Diffusion stabilises time-periodic solutions in conservation laws coupled to a relaxation oscillator

Hiroshi Horii; Julien Barr\'e; Nils Berglund

REVIEW 2 major objections 9 minor 43 references

Diffusion stabilises time-periodic solutions in conservation laws coupled to a relaxation oscillator

T0 review · 2 major / 9 minor · reviewed 2026-07-31 · grok-4.5

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Small diffusion collapses an infinite family of relaxation oscillations into one linearly stable periodic orbit.

desk verdict Clean viscosity-selection theorem for a globally coupled conservation-law–relaxation oscillator; the a-posteriori slope bound scopes the basin but does not break the stated claim. read the letter →

arxiv 2607.24994 v1 pith:PM5OA36K submitted 2026-07-27 math.AP math.DS

Julien Barré , Nils Berglund , Hiroshi Horii This is my paper · ORCID

classification math.APmath.DS MSC 35Q4934C2634K3335P15

keywords slow-fastsystemrelaxationoscillationconservationlawaveragingspectralstabilityviscousregularisationtransport-ODEcouplingplasmavoid

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

A viscous conservation law for a particle density is coupled to a fast ODE that can sustain relaxation oscillations. When viscosity vanishes, every admissible initial density profile evolves into its own time-periodic solution, all sharing the same period set by the slow–fast ODE. The paper proves that a small positive viscosity, acting on symmetric monotone initial data, selects a single isolated periodic orbit and makes it linearly asymptotically stable, with attraction rate proportional to the viscosity. The argument rewrites the profile via its reciprocal function, averages over the relaxation cycle, and obtains a stationary averaged equation whose unique admissible steady state is shown stable by spectral methods. The result matters because it shows how weak diffusion can lift a massive degeneracy and pick a preferred oscillatory pattern in a transport–ODE model motivated by dusty-plasma voids.

What carries the argument

Averaging of the reciprocal-profile equation over one relaxation cycle. After a slow-time change of variables the leading-order averaged dynamics is autonomous; its unique stationary solution compatible with a conserved mass integral is linearly stable by a Sturm–Liouville spectral-gap argument on an invariant hyperplane, and a Poincaré-map/implicit-function argument lifts the conclusion back to the original system.

What would settle it

Simulate or analyse the system from a symmetric monotone initial profile whose midpoint slope is large enough that the viscous term cannot prevent steepening: if a shock forms or the orbit fails to approach the unique averaged steady state predicted by the theory, the selection-and-stability claim fails for that regime.

Watch

Extended reading notes

Core claim

Under the paper’s structural assumptions on the nullclines and for symmetric monotone initial data, sufficiently small positive viscosity ε (with time-scale ratio η at most order 1/log(1/ε)) produces an isolated time-periodic solution of the coupled system that is linearly asymptotically stable in H²×ℝ, attracts nearby solutions at rate O(ε), and whose period converges to the common inviscid period as ε→0. Without viscosity the same system admits an uncountable family of such periodic solutions, one for each initial profile.

Load-bearing premise

The inverse density profile must keep a strictly positive slope at the midpoint for all time, so that no shock forms and the averaging hypotheses stay valid; this is only guaranteed when the initial data are already close to the selected orbit.

Editorial extensions

If this is right

Weak diffusion is enough to select a unique relaxation-oscillation density profile among the continuum present at zero viscosity.
The attraction rate to that orbit scales linearly with the diffusion coefficient ε.
The selected period converges to the period of the inviscid slow–fast limit cycle as ε→0.
Symmetric monotone data near the orbit remain strictly increasing for all time, so the reciprocal-profile representation never breaks.
Changing the relative placement of the slow and fast nullclines can replace large-amplitude relaxation cycles by small-amplitude or mixed-mode patterns.

Reading between the lines

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

The same averaging-plus-spectral strategy may apply to other scalar conservation laws globally coupled to a relaxation oscillator, not only the dusty-plasma idealisation.
Non-symmetric initial data appear numerically to select different limit cycles labelled by total mass; a full uniqueness theory without symmetry remains open.
If the midpoint slope is allowed to approach zero, folded-node or canard mechanisms could generate mixed-mode oscillations, consistent with the paper’s numerical canard-like transients.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

The new result is Theorem 2.7: small viscosity collapses an infinite-dimensional family of inviscid relaxation cycles (Prop 2.5) to one isolated, linearly stable periodic orbit for symmetric monotone data, with rate O(ε) and period converging to the inviscid one. That combination—non-propagating, globally mass-coupled viscous conservation law plus fast relaxation ODE—is not covered by classical viscous-front theory, periodic travelling-wave stability, or standard FHN slow–fast ODEs. The architecture is standard pieces newly assembled: reciprocal profile X, polar-like (r,θ) coordinates outside the central rectangle, GSPT radial control via Szmolyan–Wechselberger, cycle averaging to a stationary problem for G, conserved K fixing the unique averaged state, rank-one Sturm–Liouville spectral gap, then Poincaré + IFT through the O(ε², η^{1/3}ε) remainders.

What works: global existence and symmetry are clean; the averaging and spectral analysis are careful; the hypotheses (S-shaped critical manifold, slow nullcline geometry) are structural rather than fitted; numerics (finite-domain Lax scheme) illustrate the selection, the O(ε) decay, and the canard-like transition when the slow nullcline nears a fold. The authors flag the open questions (non-symmetric data, large slope, possible longer-period orbits, MMOs) instead of hiding them.

The real soft spot is exactly the one the stress-test flags: every averaging/GSPT step needs ∂_m X(θ,0) ⩾ q_0 > 0 so shocks never form and normal switching stays valid. That bound is closed only a posteriori inside an H² ball around the selected orbit (Cor 5.14). Logically the theorem is fine—it assumes the neighbourhood (2.9) and verifies persistence—but the abstract’s “selects a unique” language is stronger than the proved basin. Large-slope or non-monotone data remain open, and Remark 4.7 correctly notes folded-node/MMO possibilities once the bound fails. Minor: numerics are code-free and finite-domain, so they support rather than prove.

This is for people working on slow–fast PDEs, viscous conservation laws with global coupling, or mathematical plasma models. The math is solid enough for a serious referee. I would send it to peer review; the scoped limitations are honest, not fatal.

Referee Report

2 major / 9 minor

Summary. The paper studies a viscous scalar conservation law on R, ∂_t n = −a(E,M)∂_x[n(1−n)] + ε∂_{xx}n, globally coupled through M(t) = −∫_{−∞}^0 n dx to a fast ODE ηĖ = g(E,M) of van der Pol type. For ε = 0 the (E,M) dynamics closes and exhibits a relaxation limit cycle, and each monotone initial profile n_0 produces its own periodic modulation: an uncountable family of cycles with a common period (Prop. 2.5). The main result (Thm. 2.7) is that for 0 < ε ≪ 1 and 0 ≤ η < c/log(ε⁻¹), diffusion selects an isolated periodic orbit, linearly asymptotically stable with rate O(ε) for symmetric, monotone, H²-nearby data, with period converging to the inviscid one. The proof uses a reciprocal-function formulation (Prop. 4.1), polar-like coordinates and GSPT control of the radial variable (Prop. 4.5, via [40]), a θ-time change (Prop. 4.8), averaging to an autonomous equation for Ȳ (Props. 4.4/4.11), a unique stationary state selected by the conserved quantity K (Prop. 5.3), a spectral-gap analysis of a rank-one Sturm–Liouville perturbation (Props. 5.6–5.10), and an implicit-function/Poincaré argument for the untruncated system (§6). Global existence is proved in Appendix A; numerics in §3 support the O(ε) decay rate.

Significance. If correct, this is a strong and, to my knowledge, first rigorous result showing that vanishing viscosity collapses an infinite-dimensional continuum of relaxation-oscillation cycles to a single stable orbit in a PDE–ODE coupled conservation law, a selection mechanism relevant to void oscillations in dusty plasmas. The argument is a forward derivation: the averaged nonlinearity G (eqs. (4.9)/(4.38)) and the conserved quantity K are computed from the model, not fitted. The spectral analysis of §5.2 — reducing stability of a non-self-adjoint rank-one perturbation to the Laplace transform of a positive measure — is elegant and self-contained. The paper ships a quantitative, falsifiable prediction (decay rate O(ε) in (2.10)) that is tested numerically (Fig. 7: fitted slopes s±(ε) ≈ −2ε, −2.3ε). Limitations (basin restriction, possible MMOs near shock formation) are disclosed honestly in §2.3 and Remark 4.7.

major comments (2)

[§4.2.2, Proposition 4.5] The application of the Szmolyan–Wechselberger results [40] is not directly justified: [40] concerns autonomous slow–fast ODEs, whereas here the slow vector field g_1 = a/4 − εp(y) contains p(y) = 1/∂_m X(θ,0), which is not a function of (x,y,z) alone but is slaved to the PDE component X. The lower bound (4.19) controls |p|, but the GSPT estimates (slow-manifold attraction, fold extension, Poincaré map) are C^1-robust statements and require control of ∂_θ p as well, i.e. of ∂_θ∂_m X(θ,0). Since the radial asymptotics (4.18) underpins both the time change (Prop. 4.8) and the averaging (Prop. 4.11), the uniformity of [40]'s estimates in p (and its θ-derivative, controllable via (4.26) and the H²-closeness) should be stated and verified explicitly.
[§5.3–§6, Remark 6.1] This remark performs load-bearing work: Corollary 5.14 establishes persistence of the bound ∂_m X(θ,0) ≥ q_0 > 0 (equivalently (5.23)) only for the truncated averaged equation (5.1), yet assumptions (4.19)/(4.23) are imposed on the full O(ε²)- (resp. O(ε²+η^{1/3}ε)-) perturbed dynamics. The transfer of the H²-ball attraction and of the pointwise lower bound to the untruncated system is exactly what closes the a-posteriori shock-free hypothesis inside the basin of Thm. 2.7. A one-sentence 'perturbation argument' is not sufficient here; please supply a proof or a detailed sketch (e.g. a Duhamel estimate for the perturbed semigroup combined with Lemma 5.12).

minor comments (9)

[§2.1 vs §2.3] Notation collision: n∗ denotes the fixed tanh-type reference profile in §2.1 (Thm. 2.1) and the periodic orbit n∗(t,x) in Thm. 2.7. Please distinguish them.
[§5.1, after Proposition 5.3] The translation of K(0) > z_0/2 appears to have the inequality reversed: K = −2M− + O(ε) > z_0/2 = 2(M+ − M−) + O(η^{1/3}) yields 0 > 2M+, i.e. M+ < O(ε) + O(η^{1/3}). The displayed '−2M− + O(ε) ≤ 2(M+−M−) + O(η^{1/3})' should be '≥'. The final conclusion (M+ negative and small) is correct.
[§5.2, proof of Proposition 5.10] 'Since q(0)=1' should read 'q(0)>0' (Prop. 5.9 gives q(0)=H(0)>0). The argument is unaffected.
[§5, proof of Proposition 5.1] 'The integral (5.3) is well-defined' — (5.3) is the list of limit properties; the reference should be to (4.9)/(4.38).
[§5.2, Corollary 5.11] The exclusion of eigenvalue sequences accumulating on the imaginary axis works via the analytic function F, but a sentence confirming that eigenvalues of A outside spec(A_0) are exactly the zeros of F−1/2 (and hence isolated) would close the argument, given the degeneracy H(1)=0 of the Sturm–Liouville weight.
[Appendix A, Proposition A.2] R = 2(‖ũ_0‖_{H²} ∨ ẽ_0 ∨ m̃_0) should use |ẽ_0|, |m̃_0|, since these can be negative.
[§4.2.2, point 2] The normal switching condition (4.21) is said to hold 'if ε|p(y)| is small enough'; since a(E±,M∓) ≠ 0 by Assumption 2.4, please state the explicit smallness condition on ε/q_0.
[§2.3] Suggest adding one sentence after Thm. 2.7 clarifying that the constants C, M_0, Δ_0 in (2.9) depend on the uniform lower bound q_0 in (4.19), and that persistence of the shock-free bound is proved only within the basin (forward reference to §5.3/Remark 6.1). The conditional logic is correct as stated, but easy to miss on first reading.
[General] Typos: 'Lipshitz' → 'Lipschitz' throughout (Thm. 2.1, §A); the abstract says 'a unique periodic solution' while Thm. 2.7 asserts an isolated one — align wording; Fig. 4 axis labels render as placeholder glyphs in the arXiv PDF; Δx, Δt are specified only for Figs. 6–7, please add them for Figs. 2–5.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: forward derivation from structural hypotheses to existence/stability via averaging and spectral theory

full rationale

Theorem 2.7 is obtained by a self-contained chain: Assumptions 2.3–2.4 fix the geometry; the reciprocal-profile formulation (Prop. 4.1) and polar-like (r,θ) coordinates put the system in slow–fast form; averaging (Props. 4.4, 4.11) produces an autonomous equation whose unique stationary solution is fixed by the model’s conserved quantity K (Props. 4.3, 4.9, 5.2–5.3), not by a fit; linear stability follows from a Sturm–Liouville analysis of the rank-one perturbation A=A0+A1 on the invariant hyperplane S0 (Props. 5.6–5.10, Cor. 5.11); the Poincaré-map/IFT argument (Sec. 6) lifts the truncated averaged orbit to the full system. External citations ([40], [33], [16]) supply standard GSPT facts about relaxation orbits and are not author-overlapping load-bearing uniqueness claims. The a-posteriori slope bound (4.19)/(4.23) closed in §5.3 is a basin-scoping hypothesis already stated in (2.9), not a definitional reduction of the claim to its inputs. No fitted parameter is renamed a prediction; no self-citation forces the result.

Assumptions & free parameters 3 free parameters · 7 assumptions · 3 invented entities

The result rests on standard slow–fast and conservation-law machinery plus several structural modelling hypotheses that encode relaxation oscillations and shock-free symmetric profiles. No numerical free parameters are fitted to produce the existence/stability theorem; ε and η are asymptotic smallness parameters. The main invented modelling objects are the specific global coupling through M and the polar-like (r,θ) chart used for averaging.

free parameters (3)

ε (viscosity) = 0 < ε < ε_0 (ε_0 existential)
Asymptotic small parameter; existence/stability claimed for all sufficiently small ε>0, not fitted to data.
η (time-scale separation) = 0 ≤ η < c/log(ε^{-1})
Must satisfy η = O(1/log(ε^{-1})) for averaging remainders; not fitted.
Numerical scheme parameters (L, μ, Δx, Δt, a0, E0, b3, b1, b0, c, E⋆) = e.g. L=10, ε=0.05, η=0.01, a0=7, ...
Chosen by hand for finite-domain illustrations in §3; they do not enter the proof of Theorem 2.7.

assumptions (7)

domain assumption Flux f(n)=n(1-n) on [0,1] and f(n)=f(1-n), so symmetric profiles remain symmetric and support stationary steps between 0 and 1.
Assumption 2.3 / Prop 2.2; standard Burgers-type flux chosen for dusty-plasma idealisation.
domain assumption S-shaped critical manifold M=h(E) with two folds satisfying nondegeneracy ∂_M g ≠ 0, ∂_{EE} g ≠ 0, and a slow nullcline E=k(M) intersecting once between the folds (Assumption 2.4).
Encodes van der Pol–type relaxation oscillations for (E,M) when ε=0; taken as modelling hypothesis.
domain assumption Initial density is H^2∩L^1 relative to a fixed background n_*, centrally symmetric, and strictly monotone (Assumption 2.3).
Used throughout to keep n(t,0)=1/2, define the inverse profile X, and avoid shocks.
standard math Geometric singular perturbation results on slow manifolds near folds and Poincaré maps for relaxation oscillations in R^3 (Szmolyan–Wechselberger and related GSPT).
Invoked in Prop 4.5 via [40] to control r(θ) asymptotics for small η.
standard math Averaging and implicit-function theorem for Poincaré maps of periodically forced ODEs / evolution equations with spectral gap.
Props 4.4, 4.11 and §6; classical perturbation theory once the averaged linearisation has a gap on the conserved hyperplane.
standard math Hille–Yosida / analytic-semigroup bounds and Laplace-transform representation of the resolvent for the Sturm–Liouville operator A_0.
§5.2, used to analyse F(λ) and prove the spectral gap on S_0.
standard math Global Lipschitz and boundedness of a,f,g for the bare existence theorem; later relaxed by a posteriori bounds near the periodic orbit.
Theorem 2.1 / Appendix A; standard fixed-point setup for the coupled PDE–ODE.

invented entities (3)

Global mass coupling M(t)=−∫_{-∞}^0 n(t,x) dx feeding a fast ODE for E
purpose: Idealised long-range Poisson-like feedback that closes the relaxation oscillator with the conservation law.
Modelling choice motivated by dusty plasmas; not derived from a full Poisson–Vlasov system.
Polar-like (r,θ) coordinates on eight regions of the (E,M)-plane outside the central rectangle
purpose: Convert the slow–fast (E,M) dynamics into a form where θ is monotone time and averaging applies uniformly, including corners and jumps.
Technical chart invented for the proof (§4.2.1, Table 1); no claim of independent physical meaning.
Averaged nonlinearity G(z) built from 1/|ã(s)|(z−4φ(s)) over one relaxation period independent evidence
purpose: Effective flux whose stationary inverse-profile equation selects the unique periodic density shape.
Derived object (eqs. 4.9 / 4.38), not postulated; listed because it is the key reduced entity carrying the selection mechanism.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Diffusion stabilises time-periodic solutions in conservation laws coupled to a relaxation oscillator." pith.science (2026). https://pith.science/paper/PM5OA36K

@misc{pith2026260724994,
  author       = {Pith},
  title        = {Pith review of: Diffusion stabilises time-periodic solutions in conservation laws coupled to a relaxation oscillator},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PM5OA36K}},
  note         = {Machine review of arXiv:2607.24994}
}

read the original abstract

We study a viscous one-dimensional conservation law, coupled to a fast ordinary differential equation. For a vanishing viscosity, the system has an infinite-dimensional family of time-periodic solutions, corresponding to relaxation oscillations. We show that for symmetric initial conditions, a small positive viscosity selects a unique periodic solution, which we prove to be linearly stable. The proof exploits the slow-fast structure through an averaging strategy, as well as spectral-theoretic methods. The results are illustrated by numerical simulations.

Figures

Figures reproduced from arXiv: 2607.24994 by the authors.

**Figure 1.** Geometry of the nullclines g(E,M) = 0 (or M = h(E), critical manifold) and a(E,M) = 0 (or E = k(M), slow nullcline). • Fold points: One has ∂Mg(E±, h(E±)) ≠ 0 and ∂EEg(E±, h(E±)) ≠ 0 . • Slow nullcline: There exists a C 2 function k ∶ R → R such that a(E,M) > 0 ⇔ E < k(M) , a(E,M) < 0 ⇔ E > k(M) . Furthermore, the nullclines {M = h(E)} and {E = k(M)} intersect exactly once, between the points of abscissa E− and E+.… view at source ↗

**Figure 2.** Time evolution of the density profile n(t, x) for the transport-ODE model. The profiles are shown at the selected times t = 0, 5.0, 5.3, 20.0, 20.3, and 100.0. The pairs t = 5.0, 5.3 and t = 20.0, 20.3 illustrate the periodic behaviour of the density profile, while the profile at t = 100.0 shows its long-time behaviour. Parameters: L = 10, µ = 0.1, ε = 0.05, a0 = 7, E0 = 1.56, b3 = 2, b1 = −0.4, b0 = 0.92, c = 0.8, … view at source ↗

**Figure 3.** Phase portraits in the E-Msim plane for different values of the time-scale separation parameter η. Panels (a), (b), and (c) correspond to η = 0.01, η = 0.1, and η = 1.0, respectively. The curve shows the trajectory of (E(t),Msim(t)), and the dashed curve is the critical manifold g(E,Msim) = 0, corresponding to the cubic nonlinearity in (3.3). Parameters: L = 10, µ = 0.1, ε = 0.05, a0 = 7, E0 = 1.56, b3 = 2, b1 = −0.… view at source ↗

Figures from the paper (10 more)

**Figure 4.** Figure 4: Time series of Msim(t) and convergence of the maximum and minimum values taken over each oscillation period, for different values of η. Panels (a)-(c) show the time series of Msim(t) for η = 0.01, η = 0.1, and η = 1.0, respectively. For each oscillation period, we take…

**Figure 5.** Figure 5: Density profiles near the local maximum and local minimum values of Msim(t) along the cycle, for different values of the time-scale separation parameter η. Panels (a), (b), and (c) show the profiles near the local maximum side, denoted by Msim + , for η = 0.01, η = 0.1…

**Figure 6.** Figure 6: Differences ∣Msim ± (t) − Msim ±,final∣ between the maximum and minimum values of Msim(t) in each oscillation period and their corresponding final values, computed as in [PITH_FULL_IMAGE:figures/full_fig_p014_6.png]

**Figure 7.** Figure 7: Dependence of the fitted exponential-decay exponents s± on the diffusion parameter ε for fixed time scale separation η = 0.01. The points correspond to ε ∈ {0.01, 0.025, 0.04, 0.05, 0.06, 0.075, 0.10}, and the dashed lines show the fits s±(ε) = a±ε + b±. Panel (a) sho…

**Figure 8.** Figure 8: Comparison of density profiles starting from different initial conditions. Panels (a) and (c) show the initial profiles: hyperbolic tangent and linear profiles in panel (a), and hyperbolic tangent and piecewise-linear profiles in panel (c). Panels (b) and (d) show the…

**Figure 9.** Figure 9: Time series of Msim(t) for three nearby values of E0. Panels (a), (b), and (c) correspond to E0 = 1.1925, E0 = 1.193, and E0 = 1.195, respectively. The initial conditions are n tanh 0 and E(0) = 0.4 in all three panels, and all parameters other than E0 are kept fixed.…

**Figure 10.** Figure 10: Decomposition of the (E,M)-plane into regions used to define the polar-like (θ, r) coordinates. 4.2 The case η > 0 We now extend the above approach to the case of small positive η. We will proceed in several steps. 1. Replace the variables (M, E) by polar-like coordin…

**Figure 11.** Figure 11: Schematic representation of the singular periodic orbit Γ(0), and of its deformation Γ(η) for η > 0. 4.2.2 Behaviour of the radial coordinate In this section, we show that the radial variable r(t) behaves in a similar way for ε = 0 and for small positive ε. When ε = …

**Figure 12.** Figure 12: Radial behaviour of the periodic solution for ε = 0, both in the singular limit η = 0, and for small, positive η. Only the θ-interval ∣0, T5] is shown, the behaviour on [T5, T8] is similar to the behaviour on [T1, T4]. Proposition 4.5. Assume there exists a constant q…

**Figure 13.** Figure 13: The spectrum of A0 belongs to [δ,∞). Corollary 5.11 shows that the spectrum of A restricted to the invariant plane S0 is located in the shaded region. The sector with vertex (δ1, 0) is used in Section 5.3 to show that the semigroup e −tA is analytic. Corollary 5.11 (S…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 1 linked inside Pith

[40]

Szmolyan and M

P. Szmolyan and M. Wechselberger. Relaxation oscillations inR 3.J. Diﬀer. Equations, 200(1):69–104, 2004

2004
[1]

Beck and C

M. Beck and C. E. Wayne. Invariant manifolds and the stability of traveling waves in scalar viscous conservation laws.J. Diﬀer. Equations, 244(1):87–116, 2008

2008
[2]

Beckers, J

J. Beckers, J. Berndt, D. Block, M. Bonitz, P. Bruggeman, L. Couëdel, G. L. Delzanno, Y. Feng, R. Gopalakrishnan, F. Greiner, et al. Physics and applications of dusty plas- mas: The perspectives 2023.Physics of Plasmas, 30(12), 2023

2023
[3]

Berglund, B

N. Berglund, B. Gentz, and C. Kuehn. Hunting French ducks in a noisy environment. J. Diﬀerential Equations, 252(9):4786–4841, 2012

2012
[4]

Berglund, B

N. Berglund, B. Gentz, and C. Kuehn. From random Poincaré maps to stochastic mixed-mode-oscillation patterns.J. Dynam. Diﬀerential Equations, 27(1):83–136, 2015

2015
[5]

Berglund and D

N. Berglund and D. Landon. Mixed-mode oscillations and interspike interval statis- tics in the stochastic FitzHugh–Nagumo model.Nonlinearity, 25(8):2303–2335, 2012

2012
[6]

Borsche, R

R. Borsche, R. M. Colombo, and M. Garavello. Mixed systems: ODEs–balance laws. Journal of Diﬀerential equations, 252(3):2311–2338, 2012

2012
[7]

Couëdel, M

L. Couëdel, M. Mikikian, A. A. Samarian, and L. Boufendi. Self-excited void instabil- ity during dust particle growth in a dusty plasma.Physics of plasmas, 17(8), 2010

2010

Show all 43 references

[8]

Desroches, J

M. Desroches, J. Guckenheimer, B. Krauskopf, C. Kuehn, H. M. Osinga, and M. Wech- selberger. Mixed-mode oscillations with multiple time scales.SIAM Rev., 54(2):211– 288, 2012

2012
[9]

Engel, C

M. Engel, C. Kuehn, and B. de Rijk. A traveling wave bifurcation analysis of turbulent pipe ﬂow.Nonlinearity, 35(11):5903, oct 2022

2022
[10]

Fenichel

N. Fenichel. Geometric singular perturbation theory for ordinary diﬀerential equa- tions.J. Diﬀerential Equations, 31(1):53–98, 1979. 51

1979
[11]

Gameiro, T

M. Gameiro, T. Gedeon, W. Kalies, H. Kokubu, K. Mischaikow, and H. Oka. Topolog- ical horseshoes of traveling waves for a fast-slow predator-prey system.J. Dyn. Diﬀer. Equations, 19(3):623–654, 2007

2007
[12]

Godlewski and P.-A

E. Godlewski and P.-A. Raviart.Hyperbolic systems of conservation laws, volume 3/4 of Math. Appl. (Paris). Paris: Ellipses, 1991

1991
[13]

Greenberg and M

J. Greenberg and M. Rascle. Time-periodic solutions to systems of conservation laws. Archive for rational mechanics and analysis, 115(4):395–407, 1991

1991
[14]

Guckenheimer and R

J. Guckenheimer and R. Haiduc. Canards at folded nodes.Mosc. Math. J., 5(1):91–103, 2005

2005
[15]

Guckenheimer, K

J. Guckenheimer, K. Hoﬀman, and W. Weckesser. Bifurcations of relaxation oscilla- tions near folded saddles.Internat. J. Bifur. Chaos Appl. Sci. Engrg., 15(11):3411–3421, 2005

2005
[16]

Haberman

R. Haberman. Slowly varying jump and transition phenomena associated with alge- braic bifurcation problems.SIAM J. Appl. Math., 37(1):69–106, 1979

1979
[17]

M. Hairer. An introduction to stochastic PDEs. Lecture notes, University of Warwick, 2009.http://arxiv.org/abs/0907.4178

2009 arXiv
[18]

Z. Hu, Y. Chen, X. Zheng, F. Huang, G.-f. Shi, and M. Yu. Theory of void formation in dusty plasmas.Physics of Plasmas, 16(6), 2009

2009
[19]

M. A. Johnson, P. Noble, L. M. Rodrigues, and K. Zumbrun. Behavior of periodic so- lutions of viscous conservation laws under localized and nonlocalized perturbations. Inventiones mathematicae, 197(1):115–213, 2014

2014
[20]

M. A. Johnson and K. Zumbrun. Nonlinear stability of periodic traveling-wave so- lutions of viscous conservation laws in dimensions one and two.SIAM J. Appl. Dyn. Syst., 10(1):189–211, 2011

2011
[21]

Kato.Perturbation theory for linear operators.Class

T. Kato.Perturbation theory for linear operators.Class. Math. Berlin: Springer-Verlag, reprint of the corr. print. of the 2nd ed. 1980 edition, 1995

1980
[22]

Krupa and P

M. Krupa and P. Szmolyan. Extending geometric singular perturbation theory to non- hyperbolic points—fold and canard points in two dimensions.SIAM J. Math. Anal., 33(2):286–314 (electronic), 2001

2001
[23]

Krupa and M

M. Krupa and M. Wechselberger. Local analysis near a folded saddle-node singularity. J. Diﬀ. Eq., 248(12):2841–2488, 2010

2010
[24]

Li and X

Z. Li and X. Liu. Existence and stability of slow-fast traveling pulses in a chemical system with quintic nonlinearity.J. Dyn. Diﬀer. Equations, 38(1):143–181, 2026

2026
[25]

I. P. Longo, E. Queirolo, and C. Kuehn. On the transition between autonomous and nonautonomous systems: the case of FitzHugh-Nagumo’s model.Chaos, 34(12):15,
[26]

Matsumura and K

A. Matsumura and K. Nishihara. Asymptotic stability of traveling waves for scalar viscous conservation laws with non-convex nonlinearity.Commun. Math. Phys., 165(1):83–96, 1994. 52

1994
[27]

Mikikian, M

M. Mikikian, M. Cavarroc, L. Couëdel, Y. Tessier, and L. Boufendi. Mixed-mode oscillations in complex-plasma instabilities.Physical Review Letters, 100(22):225005, 2008

2008
[28]

Mikikian, L

M. Mikikian, L. Couëdel, M. Cavarroc, Y. Tessier, and L. Boufendi. Threshold phe- nomena in a throbbing complex plasma.Physical review letters, 105(7):075002, 2010

2010
[29]

Oh and K

M. Oh and K. Zumbrun. Stability of periodic solutions of conservation laws with viscosity: Analysis of the Evans function.Archive for rational mechanics and analysis, 166(2):99–166, 2003

2003
[30]

Pan and G

J. Pan and G. Warnecke. Asymptotic stability of traveling waves for viscous conser- vation laws with dispersion.Adv. Diﬀer. Equ., 9(9-10):1167–1184, 2004

2004
[31]

Pikalev, M

A. Pikalev, M. Pustylnik, C. Räth, and H. M. Thomas. Heartbeat instability as auto- oscillation between dim and bright void regimes.Physical Review E, 104(4):045212, 2021

2021
[32]

Pikalev, I

A. Pikalev, I. Semenov, M. Pustylnik, C. Räth, and H. Thomas. Dim and bright void regimes in capacitively-coupled rf complex plasmas.Plasma Sources Science and Tech- nology, 30(3):035014, 2021

2021
[33]

L. S. Pontryagin and L. V. Rodygin. Approximate solution of a system of ordinary diﬀerential equations involving a small parameter in the derivatives.Soviet Math. Dokl., 1:255–258, 1960

1960
[34]

Rinzel and D

J. Rinzel and D. Terman. Propagation phenomena in a bistable reaction-diﬀusion system.SIAM J. Appl. Math., 42:1111–1137, 1982

1982
[35]

Sandstede

B. Sandstede. Stability of travelling waves. In B. Fiedler, editor,Handbook of dynamical systems. Volume 2, pages 983–1055. Amsterdam: Elsevier, 2002

2002
[36]

Sandstede and A

B. Sandstede and A. Scheel. Hopf bifurcation from viscous shock waves.SIAM journal on mathematical analysis, 39(6):2033–2052, 2008
[37]

Serre.Systems of conservation laws

D. Serre.Systems of conservation laws. I: Hyperbolicity, entropies, shock waves. Translated from the French by I. N. Sneddon. Cambridge: Cambridge University Press, 1999

1999
[38]

D. Serre. Spectral stability of periodic solutions of viscous conservation laws: large wavelength analysis.Communications in Partial Diﬀerence Equations, 30(1-2):259–282, 2005

2005
[39]

Szmolyan and M

P. Szmolyan and M. Wechselberger. Canards inR3.J. Diﬀer. Equations, 177(2):419–453, 2001

2001
[41]

van der Pol

B. van der Pol. On relaxation oscillation.Phil. Mag., 2:978–992, 1926

1926
[42]

Vladimirov, V

S. Vladimirov, V. Tsytovich, and G. Morﬁll. Stability of dust voids.Physics of Plasmas, 12(5), 2005. 53

2005
[43]

S. K. Zhdanov, M. Schwabe, R. Heidemann, R. Sütterlin, H. M. Thomas, M. Rubin- Zuzic, H. Rothermel, T. Hagl, A. V. Ivlev, G. E. Morﬁll, et al. Auto-oscillations in complex plasmas.New Journal of Physics, 12(4):043006, 2010. Julien Barré Institut Denis Poisson (IDP) Université ...

2010

Pith tools

Reviewed July 31, 2026 · model on record in the stance chip above.

[1] [40]

Szmolyan and M

P. Szmolyan and M. Wechselberger. Relaxation oscillations inR 3.J. Diﬀer. Equations, 200(1):69–104, 2004

2004

[2] [1]

Beck and C

M. Beck and C. E. Wayne. Invariant manifolds and the stability of traveling waves in scalar viscous conservation laws.J. Diﬀer. Equations, 244(1):87–116, 2008

2008

[3] [2]

Beckers, J

J. Beckers, J. Berndt, D. Block, M. Bonitz, P. Bruggeman, L. Couëdel, G. L. Delzanno, Y. Feng, R. Gopalakrishnan, F. Greiner, et al. Physics and applications of dusty plas- mas: The perspectives 2023.Physics of Plasmas, 30(12), 2023

2023

[4] [3]

Berglund, B

N. Berglund, B. Gentz, and C. Kuehn. Hunting French ducks in a noisy environment. J. Diﬀerential Equations, 252(9):4786–4841, 2012

2012

[5] [4]

Berglund, B

N. Berglund, B. Gentz, and C. Kuehn. From random Poincaré maps to stochastic mixed-mode-oscillation patterns.J. Dynam. Diﬀerential Equations, 27(1):83–136, 2015

2015

[6] [5]

Berglund and D

N. Berglund and D. Landon. Mixed-mode oscillations and interspike interval statis- tics in the stochastic FitzHugh–Nagumo model.Nonlinearity, 25(8):2303–2335, 2012

2012

[7] [6]

Borsche, R

R. Borsche, R. M. Colombo, and M. Garavello. Mixed systems: ODEs–balance laws. Journal of Diﬀerential equations, 252(3):2311–2338, 2012

2012

[8] [7]

Couëdel, M

L. Couëdel, M. Mikikian, A. A. Samarian, and L. Boufendi. Self-excited void instabil- ity during dust particle growth in a dusty plasma.Physics of plasmas, 17(8), 2010

2010