Oscillation probabilities for a PT-symmetric non-Hermitian two-state system

1. Problem Statement

The paper addresses a longstanding theoretical challenge in quantum field theory: formulating a consistent description of transition probabilities (specifically flavor oscillations) in systems governed by non-Hermitian but PT-symmetric Hamiltonians.

• The Conflict: While non-Hermitian quantum theories with PT-symmetry (where $[H, PT] = 0$) are known to possess real energy spectra and unitary time evolution under specific conditions, previous attempts to calculate transition probabilities in such systems have yielded unphysical results. Specifically, earlier analyses produced probabilities that were either negative or exceeded unity, violating the fundamental principles of probability theory and perturbative unitarity.
• The Root Cause: The failure of previous formulations stemmed from an incorrect choice of basis and inner product. Standard approaches often relied on the Dirac inner product (Hermitian conjugation) or failed to ensure that the interaction eigenstates (flavor states) and energy eigenstates were orthonormal with respect to the same positive-definite inner product. This led to a violation of time-translation invariance in the calculated probabilities.

2. Methodology

The authors develop a rigorous framework for calculating oscillation probabilities in a two-state system using a specific non-Hermitian scalar field theory model, which they then generalize to fermion mixing (neutrinos).

• Model System: They begin with a Lagrangian density for two complex scalar fields $\phi_1$ and $\phi_2$ with a non-Hermitian mass matrix $M^2 \neq (M^2)^\dagger$. The mass matrix contains off-diagonal terms $\pm \mu^2$ that break Hermiticity but preserve PT-symmetry.
• Inner Product Formalism:
  • The authors utilize the C'PT inner product (where C' is an additional discrete symmetry operator distinct from charge conjugation) to define norms. This ensures positive-definite norms and unitarity.
  • They demonstrate that the standard Dirac inner product fails to maintain orthogonality between flavor states at $t \neq 0$, leading to time-dependent norms and unphysical probabilities.
• Basis Construction: The core methodological breakthrough is the identification of the correct basis to span the state space.
  • Instead of using the standard flavor kets $|\phi_1\rangle$ and $|\phi_2\rangle$, the authors propose spanning the space with $\{|\phi_1\rangle, |\phi_2^{C'}\rangle\}$ (or equivalently $\{|\phi_1^{C'}\rangle, |\phi_2\rangle\}$).
  • This specific choice ensures that the flavor states and the energy (mass) eigenstates are orthonormal with respect to the same positive-definite C'PT inner product.
• Probability Calculation: Transition and survival probabilities are calculated using projection operators constructed from these correctly normalized states: $\hat{\pi}_j(t) = |\phi_j(t)\rangle \langle \phi_j^{C'PT}(t)|$. The probability is defined as $P_{i \to j} = \text{tr}(\hat{\rho}_i(t_0)\hat{\pi}_j(t))$.

3. Key Contributions

• Resolution of the Unitarity Paradox: The paper provides the first formulation of transition matrix elements in PT-symmetric non-Hermitian theories that strictly satisfies positivity ($0 \le P \le 1$) and perturbative unitarity.
• Identification of the Correct Basis: The authors prove that the choice of basis is not arbitrary; using the C'-conjugate of one of the flavor states is necessary to maintain orthogonality and time-translation invariance.
• Distinction from Hermitian Limits: They clarify that the non-Hermitian results are not simple analytic continuations (via $\mu^4 \to -\mu^4$) of Hermitian results. The behavior of the system near the "exceptional point" is fundamentally different.
• Application to Neutrino Physics: The framework is successfully applied to two-neutrino mixing, demonstrating compatibility with the seesaw mechanism in a non-Hermitian context.

4. Key Results

• Oscillation Probabilities:
  For a two-state system with mixing parameter $\eta$ (related to the non-Hermitian coupling) and phase difference $\Delta \omega \Delta t$, the transition and survival probabilities are:
  $$P_{1 \to 2} = \eta^2 \sin^2\left(\frac{\Delta \omega \Delta t}{2}\right)$$
  $$P_{1 \to 1} = 1 - \eta^2 \sin^2\left(\frac{\Delta \omega \Delta t}{2}\right)$$
  These probabilities are strictly bounded between 0 and 1 for all $\eta \in [0, 1]$.
• The Exceptional Point:
  • In the Hermitian case, probabilities saturate only as the mixing parameter $\eta \to \infty$, and the mass eigenvalues diverge (one becomes negative/tachyonic).
  • In the PT-symmetric non-Hermitian case, probabilities saturate at the exceptional point where $\eta \to 1$. At this point, the mass eigenvalues coalesce (become degenerate), but the system remains stable with real eigenvalues.
• Neutrino Oscillations:
  • Applying the formalism to neutrinos, the mixing angle factor $\sin^2(2\theta)$ found in the standard Hermitian case is replaced by $\tanh^2(2\theta)$.
  • The transition probability for $\nu_\mu \to \nu_e$ becomes:
    $$P_{\nu_\mu \to \nu_e} = \tanh^2(2\theta) \sin^2\left(\frac{\Delta \omega \Delta t}{2}\right)$$
  • This result is consistent with the seesaw mechanism, where the mass eigenvalues $m_\pm$ remain real and degenerate at the exceptional point, avoiding tachyonic instabilities.
• Mixing Matrix Structure:
  The non-unitary mixing matrix $V$ in the non-Hermitian case is constrained by pseudo-Hermiticity relations ($VP = PV^{-1}$), resulting in a structure with one mixing angle and one phase, distinct from the standard unitary CKM/PMNS matrices.

5. Significance

• Theoretical Foundation: This work lays the rigorous mathematical groundwork for extending the Standard Model of particle physics to include non-Hermitian sectors. It resolves the "negative probability" issue that has hindered the phenomenological application of PT-symmetric theories.
• Phenomenological Viability: By showing that PT-symmetric theories can accommodate the seesaw mechanism and produce physically valid oscillation probabilities, the paper opens the door to testing non-Hermitian extensions of the Standard Model against experimental data (e.g., meson mixing in $K^0, D^0, B^0$ systems and neutrino oscillations).
• New Physics Signatures: The paper predicts distinct signatures, such as the saturation of oscillation probabilities at the exceptional point ($\eta=1$) rather than at infinite mixing, and the specific behavior of mass eigenvalues near this point. These differences offer potential experimental handles to distinguish between Hermitian and non-Hermitian physics in future high-precision experiments.
• Generalizability: The methodology is not limited to scalars or neutrinos; the authors argue it applies to any two-state system with a PT-symmetric Hamiltonian, providing a template for analyzing CP violation and flavor mixing in broader non-Hermitian quantum field theories.