Imagine you're running a restaurant kitchen with infinite possible ingredient combinations. Right now, you can only certify dishes as 'good' if they follow a narrow Gaussian recipe — the standard toolkit handles bell curves beautifully but chokes on anything exotic. This paper hands you a finite checklist that works for a much larger menu: any dish assembled by polynomial combinations of your base ingredients. The committed claim: for polynomially excited quantum states built from creation and annihilation operators acting on Gaussian states, the authors reduce the infinite-dimensional problem of certifying NPT (negative partial transpose) entanglement to checking whether a finite-dimensional bilinear representation matrix fails to be positive semi-definite. That's the core move — collapsing an intractable positivity question into linear algebra you can actually compute. The structural result is a monotonicity theorem: polynomial excitations can never destroy NPT entanglement that already exists between any two subsystems, and they can create it between previously separable ones. The number of NPT-entangled pairs is a one-way ratchet under these operations. This is clean and general — it holds for arbitrary multipartite Gaussian states, pure or mixed, with no restriction on the number of modes. The architecture is pure mathematical physics: symplectic diagonalization of covariance matrices, spectral decomposition, and an explicit Fock-to-position-space transformation. No numerics, no variational circuits, no hardware. The method is algebraic and constructive — they build the transformation matrices rather than parameterizing and optimizing them. The 46-page treatment includes a concrete application: quadratic excitations of a three-mode mixed Gaussian state, tracked through unitary time evolution as a function of a non-locality parameter κ, with quantum negativity evaluated explicitly. The time-evolution analysis reveals a striking pattern: under the chosen quadratic Hamiltonian, NPT entanglement persists for long intervals, punctuated only by brief PPT (positive partial transpose) windows. They also derive a closed-form covariance matrix for general quadratic operators across arbitrary degrees of freedom in phase space — a workhorse formula that other groups can plug into immediately. Integrity-wise, this is a theorem paper, not an experimental one. The validation is mathematical proof, which is the gold standard for this type of claim — no cherry-picking risk, no benchmark shopping. The limitation is that the connection to experiment is forward-looking: the authors claim their framework connects to 'experimentally feasible implementations' but don't report any experimental data. The three-mode example is a demonstration of the formalism's tractability, not an independent validation. The gap between this framework and experimental deployment is real but navigable. Photonic quantum information platforms already generate non-Gaussian states via photon addition and subtraction — exactly the polynomial excitations this paper treats. The next milestone is applying these criteria to states actually prepared in a lab and comparing the finite-matrix entanglement verdict against tomographic reconstruction. That's likely a 2-3 year horizon given current photonic state preparation capabilities.
