Mesoscopic Energetics: Bridging Micro and Macro

• Mesoscopic energetics is a framework that bridges microscopic dynamics and macroscopic thermodynamics through coarse-grained free-energy functions and stochastic trajectories.
• It employs Markovian dynamics and large-deviation principles to quantify entropy production, dissipation, and the stability of mesoscopic states.
• The approach underpins diverse applications, from chemical kinetics and DNA landscapes to superconductivity and active matter, by retaining key finite-size effects.

Searching arXiv for recent and foundational papers on mesoscopic energetics across stochastic thermodynamics, coarse-graining, chemical kinetics, active matter, and related mesoscopic frameworks. Mesoscopic energetics denotes a family of intermediate-scale energetic descriptions that sit between microscopic dynamics and macroscopic thermodynamics. In this regime, the relevant objects are neither exact microstates nor purely bulk state variables, but coarse-grained occupancies, stochastic trajectories, basin probabilities, shell configurations, or finite-system order parameters whose fluctuations remain thermodynamically consequential. Across chemical reaction networks, DNA free-energy landscapes, confined dipolar clusters, mesoscopic superconductors, active bacterial swarms, and multiscale thermodynamic reductions, the central aim is to construct free-energy-like functions, entropy-production balances, or dissipation principles that retain finite-size structure, stochasticity, and nonequilibrium driving rather than averaging them away (Ge et al., 2016, Qian et al., 2016, Grmela, 2020, Falasco et al., 2023).

1. Mesoscopic scale and conceptual scope

The mesoscopic scale is consistently presented as an intermediate domain between microscopic and macroscopic descriptions. In multiscale thermodynamics, a level of investigation is defined by both a state space and a theory of observations, and a mesoscopic level is characterized by having both a reduced structure inherited from a more microscopic level and a reducing structure that prepares a lower-level description. In that setting, mesoscopic energetics is inseparable from reduction: dissipation removes detail, while an emergent thermodynamic relation appears at the coarser level (Grmela, 2020).

A complementary formulation appears in coarse-grained classical many-body theory, where the mesoscopic state is built from occupation variables attached to product cells $V_i\times\Omega_\alpha$, with a scale separation

$\xi \ll \ell \ll L,$

so that each cell is large enough to contain many particles but small enough to resolve spatial heterogeneity. The mesoscopic description is therefore neither a purely spatial density field nor a full phase-space density, but a combined spatial and phase-space coarse-graining (Osano, 1 May 2026).

In stochastic thermodynamic formulations, the mesoscopic level is instead the level of Markovian state-to-state dynamics. For chemical systems this is a continuous-time Markov jump process over copy-number vectors; for nonequilibrium thermodynamics more broadly it is a probability distribution over states in phase space or over discrete states; and for macroscopic stochastic thermodynamics it is a Markov jump process whose local detailed balance structure survives scaling to the macroscopic limit (Ge et al., 2016, Qian et al., 2016, Falasco et al., 2023).

Taken together, these works suggest that mesoscopic energetics is not a single formalism but a family of intermediate descriptions linked by a common methodological commitment: fluctuations, finite-size effects, and coarse-grained nonequilibrium structure remain part of the thermodynamic bookkeeping rather than being treated as negligible corrections.

2. State functions, landscapes, and energetic potentials

A recurring feature of mesoscopic energetics is the replacement of a single equilibrium free energy by a broader class of free-energy-like objects. In stochastic chemical kinetics, a generalized mesoscopic free energy for the probability distribution over copy numbers converges, after $V\to\infty$, to a macroscopic chemical energy function $\varphi^{ss}(\mathbf{x})$. This function is not postulated phenomenologically: it emerges as the large-deviation rate function of the stationary distribution, and in the macroscopic limit it governs deterministic chemical relaxation. For detailed-balance systems with mass-action kinetics, $\varphi^{ss}$ becomes the Gibbs function for ideal solutions; for complex-balanced systems it takes the relative-entropy form

$$\varphi^{ss}(\mathbf{x}) = \sum_{i=1}^N \left[ x_i\ln\left(\frac{x_i}{x_i^*}\right)-x_i+x_i^* \right].$$

These forms make mesoscopic energetics a bridge from stochastic kinetics to macroscopic chemical thermodynamics (Ge et al., 2016).

In macroscopic stochastic thermodynamics derived from mesoscopic Markov jumps, the corresponding nonequilibrium state function is the quasi-potential $I_{\rm ss}(c)$. Far from equilibrium, the paper explicitly states that free energy is replaced by this dynamically generated quasi-potential, or self-information, which is a Lyapunov function for the macroscopic deterministic dynamics. Rare transitions between attractors are then controlled by quasi-potential barriers, so stability, escape, and dissipation are all constrained by the same mesoscopic energetic structure (Falasco et al., 2023).

A different but closely related construction appears in sequence-dependent DNA models. There, a coupled Peyrard-Bishop-Dauxois chain and Brownian probe are converted into a conformational Markov network, and a free-energy value is assigned to each node by

$\frac{F_i}{kT}=\log(P_w)-\log(P_i).$

After coarse-graining into basins of attraction, the free-energy difference of a candidate site relative to a nonspecific basin is

$$\frac{\Delta F_\alpha}{k_BT}=-\log\left(\frac{P_\alpha}{P_\beta}\right).$$

This turns sequence-dependent opening fluctuations into a free-energy landscape that can identify transcription start sites and TATA-box-associated basins and compare their relative strengths (Tapia-Rojo et al., 2012).

These examples share a common structural theme: the energetic object is defined on mesoscopic states or distributions rather than on exact microstates, and its principal role is organizational. It selects attractors, ranks basin stability, quantifies driving versus dissipation, or determines the barrier structure of rare transitions.

3. Entropy production, detailed balance, and nonequilibrium driving

Mesoscopic energetics is centrally concerned with entropy production. In mesoscopic stochastic nonequilibrium thermodynamics, the entropy balance is derived directly from Markovian dynamics. For continuous Markov dynamics with probability density $f(x,t)$ and flux $\xi \ll \ell \ll L,$0, the internal entropy production is

$\xi \ll \ell \ll L,$1

while for discrete-state jump processes entropy production admits a cycle representation in which nonequilibrium energy transduction is carried by forward and backward cycle fluxes. In this formulation, chemical potentials, temperature differences, and mechanical forces all appear as cycle affinities driving mesoscopic currents (Qian et al., 2016).

The chemical-kinetic emergence theory makes the same point in a different language. Along the macroscopic dynamics induced by mesoscopic Markov kinetics,

$\xi \ll \ell \ll L,$2

with $\xi \ll \ell \ll L,$3 the entropy production rate and $\xi \ll \ell \ll L,$4 the chemical motive force. For systems with detailed balance, $\xi \ll \ell \ll L,$5, so free energy decreases purely by dissipation. For driven nonequilibrium steady states, $\xi \ll \ell \ll L,$6, and steady fluxes are maintained by a balance of input and dissipation rather than by relaxation to equilibrium (Ge et al., 2016).

Biological nonequilibrium systems sharpen this distinction. In the stochastic thermodynamics of voltage-gated ion channels, total heat is decomposed into excess and housekeeping parts,

$\xi \ll \ell \ll L,$7

where excess heat is associated with adaptation to changing conditions and housekeeping heat is the homeostatic cost of maintaining a nonequilibrium steady state. The trajectory-class fluctuation theorem extends Crooks/Jarzynski-type reasoning to nondetailed-balanced mesoscopic systems, and the application to sodium and potassium channels shows that ignoring housekeeping dissipation can invalidate fluctuation-theorem predictions (Semaan et al., 2022).

A common misconception is that mesoscopic energetics is simply equilibrium thermodynamics with noise added. The large-deviation framework for macroscopic stochastic thermodynamics argues against this directly: replacing the underlying jump process by a Langevin equation with Gaussian white noise is, in general, only a Gaussian approximation and can be thermodynamically inconsistent away from equilibrium, particularly for housekeeping dissipation and current fluctuation symmetries (Falasco et al., 2023).

4. Coarse-graining, reduction, and extensivity

Coarse-graining is not merely a practical simplification in mesoscopic energetics; it is often the source of the thermodynamic structure itself. In the mesoscopic partition-function framework for classical many-body systems, the canonical integral is replaced by a discrete sum over occupation numbers $\xi \ll \ell \ll L,$8 associated with product cells $\xi \ll \ell \ll L,$9, subject to $V\to\infty$0. The central structural equivalence is

$V\to\infty$1

so cellwise factorisation is equivalent to extensivity of the coarse-grained free energy. Departures from additivity are encoded by inter-cell correlation terms $V\to\infty$2, interpreted as mutual-information-like corrections, and they enter a generalized Euler relation

$V\to\infty$3

with $V\to\infty$4 as a subextensive boundary-and-correlation term (Osano, 1 May 2026).

Multiscale thermodynamics provides a more general reduction architecture. Its mesoscopic dynamics is organized by GENERIC, in which reversible Hamiltonian motion and irreversible gradient dynamics coexist:

$V\to\infty$5

The reduction from an upper level to a lower one can be realized dynamically through time evolution or statically through Maximum Entropy. In this framework, entropy is described as the “ambassador of the lower level,” and dissipation is the mechanism that removes microscopic detail and reveals a lower-level thermodynamic pattern (Grmela, 2020).

The DNA promoter analysis offers a concrete algorithmic realization of this logic. Long stochastic trajectories of the coupled DNA-particle system are first generated by Langevin dynamics; the DNA coordinates are then reduced by principal component analysis, with the first five principal components accounting for about $V\to\infty$6 of chain fluctuations; the reduced DNA space is discretized into $V\to\infty$7 bins and particle position into $V\to\infty$8 bins; and the resulting conformational Markov network is decomposed into basins of attraction and represented by a dendrogram. The crucial point is that the free-energy landscape is built from the actual stochastic dynamics of the coupled system rather than from static sequence information alone (Tapia-Rojo et al., 2012).

5. Finite-size structure and experimental manifestations

In finite quantum systems, mesoscopic energetics often appears as a competition among confinement, interaction energy, and quantum fluctuations. For two-dimensional trapped dipolar boson clusters with $V\to\infty$9, increasing confinement drives a sequence from superfluid to supersolid to insulating crystal. In the strong-confinement regime, classical potential-energy minimization predicts the shell structure, but for $\varphi^{ss}(\mathbf{x})$0 and $\varphi^{ss}(\mathbf{x})$1 an intermediate confinement window stabilizes non-classical crystalline ground states through quantum zero-point motion. The result is a specifically mesoscopic form of crystallization in which discrete shell occupancies and finite-size effects change the energetic ordering of structures that are nearly degenerate classically (Boninsegni, 2013).

In mesoscopic nanowires with attractive pairing, the relevant energetic quantity is the parity parameter,

$\varphi^{ss}(\mathbf{x})$2

which measures the odd-even ground-state energy difference. This quantity generalizes the bulk pair-binding scale across the simultaneous bulk-to-mesoscopic and BCS-BEC crossover. In SrTiO$\varphi^{ss}(\mathbf{x})$3 nanowires, the conductance-peak splitting field is identified with the parity energy through

$\varphi^{ss}(\mathbf{x})$4

and the fitted parameters place the device in a fluctuation-dominated mesoscopic regime on the BCS side, with $\varphi^{ss}(\mathbf{x})$5, $\varphi^{ss}(\mathbf{x})$6, and $\varphi^{ss}(\mathbf{x})$7 (Hofmann, 2017).

Near the superconducting transition in disordered thin films, mesoscopic energetics is amplified by emergent superconductivity. As the Ginzburg-Landau coherence length diverges, short-range disorder fluctuations become effective local fluctuations of $\varphi^{ss}(\mathbf{x})$8, producing a singular mesoscopic component of the pair susceptibility. In two dimensions, the rms Aslamazov-Larkin conductivity fluctuation is estimated as

$\varphi^{ss}(\mathbf{x})$9

and the paper emphasizes that mesoscopic conductivity fluctuations cease to be universal and can exceed the quantum normal-state scale by a large factor while remaining below the Drude conductivity. Conductivity, magnetic susceptibility, and transverse thermomagnetic response then share a common singular temperature dependence in the Ginzburg region (Hettinger et al., 2019).

In active matter, mesoscopic energetics becomes directly measurable. Confined Proteus mirabilis swarm clusters in optical tweezers are treated as nonequilibrium stationary mechanical systems whose forces are measured by the Photon Momentum Method and trajectories by Multiple Particle Tracking. The trap stiffness is reported as $\varphi^{ss}$0, and the rise of fluctuating force with cluster size follows a Hill-type law with cooperative onset size $\varphi^{ss}$1. Coarse-graining in force space reveals vortex-like probability currents and broken detailed balance in living clusters but not in dead ones. The work, rotational work, and entropy production are then quantified within a stochastic-thermodynamic framework, with an Ohmic-like dissipative law

$\varphi^{ss}$2

used to describe steady active dissipation (Luque-Rioja et al., 17 Jun 2025).

DNA promoter landscapes provide a further biological manifestation. Strong promoters such as collagen and P5 produce a small number of highly occupied basins associated with the transcription start site and TATA box, with dominant $\varphi^{ss}$3 values around $\varphi^{ss}$4; the lac operon regulatory region behaves as a weak promoter with many more basins and smaller free-energy gaps; and a shuffled P5 control with the same base composition but random order lacks an organized promoter-like hierarchy (Tapia-Rojo et al., 2012).

6. Interpretive extensions and open questions

The phrase “mesoscopic energetics” has also been extended into domains that are conceptually more ambitious than standard stochastic thermodynamics. In mesoscopic spacetime kinetic theory, spacetime is treated analogously to a kinetic medium with a density of states $\varphi^{ss}$5, where $\varphi^{ss}$6 is an internal fluctuating vector associated with each event. A zero-point length yields

$\varphi^{ss}$7

and extremizing the total density of states of geometry plus matter leads to Einstein’s equations, reinterpreted as a zero-heat-dissipation principle through

$\varphi^{ss}$8

In this usage, mesoscopic energetics is a thermodynamic-statistical reinterpretation of gravitational dynamics rather than a theory of matter in the usual condensed or chemical sense (Padmanabhan, 2018).

A different extension appears in irreversible thermodynamics near equilibrium. There, thermodynamic forces and their conjugate variables are promoted to discrete operators obeying canonical commutation rules at the mesoscopic scale, and the total entropy production becomes an operator with a discrete spectrum,

$\varphi^{ss}$9

The paper interprets these quanta as “thermodynamic units of information” and estimates $$\varphi^{ss}(\mathbf{x}) = \sum_{i=1}^N \left[ x_i\ln\left(\frac{x_i}{x_i^*}\right)-x_i+x_i^* \right].$$0 from experiment, with $$\varphi^{ss}(\mathbf{x}) = \sum_{i=1}^N \left[ x_i\ln\left(\frac{x_i}{x_i^*}\right)-x_i+x_i^* \right].$$1 (Sonnino, 2023).

A neo-Gibbsian statistical energetics pushes the intermediate-scale interpretation in yet another direction. It identifies $$\varphi^{ss}(\mathbf{x}) = \sum_{i=1}^N \left[ x_i\ln\left(\frac{x_i}{x_i^*}\right)-x_i+x_i^* \right].$$2 with classical energetics, derives the dual variational relations

$$\varphi^{ss}(\mathbf{x}) = \sum_{i=1}^N \left[ x_i\ln\left(\frac{x_i}{x_i^*}\right)-x_i+x_i^* \right].$$3

and argues that if one differentiates the statistical free energy before taking the $$\varphi^{ss}(\mathbf{x}) = \sum_{i=1}^N \left[ x_i\ln\left(\frac{x_i}{x_i^*}\right)-x_i+x_i^* \right].$$4 limit, mesoscopic energetics with fluctuations emerges in the form of Shannon entropy and relative entropy. The irreversible thermodynamic potential

$$\varphi^{ss}(\mathbf{x}) = \sum_{i=1}^N \left[ x_i\ln\left(\frac{x_i}{x_i^*}\right)-x_i+x_i^* \right].$$5

is then interpreted as a distance-like measure for nonequilibrium states and extended to nonequilibrium cells under constant $$\varphi^{ss}(\mathbf{x}) = \sum_{i=1}^N \left[ x_i\ln\left(\frac{x_i}{x_i^*}\right)-x_i+x_i^* \right].$$6, $$\varphi^{ss}(\mathbf{x}) = \sum_{i=1}^N \left[ x_i\ln\left(\frac{x_i}{x_i^*}\right)-x_i+x_i^* \right].$$7, and $$\varphi^{ss}(\mathbf{x}) = \sum_{i=1}^N \left[ x_i\ln\left(\frac{x_i}{x_i^*}\right)-x_i+x_i^* \right].$$8 (Miao et al., 24 Aug 2025).

These extensions underline an important interpretive point. Mesoscopic energetics is not restricted to one ontology, one set of variables, or one mathematical apparatus. In some papers it denotes stochastic thermodynamics of Markov jump or Langevin-type systems; in others it denotes coarse-grained partition functions, dynamic free-energy landscapes, finite-size quantum stability criteria, operator formulations of entropy production, or thermodynamic reinterpretations of spacetime. A plausible implication is that the unity of the subject lies less in a single formalism than in a recurrent problem: how to formulate energetics when microscopic detail is inaccessible, macroscopic averaging is too severe, and fluctuations, correlations, or finite-size structure remain dynamically decisive.