Partial independence suffices to rule out Real Quantum Theory experimentally

Basic Information

• Paper ID: 2502.20102
• Title: Partial independence suffices to rule out Real Quantum Theory experimentally
• Authors: Mirjam Weilenmann (Inria, Télécom Paris), Nicolas Gisin (University of Geneva, Constructor University), Pavel Sekatski (University of Geneva)
• Classification: quant-ph (Quantum Physics)
• Publication Date: February 2025 (arXiv v2: November 21, 2025)
• Paper Link: https://arxiv.org/abs/2502.20102

Abstract

This paper investigates the necessity of complex numbers in quantum theory. Building on the 2021 Nature work by Renou et al. demonstrating that Real Quantum Theory (RQT) cannot explain predictions of complex quantum theory, the authors revisit the source independence assumption. The paper proves that partial independence suffices to establish the insufficiency of real quantum theory, derives trade-off relations between source independence and achievable Bell values in RQT, and demonstrates that 1 ebit of entanglement is both necessary and sufficient as a resource to recover complex quantum correlations using real quantum theory. Finally, based on McKague et al.'s work, the authors construct a scheme using m real qubit entangled states to simulate arbitrary m-source complex quantum experiments.

Research Background and Motivation

1. Core Question

Does quantum theory truly require complex number structure? This represents a long-standing puzzle in quantum mechanics foundations. Although classical physics (mechanics, thermodynamics, electromagnetism) is based on real numbers, the first postulate of quantum theory introduces complex Hilbert space $\mathbb{C}^d$.

2. Problem Significance

• Physical necessity of mathematical structure: Complex numbers in quantum theory are not merely computational tools; quantum states themselves are complex-valued
• Ontological question: If real Hilbert space $\mathbb{R}^d$ suffices, why require "imaginary numbers" that Descartes called "imaginary" and Euler deemed "impossible"?
• Experimental testability: Can this foundational question be determined experimentally?

3. Limitations of Existing Approaches

• McKague et al. (2009): Proved that any complex quantum experiment can be simulated using real quantum theory by attaching an auxiliary real qubit (rebit) to each subsystem, but requires these auxiliary qubits to share an entangled state
• Physical interpretation difficulty: When subsystems are spatially separated, how can they all access the same auxiliary rebit? Where does this entanglement originate?
• Renou et al. (2021): In the network scenario, assuming complete source independence, proved that real quantum theory cannot achieve the Bell value upper bound of complex quantum theory ($B_C^{sup}=6\sqrt{2}\approx 8.49$ vs $B_R^{ub}\leq 7.66$)

4. Paper Motivation

Relaxing the independence assumption: What happens if sources are only partially independent? Can we quantitatively characterize how much quantum correlation is needed to explain experimental results using real quantum theory?

Core Contributions

1. Relaxed independence assumption: Proves that partial source independence suffices to rule out real quantum theory, without requiring complete independence
2. Quantitative trade-off relations: Derives precise trade-off relations between the degree of source independence and achievable Bell values in RQT (via semidefinite programming hierarchy)
3. Necessary and sufficient resources: Proves that in the Renou et al. scenario, 1 ebit of entanglement (corresponding to maximal entanglement of two rebits) is both necessary and sufficient:
   • Necessity: Via self-testing, achieving maximum Bell value $6\sqrt{2}$ requires state $\bar{\varrho}_{S_1S_2}=\frac{1}{2}(\Phi^-+\Psi^+)$ with entanglement of formation $E_F=1$ and distance to separable states $D_{Sep}=\frac{1}{2}$
   • Sufficiency: This entanglement suffices to simulate arbitrary quantum experiments in the network
4. General simulation theorem: Extends the McKague construction, proving that m rebit entangled states $\bar{\varrho}_S^{(m)}$ suffice to simulate complex quantum experiments with arbitrary m independent sources
5. Quantified experimental significance: Assigns quantitative meaning to existing experimental results (e.g., Bell value 8.09 requires $D_{Sep}\geq 0.2$)

Detailed Methodology

Task Definition

In the three-party network scenario (Figure 1):

• Setup: Three parties Alice (A), Bob (B), Charlie (C), with two sources $S_1$ (distributing to AB) and $S_2$ (distributing to BC)
• Measurements: Alice inputs $x\in\{1,2,3\}$ outputs $a\in\{1,-1\}$, Bob outputs $b\in\{00,01,10,11\}$, Charlie inputs $z\in\{1,...,6\}$ outputs $c\in\{1,-1\}$
• Objective: Analyze the upper bound on Bell values achievable by real quantum theory when sources are allowed to share partial quantum correlations $\varrho_{S_1S_2}$

Bell Test Expression

Uses the Bell functional from Renou et al.: $B(P) = \sum_{b\in\{0,1\}^2} B_b(P)$ where each $B_b(P)$ is a combination of three CHSH tests between Alice and Charlie (Eq. 2). Key values:

• Maximum in complex quantum theory: $B_C^{sup}=6\sqrt{2}\approx 8.49$
• Upper bound for independent sources in RQT: $B_R^{ub}\leq 7.66$

Core Technical Methods

1. Entanglement Measures (Appendix A)

Defines two geometric measures:

• Distance to separable set: $D_{Sep}(\varrho)=\inf_{\sigma\in Sep}\frac{1}{2}\|\varrho-\sigma\|$
• Distance to independent set: $D_{Ind}(\varrho)=\inf_{\sigma\in Ind}\frac{1}{2}\|\varrho-\sigma\|$
• Entanglement of Formation: $E_F(\varrho)$

Key properties:

• Monotone non-increasing under local operations in real quantum theory
• For two-rebit states $\bar{\varrho}_{S_1S_2}$: $D_{Ind}=D_{Sep}=\frac{1}{2}$, $E_F=1$ (all achieve maximum values)

2. Semidefinite Programming Hierarchy (SDP Hierarchy, Eq. 6)

Constructs a constrained optimization problem: $\max B(P) \quad \text{s.t.} \quad D_{Sep}(\tau_{AC})\leq \epsilon$

Innovation: Incorporates trace distance constraints into the moment problem:

• Uses moment matrix $\Gamma^n(\tau)$ (n-th level hierarchy)
• Exploits positive semidefiniteness constraints and separability condition $\sigma^{T_A}=\sigma$
• Trace distance condition transforms to: $\begin{pmatrix} \Gamma^n(M) & \Gamma^n(\tau-\sigma) \\ \Gamma^n(\tau-\sigma)^T & \Gamma^n(N) \end{pmatrix}\geq 0, \quad \Gamma^n(M)_{(1,1)}+\Gamma^n(N)_{(1,1)}\leq 4\epsilon$

Computation: 2nd level hierarchy, using YALMIP + Mosek, requires 500GB memory, 5-6 days runtime

3. Linear Bound (Eq. 7, Appendix C)

Derives a simple bound via operator norms: $D_{Sep}(\tau_{AC}) \geq \frac{B(P)-B_R^{ub}}{B_C^{sup}-B_C^{inf}} = \frac{B(P)-7.66}{12\sqrt{2}}$

Proof strategy:

• Bell operator satisfies $\text{tr}(B\rho)=B(P)$
• Uses $\text{tr}(B(\rho-\sigma))\leq D(\rho,\sigma)(B_C^{sup}-B_C^{inf})$
• For any separable state $\sigma$, $\text{tr}(B\sigma)\leq B_R^{ub}$

4. Real Simulation Construction (Appendix D)

Reference frame encoding:

• Each $d$-dimensional complex system $C_i$ corresponds to $2d$-dimensional real system $\bar{R}_i=R_iL_i$
• $L=L_1...L_n$ acts as a "complex reference frame," encoded in the logical subspace: $|R\rangle_L = \frac{1}{\sqrt{2}}(\bigotimes_i|y^+\rangle_{L_i}+\bigotimes_i|y^-\rangle_{L_i})$$|I\rangle_L = \frac{i}{\sqrt{2}}(\bigotimes_i|y^+\rangle_{L_i}-\bigotimes_i|y^-\rangle_{L_i})$ where $|y^\pm\rangle=(|0\rangle\pm i|1\rangle)/\sqrt{2}$

Key property: Real unitary operator $J=i\sigma_Y$ acting on any $L_j$ simulates the imaginary unit: $J_{L_j}|R\rangle_L=|I\rangle_L, \quad J_{L_j}|I\rangle_L=-|R\rangle_L$

State representation (phase-invariant): $\varrho_C \sim \varrho_{RL} = \varrho_R^{Re}\otimes\frac{|R\rangle\langle R|+|I\rangle\langle I|}{2} + \varrho_R^{Im}\otimes\frac{|I\rangle\langle R|-|R\rangle\langle I|}{2}$

Local broadcastability:

• Reference frame state $\bar{\varrho}_L^{(n)}$ can be extended via local operations: $\bar{\varrho}_L^{(n)}\xrightarrow{LO}\bar{\varrho}_{LL'}^{(2n)}$
• Marginal states remain unchanged: $\text{tr}_{L'}\bar{\varrho}_{LL'}^{(2n)}=\bar{\varrho}_L^{(n)}$
• This is impossible in complex quantum theory (no-local-broadcasting theorem), demonstrating operational differences between the two theories

Technical Innovations

1. Formalization of partial independence: First quantitative characterization of source correlation degree (via $D_{Sep}$, $E_F$) and its relation to Bell violation
2. SDP constraint innovation: Embeds continuous trace distance constraints into discrete moment problems, high technical difficulty
3. Tightness proof: Ideal bounds are tight ($\epsilon=0.5$ exactly achieves $6\sqrt{2}$)
4. Resource theory perspective: Places the problem within quantum resource theory framework, clarifying free states (Sep/Ind) and free operations (LOCC/LO)
5. Bound entanglement application: Exploits locally broadcastable bound entanglement in real quantum theory, which does not exist in complex theory

Experimental Setup

Experimental Scenario

• Network topology: Bilocality scenario with two sources and two receivers
• Measurement configuration: Alice 3 inputs, Charlie 6 inputs, Bob fixed measurements (4 outputs)
• Quantum states: In complex theory, uses maximally entangled states and specific measurements to achieve $6\sqrt{2}$

Existing Experimental Data

The paper analyzes three published experiments:

1. Chen et al. (2022, PRL 128:040403): Bell value 8.09(1) → requires $D_{Sep}\geq 0.2$
2. Li et al. (2022, PRL 128:040402): Uses different Bell expression
3. Wu et al. (2022, PRL 129:140401): Bell value 7.83(3) → requires $D_{Sep}\geq 0.05$

Computational Resources

• Software: MATLAB + YALMIP + Mosek solver
• Hardware: High-performance computing cluster, 500GB RAM
• Runtime: 5-6 days per $\epsilon$ value
• Alternative: SCS solver can run on desktop (1-2 hours) but with lower precision

Experimental Results

Main Results (Table I)

Key findings:

1. SDP bound significantly outperforms linear bound: For example, at $B=8.06$, SDP gives $\epsilon_1=20\%$ vs linear bound $\epsilon_2=2.4\%$
2. Tightness verification: At $\epsilon_1=50\%$, $B=8.4953\approx 6\sqrt{2}$ (equal within numerical precision)
3. Experimental implications:
   • Chen experiment (8.09): RQT requires $D_{Sep}\geq 0.2$ source correlation
   • Wu experiment (7.83): Requires $D_{Sep}\geq 0.05$

Ideal Case Analysis

Self-testing result: When achieving $B=6\sqrt{2}$, the Alice-Charlie reduced state self-tests to: $\bar{\varrho}_{AC} = \frac{1}{2}(\Phi^-+\Psi^+) = \frac{1}{4}\begin{pmatrix}1&0&0&-1\\0&1&1&0\\0&1&1&0\\-1&0&0&1\end{pmatrix}$

Entanglement measures (Appendix A):

• $D_{Sep}(\bar{\varrho})=D_{Ind}(\bar{\varrho})=\frac{1}{2}$ (maximum values)
• $E_F(\bar{\varrho})=1$ ebit (maximal entanglement of two rebits)
• By self-testing, source state $\varrho_{S_1S_2}$ must satisfy the same bounds

Sufficiency (Proposition): This state suffices to simulate any complex quantum experiment in the network

General Simulation Theorem

Theorem: Any complex quantum experiment with m independent sources can be simulated using real quantum theory if sources initially share: $\bar{\varrho}_S^{(m)} = \frac{1}{2}\left(\bigotimes_{i=1}^m|y^+\rangle\langle y^+|_{S_i} + \bigotimes_{i=1}^m|y^-\rangle\langle y^-|_{S_i}\right)$

Proof outline:

1. This state contains 1 ebit of entanglement (across any bipartition)
2. Simultaneously locally broadcastable: $\bar{\varrho}_S^{(m)}\xrightarrow{LO}\bar{\varrho}_{SL}^{(m+n)}$
3. Measurements do not consume reference frame entanglement
4. For $m=2$, exactly the self-tested state

Related Work

Historical Context

1. Stueckelberg (1960): First proposed real Hilbert space quantum theory
2. Wootters (1990), Caves et al. (2002): Studied local accessibility of real quantum states
3. McKague et al. (2009): Proved real quantum theory can simulate complex quantum theory (requiring auxiliary entanglement)
4. Hardy & Wootters (2012): Finite wholeness and real vector space quantum theory
5. Chiribella et al. (2010), Barnum et al. (2020): Research within probabilistic theory framework

Network Scenario Breakthrough

• Renou et al. (2021, Nature): First experimental distinction between theories in network scenario
  • Key: Source independence assumption excludes shared entanglement
  • Bell value bounds: $B_R^{ub}=7.66 < B_C^{sup}=6\sqrt{2}$
  • Shared randomness cannot rescue real quantum theory

Relation to Prior Work

• Extends Renou work: Relaxes independence, quantitatively analyzes partial independence
• Extends McKague work: From single experiment to network scenario, clarifies minimal resources
• Distinction from recent work:
  • Barrios et al. (2025, arXiv:2503.17307): Modifies state definition to make reference frame "appear" independent
  • Hoffreumon & Woods (2025, arXiv:2504.02808): Modifies tensor product rules to embed complex theory
  • This paper's stance: Focuses on experimental distinction between Real vs Complex theories, rather than reformulation

Conclusions and Discussion

Main Conclusions

1. Partial independence suffices: Complete source independence unnecessary; partial independence suffices to rule out RQT
2. Quantitative trade-offs: Establishes precise relationship between Bell value and required source correlation (Table I)
3. Minimal resources: 1 ebit entanglement is necessary and sufficient (for two-source scenario)
4. Universal construction: m-rebit entangled state can simulate arbitrary m-source experiments
5. Experimental significance: Existing experimental results imply quantitative lower bounds on source correlation

Limitations

1. Computational complexity: SDP hierarchy requires enormous computational resources, limiting higher precision analysis
2. Scenario specificity: Results apply to specific network topology (two sources, three parties); other networks require separate analysis
3. Idealized assumptions:
   • Perfect measurements (noiseless)
   • Finite-dimensional Hilbert space
   • Non-relativistic quantum theory
4. Experimental challenges:
   • Existing experiments not loophole-free
   • Source independence difficult to guarantee in chip implementations
   • Bell values still gap from theoretical maximum

Future Directions

1. Other network topologies: Generalize to star, chain, and other networks
2. Noise robustness: Analyze impact of realistic noise on conclusions
3. More efficient algorithms: Reduce SDP solving complexity
4. Experimental improvements:
   • Increase Bell values (currently 8.09 vs theoretical 8.49)
   • Implement loophole-free experiments
   • Better control of source independence
5. Philosophical implications: Deeper exploration of "entanglement space" ontological interpretation

In-Depth Evaluation

Strengths

1. Theoretical rigor:
   • Complete mathematical framework (resource theory, SDP, self-testing)
   • Bidirectional proofs of necessity and sufficiency
   • Numerical verification of tight bounds
2. Novelty:
   • First quantitative characterization of partial independence
   • Embedding trace distance constraints into moment problems (high technical difficulty)
   • Reveals unique properties of real theory (locally broadcastable bound entanglement)
3. Experimental relevance:
   • Assigns new meaning to existing experiments
   • Guides future experimental design
   • Provides operational criteria
4. Clear presentation:
   • Detailed historical background
   • Complete technical details (appendices)
   • Clear comparison with related work

Weaknesses

1. Computational bottleneck:
   • 500GB memory requirement limits accessibility
   • Only computed to 2nd level hierarchy (possibly not tight)
   • Large gap between linear and SDP bounds (Table I suggests linear bound insufficient)
2. Scenario specificity:
   • Results depend on specific Bell expression
   • Fixed network topology
   • Generalizability requires further verification
3. Experimental gap:
   • Existing experimental Bell values significantly below theoretical maximum
   • Experimental verification of source independence difficult
   • Loophole-free experiments not yet realized
4. Philosophical controversy:
   • Ontological status of "entanglement space" interpretation insufficiently discussed
   • Fundamental disagreement with Barrios, Hoffreumon work (definition vs experimental question)

Impact

1. Theoretical contribution:
   • Deepens understanding of complex numbers' role in quantum theory
   • Establishes quantitative framework for partial independence
   • Connects network scenarios to resource theory
2. Experimental guidance:
   • Provides quantitative targets for experimental design
   • Quantifies source independence requirements
   • New perspective on experimental result interpretation
3. Methodology:
   • SDP + trace distance constraint technique applicable to other problems
   • Paradigm combining self-testing with resource theory
4. Reproducibility:
   • Code open-sourced (GitHub)
   • Detailed method description
   • High computational resource requirements

Applicable Scenarios

1. Quantum foundations research:
   • Experimental verification of quantum theory mathematical structure
   • Network scenario quantum correlations
   • Other scenarios for complex number necessity
2. Quantum information:
   • Quantum network design
   • Entanglement resource analysis
   • Device-independent protocols
3. Experimental physics:
   • Photonic quantum network experiments
   • Superconducting quantum chip networks
   • Ion trap networks
4. Theoretical extensions:
   • Quantum theory over other number fields (quaternions, etc.)
   • Generalized probabilistic theories
   • Network structures in quantum gravity

References

Key Citations

1. Renou et al., Nature 600, 625 (2021): Original network scenario experimental proposal
2. McKague et al., PRL 102, 020505 (2009): Foundation of real simulation construction
3. Chen et al., PRL 128, 040403 (2022): First experimental implementation
4. Caves et al., Found. Phys. Lett. 14, 199 (2001): Real entanglement measures
5. Bancal et al., Quantum 5, 401 (2021): Finite statistics self-testing

Recent Controversial Literature

6. Barrios et al., arXiv:2503.17307 (2025): Real quantum theory with modified state definition
7. Hoffreumon & Woods, arXiv:2504.02808 (2025): Real quantum theory with modified tensor product

Overall Assessment: This is a high-quality paper of significant theoretical and experimental importance in quantum foundations. Through rigorous mathematical analysis and innovative SDP techniques, it advances the pioneering work of Renou et al. to the quantitative level, providing a more nuanced picture of the indispensability of complex numbers in quantum theory. Despite computational complexity and experimental challenges, the theoretical framework and methodology established have lasting value for quantum foundations research and quantum network experiments. The interesting contrast with recent theoretical reformulations (Barrios, Hoffreumon) highlights the subtle nature of the definitional question "what is real quantum theory?"