24.07.2013 Views

Experiments to Control Atom Number and Phase-Space Density in ...

Experiments to Control Atom Number and Phase-Space Density in ...

Experiments to Control Atom Number and Phase-Space Density in ...

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

phase-space density is def<strong>in</strong>ed as ρ ≡ nλ3 dB , λdB = √ √ 2π , where n is the spatial density<br />

mkBT<br />

of a<strong>to</strong>ms with mass m at temperature T. The phase-space density is limited by the<br />

temperature achievable <strong>in</strong> a MOT as well as by the spatial density of a<strong>to</strong>ms (typically<br />

on the order of ≈ 10 11 cm −3 ) [38]. Once the density exceeds a critical value a<strong>to</strong>ms<br />

start absorb<strong>in</strong>g fluorescent light emitted by other a<strong>to</strong>ms. This leads <strong>to</strong> a repulsive force<br />

between the a<strong>to</strong>ms limit<strong>in</strong>g the maximum density [35].<br />

2.10 Evaporative Cool<strong>in</strong>g<br />

The phase-space density <strong>in</strong> a magne<strong>to</strong>-optical trap is still about 6 orders of mag-<br />

nitude below degeneracy. The st<strong>and</strong>ard method of cool<strong>in</strong>g alkali a<strong>to</strong>ms further is evap-<br />

orative cool<strong>in</strong>g. This can happen <strong>in</strong> either a magnetic trap us<strong>in</strong>g an RF-knife, as is<br />

typically done for 87 Rb or 23 Na, or <strong>in</strong> an optical dipole trap, as is typically done for<br />

6 Li, where evaporative cool<strong>in</strong>g <strong>in</strong> a magnetic trap is not possible, see chapter 2.5.1. The<br />

basic pr<strong>in</strong>ciple of evaporative cool<strong>in</strong>g is illustrated <strong>in</strong> figure 2.19.<br />

Consider an a<strong>to</strong>mic cloud trapped <strong>in</strong> an optical dipole trap of trap depth U. If<br />

the thermal tail extends above the trap depth, the hottest a<strong>to</strong>ms with E > U will leave<br />

the trap, effectively lower<strong>in</strong>g the average energy of the rema<strong>in</strong><strong>in</strong>g a<strong>to</strong>ms. The rema<strong>in</strong><strong>in</strong>g<br />

a<strong>to</strong>ms then rethermalize through collisions, thus lower<strong>in</strong>g the average temperature of the<br />

trapped a<strong>to</strong>ms. Both of these processes happen simultaneously, until the thermal tail<br />

no longer extends above the trap depth <strong>and</strong> evaporation s<strong>to</strong>ps. The evaporative cool<strong>in</strong>g<br />

process can be forced <strong>to</strong> cont<strong>in</strong>ue by reduc<strong>in</strong>g the trap depth U.<br />

Evaporative cool<strong>in</strong>g can only be applied successfully if elastic collisions between<br />

a<strong>to</strong>ms dom<strong>in</strong>ate over <strong>in</strong>elastic collisions. This is true for both fermions <strong>and</strong> bosons.<br />

However, fermions cannot scatter with other fermions <strong>in</strong> the same state, due <strong>to</strong> the<br />

anti-symmetric nature of their wavefunction. Evaporative cool<strong>in</strong>g can therefore only be<br />

done us<strong>in</strong>g either a<strong>to</strong>ms <strong>in</strong> two different states [39], us<strong>in</strong>g two different a<strong>to</strong>ms, or by<br />

sympathetic cool<strong>in</strong>g [14]. The condition of a dom<strong>in</strong>at<strong>in</strong>g elastic cross section is satisfied<br />

for the two lowest states of 6 Li, |F = 1/2,mF = 1/2〉 <strong>and</strong> |F = 1/2,mF = −1/2〉, <strong>and</strong><br />

it is therefore possible <strong>to</strong> use these states for evaporative cool<strong>in</strong>g.<br />

26

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!