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 ...
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
5.2 Branch<strong>in</strong>g Ratios <strong>and</strong> Population Distribution<br />
The irreversible step <strong>in</strong> the implementation of s<strong>in</strong>gle-pho<strong>to</strong>n cool<strong>in</strong>g <strong>in</strong> rubidium<br />
lies <strong>in</strong> the spontaneous scatter<strong>in</strong>g event that occurs after the a<strong>to</strong>m has been excited <strong>to</strong> the<br />
|F ′ = 1〉 state. Because of the branch<strong>in</strong>g ratios it is impossible <strong>to</strong> transfer all the a<strong>to</strong>ms<br />
from the |F = 2,mF = 2〉 state <strong>in</strong><strong>to</strong> one f<strong>in</strong>al trapped state <strong>in</strong> the gravi<strong>to</strong>-optical trap.<br />
Instead they are evenly distributed between the |F = 1,mF = 1〉 <strong>and</strong> |F = 1,mF = 0〉<br />
states. After completion of the s<strong>in</strong>gle-pho<strong>to</strong>n cool<strong>in</strong>g sequence, the distribution of a<strong>to</strong>ms<br />
between the two f<strong>in</strong>al states is determ<strong>in</strong>ed by apply<strong>in</strong>g a magnetic field gradient such<br />
that a<strong>to</strong>ms <strong>in</strong> the |F = 1,mF = 1〉 state are ejected out of the gravi<strong>to</strong>-optical trap,<br />
leav<strong>in</strong>g only the magnetically decoupled a<strong>to</strong>ms <strong>in</strong> the |F = 1,mF = 0〉 state <strong>in</strong> the trap.<br />
As expected about 50% of the <strong>to</strong>tal number of a<strong>to</strong>ms are <strong>in</strong> this state. Only a<strong>to</strong>ms <strong>in</strong><br />
the |F = 1,mF = 0〉 state are considered when calculat<strong>in</strong>g phase-space density.<br />
About 10% of the a<strong>to</strong>ms that encounter the demon will fall back <strong>in</strong><strong>to</strong> the |F =<br />
2,mF = 2〉 state, <strong>and</strong> about 5% will fall <strong>in</strong><strong>to</strong> the |F = 2,mF = 1〉 state. These a<strong>to</strong>ms<br />
are thus transferred back <strong>in</strong><strong>to</strong> the reservoir <strong>and</strong> can undergo another cycle of excitation<br />
<strong>and</strong> decay when encounter<strong>in</strong>g the demon beam. Only about 5% of the a<strong>to</strong>ms decay <strong>in</strong><strong>to</strong><br />
the untrapped |F = 2,mF = 0〉 state <strong>and</strong> are lost from the system.<br />
5.3 Effect of the Demon Beam Detun<strong>in</strong>g<br />
As mentioned before, the demon beam is slightly detuned from resonance. Naively<br />
one would assume that the beam should ideally be on resonance. This is not true, as<br />
figure 5.5 shows. The maximum number of a<strong>to</strong>ms is transferred <strong>in</strong><strong>to</strong> the trough not at<br />
zero detun<strong>in</strong>g, but rather at a power-dependent optimum value.<br />
There are two reasons for this phenomenon. First, consider the scattered pho<strong>to</strong>n.<br />
As the density of a<strong>to</strong>ms <strong>in</strong> the trap <strong>in</strong>creases, the probability of reabsorption of this<br />
pho<strong>to</strong>n by a<strong>to</strong>ms <strong>in</strong> the optical trough <strong>in</strong>creases as well. If this pho<strong>to</strong>n is absorbed by<br />
an a<strong>to</strong>m <strong>in</strong>side the trap, this additional scatter<strong>in</strong>g event can lead <strong>to</strong> trap loss. If the<br />
demon beam detun<strong>in</strong>g ∆ is away from resonance, the probability of reabsorption of this<br />
pho<strong>to</strong>n is dramatically decreased, as the scatter<strong>in</strong>g rate is roughly proportional <strong>to</strong> 1/∆.<br />
72