Tripartite Quantum Graph States

• Tripartite quantum graph states are three-party entangled states formed by controlled-phase interactions along graph edges, distinguishing pairwise and higher-body couplings.
• They are classified via LU and SLOCC criteria, revealing distinct structures such as GHZ-type and W-like behavior and highlighting differences between ordinary and hypergraph states.
• Their practical applications include quantum secret-sharing protocols, network generation, and weighted graph models, bridging theoretical insights with operational resources.

Tripartite quantum graph states are three-party entangled states associated with graphs or graph-like calculi, usually constructed by placing one quantum system at each of three vertices and applying controlled-phase interactions along pairwise edges. In the discrete-variable qubit setting, a tripartite graph state is the ordinary three-vertex, two-body case; tripartite hypergraph states enlarge this by allowing a genuine three-body hyperedge, while categorical and diagrammatic approaches recast the same tripartite entanglement structure in terms of Frobenius algebras, spiders, and Z/X-calculus. Across these formalisms, the tripartite case is the minimal setting in which GHZ-type versus W-type behavior, local-unitary versus SLOCC equivalence, stabilizer structure, secret-sharing protocols, and network-generation limits can all be stated sharply (Qu et al., 2013, Salem, 2 Dec 2025, Coecke et al., 2010).

1. Definitions and basic constructions

In the standard discrete-variable preparation picture, graph states are built by initializing each qubit in the plus state and applying controlled-phase entangling gates along graph edges. The review literature treats graph states as the 2-uniform, edge-based special case of hypergraph states: graph states use only pairwise edges, whereas hypergraph states allow hyperedges of arbitrary size $k \ge 2$ and therefore higher-body phase interactions (Salem, 2 Dec 2025).

For three qubits, the hypergraph formalism fixes

$V=\{A,B,C\},$

and defines the state associated with a hypergraph $g=(V,E)$ by

$$|g\rangle=\prod_{e\in E} Z_e\, |+\rangle^{\otimes 3}.$$

This construction reduces to an ordinary graph state when only 2-edges are present. The same formalism admits a Boolean-function representation as a real equally weighted state, so every three-qubit hypergraph state can be written with phases determined by a Boolean polynomial $u(g)$ (Qu et al., 2013).

A useful distinction is therefore between the ordinary tripartite graph-state regime and the genuinely tripartite hypergraph regime. The review makes this caveat explicit: a 3-vertex graph state is not the same object as the 3-vertex single-hyperedge state $\ket{H_3}$, because the latter is produced by a genuine 3-body phase gate and belongs to the hypergraph-state family rather than the ordinary graph-state family (Salem, 2 Dec 2025).

Formalism	Tripartite object	Defining feature
Ordinary graph state	3-vertex graph with pairwise edges	2-body controlled-phase preparation
Hypergraph state	3 vertices with possible 3-body hyperedge	Higher-body phase gate allowed
Categorical graphical state	Tripartite state as diagrammatic operation	Frobenius/spider structure

This hierarchy is central: ordinary tripartite graph states provide the baseline stabilizer-like model, while tripartite hypergraph states and graphical calculi expose which aspects of tripartite entanglement are genuinely beyond the pairwise graph-state setting.

2. Three-qubit entanglement classification

The most detailed finite classification in the supplied literature is the three-qubit hypergraph-state analysis. Using local entropic measures $E_A,E_B,E_C$, 3-tangle $\tau$, Schmidt measure, and concurrences, the three-vertex hypergraphs are partitioned into six LU-equivalence classes

$G_0,\ G_1,\ G_2,\ G_3,\ G_4,\ G_5.$

Within this taxonomy, $G_5$, the class containing the genuine three-hyperedge state $V=\{A,B,C\},$0, is the only LU class not equivalent to any ordinary graph state, whereas the GHZ state is LU-equivalent to a representative of $V=\{A,B,C\},$1 (Qu et al., 2013).

Under SLOCC, the same family collapses to five classes,

$V=\{A,B,C\},$2

which coincide with the usual Dür–Vidal–Cirac classes $V=\{A,B,C\},$3, $V=\{A,B,C\},$4, $V=\{A,B,C\},$5, $V=\{A,B,C\},$6, and GHZ-type. The paper reports

$V=\{A,B,C\},$7

so both $V=\{A,B,C\},$8 and $V=\{A,B,C\},$9 are GHZ-type under SLOCC, not W-type (Qu et al., 2013).

This behavior aligns with the broader tripartite qubit classification in which the only genuinely tripartite SLOCC-maximal classes are GHZ and W. In the compositional literature, these are treated as the two primitive tripartite states, represented respectively by

$g=(V,E)$0

and used as generators of a graphical theory of multipartite entanglement (Coecke et al., 2010).

The resulting structural lesson is precise. Hypergraph states strictly enlarge graph states at the level of LU structure, because $g=(V,E)$1 is new relative to ordinary graph states, but they do not enlarge the available three-qubit SLOCC entanglement types. The corollary stated explicitly is that any three-qubit hypergraph state cannot be converted into the W state by SLOCC (Qu et al., 2013).

3. Bipartite entanglement patterns and W-like behavior

A separate classification arises when three-qubit states are reduced to their entangled-graph structure, where vertices represent qubits and an edge indicates nonzero pairwise concurrence. In this picture, three-qubit hypergraph states fall into five classes: one with no edges, three with exactly one edge, and one with all three edges. Specifically, $g=(V,E)$2 has no edge, $g=(V,E)$3, $g=(V,E)$4, and $g=(V,E)$5 each have one edge, and $g=(V,E)$6 yields the complete triangle (Qu et al., 2013).

The key formulas behind this analysis are

$g=(V,E)$7

and the Coffman–Kundu–Wootters monogamy relation

$g=(V,E)$8

together with permutation symmetry. These relations determine the pairwise concurrences and therefore the entangled graph of each class (Qu et al., 2013).

The complete-triangle class $g=(V,E)$9 is especially notable because its entangled graph matches that of the W state: every pair of qubits is entangled. The supplied material emphasizes that this makes $$|g\rangle=\prod_{e\in E} Z_e\, |+\rangle^{\otimes 3}.$$0 W-like only in the sense of bipartite entanglement distribution. It is not SLOCC-equivalent to W, since $$|g\rangle=\prod_{e\in E} Z_e\, |+\rangle^{\otimes 3}.$$1 belongs to the GHZ-type class under the 3-tangle classification (Qu et al., 2013).

This distinction addresses a common misconception. Equality of pairwise entangled-graph structure does not imply equality of genuine tripartite entanglement type. The diagrammatic literature makes the same separation in different language: GHZ-class states exhibit global tripartite entanglement, whereas W-class states exhibit pairwise or local entanglement. In the Z/X-calculus, GHZ is represented by a single connected spider, while W requires a different phase-sensitive construction characterized by supplementarity rather than a single-spider normal form (Coecke et al., 2011).

A second misconception concerns expressivity. Hypergraph states do not realize every possible three-qubit entangled graph; in particular, the analysis states that they do not produce states with exactly two edges in the entangled-graph picture (Qu et al., 2013).

4. Stabilizers, spiders, and graphical calculi

The tripartite regime is also a testing ground for how graph-state ideas change when one moves beyond ordinary pairwise stabilizer structure. For the canonical three-qubit hypergraph state

$$|g\rangle=\prod_{e\in E} Z_e\, |+\rangle^{\otimes 3}.$$2

the stabilizers are

$$|g\rangle=\prod_{e\in E} Z_e\, |+\rangle^{\otimes 3}.$$3

These operators are explicitly nonlocal in the hypergraph case: each local $$|g\rangle=\prod_{e\in E} Z_e\, |+\rangle^{\otimes 3}.$$4 is accompanied by a controlled-phase gate acting on the other two qubits (Salem, 2 Dec 2025).

In categorical quantum mechanics, the same tripartite distinction is encoded algebraically. A symmetric tripartite state satisfying the appropriate maximality and symmetry conditions is a Frobenius state and canonically induces a commutative Frobenius algebra with multiplication $$|g\rangle=\prod_{e\in E} Z_e\, |+\rangle^{\otimes 3}.$$5, unit $$|g\rangle=\prod_{e\in E} Z_e\, |+\rangle^{\otimes 3}.$$6, comultiplication $$|g\rangle=\prod_{e\in E} Z_e\, |+\rangle^{\otimes 3}.$$7, and counit $$|g\rangle=\prod_{e\in E} Z_e\, |+\rangle^{\otimes 3}.$$8. In the qubit case, GHZ corresponds to a special commutative Frobenius algebra and W to an anti-special commutative Frobenius algebra; every SLOCC-maximal tripartite qubit state is SLOCC-equivalent to a Frobenius state (Coecke et al., 2010).

The Z/X-calculus sharpens this contrast diagrammatically. GHZ appears as a single green spider with three legs and is therefore naturally graph-state-like. W can also be represented, but only as a mixed-color diagram with specific phase conditions; it is not generated by the same single-spider structure. The supplementarity conditions identify when such diagrams fall out of the GHZ class and into the W class, thereby linking phase singularities to tripartite entanglement type (Coecke et al., 2011).

The qutrit generalization modifies the landscape further. Although tripartite qutrits have infinitely many SLOCC classes overall, the symmetry requirements associated with Frobenius-state structure select exactly three classes,

$$|g\rangle=\prod_{e\in E} Z_e\, |+\rangle^{\otimes 3}.$$9

corresponding respectively to an SCFA, an ACFA, and an ISCFA. This is the qutrit analogue of the qubit GHZ/W dichotomy, but with an intermediate class that has no qubit counterpart (Honda, 2012).

The review literature also situates these discrete-variable structures relative to continuous-variable graph states. There the graph-state stabilizers are written as

$u(g)$0

with nullifiers

$u(g)$1

By contrast, continuous-variable hypergraph states are described as more recent and non-Gaussian, and the tripartite continuous-variable hypergraph case is explicitly less mature than the discrete-variable theory (Salem, 2 Dec 2025).

5. Tripartite graph states in secret sharing and threshold protocols

Tripartite graph states are not only classification objects; they are also operational resources. In prime dimension $u(g)$2, qudit graph states provide a unified stabilizer-based language for threshold secret sharing with classical and quantum secrets. The tripartite case is the simplest nontrivial setting and is central to the $u(g)$3 CC, CQ, and QQ protocols (Keet et al., 2010).

In the $u(g)$4 CC protocol, a classical secret $u(g)$5 is encoded into graph-state labels so that no single player can access $u(g)$6, while any pair can recover it via explicit stabilizer measurements. In the $u(g)$7 CQ protocol, the dealer is attached to players 2 and 3 with edge weights 2 and 1, and security is certified by

$u(g)$8

which shows that an eavesdropper’s reduced state is uncorrelated with the dealer’s measurement result (Keet et al., 2010).

The tripartite $u(g)$9 QQ protocol is singled out as a perfect threshold scheme. The encoded state is a superposition of labelled graph states, and for $\ket{H_3}$0 the construction is exactly the known $\ket{H_3}$1 perfect qutrit secret-sharing code of Cleve–Gottesman–Lo. Any two players can recover the secret by local operations, whereas no single player can obtain any information (Keet et al., 2010).

These protocols clarify the operational meaning of tripartite graph-state entanglement. In this context the graph-state structure does not merely witness multipartite correlations; it implements the access structure itself through stabilizer products, weighted adjacency, and label shuffling (Keet et al., 2010).

6. Network generation, weighted tripartite graphs, and structural observables

Recent work studies tripartite graph states from two complementary directions: which tripartite graph-like entanglement can be generated inside constrained quantum networks, and how graph-theoretic structure can be inferred from observables of tripartite quantum graph states.

In the triangle network with bipartite sources and shared randomness only, explicit protocols generate genuinely multipartite entangled states with nontrivial fidelity to GHZ targets. The reported fidelities are

$\ket{H_3}$2

and, for square dimensions $\ket{H_3}$3,

$\ket{H_3}$4

together with the general lower bound

$\ket{H_3}$5

Since fidelity above $\ket{H_3}$6 certifies genuine multipartite entanglement, these constructions show that the triangle network can generate genuine tripartite entanglement from pairwise resources alone (Zhou et al., 12 Mar 2025).

The same work also derives an upper-bound criterion for fidelity with connected graph states under LOSR networks with bipartite sources: $\ket{H_3}$7 This uses local complementation to reduce graph states to a standard form indexed by

$\ket{H_3}$8

and thereby sharpens the boundary between realizable and non-realizable tripartite graph-state targets (Zhou et al., 12 Mar 2025).

A different recent direction associates multi-qubit quantum graph states to weighted tripartite graphs with vertex partition $\ket{H_3}$9, allowing edges only between distinct parts. In that framework, edge weights are encoded in the two-qubit gates

$E_A,E_B,E_C$0

and the entanglement of a qubit with the rest of the system is quantified by the entanglement distance

$E_A,E_B,E_C$1

The paper’s main conclusion is that this entanglement is determined by the weights of the edges in the closed neighborhood of the corresponding vertex and by its degree with respect to other sets (Gnatenko, 30 Apr 2026).

Two-qubit correlators in the same framework depend on graph-structural quantities such as the number of non-overlapping neighbors,

$E_A,E_B,E_C$2

the number of common neighbors,

$E_A,E_B,E_C$3

and the number of 4-cycles,

$E_A,E_B,E_C$4

For the simplest triangle tripartite graph, noisy simulations on AerSimulator agree well with the analytical entanglement-distance formulas (Gnatenko, 30 Apr 2026).

Taken together, these results place tripartite quantum graph states at the intersection of entanglement theory, graphical algebra, and graph-theoretic modeling. Ordinary tripartite graph states remain the canonical pairwise-edge construction; hypergraph states show exactly how genuine three-body interactions extend that picture under LU but not under SLOCC; categorical and Z/X calculi recast the same tripartite distinctions as algebraic and topological ones; and qudit, network, and weighted-graph generalizations show that tripartite graph-state structure is simultaneously a classification tool, a protocol resource, and a probe of underlying graph structure (Qu et al., 2013, Salem, 2 Dec 2025, Coecke et al., 2010, Zhou et al., 12 Mar 2025).