Quantum Computation with Superconducting Quantum Devices T.P. Orlando, S. Lloyd, L. Levitov, J.E. Mooij - MIT

M. Tinkham – Harvard; M. Bocko, M. Feldman – U. of Rochester orlando@mit.edu web.mit.edu/superconductivity

Objective: To use superconducting loops and Josephson junctions

1. To model the measurement process, understand decoherence, and to develop scalable algorithms,

2. To combine these qubits with classical on-chip, high-speed superconducting control electronics,

3. To implement the fabrication and testing of the superconducting qubits.

Objective Approach:Theory: To understand the measurement and control processes, develop algorithms and guide the experimental design and testing.

Circuits: To design, analyze and demonstrate superconducting circuitry for the on-chip input and the required control functions for qubit manipulation.

Implementation: To test and analyze results from each integration step; oversee fabrication and improve junction quality.

Status: 1. Measurements of the two states in a Nb qubit with

0.45m junctions an underdamped Nb dc-SQUID :• Energy landscape determined from thermal

activation measurements for T> 300mK• A Q factor of 106 which agrees with

measurements of the Rsubgap > 1 M2. Al qubits: Measured relaxation time ~ 1 s

3. SFQ components (delay lines, DC/SFQ, T-flip-flops) measured at low current density and low temperature.

4. Modeling the environmental coupling to the qubit and the measurement process

5. Scalable architecture for adiabatic quantum computing

7/28/02

qubit&

readout

5 m

(Put collaborative UR/MIT experiment here)

Put 30mk data here

Participants and CollaboratorsMITSeth Lloyd: Lin Tian, Bill KaminskyLeonid Levitov;Terry Orlando: Ken Segall Donald Crankshaw Daniel Nakada, Janice Lee Bhuwan Singh, David Berns

TU DELFTJohan Mooij & Kees Harmans Alexander ter Haar

MIT Lincoln LaboratoryKarl Berggren & Jay Sage

Harvard UniversityMichael Tinkham: Nina Markovic, Sergio Valenzuela

University of RochesterMark Bocko & Marc Feldman Jon Habif, Pavel Rott Xingxiang Zhou Gui-Zhen Zhang, Michael Wulf

This work is supported in part by the AFOSR grant F49620-01-1-0457 under the DoD University Research Initiative on Nanotechnology (DURINT) program and ARDA, and in part by the AFOSR/NM and also by the NSA and ARDA under ARO grant number DAAG55-998-1-0369. The Type II computing is funded by AFOSR/NM.

University of MunichFrank Wilhelm: Markus Storcz

AFRL

Jeff Yepez

Collaborations

• Lincoln Laboratory (fabrication and type-II computing)

• Delft (off-chip experiments, Al qubits, tight collaboration, theory)

• TRW (fabrication source)• MIT (MIT/Cambridge Consortium, NSF Center,

Type II computing)• Univ. Munich (Frank Wilhelm) Theory • AFRL (Yepez) Type II computing• SQUBIT European Project

Outline

• General Overview (Terry Orlando)• Introduction

• Highlights of recent results

• Future work

• Implementation Review (Ken Segall)

• Circuits Review (Marc Feldman)

• Experiments from Delft (Kees Harmans)

• Theory Review (Seth Lloyd)

Persistent Current QubitThis qubit design uses a superconducting loop interrupted by three Josephson Junctions.The two lowest energy states, which serve as the |0> and |1> states of the qubit, have circulating currents in opposite directions, with opposite magnetic fields of ~0.001 0.

Rotating the qubit will require flux oscillations at the frequency of the energy difference.The Rabi frequency depends on the magnitude of the flux oscillations.

ƒ1

1 2

j3

j2j1

|0

|1>

current

Ici

r

Ei

ext (0)0.5

v

Quantum Computation with Superconducting Quantum Devices T.P. Orlando, S. Lloyd, L. Levitov, J.E. Mooij - MIT

M. Tinkham – Harvard; M. Bocko, M. Feldman – U. of Rochesterin collaboration with K. Berggren, MIT Lincoln Laboratory 7/28/02

qubit&

readout

5 m

Fabricationmodeling, and measurements

Ic CJ

Ib

CsZ0qubit

2

~1

~

• Persistent current qubit fabricated in Nb with submicron junctions

• Two states seen in measurement (thermal activations and energy levels)

1pF 1pF

0.45m

1.1m

0.55m

1.1m

I- V-

I+ V+

Thermal Activation Theory

1

kTU

ekT

UQ

2

2.7 0

rate

f

T

f ssln

,

Condition for <Isw> = 0

Thermal rate with damping(Energy diffusion regime)

02

12.4 UfEU J Energy barrier linear in flux

EJ = 4200 eV

Q = 2x106

• EJ indicates junctions are small (0.55 m)• Q suggests long relaxation times (T1 ~ Q/0 ~ 4 s)

Lincoln Lab: Rsubgap measurement

Delft University of Technology & DIMES The Netherlands MIT CambridgeCaspar van der Wal, A. Ter Haar, Kees Harmans, Hans Mooij T. Orlando, L. Levitov, S. Lloyd

Macroscopic quantum superposition in a Josephson junction loop

MIT

T U D elftTechnische U nivers ite it D elft

0.498 0.500 0.502

~

9.711

8.650

6.985

5.895

4.344

3.208

2.013

1.437

1.120

0.850

SW

(0

.4 n

A p

er

div

isio

n)

ext

Ipc

3 m

• Superposition of states observed• Relaxation time 5 sec, •Dephasing time 0.1 sec

0 1000 2000 3000 4000 5000 6000 70000.0

0.2

0.4

0.6

0.8

1.0

0.5010 0.5012

a)

8.42 GHz 11.39 GHz 18.96 GHz

Dip

Am

plitu

de (

a.u)

Delay Time (ns)

b)11.39 GHz

4200 ns

4700 ns

3700 ns

3200 ns

2700 ns

1200 ns

1700 ns

2200 ns

700 ns

200 ns

ext

(0)

<Isw

> -

linea

r tr

end

(

a.u)

Relaxation time

Measured relaxation time ~ 1 s

New Slide(s) from Delft

SFQ Results on QC2

• Inductance measurement - it’s exactly right

• Tested (4.2 K) analog and digital devices on LL fabricated chips (500 A/cm2 nominal)

– dc-SQUID coupled to large inductive loop

– Small junction I-V’s ~ 0.4 x 0.4 µm– RSFQ test circuit

• dc-sfq, JTL’s, confluence buffer, splitter, JTL clock ring, sfq-dc

Add UR logos, reference here

Test Results

Operation of the circuit on the previous page. Each cycle of the input waveform introduces one SFQ pulse to the circuit. The output flips its voltage state at each arriving pulse.

~350 bits/sec.

~3.5 kbits/sec.Add UR logos, reference here

On-chip Control for an RF-SQUIDM.J. Feldman, M.F. Bocko, Univ. of Rochester

Sources of Error in Superconducting Qubits

• Offset charge fluctuations• Quasiparticles Q > 104

• Bias current fluctuations

Decoherence from the environment (use error correction)

Dephasing sources (use “spin echo” techniques)• Coupling to nuclear spins• Diople-dipole coupling

Coherent error sources (use dynamic pulse control)• Coupling to higher levels• Two-bit gate coupling

Lin Tian, L. Levitov, et al., “General Theory of Dephasing for the Qubit,” Quantum Mesoscopic Phenomena (2000)

Model of Measurement Induced Decoherence

Tk

tJ

TkJ

Tk

tJ

r

Beff

Beff

Beff

2coth)(

2

1

2coth)(

1

2coth

2

1)(

1

2

0

00

0

2

0

0

2

0

00

Spin-Boson Model gives

2)( effJ

Ic CJ

I b

Cs Z0qubit

1~ 2

~

Lin Tian, Seth Lloyd, T. Orlando PRB (2001)

More theory slide

• Adiabatic QC

Type II Quantum Computing:1-D Algorithm

 

Ψ1,a Ψ2,a Ψ3,a Ψ4,a Ψ5,a Ψ6,a

· · · · · · Ψ1,b Ψ2,b Ψ3,b Ψ4,b Ψ5,b Ψ6,b

Φ1 Φ2 Φ3 Φ4 Φ5 Φ6

· · · · · ·

Initialize

Collide

Measure

Stream* P’1a ← P’2a←P’3a← P’4a ← P’5a← P’6a

P’1b → P’2b →P’3b→ P’4b →P’5b →P’6b

P’1a P’2a P’3a P’4a P’5a P’6a

· · · · · · P’1b P’2b P’3b P’4b P’5b P’6b

101 PP where P is occupancy probability

Measurement

Imeas bias

Iqubit bias

f1 f2

Vosc biasfosc

• Progress on last year’s objectives─ Measurements of the two states in a Nb qubit with 0.45m junctions and

underdamped Nb dc-SQUID : Energy landscape determined from thermal activation measurements for T> 300mK, and a Q factor of 106 and Rsubgap > 1M.

─ SFQ devices at 300 mK and for current densities < 200 Amps/cm2

─ Al qubits: Measured relaxation time ~ 1 ms─ Scalable architecture for adiabatic quantum computing with superconductors Research plan for the next 12 months─ Measurement of on-chip spectroscopy of a single qubit─ On-chip timed oscillator control of a single qubit ─ Spectroscopy of two-coupled qubits─ Resonance method of measurement of the state of the qubit (with Delft)─ Set up Dilution Refrigerators─ Theory here Long term objectives (demonstrations)- Combine 3 to 5 superconducting qubits with on-chip control electronics- Measure decoherence in multiple-qubit systems- Develop algorithms adapted to superconducting electronics- Explore quantum control to correct qubit dynamics

Quantum Computation with Superconducting Quantum Devices T.P. Orlando, S. Lloyd, L. Levitov, J.E. Mooij - MIT

M. Tinkham – Harvard; M. Bocko, M. Feldman – U. of Rochester orlando@mit.edu web.mit.edu/superconductivity

7/28/02

Summary Slides of Results, Circuits, and Publications

Not to be presented

Results to Date1. Implementation:

• Subgap resistance of submicron Nb junctions > 1 Mat low temperatures

• LL Resistors remain at 30 mK • Measurements of the two states in a Nb qubit with 0.45mm

junctions an underdamped Nb dc-SQUID : Energy landscape determined from thermal activation

measurements for T> 300mK A Q factor of 106 which agrees with measurements of the

Rgap > 1 M.• Delft Experiments: Spectroscopy of superposition states• Developing of gradiometer qubits to lessen flux noise • Experiments on decoherence times and noise (Delft)• Installation of Dilution Refrigerators underway at MIT and UR

2. Circuits:

• SFQ T-Flip-Flop demonstrated at 300 mK and for current densities < 200 Amps/cm2

• Demonstration of Flip-Chip inductive coupling • On-chip coupling of JJ Oscillator• Design of MQC experiments on-chip• Developing resonant measurement scheme• Other results here

3. Theory:

• Theory of persistent current qubit• Calculation of intrinsic decoherence mechanisms and

sources of errors• Method to overcome off-resonant excitations • Modeling of decoherence of coupling and measuring

circuits- circuit model formulation• Modeling of measurement process with DC SQUID• Exploration of coupling schemes for qubits• Scalable architecture for adiabatic quantum computing

with superconducting

I. Circuits and Components That Have Already Been Tested

A. Simple Control Circuits:

1. On-chip DC-SQUID oscillators have been tested and sufficient inductive coupling to another circuits has been demonstrated. These oscillators however only operated around 3 GHz, so oscillators with variable frequency are now being fabricated. (MIT)

2. Demonstration of inductive coupling between separate chips for use in coupling qubits with control circuits fabricated on different chips (Lincoln).

3. Theoretical modeling of the effect of these simple control circuits on the decoherence of the qubit was included in the designs of the oscillators and measuring system.(MIT/Delft/Rochester)

4. Test at 4.2K of an RF SQUID coupled to a superconductive comparator with readout to room temperature (Lincoln).

5. Fixed-current superconducting loops, for magnetic flux biasing (Rochester)

B. Complex Circuit and components

1. The following components have been designed, fabricated,and tested at 4.2 K.

a. DC/SFQ and SFQ/DC converters (Rochester)

b. DRO memory cells (Rochester)

c. T-Flip-flops (Rochester)

d. Chains of up to 16 T-Flip-flops as counters. (Rochester)

e. SFQ clocks (pulse oscillators) of fixed frequencies designed from 5 to 40 GHz. (Rochester)

f. Pulse splitters and combining buffers (Rochester)

II. Types of Circuits and components that are being fabricated on QC3 (scheduled for completion in later this year.)

A. Simple Circuits:

1. On-chip DC-SQUID oscillators to work in the 5-15 GHz regime (some connected to detectors and some to qubits to do on-chip spectroscopy) (MIT)

2. On-chip SFQ microwave oscillator to work at 8 GHZ regime. (some connected to detectors and some to qubits to do on-chip spectroscopy) (Rochester/MIT)

3. A qubit coupled inductively to a coplanar waveguide. Using an external microwave generator, operating at 1-20GHz, it is possible to map the energy separation between the lowest two energy levels. (Lincoln and MIT)

B. Complex Experiments

List experiments on QC3 and explain briefly whose circuit and why the circuit is important.

1. An NDRO memory cell, similar to a DRO cell but with a non-destructive read-out, is being fabricated. Asuccessful test will allow the timed oscillator experiment (Rochester)

2. Timed oscillator experiment -- by using two out-of-phase counter and an NDRO memory cell, we can make a variable duty cycle oscillator to drive a qubit with a SQUID detector controlled off-chip. (Rochester)

3. Qubit readout experiments

a. QFP Comparators coupled with varying strengths to RF SQUID qubits (Lincoln)

b. QFP Comparators coupled to persistent-current qubits (Lincoln)

c. SFQ Comparators coupled to rf SQUID qubits and to testlines (Rochester)

C. Full Quantum Experiments Circuits (more-than-complex circuits)

1. Superposition-state time evolution experiment using rf SQUID qubit. This design has all-SFQ inputs and outputs, all on-chip; but off-chip timing. (Rochester)

2. Superposition-state time evolution experiment using single-Josephson-junction qubit (inspired by Martinis and Han experiments), with on-chip SFQ control circuits. Specifically designed to be portable to TRW fabrication. (Rochester)

Publications1. Design of Persistent Current Qubit J. E. Mooij, et al., Science, 285, 1036, (1999) T. P. Orlando, et al., PRB 60, 15398 (1999)

2. General Theory of Dephasing for the Qubit Lin Tian, L. Levitov, et al., in Quantum Mesoscopic Phenomena (2000)

3. Pulse Scheme to Decouple Higher Levels Lin Tian and S. Lloyd, PRA 62, 50301 (2000)

4. Measurements of the Qubit Energy Levels C. van der Wal, C. Harmans, J. E. Mooij, et al. Science 290, 773, (2000)

5. Inductance Effects on the Qubits D. Crankshaw, E. Trias, et al. IEEE Trans. Applied Supercond. 11, 1223, (2001)

6. Fabrication of Nb Qubits and Circuits K. Berggren, D. Nakada, et al. Proceedings of the International Conference on Experimental Methods in Quantum Computation, 2001. Rinton Press.

7. Modeling of the Measurement Process C. van der Wal, F. Wilhelm, et al., to be published Lin Tian, S. Lloyd, and T. Orlando, et al., to be published D. Crankshaw, et al., to be published