[Paper Review] Geometry of entanglement witnesses for two qutrits

Dariusz Chruściński, Filip Wudarski|arXiv (Cornell University)|May 24, 2011

Quantum Computing Algorithms and Architecture2 references4 citations

TL;DR

This paper provides a geometric characterization of entanglement witnesses for two qutrits by analyzing a convex set of positive maps in $M_3(\mathbb{C})$, revealing that part of its boundary is parameterized by $SO(2)$ rotations. It conjectures these boundary maps are optimal for detecting entanglement and introduces a new class of decomposable entanglement witnesses parameterized by $O(2)$ improper rotations.

ABSTRACT

We characterize a convex subset of entanglement witnesses for two qutrits. Equivalently, we provide a characterization of the set of positive maps in the matrix algebra of 3 x 3 complex matrices. It turns out that boundary of this set displays elegant representation in terms of SO(2) rotations. We conjecture that maps parameterized by rotations are optimal, i.e. they provide the strongest tool for detecting quantum entanglement. As a byproduct we found a new class of decomposable entanglement witnesses parameterized by improper rotations from the orthogonal group O(2).

Motivation & Objective

• To characterize the convex set of entanglement witnesses for two qutrits via the geometry of positive maps in $M_3(\mathbb{C})$.
• To investigate the role of $SO(2)$ rotations in parameterizing the boundary of the set of positive maps.
• To explore the (in)decomposability and optimality of these maps for detecting quantum entanglement.
• To construct a new class of decomposable entanglement witnesses using improper $O(2)$ rotations.
• To examine the structural symmetry and self-duality properties of the positive maps in this family.

Proposed method

• The authors analyze a one-parameter family of positive maps $\Phi[a,b,c]$ in $M_3(\mathbb{C})$ defined via a circulant matrix structure and normalization factor $N_{abc}$.
• They impose the constraint $a + b + c = 2$ to focus on the boundary of the positive map set, reducing the parameter space to two variables $b$ and $c$.
• The geometric structure is analyzed by mapping the parameters $(b,c)$ to an ellipse in the plane, revealing $SO(2)$ rotational symmetry on the boundary.
• The duality and decomposability of maps are assessed via self-duality conditions and the positivity of associated operators.
• A new class of entanglement witnesses is constructed using improper $O(2)$ rotations, with explicit matrix forms derived for $\widetilde{W}[a,b,c]$ and $W_U[a,b,c]$.
• The paper uses unitary transformations and convex combinations to generate new witnesses and analyze their (in)decomposability.

Experimental results

Research questions

• RQ1 How is the boundary of the convex set of positive maps in $M_3(\mathbb{C})$ geometrically structured?
• RQ2 Are the positive maps parameterized by $SO(2)$ rotations optimal for detecting entanglement?
• RQ3 What is the role of $O(2)$ improper rotations in generating new decomposable entanglement witnesses?
• RQ4 How do self-duality and decomposability relate to the geometric structure of the positive map set?
• RQ5 Can convex combinations of different witness families preserve or alter (in)decomposability?

Key findings

• The boundary of the convex set of positive maps in $M_3(\mathbb{C})$ is parameterized by $SO(2)$ rotations, with the ellipse equation $\frac{9}{4}\left(x-\frac{4}{3}\right)^2 + \frac{3}{4}y^2 = 1$ describing the parameter space.
• Maps corresponding to $SO(2)$ rotations are conjectured to be optimal, as both the Choi map and reduction map—known to be optimal—belong to this class.
• A new class of decomposable entanglement witnesses is constructed using improper $O(2)$ rotations, with explicit matrix forms provided for $\widetilde{W}[a,b,c]$.
• The witness $\widetilde{W}[a,b,c]$ is decomposable and self-dual, while $W_U[a,b,c]$—a unitary transformation of $W[a,b,c]$—is indecomposable when $b \neq c$.
• The two witness families $W[a,b,c]$ and $\widetilde{W}[a,b,c]$ intersect only at $a = b = c = \frac{2}{3}$, corresponding to the center of the ellipse.
• Convex combinations of $W[\alpha]$ and $\widetilde{W}[\beta]$ yield new entanglement witnesses, but their (in)decomposability depends nontrivially on the probability distributions used.