[Paper Review] Extremal Density Matrices for Qudit States

Armando Figueroa, Julio A. López-Saldívar+2 more|arXiv (Cornell University)|Sep 30, 2016

Algebraic structures and combinatorial models13 references3 citations

TL;DR

This paper presents an algebraic method to compute extremal density matrices for qudit systems by solving an underdetermined linear system involving Bloch vector components, SU(d) structure constants, and Hamiltonian parameters. The approach yields extremal mean energy values for both pure and mixed states, recovering the full energy spectrum in the pure case and identifying up to d! extremal mixed states for given energy expectations.

ABSTRACT

An algebraic procedure to find extremal density matrices for any Hamiltonian of a qudit system is established. The extremal density matrices for pure states provide a complete description of the system, that is, the energy spectra of the Hamiltonian and their corresponding projectors. For extremal density matrices representing mixed states, one gets mean values of the energy in between the maximum and minimum energies associated to the pure case. These extremal densities give also the corresponding mixture of eigenstates that yields the corresponding mean value of the energy. We enhance that the method can be extended to any hermitian operator.

Motivation & Objective

• To develop a systematic algebraic procedure for computing extremal density matrices in qudit systems governed by any Hamiltonian.
• To recover the complete energy spectrum of a Hamiltonian through extremal density matrices corresponding to pure states.
• To characterize mixed-state density matrices that extremize the expectation value of energy, yielding minimal and maximal mean energy values.
• To extend the method to any hermitian operator, not just Hamiltonians, using algebraic tools from SU(d) Lie algebra.
• To unify and apply algebraic techniques—such as characteristic polynomials and Bezoutian matrices—for positivity and extremality conditions on density matrices.

Proposed method

• Express the Hamiltonian and density matrix in the generalized Bloch vector formalism using SU(d) generators and traceless hermitian operators.
• Formulate an underdetermined linear system in the Bloch vector components λ_k, using the antisymmetric structure constants f_ijk of SU(d) and Hamiltonian parameters h_k.
• Solve the system by introducing free parameters and applying the characteristic polynomial of the density matrix to determine extremal solutions.
• Enforce positivity of the density matrix using the Bezoutian matrix and its principal minors, ensuring valid quantum states.
• Use the Cayley-Hamilton theorem to express power sums t_k of eigenvalues in terms of the coefficients c_k of the characteristic polynomial.
• Derive explicit positivity constraints for d=2,3,4 by analyzing the determinant and minors of the Bezoutian matrix B_d.

Experimental results

Research questions

• RQ1 How can extremal density matrices for qudit systems be systematically computed for any given Hamiltonian?
• RQ2 What is the algebraic structure linking the Bloch vector, SU(d) structure constants, and Hamiltonian parameters in determining extremal states?
• RQ3 How do the extremal density matrices for mixed states relate to the mean energy values bounded by the pure state spectrum?
• RQ4 What are the complete positivity conditions for qudit density matrices in terms of symmetric functions of eigenvalues for d=2,3,4?
• RQ5 Can the method be generalized beyond Hamiltonians to any hermitian operator using the same algebraic framework?

Key findings

• For pure states, the extremal density matrices fully reconstruct the energy spectrum of the Hamiltonian, including all eigenvalues and their projectors.
• For mixed states, the method identifies extremal mean energy values bounded between the minimum and maximum eigenvalues of the Hamiltonian.
• The number of extremal mixed states for a given energy expectation is at most d! due to the symmetry of the solution space.
• For d=2, the positivity condition reduces to c₂ ≤ 1/4, consistent with the known Bloch sphere constraint.
• For d=3, the compatibility region of (c₂, c₃) is bounded by three inequalities: 0 ≤ c₂ ≤ 1/3, 0 ≤ c₃ ≤ 1/27, and a quartic determinant inequality involving c₂ and c₃.
• For d=4, the full set of positivity conditions includes five inequalities derived from the minors of the Bezoutian matrix B₄, with det(B₄) ≥ 0 being the most restrictive, and the solution space forms a 3D solid with surfaces corresponding to degenerate eigenvalue configurations.