Abstract
We use entropy-energy arguments to assess the limitations on the running time and on the system size, as measured in qubits, of noisy macroscopic circuit-based quantum computers.
Introduction
The development of macroscopic quantum computers (MQCs) is often framed as a challenge of engineering—improving gate fidelities, extending coherence times, and scaling up cryogenic infrastructure. However, as the scale of these systems grows, they must also be understood as physical systems subject to the laws of thermodynamics. In this work, we examine the fundamental limits imposed by noise through the lens of entropy-energy balance. By treating a quantum computer as a system driven far from equilibrium, we can derive hard upper bounds on both the duration of a computation and the number of qubits a system can reliably sustain before a thermodynamic "crash" occurs.
The Quantum Computer as a Nonequilibrium System
A quantum computation is essentially a controlled sequence of states that a physical system would not naturally inhabit. In the language of statistical physics, the computer is maintained in a metastable state. While the system's natural tendency is to relax toward a state of maximum entropy (thermal equilibrium with its environment), the application of control Hamiltonians (logic gates) attempts to keep the system in a state useful for information processing.
The model employed here treats the quantum circuit as an alternating sequence of two types of operations:
- Computation Steps: Unitary gates that transform the state.
- Noise Steps: Modeled as trace-preserving completely positive linear maps (channels).
We specifically focus on local stochastic noise, where noise affects qubits individually or in small clusters. These noise channels are assumed to be bistochastic—preserving the maximally mixed state—and strictly contractive, meaning they consistently push the system toward a uniform, high-entropy distribution.
The Entropy-Energy Tug-of-War
The stability of the computation depends on the balance between the entropy generated by noise and the energy resources provided by the control system. Every noise step increases the von Neumann entropy of the system. To counteract this, the "energy" (the work done by the control Hamiltonians) must maintain the system's free energy.
Using the relationship between entropy gain and the shift in free energy ($\Delta F_\beta$), we can determine the point at which the system can no longer be held in its metastable state. If the entropy gain outweighs the energy resources available to the controller, the free energy shift becomes negative, and the useful quantum state collapses into the environment's thermal background.
Fundamental Scaling Laws
The analysis yields two primary scaling laws that define the boundaries of reliable quantum computation in the presence of noise:
1. Temporal Stability (Running Time)
The maximum number of sequential operations, $m$, that a noisy quantum computer can perform is bounded by: $$m = O(\beta E)$$ where $\beta$ is the inverse temperature of the environment and $E$ represents the energy resources available for control. This implies that for any given energy budget and environmental temperature, there is a hard "time limit" after which the computation will inevitably fail due to accumulated entropy.
2. Spatial Stability (System Size)
The maximum number of qubits, $N$, that can be managed reliably is bounded by: $$N = O(e^{\beta E}$$ While this exponential relationship suggests more flexibility in scaling the number of qubits compared to the number of operations, it still imposes a finite limit based on the thermal environment.
Practical Implications: The Case of NMR
These constraints are particularly visible in technologies like Nuclear Magnetic Resonance (NMR) quantum computing. In NMR systems, the thermal energy of the environment is high relative to the energy of the control signals (a small $\beta E$ ratio). Consequently, these systems face severe thermodynamic hurdles when attempting to scale to a large number of qubits or long execution times, aligning with the observed difficulties in scaling NMR-based architectures.
Limitations and Future Directions
The current framework focuses on circuit-based models with limited parallelization. The author suggests that the interaction between spatial and temporal limits might change significantly in massively parallel systems. Furthermore, "one-way" or measurement-based quantum computers may possess intrinsic thermodynamic stability that allows them to bypass these specific circuit-based bounds.
While error correction is the standard solution to noise, this thermodynamic perspective suggests that error correction protocols must be viewed as "entropy drains." Future work is needed to quantitatively integrate specific error-correction codes into this entropy-energy balance to see how they shift the fundamental "crash" point.
Conclusion
By applying classical statistical physics to quantum information, we demonstrate that noise is not merely an engineering nuisance but a source of fundamental thermodynamic limits. The balance between entropy production and energy control determines the ultimate scale and duration of quantum computation, providing a new lens through which to evaluate the feasibility of future macroscopic quantum architectures.
Summary updated