arxiv: 2605.03192 · v1 · submitted 2026-05-04 · ❄️ cond-mat.stat-mech

Recognition: unknown

Universal criticality of entropy production in chemical reaction networks

Keiji Saito, Kyota Tamano

Pith reviewed 2026-05-08 17:06 UTC · model grok-4.3

classification ❄️ cond-mat.stat-mech

keywords entropy productionchemical reaction networksnonequilibrium criticalitybifurcationsstochastic thermodynamicsscaling exponentsfluctuation-response relation

0 comments

The pith

Entropy production fluctuations in chemical reaction networks show universal scaling at bifurcations, diverging more readily than responses.

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

The paper examines entropy production near nonequilibrium critical points in well-mixed reversible chemical reaction networks. These points arise when the deterministic mass-action equations undergo local bifurcations. Linear noise approximations together with normal-form reductions yield specific scaling exponents for both the fluctuations and the average responses of entropy production at pitchfork, transcritical, saddle-node, and Hopf bifurcations. A separate bound derived from trajectory-space information theory establishes the general relation that the fluctuation scaling exponent is at least twice the response exponent. This relation implies that any divergence in the response forces a stronger divergence in fluctuations, positioning the latter as the more sensitive signature of criticality.

Core claim

In the macroscopic-first limit, entropy production acquires generic critical exponents at the four standard local bifurcations of mass-action dynamics. The exponents for fluctuations and responses are obtained from linear-noise formulas, center-manifold reductions, and Floquet theory. An additional trajectory-space Cramér-Rao bound supplies the universal inequality between the two exponents, showing that divergent responses require divergent fluctuations but not conversely.

What carries the argument

Linear-noise approximation combined with center-manifold normal forms and Floquet theory to extract scaling exponents, plus a trajectory-space bound that enforces the relation between fluctuation and response exponents.

If this is right

At each of the four bifurcations the entropy-production variance and response follow definite power laws whose exponents depend only on the bifurcation type.
Whenever the average entropy-production response diverges, its fluctuations must diverge at least quadratically faster.
Entropy-production fluctuations therefore remain finite only if the response also remains finite.
The classification applies to any well-mixed reversible network whose deterministic limit exhibits one of the standard local bifurcations.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same scaling relation may hold in other stochastic nonequilibrium systems whose dynamics admit a similar separation into deterministic flow and linear fluctuations.
Monitoring entropy-production fluctuations could reveal impending transitions in biological reaction networks before any divergence appears in mean observables.
The trajectory-space bound itself might be tested independently in simple driven systems without reference to chemical networks.

Load-bearing premise

Critical transitions occur as local bifurcations of the deterministic mass-action equations, and the linear-noise plus center-manifold approximations remain valid arbitrarily close to those points.

What would settle it

In an experimental chemical reaction network known to undergo one of the listed bifurcations, measure the variance and mean of entropy production while tuning a control parameter through the critical value and test whether the observed power-law exponents satisfy the predicted inequality.

Figures

Figures reproduced from arXiv: 2605.03192 by Keiji Saito, Kyota Tamano.

**Figure 1.** Figure 1: FIG. 1. (a): Parametric response view at source ↗

**Figure 2.** Figure 2: FIG. 2. Variance of the entropy production in the model view at source ↗

read the original abstract

Stochastic thermodynamics gives universal relations for microscopic entropy production, yet its critical behavior at macroscopic nonequilibrium transitions remains unclassified. We study well-mixed reversible chemical reaction networks in the macroscopic-first limit, where transitions arise as local bifurcations of mass-action dynamics. Using linear-noise formulas, center-manifold normal forms, and Floquet theory, we obtain generic exponents for entropy-production fluctuations and responses at pitchfork, transcritical, saddle-node, and Hopf bifurcations. Beyond this low-order classification, a trajectory-space Cram\'{e}r-Rao type bound yields the universal scaling inequality $\alpha - 2\beta \geq 0$. Hence divergent responses require divergent fluctuations, but not conversely, making entropy-production fluctuations a sharper probe of nonequilibrium criticality.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper extracts generic scaling exponents for entropy production at four bifurcations plus a Cramér-Rao inequality, but the linear-noise closure looks shaky exactly where fluctuations are supposed to diverge.

read the letter

The main takeaway is that Tamano and Saito give explicit scaling exponents for entropy-production fluctuations and responses near pitchfork, transcritical, saddle-node, and Hopf points in mass-action chemical networks, together with the inequality α - 2β ≥ 0 from a trajectory-space Cramér-Rao bound. The inequality is the cleaner part: it says divergent responses require divergent fluctuations but not the reverse, so fluctuations could be a sharper diagnostic than mean responses. They obtain the exponents by feeding linear-noise formulas, center-manifold normal forms, and Floquet theory into the entropy-production expression. That systematic mapping is new in this context and could organize observables in nonequilibrium chemical systems. The work is straightforward to follow if you already know the standard tools from stochastic thermodynamics and nonlinear dynamics. The soft spot is whether those tools remain valid at the critical points themselves. Linear noise assumes small Gaussian fluctuations around a stable deterministic fixed point or limit cycle, yet at a bifurcation the restoring force on the slow mode vanishes and the variance diverges. The abstract states the approximations still yield the exponents, but without error estimates or higher-order corrections shown, it is unclear whether the claimed scalings survive or whether the closure breaks down. The stress-test note flags exactly this tension, and it is worth checking in the full derivations. This paper is for readers already working on stochastic thermodynamics of reaction networks who want a low-order classification of critical behavior. It is not a broad review but a focused calculation that could be cited for the inequality or the exponent table if the approximations check out. I would send it to peer review so the validity of the linear-noise step at criticality gets proper scrutiny.

Referee Report

2 major / 0 minor

Summary. The manuscript studies entropy production in well-mixed reversible chemical reaction networks in the macroscopic-first limit, where nonequilibrium transitions occur as local bifurcations of mass-action dynamics. Combining linear-noise formulas, center-manifold normal forms, and Floquet theory, the authors derive generic scaling exponents for entropy-production fluctuations and responses at pitchfork, transcritical, saddle-node, and Hopf bifurcations. A trajectory-space Cramér-Rao bound is then used to obtain the universal inequality α − 2β ≥ 0, which implies that divergent responses require divergent fluctuations but not conversely.

Significance. If the approximations hold, the work supplies a low-order classification of critical exponents for entropy production and positions fluctuations as a sharper probe of nonequilibrium criticality than responses. The inequality α − 2β ≥ 0 is a clear strength: it follows from an external bound applied to trajectories, is parameter-free, and does not rely on fitted quantities or internal self-consistency arguments.

major comments (2)

[Derivations of exponents via linear-noise and center-manifold reduction] The linear-noise approximation is invoked to extract the scaling exponents at the bifurcation points (see the derivations for each bifurcation type). At these points the real part of the critical eigenvalue vanishes, so the deterministic restoring force for the slow mode disappears and the variance diverges; the small-noise closure therefore ceases to be self-consistent precisely where the exponents are obtained. No error estimate, radius-of-convergence analysis, or higher-order correction is supplied to justify continued use of the linear-noise formulas at criticality.
[Application of center-manifold normal forms and Floquet theory] The manuscript asserts that the center-manifold normal forms and Floquet theory remain valid under the stochastic linear-noise closure at the critical points, yet supplies no explicit verification that the reduced stochastic dynamics preserve the leading scaling after the deterministic reduction. This is load-bearing for the claimed generic exponents.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their careful and insightful review of our manuscript. We address each of the major comments in detail below and outline the revisions we will make to strengthen the presentation of our results.

read point-by-point responses

Referee: [Derivations of exponents via linear-noise and center-manifold reduction] The linear-noise approximation is invoked to extract the scaling exponents at the bifurcation points (see the derivations for each bifurcation type). At these points the real part of the critical eigenvalue vanishes, so the deterministic restoring force for the slow mode disappears and the variance diverges; the small-noise closure therefore ceases to be self-consistent precisely where the exponents are obtained. No error estimate, radius-of-convergence analysis, or higher-order correction is supplied to justify continued use of the linear-noise formulas at criticality.

Authors: We concur that the linear-noise approximation breaks down exactly at the critical point due to the divergence of fluctuations. Nevertheless, the generic scaling exponents are extracted from the singular dependence on the distance to the bifurcation point, where for sufficiently small noise the approximation captures the leading divergence. This approach is standard in the analysis of critical phenomena in stochastic systems. In the revised manuscript, we will add a new subsection discussing the validity of the linear-noise approximation in the vicinity of the bifurcations, including a scaling analysis of the error terms in the van Kampen expansion and references to supporting literature. revision: yes
Referee: [Application of center-manifold normal forms and Floquet theory] The manuscript asserts that the center-manifold normal forms and Floquet theory remain valid under the stochastic linear-noise closure at the critical points, yet supplies no explicit verification that the reduced stochastic dynamics preserve the leading scaling after the deterministic reduction. This is load-bearing for the claimed generic exponents.

Authors: The deterministic center-manifold reduction yields the normal form governing the slow dynamics, after which the linear-noise approximation is applied to compute the stochastic entropy production. Because the fast modes are adiabatically eliminated and their fluctuations do not couple to the leading-order entropy production of the slow mode, the scaling exponents remain unchanged. We will include in the revised version an appendix providing the explicit reduction of the stochastic equations and confirming that the leading scalings for α and β are preserved. revision: yes

Circularity Check

0 steps flagged

No circularity; derivation uses independent external methods and bounds

full rationale

The paper obtains generic exponents for entropy-production fluctuations and responses at pitchfork, transcritical, saddle-node, and Hopf bifurcations by applying linear-noise formulas, center-manifold normal forms, and Floquet theory to mass-action dynamics in the macroscopic limit. These are standard, externally established tools whose validity is assumed rather than derived from the paper's own results. The universal scaling inequality α - 2β ≥ 0 is produced by applying an external trajectory-space Cramér-Rao bound, which supplies an independent constraint not reducible to any fitted parameter or self-citation within the manuscript. No load-bearing step equates a claimed prediction to its own input by construction, and the central claims retain independent mathematical content beyond the paper's definitions.

Axiom & Free-Parameter Ledger

0 free parameters · 4 axioms · 0 invented entities

The central claim rests on standard domain assumptions from stochastic thermodynamics and local bifurcation theory; no free parameters or invented entities are introduced in the abstract.

axioms (4)

domain assumption Mass-action kinetics govern the macroscopic deterministic dynamics of well-mixed reversible chemical reaction networks
Invoked to define the macroscopic-first limit and the bifurcations
domain assumption Linear-noise approximation remains valid near the critical points
Used to obtain fluctuation formulas
domain assumption Center-manifold reduction and Floquet theory apply to the local bifurcations
Used to extract normal forms and exponents
standard math Trajectory-space Cramér-Rao bound holds for entropy-production observables
Basis for the universal inequality

pith-pipeline@v0.9.0 · 5418 in / 1505 out tokens · 40682 ms · 2026-05-08T17:06:12.409237+00:00 · methodology

Review history (2 revisions) →

discussion (0)

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

Singular Behavior of Observables at Hopf Bifurcations
cond-mat.stat-mech 2026-05 unverdicted novelty 7.0

Time-averaged observables generically exhibit kink or higher-order singularities in their derivatives at supercritical Hopf bifurcations because phase averaging over the limit cycle eliminates odd powers of the oscill...
Singular Behavior of Observables at Hopf Bifurcations
cond-mat.stat-mech 2026-05 unverdicted novelty 6.0

Time-averaged observables exhibit kink or higher-order derivative singularities at supercritical Hopf bifurcations because phase averaging eliminates odd powers of the limit-cycle amplitude while the squared amplitude...

Reference graph

Works this paper leans on

61 extracted references · cited by 1 Pith paper

[1]

Summary of the universality classes in the entropy production

(0 −, 0−) TABLE I. Summary of the universality classes in the entropy production. The stable solutions for the normal form equations are depicted with a blue curve. The exponentsα ± andβ ± are the diverging exponents forθ > θ c andθ < θ c, respectively. The symbol 0 − means that the exponents are nonpositive meaning no divergence.∅implies that the exponen...
[2]

L. D. Landau and E. M. Lifshitz,Statistical Physics, 3rd ed., Course of Theoretical Physics, Vol. 5 (Butterworth- Heinemann, 1980)

1980
[3]

Klages, W

R. Klages, W. Just, and C. Jarzynski, eds.,Nonequilib- rium Statistical Physics of Small Systems: Fluctuation Relations and Beyond(Wiley-VCH Verlag GmbH & Co. KGaA, 2013)

2013
[4]

Seifert,Stochastic Thermodynamics(Cambridge Uni- versity Press, 2025) p

U. Seifert,Stochastic Thermodynamics(Cambridge Uni- versity Press, 2025) p. 493

2025
[5]

A. C. Barato and U. Seifert, Thermodynamic uncertainty relation for biomolecular processes, Phys. Rev. Lett.114, 158101 (2015)

2015
[6]

T. R. Gingrich, J. M. Horowitz, N. Perunov, and J. L. England, Dissipation bounds all steady-state current fluctuations, Phys. Rev. Lett.116, 120601 (2016)

2016
[7]

Dechant and S.-i

A. Dechant and S.-i. Sasa, Current fluctuations and transport efficiency for general langevin systems, Journal of Statistical Mechanics: Theory and Experiment2018, 063209 (2018)

2018
[8]

Aslyamov, K

T. Aslyamov, K. Ptaszy´ nski, and M. Esposito, Nonequi- librium fluctuation-response relations: From identities to bounds, Phys. Rev. Lett.134, 157101 (2025)

2025
[9]

Seifert, Entropy production along a stochastic tra- jectory and an integral fluctuation theorem, Phys

U. Seifert, Entropy production along a stochastic tra- jectory and an integral fluctuation theorem, Phys. Rev. Lett.95, 040602 (2005)

2005
[10]

G. E. Crooks, Entropy production fluctuation theorem and the nonequilibrium work relation for free energy dif- ferences, Phys. Rev. E60, 2721 (1999)

1999
[11]

D. J. Evans, E. G. D. Cohen, and G. P. Morriss, Proba- bility of second law violations in shearing steady states, Phys. Rev. Lett.71, 2401 (1993)

1993
[12]

Gallavotti and E

G. Gallavotti and E. G. D. Cohen, Dynamical ensembles in nonequilibrium statistical mechanics, Phys. Rev. Lett. 74, 2694 (1995)

1995
[13]

J. L. Lebowitz and H. Spohn, A gallavotti–cohen-type symmetry in the large deviation functional for stochastic dynamics, Journal of Statistical Physics95, 333 (1999)

1999
[14]

Jarzynski, Hamiltonian derivation of a detailed fluc- tuation theorem, Journal of Statistical Physics98, 77 (2000)

C. Jarzynski, Hamiltonian derivation of a detailed fluc- tuation theorem, Journal of Statistical Physics98, 77 (2000)

2000
[15]

Gaspard, Fluctuation theorem for nonequilibrium re- actions, The Journal of Chemical Physics120, 8898 (2004)

P. Gaspard, Fluctuation theorem for nonequilibrium re- actions, The Journal of Chemical Physics120, 8898 (2004)

2004
[16]

Shiraishi, K

N. Shiraishi, K. Funo, and K. Saito, Speed limit for clas- sical stochastic processes, Phys. Rev. Lett.121, 070601 (2018)

2018
[17]

A. Dechant, Minimum entropy production, detailed balance and wasserstein distance for continuous-time markov processes, Journal of Physics A: Mathematical and Theoretical55, 094001 (2022)

2022
[18]

Van Vu and K

T. Van Vu and K. Saito, Thermodynamic unification of optimal transport: Thermodynamic uncertainty relation, minimum dissipation, and thermodynamic speed limits, Physical Review X13, 011013 (2023)

2023
[19]

J. S. Lee, S. Lee, H. Kwon, and H. Park, Speed limit for a highly irreversible process and tight finite-time landauer’s bound, Physical review letters129, 120603 (2022)

2022
[20]

K. Funo, N. Shiraishi, and K. Saito, Speed limit for open quantum systems, New Journal of Physics21, 013006 (2019)

2019
[21]

Falasco and M

G. Falasco and M. Esposito, Macroscopic stochastic ther- modynamics, Rev. Mod. Phys.97, 015002 (2025)

2025
[22]

Freitas and M

N. Freitas and M. Esposito, Maxwell demon that can work at macroscopic scales, Phys. Rev. Lett.129, 120602 (2022)

2022
[23]

J. M. R. Parrondo, The szilard engine revisited: Entropy, macroscopic randomness, and symmetry breaking phase transitions, Chaos: An Interdisciplinary Journal of Non- 6 linear Science11, 725 (2001)

2001
[24]

Nguyen and U

B. Nguyen and U. Seifert, Exponential volume depen- dence of entropy-current fluctuations at first-order phase transitions in chemical reaction networks, Phys. Rev. E 102, 022101 (2020)

2020
[25]

Remlein and U

B. Remlein and U. Seifert, Nonequilibrium fluctuations of chemical reaction networks at criticality: The schl¨ ogl model as paradigmatic case, The Journal of Chemical Physics160, 134103 (2024)

2024
[26]

Nguyen, U

B. Nguyen, U. Seifert, and A. C. Barato, Phase transition in thermodynamically consistent biochemical oscillators, The Journal of Chemical Physics149, 045101 (2018)

2018
[27]

C. E. Fiore, P. E. Harunari, C. E. F. Noa, and G. T. Landi, Current fluctuations in nonequilibrium discontin- uous phase transitions, Phys. Rev. E104, 064123 (2021)

2021
[28]

Shim, H.-M

P.-S. Shim, H.-M. Chun, and J. D. Noh, Macroscopic time-reversal symmetry breaking at a nonequilibrium phase transition, Phys. Rev. E93, 012113 (2016)

2016
[29]

Ptaszy´ nski and M

K. Ptaszy´ nski and M. Esposito, Critical heat current fluc- tuations in curie-weiss model in and out of equilibrium, Phys. Rev. E111, 034125 (2025)

2025
[30]

Oberreiter, U

L. Oberreiter, U. Seifert, and A. C. Barato, Stochas- tic discrete time crystals: Entropy production and sub- harmonic synchronization, Phys. Rev. Lett.126, 020603 (2021)

2021
[31]

Chudak, M

M. Chudak, M. Esposito, and K. Ptaszy´ nski, Syn- chronization of thermodynamically consistent stochastic phase oscillators, Phys. Rev. E113, 034129 (2026)

2026
[32]

M. J. Kewming, M. T. Mitchison, and G. T. Landi, Di- verging current fluctuations in critical kerr resonators, Phys. Rev. A106, 033707 (2022)

2022
[33]

Goldenfeld,Lectures On Phase Transitions And The Renormalization Group, 1st ed

N. Goldenfeld,Lectures On Phase Transitions And The Renormalization Group, 1st ed. (CRC Press, 1992)

1992
[34]

Nishimori and G

H. Nishimori and G. Ortiz,Elements of Phase Transi- tions and Critical Phenomena(Oxford University Press, 2010)

2010
[35]

Y. A. Kuznetsov,Elements of Applied Bifurcation The- ory, 4th ed., Applied Mathematical Sciences (Springer, Cham, 2023)

2023
[36]

Chicone,Ordinary Differential Equations with Appli- cations, 1st ed., Texts in Applied Mathematics, Vol

C. Chicone,Ordinary Differential Equations with Appli- cations, 1st ed., Texts in Applied Mathematics, Vol. 34 (Springer, New York, NY, 1999)

1999
[37]

Gardiner,Stochastic Methods, 4th ed., Springer Series in Synergetics (Springer Berlin Heidelberg, 2009)

C. Gardiner,Stochastic Methods, 4th ed., Springer Series in Synergetics (Springer Berlin Heidelberg, 2009)

2009
[38]

J. M. Horowitz, Diffusion approximations to the chem- ical master equation only have a consistent stochastic thermodynamics at chemical equilibrium, The Journal of Chemical Physics143, 044111 (2015)

2015
[39]

S. R. d. Groot and P. Mazur,Non-equilibrium thermody- namics, dover ed ed. (Dover Publications, 1984)

1984
[40]

I. R. Epstein and J. A. Pojman,An Introduction to Non- linear Chemical Dynamics: Oscillations, Waves, Pat- terns, and Chaos(Oxford University Press, 1998)

1998
[41]

T. G. Kurtz, The relationship between stochastic and deterministic models for chemical reactions, The Journal of Chemical Physics57, 2976 (1972)

1972
[42]

T. G. Kurtz, Limit theorems for sequences of jump markov processes approximating ordinary differential processes, Journal of Applied Probability8, 344–356 (1971)

1971
[43]

Rao and M

R. Rao and M. Esposito, Nonequilibrium thermodynam- ics of chemical reaction networks: Wisdom from stochas- tic thermodynamics, Phys. Rev. X6, 041064 (2016)

2016
[44]

Supplemental Material
[45]

For the choice ofa ℓ =a (0) ℓ +g(θ)ˆaℓ, the critical behav- ior shows [g(θ)−g(θ c)]−α ∼[g ′(θc)]−α(θ−θ c)−α, which results in the same exponentα
[46]

S. H. S. H. Strogatz,Nonlinear dynamics and chaos : with applications to physics, biology, chemistry, and en- gineering(Addison-Wesley Pub. Co., 1994)

1994
[47]

We emphasize that thec→ −csymmetry of the pitchfork normal form does not, by itself, implyσ(c) =σ(−c). The entropy production is a physical observable expressed in the original concentrations and reservoir affinities, and it generally contains an odd component in the center- manifold coordinate,σ(c) =σ 0 +s 1c+s 2c2 +· · ·with s1 ̸= 0. The pitchfork expon...
[48]

Prigogine and R

I. Prigogine and R. Lefever, Symmetry breaking instabil- ities in dissipative systems. ii, The Journal of Chemical Physics48, 1695 (1968)

1968
[49]

Kondepudi,Modern Thermodynamics(John Wiley & Sons, Ltd, 2014)

D. Kondepudi,Modern Thermodynamics(John Wiley & Sons, Ltd, 2014)

2014
[50]

Schl¨ ogl, Chemical reaction models for non-equilibrium phase transitions, Zeitschrift f¨ ur Physik253, 147 (1972)

F. Schl¨ ogl, Chemical reaction models for non-equilibrium phase transitions, Zeitschrift f¨ ur Physik253, 147 (1972)

1972
[51]

Vellela and H

M. Vellela and H. Qian, Stochastic dynamics and non- equilibrium thermodynamics of a bistable chemical sys- tem: the schl¨ ogl model revisited, Journal of The Royal Society Interface6, 925 (2008)

2008
[52]

Keizer, Nonequilibrium thermodynamics and the sta- bility of states far from equilibrium, Accounts of Chemi- cal Research12, 243 (1979)

J. Keizer, Nonequilibrium thermodynamics and the sta- bility of states far from equilibrium, Accounts of Chemi- cal Research12, 243 (1979)

1979
[53]

Vellela and H

M. Vellela and H. Qian, A quasistationary analysis of a stochastic chemical reaction: Keizer’s paradox, Bulletin of mathematical biology69, 1727 (2007)

2007
[54]

Universal criticality of entropy production in chemical reaction networks

D. J. Higham, An algorithmic introduction to numeri- cal simulation of stochastic differential equations, SIAM review43, 525 (2001). 7 END MATTER Variance formulas (9) and (10).—We outline the derivation (10) for the limit cycle case, which is reduced to (9) for the fixed point cases. The details are ex- plained in the SM [43]. For the size expansionnℓ(Γ)...

2001
[55]

+· · ·, ˙c2 =g c,2(c1, c2) =ω 0c1 +µc 2 −γc 2(c2 1 +c 2
[56]

+· · ·. (S.56) Writingz=re iϕ, we obtain the equations ˙r=µr−γr 3 +O(µ 2r+µr 3 +r 5), ˙ϕ=ω 0 +O(µ+r 2).(S.57) Ifγ >0, the Hopf bifurcation is supercritical, and a stable limit cycle of radius emerges forµ >0 (equivalently, θ > θ c): r= p µ/γ , T(µ) = 2π ω0 +O(µ),(S.58) whereT(µ) is the period of the limit cycle. Ifγ <0, the Hopf bifurcation is subcritical...
[57]

The stable equilibria are given by ¯c=± √µforµ >0 and ¯c= 0 forµ <0

Pitchfork and transcritical bifurcations For the pitchfork bifurcation, we use the normal form (S.52), ˙c=µc−c 3 . The stable equilibria are given by ¯c=± √µforµ >0 and ¯c= 0 forµ <0. Hence, λ+ =∂ cgc(¯c, µ) =µ−3¯c2 =−2µ(µ >0),(S.60) λ− =∂ cgc(0, µ) =µ=−|µ|(µ <0).(S.61) Consequently, the generic behavior of the fluctuation of entropy production is Varσ∝ ...
[58]

Using the normal form (S.54), ˙c=µ−c 2 , one finds the stable equilibrium ¯c= √µforµ >0

Saddle-node bifurcation In contrast to the pitchfork and transcritical bifurcations, the saddle-node bifurcation shows a sudden loss of the local stable equilibrium branch. Using the normal form (S.54), ˙c=µ−c 2 , one finds the stable equilibrium ¯c= √µforµ >0. Therefore, λ+ =∂ cgc(¯c, µ) =−2¯c=−2√µ(µ >0).(S.65) Hence, the generic divergence of the entrop...
[59]

The case ofθ > θ c We here discuss the divergence behavior for the regimeµ:=θ−θ c >0 near the Hopf bifurcation point. Since the periodic motion appears in this regime, we use the monodromy matrixM θ defined as Mθ =Texp hZ T 0 dtS θ(¯x(t)) i ,(S.67) 19 where ¯x(t) is the stable periodic solution of Eq.(S.6).Tis the period of the limit cycle which depends o...
[60]

Note that the stability matrix has finite eigenvalues as in (S.55)

The case ofθ < θ c For the regime ofθ < θ c, the stable solution is not the periodic motion, but the fixed point ¯x, and hence we use the formula in (S.37) to discuss the fluctuation of entropy production. Note that the stability matrix has finite eigenvalues as in (S.55). Eigenvalues in the stable+unstable manifold areO(1), and eigenvalues in the center ...
[61]

Proof.Letµ:=θ−θ c >0, and let ¯ζ(t) = ¯c(t)T ,0 T T 22 be theT(µ)-periodic orbit in the flattened coordinatesζ= (c T ,w T )T , wherew=s−h(c)

Proofs for the lemmas 1-3 The proof of the lemma 1 is as follows. Proof.Letµ:=θ−θ c >0, and let ¯ζ(t) = ¯c(t)T ,0 T T 22 be theT(µ)-periodic orbit in the flattened coordinatesζ= (c T ,w T )T , wherew=s−h(c). The corresponding orbit in the original coordinates is denoted by ¯x(t) = Ψ(¯ζ(t)). Here Ψ is the smooth coordinate map from the flattened coordinate...