| Landauer's principle | |
|---|---|
| Name | Landauer's principle |
| Discovered by | Rolf Landauer |
| Year | 1961 |
| Field | Thermodynamics; Information theory; Quantum physics |
| Related | Second law of thermodynamics; Maxwell's demon; Szilard engine |
Landauer's principle
Landauer's principle is a physical principle linking information processing to thermodynamics: erasure of one bit of information in a computational device requires a minimum dissipation of energy as heat. It establishes a fundamental lower bound on the work cost of logically irreversible operations and thus constrains physical implementations of computation and information storage in both classical and quantum contexts. The principle is central to discussions of the thermodynamic cost of information, the resolution of Maxwell's demon, and the foundations of statistical mechanics.
Landauer's principle, originally proposed by Rolf Landauer in 1961, states that any logically irreversible manipulation of information, such as the erasure of a bit or the merging of two computational paths, must be accompanied by an entropy increase in non-information-bearing degrees of freedom, typically manifesting as at least k_B T ln 2 of heat dissipated to a thermal reservoir per erased bit, where k_B is the Boltzmann constant and T the reservoir temperature. The result ties a logical operation (erasure) to a thermodynamic cost and implies that logically reversible computations can in principle be performed without this minimal dissipation. The principle plays a role in reconciling information processing with the second law of thermodynamics and informs limits on energy efficiency for both classical and quantum information technologies.
Landauer's statement rests on the identification of information-theoretic entropy (Shannon entropy) with thermodynamic entropy in specific physical implementations. It uses the conceptual framework of statistical mechanics to map probability distributions over logical states to phase-space ensembles. The bound k_B T ln 2 follows when a two-state memory is reset to a standard state irrespective of its initial logical state, reducing the information-bearing entropy by ln 2 and requiring compensation by the environment. Connections are made to Claude Shannon's information theory, the Boltzmann entropy formula S = k_B ln W, and concepts of thermal reservoirs, canonical ensembles, and reversible vs irreversible processes as understood in classical thermodynamics and open quantum systems.
Derivations of Landauer's bound use gedanken experiments and explicit models. The canonical model is the Szilard engine and its one-particle gas description, where measurement and erasure steps are separated; analysis shows work extraction is bounded by information changes. Other models use a particle in a double-well potential representing a memory bit, coupled to a heat bath modeled by the Langevin equation or master equations; quasi-static, isothermal protocols for resetting show the minimal work cost approaches k_B T ln 2. Formal treatments employ stochastic thermodynamics, fluctuation theorems (e.g., Jarzynski equality), and resource-theoretic approaches to nonequilibrium free energy, relating information-processing maps to changes in nonequilibrium free energy and entropy production.
In the quantum regime, Landauer's principle generalizes to quantum bits (qubits) and quantum channels. Quantum extensions consider von Neumann entropy replacing Shannon entropy and account for coherent superpositions, entanglement, and measurement back-action. Key issues include the role of quantum correlations (e.g., entanglement and quantum discord), the thermodynamic cost of quantum measurement and feedback control (as in quantum versions of Maxwell's demon), and the minimal heat for erasure in presence of coherence. Theoretical work links Landauer bounds to quantum thermodynamic resource theories, thermal operations, and one-shot information measures (e.g., min- and max-entropies). Institutions active in this research include Los Alamos National Laboratory, Perimeter Institute for Theoretical Physics, University of Oxford, and Massachusetts Institute of Technology groups working on quantum thermodynamics.
Experimental tests have probed Landauer's bound using microscopic systems: colloidal particles in optical traps, single-electron boxes, superconducting qubits, and trapped ions. Notable experiments include measurements by Tobias Bérut et al. (2012) using a colloidal particle in a double-well potential, and single-electron experiments in single-electron transistors demonstrating heat dissipation near k_B T ln 2. Superconducting circuit experiments and experiments with NV centers and trapped ions have explored erasure and feedback at the quantum limit, measuring work and heat statistics via two-point measurement schemes and fluctuation relations. Results generally confirm the Landauer bound in appropriate regimes, while highlighting practical extra dissipation from finite-time protocols and control imperfections.
Debate around Landauer's principle centers on assumptions about what counts as logical irreversibility, the identification between information and thermodynamic entropy, and the role of measurement and feedback. Critics (including some reinterpretations of Maxwell's demon arguments) have argued that erasure is not uniquely responsible for thermodynamic cost and that environmental interactions or measurement apparatus may account for dissipation. Alternative analyses emphasize the importance of global system-plus-apparatus dynamics, initial correlations, and subtle distinctions between logical and physical reversibility. Formal counterclaims are rare; most disputes concern scope, operational definitions, and extensions to non-equilibrium and quantum-coherent regimes.
Landauer's principle informs limits on energy efficiency for classical and quantum computation, motivating reversible computing architectures (e.g., Fredkin gate, Toffoli gate) and adiabatic logic to approach sub-k_B T dissipation per operation. It guides design of ultra-low-power classical electronics and prospective quantum information processors, suggesting trade-offs among speed, error rates, and heat generation. In thermodynamic engines, information-to-work conversion protocols (information engines) exploit measurement and feedback, constrained by Landauer-type bounds; implementations include feedback-cooled systems and autonomous Maxwell demon models. The principle also influences theoretical studies in nanotechnology, spintronics, and proposals for thermodynamically efficient memory technologies.
Category:Quantum physics Category:Thermodynamics Category:Information theory