Transcript spin mixtures - Theory of Quantum Gases and Quantum Coherence

COLLISIONS IN ULTRACOLD
METASTABLE HELIUM GASES
G. B. Partridge, J.-C. Jaskula, M. Bonneau, D. Boiron, C. I. Westbrook
Laboratoire Charles Fabry de l’Institut d’Optique, Palaiseau France
Outline
Motivation and Background
-Optics, atomoptics, quantum optics, quantum atom optics…
Methods, apparatus, He*.
Experiments:
Optical Trapping and
Relative Number
4-wave mixing of
Squeezing
matter waves
Spin Mixtures
2
Motivation, atom optics…
Optics : Photons, waves… wave particle
duality.
Atomic physics  atom optics :
i.e – slits, interferrometers, etc
Bec  coherent atom optics:
Atom Laser, fringes, + nonlinear atom
optics (interactions): 4wm , solitons…
Quantum atom optics?
-ex’s correlations, squeezing,
entanglement, teleportation…
T. Pfau (Stuttgart)
L. Deng et al. (NIST)
Use counting, single particles,
statistics…
-Key is detection: metastable Helium (He*).
Strecker et al. (Rice)
3
He* : What’s it hiding?
The 23S1 state of He has a decay
time ~ 8000 s !*
*single atom ~ spin polarized
(So what?)
The stored energy of the metastable
state is 19.8 eV/atom.
This energy can kick off electrons &
ionize atoms of surfaces that the
atom meets.
Add in a potential, get an
avalanche of electrons.
 High gain amplifier = single
atom sensitivity.
e+
4
Trapping and Cooling He*
Laser cooling helium?
Behaves a lot like an alkali-metal.
(Cycling Optical Transition,
magnetically trappable )
I
I
5
Single Atom Detection
Use a micro-channel plate
(many e- avalanche
detectors in parallel) to give
position information.
Gather resulting electric
pulses using crossed
delay lines.
Use relative arrival times to
reconstruct atoms’ positions
(time of flight) in 3D.
6
A new tool for He*
Statistical measurements: 1000’s of
repetitions.
magnetic trap was not engineered
for this…
(although we try anyway)
Long term: favors Optical Trap
magnetic
optical
Also, better geometry: aligns long
axis of potential ( short TOF, short
correlation length) w/ high
resolution direction, Z.
Gives freedom to try spin
mixtures…
TOF
First step towards more
complicated potentials for He*
(lattices, disorder etc.)
7
BEC of He* in the optical trap
Transfer from magnetic trap after some pre-cooling: N = 5 x 106, T = 15 K
Evaporate by reducing intensity of trap laser over ~ 4 sec.
N0 = 105
r = 1.5 kHz, z = 8 Hz
G. B. Partridge, J.-C. Jaskula, M. Bonneau, D. Boiron, C. I. Westbrook,
Phys. Rev. A 81, 053631 (2010).
8
Quantum Optics: photon pairs
9
Matter Wave FWM: atom pairs
Create an m = 0
condensate w/ raman
pulse.
k0
Split BEC into two momentum
components with Bragg
pulse: +/- k0
S-wave interactions lead to
spherical shell of scattered
atoms at k=kS
 spontaneous FWM
k0
kS
k0
kS
k0
10
The “intuitive” result
Scattered pairs are correlated…
kS
k0
k0
kS
P(Δt)
0
Δt
Like in photon pairs:
“Enhanced
coincidence rate
when phase
matching
condition is met.”
11
Beyond Optics: smaller sphere
“energy balance”
kS
(per atom)
kS
k0
kS
k0
k0
Energy gain from
removal of atom from
condensate mode
k0
<
kS
Energy Cost to put atom
into scattered mode
(still overlapped w/
condensate).
|kS| < |k0|
V. Krachmalnicoff, J.-C. Jaskula, M. Bonneau, V. Leung, G. B. Partridge, D. Boiron, C. I. Westbrook, P.
Deuar, P. Zin, M. Trippenbach, K. Kheruntsyan, Phys. Rev. Lett. 104,150402 (2010).
12
Plus, the sphere’s not a sphere
After colliding, atoms still have
to get out of the region of
the condensates.
i.e. they roll down the mean
field hill: V = 2g(r,t)
But the hill is collapsing out
from under them.
Anisotropy of BEC’s leads to directional acceleration
Lesson Learned:
Do Q.O. experiments using
atoms, but be careful about
simple 1:1 intuition. There are
differences, for better or worse…
Phys. Rev. Lett. 104,150402 (2010).
Analogy? ponderomotive force in high
harmonic generation
(Balcou et al PRA 1997)
13
Intermediate Q.O.: Relative N Squeezing
Measurement of intensity noise
between “twin” beams.
IA  IB
R
I A  IB
A
B
Heidmann et al. PRL 59 2555 (1987)
Reduction in noise, 30% below the shot noise limit!
14
New Atom Pairs
Optical Trap BEC
RF + Bragg pulse.
Back-to-Back Correlations: 3600 shots
Collision along long
axis + better
repeatability gives
improved S/N.
Now what about
squeezing?
15
Matter Wave N Squeezing
Divide scattered halo into sections,
compare number difference in
geometrically opposing zones to
that of non-opposing zones.
M ij 
Ni  N j
Ni  N j
, M 1
(for uncorrelated N, i.e. shot noise)
M
1
M
16 zones
16
Details…
Detail 1:
Raw data ~ -0.5 dB squeezing
Why isn’t it perfect?
(partly b/c its an experiment)
Detail 2:
Effect of of correlation length:
~Measurement bandwidth
Specifically, the detector efficiency,
, limits the measured variance.
Perfect correlations: M = (1- )
 = 0.6 (“open area”) : -3 dB
 = .13 (best estimate): -13 dB
17
What’s next?
18
Alternate Future: spin mixtures
With optical trap, we can think about using different spin states
(mJ = +1,-1,0)
spin mixtures, spinor condensates …
But! Trapped He* gases are prone
to loss due to Ionization-enhanced
inelastic loss processes.
Spin Polarization in the mJ = 1
provides stabilization by ~5 orders
of magnitude.
RF transfers: spin mixtures
What about other states and
combinations of states?
State specific loss constants
unconfirmed experimentally (only
mJ = 1 is magnetically trappable)
19
Loss Rates in a spin mixture
Inelastic Loss Experiment 1:
Put them all together and see what survives…
“Large” loss rate: 00, ±1
“Small” loss rate: 01, 0-1, 11, -1-1
20
G. B. Partridge et al., Phys. Rev. A 81, 053631 (2010).
Quantitative Loss Rates
Inelastic Loss Experiment 2: Make careful
measure of the dominant processes 00 ±1.
00 = 6.6(4) × 10 −10 cm3/s
±1 = 7.4(10) × 10 −10 cm3/s.
 Not necessarily prohibitive! (for certain things…)
21
G. B. Partridge et al., Phys. Rev. A 81, 053631 (2010).
Summary
1. Quantum Atom Optics: Spontaneous FWM
of deBroglie matter waves.
• Don’t forget they’re atoms.
2. Relative Number Squeezing for correlated
atom pairs.
• Atomic version of a Quantum Optics Classic.
3. Spin Mixtures in of He* ?
• Stay tuned…
22
Thanks!
Questions?
23