Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

arXiv:1703.05544 · Physical Review A

Cost of remembering a bit of information

Davide Chiuchiú, Miquel López-Suárez, Igor Neri, M. C. Diamantini, L. Gammaitoni

Published May 8, 2018• 8 citations
arXivHTMLPDF

Abstract

In 1961, Landauer [R. Landauer, IBM J. Res. Develop. 5, 183 (1961)] pointed out that resetting a binary memory requires a minimum energy of ${k}_{B}Tln(2)$. However, once written, any memory is doomed to lose its content if no action is taken. To avoid memory losses, a refresh procedure is periodically performed. We present a theoretical model and an experiment on a microelectromechanical system to evaluate the minimum energy required to preserve one bit of information over time. Two main conclusions are drawn: (i) in principle, the energetic cost to preserve information for a fixed time duration with a given error probability can be arbitrarily reduced if the refresh procedure is performed often enough, and (ii) the Heisenberg uncertainty principle sets an upper bound on the memory lifetime.

DOI
10.1103/physreva.97.052108

Related papers

The cost of information erasure in atomic and spin systems

Joan A. Vaccaro, Stephen Mark Barnett

Landauer's erasure principle, which is one of the defining elements of modern information theory, states the minimum work required to erase 1 bit of information is kT ln(2), where T is the temperature of a thermal reservoir and k is Boltzmann's constant [1]. This principle arises from the maximisation of the entropy of the whole system (memory bit plus thermal reservoir) subject to energy conservation. In this talk we explore a different kind of statistical mechanics based on more general constraints other than energy conservation. Such systems have the potential to provide new mechanisms for erasing information, initialising the state of quantum systems and providing novel "heat" engines. We consider a gas of spin-1/2 systems (e.g. energy degenerate two-state atoms) which comes to equilibrium via angular momentum conserving interactions (e.g. elastic collisions). We model the gas as a canonical ensemble of spin particles constrained by the conservation of angular momentum. The gas acts as a spin reservoir whose macrostate is characterised by the total (fixed) average angular momentum of the gas. Our 1 bit memory device is represented by one spin-1/2 particle in an arbitrary mixed state. The state of the memory is erased by performing controlled operations on the memory particle and ancillary spin-1/2 particles and letting the memory and ancilla particles come to equilibrium with the spin reservoir. The cost of erasing 1 bit of memory is shown to be the angular momentum expended in the controlled operations. In principle there is no energy cost. We then apply this erasure mechanism to the Maxwell's demon problem. Imagine the demon is using the hot molecules of a thermal gas to produce mechanical work and we want to erase the demon's memory. If another thermal reservoir is used for the erasure, the associated energy cost will be greater than the mechanical work extracted by the demon [2]. However, we can avoid this energy cost by using the spin-reservoir erasure mechanism. In principle, all the heat can be extracted from the thermal reservoir as mechanical work for zero energy cost. The second law of thermodynamics and energy conservation are not violated by this process. The cost is in terms of angular momentum expended and an increase in the entropy of the spin reservoir. Moreover, the entropy of the whole system is increased in the process. [1] R. Landauer, IBM J. Res. Lett., 67 661 (1961). [2] C.H. Bennett, Int. J. Theor. Phys. 21, 905 (1982).

•Jan 1, 2008•0 citations

A New Concept of Information Based on Heisenberg's Uncertainty Principle and Its Experimental Verification

Frank Z. Wang

Abstract This study is the first use of Heisenberg's energy-time uncertainty principle to define information quantitatively from a measuring perspective: the smallest error in any measurement is a bit of information, i.e., 1 (bit)=(2∆E ∆t)⁄ℏ. If the input energy equals the Landauer bound, the time needed to write a bit of information is 1.75x10-14 s. Newton's cradle was used to experimentally verify the information-energy-mass equivalences deduced from the aforementioned concept. It was observed that the energy input during the creation of a bit of (binary) information is stored in the information carrier in the form of the doubled momentum or the doubled “momentum mass” (mass in motion) in both classical position-based and modern orientation-based information storage. Furthermore, the experiments verified our new definition of information in the sense that the higher the energy input is, the shorter the time needed to write a bit of information is. Our study may help understand the fundamental concept of information and the deep physics behind it.

•Jul 7, 2021•1 citation

A Quantitative Landauer's Principle

Philippe Faist, Frédéric Dupuis, Jonathan Oppenheim +1 more

Landauer's Principle states that the work cost of erasure of one bit of information has a fundamental lower bound of kT ln(2). Here we prove a quantitative Landauer's principle for arbitrary processes, providing a general lower bound on their work cost. This bound is given by the minimum amount of (information theoretical) entropy that has to be dumped into the environment, as measured by the conditional max-entropy. The bound is tight up to a logarithmic term in the failure probability. Our result shows that the minimum amount of work required to carry out a given process depends on how much correlation we wish to retain between the input and the output systems, and that this dependence disappears only if we average the cost over many independent copies of the input state. Our proof is valid in a general framework that specifies the set of possible physical operations compatible with the second law of thermodynamics. We employ the technical toolbox of matrix majorization, which we extend and generalize to a new kind of majorization, called lambda-majorization. This allows us to formulate the problem as a semidefinite program and provide an optimal solution.

•Nov 5, 2012•30 citations

Discrete Fluctuations in Memory Erasure without Energy Cost

Toshio Croucher, Salil Bedkihal, Joan A. Vaccaro

According to Landauer's principle, erasing one bit of information incurs a minimum energy cost. Recently, Vaccaro and Barnett (VB) explored information erasure within the context of generalized Gibbs ensembles and demonstrated that for energy-degenerate spin reservoirs the cost of erasure can be solely in terms of a minimum amount of spin angular momentum and no energy. As opposed to the Landauer case, the cost of erasure in this case is associated with an intrinsically discrete degree of freedom. Here we study the discrete fluctuations in this cost and the probability of violation of the VB bound. We also obtain a Jarzynski-like equality for the VB erasure protocol. We find that the fluctuations below the VB bound are exponentially suppressed at a far greater rate and more tightly than for an equivalent Jarzynski expression for VB erasure. We expose a trade-off between the size of the fluctuations and the cost of erasure. We find that the discrete nature of the fluctuations is pronounced in the regime where reservoir spins are maximally polarized. We also state the first laws of thermodynamics corresponding to the conservation of spin angular momentum for this particular erasure protocol. Our work will be important for novel heat engines based on information erasure schemes that do not incur an energy cost.

arXiv:1604.05795•Physical Review Letters•Feb 9, 2017•15 citations

Quantum vs Classical Erasure: Equal Bounds but Unequal Costs

Jan Neuser, Jake Xuereb, Pharnam Bakhshinezhad +1 more

Irreversibility has a fundamental thermodynamic cost, erasing information inevitably generates heat. This connection is quantified by the Landauer bound, which gives the minimum dissipation needed to erase a single bit of information. While this bound applies in both classical and quantum settings, it is saturated only in idealised limits of infinite resources. Here, we provide a unified first principles description of finite-resource erasure in both classical and quantum systems. We begin by proving the communal folklore that in the idealised regime the erasure cost of a bit encoded in a quantum or classical system is the same. Despite this, we show that their practical implementation differs substantially: achieving comparable erasure quality in quantum systems requires more control, larger accessible energy gaps and longer operation times. Classical protocols can achieve the erasure of a comparable quantum protocol under far weaker constraints which we expose in trade-off relations. Our results explain why practical erasure schemes fall short of Landauer's bound and show that classical systems enjoy several fundamental thermodynamic advantages.

quant-pharXiv:2607.27341•Jul 31, 2026•0 citations

Qubit reset with a shortcut-to-isothermal scheme

Hongbo Huang, Geng Li, Hui Dong

Landauer's principle shows that the minimum energy cost to reset a classical bit in a bath with temperature T is k_{B}Tln2 in the infinite time. However, the task to reset the bit in finite time has posted a new challenge, especially for quantum bit (qubit) where both the operation time and controllability are limited. We design a shortcut-to-isothermal scheme to reset a qubit in finite time τ with limited controllability. The energy cost is minimized with the optimal control scheme with and without bound. This optimal control scheme can provide a reference to realize qubit reset with minimum energy cost for the limited time.

arXiv:2310.18997•Physical review. E•Jun 13, 2024•2 citations
In 1961, Landauer [R. Landauer, IBM J. Res. Develop. 5, 183 (1961)] pointed out that resetting a binary memory requires a minimum energy of kBTln(2){k}_{B}Tln(2)kB​Tln(2). However, once written, any memory is doomed to lose its content if no action is taken. To avoid memory losses, a refresh procedure is periodically performed. We present a theoretical model and an experiment on a microelectromechanical system to evaluate the minimum energy required to preserve one bit of information over time. Two main conclusions are drawn: (i) in principle, the energetic cost to preserve information for a fixed time duration with a given error probability can be arbitrarily reduced if the refresh procedure is performed often enough, and (ii) the Heisenberg uncertainty principle sets an upper bound on the memory lifetime.
Landauer's erasure principle, which is one of the defining elements of modern information theory, states the minimum work required to erase 1 bit of information is kT ln(2), where T is the temperature of a thermal reservoir and k is Boltzmann's constant [1]. This principle arises from the maximisation of the entropy of the whole system (memory bit plus thermal reservoir) subject to energy conservation. In this talk we explore a different kind of statistical mechanics based on more general constraints other than energy conservation. Such systems have the potential to provide new mechanisms for erasing information, initialising the state of quantum systems and providing novel "heat" engines. We consider a gas of spin-1/2 systems (e.g. energy degenerate two-state atoms) which comes to equilibrium via angular momentum conserving interactions (e.g. elastic collisions). We model the gas as a canonical ensemble of spin particles constrained by the conservation of angular momentum. The gas acts as a spin reservoir whose macrostate is characterised by the total (fixed) average angular momentum of the gas. Our 1 bit memory device is represented by one spin-1/2 particle in an arbitrary mixed state. The state of the memory is erased by performing controlled operations on the memory particle and ancillary spin-1/2 particles and letting the memory and ancilla particles come to equilibrium with the spin reservoir. The cost of erasing 1 bit of memory is shown to be the angular momentum expended in the controlled operations. In principle there is no energy cost. We then apply this erasure mechanism to the Maxwell's demon problem. Imagine the demon is using the hot molecules of a thermal gas to produce mechanical work and we want to erase the demon's memory. If another thermal reservoir is used for the erasure, the associated energy cost will be greater than the mechanical work extracted by the demon [2]. However, we can avoid this energy cost by using the spin-reservoir erasure mechanism. In principle, all the heat can be extracted from the thermal reservoir as mechanical work for zero energy cost. The second law of thermodynamics and energy conservation are not violated by this process. The cost is in terms of angular momentum expended and an increase in the entropy of the spin reservoir. Moreover, the entropy of the whole system is increased in the process. [1] R. Landauer, IBM J. Res. Lett., 67 661 (1961). [2] C.H. Bennett, Int. J. Theor. Phys. 21, 905 (1982).
Abstract This study is the first use of Heisenberg's energy-time uncertainty principle to define information quantitatively from a measuring perspective: the smallest error in any measurement is a bit of information, i.e., 1 (bit)=(2∆E ∆t)⁄ℏ. If the input energy equals the Landauer bound, the time needed to write a bit of information is 1.75x10-14 s. Newton's cradle was used to experimentally verify the information-energy-mass equivalences deduced from the aforementioned concept. It was observed that the energy input during the creation of a bit of (binary) information is stored in the information carrier in the form of the doubled momentum or the doubled “momentum mass” (mass in motion) in both classical position-based and modern orientation-based information storage. Furthermore, the experiments verified our new definition of information in the sense that the higher the energy input is, the shorter the time needed to write a bit of information is. Our study may help understand the fundamental concept of information and the deep physics behind it.
Landauer's Principle states that the work cost of erasure of one bit of information has a fundamental lower bound of kT ln(2). Here we prove a quantitative Landauer's principle for arbitrary processes, providing a general lower bound on their work cost. This bound is given by the minimum amount of (information theoretical) entropy that has to be dumped into the environment, as measured by the conditional max-entropy. The bound is tight up to a logarithmic term in the failure probability. Our result shows that the minimum amount of work required to carry out a given process depends on how much correlation we wish to retain between the input and the output systems, and that this dependence disappears only if we average the cost over many independent copies of the input state. Our proof is valid in a general framework that specifies the set of possible physical operations compatible with the second law of thermodynamics. We employ the technical toolbox of matrix majorization, which we extend and generalize to a new kind of majorization, called lambda-majorization. This allows us to formulate the problem as a semidefinite program and provide an optimal solution.
According to Landauer's principle, erasing one bit of information incurs a minimum energy cost. Recently, Vaccaro and Barnett (VB) explored information erasure within the context of generalized Gibbs ensembles and demonstrated that for energy-degenerate spin reservoirs the cost of erasure can be solely in terms of a minimum amount of spin angular momentum and no energy. As opposed to the Landauer case, the cost of erasure in this case is associated with an intrinsically discrete degree of freedom. Here we study the discrete fluctuations in this cost and the probability of violation of the VB bound. We also obtain a Jarzynski-like equality for the VB erasure protocol. We find that the fluctuations below the VB bound are exponentially suppressed at a far greater rate and more tightly than for an equivalent Jarzynski expression for VB erasure. We expose a trade-off between the size of the fluctuations and the cost of erasure. We find that the discrete nature of the fluctuations is pronounced in the regime where reservoir spins are maximally polarized. We also state the first laws of thermodynamics corresponding to the conservation of spin angular momentum for this particular erasure protocol. Our work will be important for novel heat engines based on information erasure schemes that do not incur an energy cost.
Irreversibility has a fundamental thermodynamic cost, erasing information inevitably generates heat. This connection is quantified by the Landauer bound, which gives the minimum dissipation needed to erase a single bit of information. While this bound applies in both classical and quantum settings, it is saturated only in idealised limits of infinite resources. Here, we provide a unified first principles description of finite-resource erasure in both classical and quantum systems. We begin by proving the communal folklore that in the idealised regime the erasure cost of a bit encoded in a quantum or classical system is the same. Despite this, we show that their practical implementation differs substantially: achieving comparable erasure quality in quantum systems requires more control, larger accessible energy gaps and longer operation times. Classical protocols can achieve the erasure of a comparable quantum protocol under far weaker constraints which we expose in trade-off relations. Our results explain why practical erasure schemes fall short of Landauer's bound and show that classical systems enjoy several fundamental thermodynamic advantages.
Landauer's principle shows that the minimum energy cost to reset a classical bit in a bath with temperature T is k_{B}Tln2 in the infinite time. However, the task to reset the bit in finite time has posted a new challenge, especially for quantum bit (qubit) where both the operation time and controllability are limited. We design a shortcut-to-isothermal scheme to reset a qubit in finite time τ with limited controllability. The energy cost is minimized with the optimal control scheme with and without bound. This optimal control scheme can provide a reference to realize qubit reset with minimum energy cost for the limited time.