Unidirectional Inter-Axial Coupling and Spontaneous Cooling in a~Non-Hermitian Dynamics of a~Levitated Particle

1. Problem Statement

Non-Hermitian physics describes open systems where energy, momentum, or information is exchanged with the environment, leading to phenomena like symmetry-breaking transitions and directional energy flow. While Parity-Time (PT) symmetry breaking and non-reciprocal interactions are well-studied in optical resonators and electronic circuits, realizing these effects in mechanical systems with high precision remains challenging.

• Limitations of existing platforms: Multi-particle optomechanical arrays often suffer from parameter inhomogeneity (variations in mass, size, charge) and complex environmental asymmetries, making it difficult to isolate intrinsic non-Hermitian dynamics.
• The Gap: There is a need for a minimal, highly controllable mechanical system where inter-modal coupling can be tuned continuously from reciprocal to non-reciprocal and unidirectional regimes, allowing for the observation of spontaneous energy redistribution without external feedback loops.

2. Methodology

The authors developed an experimental platform based on a single vacuum-levitated silica nanoparticle (diameter $d \approx 987$ nm) trapped by a standing wave of counter-propagating Gaussian laser beams ($\lambda = 1064$ nm).

• Control Mechanisms:
  • Spatial Ellipticity: A Spatial Light Modulator (SLM) adjusts the beam waist radii ($w_{0x}, w_{0y}$), controlling the trap ellipticity $\epsilon = w_{0x}/w_{0y}$.
  • Polarization: A combination of Half-Wave Plates (HWP) and Quarter-Wave Plates (QWP) precisely controls the polarization state (linear, tilted linear, circular). This dictates the nature of the optical forces acting on the particle.
• Theoretical Framework:
  • The system is modeled as two coupled harmonic oscillators (motion along $x$ and $y$ axes) governed by a linearized Langevin equation with a non-symmetric stiffness matrix $K$.
  • The stiffness matrix is constructed as a linear combination of matrices corresponding to principal polarization states ($X, Y, U, V$).
  • Key Physics: Non-conservative optical forces arise when spatial or polarization symmetries are broken, leading to off-diagonal terms $K_{xy} \neq K_{yx}$.
• Data Acquisition & Analysis:
  • Particle position is tracked via a Quadrant Photodiode (QPD) and a high-speed CMOS camera.
  • Stiffness Reconstruction: The authors derived time-dependent covariance functions from particle trajectories and fitted them to analytical solutions of the Langevin equation. This allowed for the direct, model-independent reconstruction of the full stiffness matrix $K$ and the determination of complex eigenfrequencies.

3. Key Contributions

1. Minimal Non-Hermitian Platform: Demonstrated a single-particle system where the two oscillators share identical physical properties (mass, charge, refractive index) and environment, eliminating inhomogeneity issues found in multi-particle arrays.
2. Continuous Tunability: Achieved full control over inter-axial coupling, transitioning the system through:
   • Reciprocal coupling: ($K_{xy} = K_{yx}$)
   • Non-reciprocal coupling: ($K_{xy} \neq K_{yx}$, stable)
   • Unidirectional coupling: One off-diagonal term vanishes ($K_{xy}=0$ or $K_{yx}=0$), allowing energy flow in one direction only.
   • PT-Symmetry Breaking: Transitioning from stable oscillations to regimes with effective gain and loss.
3. Spontaneous Cooling without Feedback: Demonstrated that non-reciprocal coupling can induce spontaneous cooling of one mechanical mode and heating of the other purely through the interplay of non-conservative optical forces and viscous dissipation, requiring no active feedback control or engineered reservoirs.

4. Key Results

• Stiffness Matrix Engineering:
  • Horizontal Linear Polarization ($Q=1$): The stiffness matrix is diagonal ($K_{xy}=K_{yx}=0$); modes are uncoupled.
  • Tilted Linear Polarization ($U=1$): Induces reciprocal coupling ($K_{xy}=K_{yx} > 0$).
  • Circular Polarization ($V=1$): Induces anti-reciprocal coupling ($K_{xy}K_{yx} < 0$). By tuning ellipticity $\epsilon$, the system crosses Exceptional Points (EPs) where eigenvalues coalesce, entering a PT-broken phase with one mode gaining energy (negative damping) and the other losing it.
• Unidirectional Regime:
  • By superimposing specific polarization states (linear $\pi/4$ and circular), the authors canceled one off-diagonal stiffness element while retaining the other.
  • In this regime, motion along one axis drives the orthogonal axis, but there is no reciprocal back-action.
• Spontaneous Cooling:
  • In the non-reciprocal regime between reciprocal and unidirectional limits, the authors observed a violation of the equipartition theorem.
  • One mode exhibited a kinetic energy variance $\langle v^2 \rangle < k_B T/m$ (spontaneous cooling), while the orthogonal mode was heated ($\langle v^2 \rangle > k_B T/m$).
  • This cooling is "spontaneous" as it arises from the intrinsic non-Hermitian dynamics of the optical trap, not external feedback.
• Non-Linear Instability & Phonon Lasing:
  • When the effective damping of the "gain" mode became negative, the system underwent a Hopf bifurcation, leading to limit-cycle oscillations (phonon lasing).
  • This state was characterized by a second-order coherence function $g^{(2)}(0) \approx 1$ (coherent state) and high phase synchronization ($S_r \approx 0.8$) between the $x$ and $y$ modes.

5. Significance

• Fundamental Physics: The work provides a clean, robust testbed for exploring non-Hermitian phenomena, exceptional points, and PT-symmetry breaking in mechanical systems. It clarifies that PT-symmetric spectral transitions can occur via non-conservative forces without explicit gain-loss balancing.
• Thermodynamics: It demonstrates a mechanism for autonomous thermalization and energy redistribution in minimal systems, challenging standard equilibrium thermodynamics in open systems.
• Quantum Applications: By combining these results with recent advances in ground-state cooling, the platform offers a direct route to realizing non-Hermitian phenomena in the quantum regime. This could enable new strategies for quantum state control, enhanced sensing (leveraging the sensitivity near exceptional points), and the creation of synthetic mechanical matter with programmable non-reciprocal interactions.
• Practical Control: The ability to stabilize spinning anisotropic particles via elliptical trap tuning offers immediate utility for manipulating micro-rods and birefringent particles in optical tweezers.