We study whether dissipative energy-transfer dynamics can be simulated on noisy near-term quantum hardware by treating device noise as a calibrated resource rather than purely as an error source. Focusing on a biased exciton dimer, we encode the single-excitation manifold into a two-qubit subspace and implement the coherent dynamics through a shallow Trotterized propagator, while repeated noisy identity operations provide an effective dissipative channel. We benchmark the resulting short-time population dynamics against the hierarchical equations of motion (HEOM), which serves as a numerically accurate reference for the corresponding open-system model. On IBM quantum hardware, the calibrated noisy circuit reproduces a broad range of dissipative trajectories in the tested regime, and the fitted HEOM parameters exhibit an approximately linear dependence on the noisy-gate frequency. This empirical relation enables a practically useful interpolation strategy: once calibrated by a finite set of HEOM calculations, the noisy circuit can replace repeated HEOM fitting for intermediate parameter points within the same biased-dimer family. To extend the dynamics beyond the circuit-depth limit, we combine the short-time quantum data with the transfer tensor method (TTM). In simulator studies, TTM accurately extends the dynamics well beyond the directly simulated window, whereas on real hardware its performance is limited by the instability of coherence-sensitive initial states. Our results show that noisy few-qubit devices can act as calibrated phenomenological simulators of open-system dynamics and, within a restricted but experimentally relevant regime, can provide a practical surrogate for repeated HEOM-based modeling.
We study whether dissipative energy-transfer dynamics can be simulated on noisy near-term quantum hardware by treating device noise as a calibrated resource rather than purely as an error source. Focusing on a biased exciton dimer, we encode the single-excitation manifold into a two-qubit subspace and implement the coherent dynamics through a shallow Trotterized propagator, while repeated noisy identity operations provide an effective dissipative channel. We benchmark the resulting short-time population dynamics against the hierarchical equations of motion (HEOM), which serves as a numerically accurate reference for the corresponding open-system model. On IBM quantum hardware, the calibrated noisy circuit reproduces a broad range of dissipative trajectories in the tested regime, and the fitted HEOM parameters exhibit an approximately linear dependence on the noisy-gate frequency. This empirical relation enables a practically useful interpolation strategy: once calibrated by a finite set of HEOM calculations, the noisy circuit can replace repeated HEOM fitting for intermediate parameter points within the same biased-dimer family. To extend the dynamics beyond the circuit-depth limit, we combine the short-time quantum data with the transfer tensor method (TTM). In simulator studies, TTM accurately extends the dynamics well beyond the directly simulated window, whereas on real hardware its performance is limited by the instability of coherence-sensitive initial states. Our results show that noisy few-qubit devices can act as calibrated phenomenological simulators of open-system dynamics and, within a restricted but experimentally relevant regime, can provide a practical surrogate for repeated HEOM-based modeling.
Abstract Quantum simulation represents the most promising quantum application to demonstrate quantum advantage on near-term noisy intermediate-scale quantum (NISQ) computers, yet available quantum simulation algorithms are prone to errors and thus difficult to realize with limited circuit depth on nowadays quantum devices. Herein, we propose a novel scheme to utilize intrinsic gate errors of NISQ devices to enable controllable simulation of open quantum system dynamics without ancillary qubits or explicit bath engineering, thus turning unwanted quantum noises into useful quantum resources. Specifically, we simulate the energy transfer process in a photosynthetic dimer system on IBM-Q cloud. By employing tailored decoherence-inducing gates, we show that quantum dissipative dynamics can be simulated efficiently across coherent-to-incoherent regimes with results comparable to those of the numerically exact classical method. Moreover, we demonstrate a calibration routine that enables consistent and predictive simulations of open-quantum system dynamics in the intermediate coupling regime. This work provides a new direction for quantum advantage in the NISQ era.
Open quantum systems are central to quantum optics, condensed matter, and chemistry, yet their simulation remains challenging for both classical and near-term quantum hardware. In this work we implement and execute utility-scale quantum circuits that accurately reproduce the dissipative dynamics of interacting qubits. We consider a one-dimensional chain of many qubits weakly coupled to a common Markovian bath. The Markovian time evolution of the system is implemented through Trotterized evolution with the introduction of ancilla-assisted dissipative channels, including single-qubit and two-qubit dissipators to capture collective decay. Mid-circuit measurements, conditional gates, and hardware-aware transpilation significantly reduce circuit depth. We further implement a biased Clifford data regression (biased CDR), an error mitigation strategy that outperforms the uniform Cliffordization baseline and a variety of zero-noise extrapolation protocols. We execute large-scale quantum experiments of the dynamics of chains comprising up to 86 emitters on the IBM System Two \texttt{ibm\_basquecountry}. In order to do so, we use 129 total qubits (including ancillas), with the largest circuits contain about 8000 two-qubit gates. To validate these experiments we develop a classical Monte Carlo-Time-Evolving Block-Decimation (MC-TEBD) tensor-network method that incorporates reset operations through stochastic pure-state trajectories, obtaining very good agreement. The approach presented here opens a practical route for utility-scale quantum simulation of dissipative dynamics, enabled by dynamic circuits, targeted error mitigation, and tensor-network validation, and enables to tackle complex dynamics of systems such as quantum emitters in dissipative optical cavities.
Simulating open quantum systems on digital quantum computers typically relies on the use of auxiliary qubits, resulting in overheads of noisy multi-qubit gates, that severely limit execution on near-term hardware. In this work, we explore the simulation of non-Markovian dynamics as well as memory channels leveraging the method of trajectory mixing, valid for mixed unitary channels. This allows to drastically reduce circuit depth, trading entangling gates for a statistical mixture of independent, pure state trajectories. Using a realistic noise model calibrated to modern quantum processors, we show the benefits of this approach, yielding higher state fidelity and better preservation of quantum correlations. This shows the possibility of simulating long-time non-Markovian evolutions with low noise and limited resources.
While simple spin-boson models have been realized on quantum hardware, simulating extended electronic networks with local vibrational environments remains a fundamental challenge in the presence of non-equilibrium, long-lived electronic-vibrational (vibronic) coherence. We present a framework for the digital-analog simulation of open quantum systems governed by Hamiltonians with linear-vibronic coupling (LVC) and structured vibrational environments. Our approach exploits the intrinsic dissipation of qubits in near-term quantum hardware as a resource to emulate vibrational relaxation, combined with a model-specific error mitigation scheme to filter out noise sources incompatible with the target open system. We validate our strategy by resolving the vibronic transfer spectra of a one-dimensional donor-acceptor chain on IBM superconducting processors, reproducing non-Markovian dynamics and scaling the chain length up to 10 electronic sites, an unprecedented scale for chemical dynamics on quantum computers. Our model of vibronic electron transfer offers a portable, application-oriented benchmark for simulating long-lived entangled states on NISQ computers.
While simple spin-boson models have been realized on quantum hardware, simulating extended electronic networks with local vibrational environments remains a fundamental challenge in the presence of non-equilibrium, long-lived electronic-vibrational (vibronic) coherence. We present a framework for the digital-analog simulation of open quantum systems governed by Hamiltonians with linear-vibronic coupling (LVC) and structured vibrational environments. Our approach exploits the intrinsic dissipation of qubits in near-term quantum hardware as a resource to emulate vibrational relaxation, combined with a model-specific error mitigation scheme to filter out noise sources incompatible with the target open system. We validate our strategy by resolving the vibronic transfer spectra of a one-dimensional donor-acceptor chain on IBM superconducting processors, reproducing non-Markovian dynamics and scaling the chain length up to 10 electronic sites, an unprecedented scale for chemical dynamics on quantum computers. Our model of vibronic electron transfer offers a portable, application-oriented benchmark for simulating long-lived entangled states on NISQ computers.
Abstract Simulating open quantum systems (OQS) is essential for understanding quantum thermodynamics and materials science. However, noisy intermediate-scale quantum (NISQ) devices are severely bottlenecked by limited coherence times and gate infidelities. Conventional quantum simulations of Amplitude-Damping Lindblad dynamics requires active dissipative gates and ancilla environments, leading to spurious qubit flips, state saturation, and a significant two-qubit overhead. While recent methods exploit intrinsic hardware noise for quantum simulation, they mainly rely on tailored decoherence-inducing gates or complex error mitigation. Here, we introduce a hardware-intrinsic time-scaling methodology that completely eliminates active dissipative gates. By mapping the processor's continuous energy relaxation (T1) explicitly via native delay instructions, our protocol captures targeted dissipative dynamics with a strict O(1) algorithmic overhead, bypassing conventional mitigation sampling costs. We experimentally demonstrate this efficient approach on two IBM quantum architectures, successfully emulating tight-binding parameter regimes. These include nanosecond magnon propagation in yttrium iron garnet (YIG) spin chains, as well as ultrafast picosecond phonon dynamics across 1D transport channels within 2D molybdenum disulfide (MoS2) and multi-site graphene lattices. Furthermore, to mitigate error accumulation in multi-site arrays, we implement a depth-dependent scaling protocol. Finally, the proposed method is extended to the fermionic domain, emulating the femtosecond relaxation of an isolated charge carrier across 4-qubit topologies. This methodology provides a scalable pathway for robust, hardware-tailored materials simulations.