Qutrit APPT purity for n = 6
Abstract
Literal qutrit APPT density matrices satisfy the sharp purity bound for n = 6.
Theorem 1.1 (The reusable length-18 spectral certificate).
Proof. Machine-checked in Lean as D5/S3/Quantum/Entanglement/AbsolutePPT/QutritN6PurityBound.certificate_bound (✓ std3). ∎
Source. Repository-derived.
Commentary.
For every antitone real 18-coordinate vector of mass one with nonnegative last coordinate and both qutrit boundary matrices positive semidefinite, the spectral square sum is at most 26/400. This exposes the existing kernel-checked rational gap certificate without changing its statement or proof. The four-by-four spectral interpolation consumes this certificate after middle-coordinate insertion or deletion and mass normalization. The hypotheses are spectral LMIs; the certificate itself does not assume or characterize four-by-four APPT.
Theorem 1.2 (Sharp density-matrix purity estimate).
Proof. Machine-checked in Lean as D5/S3/Quantum/Entanglement/AbsolutePPT/QutritN6PurityBound.purity_bound (✓ std3). ∎
Source. Repository-derived.
Commentary.
Every positive semidefinite trace-one complex matrix on Fin 3 × Fin 6 satisfying literal APPT has the stated real trace-square bound. APPT means that the partial transpose of U rho Uᴴ is positive semidefinite for every unitary U, as defined in QutritPerturbationAttainment. The proof applies QutritSpectralReduction.spectral_reduction to obtain ordered spectral certificate coordinates with both boundary LMIs and matching purity, then consumes the public reusable rational gap certificate. All displayed quotients are real division.