REVIEW 2 major objections 5 minor 1 cited by
Extracting and charging energy into almost unknown quantum states
T0 review · 2 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash
Pith's one-line read This paper shows that with only the mean energy of a quantum state known, the guaranteed unitary work is exactly 2(E−epsilon_mean) for antisymmetric Hamiltonians and equals the minimum ergotropy for diagonal or random-unitary protocols.
desk verdict Worth reviewing: the Emin/Amin characterization and worst-case bounds are solid, but the diagonal-state saturation proof (SM Thm. 9) has a real gap in replacing bistochastic maps by unitaries. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing object is the pair of energy-constrained minimum functionals Emin(E)=min_{ρ:Tr[Hρ]=E} E(ρ) and Amin(E)=min_{ρ:Tr[Hρ]=E} A(ρ), with E(ρ), A(ρ) the usual ergotropy and anti-ergotropy. The proof machinery is convex geometry: passive states ρ↓ and antipassive states ρ↑ form simplexes whose vertices are flat states ρ↓_k and ρ↑_k, and since ergotropy is convex and linear on those simplexes, the minimizers are convex combinations of at most two vertices, making Emin and Amin piecewise linear and computable by a linear program. Saturation is carried by the population-reversal unitary U_rev=Σ_k |ε_{d+1−k}⟩⟨ε_k| for antisymmetric spectra, and by a diagonal-optimal unitary whose coherence sensitivity is averaged away in the random-unitary setting.
What would settle it
Take a non-antisymmetric three-level Hamiltonian and an energy E>ϵmean; compute numerically max_{Λ∈U} min_{ρ∈σ_E∩Δ} ΔE(ρ,Λ) and compare it with Emin(E). If any unitary fails to reach Emin(E) on some diagonal state, or if an orthostochastic-unrealizable doubly stochastic matrix outperforms every unitary, then the diagonal-state equality fails while the antisymmetric and random-unitary results remain untouched.
Extended reading notes
Core claim
The paper's central claim is that in the near-ignorance regime where only Tr[ρH]=E is known, the guaranteed unitary work satisfies 0≤Ĕ(E)≤Emin(E) and 0≤Ā(E)≤Amin(E), with Emin and Amin the minimum ergotropy and anti-ergotropy on the energy shell. These minima are completely characterized: minimizers are passive for E≤ϵmean and antipassive for E≥ϵmean, so the problem reduces to a linear program on the simplexes of passive/antipassive states and the curves are piecewise linear. The paper then proves two saturation theorems: for antisymmetric Hamiltonians (eigenvalues paired with constant sum), Ĕ(E)=Emin(E)=2(E−ϵmean) for E≥ϵmean, with the energy-independent unitary U_rev reversing populations; and for diagonal input states or random-unitary channels, Ĕ_Δ(E)=Ĕ_RU(E)=Emin(E) for every Hamiltonian. It also derives that the maximum coherent ergotropy at fixed energy equals Emax(E)−Emin(E), quantifying the coherence contribution.
Load-bearing premise
The saturation result for diagonal input states assumes that the doubly stochastic map delivered by the minimax theorem can be replaced by a single unitary on the diagonal subspace without changing the guaranteed extraction; for dimensions above two this replacement is not automatic, because not every doubly stochastic matrix is orthostochastic.
Editorial extensions
If this is right
- For antisymmetric Hamiltonians (angular momentum, multi-qubit systems), a single fixed unitary U_rev extracts exactly 2(E−ϵmean) from every state of mean energy E>ϵmean, with no state tomography.
- For diagonal states, and for random-unitary channels applied to any state, the worst-case extractable and chargeable energies equal the minimum ergotropy and anti-ergotropy for any finite-dimensional Hamiltonian.
- The piecewise-linear curves Emin(E) and Amin(E) can be evaluated in O(d^2) operations, so the guaranteed-work bounds are easy to compute for large systems.
- The maximum coherent ergotropy at fixed energy equals Emax(E)−Emin(E), showing exactly how much extractable energy is carried by coherence rather than populations.
- No fixed unitary can guarantee more than Emin(E) in general; beating that guarantee requires antisymmetry, a diagonal input restriction, or randomization.
Reading between the lines
- Editorial extension: the proof of the diagonal-state saturation replaces the minimax-optimal doubly stochastic (population-mixing) map by a single unitary; for d≥3 this replacement is nontrivial because not every doubly stochastic matrix is orthostochastic, so a gap could remain for coherent states even if the diagonal and random-unitary claims stand.
- A testable consequence of the paper's lower bound is that as E approaches ϵmax the coherence penalty on the diagonal-optimal unitary vanishes at least as fast as the square root of the Gibbs-state entropy, so high-energy extraction should be nearly optimal on all states.
- The same convex-geometry argument should transfer to other partial information, such as fixed energy variance or a fixed expectation value of another observable, giving worst-case work bounds for those energy shells.
- For quantum battery protocols, the practical prescription is to use U_rev when the spectrum is symmetric and otherwise measure only mean energy and apply the diagonal-optimal unitary, accepting a bounded coherence penalty; randomizing over the unitary's free phases removes the penalty.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper studies guaranteed energy extraction and charging from a finite-dimensional quantum system when only the mean energy E of the initial state is known. It introduces the maximum worst-case extractable and injectable energies E-tilde(E) and A-tilde(E), proves the upper bounds E-tilde(E) <= Emin(E) and A-tilde(E) <= Amin(E), and characterizes Emin and Amin as the minima of ergotropy and anti-ergotropy over the energy shell sigma_E. The characterization is reduced to a linear program over passive and antipassive states, yielding a piecewise-linear, efficiently computable curve. The paper then gives saturation results: exact saturation for antisymmetric Hamiltonians via the population-reversal unitary, saturation for diagonal input states, and saturation via random unitary channels, followed by a detailed qutrit comparison of the different protocols.
Significance. If the results hold, the paper provides a complete and efficiently computable characterization of worst-case energy manipulation under minimal information, with parameter-free bounds and explicit saturating protocols. The strengths are the convexity/linear-programming proof of Emin and Amin, the state-independent reversal unitary for antisymmetric Hamiltonians, the clean minimax argument for random unitaries, and the explicit qutrit construction showing how random phases cancel coherent contributions. The paper is self-contained and does not rely on fitted parameters or external numerics. The main weakness is a proof gap in the diagonal-state saturation claim, which is one of the three advertised saturation scenarios.
major comments (2)
- [SM §IV B, Theorem 9, Eqs. (S130)–(S132)] The proof of the diagonal-state equality E-tilde_Delta(E) = Emin(E) is incomplete. The proof correctly shows that every unitary channel induces a bistochastic map on diagonal states, giving the inclusion U ⊆ B in Eq. (S130). It then applies the minimax theorem over the convex set B and concludes by asserting that "the sets U and B behave in the same way on Δ(H)", replacing the maximum over B by a maximum over U. The converse inclusion is not proved, and for d ≥ 3 not every doubly stochastic matrix is orthostochastic, i.e., realizable as P_ij = |<epsilon_i|U|epsilon_j>|^2 for a single unitary U. The minimax step produces a saddle point in the convex polytope B, and no argument shows that this saddle point can be chosen unitary-realizable while preserving the value. Since Theorem 4's diagonal-state saturation is a central claim, this gap is load-bearing. The authors should either prove that the optimal bistochastic map can be chosen orthostochastic (or unitary-realizable), provide a direct max-min argument over U, or restrict the statement of the theorem.
- [SM §IV B, Theorem 9 versus Theorem 10] The random-unitary saturation result in Theorem 10 does not suffer from the same gap, because the set RU is convex and the minimax theorem applies directly. However, the main-text Theorem 4 states E-tilde_Delta(E) = E-tilde_RU(E) = Emin(E), and the abstract claims that access to decoherence guarantees saturation for all Hamiltonians. If the diagonal-state half of Theorem 4 is not established, the decoherence-based saturation claim in the abstract and conclusions should be downgraded or explicitly qualified, even though the random-unitary and antisymmetric saturation results would remain intact.
minor comments (5)
- [Main text after Eq. (4)] The phrase "positive-semidefinite quantities" should be "nonnegative quantities", since the objects are real-valued functions, not matrices.
- [SM §III, proof of Lemma 7] In the anti-ergotropy part of the proof, the displayed expression after Eq. (S103) writes "epsilon_max - E - E_min^{(H)}(E)" but the quantity should be A_min^{(H)}(E); as printed, the formula is inconsistent with the statement.
- [References] References [28] and [31] are the same Francica et al. paper (Phys. Rev. Lett. 125, 180603 (2020)) and should be consolidated.
- [Figure 2 caption] The caption lists "(E(ρ↑_k), E(ρ↑_k))" for the orange points; the two entries are presumably the mean energy and the ergotropy of the vertex state, and the notation should distinguish them, as done in the SM with (E_k, E_k).
- [Throughout] There are minor typographical errors, including "quantites" in the abstract and "V asco" in the author block, which should be corrected.
Circularity Check
No significant circularity: the central derivation is definition-driven; the diagonal-state saturation proof has an unproven unitary-replacement step, which is a correctness gap rather than a circular reduction.
full rationale
The paper is analytically self-contained. The central quantities are defined as max-min expressions, and the upper bound in Eq. (5) is the elementary max-min <= min-max inequality. The characterization of Emin(E) and Amin(E) is derived from the standard closed-form formulas for ergotropy and anti-ergotropy, convexity, and linear programming, not assumed as an input. Theorem 3 is proven by an explicit computation with U_rev for antisymmetric Hamiltonians. The random-unitary part of Theorem 4 follows from the Sion minimax theorem because the set of random unitary channels and the energy shell are compact and convex and Delta E is bilinear. There are no fitted parameters renamed as predictions, no empirical pattern merely renamed, and no load-bearing self-citation: the one self-cited formula used, E(rho)=E-lambda_down·epsilon_up, is the standard definitional ergotropy identity also present in the general ergotropy literature. The genuine concern is an omitted proof, not circularity: in SM Theorem 9 (Eq. S132), after applying the minimax theorem over the convex set B of bistochastic maps, the text states that 'the sets U and B behave in the same way on Delta E(H)' and replaces max over B by max over U without proof. For d>=3 not every doubly stochastic matrix is orthostochastic, so this replacement is not automatic. However, this does not make the claimed equality equivalent to its own input: the equality E_delta(E)=Emin(E) is not already contained in the definition of Emin, and the antisymmetric saturation and random-unitary saturation results do not rely on this step. Thus the circularity score is low; the flagged issue is a proof gap to be weighed as a correctness risk, not as circularity.
Assumptions & free parameters
assumptions (4)
- standard math Sion minimax theorem and von Neumann minimax theorem allow exchanging min and max for linear/convex-concave functions on compact convex sets.
- standard math Birkhoff-von Neumann decomposition: every doubly stochastic matrix is a convex combination of permutations.
- domain assumption The set of states with fixed mean energy is convex and compact; ergotropy and anti-ergotropy are convex functionals.
- domain assumption The Hamiltonian is finite-dimensional, known, with ordered spectrum (degeneracies resolved by an arbitrary basis choice); the mean energy is known exactly.
Cite this review
Pith. "Pith review of Extracting and charging energy into almost unknown quantum states." pith.science (2026). https://pith.science/paper/5PS6FW4W
@misc{pith2026250908899,
author = {Pith},
title = {Pith review of: Extracting and charging energy into almost unknown quantum states},
year = {2026},
howpublished = {\url{https://pith.science/paper/5PS6FW4W}},
note = {Machine review of arXiv:2509.08899}
}
read the original abstract
In this work, we investigate the amount of energy that can be extracted or charged through unitary operations when only minimal information about the state is known. Assuming knowledge of only the mean energy of the state, we start by developing optimal upper bounds for the work that can be unitarily extracted or charged in this scenario. In deriving these upper bounds, we provide a complete characterization of the minimum ergotropy and anti-ergotropy for density matrices with fixed average energy, showing that the minimum states are always passive or antipassive and the problem of finding them can be mapped to a simple linear programming algorithm. Furthermore, we show that these lower bounds directly translate into upper bounds for the energy-constrained coherent ergotropy and anti-ergotropy of a state. We continue by illustrating scenarios in which these bounds can be saturated: a simple unitary protocol is shown to saturate the bounds for relevant classes of Hamiltonians, while having access to decoherence or randomness as resources, the saturation is guaranteed for all Hamiltonians. Finally, by taking a qutrit as an example, we show and compare the performances of the various protocols identified.
Figures
Forward citations
Cited by 1 Pith paper
-
Dual-use quantum hardware for quantum resource generation and energy storage
Quantum resource generation and quantum battery charging can be achieved with the same protocols, enabling integrated dual-use hardware on superconducting circuits.
Reference graph
Works this paper leans on
-
[1]
= ϵ ( 1 + δ 3 ) =ϵmean, E1 := E ( ˆH)(ρ↑
-
[2]
We have E ( ˆH) c,max(E) = max ρ∈σ( ˆH) E [ E ( ˆH)(ρ) − E ( ˆH)(∆(ρ)) ] , (S97) ≤ max ρ∈σ( ˆH) E E ( ˆH)(ρ) − min ρ∈σ( ˆH) E E ( ˆH)(∆(ρ)), (S98) =E −ϵmin − E ( ˆH) min (E), (S99) where in the last step we exploited the fact that the state that minimizes the ergotropy is already diagonal, hence the minimum of the dephased state is equal to the minimum of...
-
[3]
= 0, (S171) E2 :=E(ρ↑
-
[4]
= ϵ (3 2 + δ 2 ) = 3 2ϵmean, E2 := E ( ˆH)(ρ↑
-
[5]
= ϵ, (S172) E3 :=E(ρ↑
-
[6]
= 2ϵ, E3 := E ( ˆH)(ρ↑
-
[7]
= 2ϵ. (S173) We already know that the energy-constrained minimum ergotropy in the area [ϵ1,ϵ mean] is zero due to the existence of passive states. We want to find its value in the area [ϵmean,ϵ 3]. There are two possible scenarios
-
[8]
The energy-constrained minimum ergotropy is given by the states (1 −p)ρ↑ 1 +pρ↑ 3 (p ∈ [0, 1]) in the whole area [ϵmean,ϵ 3]
Show all 14 references
-
[9]
Also, if δ = 0 we return to the anti-symmetric scenario in which both cases return the same results
The energy-constrained minimum ergotropy is given by the states (1 −p)ρ↑ 1 +pρ↑ 2 in the area [ϵmean, 3 2ϵmean] and by the states (1 −p)ρ↑ 2 +pρ↑ 3 in the area [ 3 2ϵmean,ϵ 3]. Also, if δ = 0 we return to the anti-symmetric scenario in which both cases return the same results....
-
[10]
We now know the correct scenario based on the probability of the state
≤ E ( ˆH)(ρ↑ 2), (S174) ⇔ ¯p · 2ϵ ≤ϵ, (S175) ⇔ ¯p ≤ 1 2, (S176) where we have exploited the linearity of the ergotropy with respect to anti-passive states. We now know the correct scenario based on the probability of the state. Our aim, though, is to find the energy-constrained...
-
[11]
+ ¯pE(ρ↑ 3), (S179) = (1 − ¯p)ϵmean + ¯p · 2ϵ, (S180) =ϵmean + ¯p(2ϵ −ϵmean), (S181) S24 FIG. S5. energy-constrained minimum ergotropy with respect to the mean energy E ( ˆH) min (E) (red line) for the qutrit when the intermediate energy level ϵ2 = ϵ(1 + δ) is closer to the gr...
-
[12]
Hence, we can use these states to bound the possible structure of ˆU (+,E) max,∆
It follows that such unitary will extract exactly the ergotropy from the states ρ↑ min(p). Hence, we can use these states to bound the possible structure of ˆU (+,E) max,∆. The most general unitary that extracts the ergotropy from the stateρ↑ min(p) is an arbitrary unitary ˆU2...
-
[14]
Let us compute those populations. ˜λ1 = ⟨ϵ1| [λ1 |ψ⟩ ⟨ψ| +λ2 |ψ⊥⟩ ⟨ψ⊥| +λ3 |ϵ3⟩ ⟨ϵ3|] |ϵ1⟩ , (S280) =λ1| ⟨ϵ1|ψ⟩ |2 +λ2| ⟨ϵ1|ψ⊥⟩ |2, (S281) =λ1(1 −q) +λ2q, (S282) ˜λ2 = ⟨ϵ2| [λ1 |ψ⟩ ⟨ψ| +λ2 |ψ⊥⟩ ⟨ψ⊥| +λ3 |ϵ3⟩ ⟨ϵ3|] |ϵ2⟩ , (S283) =λ1| ⟨ϵ2|ψ⟩ |2 +λ2| ⟨ϵ2|ψ⊥⟩ |2, (S284) =λ1q +λ2(1...
1976 arXiv
-
[23]
We want to compute these populations and then optimise the free parameters of ˆU23 so that the energy we extract is at least the energy-constrained minimum ergotropy for all diagonal states. We have ˜ρ∆ = ˆU23ρ∆ ˆU † 23, (S231) =λ1 |ϵ1⟩ ⟨ϵ1| +λ2 |ψ⟩ ⟨ψ| +λ3 |ψ⊥⟩ ⟨ψ⊥| , (S232) ...
Reviewed August 15, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.