Download Quantum Computer Hardware - Trapped Ion Quantum Information

Quantum Computer
Hardware
Chris
Monroe
University of Maryland
Department of Physics
National Institute of
Standards and Technology
Quantum Mechanics and Computing
molecular-sized
transistors
2025
atom-sized
transistors
2040
“There's Plenty of Room at the Bottom” (1959)
Richard Feynman
“When we get to the very, very small world – say circuits of seven
atoms – we have a lot of new things that would happen that
represent completely new opportunities for design.
Atoms on a small scale behave like nothing on a large scale,
for they satisfy the laws of quantum mechanics…”
A new science for the 21st Century
Quantum
Mechanics
20th Century
Information
0 1 1 0 0 0 1 1…
Theory
𝝏|𝝋
𝒊ℏ
= 𝑯|𝝋
𝝏𝒕
21st Century
Quantum Information Science
𝜶|𝟎 + 𝜷|𝟏
Computer Science and Information Theory
Charles Babbage (1791-1871)
mechanical difference engine
Alan Turing (1912-1954)
universal computing machines
Claude Shannon (1916-2001)
quantify information: the bit
k
H   pi log 2 pi
i 1
ENIAC
(1946)
The first solid-state transistor
(Bardeen, Brattain & Shockley, 1947)
The classical NAND Gate
A
B
out
0
0
1
0
1
1
1
0
1
1
1
0
V0
out
A
B
32-level NAND-based flash memory
The Golden Rules
of Quantum Mechanics
𝝏|𝝋
𝒊ℏ
= 𝑯|𝝋
𝝏𝒕
Rule #1: Quantum objects are waves
and can be in superposition
qubit: 𝜑 = 𝛼 0 + 𝛽|1
𝑖ℏ
𝑑𝛼
= 0 𝐻 0 𝛼+ 0𝐻 0 𝛽
𝑑𝑡
𝑖ℏ
𝑑𝛽
= 1𝐻 0 𝛼+ 1𝐻 1 𝛽
𝑑𝑡
Rule #2: Rule #1 holds as long as you don’t look!
𝜑 = 𝛼 0 + 𝛽|1
|0
probability
𝑝 = |𝛼|2
or
|1
1 − 𝑝 = |𝛽|2
GOOD NEWS…
quantum parallel processing on 2N inputs
Example: N=3 qubits
 = a0 |000 + a1|001 + a2 |010 + a3 |011
a4 |100 + a5|101 + a6 |110 + a7 |111
f(x)
N=300 qubits: more information
than particles in the universe!
…BAD NEWS…
Measurement gives random result
e.g.,   |101
f(x)
…GOOD NEWS!
quantum interference
depends on
all inputs
…GOOD NEWS!
quantum interference
quantum
logic gates
depends on
all inputs
quantum |0  |0 + |1
NOT gate: |1  |1  |0
quantum |0 |0  |0 |0
XOR gate: |0 |1  |0 |1
|1 |0  |1 |1
|1 |1  |1 |0
e.g., (|0 + |1) |0  |0|0 + |1|1
superposition  entanglement
Quantum State: |0|0 + |1|1
John Bell (1964)
Any possible “completion” to
quantum mechanics will violate
local realism just the same
Citations to John Bell’s 1964 paper
J. Bell, "On the Einstein Podolsky Rosen Paradox," Physics 1, 195 (1964)
800
700
600
500
400
300
200
100
0
1970
1980
1990
2000
2010
500
50
5
1970
1980
1990
2000
2010
Moore’s Law of Publishing
# articles mentioning “Quantum Information”
or “Quantum Computing”
Nature
Science
Phys. Rev. Lett.
Phys. Rev.
3000
2500
2000
Quantum
Computers
and Computing
1500
Institute of
Computer Science
Russian Academy
of Science
1000
ISSN 1607-9817
500
0
1990
1995
2000
Shor’s Quantum
Factoring Algorithm
2005
2010
(Classical) Error-correction
Shannon (1948)
Redundant encoding to protect against (rare) errors
potential error: bit flip
0/1
0/1
1/0
potential error: bit flip
000/111
p(error) = p
000/111
010/101 etc..
take majority
𝑝(𝑒𝑟𝑟𝑜𝑟) = 3𝑝2 1 − 𝑝 + 𝑝3
𝑝 → 3𝑝2 1 − 𝑝 + 𝑝3
better off whenever p < 1/2
Quantum error-correction
Shor (1995)
Steane (1996)
𝜌=
𝛼|0 + 𝛽|1
𝑃0
𝐶∗
𝐶
𝑃1
Decoherence
+|00000
𝛼 +|01001
𝛼|0 + 𝛽|1 ⟹
4 −|11101
−|10001
+ |10010
− |11011
− |00011
− |01100
+ |01001
− |00110
− |11110
− |10111
+ |10100
− |11000
− |01111
+ |00101
+|11111
𝛽 +|10101
+
4 −|00010
−|01110
+ |01101
− |00100
− |11100
− |10011
+ |10110
− |11001
− |00001
− |01000
+ |01011
− |00111
− |10000
+ |11010
5-qubit code
corrects all
1-qubit errors
to first order
N=1
N=1028
Trapped Atomic Ions
Aarhus
Amherst
Basel
Berkeley
Bonn
Citadel
Clemson
Denison
Duke
Erlangen
ETH-Zurich
Freiburg
Georgia Tech
Griffith
Hannover
Honeywell
Indiana
Innsbruck
Lincoln Labs
Lockheed
Maryland/JQI
Mainz
MIT
Munich
NIST-Boulder
Northwestern
NPL-Teddington
Osaka
Oxford
Paris
Pretoria
PTB-Braunschweig
Saarbrucken
Sandia
Siegen
Simon Fraser
Singapore
Sussex
Sydney
Tokyo
Tsinghua-Beijing
UCLA
Washington-Seattle
Weizmann
Williams
Yb+ crystal
~5 mm
171Yb+
2S
1/2
| = |1,0
| = |0,0
hyperfine qubit
wHF/2p = 12 642 812 118 + 311B2 Hz
(600 Hz/G @ 1 G)
171Yb+
qubit detection
g/2p = 20 MHz
2P
1/2
2.1 GHz
Probability
1
|z
0
369 nm
2S
1/2
0
5
10
15
20
# photons collected in 800 ms
|
|
wHF/2p = 12 642 812 118 + 311B2 Hz
(600 Hz/G @ 1 G)
25
171Yb+
qubit detection
g/2p = 20 MHz
2P
1/2
2.1 GHz
Probability
1
|z
0
369 nm
2S
1/2
>99%
detection
efficiency
0
5
10
15
20
# photons collected in 500 ms
|
|
wHF/2p = 12 642 812 118 + 311B2 Hz
(600 Hz/G @ 1 G)
25
171Yb+
2P
3/2
2P
1/2
qubit manipulation
g/2p = 20 MHz
D = 33 THz
355 nm (10 psec @ 100 MHz)
2S
1/2
|
|
wHF/2p = 12 642 812 118 + 311B2 Hz
(600 Hz/G @ 1 G)
Combination of coherence and perfect measurement
:
increment
t
t
prepare
↓
measure
P(↑)
(bright or dark)
laser
beams
:
Prob(↑|↓)
1
0.8
averaged
data
0.6
0.4
0.2
0
.
0
50
100
150
200
t (ms)
250
300
350
400
Entangling Trapped Ion Qubits
r




“dipole-dipole coupling”
|↓↓
|↓↑
|↑↓
|↑↑
~5 mm
d
d ~ 10 nm
ed ~ 500 Debye
𝑒2
𝑒𝛿 2
∆𝐸 =
−
≈−
2
2
𝑟
2𝑟 3
𝑟 +𝛿
→
|↓↓
→ 𝑒 −𝑖𝜑 |↓↑
→ 𝑒 −𝑖𝜑 |↑↓
→
|↑↑
𝑒2
𝜋
∆𝐸𝑡 𝑒 2 𝛿 2 𝑡
=
𝜑=
=
2
ℏ
2ℏ𝑟 3
for full
entanglement
Cirac and Zoller (1995)
Mølmer & Sørensen (1999)
Programmable Quantum Computer… in the lab
High NA
objective
355nm
Raman beams
Dk
•
•
•
•
5-segment linear Paul trap
High NA objective (0.37)
Tightly focused Raman beams
32ch AOM and PMT for indiv.
addressing/detection
355nm
pulsed laser
Harris Corp
32channel AOM
2μm pixels
Diffractive
optic (х10)
Quantum Fourier Transform (QFT)
𝑦𝑘 =
output
amplitudes
1
𝑁−1
𝑁 𝑗=0
𝑒
2𝜋𝑖
𝑗𝑘
𝑁 𝑥𝑗
𝑁 = 2𝑛
input
amplitudes
QFT circuit (n=5 qubits)
controlled phase gate
Controlled-Phase Gate
Controlled phase gate
𝛼 = 𝑠𝑖𝑔𝑛 𝐽𝑖𝑗
𝛽 = 𝑠𝑖𝑔𝑛 𝜃
± phase of
Ising coupling
Quantum Fourier Transform (QFT)
state preparation
e.g. state with period 8 =
results
7
15
23
31
Physics: global spin-dependent force
F = F0|↑↑|  F0|↓↓|
Physics: global spin-dependent force
|
|
ADD: Independent spin flips
↑ ↓ ↑ ↓ ↑ ↓ ↑ ↓ ↑↓↑ ↓↑ ↓↑ ↓↑ ↓ ↑ ↓ ↑ ↓ ↑ ↓
↓ ↑ ↓ ↑ ↓ ↑ ↓ ↑↓ ↑↓ ↑↓ ↑↓↑ ↓ ↑ ↓ ↑ ↓ ↑ ↓ ↑
F = F0|↑↑|  F0|↓↓|
B
Adiabatic Quantum Simulation
from S. Lloyd, Science 319, 1209 (2008)
Transverse 𝐻 =
Ising model
(𝑖) (𝑗)
(𝑖)
𝐽𝑖,𝑗 𝜎𝑧 𝜎𝑧 + 𝐵
𝑖<𝑗
𝜎𝑥
𝑖
(𝑖)
𝐵
𝜎𝑥
𝑖
Initialization:
spins along x
Detection:
measure spins
along z
(𝑖) (𝑗)
𝐽𝑖,𝑗 𝜎𝑧 𝜎𝑧
𝑖<𝑗
Time (<10 msec)
Antiferromagnetic Néel order of N=10 spins
All in state 
2600 runs, a=1.12
All in state 
AFM ground state order
222 events
219 events
441 events out of 2600 = 17%
Prob of any state at random =2 x (1/210) = 0.2%
First Excited States
(Pop. ~2% each)
Second Excited States
(Pop. ~1% each)
AFM order of N=14 spins (16,384 configurations)
N=22 spins
𝐻𝑋𝑌 =
𝑖<𝑗
𝐽0
𝑖−𝑗
𝑗
𝛼
𝑗
𝜎𝑥𝑖 𝜎𝑥 + 𝜎𝑦𝑖 𝜎𝑦
a  0.6
initialmeasured
state at t=0
state
at J0t = 36
B. Neyenhuis et al., in preparation (2015)
Medium scale vision (>100 atomic spins)
a (C.O.M.)
b (stretch)
c (Egyptian)
-15
-10
-5
d
c
axial modes only
20
b+c
a+b
b
a
carrier
2a
a
40
2b,a+c
c-a
d (stretch-2)
b-a
b-a
2b,a+c
d
a+b
c
2a
b
c-a
60
b+c
Fluorescence counts
Mode competition –
example: axial modes, N = 4 ions
0
5
10
Raman Detuning dR (MHz)
Kielpinski, Monroe, Wineland, Nature 417, 709 (2002)
15
mode
amplitudes
Univ. of
Maryland
Boulder
Mapping qubits from atoms to photons
Given photon is collected
2P
1/2
|𝝋 = |↓ |𝑩 +|↑ |𝑹
171Yb+
“post-selected”
R
B
2S
1/2
|
|
success probability
𝑑Ω
𝑝=
𝐶 𝜂𝐷 = 0.10 0.20 0.25
4𝜋
= 0.005
Doubling down: remote link through photons
|↓ 𝟏 |𝑩 𝟏 +|↑ 𝟏 |𝑹
𝟏
|↓ 𝟐 |𝑩 𝟐 +|↑ 𝟐 |𝑹
𝟐
⇒ |↓ 𝟏 |↑
𝟐
− |↑ 𝟏 |↓
𝟐
Upon coincidence detection!
l/4
50/50
BS
state of the art:
l/4
1 2
𝑅 = 𝑅𝑝
2
optical
fiber
= 10/𝑠𝑒𝑐
171Yb+
ion
171Yb+
Simon & Irvine, PRL 91, 110405 (2003)
L.-M. Duan, et. al., QIC 4, 165 (2004)
Y. L. Lim, et al., PRL 95, 030505 (2005)
D. Moehring et al., Nature 449, 68 (2007)
ion
𝑝 = 0.005
R = 1 MHz
D. Hucul, et al., Nature Phys. 11, 37 (2015)
Quantum teleportation
of a single atom
unknown qubit
uploaded to
atom #1
a| + |
qubit transfered to
atom #2
a| & |
S. Olmschenk et al., Science 323, 486 (2009).
we need
more time..
and more
qubits..
Large scale modular Architecture (103 - 106 atomic spins?)
0.001 Hz before
~10 Hz now
~1 kHz soon
CM et al., Phys. Rev. A 89, 022317 (2014)
D. Hucul, et al., Nature Phys. 11, 37 (2015)
1947: first transistor
2000: integrated circuit
single module
N ion trap modules
2015: qubit collection
Large scale quantum network?
Implementation of Quantum Hardware
•
•
•
•
control &
configurability
quantum materials by design
complex optimization
“big quantum data”
quantum computing
Verification?
trapped
ions
superconductors
molecules
NV
Q-dots
neutral
atoms
# particles
Leading Quantum Computer Hardware Candidates
Trapped Atomic Ions
individual
atoms
lasers
photon
Atomic qubits connected through
laser forces on motion or photons
Superconducting Circuits
Superconducting qubit:
right or left current
Others: still exploratory
FEATURES & STATE-OF-ART
• very long (>>1 sec) memory
• 5-20 qubits demonstrated
• atomic qubits all identical
• connections reconfigurable
Investments:
IARPA
GTRI
Sandia
FEATURES & STATE-OF-ART
• connected with wires
• fast gates
• 5-10 qubits demonstrated
• printable circuits and VLSI
LARGE
Investments:
CHALLENGES
• lasers & optics
• high vacuum
• 4K cryogenics
• engineering needed
Lockheed
UK Gov’t
CHALLENGES
• short (10-6 sec) memory
• 0.05K cryogenics
• all qubits different
• not reconfigurable
Google/UCSB IBM
Lincoln Labs Intel/Delft
• NV-Diamond
• Semiconductor quantum dots
• Atoms in optical lattices
D-Wave: superconducting circuits
venture capital funding
great advertising
but is it quantum?
N=1
N=1028
Trapped Ion Quantum Information
www.iontrap.umd.edu
Grad Students
Res. Scientists
David Campos
Jonathan Mizrahi
Clay Crocker
Kai Hudek
Shantanu Debnath
Marko Cetina
Caroline Figgatt
David Hucul (UCLA)
Volkan Inlek
Kevn Landsman
Aaron Lee
Kale Johnson
Harvey Kaplan
Antonis Kyprianidis
Lexi Parsagian
Chris Rickerd
Crystal Senko ( Harvard)
Ksenia Sosnova
Jake Smith
Ken Wright
Undergrads
Eric Birckelbaw
Kate Collins
Micah Hernandez
Postdocs
Paul Hess
Marty Lichtman
Norbert Linke
Brian Neyenhuis ( Lockheed)
Guido Pagano
Phil Richerme ( Indiana)
Grahame Vittorini ( Honeywell)
Jiehang Zhang
Collaborators
Luming Duan (Michigan)
Philip Hauke (Innsbruck)
David Huse (Princeton)
LPS/NSA
Alexey Gorshkov (JQI/NIST)
Alex Retzker (Hebrew U)
ARO
Quantum Superposition
From Taking the Quantum Leap, by Fred Alan Wolf
Quantum Superposition
From Taking the Quantum Leap, by Fred Alan Wolf
Quantum Superposition
From Taking the Quantum Leap, by Fred Alan Wolf
Quantum Entanglement
“Spooky action-at-a-distance”
(A. Einstein)
From Taking the Quantum Leap, by Fred Alan Wolf
Quantum Entanglement
“Spooky action-at-a-distance”
(A. Einstein)
From Taking the Quantum Leap, by Fred Alan Wolf
Quantum Entanglement
“Spooky action-at-a-distance”
(A. Einstein)
From Taking the Quantum Leap, by Fred Alan Wolf
Quantum Entanglement
“Spooky action-at-a-distance”
(A. Einstein)
From Taking the Quantum Leap, by Fred Alan Wolf