Pith. sign in

REVIEW 2 major objections 4 minor 14 references

Entropy from scattering in weakly interacting systems

T0 review · 2 major / 4 minor · reviewed 2026-08-15 · deepseek-v4-flash

Pith's one-line read The paper proves that, under weak unitary scattering, the von Neumann entropy of a subsystem is non-decreasing at order $\lambda^2\ln(1/\lambda^2)$ whenever the initial reduced density matrix of that subsystem has a nonempty kernel and…

desk verdict A genuine extension of the scattering area law to correlated separable states, but the proof overclaims: (3) does not imply the diagonal form (4), so the explicit second-order formula needs correction or restriction. read the letter →

arxiv 2506.19127 v3 pith:NX7NK7PW submitted 2025-06-23 quant-ph cond-mat.otherhep-thmath-phmath.MP

classification quant-phcond-mat.otherhep-thmath-phmath.MP
keywords subsystemvonNeumannentropyperturbativescatteringincreasecriteriakernelofreduceddensitymatrixseparablecorrelatedstatesarealawforentanglementopticaltheoremsecondorderperturbationtheory
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper asks whether a small unitary scattering event can ever lower the von Neumann entropy of one subsystem. It proves that for initial states whose A-reduced density matrix has a zero-eigenvalue sector and which are diagonal in a common product basis, the leading change in the A entropy is of order $\lambda^2 \ln(1/\lambda^2)$ and is nonnegative; when the T-matrix couples the zero and nonzero sectors and acts nontrivially on the other subsystem, the entropy strictly increases. The result carries the pure-state area law over to a broader class of classically preparable correlated states, and it also identifies full-rank initial states for which the entropy can decrease.

What carries the argument

The engine is second-order perturbation theory for the eigenvalues of the reduced density matrix, organized around the zero-eigenvalue sector, the kernel of $\rho^A_{\rm in}$. Because zero eigenvalues cannot shift at linear order, their leading shifts are second order, carry a $\lambda^2 \ln(1/\lambda^2)$ factor from the entropy function, and can be written as a sum of absolute squares. The sum-of-squares form is what makes the sign definite, and the cancellation of the interference term encodes unitarity.

What would settle it

Take $\rho_{\rm in} = p_1 |1\rangle\langle 1|\otimes|0\rangle\langle0| + p_2 |2\rangle\langle2|\otimes|+\rangle\langle+|$ with $|+\rangle=(|0\rangle+|1\rangle)/\sqrt2$ and a $\hat T^{(1)}$ connecting $|2\rangle$ to a kernel state of $A$ while acting nontrivially on $B$; compute the order $\lambda^2\ln(1/\lambda^2)$ shift of the $A$ entropy directly from the eigenvalue shifts. A negative value would disprove sufficiency of condition (3); a nonnegative value would show the sign result survives only via the positivity argument rather than the diagonal-basis formula.

Watch

Extended reading notes

Core claim

The central claim is that, at second order in perturbation theory, the change in subsystem A entropy is $\delta S = (\lambda^2 \ln(1/\lambda^2)) \times [\text{nonnegative sum of squares}] + O(\lambda^2)$. The nonnegative sum comes from the leading shifts of the zero eigenvalues of $\rho^A_{\rm in}$, which cannot change at linear order and must be nonnegative because they are eigenvalues of a density matrix. If those shifts are nonzero, they dominate the $\lambda^2$ terms and fix the sign. Strict growth requires $\hat T^{(1)}$ to connect kernel and non-kernel states and to act nontrivially on the B subsystem; if it acts only on A, the apparent transition probability is exactly cancelled by the overcounting term, preserving unitarity.

Load-bearing premise

The proof assumes that condition (3), commuting with the eigenspace projectors of the A-reduced state, is equivalent to being diagonal in a common product basis, which is the step that licenses the explicit formula; the equivalence fails for some states satisfying (3).

Editorial extensions

If this is right

  • For any initial state in the class with a kernel, the leading entropy change is nonnegative regardless of the details of the interaction, so no sign flip of $\hat T^{(1)}$ can lower the entropy at that order.
  • The pure-state area law $\delta S_A = \lambda^2 \ln(1/\lambda^2) \sigma$ is recovered as a special case, now embedded in a larger class of correlated separable initial states.
  • If the kernel and non-kernel sectors are connected by $\hat T^{(1)}$ acting nontrivially on B, the entropy increase is strictly positive at order $\lambda^2 \ln(1/\lambda^2)$, giving a concrete rate for entropy production in weak scattering.
  • For full-rank initial reduced states the $\lambda^2 \ln(1/\lambda^2)$ term is absent, and explicit product-state examples show the order-$\lambda^2$ entropy shift can have either sign; the paper conjectures this indefiniteness is generic.
  • The class of guaranteed-increase states is a commuting subclass of separable states, so classical correlations suffice; entanglement is not required for the monotonicity.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The paper's sign argument for zero eigenvalues relies mainly on positivity of density matrices, so the monotonicity may survive for any initial state whose A-reduction has a kernel, even when the diagonal-product-basis condition is not satisfied; checking this would extend the theorem without the disputed equivalence.
  • The commuting condition makes the initial state a classical mixture in a product basis, so the result can be read as a statement about a quantum map acting on a classical ensemble; a natural next step is to ask which unital or bistochastic maps preserve monotonicity when the kernel is nonempty.
  • In scattering applications, the result suggests a measurable bound: for incoming beams whose single-particle reduced state has at least one forbidden channel, the entropy production rate is nonnegative at leading order, which could be tested in low-energy scattering or in trapped-ion unitary gates.
  • The full-rank examples show that 'entropy increase on average' fails when all A states are populated; identifying the minimal correlation structure needed for guaranteed increase is a direct open problem.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

2 major / 4 minor

Summary. The paper studies the perturbative evolution of the von Neumann entropy of subsystem A of a bipartite system under a unitary S-matrix close to the identity. It claims that if the initial reduced density matrix of A has a nonempty kernel and the initial composite state satisfies the commutation condition (3), then the change in entropy is nonnegative at order lambda^2 ln(1/lambda^2); with the additional condition that T^(1) couples kernel and non-kernel states and acts nontrivially on B, the change is strictly positive. The paper derives an explicit second-order formula generalizing the pure-state scattering area law, and it gives examples showing that when rho_A^in has full rank the entropy can either increase or decrease depending on the fine details of the state and the T-matrix.

Significance. If correct, the result extends the pure-state scattering area law to a class of correlated separable initial states and gives a simple sufficient condition for non-decrease of subsystem entropy under weak unitary evolution. The derivation is self-contained: it uses unitarity, the optical theorem, and standard perturbation theory, with no fitted parameters; the T-matrix is treated as arbitrary, so the criteria are genuine structural restrictions on the initial state. The final formulas are explicit and the worked examples are useful. The main caveat is that the proof as written covers only a diagonal-product subclass of the stated class of states, so the central claim is not yet fully established for all states admitted by the theorem as stated.

major comments (2)
  1. [Around Eqs. (3), (4), and (19)] The assertion that 'a close examination reveals that (3), (4), and (19) are all equivalent' is false. For example, take rho_in = p1|1><1| x |0><0| + p2|2><2| x |+><+| with |+> = (|0>+|1>)/sqrt(2) and add a kernel state |3> with zero weight. Then condition (3) holds, but the B-block operators |0><0| and |+><+| do not share an eigenbasis, so no product basis exists in which rho_in has the diagonal form (4); if p1 = p2, then (19) also fails because [|1><2| x 1, rho_in] = |1><2| x (|+><+| - |0><0|). Since the second-order calculation explicitly begins with 'We shall now assume that the initial state obeys (3) and is of the form (4)', formulas (20) and (21) are derived only for the diagonal subclass. The theorem as stated in the Introduction and Conclusion, however, claims the result for all states satisfying (3). The gap appears repairable by a direct positivity argument for the kernel block of rho_A^out, but as it stands the proof does not cover the stated domain.
  2. [Around Eq. (19) in 'THE LEADING ORDER ~ lambda'] The necessity argument for (19) is also too strong. Under (3), one has rho_in = sum_m |m><m| x sigma_m. The diagonal first-order matrix elements M^(1)_mm vanish identically, and for a degenerate eigenspace the trace of the restricted first-order matrix over that eigenspace vanishes; hence the linear entropy shift in (17) is zero even when (19) is violated, for example when the sigma_m differ within a degenerate eigenspace. Thus (19) is sufficient but not necessary for the linear term to vanish, and the text's claim that the linear contribution vanishes only under (19) is not correct.
minor comments (4)
  1. [Eqs. (20) and (21)] The symbol delta S is used for delta S_A in these equations, whereas the rest of the paper uses delta S_A; the notation should be made consistent.
  2. [Sentence following Eq. (18)] There is a typo: 'at lease some' should read 'at least some'.
  3. [Eq. (13) and the degenerate case] The text should clarify that for degenerate eigenvalues the second-order term in (13) is the part of the perturbation not already diagonalized within the degenerate subspace; the subsequent trace calculation in (25) relies on this restricted-matrix convention.
  4. [Eq. (21) and the sentence after it] The phrase 'if the right-hand-side of (21) is not zero, it must be positive' is mathematically correct, but calling the right-hand side 'nonnegative' in the preceding sentence would be more precise.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the entropy-increase derivation is self-contained perturbation theory; the (3)-(4) equivalence issue is a proof-scope gap, not a circular step.

full rationale

No circularity found. The central calculation starts from unitarity, S = 1 + iT, and the optical theorem, equation (8), and expands rho_out to order lambda^2 in equation (10) without assuming any entropy result. The eigenvalue-shift matrix M in equations (13)-(14) is standard second-order perturbation theory. The constraints that force the linear entropy shift to vanish, equations (16) and (19), are derived from positivity of density matrices and from the ability to flip the sign of T^(1), not from the desired conclusion. The second-order coefficient is then written in equation (21) as a weighted sum of manifestly non-negative absolute squares, so the positivity of delta S at order lambda^2 ln(1/lambda^2) follows by algebra. The pure-state area law of references [1-7], including the co-authored reference [2], is used only as background and is re-derived as the special case in equation (23), so no fitted parameter or borrowed prediction is load-bearing. The citation to the author's thesis [12] is a provenance note, not an argument. Thus the derivation chain is self-contained. Separate correctness note, not circularity: the paper asserts that condition (3) is equivalent to the diagonal-product form (4), but this equivalence is false; a state diagonal in the A basis whose B-blocks do not share eigenbases satisfies (3) but not (4). The proof of equations (20)-(21) as printed assumes the form (4), so the stated theorem's domain is broader than the proven domain. This is a proof gap and a mathematical overstatement, but it does not make the derivation circular.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

The paper introduces no new entities or fitted parameters. Its central claim rests on standard perturbation theory, unitarity, and an auxiliary assumption that conditions (3) and (4) are equivalent, which is not generally true.

assumptions (5)
  • domain assumption The S matrix is unitary and satisfies the optical theorem (8).
    Used throughout to expand ρout to order λ² (Eq. 10).
  • domain assumption The T matrix admits an asymptotic expansion in the small parameter λ with T^(1) Hermitian.
    Eq. (9) and the optical theorem at order λ²; assumed valid for weak scattering.
  • domain assumption For any first-order operator T^(1), the opposite sign -T^(1) can be realized by a consistent unitary evolution.
    Used in the sign-flip argument to rule out linear-order entropy changes; requires the set of allowed interactions to be sign-symmetric.
  • ad hoc to paper The initial state satisfying (3) can be written in the diagonal product form (4).
    The paper claims equivalence of (3) and (4) but does not prove it; it is false in general. This assumption underlies the explicit formula (20) and the sum-of-squares positivity (21).
  • standard math Standard second-order perturbation theory for eigenvalues of a Hermitian matrix applies to the reduced density matrix.
    Equation (13) uses non-degenerate and degenerate perturbation theory for eigenvalue shifts.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Entropy from scattering in weakly interacting systems." pith.science (2026). https://pith.science/paper/NX7NK7PW

@misc{pith2026250619127,
  author       = {Pith},
  title        = {Pith review of: Entropy from scattering in weakly interacting systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NX7NK7PW}},
  note         = {Machine review of arXiv:2506.19127}
}
read the original abstract

Perturbation theory is used to investigate the evolution of the von Neumann entropy of a subsystem of a bipartite quantum system under the action of a unitary matrix, in the limit where that matrix is close to the unit matrix. The physical context for such process would be scattering with weak short-ranged interactions where the unitary matrix is the S matrix. We find surprisingly simple criteria for the initial state and the S matrix that guarantee that the subsystem entropy increases. The class of initial states that meet these criteria are more correlated than simple product states of the subsystems. They form a subclass of the set of all separable states, and they can therefore be assembled by classical processes alone.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

14 extracted references · 3 canonical work pages

  1. [9]

    Elastic cross section is entangle- ment entropy,

    I. Low and Z. Yin, “Elastic cross section is entangle- ment entropy,” Phys. Rev. D111(2025) no.6, 065027 [arXiv:2410.22414 [hep-th]]

  2. [1]

    Variation of Entangle- ment Entropy in Scattering Process,

    S. Seki, I.Y. Park, S.-J. Sin, “Variation of Entangle- ment Entropy in Scattering Process,” Phys.Lett.B 743 (2015) 147-153, arXiv:1412.7894 [hep-th]

  3. [2]

    For any eigenvector|m⟩ofˆρAin, |m⟩⟨m|⊗1 B,ˆρin = 0;and (3)

  4. [3]

    Thetheleading termintheperturbativeexpansion of the ˆTmatrix, which we shall denote asλ ˆT (1), allows for transitions between the non-kernel and the kernel states ofˆρAin and it acts nontrivially on the B subsystem. When the first two of the three criteria listed above are both satisfied, the change in entropy is at least of orderλ 2 ln 1 λ2 asλ→0and the...

  5. [4]

    Scattering and momen- tum space entanglement,

    G. Grignani, G. W. Semenoff, “Scattering and momen- tum space entanglement,” Phys. Lett. B772, 699-702 (2017) [arXiv:1612.08858 [hep-th]]

  6. [5]

    Relativistic effect of entanglement in fermion-fermion scattering,

    J. Fan and X. Li, “Relativistic effect of entanglement in fermion-fermion scattering,” Phys. Rev. D97, no.1, 016011 (2018) [arXiv:1712.06237 [hep-th]]

  7. [6]

    Information loss in effective field the- ory: entanglement and thermal entropies,

    D. Boyanovsky, “Information loss in effective field the- ory: entanglement and thermal entropies,” Phys. Rev. D97, no.6, 065008 (2018) [arXiv:1801.06840 [hep-th]]

  8. [7]

    Evaluation of Entanglement Entropy in High Energy Elastic Scattering,

    R. Peschanski and S. Seki, “Evaluation of Entanglement Entropy in High Energy Elastic Scattering,” Phys. Rev. D100, no.7, 076012 (2019) [arXiv:1906.09696 [hep-th]]

Show all 14 references
  1. [8]

    An Area Law for Entanglement Entropy in Particle Scattering,

    I. Low and Z. Yin, “An Area Law for Entanglement Entropy in Particle Scattering,” [arXiv:2405.08056 [hep- th]]

  2. [10]

    The second laws of quantum thermodynam- ics

    For some recent discussions of a possible second law of entanglement, see F. Brandao, M. Horodecki, N. Ng, S. Wehner, “The second laws of quantum thermodynam- ics”, PNAS112(11) 3275-3279 (2015); S. Minagawa, M. Mohammady, K. Sakai, K. Kato, F. Buscemi, “Uni- versal validity of...

  3. [11]

    Entropy growth in perturbative scattering,

    C. Cheung, T. He and A. Sivaramakrishnan, “Entropy growth in perturbative scattering,” Phys. Rev. D108, no.4, 045013 (2023), [arXiv:2304.13052 [hep-th]]

  4. [12]

    Entanglement detection

    For a discussion and definitions, see O. Guhne, G. Toth, “Entanglement detection”, Physics Reports474, 1-75 (2009), [arXiv:0811.2803 [quant-ph]]

  5. [13]

    Momentum-space entanglement and renormal- ization in quantum field theory,

    V.Balasubramanian, M.B.McDermott, M.VanRaams- donk, “Momentum-space entanglement and renormal- ization in quantum field theory,” Phys.Rev.D 86(2012) 045014, arXiv:1108.3568 [hep-th]

  6. [14]

    Entropy Growth in Quantum Mechan- ics

    D. MacIntyre, “Entropy Growth in Quantum Mechan- ics”, B.Sc. Thesis, University of British Columbia, 2024

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.