Entanglement in C$^*$-algebras: tensor products of state spaces

1. Problem Statement

The paper addresses the relationship between the Namioka–Phelps tensor products of compact convex sets and the phenomenon of quantum entanglement within the context of $C^*$-algebras.

Specifically, the authors investigate two distinct ways to define the tensor product of the state spaces $S(A)$ and $S(B)$ of two unital $C^*$-algebras $A$ and $B$:

1. The Minimal Tensor Product ($S(A) \otimes_\wedge S(B)$): Defined via the "inner" positive cone generated by elementary tensors.
2. The Maximal Tensor Product ($S(A) \otimes_\vee S(B)$): Defined via the "outer" positive cone consisting of elements positive against all tensor products of positive functionals.

The central problem is to determine the conditions under which these two tensor products coincide. This relates directly to Barker's Conjecture, which posits that for compact convex sets $K_1$ and $K_2$, $K_1 \otimes_\wedge K_2 = K_1 \otimes_\vee K_2$ if and only if at least one of the sets is a Choquet simplex. While this was known for finite-dimensional sets, the paper seeks to resolve this for the infinite-dimensional state spaces of $C^*$-algebras and characterize the nature of entanglement in this setting.

2. Methodology

The authors employ a combination of functional analysis, convex geometry, and operator algebra theory:

• Identification of State Spaces: They identify the state space $S(A)$ with the state space of the ordered vector space $(A_{sa}, A_+, 1_A)$. This allows them to translate the Namioka–Phelps definitions of tensor products of convex sets into the language of $C^*$-algebra tensor products (algebraic, minimal/spatial, and maximal).
• Separable vs. Entangled States: They establish that the minimal tensor product of state spaces corresponds exactly to the set of separable (un-entangled) states on the spatial tensor product $A \otimes B$. The maximal tensor product corresponds to a broader set of states related to unital positive maps.
• Embedding Techniques (Glimm's Lemma): To handle general non-commutative $C^*$-algebras, the authors use a variation of Glimm's Lemma. This allows them to embed matrix algebras (or cones over them) into any non-commutative $C^*$-algebra with sufficiently high rank. This technique "lifts" known entanglement phenomena from finite-dimensional matrix algebras to infinite-dimensional algebras.
• Injectivity and Approximate Injectivity: The paper utilizes the concepts of injective and approximately injective $C^*$-algebras (e.g., nuclear algebras) to show that entanglement in subalgebras cannot "un-entangle" when passing to larger algebras.
• Trace Simplexes: They analyze the tensor products of trace simplexes (sets of tracial states), leveraging the fact that trace simplexes are always Choquet simplexes.

3. Key Contributions and Results

A. Resolution of Barker's Conjecture for $C^*$-algebras

The paper proves that for the state spaces of unital $C^*$-algebras, the minimal and maximal Namioka–Phelps tensor products coincide if and only if at least one of the algebras is commutative.

• Theorem 3.17: $S(A) \otimes_\wedge S(B) = S(A) \otimes_\vee S(B)$ if and only if $A$ or $B$ is commutative.
• Implication: If both $A$ and $B$ are non-commutative, the sets are distinct. This confirms Barker's conjecture for the specific class of state spaces of $C^*$-algebras.
• Re-proof of Classical Fact: As a byproduct, they re-prove that the state space of a $C^*$-algebra is a Choquet simplex if and only if the algebra is commutative (in which case it is a Bauer simplex).

B. Characterization of Entanglement

The authors provide a rigorous characterization of entanglement in infinite-dimensional $C^*$-algebras:

• Strict Inclusions: For non-commutative $A$ and $B$, the following chain of strict inclusions holds:
  $$S^*(A \otimes B) \subsetneq S(A \otimes B) \subsetneq S(A \otimes_{max} B) \subsetneq S(A) \otimes_\vee S(B)$$
  Where $S^*(A \otimes B)$ is the set of separable states.
• Positive Elements: They show that the cone of separable positive elements $A_+ \otimes B_+$ is strictly contained in the cone of all positive elements $(A \otimes B)_+$ if and only if both algebras are non-commutative.
• Quantifying Entanglement: They introduce a constant $\kappa(A, B)$ defined as the supremum of the maximal norms of functionals in the minimal tensor product of state spaces.
  • If $A$ and $B$ are not sub-homogeneous (i.e., contain arbitrarily large matrix algebras), $\kappa(A, B) = \infty$.
  • This implies that the maximal tensor product of state spaces is unbounded relative to the minimal tensor product in the non-commutative case, providing a new metric for the "size" of entanglement.

C. Tensor Products of Trace Simplexes

Unlike general state spaces, the tensor products of trace simplexes behave differently because traces are inherently "un-entangled" in a specific sense:

• Theorem 4.1: For any unital $C^*$-algebras $A$ and $B$ with traces, the minimal and maximal tensor products of their trace simplexes coincide and equal the trace simplex of the tensor product of the algebras:
  $$T(A) \otimes T(B) = T^*(A \otimes B) = T(A \otimes B) = T(A \otimes_{max} B)$$
• Significance: This result confirms that traces cannot be entangled. Every tracial state on a tensor product $A \otimes_{max} B$ factors through the minimal tensor product $A \otimes B$.
• Classification: They characterize when the trace simplex of a tensor product is a Bauer simplex or the Poulsen simplex (the unique non-trivial simplex with a dense extreme boundary). For instance, the trace simplex of an infinite tensor product of $C^*$-algebras is the Poulsen simplex if and only if the trace simplexes of the components are Poulsen (or trivial) and not all are trivial.

4. Significance

• Bridging Fields: The paper successfully bridges the gap between abstract convex geometry (Namioka–Phelps theory) and quantum information theory (entanglement in $C^*$-algebras).
• Infinite Dimensions: It extends finite-dimensional results on entanglement to the general infinite-dimensional setting of $C^*$-algebras, proving that non-commutativity is the sole source of entanglement in this context.
• New Metrics: The introduction of the constant $\kappa(A, B)$ and the observation of unboundedness in the non-commutative case offer a novel way to quantify the "degree" of entanglement beyond binary separable/entangled distinctions.
• Trace Theory: The result that traces are never entangled provides a powerful structural insight into the relationship between minimal and maximal tensor products of $C^*$-algebras, simplifying the analysis of tracial states in complex operator algebras.

In summary, Musat and Rørdam demonstrate that the geometric distinction between minimal and maximal tensor products of state spaces is a precise mathematical manifestation of quantum entanglement, occurring exactly when the underlying algebras are non-commutative.