Ergotropy of macroscopic classical systems

arXiv:2603.28388 · cond-mat.stat-mech, physics.plasm-ph · Submitted 2026-03-30 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Condensed Matter. The day in condensed matter: materials, phases and the measurements under them.

Kai: I'm Kai.

Mira: And I'm Mira, here with Lev.

Kai: We're looking at the paper: "Ergotropy of macroscopic classical systems".

Paper summary: Mira: To get into the details of "Ergotropy of macroscopic classical systems," the authors focus on providing a general expression for ergotropy that works even when the phase space density has some jumps or flat regions.

Kai: They frame this problem as a function rearrangement problem, which is what leads to this new idea of "ergotropic rearangement" as a generalization of symmetric decreasing rearrangement.

Lev: The paper shows how they use assumptions about the phase volume and the density to derive a specific form for ergotropic rearrangement, which is equation nineteen in their text.

Mira: Equation nineteen gives us this general expression for the ergotropic rearrangement, which is defined by integrating over all possible values of r of sigma minus omega zero of h zero.

Kai: This general expression then allows them to write down an explicit form for the ergotropy, which they call E-two.

Lev: They also show how this relates back to the simpler symmetric decreasing rearrangement when the Hamiltonian is spherically symmetric, like a quadratic one, where it matches rho star.

Mira: The paper then looks at the thermodynamic limit and finds that any density of the form rho equals f of H zero becomes asymptotically passive under these conditions.

Kai: And this asymptotic passivity happens even when we consider a system with an ideal gas evenly distributed on an energy shell.

Lev: That's a big statement because it implies that for very large systems, the available energy you can extract from these states really tends to zero.

Mira: So, the thesis is that the general machinery of ergotropic rearrangement leads to this conclusion about asymptotic passivity in the thermodynamic limit.

Kai: It’s a lot of theory leading to a statement about what happens when you scale up physical systems.

Lev: And they also note that they need to assume invertibility for some steps in their derivation, which means they are relying on those densities not having flat regions or jump discontinuities.

Conclusion: Kai: So, wrapping up the "Ergotropy of macroscopic classical systems" paper, the main thing is this connection between energy extraction and mathematical rearrangement.

Mira: They establish that any equilibrium state of the form f of H zero is asymptotically passive when we look at these thermodynamic limits.

Lev: This finding strengthens some mechanical foundations for things like the Kelvin formulation of the second law of thermodynamics and it also sets a condition on dimensionality for certain engines to work.

Kai: But they also found something interesting about what happens in that limit regarding the uniqueness of states.

Mira: They concluded that in the thermodynamic limit, the uniqueness of those ground states breaks down, and instead, there is a multiplicity of passive states emerging.

Lev: It means if you're working on error correction or state preparation for large systems, you can't just expect one single passive state to be the only one available.

Kai: It suggests that in the large N limit, an N-copy compound system typically tends to become active instead of passive.

Mira: That’s a complex picture: the theory gives us a general tool, but it tells us that when you scale up, the simple expectation of extractable energy going to zero is true.

Lev: And the caveat is that for finite systems, we still have to deal with those non-passive states before we can rely on this asymptotic behavior.

Istituto Nanoscienze-CNR · NEST Scuola Normale Superiore

cond-mat.stat-mech, physics.plasm-ph

Submitted: 2026-03-30

Updated: 2026-11-17

Comments: v2 largely revised and expanded, accepted for publication in New Journal of Physics (2026)

DOI: 10.1088/1367-2630/aec31f

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 77/100

The gist: Ergotropy, or available energy, denotes the maximal amount of energy that can be extracted from a thermally isolated mechanical system by means of a cyclic time-dependent perturbation…

Key concepts

Ergotropy (E)
Ergotropy is defined as the maximal amount of energy that can be extracted from a thermally isolated mechanical system using cyclic, time-dependent driving forces. Mathematically, it represents the maximum extractable work achievable under these specific conditions.
Passive State ($ ho^ullet$)
A passive state is a mathematical concept related to the density distribution of a system. The paper suggests that certain equilibrium states are asymptotically passive, meaning they behave like a passive state when the system size becomes infinitely large, implying limited extractable energy in that limit.
Ergotropic Rearrangement ($ ho^ullet$)
This concept generalizes symmetric decreasing rearrangement to define how the density distribution of a system should be rearranged to maximize ergotropy. It is defined by an integral involving the probability density $ ho(x)$ and its relation to energy levels $H_0(z).
Thermodynamic Limit Findings
When applied to ideal gases, the general expression shows that any state $ ho = f(H_0)$ becomes passive in the thermodynamic limit. Crucially, both characteristic energies tend toward a constant value $E$, meaning the extractable ergotropy vanishes as system size grows infinitely large.

Terminology

Summary

Ergotropy, or available energy, denotes the maximal amount of energy that can be extracted from a thermally isolated mechanical system by means of a cyclic time-dependent perturbation <ref:2603. p. 0

Where the paper says it

We apply it to investigate the fate of classical ergotropy in the thermodynamic limit, and find that any density of the form 𝜌 = 𝑓 (𝐻0) is asymptotically passive, where 𝐻0 is the system Hamiltonian and 𝑓 a generic function. INTRODUCTION Ergotropy, or available energy, denotes the maximal amount of energy that can be extracted from a thermally isolated mechanical system by means of a cyclic time-dependent perturbation.

28388#pg2> It is possible to provide the general expression of ergotropy that applies without those limitations <ref:2603. p. 0
Where the paper says it

available energy) of a classical system is known for the case when the system phase space density is continuous and with no plateaus. Here we provide the general expression of ergotropy that applies without those limitations.

28388#pg2> This leads to the notion of “ergotropic rearangement” which generalises that of “symmetric decreasing rearrangement” <ref:2603. p. 0
Where the paper says it

It easily follows upon casting the ergotropy problem as a function rearrangement problem. This leads to the notion of “ergotropic rearangement” which generalises that of “symmetric decreasing rearrangement” (an advanced topic of measure theory).

28388#pg2> Any density of the formrho = f (H0) is asymptotically passive, where H0 is the system Hamiltonian and f a generic function <ref:2603. p. 0
Where the paper says it

This leads to the notion of “ergotropic rearangement” which generalises that of “symmetric decreasing rearrangement” (an advanced topic of measure theory). We apply it to investigate the fate of classical ergotropy in the thermodynamic limit, and find that any density of the form 𝜌 = 𝑓 (𝐻0) is asymptotically passive, where 𝐻0 is the system Hamiltonian and 𝑓 a generic function.

28388#pg2>.

Definition and Mathematical Framework

Ergotropy E is defined as the maximal amount of energy that can be extracted from a thermally isolated mechanical system by means of a cyclic time-dependent driving V(z, t) <ref:2603.28388#pg2>. In formal terms, E [rho] = max ∫ dzH0 (z) [rho(z) − ρ(phi−1 (z))] <ref:2603. p. 0

Where the paper says it

Here z = (q, p) is a point in the system phase space. In formal terms: E [𝜌] = max 𝜑 ∫ 𝑑z𝐻0 (z) [𝜌(z) − 𝜌(𝜑 −1 (z)] (1) where the maximisation is over all possible Hamiltonian flows 𝜑 : R 2𝑠 → R 2𝑠 (𝑠 denotes the number of degrees of freedom).

28388#pg2>. The passive companion of rho [36, 37] is denoted as ρ˘ <ref:2603. p. 0
Where the paper says it

We shall use the symbol 𝜌˘ for the state of minimal energy expectation among all possible evolved states 𝜌(𝜑 −1 (z)). 𝜌˘ is the passive companion of 𝜌 [36, 37].

28388#pg2>. The derivation of Ref. p. 0
Where the paper says it

[34] provides an explicit formula for E valid in the case where 𝜌 is continuous and does not present flat plateaus, which we review below. The derivation of Ref.

p. 0
Where the paper says it

[34] provides an explicit formula for E valid in the case where 𝜌 is continuous and does not present flat plateaus, which we review below. The derivation of Ref.

[34] is based on Gardner’s principles which state that i) the passive state ρ˘(z) is of the form g(H0 (z)) with g a non-increasing function, ii) the measure of the set whererho > r has to be the same as the measure of the set whererho > r ˘, for all r > 0 <ref:2603. p. 0
Where the paper says it

[34] is based on Gardner’s principles [25], according to which i) the passive state 𝜌˘(z) is of the form 𝑔(𝐻0 (z)) with 𝑔 a non-increasing function, ii) the measure of the set 𝜌 > 𝑟 has to be the same as the measure of the set 𝜌 > 𝑟 ˘ , for all 𝑟 > 0. Introducing the function Σ(𝑟) = ∫ 𝑑z𝜃 [𝜌(z) − 𝑟] (2) where 𝜃 is Heaviside step function, the Gardner principles can be succintly expressed as: Σ(𝑟) = ∫ 𝑑z𝜃 [𝑔(𝐻0 (z)) − 𝑟] , (3) arXiv:2603.

28388#pg2>.

Ergotropic Rearrangement Concept

The problem of classical ergotropy is mathematically analogous to the problem of “symmetric decreasing rearrangement” discussed, e.g., in the text of Lieb and Moss [35] <ref:2603. p. 0

Where the paper says it

The problem of classical ergotropy, is, mathematically speaking, analogous to the problem of “symmetric decreasing rearrangement”, discussed, e. , in the text of Lieb and Moss [35].

28388#pg2>. The concept of “ergotropic rearrangement" leads to an explicit expression of classical ergotropy, see Eq. (15) that generalises the previous Eq. p. 0
Where the paper says it

This in turn leads to an explicit expression of classical ergotropy, see Eq. (15) that generalises the previous Eq.

(7), without restrictions on its applicability <ref:2603. p. 0
Where the paper says it

(15) that generalises the previous Eq. (7), without restrictions on its applicability.

28388#pg2>. This generalisation is achieved by defining the ergotropic rearrangement of a probability densityrho as ρ˘(z) = ∫ ∞ 0 drchi˘ x:rho(x)>r (z) <ref:2603. p. 1
Where the paper says it

Let us then quickly review the theory of “symmetric decreasing rearrangement”, and then formulate its generalisation, leading to the notion of “ergotropic rearrangement”. One begins by defining the symmetric rearrangement 𝐴 ∗ of a set 𝐴: 𝐴 ∗ = {z : |z| ≤ 𝑅(𝐴)}, (8) where 𝑅(𝐴) is the radius of the Ball that has same Lebesgue measure 𝜇(𝐴) of the set 𝐴.

28388#pg2>.

Thermodynamic Limit Findings

The application of the general expression to an ideal gas evenly distributed on an energy shell leads to the conclusion that any phase space density of the formrho = f (H0) tends to become passive as the system size increases <ref:2603.28388#pg2>. Most remarkably, in the thermodynamic limit γ → ∞ both E¯i and E¯f tend to E, thus the ergotropy vanishes in said limit <ref:2603. p. 3

Where the paper says it

Substituting in Eq. p. 3

Where the paper says it

(24) we have Ω −1 0 (𝜇(𝑆)) = [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] 1/𝛾 𝐸 ′ (31) where we used the symbol 𝐸 ′ to represents the energy of the hypersurface of constant 𝐻0 that encloses the volume 𝜇(𝑆). Substituting in Eq.

p. 3
Where the paper says it

(24) we have Ω −1 0 (𝜇(𝑆)) = [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] 1/𝛾 𝐸 ′ (31) where we used the symbol 𝐸 ′ to represents the energy of the hypersurface of constant 𝐻0 that encloses the volume 𝜇(𝑆). Substituting in Eq.

p. 3
Where the paper says it

(24) we have Ω −1 0 (𝜇(𝑆)) = [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] 1/𝛾 𝐸 ′ (31) where we used the symbol 𝐸 ′ to represents the energy of the hypersurface of constant 𝐻0 that encloses the volume 𝜇(𝑆). Substituting in Eq.

p. 3
Where the paper says it

(24) we have Ω −1 0 (𝜇(𝑆)) = [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] 1/𝛾 𝐸 ′ (31) where we used the symbol 𝐸 ′ to represents the energy of the hypersurface of constant 𝐻0 that encloses the volume 𝜇(𝑆). Substituting in Eq.

p. 3
Where the paper says it

(24) we have Ω −1 0 (𝜇(𝑆)) = [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] 1/𝛾 𝐸 ′ (31) where we used the symbol 𝐸 ′ to represents the energy of the hypersurface of constant 𝐻0 that encloses the volume 𝜇(𝑆). Substituting in Eq.

p. 3
Where the paper says it

(24) we have Ω −1 0 (𝜇(𝑆)) = [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] 1/𝛾 𝐸 ′ (31) where we used the symbol 𝐸 ′ to represents the energy of the hypersurface of constant 𝐻0 that encloses the volume 𝜇(𝑆). Substituting in Eq.

p. 3
Where the paper says it

(24) we have Ω −1 0 (𝜇(𝑆)) = [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] 1/𝛾 𝐸 ′ (31) where we used the symbol 𝐸 ′ to represents the energy of the hypersurface of constant 𝐻0 that encloses the volume 𝜇(𝑆). Substituting in Eq.

p. 3
Where the paper says it

(24) we have Ω −1 0 (𝜇(𝑆)) = [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] 1/𝛾 𝐸 ′ (31) where we used the symbol 𝐸 ′ to represents the energy of the hypersurface of constant 𝐻0 that encloses the volume 𝜇(𝑆). Substituting in Eq.

p. 3
Where the paper says it

(24) we have Ω −1 0 (𝜇(𝑆)) = [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] 1/𝛾 𝐸 ′ (31) where we used the symbol 𝐸 ′ to represents the energy of the hypersurface of constant 𝐻0 that encloses the volume 𝜇(𝑆). Substituting in Eq.

p. 3
Where the paper says it

(24) we have Ω −1 0 (𝜇(𝑆)) = [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] 1/𝛾 𝐸 ′ (31) where we used the symbol 𝐸 ′ to represents the energy of the hypersurface of constant 𝐻0 that encloses the volume 𝜇(𝑆). Substituting in Eq.

p. 3
Where the paper says it

(24) we have Ω −1 0 (𝜇(𝑆)) = [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] 1/𝛾 𝐸 ′ (31) where we used the symbol 𝐸 ′ to represents the energy of the hypersurface of constant 𝐻0 that encloses the volume 𝜇(𝑆). Substituting in Eq.

p. 3
Where the paper says it

(24) we have Ω −1 0 (𝜇(𝑆)) = [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] 1/𝛾 𝐸 ′ (31) where we used the symbol 𝐸 ′ to represents the energy of the hypersurface of constant 𝐻0 that encloses the volume 𝜇(𝑆). Substituting in Eq.

p. 3
Where the paper says it

(24) we have Ω −1 0 (𝜇(𝑆)) = [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] 1/𝛾 𝐸 ′ (31) where we used the symbol 𝐸 ′ to represents the energy of the hypersurface of constant 𝐻0 that encloses the volume 𝜇(𝑆). Substituting in Eq.

p. 3
Where the paper says it

(24) we have Ω −1 0 (𝜇(𝑆)) = [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] 1/𝛾 𝐸 ′ (31) where we used the symbol 𝐸 ′ to represents the energy of the hypersurface of constant 𝐻0 that encloses the volume 𝜇(𝑆). Substituting in Eq.

p. 3
Where the paper says it

(24) we have Ω −1 0 (𝜇(𝑆)) = [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] 1/𝛾 𝐸 ′ (31) where we used the symbol 𝐸 ′ to represents the energy of the hypersurface of constant 𝐻0 that encloses the volume 𝜇(𝑆). Substituting in Eq.

p. 3
Where the paper says it

(24) we have Ω −1 0 (𝜇(𝑆)) = [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] 1/𝛾 𝐸 ′ (31) where we used the symbol 𝐸 ′ to represents the energy of the hypersurface of constant 𝐻0 that encloses the volume 𝜇(𝑆). Substituting in Eq.

p. 3
Where the paper says it

(24) we have Ω −1 0 (𝜇(𝑆)) = [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] 1/𝛾 𝐸 ′ (31) where we used the symbol 𝐸 ′ to represents the energy of the hypersurface of constant 𝐻0 that encloses the volume 𝜇(𝑆). Substituting in Eq.

p. 3
Where the paper says it

(24) we have Ω −1 0 (𝜇(𝑆)) = [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] 1/𝛾 𝐸 ′ (31) where we used the symbol 𝐸 ′ to represents the energy of the hypersurface of constant 𝐻0 that encloses the volume 𝜇(𝑆). Substituting in Eq.

p. 3
Where the paper says it

(24) we have Ω −1 0 (𝜇(𝑆)) = [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] 1/𝛾 𝐸 ′ (31) where we used the symbol 𝐸 ′ to represents the energy of the hypersurface of constant 𝐻0 that encloses the volume 𝜇(𝑆). Substituting in Eq.

(28) we get 𝜌˘(z) = 𝜃 [𝐸 ′ − 𝐻0 (z)] 𝐶𝛾𝐸′𝛾 (32) The system energy expectation is easily calculated as: 𝐸¯ 𝑖 = ∫ 𝑑z𝐻0 (z)𝜌(z) = 𝛾 𝛾 + 1 𝐸 𝛾+1 − (𝐸 − 𝜀) 𝛾+1 𝐸𝛾 − (𝐸 − 𝜀) 𝛾 (33) The lowest energy expectation the system can reach is 𝐸¯ 𝑓 = ∫ 𝑑z𝐻0 (z)𝜌˘(z) = 𝛾 𝛾 + 1 [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] 1/𝛾 (34) The ergotropy is their difference E = 𝐸¯ 𝑖 − 𝐸¯ 𝑓 It is interesting to note that both 𝐸¯ 𝑖 and 𝐸¯ 𝑓 (therefore also E) do not depend on the mass 𝑚, nor on the volume 𝑉 of the box, Most remarkably, in the thermodynamic limit 𝛾 → ∞ both 𝐸¯ 𝑖 and 𝐸¯ 𝑓 tend to 𝐸, thus the ergotropy vanishes in said limit.

28388#pg2>. This evidences that despite the staterho(z) is not a passive state for any finite N, the energy you can extract therefrom tends to vanish as N increases <ref:2603. p. 3
Where the paper says it

(28) we get 𝜌˘(z) = 𝜃 [𝐸 ′ − 𝐻0 (z)] 𝐶𝛾𝐸′𝛾 (32) The system energy expectation is easily calculated as: 𝐸¯ 𝑖 = ∫ 𝑑z𝐻0 (z)𝜌(z) = 𝛾 𝛾 + 1 𝐸 𝛾+1 − (𝐸 − 𝜀) 𝛾+1 𝐸𝛾 − (𝐸 − 𝜀) 𝛾 (33) The lowest energy expectation the system can reach is 𝐸¯ 𝑓 = ∫ 𝑑z𝐻0 (z)𝜌˘(z) = 𝛾 𝛾 + 1 [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] 1/𝛾 (34) The ergotropy is their difference E = 𝐸¯ 𝑖 − 𝐸¯ 𝑓 It is interesting to note that both 𝐸¯ 𝑖 and 𝐸¯ 𝑓 (therefore also E) do not depend on the mass 𝑚, nor on the volume 𝑉 of the box, Most remarkably, in the thermodynamic limit 𝛾 → ∞ both 𝐸¯ 𝑖 and 𝐸¯ 𝑓 tend to 𝐸, thus the ergotropy vanishes in said limit. This evidences that despite the state 𝜌(z) is not a passive state for any finite 𝑁, the energy you can extract therefrom tends to vanish as 𝑁 increases.

28388#pg2>.

Conclusion and Implications

We have established a connection between the physical problem of finding the available energy, a.k.a. ergotropy, of a classical system and the mathematical theory of rearrangement <ref:2603. p. 3

Where the paper says it

ergotropy, of a classical system and the mathematical theory of rearrangement. By noting that the “ergotropic rearrangement” is a generalisation of the classic concept of “symmetric decreasing rearrangement”, one easily finds an expression of the minimal energy state 𝜌˘, valid under most general assumptions, including the case when the system state 𝜌 presents jump discontinuities and flat regions, thus extending the results previously obtained in Ref.

28388#pg2>. Our main conclusion, Eq. p. 3
Where the paper says it

As an interesting application, we have used the “ergotorpic rearrangement” formula to study the fate of ergotropy in the thermodynamic limit. Our main conclusion, Eq.

(35), is that any equilibrium state of the form f (H0), is asymptotically passive <ref:2603. p. 3
Where the paper says it

Our main conclusion, Eq. (35), is that any equilibrium state of the form 𝑓 (𝐻0), is asymptotically passive.

28388#pg2>. This finding strengthens the mechanical foundations of Kelvin formulation of the second Law of Thermodynamics [41] while establishing that low dimensionality is a necessary condition for microcanonical Szilard engines to work <ref:2603. p. 3
Where the paper says it

On the contrary, you should be able to do so in low dimensional systems, as was in fact demonstrated in various works [42–45]. This finding strengthens the mechanical foundations of Kelvin formulation of the second Law of Thermodynamics [39, 41] while establishing that low dimensionality is a necessary condition for microcanonical Szilard engines to work [42–45].

28388#pg2>. Furthermore, Eq. p. 3
Where the paper says it

[41] for quantum systems. Furthermore, note that any state of the form 𝜌 = 𝑓 (𝐻0), with 𝑓 a generic function can be approximated to any wanted degree, by convex linear combinations of flat distributions defined over sufficiently thin energy shells.

(35) leads to the interesting conclusion that in the thermodynamic limit, the uniqueness of the “ground state” breaks down, while a multiplicity of passive states emerges <ref:2603. p. 3
Where the paper says it

Furthermore, Eq. (35) leads to the interesting conclusion that in the thermodynamic limit, the uniqueness of the “ground state” breaks down, while a multiplicity of passive states emerges, a topic that deserves further investigation.

28388#pg2>.

Pruning redundant reasoning tokens

The expression in Eq. (19) generalises that of Eq. p. 0

Where the paper says it

It easily follows upon casting the ergotropy problem as a function rearrangement problem. This leads to the notion of “ergotropic rearangement” which generalises that of “symmetric decreasing rearrangement” (an advanced topic of measure theory).

(6) to the case of a generic phase space densityrho <ref:2603. p. 3
Where the paper says it

By noting that the “ergotropic rearrangement” is a generalisation of the classic concept of “symmetric decreasing rearrangement”, one easily finds an expression of the minimal energy state 𝜌˘, valid under most general assumptions, including the case when the system state 𝜌 presents jump discontinuities and flat regions, thus extending the results previously obtained in Ref. In “ergotropic rearrangement” a phase space density 𝜌0 (z) is rearranged in decreasing manner relative to the increase of the Hamilton function 𝐻0 (z).

28388#pg2>. The expression for the minimal energy can be conveniently written as E˘ = ∫ dzH0 (z)rho˘(z) = ∫ dzH0 (z) ∫ ∞ 0 drθ [Σ(r) − Ω0 (H0 (z))] <ref:2603. p. 2
Where the paper says it

, the quadratic Hamiltonian 𝐻0 (z) = |z| 2 , 𝜌˘ coincides with 𝜌 ∗ , so we see that ergotropic rearrangement is a generalisation of symmetric decreasing rearrangement. (19), the minimal energy can be conveniently written as: 𝐸˘ = ∫ 𝑑z𝐻0 (z)𝜌˘(z) = ∫ 𝑑z𝐻0 (z) ∫ ∞ 0 𝑑𝑟𝜃 [Σ(𝑟) − Ω0 (𝐻0 (z))] = ∫ ∞ 0 𝑑Φ 𝐸0 (Φ) ∫ ∞ 0 𝑑𝑟 𝜃 [Σ(𝑟) − Φ] = ∫ ∞ 0 𝑑𝑟 ∫ Σ(𝑟 ) 0 𝑑Φ 𝐸0 (Φ) where we made the usual changes of variables z → 𝑒 = 𝐻0 (z), 𝑒 → Φ = Ω0 (𝑒) [34, 39] , with 𝑑z = Ω′ 0 (𝑒)𝑑𝑒 = 𝑑Φ, and introduced the function 𝐸0 𝐸0 Ω −1 0 (21) such that 𝐸0 (Φ) denotes the energy of the hypersurface of 𝐻0 that encloses the volume Φ [34, 39].

28388#pg2>. The system energy expectation is easily calculated as E¯i = ∫ dzH0 (z)rho(z) = γ/γ + 1 E/γ+1 − (E − ε)/γ+1 E/γ − (E − ε)/γ <ref:2603. p. 3
Where the paper says it

Substituting in Eq. (28) we get 𝜌˘(z) = 𝜃 [𝐸 ′ − 𝐻0 (z)] 𝐶𝛾𝐸′𝛾 (32) The system energy expectation is easily calculated as: 𝐸¯ 𝑖 = ∫ 𝑑z𝐻0 (z)𝜌(z) = 𝛾 𝛾 + 1 𝐸 𝛾+1 − (𝐸 − 𝜀) 𝛾+1 𝐸𝛾 − (𝐸 − 𝜀) 𝛾 (33) The lowest energy expectation the system can reach is 𝐸¯ 𝑓 = ∫ 𝑑z𝐻0 (z)𝜌˘(z) = 𝛾 𝛾 + 1 [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] 1/𝛾 (34) The ergotropy is their difference E = 𝐸¯ 𝑖 − 𝐸¯ 𝑓 It is interesting to note that both 𝐸¯ 𝑖 and 𝐸¯ 𝑓 (therefore also E) do not depend on the mass 𝑚, nor on the volume 𝑉 of the box, Most remarkably, in the thermodynamic limit 𝛾 → ∞ both 𝐸¯ 𝑖 and 𝐸¯ 𝑓 tend to 𝐸, thus the ergotropy vanishes in said limit.

28388#pg2>. The lowest energy expectation the system can reach is E¯f = ∫ dzH0 (z)rho˘(z) = γ/γ + 1 [E/γ − (E − ε)/γ] 1/γ <ref:2603. p. 3
Where the paper says it

Substituting in Eq. (28) we get 𝜌˘(z) = 𝜃 [𝐸 ′ − 𝐻0 (z)] 𝐶𝛾𝐸′𝛾 (32) The system energy expectation is easily calculated as: 𝐸¯ 𝑖 = ∫ 𝑑z𝐻0 (z)𝜌(z) = 𝛾 𝛾 + 1 𝐸 𝛾+1 − (𝐸 − 𝜀) 𝛾+1 𝐸𝛾 − (𝐸 − 𝜀) 𝛾 (33) The lowest energy expectation the system can reach is 𝐸¯ 𝑓 = ∫ 𝑑z𝐻0 (z)𝜌˘(z) = 𝛾 𝛾 + 1 [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] 1/𝛾 (34) The ergotropy is their difference E = 𝐸¯ 𝑖 − 𝐸¯ 𝑓 It is interesting to note that both 𝐸¯ 𝑖 and 𝐸¯ 𝑓 (therefore also E) do not depend on the mass 𝑚, nor on the volume 𝑉 of the box, Most remarkably, in the thermodynamic limit 𝛾 → ∞ both 𝐸¯ 𝑖 and 𝐸¯ 𝑓 tend to 𝐸, thus the ergotropy vanishes in said limit.

28388#pg2>.

How the training signal is built

The explicit expression of ergotropy, see Eq. (7) below, which, however only applies to phase-space densitiesrho that are continuous and without flat plateaus <ref:2603. p. 0

Where the paper says it

That was possible thanks to writing an explicit expression of ergotropy, see Eq. (7) below, which, however only applies to phase-space densities 𝜌 that are continuous and without flat plateaus.

28388#pg2>. The derivation of Ref. [34] proceeds from Eq. p. 1
Where the paper says it

The derivation of Ref. [34] proceeds from Eq.

(5), by further assuming Σ is invertible <ref:2603. p. 1
Where the paper says it

[34] proceeds from Eq. (5), by further assuming Σ is invertible.

28388#pg2>. Under the above assumption, rewriting Eq. p. 1
Where the paper says it

This will be illustrated below, in Sec. Under the above assumption, rewriting Eq.

(4) as Σ = Ω0 ◦ g−1, and applying Σ−1 to the left and g to the right, one obtains g = Σ−1 ◦ Ω0 <ref:2603. p. 1
Where the paper says it

Under the above assumption, rewriting Eq. (4) as Σ = Ω0 ◦ 𝑔 −1 , and applying Σ −1 to the left and 𝑔 to the right, one obtains 𝑔 = Σ−1 ◦ Ω0.

28388#pg2>.

Where the gains come from

The paper demonstrates that any state of the formrho = f (H0), with f a generic function can be approximated to any wanted degree, by convex linear combinations of flat distributions defined over sufficiently thin energy shells <ref:2603.28388#pg2>. Any time dependent perturbation maps (in the thermodynamic limit), each shell to a set of larger or same energy expectation, so that the full distribution gets mapped to a new distribution with higher or same energy <ref:2603. p. 3

Where the paper says it

Furthermore, note that any state of the form 𝜌 = 𝑓 (𝐻0), with 𝑓 a generic function can be approximated to any wanted degree, by convex linear combinations of flat distributions defined over sufficiently thin energy shells. Any time dependent perturbation maps (in the thermodynamic limit), each shell to a set of larger or same energy expectation, so that the full distribution gets mapped to a new distribution with higher or same energy.

28388#pg2>. This follows that lim N→∞ E [f (H0)] = 0 <ref:2603. p. 3
Where the paper says it

Any time dependent perturbation maps (in the thermodynamic limit), each shell to a set of larger or same energy expectation, so that the full distribution gets mapped to a new distribution with higher or same energy. It follows that lim 𝑁→∞ E [ 𝑓 (𝐻0)] = 0 (35) namely, in the thermodynamic limit, any stationary state of the form 𝑓 (𝐻0) is asymptotically passive.

28388#pg2>. The compound state of the N-copy systemrhotot(z) = ρ(z1)rho(z2). p. 4
Where the paper says it

5 because, while the compound state of the 𝑁-copy system 𝜌tot(z) = 𝜌(z1)𝜌(z2) . 𝜌(z𝑁 ) is a stationary state relative to the total Hamiltonian 𝐻tot(z) = 𝐻(z1) + 𝐻(z2) + · · · + 𝐻(z𝑁 ) it is generally not of the form 𝑓 (𝐻tot(z)).

.. ρ(zN) is a stationary state relative to the total Hamiltonian Htot(z) = H(z1) + H(z2) + · · · + H(zN) it is generally not of the form f (Htot(z)) <ref:2603. p. 4
Where the paper says it

5 because, while the compound state of the 𝑁-copy system 𝜌tot(z) = 𝜌(z1)𝜌(z2) . 𝜌(z𝑁 ) is a stationary state relative to the total Hamiltonian 𝐻tot(z) = 𝐻(z1) + 𝐻(z2) + · · · + 𝐻(z𝑁 ) it is generally not of the form 𝑓 (𝐻tot(z)).

28388#pg2>. The paper concludes that in the large N limit, an N-copy compound system typically tends to become active <ref:2603. p. 3
Where the paper says it

This is known for both classical [36] and quantum systems [5]. Thus, in the large 𝑁 limit, an 𝑁-copy compound system typically tends to become active.

28388#pg2>.

The explicit expression of μ(S) is Ω0 (E) − Ω0 (E − ε) = Cγ [E/γ − (E - ε)/γ] <ref:2603. p. 3

Where the paper says it

4 Plugging it in Eq. (19) and recalling that by definition Ω0 ≥ 0 we swiftly obtain 𝜌˘(z) = 𝜃 [𝜇(𝑆) − Ω0 (𝐻0 (z)] 𝜇(𝑆) (28) Using that Ω−1 0 is a strictly increasing function, we obatin 𝜌˘(z) = 𝜃 [Ω−1 0 (𝜇(𝑆)) − 𝐻0 (z)] 𝜇(𝑆) (29) The explicit expression of 𝜇(𝑆) is 𝜇(𝑆) = Ω0 (𝐸) − Ω0 (𝐸 − 𝜀) = 𝐶𝛾 [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] (30) Using Eq.

p. 3
Where the paper says it

4 Plugging it in Eq. (19) and recalling that by definition Ω0 ≥ 0 we swiftly obtain 𝜌˘(z) = 𝜃 [𝜇(𝑆) − Ω0 (𝐻0 (z)] 𝜇(𝑆) (28) Using that Ω−1 0 is a strictly increasing function, we obatin 𝜌˘(z) = 𝜃 [Ω−1 0 (𝜇(𝑆)) − 𝐻0 (z)] 𝜇(𝑆) (29) The explicit expression of 𝜇(𝑆) is 𝜇(𝑆) = Ω0 (𝐸) − Ω0 (𝐸 − 𝜀) = 𝐶𝛾 [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] (30) Using Eq.

p. 3
Where the paper says it

4 Plugging it in Eq. (19) and recalling that by definition Ω0 ≥ 0 we swiftly obtain 𝜌˘(z) = 𝜃 [𝜇(𝑆) − Ω0 (𝐻0 (z)] 𝜇(𝑆) (28) Using that Ω−1 0 is a strictly increasing function, we obatin 𝜌˘(z) = 𝜃 [Ω−1 0 (𝜇(𝑆)) − 𝐻0 (z)] 𝜇(𝑆) (29) The explicit expression of 𝜇(𝑆) is 𝜇(𝑆) = Ω0 (𝐸) − Ω0 (𝐸 − 𝜀) = 𝐶𝛾 [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] (30) Using Eq.

p. 3
Where the paper says it

4 Plugging it in Eq. (19) and recalling that by definition Ω0 ≥ 0 we swiftly obtain 𝜌˘(z) = 𝜃 [𝜇(𝑆) − Ω0 (𝐻0 (z)] 𝜇(𝑆) (28) Using that Ω−1 0 is a strictly increasing function, we obatin 𝜌˘(z) = 𝜃 [Ω−1 0 (𝜇(𝑆)) − 𝐻0 (z)] 𝜇(𝑆) (29) The explicit expression of 𝜇(𝑆) is 𝜇(𝑆) = Ω0 (𝐸) − Ω0 (𝐸 − 𝜀) = 𝐶𝛾 [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] (30) Using Eq.

p. 3
Where the paper says it

4 Plugging it in Eq. (19) and recalling that by definition Ω0 ≥ 0 we swiftly obtain 𝜌˘(z) = 𝜃 [𝜇(𝑆) − Ω0 (𝐻0 (z)] 𝜇(𝑆) (28) Using that Ω−1 0 is a strictly increasing function, we obatin 𝜌˘(z) = 𝜃 [Ω−1 0 (𝜇(𝑆)) − 𝐻0 (z)] 𝜇(𝑆) (29) The explicit expression of 𝜇(𝑆) is 𝜇(𝑆) = Ω0 (𝐸) − Ω0 (𝐸 − 𝜀) = 𝐶𝛾 [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] (30) Using Eq.

p. 3
Where the paper says it

4 Plugging it in Eq. (19) and recalling that by definition Ω0 ≥ 0 we swiftly obtain 𝜌˘(z) = 𝜃 [𝜇(𝑆) − Ω0 (𝐻0 (z)] 𝜇(𝑆) (28) Using that Ω−1 0 is a strictly increasing function, we obatin 𝜌˘(z) = 𝜃 [Ω−1 0 (𝜇(𝑆)) − 𝐻0 (z)] 𝜇(𝑆) (29) The explicit expression of 𝜇(𝑆) is 𝜇(𝑆) = Ω0 (𝐸) − Ω0 (𝐸 − 𝜀) = 𝐶𝛾 [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] (30) Using Eq.

p. 3
Where the paper says it

4 Plugging it in Eq. (19) and recalling that by definition Ω0 ≥ 0 we swiftly obtain 𝜌˘(z) = 𝜃 [𝜇(𝑆) − Ω0 (𝐻0 (z)] 𝜇(𝑆) (28) Using that Ω−1 0 is a strictly increasing function, we obatin 𝜌˘(z) = 𝜃 [Ω−1 0 (𝜇(𝑆)) − 𝐻0 (z)] 𝜇(𝑆) (29) The explicit expression of 𝜇(𝑆) is 𝜇(𝑆) = Ω0 (𝐸) − Ω0 (𝐸 − 𝜀) = 𝐶𝛾 [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] (30) Using Eq.

p. 3
Where the paper says it

4 Plugging it in Eq. (19) and recalling that by definition Ω0 ≥ 0 we swiftly obtain 𝜌˘(z) = 𝜃 [𝜇(𝑆) − Ω0 (𝐻0 (z)] 𝜇(𝑆) (28) Using that Ω−1 0 is a strictly increasing function, we obatin 𝜌˘(z) = 𝜃 [Ω−1 0 (𝜇(𝑆)) − 𝐻0 (z)] 𝜇(𝑆) (29) The explicit expression of 𝜇(𝑆) is 𝜇(𝑆) = Ω0 (𝐸) − Ω0 (𝐸 − 𝜀) = 𝐶𝛾 [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] (30) Using Eq.

p. 3
Where the paper says it

4 Plugging it in Eq. (19) and recalling that by definition Ω0 ≥ 0 we swiftly obtain 𝜌˘(z) = 𝜃 [𝜇(𝑆) − Ω0 (𝐻0 (z)] 𝜇(𝑆) (28) Using that Ω−1 0 is a strictly increasing function, we obatin 𝜌˘(z) = 𝜃 [Ω−1 0 (𝜇(𝑆)) − 𝐻0 (z)] 𝜇(𝑆) (29) The explicit expression of 𝜇(𝑆) is 𝜇(𝑆) = Ω0 (𝐸) − Ω0 (𝐸 − 𝜀) = 𝐶𝛾 [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] (30) Using Eq.

p. 3
Where the paper says it

4 Plugging it in Eq. (19) and recalling that by definition Ω0 ≥ 0 we swiftly obtain 𝜌˘(z) = 𝜃 [𝜇(𝑆) − Ω0 (𝐻0 (z)] 𝜇(𝑆) (28) Using that Ω−1 0 is a strictly increasing function, we obatin 𝜌˘(z) = 𝜃 [Ω−1 0 (𝜇(𝑆)) − 𝐻0 (z)] 𝜇(𝑆) (29) The explicit expression of 𝜇(𝑆) is 𝜇(𝑆) = Ω0 (𝐸) − Ω0 (𝐸 − 𝜀) = 𝐶𝛾 [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] (30) Using Eq.

p. 3
Where the paper says it

4 Plugging it in Eq. (19) and recalling that by definition Ω0 ≥ 0 we swiftly obtain 𝜌˘(z) = 𝜃 [𝜇(𝑆) − Ω0 (𝐻0 (z)] 𝜇(𝑆) (28) Using that Ω−1 0 is a strictly increasing function, we obatin 𝜌˘(z) = 𝜃 [Ω−1 0 (𝜇(𝑆)) − 𝐻0 (z)] 𝜇(𝑆) (29) The explicit expression of 𝜇(𝑆) is 𝜇(𝑆) = Ω0 (𝐸) − Ω0 (𝐸 − 𝜀) = 𝐶𝛾 [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] (30) Using Eq.

p. 3
Where the paper says it

4 Plugging it in Eq. (19) and recalling that by definition Ω0 ≥ 0 we swiftly obtain 𝜌˘(z) = 𝜃 [𝜇(𝑆) − Ω0 (𝐻0 (z)] 𝜇(𝑆) (28) Using that Ω−1 0 is a strictly increasing function, we obatin 𝜌˘(z) = 𝜃 [Ω−1 0 (𝜇(𝑆)) − 𝐻0 (z)] 𝜇(𝑆) (29) The explicit expression of 𝜇(𝑆) is 𝜇(𝑆) = Ω0 (𝐸) − Ω0 (𝐸 − 𝜀) = 𝐶𝛾 [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] (30) Using Eq.

p. 3
Where the paper says it

4 Plugging it in Eq. (19) and recalling that by definition Ω0 ≥ 0 we swiftly obtain 𝜌˘(z) = 𝜃 [𝜇(𝑆) − Ω0 (𝐻0 (z)] 𝜇(𝑆) (28) Using that Ω−1 0 is a strictly increasing function, we obatin 𝜌˘(z) = 𝜃 [Ω−1 0 (𝜇(𝑆)) − 𝐻0 (z)] 𝜇(𝑆) (29) The explicit expression of 𝜇(𝑆) is 𝜇(𝑆) = Ω0 (𝐸) − Ω0 (𝐸 − 𝜀) = 𝐶𝛾 [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] (30) Using Eq.

p. 3
Where the paper says it

4 Plugging it in Eq. (19) and recalling that by definition Ω0 ≥ 0 we swiftly obtain 𝜌˘(z) = 𝜃 [𝜇(𝑆) − Ω0 (𝐻0 (z)] 𝜇(𝑆) (28) Using that Ω−1 0 is a strictly increasing function, we obatin 𝜌˘(z) = 𝜃 [Ω−1 0 (𝜇(𝑆)) − 𝐻0 (z)] 𝜇(𝑆) (29) The explicit expression of 𝜇(𝑆) is 𝜇(𝑆) = Ω0 (𝐸) − Ω0 (𝐸 − 𝜀) = 𝐶𝛾 [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] (30) Using Eq.

p. 3
Where the paper says it

4 Plugging it in Eq. (19) and recalling that by definition Ω0 ≥ 0 we swiftly obtain 𝜌˘(z) = 𝜃 [𝜇(𝑆) − Ω0 (𝐻0 (z)] 𝜇(𝑆) (28) Using that Ω−1 0 is a strictly increasing function, we obatin 𝜌˘(z) = 𝜃 [Ω−1 0 (𝜇(𝑆)) − 𝐻0 (z)] 𝜇(𝑆) (29) The explicit expression of 𝜇(𝑆) is 𝜇(𝑆) = Ω0 (𝐸) − Ω0 (𝐸 − 𝜀) = 𝐶𝛾 [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] (30) Using Eq.

p. 3
Where the paper says it

4 Plugging it in Eq. (19) and recalling that by definition Ω0 ≥ 0 we swiftly obtain 𝜌˘(z) = 𝜃 [𝜇(𝑆) − Ω0 (𝐻0 (z)] 𝜇(𝑆) (28) Using that Ω−1 0 is a strictly increasing function, we obatin 𝜌˘(z) = 𝜃 [Ω−1 0 (𝜇(𝑆)) − 𝐻0 (z)] 𝜇(𝑆) (29) The explicit expression of 𝜇(𝑆) is 𝜇(𝑆) = Ω0 (𝐸) − Ω0 (𝐸 − 𝜀) = 𝐶𝛾 [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] (30) Using Eq.

p. 3
Where the paper says it

4 Plugging it in Eq. (19) and recalling that by definition Ω0 ≥ 0 we swiftly obtain 𝜌˘(z) = 𝜃 [𝜇(𝑆) − Ω0 (𝐻0 (z)] 𝜇(𝑆) (28) Using that Ω−1 0 is a strictly increasing function, we obatin 𝜌˘(z) = 𝜃 [Ω−1 0 (𝜇(𝑆)) − 𝐻0 (z)] 𝜇(𝑆) (29) The explicit expression of 𝜇(𝑆) is 𝜇(𝑆) = Ω0 (𝐸) − Ω0 (𝐸 − 𝜀) = 𝐶𝛾 [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] (30) Using Eq.

p. 3
Where the paper says it

4 Plugging it in Eq. (19) and recalling that by definition Ω0 ≥ 0 we swiftly obtain 𝜌˘(z) = 𝜃 [𝜇(𝑆) − Ω0 (𝐻0 (z)] 𝜇(𝑆) (28) Using that Ω−1 0 is a strictly increasing function, we obatin 𝜌˘(z) = 𝜃 [Ω−1 0 (𝜇(𝑆)) − 𝐻0 (z)] 𝜇(𝑆) (29) The explicit expression of 𝜇(𝑆) is 𝜇(𝑆) = Ω0 (𝐸) − Ω0 (𝐸 − 𝜀) = 𝐶𝛾 [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] (30) Using Eq.

p. 3
Where the paper says it

4 Plugging it in Eq. (19) and recalling that by definition Ω0 ≥ 0 we swiftly obtain 𝜌˘(z) = 𝜃 [𝜇(𝑆) − Ω0 (𝐻0 (z)] 𝜇(𝑆) (28) Using that Ω−1 0 is a strictly increasing function, we obatin 𝜌˘(z) = 𝜃 [Ω−1 0 (𝜇(𝑆)) − 𝐻0 (z)] 𝜇(𝑆) (29) The explicit expression of 𝜇(𝑆) is 𝜇(𝑆) = Ω0 (𝐸) − Ω0 (𝐸 − 𝜀) = 𝐶𝛾 [𝐸 𝛾 − (𝐸 − 𝜀) 𝛾 ] (30) Using Eq.

28388#pg2>. Using Eq. (24) we have Ω−1 0(μ(S)) = [E/γ − (E - ε)/γ] 1/γ E' <ref:2603.28388#pg2>. Substituting in Eq. (28) we get ρ˘(z) = θ [E' − H0 (z)] Cγ E' / γ <ref:2603.28388#pg2>. The explicit expression of μ(S) is Ω0 (E) − Ω0 (E − ε) = Cγ [E/γ − (E - ε)/γ] <ref:2603.28388#pg2>. Using Eq. (24) we have Ω−1 0(μ(S)) = [E/γ − (E - ε)/γ] 1/γ E' <ref:2603.28388#pg2>. Substituting in Eq. (28) we get ρ˘(z) = θ [E' − H0 (z)] Cγ E' / γ <ref:2603.28388#pg2>. The explicit expression of μ(S) is Ω0 (E) − Ω0 (E − ε) = Cγ [E/γ − (E - ε)/γ] <ref:2603.28388#pg2>. Using Eq. (24) we have Ω−1 0(μ(S)) = [E/γ − (E - ε)/γ] 1/γ E' <ref:2603.28388#pg2>. Substituting in Eq. (28) we get ρ˘(z) = θ [E' − H0 (z)] Cγ E' / γ <ref:2603.28388#pg2>. The explicit expression of μ(S) is Ω0 (E) − Ω0 (E − ε) = Cγ [E/γ − (E - ε)/γ] <ref:2603.28388#pg2>. Using Eq. (24) we have Ω−1 0(μ(S)) = [E/γ − (E - ε)/γ] 1/γ E' <ref:2603.28388#pg2>. Substituting in Eq. (28) we get ρ˘(z) = θ [E' − H0 (z)] Cγ E' / γ <ref:2603.28388#pg2>. The explicit expression of μ(S) is Ω0 (E) − Ω0 (E − ε) = Cγ [E/γ − (E - ε)/γ] <ref:2603.28388#pg2>. Using Eq. (24) we have Ω−1 0(μ(S)) = [E/γ − (E - ε)/γ] 1/γ E' <ref:2603.28388#pg2>. Substituting in Eq. (28) we get ρ˘(z) = θ [E' − H0 (z)] Cγ E' / γ <ref:2603.28388#pg2>. The explicit expression of μ(S) is Ω0 (E) − Ω0 (E − ε) = Cγ [E/γ − (E - ε)/γ] <ref:2603.28388#pg2>. Using Eq. (24) we have Ω−1 0(μ(S)) = [E/γ − (E - ε)/γ] 1/γ E' <ref:2603.28388#pg2>. Substituting in Eq. (28) we get ρ˘(z) = θ [E' − H0 (z)] Cγ E' / γ <ref:2603.28388#pg2>. The explicit expression of μ(S) is Ω0 (E) − Ω0 (E − ε) = Cγ [E/γ − (E - ε)/γ] <ref:2603.28388#pg2>. Using Eq. (24) we have Ω−1 0(μ(S)) = [E/γ − (E - ε)/γ] 1/γ E' <ref:2603.28388#pg2>. Substituting in Eq. (28) we get ρ˘(z) = θ [E' − H0 (z)] Cγ E' / γ <ref:2603.28388#pg2>. The explicit expression of μ(S) is Ω0 (E) − Ω0 (E − ε) = Cγ [E/γ − (E - ε)/γ] <ref:2603.28388#pg2>. Using Eq. (24) we have Ω−1 0(μ(S)) = [E/γ − (E - ε)/γ] 1/γ E' <ref:2603.28388#pg2>. Substituting in Eq. (28) we get ρ˘(z) = θ [E' − H0 (z)] Cγ E' / γ <ref:2603.28388#pg2>. The explicit expression of μ(S) is Ω0 (E) − Ω0 (E − ε) = Cγ [E/γ − (E - ε)/γ] <ref:2603.28388#pg2>. Using Eq. (24) we have Ω−1 0(μ(S)) = [E/γ − (E - ε)/γ] 1/γ E' <ref:2603.28388#pg2>. Substituting in Eq. (28) we get ρ˘(z) = θ [E' − H0 (z)] Cγ E' / γ <ref:2603.28388#pg2>. The explicit expression of μ(S) is Ω0 (E) − Ω0 (E − ε) = Cγ [E/γ − (E - ε)/γ] <ref:2603.28388#pg2>. Using Eq. (24) we have Ω−1 0(μ(S)) = [E/γ − (E - ε)/γ] 1/γ E' <ref:2603.28388#pg2>. Substituting in Eq. (28) we get ρ˘(z) = θ [E' − H0 (z)] Cγ E' / γ <ref:2603.28388#pg2>. The explicit expression of μ(S) is Ω0 (E) − Ω0 (E − ε) = Cγ [E/γ − (E - ε)/γ] <ref:2603.28388#pg2>. Using Eq. (24) we have Ω−1 0(μ(S)) = [E/γ − (E - ε)/γ] 1/γ E' <ref:2603.28388#pg2>. Substituting in Eq. (28) we get ρ˘(z) = θ [E' − H0 (z)] Cγ E' / γ <ref:2603.28388#pg2>. The explicit expression of μ(S) is Ω0 (E) − Ω0 (E − ε) = Cγ [E/γ − (E - ε)/γ] <ref:2603.28388#pg2>. Using Eq. (24) we have Ω−1 0(μ(S)) = [E/γ − (E - ε)/γ] 1/γ E' <ref:2603.28388#pg2>. Substituting in Eq. (28) we get ρ˘(z) = θ [E' − H0 (z)] Cγ E' / γ <ref:2603.28388#pg2>. The explicit expression of μ(S) is Ω0 (E) − Ω0 (E − ε) = Cγ [E/γ − (E - ε)/γ] <ref:2603.28388#pg2>. Using Eq. (24) we have Ω−1 0(μ(S)) = [E/γ − (E - ε)/γ] 1/γ E' <ref:2603.28388#pg2>. Substituting in Eq. (28) we get ρ˘(z) = θ [E' − H0 (z)] Cγ E' / γ <ref:2603.28388#pg2>. The explicit expression of μ(S) is Ω0 (E) − Ω0 (E − ε) = Cγ [E/γ − (E - ε)/γ] <ref:2603.28388#pg2>. Using Eq. (24) we have Ω−1 0(μ(S)) = [E/γ − (E - ε)/γ] 1/γ E' <ref:2603.28388#pg2>. Substituting in Eq. (28) we get ρ˘(z) = θ [E' − H0 (z)] Cγ E' / γ <ref:2603.28388#pg2>. The explicit expression of μ(S) is Ω0 (E) − Ω0 (E − ε) = Cγ [E/γ − (E - ε)/γ] <ref:2603.28388#pg2>. Using Eq. (24) we have Ω−1 0(μ(S)) = [E/γ − (E - ε)/γ] 1/γ E' <ref:2603.28388#pg2>. Substituting in Eq. (28) we get ρ˘(z) = θ [E' − H0 (z)] Cγ E' / γ <ref:2603.28388#pg2>. The explicit expression of μ(S) is Ω0 (E) − Ω0 (E − ε) = Cγ [E/γ − (E - ε)/γ] <ref:2603.28388#pg2>. Using Eq. (24) we have Ω−1 0(μ(S)) = [E/γ − (E - ε)/γ] 1/γ E' <ref:2603.28388#pg2>. Substituting in Eq. (28) we get ρ˘(z) = θ [E' − H0 (z)] Cγ E' / γ <ref:2603.28388#pg2>. The explicit expression of μ(S) is Ω0 (E) − Ω0 (E − ε) = Cγ [E/γ − (E - ε)/γ] <ref:2603.28388#pg2>. Using Eq. (24) we have Ω−1 0(μ(S)) = [E/γ − (E - ε)/γ] 1/γ E' <ref:2603.28388#pg2>. Substituting in Eq. (28) we get ρ˘(z) = θ [E' − H0 (z)] Cγ E' / γ <ref:2603.28388#pg2>. The explicit expression of μ(S) is Ω0 (E) − Ω0 (E − ε) = Cγ [E/γ − (E - ε)/γ] <ref:2603.28388#pg2>. Using Eq. (24) we have Ω−1 0(μ(S)) = [E/γ − (E - ε)/γ] 1/γ E' <ref:2603.28388#pg2>. Substituting in Eq. (29) we get ρ˘(z) = θ [Ω−1 0(μ(S)) − H0 (z)] / μ(S) <ref:2603. p. 0
Where the paper says it

[34] is based on Gardner’s principles [25], according to which i) the passive state 𝜌˘(z) is of the form 𝑔(𝐻0 (z)) with 𝑔 a non-increasing function, ii) the measure of the set 𝜌 > 𝑟 has to be the same as the measure of the set 𝜌 > 𝑟 ˘ , for all 𝑟 > 0. Introducing the function Σ(𝑟) = ∫ 𝑑z𝜃 [𝜌(z) − 𝑟] (2) where 𝜃 is Heaviside step function, the Gardner principles can be succintly expressed as: Σ(𝑟) = ∫ 𝑑z𝜃 [𝑔(𝐻0 (z)) − 𝑟] , (3) arXiv:2603.

p. 0
Where the paper says it

[34] is based on Gardner’s principles [25], according to which i) the passive state 𝜌˘(z) is of the form 𝑔(𝐻0 (z)) with 𝑔 a non-increasing function, ii) the measure of the set 𝜌 > 𝑟 has to be the same as the measure of the set 𝜌 > 𝑟 ˘ , for all 𝑟 > 0. Introducing the function Σ(𝑟) = ∫ 𝑑z𝜃 [𝜌(z) − 𝑟] (2) where 𝜃 is Heaviside step function, the Gardner principles can be succintly expressed as: Σ(𝑟) = ∫ 𝑑z𝜃 [𝑔(𝐻0 (z)) − 𝑟] , (3) arXiv:2603.

28388#pg2>. The explicit expression of μ(S) is Ω0 (E) − Ω0 (E − ε) = Cγ [E/γ − (E - ε)/γ] <ref:2603.28388#pg2>. Using Eq. (24) we have Ω−1 0(μ(S)) = [E/γ − (E - ε)/γ] 1/γ E' <ref:2603.28388#pg2>. Substituting in Eq. (29) we get ρ˘(z) = θ [Ω−1 0(μ(S)) − H0 (z)] / μ(S) <ref:2603.28388#pg2>. The explicit expression of μ(S) is Ω0 (E) − Ω0 (E − ε) = Cγ [E/γ − (E - ε)/γ] <ref:2603.28388#pg2>. Using Eq.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Improved AI systems can calculate maximal energy extraction from a system by analyzing its phase space density using ergotropic rearrangement, which generalizes symmetric decreasing rearrangement.

  2. The improved AI can determine that any phase space density of the form rho = f (H0) is asymptotically passive in the thermodynamic limit, meaning any stationary state of this form is asymptotically passive.

  3. The system can predict that for a large ergodic system where only its energy with finite accuracy is known, no energy will be extractable, which strengthens the mechanical foundations of Kelvin formulation of the second law of Thermodynamics.

  4. The AI can establish that low dimensionality is a necessary condition for microcanonical Szilard engines to work by leveraging the finding that any equilibrium state of the form f (H0) is asymptotically passive in this limit.

Abstract

We address the fundamental question of what is the maximal amount of energy (a.k.a. ergotropy) that can be extracted from a macroscopic classical system by means of a cyclical variation of an external parameter. We show that if the system is prepared in a statistical state of the form ρ= F(H), with H the Hamiltonian and F a generic limited function, then the ergotropy is vanishing in the large N limit, provided the thermodynamic entropy obeys the scaling S(E) N αs (E / N β) with α,β>0 and s some strictly increasing function with certain reasonable properties. We stress that such ubiquitous condition also ensures the equivalence of the so-called "surface" and "volume" entropies. We illustrate this by introducing the notion of ``ergotropic rearrangement'' of a phase space density which we then apply to find the exact ergotropy expression of systems with constant (positive or negative) heat capacity being evenly distributed over energy shells in phase space. We discuss cases where the above mentioned equivalence does (does not) break and the ergotropy does (does not) vanish accordingly, in the large N limit. The presented results shed new light on the statistical-mechanical foundations of the second law of thermodynamics.

Sources

Related papers