Catalytic Activation of Bell Nonlocality

The correlations of certain entangled states can be perfectly simulated classically via a local model. Hence such states are termed Bell local, as they cannot lead to Bell inequality violation. Here, we show that Bell nonlocality can nevertheless be activated for certain Bell-local states via a catalytic process. Specifically, we present a protocol where a Bell-local state, combined with a catalyst, is transformed into a Bell-nonlocal state while the catalyst is returned exactly in its initial state. Importantly, this transformation is deterministic and based only on local operations. Moreover, this procedure is possible even when the state of the catalyst is itself Bell local, demonstrating a new form of superactivation of Bell nonlocality, as well as an interesting form of quantum catalysis. On the technical level, our main tool is a formal connection between catalytic activation and many-copy activation.

💡 Research Summary

The paper “Catalytic Activation of Bell Nonlocality” introduces a novel protocol that transforms a Bell‑local entangled state into a Bell‑nonlocal one by employing a quantum catalyst, while preserving the catalyst’s state exactly. The authors consider two distant parties, Alice and Bob, who share a bipartite state ρ_AB that is entangled yet admits a local hidden‑variable (LHV) model for all possible local measurements. They also share an ancillary bipartite system ω_CA CB, the catalyst. By performing only deterministic local operations on their respective subsystems (Alice on A C_A, Bob on B C_B) and without any classical communication, they map the joint state ρ_AB ⊗ ω_CA CB to a new joint state τ_A′B′ ⊗ ω_CA CB. The marginal on the catalyst remains unchanged, satisfying the catalytic condition, while the marginal τ_A′B′ is shown to violate a Bell inequality, thus exhibiting Bell‑nonlocality.

The central theoretical result (Theorem 1) states that if a Bell‑local state ρ_AB becomes Bell‑nonlocal when taken in n copies (i.e., ρ_AB^{⊗ n} violates some Bell inequality for a finite n), then a single copy of ρ_AB can be catalytically activated using a catalyst ω defined as a convex mixture of lower‑order tensor powers of ρ and an arbitrary product state σ. The construction proceeds in two steps. Lemma 1 demonstrates that, through conditional swapping and copying operations controlled by classical registers, the initial state ρ_AB ⊗ ω can be locally transformed into a mixture of the form (1/n) ρ_AB^{⊗ n} ⊗ |00⟩⟨00| + ((n−1)/n) σ^{⊗ n} ⊗ |11⟩⟨11|. Lemma 2 then shows that any state of this mixed form retains the Bell‑inequality violation of the original ρ_AB^{⊗ n}, because the parties can condition their measurement strategy on the value of the classical register: when the register is 0 they apply the nonlocal measurements that achieve the violation, and when it is 1 they use a deterministic local strategy that attains the local bound. The overall Bell score exceeds the local bound by a factor proportional to the probability weight of the nonlocal branch, guaranteeing nonlocality of τ_A′B′.

A striking implication is “super‑activation” with a Bell‑local catalyst. If n is the smallest number of copies for which ρ_AB^{⊗ n} is nonlocal, the catalyst ω constructed in Theorem 1 consists solely of mixtures of ρ_AB^{⊗ m} for m < n (all Bell‑local) and product states, making ω itself Bell‑local. Yet, when combined with a single copy of ρ_AB, the pair yields a Bell‑nonlocal output. This contrasts with previous quantum catalysis results, where the catalyst must already possess the resource it helps to generate.

The authors further derive Corollary 1: any bipartite state with local dimension d and singlet fraction F(ρ) > 1/d can be catalytically activated. This includes isotropic two‑qubit states ρ(V)=V|Φ⁺⟩⟨Φ⁺|+(1−V)I/4 with visibility 1/3 < V ≤ 1/2, which are known to be Bell‑local for arbitrary measurements but become nonlocal after catalytic activation. The paper also discusses a concrete protocol for the CHSH inequality, showing how the catalytic scheme can be tailored to the most widely used Bell test, with direct relevance to device‑independent randomness certification and quantum key distribution.

Overall, the work establishes a rigorous equivalence between many‑copy activation and catalytic activation of Bell nonlocality, expands the toolbox for generating nonlocal correlations from weakly entangled states, and opens new avenues for experimental implementation of catalytic nonlocality without the need for multiple copies or communication. The deterministic, fully local nature of the protocol, together with the possibility of using a Bell‑local catalyst, makes this approach both conceptually surprising and practically promising for future quantum information tasks.