Unique determination and generation of rank two single-qubit quantum channels

AI summary

Overview

Research area: Quantum control theory and quantum information — specifically the characterization (unique determination) and controlled generation of non-unitary quantum channels, with focus on rank-two single-qubit channels.

Technical level: Advanced. The paper is built on linear-algebraic proofs involving Choi matrices, affine subspaces of trace-preserving maps, Schur complements, and the Bloch/Pauli representation of qubit channels, plus a numerically implemented gradient-based control algorithm.

Scope (one sentence): The paper proves a general lower bound on how many input states are needed to uniquely identify a full-rank quantum channel, proves that three suitably chosen states suffice for any rank-two single-qubit channel, and then numerically studies generation of non-unitary single-qubit channels using coherent and incoherent controls.

What This Paper Is About

To claim that a laboratory realization matches a desired quantum channel, one must compare the actual evolution with the target. The practical question is how many input states must be prepared and measured to guarantee that agreement on those states implies agreement of the entire channel. A previous result by Goerz, Reich and Koch (New J. Phys. 16, 055012 (2014)) established that three special density matrices suffice for unitary channels; this paper asks the analogous question for non-unitary channels, the simplest nontrivial class being rank two.

Key Contributions

1. A general lower bound (Theorem 1). For a channel Φ from n×n to m×m matrices with m > 1 whose Choi matrix has full rank n·m, no set of fewer than n² states can uniquely determine it. The proof counts dimensions of the affine space of trace-preserving maps agreeing with Φ on a given set of states.

2. A sharp three-state result for rank-two qubits (Theorem 2). Every single-qubit quantum channel of Kraus rank two is uniquely determined by its action on one explicit set of three density matrices, so that in this case the required number of states is k = 3.

3. A structural theorem for the ambiguity set (Proposition 2). For the specific three-state set used, all channels agreeing with a given qubit channel on those states form an affine space obtained by adding an explicitly described three-parameter family of matrices; the corresponding Choi-matrix image of that family is computed, and the positive-semidefinite part of this affine set is shown to be the single point of the target channel.

4. A control-theoretic and numerical framework. The paper formulates generation of non-unitary single-qubit channels under a GKSL master equation with a coherent control u(t) and an incoherent control n(t), describes the incoherent GRAPE (inGRAPE) gradient method for this setting, and reports numerical investigation of replacement channels and a phase damping channel, with analytic gradient formulas given in appendices.

Main Findings

• Full-rank channels need n² states: Theorem 1 states that a channel Φ ∈ Chan(n,m) with m > 1 and rank Φ = n·m cannot be uniquely determined by fewer than n² states. The argument shows the affine set of trace-preserving maps matching Φ on fewer than n² states has positive dimension, and since a full-rank Choi matrix lies in the interior of the positive semidefinite cone, the affine set intersects the cone in infinitely many points.

• Three states suffice in the rank-two qubit case: Theorem 2 proves that any Φ₀ ∈ Chan(2,2) of the canonical form with parameters u ∈ [0, 2π) and v ∈ [0, π) is uniquely determined by the set S₃ = {ρ₁, ρ₂, ρ₃} with ρ₁ = ½[[1,1],[1,1]], ρ₂ = ½[[1,−i],[i,1]], and ρ₃ = ½·I.

• Unitary invariance transfers the result: Proposition 1 shows that if Φ₁ is uniquely determined by {ρⱼ}, then U Φ₁ V(·)V†U† is uniquely determined by {V†ρⱼV}, so it suffices to prove the theorem for the canonical diagonal-plus-translation normal form of rank-one and rank-two qubit channels.

• The residual freedom is a three-parameter family: Proposition 2 shows that the set of channels agreeing with an arbitrary qubit channel on S₃ is the affine space T_N + W, where W consists of 3×3 blocks with only a nonzero third column; the Choi image of W is a four-by-four matrix with entries built from three real parameters c₁, c₂, c₃. Positive semidefiniteness of the perturbed Choi matrix forces c₁ = c₂ = c₃ = 0.

• Degenerate cases excluded: The Choi matrix of the canonical channel splits into two 2×2 blocks each of rank at most one, and the Kraus rank drops to one at (u, v) = (π, 0) and at (u, v) = (0, 0); these rank-one cases are excluded from the analysis of rank exactly two.

• Numerical generation is target-dependent: The abstract and introduction state that, as one may expect, not all target quantum channels can be generated with high precision, because controllability on the set of quantum channels is generally not established; controllability of open qubit reachable sets has been studied, but controllability on the set of channels still requires investigation.

• Quantitative numerical results: The specific numerical values, fidelities and control parameters from the generation of replacement and phase damping channels are not reported in the truncated content provided here; they appear in Sec. 6 of the full paper, and the notice states that the Version of Record (J. Phys. A: Math. Theor. 59, 355309 (2026)) has a significantly polished numerical part.

Methodology in Plain English

The authors work with three interchangeable descriptions of a quantum channel: as an abstract linear map, as a real matrix acting on Bloch-vector-like coordinates (with an identity component plus generalized Gell-Mann or Pauli matrices), and as a Choi matrix.

The uniqueness question becomes a counting question. Requiring a channel to match a target on one input state imposes a fixed number of linear conditions on the channel's matrix; trace preservation imposes another set. If the number of states is too small, the leftover solution space has positive dimension. For full-rank channels the target Choi matrix sits inside the interior of the positive semidefinite cone, so a small ball around it is entirely made of valid channels, and any line through it inside the solution space gives infinitely many valid channels agreeing on the given states.

For rank-two qubits the counting argument alone is not enough, because the target Choi matrix lies on the boundary. The authors first reduce the problem to a canonical form using unitary freedom on the input and output, then write the generic solution of the matching conditions, and finally check positivity block by block using the Schur complement criterion for Hermitian block matrices. They split the parameters into three cases (u ≠ π−v and u ≠ 2π−v; u = 2π−v; u = π−v) and show in each that any nonzero perturbation either makes a diagonal block indefinite or leaves a block with zero diagonal corner but nonzero off-diagonal coupling, which violates the criterion.

For the control part, the system is an open two-level atom described by a GKSL equation with free Hamiltonian H₀ = ½ω(I₂ − σ_z), coherent drive V = μσ_x with control u(t), and a dissipator whose rates depend on the incoherent control n(t) (the analogous photon-occupation number). Generation of a target channel is posed as optimizing controls so that the evolved outputs on the three determining states match the target outputs; gradients of the objective functionals are derived analytically and used in the inGRAPE gradient-ascent method.

Why This Matters

The result directly reduces the experimental cost of verifying that a realized process equals a desired one. Full channel tomography requires many state preparations and measurements; knowing that three states suffice for rank-two qubit channels cuts the cost by a large factor, since the paper notes that total tomography cost is the cost per state multiplied by the number of states. It also supplies a correct objective functional for numerical optimal-control searches over non-unitary processes.

Real-world applications implied by the paper's cited literature and framing:

• Noisy intermediate-scale quantum computing: characterizing and calibrating non-unitary gates and mixed-state evolution, including measurement-assisted control and feedback on qubits.
• Engineered dissipation: preparing entangled many-body states and stabilizing pointer states through controlled decoherence, where the target object is precisely a non-unitary channel.
• Robust state preparation and polarization control: replacement (all-to-one) channels were proposed for robust control and implemented experimentally as Kraus-map-based transformations.
• Molecular and material control: incoherent control of processes such as retinal isomerization and tailoring kinetic processes in complex materials.

Industry relevance: channel characterization underpins benchmarking and validation of quantum processors and quantum communication links; reducing the number of required probe states lowers the calibration overhead of hardware providers, and the incoherent-control formulation connects to physically motivated noise-engineering schemes rather than idealized unitary gate design.

Future Directions

• Higher-dimensional and higher-rank channels: extend the three-state result beyond single-qubit rank two, for example to qudits or to qubit channels of other Kraus ranks, where the paper notes the general problem is nontrivial and full-rank channels provably require n² states.
• Controllability on the set of quantum channels: the authors explicitly flag that controllability criteria for channel generation in a specific physical model are largely unstudied, in contrast to controllability on the set of density matrices.
• Improving numerical reachability: since not all targets were generated with high precision, identifying which channels lie in the reachable set and designing better control landscapes or functionals is an open task.
• Experimental implementation: the three-state criterion is motivated by reducing tomography cost in the laboratory, so testing it on an actual open qubit with coherent and incoherent controls is the natural next step.

Target Audience

Researchers in quantum control and open quantum systems, mathematical physicists working on channel theory and tomography, and graduate students with a linear-algebra and quantum-information background who are interested in how many measurements are needed to certify a quantum process. Experimentalists building calibration and benchmarking protocols for qubits will find the three-state criterion practically relevant, though the proofs themselves require comfort with Choi matrices, affine geometry, and the Schur complement.

Authors’ abstract

An important problem in quantum technologies is the generation of a target evolution of an open quantum system. A core component of this task is the ability to determine whether the actual evolution of the system coincides with the target evolution. For the generation of unitary quantum channels in open quantum systems, it was shown by Goerz, Reich and Koch [New J. Phys. {\bf 16} 055012 (2014)] that for determining whether the actual evolution coincides with a desired unitary it is sufficient to compare their action on three special density matrices. In this work, we consider controlled generation and unique determination of non-unitary quantum channels in open quantum systems with particular emphasis on rank two single-qubit quantum channels. We prove that to uniquely determine such quantum channels it is sufficient to consider their action on only three suitably chosen density matrices. Based on this theoretical result, we numerically investigate generation of various non-unitary target single-qubit quantum channels in open quantum systems using coherent and incoherent controls.