1 of 57

Multi-Mode Quantum Memories for High-Throughput Satellite Entanglement Distribution

Connor Casey

University of Massachusetts Amherst | Department of Physics

2 October 2025

136

28

27

191

0

255

2 of 57

Meet the Team

2

Eugene

Rotherham

Catherine

McCaffrey

Connor

Casey

Albert

Williams

Nathan

Darby

136

28

27

191

0

255

3 of 57

Scope of Talk

3

Pillars of Quantum Science & Technology
Motivation & Objectives
Why Use Satellites in Quantum Networking?
Motivating The Need for Quantum Memories
Our Proposed System Architecture
Memory Performance & Results
Conclusion & Outlook

136

28

27

191

0

255

4 of 57

I. Pillars of Quantum Science & Technology

136

28

27

191

0

255

5 of 57

Quantum Science & Technology Pillars

5

The Physics of the Quantum Internet. CQN Winter School. 2024

Quantum Computing

Optimization
Designing Molecules
Materials Design
Pattern Recognition
Machine Learning
Decryption

Quantum Sensing

Magnetic Fields
Gravitational Fields
Biomedical Imaging
Distributed Sensing

Quantum Networking

Remote Quantum Computing
Distributed Quantum Computing
Distributed Sensing
Multipart Entangled Protocols

136

28

27

191

0

255

6 of 57

Quantum Science & Technology Pillars

6

The Physics of the Quantum Internet. CQN Winter School. 2024

Quantum Computing

Optimization
Designing Molecules
Materials Design
Pattern Recognition
Machine Learning
Decryption

Quantum Sensing

Magnetic Fields
Gravitational Fields
Biomedical Imaging
Distributed Sensing

Quantum Networking

Remote Quantum Computing
Distributed Quantum Computing
Distributed Sensing
Multipart Entangled Protocols

Quantum Networking Enables and Links Together Diverse Quantum Technologies!

136

28

27

191

0

255

7 of 57

II. Motivation & Objectives

136

28

27

191

0

255

8 of 57

Motivation & Overview

8

Challenge

Global quantum networks using satellites are a key pathway to secure communication and distributed quantum computing. However, current satellite-based systems face two bottlenecks:

Single-mode quantum memories that restrict throughput to one photon mode at a time.
Short coherence times of practical platforms, which limit reliable storage and retrieval over orbital timescales.

136

28

27

191

0

255

9 of 57

Motivation & Overview

9

What We Propose

We propose a hybrid alkali–noble-gas multimode quantum memory architecture which enables:

Multimode storage of entangled photons across spectral and temporal modes.
Extended coherence times from minutes to hours, sufficient for satellite passes.
Compatibility with near-term satellite technology, making scalable global entanglement distribution feasible without prohibitive hardware costs.

136

28

27

191

0

255

10 of 57

Study Objectives

10

This Work

Building on physics defined in previous terrestrial studies, we model the memory dynamics in the environment of space and evaluate the feasibility of realizing this architecture on LEO Satellites.

Key Metrics

Link Establishment Probability
Memory Efficiency
SKR

136

28

27

191

0

255

11 of 57

III. Why Use Satellites in Quantum Networking?

136

28

27

191

0

255

12 of 57

Transmitting Qubits Terrestrially

12

Bob

136

28

27

191

0

255

13 of 57

Transmitting Qubits Terrestrially

13

Photon Loss

: length of optical-fiber path between A and B

: fiber loss coefficient (depends on fiber quality)

No-Cloning Theorem

QM Prevents Amplification of Quantum States

Bob

136

28

27

191

0

255

14 of 57

Transmitting Qubits Terrestrially

14

Photon Loss

: length of optical-fiber path between A and B

: fiber loss coefficient (depends on fiber quality)

No-Cloning Theorem

QM Prevents Amplification of Quantum States

We Need Something to Combat This…

Bob

136

28

27

191

0

255

15 of 57

Transmitting Qubits Terrestrially

15

Bob

136

28

27

191

0

255

16 of 57

Loss Translates to Achievable Rate

16

A

B

Rate

136

28

27

191

0

255

17 of 57

Quantum Repeater Mechanics

17

Photon 1

Photon 2

Photon 3

Photon 4

Initial

Alice

Bob

136

28

27

191

0

255

18 of 57

Quantum Repeater Mechanics

18

Initial

Entangle

Entangled

Photon 1

Photon 2

Photon 3

Photon 4

136

28

27

191

0

255

19 of 57

Quantum Repeater Mechanics

19

Initial

Entangle

BSM

Entangled

BSM

Entanglement Swap

Photon 1

Photon 2

Photon 3

Photon 4

136

28

27

191

0

255

20 of 57

Quantum Repeater Mechanics

20

Initial

Entangle

BSM

Entangled

BSM

Entangled A + B

Photon 1

Photon 2

Photon 3

Photon 4

136

28

27

191

0

255

21 of 57

21

All we need to do is encode quantum states into optical signals.

Why Photons?

Photons have multiple degrees of freedom suitable for quantum encoding (polarization, frequency, time-bin)
Signals travel at the speed of light
Optical components are mature
Compatible with telecom infrastructure

Fiber Optics → Free Space Optics (FSO)

136

28

27

191

0

255

22 of 57

A Satellite Perspective of this Process

22

Mustafa Gündoğan. Proposal for Space-borne Quantum Memories for Global Quantum networking . Nature Communications. 2021.

Bob

136

28

27

191

0

255

23 of 57

A Satellite Perspective of this Process

23

Mustafa Gündoğan. Proposal for Space-borne Quantum Memories for Global Quantum networking . Nature Communications. 2021.

- Uplink

- Downlink

Bob

136

28

27

191

0

255

24 of 57

A Satellite Perspective of this Process

24

Mustafa Gündoğan. Proposal for Space-borne Quantum Memories for Global Quantum networking . Nature Communications. 2021.

- Uplink

- Downlink

Bob

136

28

27

191

0

255

25 of 57

An Aside On Uplink vs Downlink

25

Mustafa Gündoğan. Proposal for Space-borne Quantum Memories for Global Quantum networking . Nature Communications. 2021.

136

28

27

191

0

255

26 of 57

An Aside On Uplink vs Downlink

TLDR; A downlink architecture achieves higher channel transmission because atmospheric losses occur at the end of transmission, and the ground station can host a much larger telescope.

26

136

28

27

191

0

255

27 of 57

A Satellite Perspective of this Process

27

Mustafa Gündoğan. Proposal for Space-borne Quantum Memories for Global Quantum networking . Nature Communications. 2021.

Bob

Now focusing on Quantum Memories!

136

28

27

191

0

255

28 of 57

IV. Motivating The Need for Quantum Memories

136

28

27

191

0

255

29 of 57

Mechanism of Transferring Qubits

29

Initial

OGS

136

28

27

191

0

255

30 of 57

Mechanism of Transferring Qubits

30

Initial

Entangle (EPS)

Entangled

OGS

136

28

27

191

0

255

31 of 57

Mechanism of Transferring Qubits

31

Initial

Entangle

Entangled

Load Photon Into QM

OGS

136

28

27

191

0

255

32 of 57

Mechanism of Transferring Qubits

32

Initial

Entangle

Entangled

Quantum Memory

OGS

Sits in Memory Until Successfully Entangled (or decoherence)

136

28

27

191

0

255

33 of 57

Mechanism of Transferring Qubits

33

Initial

Entangle

Entangled

OGS

FSO Channel

Quantum Memory

Sits in Memory Until Successfully Entangled (or decoherence)

136

28

27

191

0

255

34 of 57

Mechanism of Transferring Qubits

34

Initial

Entangle

Entangled

OGS

FSO Channel

Quantum Memory

Sits in Memory Until Successfully Entangled (or decoherence)

136

28

27

191

0

255

35 of 57

Mechanism of Transferring Qubits

35

Sits in Memory Until Successfully Entangled (or decoherence)

Initial

Entangle

Entangled

Transmit

FSO Channel

Quantum Memory

The Problem: Low SKR rate due to low transmission success probability & single photon focus

136

28

27

191

0

255

36 of 57

The Solution: Multi-Mode Memories

36

Initial EPS Generation in Parallel

Uses Different Degrees of Freedom to Achieve This

136

28

27

191

0

255

37 of 57

The Solution: Multi-Mode Memories

37

OGS

Multi-Mode QM

Load Each Photon Into a Different Mode in the QM

***Remember this is One Quantum Memory With Multiple Modes***

136

28

27

191

0

255

38 of 57

The Solution: Multi-Mode Memories

38

OGS

Multi-Mode QM

FSO Channel

136

28

27

191

0

255

39 of 57

The Solution: Multi-Mode Memories

39

OGS

Multi-Mode QM

FSO Channel

Photon Loss

Entangled

Photon Loss

136

28

27

191

0

255

40 of 57

The Solution: Multi-Mode Memories

40

OGS

Multi-Mode QM

FSO Channel

Photon Loss

Entangled

Photon Loss

Since entanglement creation can be attempted multiple times in each clock cycle the success increases to:

136

28

27

191

0

255

41 of 57

V. Our Proposed System Architecture

136

28

27

191

0

255

42 of 57

Scheme to Beat: Dual Downlink

42

Connor Casey et al. Multi-Mode Quantum Memories for High-Throughput Satellite Entanglement Distribution. 2025.

136

28

27

191

0

255

43 of 57

Scheme to Beat: Dual Downlink

43

Connor Casey et al. Multi-Mode Quantum Memories for High-Throughput Satellite Entanglement Distribution. 2025.

The Best-Case Scenario for Dual Downlink is when each station is at a 20 Degree Elevation Angle on the Radius of the Satellite Projection

136

28

27

191

0

255

44 of 57

Our Proposed Scheme: A Buffer

44

Connor Casey et al. Multi-Mode Quantum Memories for High-Throughput Satellite Entanglement Distribution. 2025.

136

28

27

191

0

255

45 of 57

Our Proposed Scheme: A Buffer

45

Connor Casey et al. Multi-Mode Quantum Memories for High-Throughput Satellite Entanglement Distribution. 2025.

If we meet the buffer requirement this scenario has ~3x less travel distance through the atmosphere

136

28

27

191

0

255

46 of 57

Physical Implementation

46

Connor Casey et al. Multi-Mode Quantum Memories for High-Throughput Satellite Entanglement Distribution. 2025.

AFC Protocol

Key Advantages

Multi-Mode
Non-cryogenic
Interface with Nuclear Spins
Well-established

1 CM

Uses

136

28

27

191

0

255

47 of 57

VI. Memory Performance & Results

136

28

27

191

0

255

48 of 57

Memory Efficiency

48

Connor Casey et al. Multi-Mode Quantum Memories for High-Throughput Satellite Entanglement Distribution. 2025.

136

28

27

191

0

255

49 of 57

Memory Efficiency

49

Connor Casey et al. Multi-Mode Quantum Memories for High-Throughput Satellite Entanglement Distribution. 2025.

Overall, we achieved a Read / Write efficiency of

136

28

27

191

0

255

50 of 57

Putting Efficiency Into Context

50

Connor Casey et al. Multi-Mode Quantum Memories for High-Throughput Satellite Entanglement Distribution. 2025.

With ~3x Less Travel Time (slant distance) We See A Major Boost to SKR

With 100 Modes This Protocol Achieves 3.5 Mb / S SKR using the BB-84 Encoding Scheme

136

28

27

191

0

255

51 of 57

VI. Conclusion & Outlook

136

28

27

191

0

255

52 of 57

Conclusions & Outlook

52

Throughput via multimode buffering: The proposed scheme delivers substantial SKR gains.

LEO-ready hardware: AFC + alkali↔nuclear-spin interface is compact, non-cryogenic, and payload-compatible; needs minutes-scale coherence and ≥70% memory efficiency.

Near-term demonstrability: We believe LEO satellite implementation is experimentally achievable with near-term components.

1 cm cavity footprint: Small vacuum cavities enable high memory-cell density per satellite.

136

28

27

191

0

255

53 of 57

Questions?

136

28

27

191

0

255

54 of 57

Appendix A: Quantum Memory Architecture

136

28

27

191

0

255

55 of 57

A 10,000 Foot View of the Memory

55

Connor Casey et al. Multi-Mode Quantum Memories for High-Throughput Satellite Entanglement Distribution. 2025.

136

28

27

191

0

255

56 of 57

Appendix B: Key Results Summary

136

28

27

191

0

255

57 of 57

Key Results Summarized

57

Connor Casey et al. Multi-Mode Quantum Memories for High-Throughput Satellite Entanglement Distribution. 2025.

136

28

27

191

0

255