Transcript pres - Division of Engineering and Applied Science

Disclaimer
• This collection of slides is for reference only, in
connection with the workshop at which it was
presented. It is not intended as a complete
review or a definitive overview of the field,
but only as a presentation of some material
that might be helpful. The opinion slides that
follow the “End” are the opinion of the author
and are not necessarily those of NIST or of the
US government. Please do not use or
reproduce any of the material from this
presentation without the permission of the
author.
Please see disclaimer at beginning of slides
Quantum information, computing,
and simulation with cold atoms
William D. Phillips
Joint Quantum Institute
National Institute of Standards and Technology, Gaithersburg, MD
University of Maryland, College Park, MD
NIST/JQI Laser Cooling and Trapping Group:
Kris Helmerson, Paul Lett, Trey Porto, Ian Spielman
I will present some of the state-of-the-art internationally,
including work in the NIST/JQI Laser Cooling and Trapping
Group, along with my personal perspectives.
$upport: DARPA, IARPA, JQI, NIST, NSF (PFC), ONR
Please see disclaimer at beginning of slides
Quo Vadis Cold Atoms?
Enticing themes for the future with cold atoms:
Many-body entanglement
Strongly correlated systems
Large scale systems married with control and measurement that
spans bulk properties and individual particles and correlations.
Engineering a different class of Hamilltonian—exotic kinds of
matter (e.g., topological matter)
Please see disclaimer at beginning of slides
Why Cold (Neutral) Atoms?
As Qubits
As Simulators
• Excellent coherence—clocks
• Optical lattices are
flexible and reconfigurable
• Known Hamiltonian
• Choice of dimensionality,
• Optical information transfer
quantum statistics, lattice
• High qubit density
type (even quasi-periodic)
• Mobility: arbitrary pair
•Tunneling and interactions
entanglement
are tunable
• Both individual and massively parallel
• Lattices are near-perfect:
operations
no phonons, no dislocations
•Plentiful fresh qubits—important for
• Non-lattice applications
error correction
• Measurements: velocity,
•Choice of qubit basis
coherence, correlations, …
• High-fidelity, fast single-qubit
Simulation can be to realize a
operations, detection
mathematical model or to
• etc.
Please see disclaimer at beginning of slides
mimic a physical system.
Optical
Lattice
Optical
Tweezers
• Versatile and agile
• Address a single atom or
several nearby atoms
• Induce state changes or
entangle atoms
• Prepare large numbers of
initialized atomic qubits
• Store a large number of
atoms in a small space
• Perform identical operations
on large numbers of atoms
• Move atoms to distant
Please see disclaimer at beginning of slides
locations
DiVincenzo Criteria (5+2)
1. Well defined extendible qubit array; stable memory
2. Initialization of the qubits
3. Long decoherence time—much longer than gate time
4. Universal set of gate operations
5. Qubit-specific quantum measurements
(Cold atoms have, with varying degrees of fidelity, in various
laboratories, achieved all of these)
6.Interconvert stationary and flying qubits
7.Transmit flying qubits from place to place
(Information exchange from atoms to photons is more natural
than with some other kinds of qubits. See “The Quantum
Internet,” H. J. Kimble, Nature (2008), and his talk here.
Please see disclaimer at beginning of slides
Initializing lots of cold-atom qubits
Mott insulator
“superfluid” phase
first observed with cold atoms
by I. Bloch et al. in Munich ‘02
BEC
( commensurate filling )
Bose-Einstein Condensate (1995 NIST/JILA and MIT;
Nobel 2001); millions of atoms in the same quantum state.
deeper lattice
NIST/JQI
2-D Mott insulator,
Spielman et al. 2007
lattice spacing
~ 0.4 mm
Please see disclaimer at beginning of slides
Entangling adjacent atoms in a lattice
(Hansch group (Munich/Garching); Bloch, Greiner, et al.)
(Following a suggestion by Cirac and Zoller)
In the long chain of atoms—one per site due to the Mott
transition—not pairwise entanglement, but cluster
entanglement results (useful for measurement-based
quantum computation).
Please see disclaimer at beginning of slides
Swap Movie
quantum SWAP of isolated pairs NIST/JQI
SWAP movie
Please see disclaimer at beginning of slides
Swap oscillations: Porto Group at NIST/JQI
no decoherence during many SWAP oscillations
Fidelity > 64%
0.5
1.0
milliseconds
SWAP is a “universal” 2-qubit entangling gate

Please see disclaimer at beginning of slides
1.5
Swap oscillations: Porto Group at NIST/JQI
no decoherence during many SWAP oscillations
Fidelity > 64%
0.5
1.0
milliseconds
1.5
SWAP is a “universal” 2-qubit entangling gate
But all of these lattice entangling operations detect many qubits or pairs of qubits at once

Please see disclaimer at beginning of slides
Quantum gas microscope: Greiner Group, Harvard
High-resolution imaging just resolves atoms in adjacent lattice sites.
lattice spacing = 600 nm
Please see disclaimer at beginning of slides
3D Optical Lattice with Large Spacing
(Weiss Group, Penn State)
θ =10°
4.9 μm
With these large lattice spacings,
atoms can be individually addressed
and pairwise entangled using focused
Please see
disclaimer
at beginning coupling
of slides
laser beams
(but
tunnel
is low)
Atoms in a large-spacing lattice at Penn State
This is one plane of a 3D
lattice, with random half-filling
due to destruction at multiplyfilled sites.
Atom
Please see disclaimer at beginning of slides
Repairing and manipulating an atom storage
register (Meschede Group, Bonn)
Two perpendicular conveyor-belts
• "sort" atoms
• coupling between arbitrary pairs of atoms
Please see disclaimer at beginning of slides
Sorting Atoms
False color tagging to guide the eye
Please see disclaimer at beginning of slides
Sorting Atoms
Please see disclaimer at beginning of slides
Sorting Atoms
Please see disclaimer at beginning of slides
Sorting Atoms
Please see disclaimer at beginning of slides
Sorting Atoms
Please see disclaimer at beginning of slides
Sorting Atoms
Please see disclaimer at beginning of slides
Sorting Atoms
Please see disclaimer at beginning of slides
Sorting Atoms
Please see disclaimer at beginning of slides
Sorting Atoms
Dx ~ 15 mm
Dx
Dx
Please see disclaimer at beginning of slides
Optical Tweezers
Another, and more flexible, way to have
individual addressing of single qubits is with
optical tweezers—single focused laser
beams that trap or light-shift specific atoms.
Please see disclaimer at beginning of slides
Trapping single atoms in optical tweezers
(Grangier group—Palaiseau)
Dipole trap 1
lens
Atom A
Atom B
Vacuum
CCD image
Dipole trap 2
Single qubit gate: 2 ms p-pulse,
99% Fidelity (prep.+ gate + detection);
40 ms decoherence time
Move single qubits, transfer qubits between tweezers:
no loss, no decoherence
Please see disclaimer
at beginning
slides
(in Jessen group:
Fidelity
(π ofpulse)
= 0.996(3))
Tweezer arrays: Birkl group, Darmstadt
single atoms,
separately
addressable and
detectable
microlens array
illuminated
uniformly with
one or multiple
beams or
individually
1
2
4
7
3
6
8
9
Shift register: atoms
are shifted by one array
of tweezers,
transferred to another,
then shifted again.
Please see disclaimer at beginning of slides
Another approach: magnets+tweezers
B0
Spreeuw group, U. Amsterdam
trapped atoms:
mK 87Rb
“atom chip”
(room
temperature)
hundreds of sites with
10-1000 atoms each
10 mm
Lithographically microfabricated array of
1250 traps/mm2
Magnetized film:
hard-disc material
shift-register applied to
tweezer-addressed defect
Please see disclaimer at beginning of slides
Entangling distant atoms in tweezers
or large-period lattices
Ground-state atoms have short range, weak
interactions—mainly contact interactions, so that
atoms need to be on the same site to become
entangled. For atoms at different sites, excitation to
high Rydberg states allows fast, distant entanglement.
V ~ d2/R3
R ~ several mm
R
A
B
Please
at beginning
Time to entangle
~see1disclaimer
/ V <<
1 msof slides
(for high Ry states)
Demonstrations of Ry Blockade
common excitation
Palaiseau: Grangier, Browaeys…
separate excitation
U. Wisc. Madison: Saffman, Walker,…
4 mm separation
10 mm separation
Note ~ 50 ns “gate time”
In both cases, the presence of an atom, some microns
distant, prevents the excitation of its neighbor, because
of the strong, long-range, Ry-Ry interaction energy.
Please see disclaimer at beginning of slides
Perspectives on Quantum Computing with cold atoms
• All the DiVincenzo Five criteria have been addressed. Need a lot
more fidelity and scaling. No obvious roadblocks.
• Need to put the pieces together in a scalable way.
• There is a long way to go before a “competent” quantum computer is
realized (e.g., factor numbers no other computer can factor) if ever.
(remember the Phillips 50/50 analysis)
•There are a lot of interesting things to do with quantum processing
that do not require a competent, general purpose computer:
• Quantum repeaters. This involves the DiVincenzo +2 criteria.
• Quantum devices—sub-shot-noise measurement.
• Testing quantum mechanics in large entangled systems.
Quantum/classical interface; e.g., can 400 qubits be entangled?
• Quantum simulation (we are doing it NOW!)
Please see disclaimer at beginning of slides
Quantum Simulation
Feynman legendarily noted that no classical
computer can efficiently simulate a quantum
system (although clever approximations abound).
• Direct calculation: a competent, error corrected quantum
computer will be able to do interesting quantum problems.
• More modest problems might be done without error correction
(even several tens of entangled objects are beyond the capability
of classical computers).
• The “native” Hamiltonians for cold atoms can realize models
that are too hard to solve mathematically.
• “New” Hamiltonians can be engineered to solve other models.
• Physical systems can be approximated with cold atoms, where
measurement and control capabilities provide advantages over
the original systems.
Please see disclaimer at beginning of slides
The native Hamiltonian for atoms in an
optical lattice is the Hubbard model of
condensed matter physics
t ... tunneling
U ... onsite interaction
With cold atoms, one can change
t/U in ways impossible in real
condensed matter systems. One
can undergo the Mott transition.
H  t b b  12 U  ni ni 1 mi
†
i j
i
i,j
Please see disclaimer
at beginning
of slides
Jaksch, Bruder, Cirac, Gardiner,
Zoller,
Phys.
Rev. Lett. 81, 3108 (1998)
The Mott Insulator Transition
“superfluid” phase
first observed with cold atoms
by I. Bloch et al. in Munich ‘02
“insulator” phase
deeper lattice
Precision measurement at NIST/
JQI (Spielman et al. 2008)
locates the 2D Bose-Hubbard
phase transition that agrees with
an-only-recently available high-
accuracy Quantum Monte Carlo
Please see disclaimer at beginning of slides
calculation.
Another quantum phase transition, in spinor BEC
calculated
phase diagram
1
0
0
0
B
m
1 0
Please see disclaimer at beginning of slides
(Lett group, NIST/JQI)
Engineered Hamiltonian:
a charge in a magnetic vector potential
r
r
p  eA

H

2
2M
(There are lots of other
Hamiltonians to engineer in
lots of ways.)
Why? Study the equivalent to quantum Hall and fractional quantum Hall effects in a
2-D system: an exotic state of matter that is easily controlled and measured.
How? One way is to rotate. In the rotating frame the Coriolis force mimics the
Lorentz Force, but
the centrifugal force causes problems.
Please see disclaimer at beginning of slides
quasimomentum (kphot)
Demonstration of a synthetic E-field
r
r
A
E 
t
Switch off the synthetic A-field, producing an impulse that kicks the atoms

before
kick
after
kick
mechanical momentum
Next: give the A-field a curl, to simulate a magnetic field;
look for a vortex lattice,
go to
the ofquantum
Hall regime.
Please see disclaimer
at beginning
slides
Cold atoms mimic a condensed system
Berezinskii-Kosterlitz-Thouless transition in 2-D
Dalibard group, ENS, interferes
two planes of atomic gas.
E. Cornell group, NIST/JILA
realizes the X-Y model of KT.
Helmerson
group, NIST/JQI
Single plane
of atoms Please see disclaimer at beginning of slides
Simulating a superconducting loop with cold atoms
Helmerson and colleagues @ JQI
without
plug
Plugged trap allows persistent current in a toroidal geometry
Next: a simulated Josephson junction
added to the circuit—atom SQUID.
•No Meissner effect
•Dynamically variable tunnel barrier
•Vary interactions
Please see disclaimer at beginning of slides
with plug
Something new: Atom-friendly,
multi-spatial-mode squeezed light
Hot atoms
Group of Paul Lett
at NIST/JQI
Please see disclaimer at beginning of slides
Something new: Atom-friendly,
multi-spatial-mode squeezed light
Hot atoms
Please see disclaimer at beginning of slides
Something new: Atom-friendly,
multi-spatial-mode squeezed light
• transfer non-classicality to
atoms
• continuous-variable quantum
communications
• faint image recovery
• supersensitive detection
• superresolution imaging
• high-density data storage
• position sensing
•parallel quantum data storage
• (storage of entangled images)
Please see disclaimer at beginning of slides
Perspectives on quantum simulation
Simulations with cold atoms hold the promise of learning things about manybody quantum systems that we have not been able to learn in other ways.
(Because the problems are too hard, or other systems are not amenable to
measurement and control, or similar systems simply do not exist.)
• Simulations of all kinds (native and non-native realizations of models, plus analogs of
CM systems) show great promise for important and interesting results.
• Important contributions from simulations are already in hand (e.g. accurate quantum
phase transitions, fresh look at 2-D superfluid transition)
• Synthetic magnetic fields should allow study of interesting quantum Hall physics,
topological matter.
• Fermi-Hubbard and similar studies may elucidate high-Tc superconductivity
• SQUID analogs could bring new insights into superfluid behavior and applications.
• Cold atoms allow a new look at the role of dimensionality in quantum systems.
• Getting lower temperature will be an important issue, with new approaches needed—
bosonic cooling of fermions, algorithmic cooling, using QI techniques.
• exotic spins systems (e.g.,frustrated systems, spin-glass, topological states,….)
• Control the dissipation/bath as well as the conservative Hamiltonian
•…
Please see disclaimer at beginning of slides
Final Thoughts on Simulations
Why cold-atom quantum simulation is interesting:
• Some problems may be intractable by other means.
• Interesting CM problems may be better addressed with particles
whose momentum is easily measured, whose Hamiltonians and
states are easily manipulated, etc.
• Other kinds of physics (high-energy, cosmology,…) may have
useful simulations with cold atoms in lattices or otherwise.
• Things not seen in “Nature” can be easy in cold-atom system, and
may teach us interesting things (e.g., certain kinds of quasiperiodic
structures, fermion/boson choices, interaction choices,…)
Please see disclaimer at beginning of slides
The End
Please see disclaimer at beginning of slides
Some thoughts about the pursuit of research in quantum information
I believe it is likely that a “really” large scale quantum computer (one that can factor
numbers that cannot otherwise be factored) will be hybrid—to a larger degree than
Lukin discussed in the context of NV centers. It would be a mistake to discount
certain QC platforms at this stage, and probably for quite some time, as they might
be key contributors to some aspect of a QI system. It would be a mistake to only
support the apparent function of a particular platform that appears to be best
adapted to it—Nature is perverse and experimenters are both lucky and clever.
Corollary: Don’t think too hard about an experiment before you do it—you may
become discouraged because you cannot imagine how clever and lucky you will
be when the problem you anticipate is actually at hand.
Stable funding: don’t dismiss the multiple appeals for stable funding as whining on
the part of researchers who don’t like being cut. It makes a big difference in the
ability to accomplish missions. Jerking people around with funding is a way to
alienate some of the best people and to select in favor of mediocrity. QC in
particular is a long-haul proposition. It is not the Manhattan project, and even if it
could be accomplished by assigning a tremendous amount of money and people,
that would probably not be the best way to do it. That doesn’t mean it is not very
important. Nor do I mean that unproductive research should have continued
Please could
see disclaimer
at beginning
of slides of the mission.
funding. But volatility in funding
insure
the failure
Some thoughts about the pursuit of research in quantum information
Diversity of funding styles at different agencies is one of the great strengths of US
science, and one that will serve the development of quantum information very well.
The styles of places like NSF, DARPA, ONR,.. are very different and that is a good
thing. The success of any one of those is not in itself a good reason for others to
copy that style. The differences themselves have high value intrinsically. It is one
of the things that helps to ensure that a brilliant, innovative idea will find a home.
There are some agencies where a single program director can decide that
something should be funded, and others where half a dozen outside reviewers
have to rank a proposal as “excellent” for it to be funded. We are well served by
having both kinds, particularly with QI, which is important and has important
aspects that are not at all well-understood.
Please see disclaimer at beginning of slides