Inverse Purcell Suppression of Decoherence in Majorana Qubits via Environmental Engineering

Abstract

We propose a novel approach for optimizing topological quantum devices: instead of merely isolating qubits from environmental noise, we engineer the environment to actively suppress decoherence. For a Majorana qubit in a topological superconducting wire, the exponentially small energy splitting $\epsilon \sim e^{-L/\xi}$ provides protection against local perturbations but renders it highly susceptible to pure dephasing from low-frequency environmental noise. We show that coupling via a parity-conserving operator ($i\gamma_L\gamma_R$) to a bosonic environment yields a dephasing rate $\Gamma_\phi \propto S(\epsilon)$, where $S(\epsilon)$ is the environmental noise power at the qubit splitting frequency. In the experimentally relevant regime where $k_B T \gtrsim \hbar\epsilon$ (with $T \sim 10-100$ mK), the noise power scales as $S(\epsilon) \propto \rho(\epsilon) k_B T/\hbar\epsilon$, leading to a dephasing rate $\Gamma_\phi \propto \rho(\epsilon) T/\epsilon$. This exposes a fundamental challenge: the dephasing rate diverges as $1/\epsilon$ for a standard environment, e.g., a 1D system with linear dispersion where $\rho(\epsilon)$ is constant. We overcome this by designing environments with a suppressed density of states following $\rho_{\text{engineered}}(\epsilon) = \rho_{\text{free}}(\epsilon) (\epsilon/\omega_c)^\alpha$. This creates an ``inverse Purcell effect'' that yields a temperature-independent suppression factor $F_P = (\epsilon/\omega_c)^\alpha$. For $\alpha > 1$, the engineered dephasing rate decreases exponentially with wire length, $\Gamma_{\phi,\text{engineered}} \propto e^{-(\alpha-1)L/\xi}$, meaning longer wires provide better coherence protection. This provides a quantitative design principle where environmental engineering transforms detrimental noise into a tool for coherence stabilization, while respecting fermion parity superselection rules.