap is still clear.

Figure 4

Radial trap frequency vs applied radial magnetic field. Atoms were loaded into a single trap site on the surface and evaporated to BEC. The trap frequencies were measured with parametric heating, but above $5\mathrm{kHz}$ other heating effects made it difficult to resolve reliably. The deviation from linearity at low applied fields is most likely due to a small, off-axis, residual bias field.

Figure 5

Breakup of the atomic cloud as the atoms approached the surface. The thermal cloud was evaporated to $T\approx T_c$, then the axial trapping frequency was reduced from $10\mathrm{Hz}$ to $1\mathrm{Hz}$ to allow the atoms to expand in one dimension. The axial bias field was then increased to push the atoms closer to the surface, and the atoms were imaged in trap. The grazing incidence imaging used here produces a double image of the atoms from the primary and secondary reflections of the imaging beam off the disk. The field of view is $0.29\mathrm{mm}\times 2.3\mathrm{mm}$, and each frame is an average of five distinct images to reduce noise for clarity. At distances smaller than $40\mathrm{\mu }\mathrm{m}$, significant breakup was observed. Heights for the images shown are (a) $35\mathrm{\mu }\mathrm{m}$, (b) $29\mathrm{\mu }\mathrm{m}$, (c) $25\mathrm{\mu }\mathrm{m}$, (d) $22\mathrm{\mu }\mathrm{m}$, and (e) $19\mathrm{\mu }\mathrm{m}$.