Pure-State Entanglement in Families

• The paper reveals that pure-state entanglement in families emerges as a global geometric obstruction, quantified via Azumaya algebras and Severi–Brauer schemes.
• It employs torsor reduction theory and Hilbert scheme moduli to demonstrate how local subsystem factorizations can fail to globalize due to cohomological constraints.
• The study unifies conventional entanglement concepts with advanced geometric methods to enhance classification under LOCC, SLOCC, and operational criteria.

Pure-state entanglement in families concerns the structure of entanglement in parameterized collections (“families”) of pure quantum states, most notably when the global Hilbert space and potential subsystem structures vary over a base space. This unifies conventional entanglement theory, which is rooted in a fixed tensor product decomposition, with the algebro-geometric, operational, and categorical frameworks necessary to capture situations where subsystem decompositions do not globalize due to geometric, topological, or cohomological obstructions. The mathematical formalism centers around Severi-Brauer schemes, Azumaya algebras, torsor reduction theory, and moduli/Hilbert schemes, providing geometric and cohomological invariants for the existence or failure of product-state loci in families (Ikeda, 20 Jan 2026). The broader context of “entanglement in families” also provides a systematic perspective on the local-versus-global dichotomy in quantum information, and underpins classification schemes for pure-state entanglement under LOCC, SLOCC, and operational criteria.

1. Geometric Formalism: Severi–Brauer Schemes and Azumaya Algebras

The core geometric object for formulating pure-state entanglement in families is the Severi–Brauer fibration

$SB(A) = P \times^{PGL_n} \mathbb{P}^{n-1} \to X,$

where $A$ is a degree-$n$ Azumaya algebra over a base variety $X$ ($A$ is locally isomorphic to $Mat_n(\mathcal{O}_U)$ on the fppf site), and $P \to X$ is its associated principal $PGL_n$-torsor. Locally, $SB(A)$ is a $\mathbb{P}^{n-1}$-bundle, but globally, it may be twisted. The failure of $A$ to admit a global splitting (i.e., to be isomorphic to $Mat_n$ globally) is measured by its Brauer class $[A] \in \mathrm{Br}(X)$.

A fixed factorization type $\mathbf{d} = (d_1, \dots, d_s)$ with $n = \prod_i d_i$ encodes a putative tensor product structure $H \cong H_1 \otimes \cdots \otimes H_s$ for the Hilbert space. The Segre variety $\Sigma_{\mathbf{d}} \subset \mathbb{P}^{n-1}$ specifies product states of this type. The subscheme locus representing product (i.e., unentangled) states of type $\mathbf{d}$ is the image of $\Sigma_{\mathbf{d}}$ under twisting by $P$. However, globalizing these local charts requires the transition cocycle $g_{ij} \in PGL_n$ to land entirely in the stabilizer subgroup $G_{\mathbf{d}}$ of $\Sigma_{\mathbf{d}}$.

Table 1: Geometric Data in Pure-State Entanglement in Families

Object	Description	Role in Entanglement
$A$	Degree-$n$ Azumaya algebra on $X$	Encodes twisted state space
$P$	Principal $PGL_n$-torsor associated to $A$	Transition data
$SB(A)$	Severi–Brauer scheme: $P \times^{PGL_n} \mathbb{P}^{n-1}$	Family of projective state spaces
$G_{\mathbf{d}}$	Stabilizer subgroup of Segre variety $\Sigma_{\mathbf{d}}$	Describes subsystem structure
$P/G_{\mathbf{d}}$	Moduli of subsystem reductions	Classifies product-state loci
$\mathrm{Br}(X)$	Brauer group of $X$	Cohomological obstruction

The existence of a global $\mathbf{d}$–subsystem structure is thus a geometric problem of torsor reduction: $$[P]\in H^1(X, PGL_n) \ \text{lies in the image of}\ H^1(X, G_{\mathbf{d}}) \to H^1(X, PGL_n),$$ with the commutative diagram linking $H^1$ and the Brauer group via the boundary map.

2. Entanglement as a Geometric Obstruction

In this framework, pure-state entanglement in families is conceptualized as an obstruction to the existence of a global product-state locus of factorization type $\mathbf{d}$ in $SB(A)$. Specifically, the lack of a global Segre subfibration $\Sigma_{\mathbf{d}}(A) \hookrightarrow SB(A)$ is precisely the failure to reduce $P$ to $G_{\mathbf{d}}$. If $[A] \notin \mathrm{Br}_{\mathbf{d}}(X)$, then no global subsystem structure of type $\mathbf{d}$ exists, and the entire family is entangled relative to this type: locally, one can select charts in which the Hilbert spaces are factorized, but patching together globally is obstructed.

If a reduction exists, the family becomes trivializable as a $\mathbf{d}$–type, and all usual bipartite or multipartite entanglement invariants (e.g., Schmidt rank loci, stratifications) descend to the family level and are compatible under base change (Ikeda, 20 Jan 2026).

3. Moduli Spaces and Hilbert Scheme Realization

The moduli problem of subsystem choices is represented by the quotient $P/G_{\mathbf{d}}$, parameterizing all possible $G_{\mathbf{d}}$-reductions of $P$. This quotient is identified with a locally closed subscheme of the relative Hilbert scheme: $$\text{Hilb}^{\Sigma_{\mathbf{d}}}\left(SB(A)/X\right) \subset \text{Hilb}\left(SB(A)/X\right),$$ which parameterizes all flat closed $T$-subschemes of $SB(A) \times_X T$ (for $T \to X$) locally isomorphic to $\Sigma_{\mathbf{d}} \times T$. The main theorem (subsystem–Hilbert isomorphism) states that the moduli functor of $G_{\mathbf{d}}$-reductions is represented precisely by this Hilbert subspace: $$\text{Hilb}^{\Sigma_{\mathbf{d}}}(SB(A)/X) \cong P/G_{\mathbf{d}}.$$ A global section of $P/G_{\mathbf{d}}$ provides a global subsystem structure; the absence of such a section provides a geometric certificate of family-level entanglement (Ikeda, 20 Jan 2026).

4. Obstruction Theory and Cohomological Constraints

The obstructions to reduction, and hence to the global existence of a product-state locus, are captured cohomologically. The necessary condition for a degree-$n$ Azumaya algebra $A$ to admit a $\mathbf{d}$-type subsystem decomposition is that its class $[A]\in \mathrm{Br}(X)$ lies in the subgroup $\mathrm{Br}_{\mathbf{d}}(X)$, the image of $H^1(X, G_{\mathbf{d}})$ in the Brauer group. The minimal torsion constraint is

$$\mathrm{Br}_{\mathbf{d}}(X) \subset \mathrm{Br}(X)[\ell],$$

where $\ell = \mathrm{lcm}(d_i)$. If the obstruction vanishes, reductions exist; if not, the family is intrinsically and unavoidably entangled.

These cohomological characterizations connect entanglement in families to the structure theory of torsors, nonabelian Galois cohomology, and descent in algebraic geometry, integrating quantum information with advanced algebro-geometric tools (Ikeda, 20 Jan 2026).

5. Compatibility with Quantum Information Structures

Whenever a global subsystem structure exists, familiar quantum-information theoretic notions, such as product-state loci, determinantal varieties (for Schmidt rank), stratification by degeneracy type, and numerical invariants, all extend uniformly and in a base-compatible way over $X$. These data can be trivialized locally on the base, but their global structure is mediated by the moduli stack $P/G_{\mathbf{d}}$.

This geometric machinery enables the analysis of families where parameter space $X$ could be a moduli space itself (e.g., varying external parameters, families of physical systems, or even geometric phases), and where global choices of subsystems correspond to reductions of structure group, a paradigm familiar from differential and algebraic geometry but novel in standard quantum information.

6. Implications and Broader Context

The algebro-geometric approach to pure-state entanglement in families demonstrates that entanglement is not merely a property of vectors within a fixed Hilbert space, but may be a “global” property of a vector bundle (or $SB(A)$ fibration) over a parameter variety, tied intricately to the nontriviality of torsors and associated reduction problems (Ikeda, 20 Jan 2026). This perspective connects with operational and categorical viewpoints in quantum information:

• In the finite-dimensional case, all pure states are equivalent as entanglement resources once unrestricted choice of observables is allowed, as proven by the tailored-observables theorem (Harshman et al., 2011). However, in the geometric/family setting, such freedom is constrained by global topology and nontriviality of the underlying torsors.
• In the context of universally entangled sets, the existence or nonexistence of global product-state loci ties to whether collections of states remain entangled for all possible subsystem definitions (Yu et al., 2021).

This geometric formalism supports a rigorous theory of “entanglement incompatibility” hardwired into the global structure of parameterized quantum systems, fundamentally extending the landscape of quantum entanglement theory.

• Quantum Entanglement Geometry on Severi-Brauer Schemes: Subsystem Reductions of Azumaya Algebras (Ikeda, 20 Jan 2026)
• Observables can be tailored to change the entanglement of any pure state (Harshman et al., 2011)
• Absolutely entangled sets of pure states for bipartitions and multipartitions (Yu et al., 2021)