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.

The magnetic field gradients <strong>in</strong>troduce a position-dependent scatter<strong>in</strong>g force on<br />

the a<strong>to</strong>ms, conf<strong>in</strong><strong>in</strong>g the a<strong>to</strong>ms with<strong>in</strong> the magne<strong>to</strong>-optical trap radius. For simplicity<br />

consider a one-dimensional setup <strong>and</strong> an a<strong>to</strong>m with the ground state J = 0 <strong>and</strong> an<br />

excited state J = 1. In the presence of the quadrupole magnetic field, the Zeeman effect<br />

splits the excited state <strong>in</strong><strong>to</strong> three states, denoted by m = +1, m = 0 <strong>and</strong> m = −1, see<br />

figure 2.16. The transition frequency between the ground <strong>and</strong> the excited states m±1<br />

is thus vary<strong>in</strong>g along the quadrupole field. The laser beams are red-detuned from the<br />

transition frequency (ω0) at zero magnetic field by an amount ∆ . At the position z = z ′<br />

the transition frequency from the ground state <strong>to</strong> the excited state with m = −1 is then<br />

resonant with the laser frequency, ω(z = z ′ ) = ω(z = 0)−∆. This transition is driven by<br />

σ − circularly polarized light, the beam imp<strong>in</strong>g<strong>in</strong>g from the right h<strong>and</strong> side. Because the<br />

detun<strong>in</strong>g for the transition between the ground state <strong>and</strong> the m = +1 state is 2∆, the<br />

a<strong>to</strong>m is more likely <strong>to</strong> scatter pho<strong>to</strong>ns from the laser beams com<strong>in</strong>g from the right than<br />

from the beam com<strong>in</strong>g from the left h<strong>and</strong> side hav<strong>in</strong>g σ + polarization. This imbalance <strong>in</strong><br />

scatter<strong>in</strong>g events moves the a<strong>to</strong>m back <strong>to</strong>wards the center of the quadrupole field. If the<br />

a<strong>to</strong>m is displaced along the other direction, the selection rules for driv<strong>in</strong>g a transition<br />

now imply that the a<strong>to</strong>m will scatter more pho<strong>to</strong>ns from the beam com<strong>in</strong>g from the left<br />

than from the beam com<strong>in</strong>g from the right, aga<strong>in</strong> mov<strong>in</strong>g the a<strong>to</strong>m back <strong>to</strong>wards the<br />

center of the trap. This scheme only works as long as a<strong>to</strong>ms return <strong>to</strong> the same ground<br />

state. This type of transition is then called a cycl<strong>in</strong>g transition.<br />

equation 2.30.<br />

Quantitatively the res<strong>to</strong>r<strong>in</strong>g force is described by add<strong>in</strong>g the Zeeman shift <strong>in</strong><strong>to</strong><br />

where β = gµB<br />

<br />

FMOT = F σ+<br />

scatt(ω −(ω0 +βz)−kv)−F σ−<br />

scatt(ω −(ω0 +βz)+kv), (2.32)<br />

dB.<br />

Near the trap center it follows that<br />

dz<br />

FMOT ≈ −2 ∂F ∂F<br />

kv +2 βz (2.33)<br />

∂ω ∂ω0<br />

= −2 ∂F<br />

(kv +βz) (2.34)<br />

∂ω<br />

= −αv − αβ<br />

z. (2.35)<br />

k<br />

23

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

Saved successfully!

Ooh no, something went wrong!