Triangle-Free Operator Systems

• Triangle-free operator systems are defined by associating a graph’s zero pattern on matrix entries with operator structures, bridging graph theory and quantum information.
• They enable entanglement detection by leveraging local diagonal orthogonal invariance and comparison matrices to transform complex separability tests into tractable positivity criteria.
• Their framework extends to noncommutative graphs, distinguishing between triangle-free and strongly triangle-free systems, and guides the construction of families of PPT entangled states.

Searching arXiv for the specified papers to ground the article and citations. arXiv search query: id:(Singh, 2020) OR id:(Weaver, 22 Sep 2025) Triangle-free operator systems arise in two related but distinct senses in recent research. In bipartite quantum information, a triangle-free graph $G$ on $[d]$ determines an operator system $S_G \subseteq M_d \otimes M_d$ through a prescribed zero pattern on diagonal entries in the product basis, and this structure supports an entanglement-detection method based on local averaging and comparison matrices (Singh, 2020). In noncommutative graph theory, an operator system $\mathcal{V} \subseteq M_n(\mathbb{C})$ is treated as a quantum graph; in that setting, “triangle-free” means the absence of a $3$-clique, while “strongly triangle-free” requires that every rank-$3$ compression contain a rank-$2$ anticlique (Weaver, 22 Sep 2025). The subject therefore connects graph-theoretic sparsity, operator-system structure, and separability or clique-anticlique phenomena in matrix algebras.

1. Graph-supported operator systems from triangle-free graphs

Let $G=(V,E)$ be a simple undirected graph on the vertex set $V=[d]$. It is $\Delta$-free, or triangle-free, if it contains no $[d]$0-cycle. The associated operator system is

$[d]$1

Equivalently, in the product basis $[d]$2, all diagonal matrix entries $[d]$3 vanish away from positions $[d]$4 corresponding either to an edge of $[d]$5 or to $[d]$6 (Singh, 2020).

This construction isolates a graph-controlled support condition inside $[d]$7. In practice, one often works with the larger subspace of local diagonal orthogonal-invariant matrices, but the diagonal-entry pattern remains dictated by $[d]$8. The triangle-free hypothesis is not merely combinatorial: it is the condition under which the separability test developed for these systems collapses to a tractable positivity criterion on auxiliary matrices.

2. Local averaging and the LDOI normal form

Let $[d]$9 be the group of diagonal orthogonal matrices, that is, matrices of the form $S_G \subseteq M_d \otimes M_d$0. The local-averaging map is

$S_G \subseteq M_d \otimes M_d$1

This map is the orthogonal projection onto the subspace of local diagonal orthogonal-invariant (LDOI) matrices. For arbitrary $S_G \subseteq M_d \otimes M_d$2, the averaged state has the form

$S_G \subseteq M_d \otimes M_d$3

with

$S_G \subseteq M_d \otimes M_d$4

Because $S_G \subseteq M_d \otimes M_d$5 is a separable, in fact LOCC, operation, it preserves separability: $S_G \subseteq M_d \otimes M_d$6 (Singh, 2020).

The significance of this reduction is structural. It replaces an arbitrary bipartite state by a canonical LDOI representative whose relevant data are carried by the three matrices $S_G \subseteq M_d \otimes M_d$7. In the triangle-free regime, the resulting separability question is governed by comparison-matrix positivity rather than by a full direct analysis of the original state.

3. Comparison matrices and the graph-theoretic detection criterion

For any matrix $S_G \subseteq M_d \otimes M_d$8, the comparison matrix $S_G \subseteq M_d \otimes M_d$9 is defined by

$\mathcal{V} \subseteq M_n(\mathbb{C})$0

Two graph-theoretic propositions drive the entanglement test. First, if $\mathcal{V} \subseteq M_n(\mathbb{C})$1 is acyclic, that is, a forest, and $\mathcal{V} \subseteq M_n(\mathbb{C})$2 is any matrix-realization of $\mathcal{V} \subseteq M_n(\mathbb{C})$3 with $\mathcal{V} \subseteq M_n(\mathbb{C})$4 exactly on the edges of $\mathcal{V} \subseteq M_n(\mathbb{C})$5, then

$\mathcal{V} \subseteq M_n(\mathbb{C})$6

Second, if $\mathcal{V} \subseteq M_n(\mathbb{C})$7 is triangle-free but contains at least one cycle of length $\mathcal{V} \subseteq M_n(\mathbb{C})$8, then there exists a positive semidefinite $\mathcal{V} \subseteq M_n(\mathbb{C})$9 with $3$0 such that

$3$1

From these facts one obtains the main graph test: $3$2 and, more sharply, for a triangle-free graph $3$3, there exist entangled PPT-LDOI states detectable by this criterion if and only if $3$4 contains an induced cycle of length $3$5. Equivalently, nontrivial detection on $3$6 occurs exactly when $3$7 is not a forest (Singh, 2020).

The criterion therefore separates triangle-free graphs into two classes. Forests never yield a nontrivial comparison-matrix obstruction, whereas triangle-free cyclic graphs do. In that sense, induced cycles of length at least $3$8 are the precise graph-theoretic source of detectable PPT entanglement within this operator-system framework.

4. Construction of PPT entangled families on triangle-free supports

Fix a triangle-free cyclic graph $3$9 on $3$0 vertices. There exists a positive matrix

$3$1

Choose matrices $3$2 and $3$3 such that

• $3$4,
• $3$5,
• for all $3$6,

$3$7

Then

$3$8

is a PPT LDOI state, and because $3$9, it fails the comparison-matrix separability test. This yields a continuous family of PPT entangled states supported on $2$0. By varying over all triangle-free cyclic graphs and all admissible choices of $2$1, one obtains infinitely many inequivalent PPT entangled families in every dimension (Singh, 2020).

A concrete example is the $2$2 $2$3 family. For the $2$4-cycle $2$5, one may take

$2$6

and then choose any $2$7, $2$8 with

$2$9

Each resulting $G=(V,E)$0 is PPT but entangled. The construction is notable because it does not merely prove existence; it gives a recipe for generating large families parameterized by graph support and admissible positive data.

5. Extension beyond the triangle-free case and complexity

For an arbitrary $G=(V,E)$1, not necessarily itself $G=(V,E)$2-free, define

$G=(V,E)$3

If some principal submatrix $G=(V,E)$4 of $G=(V,E)$5, of size $G=(V,E)$6, has graph $G=(V,E)$7 that is triangle-free, then one may project to the corresponding $G=(V,E)$8 subsystem,

$G=(V,E)$9

and apply the same test to $V=[d]$0 and $V=[d]$1. The resulting theorem states that if there exists such a principal submatrix $V=[d]$2 with

$V=[d]$3

then $V=[d]$4 is entangled (Singh, 2020).

The search problem underlying this extension is computationally hard. Finding a triangle-free induced subgraph of maximal size is a known NP-complete problem, since it generalizes the independent-set problem. The worst-case complexity of the associated detection method is therefore NP-hard. This places the method in a common pattern of entanglement theory: the underlying certificate is explicit once the right subsystem is found, but locating that subsystem may itself dominate the computational cost.

6. Triangle-free and strongly triangle-free operator systems as quantum graphs

In the noncommutative setting, a quantum graph on an $V=[d]$5-dimensional Hilbert space is an operator system

$V=[d]$6

that is, a linear subspace containing the identity $V=[d]$7 and closed under the adjoint operation. The ambient space $V=[d]$8 is equipped with the Hilbert–Schmidt inner product

$V=[d]$9

The orthogonal projection onto $\Delta$0 is denoted

$\Delta$1

characterized by

$\Delta$2

The noncommutative adjacency relation is defined by declaring two unit vectors $\Delta$3 adjacent if there exists $\Delta$4 such that

$\Delta$5

For a rank-$\Delta$6 projection $\Delta$7, the compression is

$\Delta$8

A $\Delta$9-clique is a rank-$[d]$00 projection $[d]$01 with

$[d]$02

while a $[d]$03-anticlique is a rank-$[d]$04 projection $[d]$05 with

$[d]$06

An operator system $[d]$07 is triangle-free if it admits no $[d]$08-clique, equivalently if every rank-$[d]$09 projection $[d]$10 satisfies

$[d]$11

It is strongly triangle-free if every rank-$[d]$12 projection $[d]$13 dominates a $[d]$14-anticlique: there exists a rank-$[d]$15 projection $[d]$16 such that

$[d]$17

(Weaver, 22 Sep 2025).

These definitions separate two notions that coincide in the classical graph setting but not in the quantum one. Triangle-freeness forbids full $[d]$18 matrix behavior on a $[d]$19-dimensional subspace; strong triangle-freeness requires, in addition, that every $[d]$20-dimensional compression contain a $[d]$21-dimensional scalar compression.

7. Classification, examples, and the classical–quantum divide

A complete characterization is known for strongly triangle-free operator systems. Weaver’s theorem states that an operator system $[d]$22 is strongly triangle-free if and only if one of the following mutually exclusive cases holds:

1. $[d]$23.
2. $[d]$24.
3. $[d]$25 and, up to unitary equivalence,

$[d]$26

1. $[d]$27 for some orthonormal basis, where

$[d]$28

Several examples and non-examples delimit the scope of these notions. If $[d]$29, then $[d]$30 is triangle-free. If $[d]$31, then $[d]$32 is strongly triangle-free. The diagonal system $[d]$33 has no $[d]$34-anticlique for $[d]$35, so it fails strong triangle-freeness for $[d]$36, and it is triangle-free exactly when $[d]$37. For complete bipartite operator systems $[d]$38, $[d]$39 is strongly triangle-free for every $[d]$40, whereas $[d]$41 with $[d]$42 has a $[d]$43-clique and hence is not triangle-free. There are also nine-dimensional triangle-free systems in $[d]$44: the block-diagonal eight-dimensional system $[d]$45 admits a one-dimensional extension that remains triangle-free, and the skew version

$[d]$46

likewise. The exceptional three-dimensional $[d]$47 system displayed above is strongly triangle-free; every rank-$[d]$48 projection in $[d]$49 compresses it to a system admitting a rank-$[d]$50 anticlique (Weaver, 22 Sep 2025).

The proof strategy for the classification proceeds through compressions to $[d]$51, eigenvalue constraints, and structural reductions. Rank-$[d]$52 anticliques for Hermitian matrices in $[d]$53 are classified first; subsequent lemmas derive balance conditions from $[d]$54-dimensional compressions, show that the presence of a rank-$[d]$55 projection forces $[d]$56, isolate the exceptional two-eigenvalue case with $[d]$57, and then exclude more complicated spectral behavior. In the comparison with classical theory, a graph $[d]$58 may be encoded by an operator system of diagonal matrices plus edge matrix units $[d]$59. Classically, “no triangle” means no induced $[d]$60, equivalently every $[d]$61-subset has a missing edge, hence a $[d]$62-anticlique on that $[d]$63-set. In the quantum case, this equivalence splits: triangle-free means no $[d]$64-dimensional quantum clique, while strongly triangle-free means every $[d]$65-dimensional subspace contains a $[d]$66-dimensional anticlique. There is no simple duality between cliques and anticliques in the quantum setting, and most quantum bipartite operator systems $[d]$67 with $[d]$68 fail triangle-freeness by admitting $[d]$69-cliques.