Topological Charge of Causality at a PT-Symmetric Exceptional Point

Kejun Liu · 2026

Read Topological Charge of Causality at a PT-Symmetric Exceptional Point by Kejun Liu (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-00117.txt.

Introduction

Kejun Liu rethinks the Kramers-Kronig relations, the classic links between the absorptive and dispersive parts of a response. Those relations rest on a binary premise: a response function either is analytic in the upper half-plane or is not. Liu shows a single reduced-state transform organizes both outcomes into one sharp dichotomy controlled by microscopic unitarity.

One closed-form function carries both a zero and a pole in the upper half-plane, and a spectral quantity of the reduced propagator decides which appears. On the unitary side the zero is protected and measurable directly from a finite-time record, with the relations intact. On the gain-driven side the zero becomes a genuine pole, the winding number jumps, and the relations gain a correction term. Solvable models realize both sides, and a terahertz signature is proposed.

Catalog record

CREATORKejun Liu
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE17 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.00117

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Topological Charge of Causality at a PT-Symmetric Exceptional Point Kejun Liu Abstract Causality in linear response is conventionally treated as a binary property: a response function is either analytic in the upper half-plane or it is not. We show that in a PT-symmetric open dimer it instead carries a topological charge. As the gain–loss parameter crosses the exceptional point, a single pole of the reflection coefficient migrates into the upper half-plane, the Blaschke winding number jumps from 0 0 to 1 1 , and standard Kramers–Kronig reconstruction acquires a Lorentzian residual fixed by the pole residue. The transition is sharp, protected by the codimension-one structure of the exceptional point, and directly measurable in a one-port reflection experiment. Most strikingly, the violation magnitude scales as Δ KK ∼ | γ − γ c | ν \Delta_{\rm KK}\sim|\gamma-\gamma_{c}|^{\nu} with ν ≈ − 1.08 \nu\approx-1.08 in the single-port geometry: the breakdown of standard KK is strongest at threshold and weakens deeper in the broken phase. We derive the exact reflection coefficient, verify the residue-corrected dispersion relation, and propose a THz time-domain spectroscopy protocol that detects the topological charge through the residual itself. Causality, in the form of analyticity of the response function in the upper half-plane \mathbb{H}_{+} , is one of the most universal constraints in linear-response physics Toll 1956 ; Nussenzveig 1972 . It is usually treated as a binary property: a response either obeys the upper-half-plane analyticity needed for standard Kramers–Kronig (KK) relations, or it fails to do so. In passive systems this binary picture is consistent with thermodynamics. In non-Hermitian systems with engineered gain, from \mathcal{PT} -symmetric photonic dimers Rüter et al. 2010 ; Peng et al. 2014 to driven–dissipative cold-atom and superconducting platforms Özdemir et al. 2019 ; El-Ganainy et al. 2018 , it is not. Recent work has shown two ways in which reduced descriptions can display apparent acausality. Zhang et al. found internal causality breaking in a reduced subsystem driven by initial-state structure rather than by gain/loss dynamics Zhang et al. 2024 . Gavassino showed, in a classical coarse-graining setting, that causal microscopic dynamics can generate effective equations with acausal form Gavassino 2026 . Our question is sharper: can the failure of standard KK itself be organised as a topological transition in an open non-Hermitian system, with an integer invariant that counts how many poles have crossed into the upper half-plane? For the simplest solvable model, a PT-symmetric dimer coupled to a one-port waveguide, the upper-half-plane pole count of the reflection coefficient is a Blaschke winding number \mathcal{N}_{B} . It jumps from 0 to 1 as the gain–loss parameter crosses the exceptional point and is therefore a parameter-space topological charge of causality. This is not just another exceptional-point story: exceptional points can coexist with causality in driven bound-state systems Kulkarni and Joglekar 2025 , so the mechanism must be traced to the open gain–loss dynamics rather than to spectral degeneracy alone. Below we draw this distinction and make it measurable. Model. —The bare PT-symmetric dimer Hamiltonian is H(\gamma)=\bigl(\begin{smallmatrix}-i\gamma&\kappa\\ \kappa&i\gamma\end{smallmatrix}\bigr) with [\mathcal{PT},H]=0 . Coupling site 1 to a one-port bus waveguide via the Gardiner–Collett interaction Gardiner and Collett 1985 ; Walls and Milburn 2008 and applying the Markov approximation yields the single-port effective Hamiltonian H_{\mathrm{eff}}^{\mathrm{SP}}=H(\gamma)-i(\gamma_{\rm ex}/2)|1\rangle\!\langle 1| (derivation in the Supplemental Material SM ). The input–output relation gives the exact reflection coefficient r(\omega;\gamma)=1+i\gamma_{\rm ex}\,G_{11}^{\rm eff}(\omega),\quad G^{\rm eff}=(\omega I-H_{\mathrm{eff}}^{\mathrm{SP}})^{-1}, (1) which is rational of degree two over two with poles at the roots of D^{\rm SP}(\omega)=\omega^{2}+i(\gamma_{\rm ex}/2)\omega+\gamma^{2}+\gamma\gamma_{\rm ex}/2-\kappa^{2} . The upper pole sits in \mathbb{H}_{+} iff \gamma\bigl(\gamma+\tfrac{\gamma_{\rm ex}}{2}\bigr)>\kappa^{2}, (2) with leading-order imaginary part \Imag\,\omega_{+}^{\rm SP}\approx\sqrt{\gamma^{2}-\kappa^{2}}-\gamma_{\rm ex}/4 . This differs by a factor of two from the heuristic formula commonly quoted in the literature, which corresponds to a symmetric two-port convention (see SM SM for the comparison). Figure 1 (a) shows the pole trajectory crossing into \mathbb{H}_{+} at the EP; Fig. 1 (b) maps the phase diagram. Topological invariant and residue-corrected KK. —A response R(z) meromorphic in \mathbb{H}_{+} with finitely many simple poles \{z_{j}\} and residues \{\rho_{j}\} admits a Blaschke factorisation R=B\cdot R_{\rm reg} , where B(z)=\prod_{j}(z-z_{j})/(z-z_{j}^{*}) has unit modulus on \mathbb{R} and R_{\rm reg}\in H^{p}_{+} is the Hardy core on which standard KK applies. The deviations from standard KK are captured by the residue-corrected relations \displaystyle R^{\prime}(\omega) \displaystyle=\frac{1}{\pi}\,\mathrm{P}\!\!\int\!\frac{R^{\prime\prime}(\omega^{\prime})}{\omega^{\prime}-\omega}\,d\omega^{\prime}+2\sum_{j}\Real\!\biggl(\frac{\rho_{j}}{\omega-z_{j}}\biggr), (3) \displaystyle R^{\prime\prime}(\omega) \displaystyle=-\frac{1}{\pi}\,\mathrm{P}\!\!\int\!\frac{R^{\prime}(\omega^{\prime})}{\omega^{\prime}-\omega}\,d\omega^{\prime}+2\sum_{j}\Imag\!\biggl(\frac{\rho_{j}}{\omega-z_{j}}\biggr), (4) derived by Cauchy integration with UHP pole contributions (proof and a correction to a previously stated \Real\leftrightarrow\Imag swap in SM SM ). The Blaschke winding number \mathcal{N}_{B}=(2\pi i)^{-1}\oint R^{\prime}/R\,dz counts UHP poles minus zeros; for the PT-dimer ( 1 ) the numerator has no UHP zeros for \gamma_{\rm ex}\ll\kappa , so \mathcal{N}_{B} counts poles directly. It jumps from 0 to 1 at the EP and is topologically protected by the codimension-one structure of the exceptional point within PT-symmetric matrices Kawabata et al. 2019 . Numerical verification. —We verify Eq. ( 3 ) for the broken-phase reference point (\gamma,\gamma_{\rm ex},\kappa)=(1.5,0.10,1.0) in the SP convention Liu 2026a . The closed-form roots give z_{+}\approx 1.1264\,i (in \mathbb{H}_{+} ) with residue \rho_{+}\approx-0.016\,i . The standard-KK residual exhibits a Lorentzian peak; subtracting the residue contribution reduces the L^{2} norm by a factor of \sim 21 on \omega/\kappa\in[-5,5] (Fig. 3 (a)), confirming the residue-corrected identity. Scaling near the EP. —The EP bifurcation propagates into the standard-KK residual \Delta_{\rm KK}(\gamma)\equiv\|\Real r-\mathcal{H}[\Imag r]\|_{2} . Scanning \gamma/\kappa\in[1.02,1.50] at \gamma_{\rm ex}/\kappa=0.05 yields \Delta_{\rm KK}\sim|\gamma-\gamma_{c}|^{\nu},\quad\nu=-1.080\pm 0.011\ (\text{SP}), (5) with R^{2}>0.997 (Fig. 3 (b); the SYM exponent \nu=-1.195\pm 0.005 is reported in SM SM ). The negative sign corrects the heuristic \nu=+1/2 : as \gamma increases above \gamma_{c} , the pole migrates deeper into \mathbb{H}_{+} and its residue decreases, so the violation peaks at threshold and decays. The residue-corrected residual reduces \|\Delta_{\rm KK}\|_{2} by \geq 15\times at every \gamma>\gamma_{c} . Figure 1: Pole trajectory of the PT-dimer reflection coefficient in the complex z -plane as \gamma/\kappa sweeps across [0,2] , for two values of \gamma_{\rm ex}/\kappa . (a) Single-port coupling. (b) Symmetric coupling. A single pole crosses from the LHP into \mathbb{H}_{+} at the EP. Figure 2: Phase diagram in the (\gamma/\kappa,\gamma_{\rm ex}/\kappa) plane. (a) Single-port: causal ( \mathcal{N}_{B}=0 , white) vs acausal ( \mathcal{N}_{B}=1 , blue). (b) Symmetric: three regions including an overdamped re-entrant causal phase. Figure 3: KK residual at (\gamma,\gamma_{\rm ex})=(1.5,0.1)\kappa . (a) \Real r(\omega) . (b) \Imag r(\omega) . (c) Standard (orange) vs residue-corrected (green) residual; the correction reduces the L^{2} norm by \sim 21\times . Figure 4: Log-log scaling of \|\Delta_{\rm KK}\|_{2} vs \gamma-\gamma_{c} . Circles: SP; squares: SYM. The extracted slope gives \nu_{\rm SP}=-1.08 . Experimental proposal. —The topological charge is directly testable with THz time-domain spectroscopy on a PT-symmetric plasmonic metasurface Bhardwaj et al. 2025 . The protocol: (1) fabricate samples spanning the EP; (2) measure complex transmission and extract \chi_{\rm eff}(\omega) ; (3) compute the standard-KK prediction; (4) fit the residual to the Blaschke ansatz \Delta(\omega)\approx 2\Real[\rho/(\omega-z_{0})] to extract the UHP pole; (5) plot \|\Delta\|_{2} vs |S-S_{\rm EP}| to verify \nu<0 . A passive reference sample serves as null control. A faster proof-of-concept on an electronic RLC dimer with a negative impedance converter Schindler et al. 2011 can validate the pipeline before metasurface fabrication. Discussion. —The Blaschke winding number \mathcal{N}_{B} is a topological invariant in parameter space, in contrast to

Read the complete text