Non-Stationary Decoherence in Superconducting Qubits: Memory Multi-Fractional Brownian Motion and a Time-Dependent Quantum Brownian Motion Extension

2026

Read Non-Stationary Decoherence in Superconducting Qubits: Memory Multi-Fractional Brownian Motion and a Time-Dependent Quantum Brownian Motion Extension (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-2605-18914.txt.

Introduction

Mahboob Ul Haq models decoherence in superconducting charge qubits with a drift model built on memory multi-fractional Brownian motion. A time-varying Hurst exponent and memory kernel capture shifting noise and long-range correlations that standard models miss. The quantum extension uses a time-dependent Caldeira-Leggett environment. Four results follow: relaxation and noise act independently on energy decay, varying exponents fit observed spectra better than fixed ones, and adaptive kernels preserve correlations with no added damping. Simulations predict coherence times matching theory when charge noise dominates.

Catalog record

YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE39016
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.18914

Browse the full library · About the Tower of Babel

Full text

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

Non-Stationary Decoherence in Superconducting Qubits: Memory Multi-Fractional Brownian Motion and a Time-Dependent Quantum Brownian Motion Extension Abstract Building upon our prior work [ 1 ] , we present a unified stochastic drift model (SdM) for superconducting charge qubits based on memory multi-fractional Brownian motion (mmfBm). The classical sector employs a time-dependent Hurst exponent H ⁡ ( t ) H(t) and adaptive memory kernel K ⁡ ( t , s ) K(t,s) , capturing non-stationary 1 / f β 1/f^{\beta} noise and long-range temporal correlations inaccessible to conventional models. The quantum extension is formulated via a time-dependent Caldeira–Leggett environment with spectral density J ( ω ; t ) = η ( t ) ω c 1 − s ⁡ ( t ) ω s ⁡ ( t ) e − ω / ω c J(\omega;t)=\eta(t)\,\omega_{c}^{1-s(t)}\omega^{s(t)}e^{-\omega/\omega_{c}} , where s ⁡ ( t ) = 2 ​ H ​ ( t ) − 1 s(t)=2H(t)-1 , consistently reproducing β ⁡ ( t ) = 2 ​ H ​ ( t ) − 1 \beta(t)=2H(t)-1 . Four central results emerge: (1) relaxation and noise amplitudes act independently on energy decay; (2) time-varying H ⁡ ( t ) H(t) matches experimental 1 / f 1/f spectra more accurately than any constant exponent; (3) adaptive kernel dynamics preserve correlations without artificial damping; and (4) simulations predict coherence times ( T 1 ∼ 5.00 × 10 6 T_{1}\sim 5.00\times 10^{6} ns, T 2 ∼ 4.18 × 10 5 T_{2}\sim 4.18\times 10^{5} ns) consistent with theory when charge noise dominates. The qubit exhibits stretched-exponential Ramsey and echo decay, non-Markovian dephasing, and a temperature-driven quantum-to-classical crossover. We derive the effective time-local Lindblad master equation, establish the classical mmfBm limit at high temperatures, and provide experimentally testable scaling relations. The non-exponential decay patterns reveal fundamental limitations of Markovian approaches, and the framework guides the design of noise-resilient qubit architectures. Keywords: qubit decoherence, multi-fractional Brownian motion, 1 / f 1/f noise, Caldeira–Leggett model, open quantum systems, non-Markovian dynamics. Mahboob Ul Haq August 10, 2026 1 Introduction Quantum information can be encoded in a variety of physical platforms, including superconducting circuits [ 2 ] , trapped ions [ 3 ] , nuclear magnetic resonance systems [ 4 ] , nitrogen-vacancy centers in diamond [ 5 ] , semiconductor quantum dots [ 6 ] , and neutral atom arrays [ 7 ] . Superconducting qubits are exquisitely sensitive to low-frequency environmental fluctuations originating from dielectric defects, fluctuating two-level systems (TLS), non-equilibrium quasiparticles, and electromagnetic circuitresonances [ 8 , 9 , 10 ] . Measurements of Ramsey and spin-echo coherence consistently reveal non-exponential decay and long-memory temporal correlations that are incompatible with white-noise or purely Markovian models [ 11 ] . Any realistic quantum system interacts with its environment, becoming entangled with numerous external degrees of freedom. This entanglement drives decoherence—the dynamical suppression of quantum coherence as information leaks irreversibly into the environment [ 12 , 13 , 14 ] . Experimental noise spectroscopy typically reports an approximate power-law spectral density S(\omega)\propto\frac{1}{\omega^{\beta}},\qquad 0.8\lesssim\beta\lesssim 1.3, (1) characteristic of generalized 1/f noise. In this manuscript \beta always denotes the exponent of the power spectral density (PSD) of the effective noise process that directly modulates the qubit frequency. Two features observed in modern devices cannot be captured by stationary Gaussian models alone: (i) slow temporal drift of the spectral exponent \beta(t) , (ii) long-range temporal correlations extending over many gate cycles. To address both aspects we introduce memory multi-fractional Brownian motion (mmfBm), a Gaussian process with a slowly varying local Hurst exponent H(t)\in(1/2,1) . The model interpolates continuously between different long-memory regimes while retaining non-local temporal correlations. The second part of this work embeds the stochastic model into a microscopic quantum environment via a time-dependent Caldeira–Leggett construction [ 15 , 16 , 17 , 13 ] . This establishes the classical mmfBm description as the high-temperature adiabatic limit of a non-stationary quantum bath and provides a unified picture of low-frequency decoherence. 2 Classical Memory Multi-Fractional Brownian Motion Model 2.1 Deterministic qubit dynamics We describe the effective qubit coordinate by \chi(t) , a reduced collective variable such as gate charge or flux. In the absence of environmental perturbations the deterministic evolution is \frac{d\chi}{dt}=-\frac{\partial V(\chi)}{\partial\chi},\qquad V(\chi)=E_{J}\cos\chi+\frac{E_{C}}{2}\,\chi^{2}, (2) where E_{J} and E_{C} are the Josephson and charging energies, respectively. 2.2 Definition and properties of mmfBm Memory multi-fractional Brownian motion M(t) is defined through the Riemann–Liouville fractional integral of standard Wiener noise [ 18 , 19 , 20 ] : \boxed{M(t):=\frac{1}{\Gamma\!\big(H(t)+\tfrac{1}{2}\big)}\int_{0}^{t}(t-s)^{H(t)-1/2}\,dW(s),} (3) where W(t) is standard Brownian motion, H(t)\in(1/2,1) is the time-dependent Hurst exponent, and \Gamma(\cdot) is the Gamma function. The kernel (t-s)^{H(t)-1/2} generates algebraically decaying memory, while the slow variation of H(t) introduces non-stationarity. Remark 2.1 (Local roughness) . For a fixed time window centered at t , the process behaves locally as fractional Brownian motion with Hurst index H(t) . For the special case H(t)=H=\text{constant} , the mmfBm reduces to standard fractional Brownian motion (fBm) [ 18 ] . Its covariance is exactly \mathbb{E}[M(u)M(v)]=\frac{1}{2}\Bigl(u^{2H}+v^{2H}-|u-v|^{2H}\Bigr), (4) which serves as an essential consistency check. For general slowly varying H(t) , the rigorous covariance kernel follows from the Riemann–Liouville definition [ 21 ] : C_{M}(u,v)=\frac{\min(u,v)^{H(u)+H(v)}}{2\,\Gamma\!\big(H(u)+\tfrac{1}{2}\big)\,\Gamma\!\big(H(v)+\tfrac{1}{2}\big)}\;{}_{2}F_{1}\!\Bigl(\tfrac{1}{2}-H(u),\,H(v)-\tfrac{1}{2};\,H(u)+\tfrac{1}{2};\,\tfrac{\min(u,v)}{\max(u,v)}\Bigr), (5) where {}_{2}F_{1}(a,b;c;z) is the Gauss hypergeometric function. Equation ( 5 ) is exact but computationally demanding. When the relative variation of H(t) is small, \bigl|H^{\prime}(t)\bigr|\,t\ll H(t), (6) we may employ the locally stationary approximation C_{M}(u,v)\;\approx\;\frac{1}{2}\Bigl(u^{2\overline{H}}+v^{2\overline{H}}-|u-v|^{2\overline{H}}\Bigr),\qquad\overline{H}\equiv\frac{H(u)+H(v)}{2}. (7) 2.3 Spectral properties of the noise process The power spectral density of M(t) is obtained from the Fourier transform of the stationary covariance in the constant- H limit. For fractional Brownian motion the PSD scales as \omega^{-(2H+1)} . However, the physically relevant quantity for qubit dephasing is the increment process \Delta M(t)=M(t+\delta t)-M(t) , which has the character of fractional Gaussian noise (fGn). Its PSD behaves as [ 18 , 22 ] \boxed{S_{M}(\omega;t)\;\propto\;|\omega|^{-(2H(t)-1)}\;\equiv\;|\omega|^{-\beta(t)},}\qquad\beta(t)=2H(t)-1. (8) Thus, H(t) and the experimentally measured noise exponent \beta(t) are directly linked. A value H\approx 0.7 corresponds to \beta\approx 0.4 , consistent with typical 1/f observations [ 9 ] . 2.4 Stochastic evolution equation Since H(t)>1/2 , the sample paths of M(t) are Hölder continuous with exponent arbitrarily close to H_{\min}-1/2 . Consequently, the stochastic integral \int G\,dM is well defined in the Young sense [ 23 ] . The reduced qubit variable evolves as \chi(t)=\chi(0)+\int_{0}^{t}F\bigl(s,\chi(s)\bigr)\,ds+\sigma\int_{0}^{t}G\bigl(s,\chi(s)\bigr)\,dM(s), (9) where F is the deterministic drift from Eq. ( 2 ), G modulates the noise coupling, and \sigma sets the overall noise amplitude. 2.5 Variance scaling For small deviations about equilibrium, G(s,\chi(s))\approx\text{const} , and the variance becomes \langle\chi(t)^{2}\rangle\;\approx\;\sigma^{2}\int_{0}^{t}\frac{(t-s)^{2H(s)-1}}{\Gamma^{2}\!\big(H(s)+\tfrac{1}{2}\big)}\,ds. (10) For slowly varying H(t) , a saddle-point evaluation yields the asymptotic scaling \boxed{\langle\chi(t)^{2}\rangle\;\approx\;\frac{\sigma^{2}}{2H(t)\,\Gamma^{2}\!\big(H(t)+\tfrac{1}{2}\big)}\,t^{2H(t)}\Bigl[1+\mathcal{O}\!\Bigl(\frac{H^{\prime}(t)}{H(t)}\,t\ln t\Bigr)\Bigr].} (11) The logarithmic correction reflects the non-stationary drift of the local roughness exponent. 2.6 Stochastic Simulation of Energy Fluctuations in Superconducting Charge Qubits To investigate decoherence in superconducting charge qubits, we modeled the energy evolution \epsilon(t) using a stochastic integro-differential equation driven by memory multi-fractional Brownian motion (mmfBm): d\epsilon(t)=-\lambda\epsilon(t)\,dt+\sigma(t,\epsilon(t))\int_{0}^{t}K(t,s)\,dB^{H}(s), (12) where \lambda denotes the relaxation rate and the kernel K(t,s)=\begin{cases}(t-s)^{H(t)-1/2},&t>s,\\ 0,&\text{otherwise},\end{cases} (13) introduces time-dependent long-range memory through the variable Hurst exponent H(t) . The simulations demonstrate that the decoherence dynamics are highly sensitive to the interplay between drift, stochastic noise, and memory effects. Increasing the relaxation rate \lambda accelerated energy dissipation Fig 1 , while larger

Read the complete text