Phase separation of mixtures after a second quench: composition heterogeneities

Pablo de Castro; Peter Sollich

REVIEW 3 major objections 4 minor 14 references

Phase separation of mixtures after a second quench: composition heterogeneities

T0 review · 3 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read After a second, deeper quench, dense binary mixtures develop long-lived composition patterns that one-component fluids cannot produce.

desk verdict A careful model paper that predicts genuinely new double-quench morphologies from slow composition equilibration, but the headline effects rest on a noiseless limit whose physical relevance is not quantified. read the letter →

arxiv 1908.08890 v1 pith:DEHWJVCK submitted 2019-08-23 cond-mat.soft cond-mat.mtrl-sci

Pablo de Castro , Peter Sollich This is my paper · ORCID

classification cond-mat.softcond-mat.mtrl-sci

keywords phaseseparationsecondtemperaturequenchbinarycolloidalmixturescompositionheterogeneitiesfractionationcrowdingspinodaldecompositiondouble

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 predicts how a dense binary colloidal mixture behaves when it is quenched a second time, to a lower temperature, after a first quench has already produced coarse liquid and gas domains. Because the dense liquid is crowded, the mixture's species composition can equilibrate only slowly, so the second quench does not resume normal coarsening. Instead, secondary gas bubbles form inside the liquid, often in regularly spaced rows, and the fractionated interfaces left behind when bubbles shrink survive for very long times as patches of unequal composition. The paper identifies five signatures of this slow fractionation: long-lived regular arrays of secondary domains, wetting of interfaces by oppositely fractionated layers, surface-directed spinodal waves travelling inward from primary interfaces, a dead zone with no secondary phase separation, and filamentous morphologies in three-phase coexistence. All of these disappear when particle-particle swaps are switched on, which is the paper's evidence that crowding-induced slow composition changes, not the second quench itself, produce the effects.

What carries the argument

The central object is the Polydisperse Lattice-Gas (PLG) kinetic theory, a mean-field lattice model in which each colloidal species has its own local density and particles move either by jumping into vacancies or by direct exchange with particles of another species. The exchange rate $w_s$ acts as the control knob: $w_s=0$ forces composition changes to proceed through slow vacancy-mediated moves, producing the crowding slowdown, while $w_s=1$ allows fast composition equilibration. The conceptual engine is the two-stage scenario for dense mixtures: total density equilibrates quickly because particles of any species can move, while species composition equilibrates slowly because it requires particles of different species to swap places; the system therefore first follows a 'quenched' phase diagram and only later relaxes towards the equilibrium 'annealed' phase diagram. The paper isolates slow fractionation by comparing $w_s=0$ and $w_s=1$ under identical quench protocols and geometries. In the slab geometry, the equilibrated noiseless initial state makes the liquid-vapour interface the only source of fluctuations, so spinodal decomposition starts at the interface and propagates inward as a front, which is what produces the regular long-lived arrangement of secondary domains.

What would settle it

Repeat the same two-step quench protocol with thermal noise present throughout the dynamics, for example by adding stochastic forcing to the kinetic equations or by performing off-lattice Brownian dynamics simulations at the same densities, polydispersity, and quench temperatures. If the surface-directed spinodal waves, the long-lived regular arrangement of secondary domains, and the dead zone fail to appear, then the predicted effects depend on the noiseless mean-field evolution rather than on the crowding mechanism itself.

Watch

Extended reading notes

Core claim

The central claim is that after a second, deeper quench, slow composition equilibration caused by crowding creates long-lived, spatially heterogeneous states that are absent when composition changes are fast. For a binary mixture with particle-particle swaps turned off, the paper shows that secondary gas bubbles form inside the dense primary liquid through spinodal waves that start at the interfaces of the primary domains; the bubbles settle into a persistent regular spatial arrangement; the primary domains stop coarsening for an extremely long time; and the interfaces of shrinking secondary bubbles remain as B-rich patches while the surrounding liquid becomes A-rich. In a slab geometry, where an equilibrated liquid slab is surrounded by vapour, the spinodal waves travel inward from the slab-vapour interface and are suppressed if noise is added to the starting state. At higher temperatures an A-rich layer forms next to the interface and a dead zone appears in which density has equilibrated so quickly that composition fluctuations are damped rather than grown. In the three-phase case, secondary bubbles mediate B-rich filaments that wet and connect primary domains and eventually become the third equilibrium phase. Because the same simulations with fast particle-particle exchange lose all these long-lived structures, the paper concludes that slow fractionation is the mechanism.

Load-bearing premise

The predictions rely on the dynamics being deterministic and essentially noiseless after the second quench: in the slab geometry the equilibrated liquid has no thermal fluctuations, so instabilities can start only at the slab-vapour interface, and the paper itself shows that adding noise to the initial state suppresses the spinodal waves and that nucleation and growth, which require noise, could destroy the dead zone.

Editorial extensions

If this is right

A deep second quench should interrupt the coarsening of primary domains for a very long time, with the usual $t^{1/3}$ growth resuming only after all secondary bubbles have disappeared and with a much reduced prefactor.
Long-lived regular arrays of secondary domains should be a generic signature of slow fractionation, and they should be destroyed whenever particles can exchange species quickly.
In a slab geometry, surface-directed spinodal waves should propagate inward from an equilibrated interface, and stronger initial noise should reduce how far they travel before bulk spinodal modes take over.
At higher temperatures, an A-rich layer should form next to the slab-vapour interface and a dead zone should appear where the liquid is too dense to be unstable, because density equilibration beats composition equilibration.
A second quench into a three-phase region should produce B-rich filaments that connect primary domains and eventually become the third phase, with filament formation much slower when particle-particle swaps are disabled.

Reading between the lines

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

Because the mechanism is crowding-based rather than specific to two species, the same second-quench effects should appear in mixtures with more than two components and in off-lattice models; Brownian dynamics of bidisperse attractive colloids would be a direct test.
The dead zone suggests a control strategy: by letting density equilibrate near an interface before the second quench, one could create a compositionally quiet barrier that blocks secondary phase separation, allowing patterned placement of secondary domains.
For weak but nonzero thermal noise, the surface-directed waves should still form but with a finite penetration depth; measuring how the regular stripes extend as a function of noise amplitude would quantify the mechanism.
The three-phase filaments could provide a route to percolating minority-phase networks, since the filaments connect primary domains before they mature into the equilibrium third phase.

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.

Desk Editor's Note

read the letter

First thing you should know: this is a real extension of the authors' polydisperse lattice-gas kinetic theory to double quenches, and it yields several concrete, falsifiable morphological predictions — long-lived regular secondary domains, a dead zone, surface-directed spinodal waves, and three-phase filaments. Prior double-quench work on incompressible symmetric polymer blends did not include slow composition equilibration, so the compressible, crowding-driven effects here are new. The slab-geometry setup is an effective way to isolate the surface-directed waves, and the ws=0 vs ws=1 comparison clearly exposes the role of slow fractionation.

The paper does the core science well. The LSW argument in Section 3.1 gives a reasonable explanation for why secondary domains do not reach asymptotic coarsening, and the quantitative data on domain areas and density profiles back up the visual claims. The citation pattern looks appropriate; the authors distinguish their work from the existing double-quench literature.

Where it gets soft: the most distinctive predictions live in a noiseless deterministic limit. The equilibrated initial states are noise-free, so instabilities start at interfaces. The authors show that added initial noise suppresses the spinodal waves (Fig. 10), and they concede in the Conclusions that nucleation and growth could destroy the dead zone. What is missing is any estimate of the noise amplitude that corresponds to physical thermal fluctuations at the final temperature, or a determination of whether bulk modes would outcompete interface modes. Without that, the practical relevance of the long-lived regularity and dead zone remains illustrative rather than established. That's a significant caveat, but not a fatal one — the paper openly acknowledges it, and the qualitative predictions can still be tested in experiments or in noisy simulations.

A second, moderate weakness is reproducibility: no code or data is provided, and some quantitative statements rest only on 'data not shown' (the long arrested ws=0 behaviour and the primary-domain area for ws=1). For a simulation-based paper that is frustrating and should be fixed in revision.

Who is this for? Soft-matter theorists interested in phase-separation kinetics, polydispersity, and history-dependent structure formation. Experimentalists may find testable predictions. It deserves a serious referee: the model work is careful, the effects are new, and the limitations are honestly stated. The main referee asks should be for a noise-amplitude estimate and for code/data, not for redoing the physics.

Referee Report

3 major / 4 minor

Summary. The manuscript studies the kinetics of phase separation in binary colloidal mixtures after a second, deeper temperature quench, using the mean-field Polydisperse Lattice-Gas (PLG) kinetic theory of ref. 15. The authors report simulation results for quenches into two- and three-phase coexistence regions and identify a set of new effects: long-lived regular arrangements of secondary domains, interruption of primary-domain coarsening, wetting of interfaces by fractionated material, surface-directed spinodal waves emanating from primary slab or bubble interfaces, a region of suppressed phase separation ('dead zone'), and filamentous morphologies in the three-phase case. The results are interpreted via Warren's two-stage fractionation scenario, in which crowding slows composition equilibration ('fractionation') relative to density equilibration. The paper also presents a Lifshitz-Slyozov-Wagner-style argument in Section 3.1 to explain the absence of an asymptotic coarsening regime for secondary domains. The computational evidence is based on deterministic mean-field simulation of a single lattice system, with parameters chosen to contrast slow (ws = 0) and fast (ws = 1) particle-swap kinetics.

Significance. The claimed effects, if robust, would be genuinely new: they have no counterpart in one-component fluids and are not captured by incompressible double-quench models. The paper is clearly written and the simulation snapshots (Figs. 1, 3, 6, 8, 11, 12, 15-17) are suggestive. The quantitative data that are shown (primary-domain area in Fig. 2, secondary-domain area in Fig. 4) support the qualitative story of interrupted coarsening and eventual secondary-bubble reabsorption. The LSW-type argument in Section 3.1 is a useful analytic supplement. The main weakness is that the most distinctive predictions--surface-directed waves, long-lived regularity, and the dead zone--are computed in a deterministic noiseless limit, and the physical noise amplitude is not quantified; the authors themselves acknowledge that nucleation and growth could destroy the dead zone and that stronger initial noise suppresses the waves (Fig. 10). Because these effects are the central new claims, the paper requires additional analysis or explicit discussion of physical noise levels before the predictions can be considered robust.

major comments (3)

[Section 4.1 (Fig. 10) and Section 4.2; Eqs. (3)-(5)] The surface-directed spinodal waves, the long-lived regularity of secondary domains, and the dead zone are obtained from the deterministic mean-field equations (3)-(5) starting from equilibrated, noiseless initial states. The paper states that the equilibrated interfaces are the 'sole source of fluctuations', but in a noiseless deterministic system there are no fluctuations; the interfacial instability must be seeded by uncontrolled numerical round-off. Figure 10 shows that adding initial density noise with standard deviations as small as 0.01%-1% already reduces the distance the spinodal waves propagate, and the conclusions concede that nucleation and growth could destroy the dead zone. The manuscript does not estimate the amplitude of thermal fluctuations at the final quench temperature, nor does it compare the growth of bulk modes seeded by such fluctuations with that of the interface-initiated modes. These predictions are among the headline new effects, so their robustness to physical noise is a load-bearing issue that should be addressed quantitatively, for example by estimating the noise strength from the stochastic dynamics underlying the kinetic model or by running simulations with conserved noise added throughout the dynamics.
[Section 3 (discussion of Figs. 2 and 4)] Two quantitative statements central to the interrupted-coarsening and long-lived-secondary-structure claims are documented only as 'data not shown': the long arrested behaviour of the average secondary area for ws = 0, and the qualitatively similar evolution of the primary domain area for ws = 1. Without these data, the reader cannot verify the plateau extent or the claim that the primary-area behaviour is insensitive to ws. The authors should include the missing panels or make them available as supplementary material, and should state the simulation time window over which the arrest is observed.
[Section 3 (after Fig. 1) and Section 4.1 (Fig. 6)] The paper states that the secondary-domain spacing and the decomposition-front spacing are 'consistent with' the predicted spinodal length, but no quantitative comparison is shown. Since the regular arrangement of secondary domains is a central claimed effect, the authors should provide a quantitative match between the observed wavelength (e.g., from the structure factor or from the spacing of the density-profile oscillations in Fig. 7) and the theoretical value 2pi/kmax from eq. (24) of ref. 15. This would also help separate the surface-directed wave contribution from the bulk spinodal modes.

minor comments (4)

[Section 1] 'it is has been' should read 'it has been'.
[Fig. 1 caption] 'as from numerics' is awkward; consider 'from the numerics'.
[Fig. 10] The snapshots are taken at different times for different noise strengths, so the comparison of wave-penetration depths is indirect; a plot of the wavefront position as a function of time for each noise strength would make the trend quantitative.
[Section 3.1] The LSW analysis in eqs (7)-(8) is presented with the caveat that the underlying assumptions (dilute bubbles, fast chemical-potential equilibration) are not obviously satisfied; it would be helpful to state explicitly that this is a qualitative argument, not a quantitative fit to the simulation data.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the reported effects are simulation outputs of a fixed-parameter kinetic model, not fitted parameters or self-referential reductions.

full rationale

No circular step is identifiable. The kinetic PLG equations (1)-(5) are imported from the authors' earlier work (ref. 15), but that prior derivation is an independent model input: it follows from a stated Hamiltonian, mean-field dynamical equations, and Glauber-like rates, and it does not contain the present paper's target effects as construction inputs. The spinodal criterion quoted from ref. 15 and used in Section 3 is a derived diagnostic for choosing and interpreting quench parameters; the claimed new effects—long-lived regular secondary domains, wetting layers, surface-directed spinodal waves, the dead zone, and three-phase filaments—are read off from simulations with fixed parameters rather than obtained by fitting those parameters to the phenomena. The LSW argument in Section 3.1 is standard and parameter-free. The paper's self-citations to refs. 12, 13, and 15 are continuous-program citations and do not carry the argument by fiat. The genuine weaknesses are evidentiary and robustness limitations, not circularity: the deterministic dynamics put noise only into initial conditions, Fig. 10 shows that stronger initial noise suppresses the spinodal waves, and the Conclusions explicitly concede that 'nucleation and growth' would 'require including noise throughout the dynamics' and could 'effectively destroy the dead zone'. Some quantitative claims are documented only as 'data not shown' (e.g. the long arrested ws=0 behaviour and the primary-domain area for ws=1). These limitations reduce confidence in the physical robustness of the predictions, but none of them makes the derivation equivalent to its own inputs.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

The model parameters (d, T1, T2, pA, pB, ws) are control parameters chosen to access specific regimes; none is fitted to data or tuned to reproduce the observed morphologies. The polydispersity d and attempt rates w0 and ws are inputs from the model, not free parameters of a fit. The axioms listed are the load-bearing modeling assumptions on which the predicted effects rest.

assumptions (6)

domain assumption The dynamics are governed by deterministic mean-field kinetic equations (eqn 3) with Glauber jump rates (eqn 5), ignoring thermal noise and correlations beyond local averages.
This is the core modeling assumption. It allows the simulations but also underlies the surface-directed spinodal waves and the dead zone, which the authors show depend on the absence of bulk noise.
domain assumption The initial equilibrated phases in the slab geometry are noiseless, so fluctuations only arise at the slab-vapour interface or from added initial noise.
Used in Section 4.1 to explain the surface-directed spinodal waves and the long-lived regularity of secondary domains; the authors note that stronger initial noise suppresses the waves (Fig. 10).
domain assumption Warren's scenario: in dense mixtures, density equilibrates fast via particle-vacancy exchanges while composition equilibrates slowly because it requires particle-particle exchanges.
This mechanism, adopted from ref. 16, is the interpretative framework for the slow-fractionation effects and the dead zone. If it did not apply, the ws=0 results would lose their stated meaning.
domain assumption The PLG Hamiltonian (eqn 1) with pairwise energies epsilon_alpha_beta = sigma_alpha sigma_beta and the hard-core constraint (eqn 2) captures compressible colloidal mixtures, with vacancies representing passive solvent.
Central to the model; solvent is passive and hydrodynamic interactions are ignored, as stated in Section 2.
domain assumption The spinodal criterion for the second quench instability uses the annealed spinodal temperature T = z(rho_2 - rho_1^2) evaluated at the equilibrated primary liquid composition (eqn 6).
This maps the second-quench temperature to the unstable region of the PLG model; it comes from eqn (6) of ref. 15 and is used in Section 3 to justify the choice of T2.
domain assumption The qualitative effects observed for a binary mixture are expected to generalize to a broad class of polydisperse systems.
Stated in the Introduction; the generalization claim is not directly tested here and rests on the authors' previous work (ref. 15).

how reviews work

0 comments

Cite this review

Pith. "Pith review of Phase separation of mixtures after a second quench: composition heterogeneities." pith.science (2026). https://pith.science/paper/DEHWJVCK

@misc{pith2026190808890,
  author       = {Pith},
  title        = {Pith review of: Phase separation of mixtures after a second quench: composition heterogeneities},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DEHWJVCK}},
  note         = {Machine review of arXiv:1908.08890}
}

read the original abstract

We investigate binary mixtures undergoing phase separation after a second (deeper) temperature quench into two- and three-phase coexistence regions. The analysis is based on a lattice theory previously developed for gas-liquid separation in generic mixtures. Our previous results, which considered an arbitrary number of species and a single quench, showed that, due to slow changes in composition, dense colloidal mixtures can phase-separate in two stages. Moreover, the denser phase contains long-lived composition heterogeneities that originate as the interfaces of shrunk domains. Here we predict several new effects that arise after a second quench, mostly associated with the extent to which crowding can slow down 'fractionation', i.e. equilibration of compositions. They include long-lived regular arrangements of secondary domains; wetting of fractionated interfaces by oppositely fractionated layers; 'surface'-directed spinodal 'waves' propagating from primary interfaces; a 'dead zone' where no phase separation occurs; and, in the case of three-phase coexistence, filamentous morphologies arising out of secondary domains.

Figures

Figures reproduced from arXiv: 1908.08890 by the authors.

**Figure 10.** The above discussion shows the importance of considering noise in double quench theories, as otherwise the primary phases have no noise and therefore the primary domains would act as the sole sources of fluctuations. The exact noise strength will change with time, depending on temperature and composition, and this will dictate whether spinodal waves will develop or not, and consequently the structure of the set of s… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

14 extracted references · 14 canonical work pages

[1]

Finally , it will also be worth seeing how the effects investigated here manifest themselves in off-lattice models

More speciﬁcally , it would be interesting to look into the dynamical process by which two primary liquid domains become connected. Finally , it will also be worth seeing how the effects investigated here manifest themselves in off-lattice models. Conﬂicts of interest There are no conﬂicts to declare. Acknowledgements PdC acknowledges ﬁnancial support fro...

work page 2014
[4]

The second quench was performed at t = 20000

From top left to bottom right, the snapshots are taken at t = 349, 1992, 20000, 20004, 26177, and 408307. The second quench was performed at t = 20000. Fig. 17 As Fig. 16, but for ws =

work page 1992
[10]

The above discussion shows the importance of considering noise in double quench theories, as otherwise the primary phases have no noise and therefore the primary domains would act as the sole sources of ﬂuctuations. The exact noise strength will change with time, depending on temperature and composition, and this will dictate whether spinodal waves will d...

work page
[15]

It is plotted in triangular colour space in (pA, pB, 1− pA− pB), dropping the site index i. For example, if the concentration of particles of species A at one site is high (low), and the concentration of particles of species B at the same site is low (high), then the site colour will tend towards blue (red); if the concentrations of all species are all lo...

work page 2017
[16]

and (ii) the denser phase con- tains long-lived ‘composition heterogeneities’. The effect is ob- viously impossible in one-component ﬂuids because in that case the composition is the same everywhere (all particles are of the same kind) and only the total local density ﬂuctuates. Firstly we observed that gas bubbles—or more generally domains—can be formed ...

work page Pith review arXiv 2008
[50]

36 D. A. Huse, Physical Review B, 1986, 34, 7845–7850. 37 P. Streitenberger, Acta Materialia, 2013, 61, 5026 –

work page 1986
[128]

The second quench was performed at t =

From top left to bottom right, the snapshots are taken at t = 1500, 1501, 1510, 1597, 1914, and 225702. The second quench was performed at t =

work page 1914
[573]

Evans, D

4 R. Evans, D. Fairhurst and W. Poon, Physical Review Letters , 1998, 81,

work page 1998

Show all 14 references

[807]

Singh, A

18 A. Singh, A. Mukherjee, H. M. Vermeulen, G. T. Barkema and S. Puri, The Journal of Chemical Physics, 2011, 134, 044910. 19 L.-T. Yan and X.-M. Xie, The Journal of Chemical Physics, 2008, 128, 034901. 20 M. Fialkowski and R. Holyst, The Journal of Chemical Physics, 2002, 117...

2011
[1326]

5 N. B. Wilding, P. Sollich and M. Buzzacchi, Physical Review E, 2008, 77, 011501. 6 J. Viéville, M. Tanty and M.-A. Delsuc, Journal of Magnetic Resonance, 2011, 212, 169–173. 7 S. Auer and D. Frenkel, Nature, 2001, 413, 711–713. 8 S. Belli, A. Patti, M. Dijkstra and R. Van Ro...

2008
[1344]

Sollich, Journal of Physics: Condensed Matter , 2002, 14, R79

11 P. Sollich, Journal of Physics: Condensed Matter , 2002, 14, R79. 12 P. de Castro and P. Sollich, The Journal of Chemical Physics , 2018, 149, 204902. 13 P. S. de Castro Melo, Phase separation of polydisperse ﬂuids ,

2002
[2019]

Zhang, Y

14 L. Zhang, Y. Peng, L. Zhang, X. Lei, W. Yao and N. Wang, RSC Advances, 2019, 9, 10670–10678. 15 P. de Castro and P. Sollich, Physical Chemistry Chemical Physics, 2017, 19, 22509–22527. 16 P. B. Warren, Physical Chemistry Chemical Physics , 1999, 1, 2197–2202. 17 T. C. B. Mc...

2019
[2397]

40 K. Tafa, S. Puri and D. Kumar, Physical Review E , 2001, 64, 056139. + P V S O B M / B N F < Z F B S > < W P M > 1–13 | 13

2001
[5035]

Wagner, Zeitschrift für Elektrochemie, Berichte der Bunsen- gesellschaft für physikalische Chemie, 1961, 65, 581–591

38 C. Wagner, Zeitschrift für Elektrochemie, Berichte der Bunsen- gesellschaft für physikalische Chemie, 1961, 65, 581–591. 39 L.-T. Yan and X.-M. Xie, Macromolecules, 2006, 39, 2388–

1961

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

Finally , it will also be worth seeing how the effects investigated here manifest themselves in off-lattice models

More speciﬁcally , it would be interesting to look into the dynamical process by which two primary liquid domains become connected. Finally , it will also be worth seeing how the effects investigated here manifest themselves in off-lattice models. Conﬂicts of interest There are no conﬂicts to declare. Acknowledgements PdC acknowledges ﬁnancial support fro...

work page 2014

[2] [4]

The second quench was performed at t = 20000

From top left to bottom right, the snapshots are taken at t = 349, 1992, 20000, 20004, 26177, and 408307. The second quench was performed at t = 20000. Fig. 17 As Fig. 16, but for ws =

work page 1992

[3] [10]

The above discussion shows the importance of considering noise in double quench theories, as otherwise the primary phases have no noise and therefore the primary domains would act as the sole sources of ﬂuctuations. The exact noise strength will change with time, depending on temperature and composition, and this will dictate whether spinodal waves will d...

work page

[4] [15]

It is plotted in triangular colour space in (pA, pB, 1− pA− pB), dropping the site index i. For example, if the concentration of particles of species A at one site is high (low), and the concentration of particles of species B at the same site is low (high), then the site colour will tend towards blue (red); if the concentrations of all species are all lo...

work page 2017

[5] [16]

and (ii) the denser phase con- tains long-lived ‘composition heterogeneities’. The effect is ob- viously impossible in one-component ﬂuids because in that case the composition is the same everywhere (all particles are of the same kind) and only the total local density ﬂuctuates. Firstly we observed that gas bubbles—or more generally domains—can be formed ...

work page Pith review arXiv 2008

[6] [50]

36 D. A. Huse, Physical Review B, 1986, 34, 7845–7850. 37 P. Streitenberger, Acta Materialia, 2013, 61, 5026 –

work page 1986

[7] [128]

The second quench was performed at t =

From top left to bottom right, the snapshots are taken at t = 1500, 1501, 1510, 1597, 1914, and 225702. The second quench was performed at t =

work page 1914

[8] [573]

Evans, D

4 R. Evans, D. Fairhurst and W. Poon, Physical Review Letters , 1998, 81,

work page 1998