Thermodynamic Uncertainty Relations

• Thermodynamic uncertainty relations (TURs) are universal bounds that link the precision of fluctuating currents to the entropy production in nonequilibrium systems.
• They are derived using tools from fluctuation theorems, the Cramér–Rao inequality, and information theory, setting minimal limits on relative fluctuations.
• Extensions to non-steady state, feedback, memory effects, and quantum regimes make TURs essential for optimizing nanoscale engines and transport processes.

Thermodynamic uncertainty relations (TURs) are a class of inequalities that quantify universal trade-offs between the precision of fluctuating currents—such as particle, energy, or entropy fluxes—and the corresponding thermodynamic cost, typically characterized by entropy production, in nonequilibrium systems. Fundamentally, they set lower bounds on the relative magnitude of current fluctuations as functions of dissipation, providing an operational link between information-theoretic constraints, fluctuation theorems, and nonequilibrium thermodynamics in classical, quantum, and active-matter contexts.

1. Foundational Principles and General Formulations

The archetypal TUR for a time-integrated current $J$ over an observation time $T$ takes the form

$$\frac{\mathrm{Var}(J)}{\langle J \rangle^2} \ge \frac{2}{\Sigma}$$

where $\langle J \rangle$ and $\mathrm{Var}(J)$ are the mean and variance of the current, and $\Sigma$ is the total entropy production over $[0,T]$ (Macieszczak et al., 2018, Pal et al., 2019). This universal lower bound, derived for steady-state Markov jump processes and diffusions, quantifies a trade-off: higher precision (lower relative fluctuations) unavoidably entails increased dissipation. In linear response (LR) and for time-reversible systems, the bound becomes tight and plays a critical role in setting ultimate thermodynamic limits (Macieszczak et al., 2018, Falasco et al., 2019).

Beyond the steady-state, the finite-time generalization for arbitrary initial states is (Liu et al., 2019): $$Q_T \equiv \frac{\mathrm{Var}(J)}{[T j(T)]^2} \, \sigma \ge 2$$ where $j(T)$ is the instantaneous current at the final time $T$ and $T$0 is the time-integrated EP. This introduces a "boundary constrained by the bulk" principle, wherein large boundary currents at fixed time require a commensurate increase in either total EP or integrated fluctuations over the entire interval.

A generic family of TURs, valid in any protocol for which a Crooks/Tasaki-type fluctuation theorem holds, reads (Timpanaro, 2024): $T$1 with $T$2 and $T$3, where $T$4 are current means in the forward/backward protocols, and $T$5 is the stochastic EP.

2. Theoretical Foundations: Cramér–Rao, Fluctuation Theorems, and Symmetry

The mathematical derivation of TURs rests on three pillars: Cramér–Rao–type information inequalities, nonequilibrium fluctuation theorems (FTs), and path-space symmetry or involution structures.

• Information-theoretic basis: The Cramér–Rao inequality provides a lower bound on the variance of an unbiased estimator in terms of the Fisher information. In nonequilibrium systems, this machinery is employed by introducing infinitesimal deformations of the dynamics (via parameterized rates or operators), yielding the variance-entropy production bound as a consequence of estimation theory (Liu et al., 2019, Lee et al., 2021, Han et al., 2024).
• Fluctuation theorems: FTs, such as Crooks, Tasaki–Crooks, or exchange FTs, encode symmetry relations for the probabilities of time-reversed trajectories. These enable exact derivations of TURs without recourse to large-deviation asymptotics. The tightest saturable bounds emerge from minimal two-point distributions constrained by these symmetries (Timpanaro, 2024, Ray et al., 2022, Timpanaro et al., 2019).
• Involution and symmetry: Even absent explicit FTs, the existence of a process-inverting involution suffices to enforce a general TUR. For any observable $T$6 odd under this involution, the scaled variance is bounded in terms of path-space Kullback–Leibler divergences, which reduce to entropy production for physical time-reversal (Salazar, 2022, Falasco et al., 2019).

3. Extensions: Non-Steady State, Feedback, Memory, and Non-Markovianity

TURs have been extended beyond stationary Markov setups to cover a wide range of dynamical regimes and constraints:

• Non-steady-state and finite-time protocols: TURs remain valid for time-dependent protocols and arbitrary initial conditions by properly accounting for the time-dependent bulk fluctuation and cumulative dissipation. Notably, the "boundary by bulk" TUR (Liu et al., 2019) provides exponential tightening compared to older exponential bounds in discrete-time processes.
• Feedback control and information thermodynamics: Including feedback-modified dynamics introduces an additional mutual information term to the entropy production, lifting violations observed in experiments on feedback-controlled engines. The modified bound (Liu et al., 2019): $T$7 correctly attributes the apparent violation to the missing information cost.
• Memory and underdamped degrees of freedom: In the presence of non-Markovian friction or underdamped Langevin dynamics, the standard TUR no longer holds in the naive form. The cost-function bounding precision includes both dissipative and relaxation (history-dependent) terms (Terlizzi et al., 2020, Lee et al., 2021). In the memoryless or long-time limit, the classical TUR is recovered.
• Active and renewal processes: For systems driven by colored or self-generated noise (as in active Ornstein–Uhlenbeck processes or renewal-reward processes), the TUR must incorporate both regular entropy production and the energetic cost of activity or renewal variability. Recent results show that cycle-to-cycle fluctuations and random waiting times introduce a precision penalty factor, fundamentally limiting achievable performance in transport and search processes (Han et al., 2024, Karimi et al., 15 Jun 2026).

4. Quantum, Multipartite, and Generalized TURs

Quantum and multipartite generalizations have revealed both stronger constraints and striking violations in specific regimes:

• Quantum coherence and TUR violations: Quantum-coherent conductors, SWAP engines, and harmonic systems can violate the classical TUR by arbitrarily large factors at strong driving or in synchronized, non-Markovian conditions (Timpanaro et al., 2021, Shende et al., 2024, Razzoli et al., 2024, Moustos et al., 12 Nov 2025). Such violations occur when quantum correlations, time-reversal symmetry breaking, or noncommuting observables decouple relative fluctuations from entropy production.
• Generalized and saturable TURs: The tightest possible TURs are derived by variational minimization, showing that only when the current is a function of the trajectory’s entropy production (e.g., $T$8) does the bound reduce to an equality ("thermodynamic uncertainty theorem") (Ray et al., 2022, Timpanaro, 2024). These results incorporate all cumulants of $T$9 and permit continuous interpolation between established TUR bounds.
• Multipartite networks: Multipartite or interacting systems require TURs that constrain global entropy production via local current precisions, possibly augmented by information-theoretic coupling terms, providing bounds even when subsystems violate standard TUR requirements (Kardeş et al., 2021).
• Matrix-valued TURs and current correlations: In systems with several thermodynamic fluxes (e.g., multi-terminal transport), the exchange FT implies matrix-valued TURs that bound not only variances but covariances, with tight, saturable bounds for the entire current covariance matrix (Timpanaro et al., 2019).

5. Methodological Landscape

A broad array of methodologies underpin modern TURs:

• Information geometry: Hilbert-space approaches naturally unify and generalize TURs, elegantly capturing the role of symmetry breaking and the precision bound for odd observables (Falasco et al., 2019).
• Renewal-reward and queueing frameworks: Renewal theory allows for TURs incorporating random event timings, rewards, and coarse-grained phases, with essential corrections for real biological and artificial transport systems (Karimi et al., 15 Jun 2026).
• Fluctuation-dissipation and estimation-theoretic viewpoints: TURs are linked to quantum and thermal fluctuation-dissipation inequalities, which clarify quantum-to-classical crossovers and set fundamental lower bounds on estimation and control (Reiche et al., 2022, Motta et al., 5 Mar 2025).

6. Applications and Physical Implications

TURs are central in establishing fundamental performance bounds and guiding optimization in nonequilibrium statistical physics, molecular biophysics, and quantum thermodynamics:

• Nanoscale machines and energy harvesters: TURs provide design criteria for achieving high current precision at minimum energetic cost in quantum dots, molecular motors, and photonic/quantum devices (Timpanaro et al., 2021, Motta et al., 5 Mar 2025).
• Precision trade-offs in biological systems: Transport precision in run-and-tumble bacteria, chromatin-locus diffusion, and synthetic active matter is fundamentally constrained by the interplay of dissipation and stochasticity as articulated by TURs (Han et al., 2024, Karimi et al., 15 Jun 2026).
• Novel operating regimes: Engineering correlated baths, feedback, or relativistic motion enables both deep TUR violations and nontrivial enhancements of efficiency/power trade-offs, as in relativistic quantum heat engines and synchronized oscillators (Moustos et al., 12 Nov 2025, Razzoli et al., 2024).

7. Limitations, Exceptions, and Future Directions

While TURs are universal under minimal assumptions (FT symmetry or involution), there are important caveats:

• Violation regimes: Quantum coherence, strong non-Markovianity, and certain forms of time-dependent or nonlinear driving can yield arbitrarily large violations of the classical bound (Shende et al., 2024, Timpanaro et al., 2021, Razzoli et al., 2024).
• Beyond entropy production: When memory effects or renewal variability dominate, the cost-function in the TUR cannot be interpreted entirely as physical entropy production, and must be operationally reconstructed from path statistics (Terlizzi et al., 2020, Karimi et al., 15 Jun 2026).
• Tightness and saturation: The tightest TURs are not given by fixed functions of the mean entropy production, but involve the full distribution of entropy production; higher cumulants may raise the lower bound on relative fluctuations (Ray et al., 2022).
• Open questions: Extensions to non-Markovian quantum systems, multidimensional and feedback-controlled currents, and ultimate bounds in systems with dynamical symmetries or memory remain active topics of research. Experimental progress now enables direct tests of TURs in quantum and classical platforms, fostering rapid feedback between theory and measurement (Pal et al., 2019, Shende et al., 2024).

---

Selected References

Main Result or Regime	Key Reference	arXiv ID
General TUR (Markov/Steady)	Barato & Seifert, Gingrich et al.	(Macieszczak et al., 2018)
Finite-time, arbitrary initial	Liu, Guarnieri, Polettini, et al.	(Liu et al., 2019)
Saturable bounds and cumulants	Ray, Landi, Polettini, et al.; Timpanaro	(Ray et al., 2022, Timpanaro, 2024)
Matrix/Covariance-valued bounds	Horowitz et al.	(Timpanaro et al., 2019)
Feedback control and info cost	Liu, Guarnieri, Polettini, et al.	(Liu et al., 2019)
Quantum-coherent violations	Carollo et al.; Sacchi et al.; Razzoli et al.	(Timpanaro et al., 2021, Shende et al., 2024, Razzoli et al., 2024)
Renewal/active process TUR	Karimi & Jafarpour	(Karimi et al., 15 Jun 2026)
Underdamped, Langevin memory	Kim, Chun, & Noh; Di Terlizzi & Baiesi	(Lee et al., 2021, Terlizzi et al., 2020)
Unidirectional transitions	Pal, Reuveni, & Rahav	(Pal et al., 2020)
Quantum foundations, FDI	Dong, Wang, Liu, et al.	(Reiche et al., 2022)