Tomography of a Macroscopic Quantum State influenced by Classical Self-Gravity

Technical Summary: Tomography of a Macroscopic Quantum State influenced by Classical Self-Gravity

Problem Statement
This work investigates the operational distinguishability between two competing frameworks for the dynamics of a macroscopic test mass: standard Quantum Gravity (QG), which is operationally identical to standard quantum mechanics for a single rigid body, and the Schrödinger-Newton (SN) theory. SN theory represents the non-relativistic limit of semi-classical gravity, where a quantum matter distribution couples to a classical gravitational potential generated by its own expectation value. The central problem addressed is how the state-dependent self-gravity term in SN theory affects continuous quantum state tomography of a macroscopic mechanical oscillator. Specifically, the authors ask: if a measurement record is generated by SN dynamics but processed using the standard QG/linear quantum tomography reconstruction map, what are the resulting signatures of model mismatch?

Methodology
The authors analyze a continuous optomechanical system where a macroscopic mirror (mass $M$) is coupled to an optical field via radiation pressure. The system is monitored using variable-angle homodyne detection.

• Theoretical Framework: The dynamics are described using the stochastic master equation (quantum trajectory formalism) in the Schrödinger picture. This choice is preferred over the Heisenberg picture because the SN Hamiltonian contains a conditional mean term ($x_c = \langle \psi_c | \hat{x} | \psi_c \rangle$) that makes the dynamics state-dependent and non-stationary during the finite-time readout.
• Dynamics: The SN Hamiltonian includes a self-trapping term $\frac{1}{2}M\omega_{SN}^2(\hat{x} - \langle \hat{x} \rangle)^2$, which modifies the conditional covariance evolution (Riccati equation) to oscillate at a frequency $\omega_q = \sqrt{\omega_m^2 + \omega_{SN}^2}$, distinct from the bare mechanical frequency $\omega_m$.
• Reconstruction Protocol: The authors adopt a "fixed-map" approach. They assume the data analysis uses the optimal linear filter derived for standard QG (minimizing tomographic error for linear dynamics). This same filter is applied to measurement records generated by both QG and SN dynamics.
• Quantification: The distinguishability between the reconstructed states is quantified using the Hellinger distance between the reconstructed covariance matrices. The analysis explores the dependence of these signatures on measurement strength ($\Omega_q$) and environmental temperature ($T$).

Key Contributions and Results
The paper identifies three primary signatures that arise when an SN-generated record is reconstructed using a QG-optimized map:

1. Angle-Set Dependence of Reconstructed Covariance:
   In standard linear tomography (QG), the reconstructed covariance matrix is independent of the specific set of homodyne angles chosen for the measurement (provided the set is non-singular). In contrast, under SN dynamics, the correction to the tomographic error depends on the full history of the conditional covariance trajectory, which itself is shaped by the homodyne angle. Consequently, applying the fixed QG map to SN data yields different reconstructed uncertainty ellipses depending on the chosen set of tomographic angles. This violates the consistency expected in linear quantum mechanics.

2. Violation of the Gaussian-Covariance Domain:
   In standard QG, the tomographic error is a positive semi-definite noise covariance, ensuring the reconstructed covariance matrix satisfies the Heisenberg uncertainty principle ($\det V \geq \hbar^2/4$). However, the SN-induced correction to the tomographic error is not strictly positive; it can be negative depending on the filter and conditional trajectory.

   • Result: When the QG map is applied to SN data, the inferred covariance matrix can have a determinant below the Heisenberg bound or even become non-positive. This "below-boundary" feature serves as a direct diagnostic that the assumed linear reconstruction model is incompatible with the underlying state-dependent dynamics.

3. Parameter-Dependent Distinguishability:
   The Hellinger distance between the QG and SN reconstructions exhibits a non-monotonic dependence on experimental parameters:

   • Measurement Strength: Strong measurements suppress the SN contribution (by reducing the conditional variance $V_{xx}^c$), causing the SN reconstruction to converge to the QG result. Conversely, weak measurements introduce excessive noise that masks the SN effect. The maximum distinguishability occurs in an intermediate regime where the measurement is strong enough to perform tomography but not so strong as to suppress the SN term.
   • Temperature: High thermal noise increases the standard QG tomographic error without affecting the SN correction (which depends on quantum back-action and conditional variance), thereby reducing the relative distinguishability. Low temperatures enhance the visibility of the SN signatures.

Significance and Claims
The paper positions these findings within the broader context of nonlinear quantum mechanics (NLQM). The authors argue that the SN example illustrates a general obstruction in the tomography of any system where the physical dynamics during readout depend on the state parameters being inferred.

• General Principle: In such nonlinear settings, the measurement process is not a passive projection through a state-independent kernel. Instead, the reconstruction map acquires nonlinear, model-dependent corrections. This leads to a failure of the standard Radon consistency condition (where marginals at different angles must project from a single underlying phase-space distribution).
• Operational Limit: The authors clarify that a "self-consistent" SN tomography protocol for an unknown state cannot be achieved by simply adding a noise term to the standard filter. Because the SN correction depends on the unknown initial covariance itself, the problem becomes a nonlinear inverse problem requiring adaptive or Bayesian estimation strategies, which are left for future work.
• Experimental Feasibility: The paper acknowledges that the required parameters (macroscopic mass in the mHz regime, high quality factor $Q \sim 10^7$, and millikelvin temperatures) are currently at the edge of experimental feasibility. However, it suggests that recent progress in low-frequency optomechanics (e.g., TOBAR) makes this a viable target for future tests.

In summary, the work provides a concrete theoretical framework for distinguishing semi-classical gravity from quantum gravity via state tomography, highlighting that the "mismatch" between the data generation model and the reconstruction model manifests as angle-dependent inconsistencies and violations of standard uncertainty bounds.