1: \documentclass[aps,amssymb,twocolumn]{revtex4}

2: %\input epsf

3: \usepackage{graphics}

4: 

5: \def\epsfig{\special}

6: \def\D {{\rm D}}

7: \def\d {{\rm d}}

8: \def\vector#1{{{\underline {#1}}}}

9: \def\avector#1{{\overline{\underline {#1}}}}

10: \def\tensor#1{{\underline{\underline {#1}}}}

11: \def\atensor#1{{\overline{\underline{\underline {#1}}}}}

12: \def\div {{\rm div}}

13: \def\tr{{\rm tr}}

14: \def\dd{{\xi}}

15: \def\ss{{\sigma}}

16: \def\dpl{\gamma}

17: \def\ddpl{\dot\dpl}

18: \def\tdpl{\tensor\epsilon^{pl}}

19: \def\dtdpl{\dot\tensor\epsilon^{pl}}

20: \def\dtd{\dot\tensor\epsilon}

21: \def\din{\epsilon^{in}}

22: \def\tdin{\tensor\epsilon^{in}}

23: \def\del{\epsilon^{el}}

24: \def\tdel{\tensor\epsilon^{el}}

25: \def\sdef{\tau}

26: \def\sigp{\zeta}

27: 

28: \begin{document}

29: %\bibliographystyle{prsty}

30: \title {A dynamical approach to glassy materials}

31: \author{Ana\"el Lema\^{\i}tre}

32: \affiliation{Department of physics, University of California, Santa Barbara, 

33: California 93106, U.S.A.}

34: \affiliation{CEA --- Service de Physique de l'\'Etat Condens\'e,

35: Centre d'\'Etudes de Saclay, 91191 Gif-sur-Yvette, France}

36: \date{\today}

37: \begin{abstract}

38: Typical properties of glassy materials are shown to be captured by a mean-field

39: free-volume theory. Relaxation processes are supposed to be free-volume activated,

40: and different entropy barriers are associated with density relaxation

41: and shear motion.

42: Free-volume time logarithmic relaxation, Kohlrausch-Williams-Watts,

43: and power law viscosity result from

44: the non-linear dynamics of spatially averaged quantities.

45: The exponents associated with these phenomena are related to a single parameter

46: of the theory.

47: The theory also accounts for coexistence of jamming transitions and non-linear rheology.

48: %The study in complemented with the analysis of loss and storage moduli in the linear

49: %and non-linear regime.

50: \end{abstract}

51: \pacs{05.90.+m,83.10.Gr,83.60.-a,62.20.-x}

52: 

53: 

54: \maketitle

55: 

56: %%

57: %%

58: 

59: \section{Introduction}

60: 

61: %(Recent characterization  of aging have been on PMMA 

62: %by low frequency dielectric measurements~\cite{bellon00}

63: %and on microgel pastes~\cite{cloitre00}

64: %Also in the context of solid friction~\cite{bureau02})

65: 

66: The emergence of slow modes of relaxations is a major signature 

67: of the glassy behavior.~\cite{goetze92}

68: The first consequence of large equilibration time is that the properties 

69: of a glass evolve on experimental timescales, the system ages.

70: Despite the huge amount of work that has been devoted to glassy systems,

71: a general understanding of slow relaxation processes and of their

72: relation to macroscopic rheology is still lacking.

73: For example, experimental and numerical data converge to a stretched exponential 

74: shape of relaxation phenomena, the so-called Kohlrausch-Williams-Watts

75: (KWW) relaxation,~\cite{ediger96,larson99,angell00}

76: while mode coupling theory predicts power law behavior.~\cite{goetze92}

77: When it comes to extend these results to non-linear rheology, 

78: the situation is even more difficult from the theoretical point of view,

79: since the system is driven far from equilibrium.

80: At an applied strain rate, power law viscosities or 

81: the emergence of a yield stress, are common features 

82: of complex fluids.~\cite{larson99} 

83: %and can be roughly summarized as the so-called

84: %Herschel-Bulkeley relation $\sigma=\sigma_y+\dot\epsilon^n$ between stress

85: %and strain rate.

86: MD simulations of Lennard-Jones systems~\cite{berthier01,yamamoto97,yamamoto98}

87: and recent experiments on colloidal glasses~\cite{bonn02a,bonn02b}

88: are also consistent with such properties.

89: 

90: Recent theoretical approaches to non-linear

91: rheology~\cite{sollich97,sollich98,hebraux98,berthier00,fielding00} 

92: have called on ideas originating in mode coupling theory, 

93: or trap models of spin glasses.~\cite{goetze92,bouchaud96,monthus96}

94: These approaches, however, fail to provide a description of non-linear

95: rheology that is consistent with experimental observations;~\cite{bonn02a,bonn02b}

96: moreover, these models, based only on a phase-space picture of glassiness, 

97: do not lead to a clear identification of real-space physical mechanism 

98: that are responsible for an observed macroscopic behavior.

99: Some other attempts have been directed towards phenomenological 

100: approaches,~\cite{derec01} but those are still unsatisfactory, 

101: and do not offer a microscopic picture, even heuristic, of the undergoing physical 

102: mechanisms.

103: 

104: The current work lies at the convergence point of several recent theoretical

105: ideas in various fields.

106: Firstly, it is based on the idea by Liu, Nagel and coworkers, 

107: that some unique mechanism lies 

108: behind the profound similarities displayed by structural glassy materials,

109: be they colloidal glasses, granular materials, foams,\dots~\cite{liu98}

110: %The observed similarity of the force distribution between various structural systems 

111: %indicates that the structure of a material at low temperature is essentially 

112: %dominated by large forces between molecules: relaxation processes are 

113: %therefore expected to be controlled by the force network which acts 

114: %as a mechanical skeleton.~\cite{ohern01}

115: Drawing on this idea, the fundemental assumption underlying the

116: current work is that macroscopic 

117: properties of dense materials results from major {\it dynamical} properties 

118: of structural rearrangement of the contact network.

119: 

120: Another important element of the current work comes from the study 

121: of density relaxation 

122: in granular materials.~\cite{knight95,boutreux97,nowak98}

123: Under vertical tapping, granular materials present a time-logarithmic relaxation 

124: of the occupied volume; this is in strong analogy with free-volume relaxation, 

125: commonly observed in glasses, and of central importance for aging.~\cite{larson99}

126: However, in the case of tapped granular materials, 

127: density relaxation cannot result from thermal activation

128: since there is no thermal bath:

129: the logarithmic density relaxation has been

130: explained by introducing an equation of motion for free-volume,

131: which characterizes the internal state of the granular material.

132: 

133: Finally, this work uses a theory introduced recently by Falk and Langer

134: to account for elasto-plastic transition in amorphous solids.~\cite{falk98,falk00} 

135: The so-called shear transformation zone (STZ) theory, 

136: originates from ideas by Argon, Spaepen and others

137: to describe creep in metallic

138: alloys;~\cite{spaepen77,argon79a,argon79b,spaepen81,argon83}

139: it has been primarily developed in the framework of fracture mechanics,

140: and provides a general scheme for jamming as resulting from a structuration 

141: of the material by creation of arrangements oriented along principal directions

142: of an applied stress. 

143: STZ theory calls to free-volume arguments, but in fact,

144: free-volume remains a parameter and enters constants of the theory.

145: 

146: These ideas lead me to propose a general approach to the rheology of

147: dense materials, leading to a simple set of constitutive equations.~\cite{lemaitre01b}

148: The purpose of the this article is to  elaborate the theoretical arguments 

149: underlying these equations a to detail some of their consequences.

150: 

151: Two types of internal state variables are used: free-volume, 

152: which is an isotropic property of a molecular packing, and populations of arrangements, 

153: which are related to the distortion of the contact network between molecules  

154: (hence introduce anisotropy).

155: Those variables evolves dynamically by free-volume activated processes:

156: two types of rearrangements are associated to compaction and shear motion.

157: Free-volume enters activation factors that control both types 

158: of elementary rearrangements, and it is also a dynamical quantity. 

159: 

160: KWW relaxation has been shown to result from 

161: the existence of a distribution of timescales;~\cite{larson99,angell00}

162: and it seems that this is often thought to be a necessary requirement.

163: The first contribution of the current work is to show that it is not necessary,

164: and that KWW relaxation also result from a more simple mechanism:

165: %shear motion and free-volume relaxation result from different types 

166: %of collective rearrangements, hence are activated processes with 

167: %different entropy barrier heights;

168: %KWW results directly from 

169: the interplay between free-volume logarithmic relaxation and

170: free-volume activated shear deformation.

171: 

172: The second contribution of this work is to establish a relation 

173: between KWW relaxation, time-logarithmic density relaxation, 

174: and power law viscosity of a strongly sheared glassy material.

175: A single parameter of the theory determines these three types of phenomena.

176: Power law viscosity will be shown to result from 

177: the existence of shear induced dilatancy. 

178: I propose one of the simplest set of equations that can account 

179: for those three properties is a single framework.

180: 

181: The third contribution of this work is to present a theoretical framework

182: which accounts for two coexisting mechanisms for jamming.

183: In mode-coupling theory, jamming results from an entropy crisis:

184: at low temperature, the system supports any applied stress because 

185: no motion of molecules is allowed.

186: Resistance to shear is therefore an intrinsic property of a material,

187: determined by the thermodynamic parameters of the system,

188: and results from a phase transition.

189: In STZ theory, jamming results from a structuration 

190: of the contact network:

191: jamming requires some amount of plastic deformation

192: to induce the creation of anisotropic structures that 

193: can support a static stress.

194: Those two mechanisms are shown here to be incorporated in a unique theory. 

195: The difference between hard and soft glassy materials

196: will be identified with the cases when one or the other process dominates.

197: The interplay between those two mechanisms

198: lead to a novel interpretation of brittleness and ductility,

199: and provide age-dependent dynamical yield criteria.

200: 

201: \section{Constitutive equations}

202: \subsection{Preliminary settings}

203: %\subsubsection{A mean-field approach}

204: 

205: The main assumption made in this work is that the out-of-equilibrium

206: dynamics of a glass can be modeled in the framework of a mean-field approximation.

207: By mean-field, I mean that spatially averaged quantities account 

208: for the local state of the material. 

209: Constitutive equations are proposed that relate the average shear stress,

210: to the average shear strain.

211: The internal state of the material is determined by free-volume

212: and by densities of arrangements (to be defined further).

213: %Since the distribution of voids in the material

214: %is characterized by the average free-volume only, 

215: %an additional Ansatz will be required for this distribution.

216: 

217: %\subsubsection{Stress, strains and inherent states}

218: 

219: The forces in the material are supposed to be characterized by a stress tensor, 

220: which is written

221: $$

222: \left(

223: \matrix{ -P& \sigma\cr

224: \sigma &-P\cr}

225: \right)

226: $$

227: where $P$ is the pressure, and $\sigma$ the shear stress.

228: The deformation tensor is defined accordingly, and composed of shear,

229: and of isotropic deformations, directly related to dilatancy.

230: The average free-volume per molecule is denoted $v_f$,

231: and is related to the total volume $V$ by, 

232: $$

233: V=N v_f + N v^{\rm rcp}

234: \quad,

235: $$

236: where $v^{\rm rcp}$ is the volume per molecule of the material in random close packing,

237: and $N$ is the number of molecules.

238: 

239: Under strong shear, soft glassy materials, like clays, foams, or granular materials, 

240: can achieve large deformations; the evolution of the system

241: is more similar to the flow of a liquid than to the deformation of a solid.

242: In this work I present a theory of deformation that is expected to hold

243: for large deformations of amorphous solids or soft glassy materials.

244: Since there is a long-standing matter of debate in the mechanical community

245: about linear versus non-linear deformations,~\cite{lubliner90}

246: it seems necessary to clarify this issue.

247: I show here that the notion of inherent states~\cite{stillinger84} 

248: offers a simple physical interpretation of small and large deformations.

249: 

250: At low temperature, a supercooled liquid or a glassy material evolves in a complex

251: energy landscape which can be partitioned in domains of influence of local minima.

252: For almost all geometrical configuration of the molecules, a unique inherent state 

253: is defined as the local minimum to which the systems relaxes

254: if suddenly quenched to zero temperature,~\cite{stillinger84} 

255: and in the absence of forcing.

256: When a material is forced, a constant stress or strain pulls the system

257: through the energy landscape.

258: Each geometrical configuration reached lies in the domain associated with

259: an instantaneous inherent state.

260: The time series of instantaneous inherent states provides a coarse-grained

261: description of the deformation of the material, and is taken here as the definition

262: of irreversible, plastic, deformation.

263: 

264: The instantaneous inherent state is a natural reference state for 

265: a given molecular configuration, and deviation from the instantaneous inherent state 

266: defines an {\it inherent (shear) deformation} $\epsilon^{\rm in}$.

267: Starting from a known initial configuration, the total shear deformation, 

268: $\epsilon= \epsilon^{\rm in} + \gamma$, results both from inherent deformation, 

269: and from the {\it flow} of inherent states in the phase space.

270: %, determined by the rate of plastic deformation $\dot\gamma$.

271: By definition, the deformation $\gamma$ is measured between two inherent states:

272: it is not a state variable, since it is defined only from the knowledge 

273: of a configuration designed as ``initial''. $\epsilon^{\rm in}$, however, is a state

274: variable, since it is well defined for any instantaneous configuration of the system.

275: Moreover, $\epsilon^{\rm in}$ is expected to be small, since it

276: measures the deformation between two nearby configurations in phase space;

277: elasticity results from a harmonic approximation for every local minimum:

278: it permits to write the stress as given by a Hooke law, 

279: $\sigma=\mu\,\epsilon^{\rm in}$. 

280: The equation that governs $\sigma$ can finally be written:

281: \begin{equation}

282: \dot\sigma = \mu\,(\dot\epsilon-\dot\gamma)

283: \quad.

284: \label{eqn:sigma:0}

285: \end{equation}

286: The rate of plastic deformation $\dot\gamma$ will come out of a statistical analysis

287: of hopping motion in the phase-space. 

288: It will be written as a function of the state variables of the system, 

289: $\sigma$, $v_f$, an others, to define the constitutive law of the material.

290: 

291: \subsection{Assumptions}

292: \subsubsection{Free-volume}

293: 

294: This work relies on the idea that, in dense materials,

295: macroscopic deformation results from free-volume activated rearrangements.

296: This is expected because temperature is so low that thermal activation 

297: becomes irrelevant, and activated processes are controlled by entropic fluctuations

298: of the free-volume.

299: A rearrangements occurs if sufficient space exists in the neighborhood

300: of a given point; the transition probabilities are thus

301: determined by the size distribution of voids.

302: In order to work at a mean-field level, the volume distribution is characterized

303: by the single scalar value $v_f$; in order to write activation factors,

304: it is necessary to provide an Ansatz for the distribution of voids in the material.

305: 

306: In the original works by Cohen and Turnbull on free-volume theory,

307: it was argued that the voids in a material are given by 

308: a Poisson distribution.~\cite{cohen59,turnbull61,turnbull70}

309: In their picture, the motion of a molecule in a glassy material is permitted only 

310: if it can move into a hole of size, say $v_1$.

311: The probability to find a hole larger than $v_1$, is proportional to $\exp(-v_1/v_f)$:

312: this leads to estimate a diffusion constant $D\propto\exp(-v_1/v_f)$,

313: which is a basis of free-volume approaches. 

314: Important differences, however, exist between various 

315: uses and interpretations of free-volume theory.

316: 

317: Falk and Langer use free-volume activation factors in the definition 

318: of STZ theory.~\cite{falk00}

319: Their argument borrowed from original ideas by Edwards and coworkers

320: for granular materials:~\cite{edwards89,mehta89,edwards94} 

321: the only extensive variable is supposed to be the volume $V$;

322: the number of states available to the system

323: is roughly proportional to $(v_f/h)^N$, where $h$ is an arbitrary constant

324: with dimension of a volume; the entropy is defined as,

325: $$

326: S(V,N) \simeq N\,\ln\left({v_f\over h}\right) = N\,\ln\left({V- N v^{\rm rcp}\over N h}\right)

327: $$

328: and the intensive quantity, analogous to temperature, is $\chi$:

329: $$

330: {1\over\chi}\equiv {\partial S\over\partial V} = {1\over v_f}

331: \quad.

332: $$

333: Free-volume thus enters activation factors of the form $\exp(-v_1/v_f)$.

334: 

335: In STZ theory, this discussion turns out to be somewhat formal

336: since free-volume is taken constant:

337: $v_f$ is absorbed in the phenomenological constants of the resulting equations.

338: It also leads to complicated expressions for transition rates as functions

339: of the applied stress (because $v_1$ is supposed to be a function of $\sigma$).

340: %At this stage, however, there is no profound reason to argue of the precise 

341: %analytical expression given for those rates and I expect the physics of dense

342: %materials to come out from rather general features.

343: 

344: However, free-volume should vary when the material dilates or contracts.

345: This idea has been considered recently by several authors in the analysis 

346: of density relaxation of granular materials subjected 

347: to vertical tapping.~\cite{knight95,boutreux97,nowak98}

348: In these works, free-volume is understood as a purely dynamical quantity,

349: and relaxes by the motion of single grains in holes created by volume fluctuations.

350: A Poisson distribution of volume fluctuations is assumed at all times:

351: this leads to an equation of motion of the form,

352: \begin{equation}

353: \dot v_f = -R_1\,\exp\left[{-{v_1\over v_f}}\right]

354: \quad.

355: \label{eqn:vf:0}

356: \end{equation}

357: The activation factor $\exp\left[{-{v_1/v_f}}\right]$ is the probability that

358: a hole of size larger than $v_1$ is present at a given point in the material;

359: $R_1$ is determined by the tapping frequency.

360: 

361: The current work borrows from those different lines of thought.

362: Drawing on the works initiated on granular compaction, free-volume is a dynamical 

363: quantity; voids are expected to redistribute fast enough in the material;

364: fast means, faster than free-volume activated collective rearrangements themselves.

365: This is justified because redistribution of free-volume requires very tiny 

366: displacements of the molecules, and voids, considered as 

367: particles, are expected to diffuse faster than molecules themselves:

368: around a void, there is by definition an excess of free-volume.

369: 

370: There is an important difference, however, from the picture used 

371: in~\cite{knight95,boutreux97,nowak98}

372: where compaction results from the motion of a single grain in a hole.

373: Here, borrowing on the argument of STZ theory,

374: compaction results from elementary rearrangements, involving several

375: molecules at a mesoscopic scale. 

376: Probabilities of transitions are estimated by entropic arguments.

377: Due to the redistribution of voids, the number of states available 

378: to the system is supposed to be proportional to $(v_f/h)^N$ at all times:

379: the fast redistribution of voids allows the system to 

380: realize a quasi-equilibrium thermodynamic ensemble, 

381: determined by an intensive variable, $v_f$.

382: 

383: The existence of an intensive variable, associated to 

384: the fluctuations of molecular configurations, is supported

385: by the observation of an effective fluctuation-dissipation theorem

386: in sheared fluids.~\cite{berthier01}

387: For reasons of clarity, and to be specific, 

388: I reserve the word temperature for the intensive quantity associated

389: with the fluctuations of energy, or for the specific kinetic energy;%~\cite{langer_priv}

390: therefore, I will not comply with to the recent use of ``effective temperature''

391: for something that is neither thermodynamic nor granular temperature.

392: Nevertheless, the observation of an effective fluctuation-dissipation theorem

393: indicates that such an intensive quantity exists, without identifying 

394: what this quantity actually is.

395: Here, it is identified to $v_f$ and is associated with entropic fluctuations of 

396: molecular configurations at a mesoscopic scale.

397: 

398: The important novelty introduced in the current work it that 

399: $v_f$ enters activation factors and is also a dynamical quantity.

400: This is in weak analogy with granular temperature, which is a temperature, 

401: and evolves dynamically.

402: But granular temperature couples with properties of the material only as far as

403: it determines a collision frequency,~\cite{savage79,lemaitre01a}

404: while free-volume enters exponential prefactors in transformation rates.

405: Another important element of the current theory is that the dynamics of 

406: free-volume results both from density relaxation by activated rearrangements 

407: -- which exists in the absence of forcing --

408: and from dilatancy induced by the macroscopic shear flow 

409: -- which occurs when the system is driven out-of-equilibrium.

410: The non-linear coupling with the mean-flow will be addressed later;

411: I will first continue the description of rearrangement processes.

412: 

413: \subsubsection{Two types of activated processes}

414: The typical elementary compaction process,

415: corresponds to the system going through a saddle point in phase space.

416: The rate of such transformations is controlled by free-volume fluctuations 

417: and this saddle is characterized by a typical activation volume $v_1$,

418: which depends on the shape of the molecules, 

419: and on the details of microscopic interactions.

420: The activated hopping of the system through this type of saddle  

421: permits density relaxation, 

422: and leads to an equation of motion for $v_f$ of the form~(\ref{eqn:vf:0}).

423: 

424: Other types of elementary rearrangements occur in the material, 

425: in particular leading to elementary shear motion.

426: Before presenting a more elaborate scheme for macroscopic shear deformation,

427: let me use a very minimal argument:

428: shear strain is supposed to be proportional to shear stress: $\dot\gamma = D\sigma$;

429: where $D$ is proportional to a free-volume activation factor.

430: 

431: The current approach relies on the remark that the relative motion of molecules 

432: that permit compaction and shear are different;

433: different types of collective rearrangements,

434: should require different activation volumes.

435: In phase space, shear and compaction correspond to two types of saddle points,

436: and those saddle points are entropic barriers of different heights.

437: The activation volume for an elementary shear is denoted $v_0$, 

438: and {\it a priori} differs from $v_1$. This leads to

439: \begin{equation}

440: \dot\gamma \propto \exp\left[{-{v_0\over v_f}}\right]\,\sigma

441: \quad.

442: \label{eqn:dotgamma:0}

443: \end{equation}

444: 

445: One major purpose of this work is to study the consequences of

446: introducing complexity, not through a non-trivial distribution of 

447: timescales,~\cite{angell00,schlesinger84}

448: but through the existence of two types of entropy barriers of unequal heights.

449: The set of equations~(\ref{eqn:sigma:0}),~(\ref{eqn:vf:0}), and~(\ref{eqn:dotgamma:0}) 

450: will be shown to account for logarithmic relaxation of $v_f$ 

451: and for a KWW relaxation of the stress after a strain increment.

452: 	

453: 

454: \subsubsection{Shear deformation}

455: 

456: \paragraph{Shear transformation zones.}

457: Equation~(\ref{eqn:dotgamma:0}) accounts for a free-volume dependent viscosity:

458: the material described is a liquid (in fact a non-Newtonian liquid).

459: It is however, a very rough description of shear motion.

460: 

461: The theoretical description of shear deformation that I present

462: is directly inspired by STZ theory.~\cite{falk98,falk00}

463: The first assumption on which STZ theory is based is that 

464: elementary shear deformation does not occur just anywhere in the material, 

465: but in zones, where a rearrangement is permitted by the local configuration of 

466: molecules.

467: A shear transformation zone (STZ) is thus defined as a locus within a material,

468: where an elementary shear is possible.

469: Another essential assumption is that STZ's are two-state systems. 

470: This can be understood as follows:

471: An elementary arrangement corresponds to the opening and closing of some 

472: contacts between molecules; once such a rearrangement has occurred 

473: somewhere in the material, these molecules cannot shear further in the same direction, but they can shear backwards.

474: A symmetry is thus introduced by shearing motion, and (at least) two types 

475: of arrangements can be identified that are transformed into one another,

476: as pictured here:

477: \begin{center}

478: \unitlength = 0.0011\textwidth

479: \begin{picture}(100,60)(0,0)

480: \put(50,40){\makebox(0,0){\large$R_+$}}

481: \put(50,0){\makebox(0,0){\large$R_-$}}

482: \put(-50,0){\resizebox{200\unitlength}{!}{\includegraphics{stz.ps}}}

483: \end{picture}

484: \end{center}

485: Note that an STZ is not supposed to contain only four molecules, 

486: but it is a mesoscopic object that accounts for the structure of several 

487: neighboring molecules (say of order 10).

488: To simplify the analysis, only one pair of types of arrangements is considered,

489: aligned along the principal directions of the stress tensor.

490: Macroscopic motion results from the statistics of elementary rearrangements,

491: and reads:

492: \begin{equation}

493: \dot\gamma = {\cal A}_0\,(R_+\,n_+-R_-\,n_-)

494: \label{eqn:stz:0}

495: \end{equation}

496: where $n_\pm$ denote the number density of arrangements, and where

497: $R_\pm$ denote the rate of transition $\pm\to\mp$.

498: 

499: Those rates result from activation processes determined by free-volume, 

500: but also by force fluctuations. 

501: A STZ shears if there is sufficient free-volume at its location,

502: and if the local bias of the force network triggers the shear in the appropriate 

503: direction.

504: In this work, force and volume fluctuations are supposed to be uncorrelated;

505: the probabilities associated with those two types of fluctuations factorize:

506: $R_\pm = R^v(v_f) R_\pm^\sigma(\sigma)$.

507: From the preceding discussion, $R^v(v_f) = \exp(-v_0/v_f)$.

508: In order to estimate the factors $R_\pm^\sigma$ associated with the fluctuations

509: of the force network, let me first remark that $R_\pm^\sigma(\sigma)$ is a positive, 

510: increasing function of $\sigma$.

511: Moreover, an elementary shear is expected to be triggered by {\em large forces}.

512: The distributions of static forces in glassy systems have been shown to 

513: share strong similarities with the force distribution in

514: a granular material,~\cite{ohern01} in which case, large forces are 

515: distributed exponentially.~\cite{coppersmith96,radjai98a,radjai98b}

516: This justifies an exponential dependency of $R_\pm^\sigma$ on $\sigma$:

517: $$

518: R_\pm = R_0\,\exp\left[{-{v_0\over v_f}}\right]\,\exp\left[{\pm{\sigma\over\bar\mu}}\right]

519: \quad,

520: $$

521: with $R_0$, the update frequency of microscopic processes.

522: Note however, that the exponential form for the activation factors $R_\pm^\sigma$

523: plays no role in the following, since these factors will be linearized.

524: The parameter $\bar\mu$ measures the typical stress that must be overcome 

525: to trigger a rearrangement. 

526: In hard-sphere systems, it is expected to be proportional

527: to the pressure $P$. In general, $\bar\mu$ is essentially the energy

528: required to break a bond between two molecules, times the density of bonds. 

529: Here, $\bar\mu$ will be taken constant, either because $P$ fixed, 

530: or because it depends on some interaction potential.

531: At some point, $\bar\mu$ will be set to unity, which fixes the unit of forces.

532: 

533: \paragraph{Dynamics of arrangements.}

534: 

535: Following~\cite{falk98,falk00}, the equation of motion for the populations $n_\pm$ is written,

536: \begin{equation}

537: \label{eqn:stz:npm}

538: \dot n_\pm = R_\mp n_\mp - R_\pm n_\pm +\sigma\,\ddpl\,({\cal A}_{c} - {\cal A}_{a}\,n_\pm)

539: \quad.

540: \end{equation}

541: The first two terms of the rhs account for the transitions between both

542: types of STZ. The last term accounts for the renewal of 

543: molecular configuration by the macroscopic flow:

544: STZ constantly appear and disappear during plastic deformation.

545: The renewal of configurations is supposed to be proportional to 

546: the work of plastic deformation, $\sigma\,\ddpl$, which 

547: is thus a common factor of creation and annhilitation terms.

548: 

549: \paragraph{Shear-induced dilatancy.}

550: In this work, the dynamics of rearrangements is coupled to free-volume dynamics.

551: Part of this coupling results from the factor $R^v(v_f)$ that controls

552: shear transformation, but the macroscopic flow is also expected

553: to influence the relaxation of free-volume.

554: 

555: Free-volume dynamics is derived in analogy with the dynamics 

556: of the populations $n_\pm$:

557: plastic deformations corresponds to the evolution of the system in the phase-space

558: along trajectories that do not necessarily permit a minimization of free-volume.

559: In real space, plastic deformation requires the constant renewal of local

560: arrangements, and new arrangements are local configurations of large entropy;

561: by definition, their density is not optimized. 

562: Therefore, the renewal of configurations by the shear flow leads to dilatancy.

563: In experiments on suspensions, for example, the system is fluidized by 

564: applying a strong shear which rejuvenates the glassy structure.~\cite{cloitre00}

565: 

566: In order to quantify the average dilatancy induced by the shear flow,

567: I assume that a fraction of the work of plastic deformation, $\sigma\,\dot\gamma$

568: in used for enthalpy production, $P \dot v_f$.

569: The equation for $v_f$ is:

570: \begin{equation}

571: \dot v_f = - R_1\,\exp\left[-{v_1\over v_f}\right] + {{\cal A}_v\over P}\,\sigma\,\dot\gamma

572: \label{eqn:vf:1}

573: \end{equation}

574: This equation is an essential element of the current theory. 

575: The non-linear coupling with the macroscopic flow shows up when the system 

576: is strongly driven out-of-equilibrium.

577: The competition between self-relaxation

578: and shear induced dilatancy

579: will be shown to lead to a power law viscosity.

580: 

581: \subsection{Equations of motion}

582: 

583: From~(\ref{eqn:stz:npm}), the equation of motion for the total density, 

584: $n_T=n_++n_-$, of arrangements reads,

585: $$

586: \dot n_T = \sigma\,\ddpl\,( 2\,{\cal A}_{c} - {\cal A}_{a}\,n_T)

587: \quad.

588: $$

589: The creation and destruction terms account for an entropic mixing 

590: of molecular configurations. 

591: The previous equation shows that this mixing leads to a convergence of

592: $n_T$ to an asymptotic value, $n_\infty=2\,{\cal A}_{c}/{\cal A}_{a}$;

593: therefore, $n_T=n_\infty$ defines a state of maximal entropy 

594: of molecular configurations.

595: For this reason, I do not expect $n_T$ to be different from $n_\infty$ unless

596: in very special circumstances, {\it e.g.} if some level of crystallization

597: exists in the material.

598: In particular, an amorphous solid is usually produced by a rapid quench

599: of a liquid at low temperature: the system comes from a state of high entropy,

600: therefore, in the resulting glass, $n_T$ is expected to achieve its asymptotic value.

601: In this work, I assume that $n_T=n_\infty$ at all times.

602: The dynamics of local arrangements is thus determined only by the bias $n_--n_+$

603: between the populations of STZ.

604: 

605: Equations~(\ref{eqn:stz:0}) and~(\ref{eqn:stz:npm})

606: are written in a more suitable form by introducing the variable

607: $$

608: \Delta = {n_--n_+\over n_\infty}

609: \quad,

610: $$

611: and parameters,

612: $\epsilon_0 = {{\cal A}_0\,{\cal A}_c/{\cal A}_a}$,

613: $\mu_0 = {\cal A}_0\,{\cal A}_c$, and $E_0 = 2\epsilon_0\,R_0$.

614: It is also convenient to introduce $\kappa=v_1/v_0$, $E_1=R_1/v_0$

615: and $\alpha = {\cal A}_v/(P v_0)$, and to rescale free-volume as,

616: $$

617: \chi = {v_f\over v_0}

618: \quad.

619: $$

620: 

621: The complete set of constitutive equations is finally:

622: \begin{eqnarray}

623: \label{eqn:stzdil:1}

624: \ddpl&=&E_0\,\exp\left[{-{1\over\chi}}\right]\,

625: \left(\sinh\left({\sigma\over\bar\mu}\right)-\Delta\,\cosh\left({\sigma\over\bar\mu}\right)\right)\\

626: \label{eqn:stzdil:2}

627: \dot\Delta&=&{\ddpl\over\epsilon_0}\,\left(1- \mu_0\,\sigma\,\Delta\right)\\

628: \label{eqn:stzdil:3}

629: \dot\chi &=& -E_1 \exp\left[-{\kappa\over\chi}\right] + \alpha\,\sigma\,\ddpl

630: \quad.

631: \end{eqnarray}

632: Those equations involve two state variables, $\Delta$ and $\chi$ which

633: play very different roles.

634: $\Delta$ evolves only when the material is sheared, therefore accounts

635: for some sort of structural memory induced by deformation imposed on the material.

636: On the contrary, $\chi$ evolves spontaneously, and relaxes towards 0.

637: Since this relaxation is slow, $\chi$  also encodes some sort of memory, 

638: but a very different type of memory than $\Delta$, 

639: a memory which constantly evolves as the system ages.

640: 

641: Note that the asymptotic value for the relaxation of $\chi$ (which here is 0)

642: may, in fact, be temperature dependent;

643: this question is eluded in the current work:

644: it seems a very reasonable assumption for materials like suspensions 

645: if not for glasses below the glass transition, because the logarithmic

646: slow-down of density relaxation around $\kappa$ prevents it to reach 

647: asymptotic values far below ($\chi_\infty<<\kappa$).

648: The parameters involved in the theory should depend on pressure and temperature,

649: and this dependency results from particularities of a given material;

650: for example the update frequencies $R_1$ and $R_0$ are expected to 

651: be proportional to $\sqrt{T}$ (where $T$ is the temperature)

652: in a dense hard-sphere material.~\cite{lemaitre01a}

653: Such questions, however, are not the focus of this study;

654: in this article, the values of the parameters are constant and can 

655: be understood as a given thermodynamical state of the system. 

656: 

657: \section{Interlude}

658: Equations~(\ref{eqn:stzdil:1}-\ref{eqn:stzdil:3}) present non-linear couplings

659: between shear motion and density relaxation.

660: Before studying the complete set of equations in various tests,

661: I start with two simple cases, when one or the other process dominates.

662: I show here that 

663: jamming transitions are captured by equations~(\ref{eqn:stzdil:1}-\ref{eqn:stzdil:2}),

664: and free-volume relaxation by equation~(\ref{eqn:stzdil:3}).

665: 

666: \subsection{Free-volume relaxation} 

667: In the absence of shear stress, $\sigma=0$ and $\dot\gamma=0$;

668: the system~(\ref{eqn:stzdil:1}-\ref{eqn:stzdil:3}) reduces to the single equation:

669: $$

670: \dot \chi = -E_1 \exp\left[-{\kappa\over\chi}\right]

671: \quad,

672: $$

673: which accounts for density relaxation.

674: 

675: At short time, this process is dominated by the initial value, $\chi_0$:

676: $$

677: \dot \chi \simeq -E_1 \exp\left[-{\kappa\over\chi_0}\right]

678: \quad;

679: $$

680: this defines a time scale for short-time free-volume relaxation:

681: $$

682: \tau_\chi = {\chi_0\over E_1}\;\exp\left[{\kappa\over\chi_0}\right]

683: \quad.

684: $$

685: 

686: The long time behavior is estimated from the integral expression:

687: $$

688: E_1 t = \int_{\chi(t)}^{\chi(0)} \;\exp\left[{\kappa \over \chi}\right] \d \chi \simeq 

689: \kappa\,\exp\left[{\kappa\over \chi(t)}\right]

690: $$

691: At long time, the integral is dominated by the small values of $\chi$, 

692: {\it i.e.} by the ultimate value reached, $\chi(t)$. Therefore,

693: \begin{equation}

694: \chi(t) \simeq {\kappa\over \log(E_1 t/\kappa)}

695: \quad.

696: \label{eqn:vf:relax}

697: \end{equation}

698: Free-volume relaxes logarithmically with time:

699: in the absence of any forcing, the system ages.

700: 

701: \subsection{Jamming}

702: In order to isolate the STZ mechanism for jamming,

703: I now assume that free-volume is constant, at some value $\chi$;

704: the evolution of the system is governed by 

705: equations~(\ref{eqn:stzdil:1}) and~(\ref{eqn:stzdil:2}) only,

706: and a constant stress $\sigma$ is applied.

707: 

708: These equations admit two stationary solutions:

709: a jammed state for which, $\dot\epsilon=\dot\gamma=0$, and

710: $$

711: \Delta = \tanh\left({\sigma\over\bar\mu}\right)

712: \quad;

713: $$

714: a steady deformation regime, for which,

715: $$

716: \Delta = {1\over\mu_0 \sigma}

717: \quad,

718: $$

719: and,

720: $$

721: \dot\epsilon=E_0\,\exp\left[{-{1\over\chi}}\right]\,

722: \left(\sinh\left({\sigma\over\bar\mu}\right)-{1\over\mu_0 \sigma}\,\cosh\left({\sigma\over\bar\mu}\right)\right)

723: \quad.

724: $$

725: The latter solution is unstable at small stresses, in which case, the material jams.

726: For large stresses, the jammed state loose stability, and steady deformation results.

727: The two solutions exchange stability at the yield stress $\sigma_y$,

728: which is the solution of:

729: $$

730: \tanh\left({\sigma_y\over\bar\mu}\right) = {1\over\mu_0 \sigma_y}

731: \quad.

732: $$

733: 

734: \subsection{Various limits}

735: 

736: In the following, I will consider several limiting cases of the general constitutive

737: equations~(\ref{eqn:stzdil:1}-\ref{eqn:stzdil:3}).

738: 

739: In particular, the exponential dependency of $\ddpl$ as a function of $\sigma$

740: will not be studied, although this non-linearity might be 

741: important in some cases.~\cite{saulnier02,bureau02}

742: Activation factors involving $\sigma$ will thus be linearized;

743: this is valid under the assumption that $\sigma<<\bar\mu$;

744: $\bar\mu$ is then incorporated into constants by taking it to unity.

745: It leads to:

746: \begin{eqnarray}

747: \label{eqn:stzdil:lin:1}

748: \ddpl&=&E_0\,\exp\left[{-{1/\chi}}\right]\,\left(\sigma-\Delta\right)\\

749: \label{eqn:stzdil:lin:2}

750: \dot\Delta&=&{\ddpl\over\epsilon_0}\,\left(1- \mu_0\,\sigma\,\Delta\right)\\

751: \label{eqn:stzdil:lin:3}

752: \dot \chi &=& -E_1 \exp\left[-{\kappa/\chi}\right]+\alpha\,\sigma\ddpl

753: \quad.

754: \end{eqnarray}

755: 

756: I will also consider an {\em isotropic (liquid) limit}, 

757: where the structural anisotropy, measured by $\Delta$ plays no role. 

758: This is the case either at timescales when the dynamics of $\Delta$ is too slow,

759: $1/\epsilon_0\to0$ (and $\Delta=0$) or because the typical stress $1/\mu_0$ vanishes. 

760: In this case, the rheology is determined by a single state variable:

761: \begin{eqnarray}

762: \label{eqn:stzdil:iso:1}

763: \ddpl&=&E_0\,\exp\left[{-{1/\chi}}\right]\,\sigma\\

764: \label{eqn:stzdil:iso:2}

765: \dot \chi &=& -E_1 \exp\left[-{\kappa/\chi}\right]+\alpha\,\sigma\ddpl

766: \quad.

767: \end{eqnarray}

768: The sets of equations~(\ref{eqn:stzdil:lin:1}-\ref{eqn:stzdil:lin:3})

769: and~(\ref{eqn:stzdil:iso:1}-\ref{eqn:stzdil:iso:2})

770: will be helpful to emphasize the roles of $\Delta$ versus $\chi$. 

771: Many aspects of glassiness will be shown to be captured by

772: equations~(\ref{eqn:stzdil:iso:1}) and~(\ref{eqn:stzdil:iso:2}).

773: Jamming, however, is not captured at this level of simplification,

774: and requires the existence of the variable $\Delta$.

775: 

776: \subsection{Experimental protocol}

777: 

778: A word about the preparation of the sample: 

779: at time $t=0$, the system is quenched from a highly 

780: mixed state, in which, $\Delta=0$ and $\chi$ is large.

781: The sample then ages during a waiting time, $t_w$ without any forcing.

782: After this relaxation, some test is performed on the system.

783: 

784: Two main experimental procedures consist in forcing the stress,

785: or forcing the deformation. 

786: If a stress, $\sigma$, is applied, the plastic deformation $\gamma(t)$ 

787: is determined by constitutive equations, 

788: and the total deformation  is, $\epsilon = \gamma + \sigma/\mu$.

789: If the deformation, $\epsilon(t)$, is forced, the evolution 

790: of $\sigma$ is given by equation~(\ref{eqn:sigma:0}), 

791: and couples to constitutive equations.

792: 

793: If a small perturbation is applied to the sample,

794: the response of the system results from first order terms in the 

795: linearized constitutive equations~(\ref{eqn:stzdil:lin:1}-\ref{eqn:stzdil:lin:3}),

796: or~(\ref{eqn:stzdil:iso:1}-\ref{eqn:stzdil:iso:2}).

797: %The non-linear terms, which introduce a coupling between free-volume relaxation 

798: %and shear motion are expected to be negligible.

799: In relaxation processes, however, the linear approximation 

800: is not always valid at long times,~\cite{yoshino98,derec01} 

801: because, the activation factors can become small, 

802: possibly smaller than non-linear terms.

803: In the current theory, all non-linear terms involve the work of plastic

804: deformation $\sigma\dot\gamma$. Therefore, the linear analysis is valid

805: so long as this work is small compared to relaxation processes.

806: 

807: \section{Stress relaxation after a strain increment}

808: Stress relaxation is monitored after a small strain step.

809: The step can be understood as a ramp in the deformation, which occurs

810: on a very small time interval around $t_w$.

811: The strain $\epsilon$ determines the initial value of 

812: the stress, $\sigma(t_w^+)=\mu\,\epsilon$.

813: During the relaxation, $\dot\epsilon=0$;

814: the subsequent evolution of the stress is governed by $\dot\sigma=-\mu\dot\gamma$, 

815: and accounts for the plastic adaptation of the material to the imposed strain.

816: 

817: The relaxation modulus, $G(t)=\sigma(t)/\sigma(t_w)$ is the correlator associated 

818: with rearrangement processes: it measures the correlation between the 

819: initial state of the material, and its state at any following time, $t$.

820: Assuming that the diffusion of molecules is dominated by collective rearrangements,

821: stress relaxation can then be compared with the decay of 

822: self-intermediate scattering functions, which measure the spatial 

823: decorrelation resulting from the diffusion of particles in the dense medium.~\cite{boon91}

824: The long-time behavior of correlation functions is one of the major

825: properties of glassy materials.~\cite{goetze92}

826: (See also recent reviews by Debenedetti and Stillinger,~\cite{debenedetti01}

827: or Angell {\it et al}.~\cite{angell00}).

828: 

829: %For a small strain step, the non-linear term in~(\ref{eqn:stzdil:lin:3}) 

830: %or~(\ref{eqn:stzdil:iso:2}) is negligible;

831: %at long times, the logarithmic relaxation of $\chi$ is given by~(\ref{eqn:vf:relax}),

832: %and controls the timescale of stress relaxation.

833: 

834: \subsection{Isotropic limit}

835: 

836: In the isotropic limit, the constitutive equations 

837: are~(\ref{eqn:stzdil:iso:1}) and~(\ref{eqn:stzdil:iso:2}).

838: Stress relaxation is governed by the linearized equations:

839: \begin{eqnarray*}

840: \dot\sigma &=& -\mu\,E_0\,\exp\left[-{1\over\chi}\right]\sigma\\

841: \dot\chi &=& -E_1 \exp\left[-{\kappa\over\chi}\right]

842: \quad.

843: \end{eqnarray*}

844: To determine the relaxation modulus, $G=\sigma/\sigma(t_w)$,

845: the only relevant initial condition is the initial value of $\chi$,

846: when the strain is applied; this value is also determined by the initial free-volume

847: $\chi_0$ at time $t=0$ and by the waiting time $t_w$.

848: Moreover, either $E_1$ or $\mu\,E_0$ can be incorporated in the time scale,

849: therefore, only two effective parameters determine all the possible outcomes

850: of those equations: $\kappa$, and $\mu\,E_0/E_1$. While $\kappa$

851: measures the relative heights of entropy barriers, $\mu\,E_0/E_1$ measures 

852: the relative update frequencies of shear and compaction processes.

853: 

854: 

855: \subsubsection{Long time relaxation}

856: 

857: At long time, the relaxation of $\chi$ is given by~(\ref{eqn:vf:relax}),

858: and stress relaxation verifies:

859: $$

860: \dot\sigma = -\mu\,E_0\,\exp\left[-{1\over\chi}\right]\sigma

861: \simeq -\mu\,E_0\, \left({E_1\,t\over\kappa}\right)^{-1/\kappa}\sigma

862: \quad.

863: $$

864: 

865: The response modulus $G(t) = \sigma(t)/\sigma(t_w)$ is,

866: $$

867: G(t) \simeq \exp\left[{A\,\left(t_w^\beta-t^\beta\right)}\right]

868: $$

869: with an exponent $\beta$ which is directly related

870: to the ratio $\kappa$ between the heights of entropy barriers:

871: $$

872: \beta = 1-{1\over\kappa}\quad {\rm and}\quad A={\mu\, E_0\over\beta}

873: \,\left({E_1\over\kappa}\right)^{-1/\kappa}

874: \quad.

875: $$

876: 

877: The validity of linear approximation is checked by considering the time-dependency

878: of the work of plastic deformations $\sigma\dot\gamma$;

879: this term should be compared with the first term is free-volume relaxation,

880: $$

881: E_1\,\exp\left[-\kappa/\chi\right]\propto {\kappa\over t}

882: \quad.

883: $$

884: The stress relaxation, leads to the following evolution for the work of plastic 

885: deformations:

886: $$

887: \sigma\dot\gamma \propto t^{-{1/\kappa}}\;\exp\left[-2At^\beta\right]

888: \quad.

889: $$

890: If $\kappa>1$, or $\beta>0$, the relaxation of $\sigma\dot\gamma$ is dominated

891: by the exponential factor;

892: if $\kappa<1$, the stress goes to a constant, but $t^{-{1/\kappa}}$ decays

893: faster than $1/t$.

894: In all cases, the non-linear term decays faster than the linear term,

895: which justifies the linear approximation at all times.

896: 

897: For $\beta>0$, {\it i.e.} $\kappa>1$, the stress undergoes KWW relaxation. 

898: This property results directly from the existence of two 

899: types of saddle points of unequal heights,

900: when hopping motion is controlled by a logarithmically relaxing free-volume.

901: Measured values of exponent $\beta$ in glassy materials, 

902: ranging from $0.2$ to $0.5$~\cite{larson99,angell00}

903: are consistent with, $\kappa=v_1/v_0\in[1.25,2]$.

904: The ratio $\kappa$ is related to the shape of molecules,

905: and to the details of their interactions.

906: There is no {\it a priori} reason, from a phase space point of view,

907: to assume a particular value for $\kappa$. In this work I will systematically

908: study both cases in order to draw a complete picture of the phenomenology 

909: described by the proposed constitutive equations.

910: 

911: If $\kappa<1$, then $\beta<0$, and

912: the previous calculation shows that the stress assumes a non-vanishing 

913: asymptotic value after relaxation. 

914: In this case, $A<0$, and 

915: the stress relaxation towards its ultimate constant value can 

916: be further estimated as

917: a power law:

918: \begin{eqnarray*}

919: G(t)

920: &\simeq& \exp\left[ |A|\,\left({t^{-|\beta|}}-{t_w^{-|\beta|}}

921: \right)\right]\\

922: &\simeq& \exp\left[-|A|\,t_w^{-|\beta|}\right]\;\left(1+|A|\,t^{-|\beta|}\right)

923: \quad.

924: \end{eqnarray*}

925: %The existence of a non-vanishing stress at long times is due to the fast

926: %free-volume relaxation, faster than elementary shear motion, on a logarithmic scale:

927: %the voids in the material disappear faster than plastic deformation is permitted,

928: %the system is trapped in a state where some residual shear exists.

929: Interpreting $G$ as a correlation function, it also means that the material 

930: never completely decorrelates from its initial state. 

931: This non-ergodicity is an indication of the glass transition as understood

932: in mode-coupling theory~\cite{goetze92};

933: it can be characterized by the asymptotic value $G(t\to\infty)$, 

934: the Edwards-Anderson parameter:

935: $$

936: G(t\to\infty) \simeq \exp\left[{-|A|\,t_w^{-|\beta|}}\right]

937: \quad,

938: $$

939: which increases with the age of the sample. 

940: 

941: Such a dependency of the Edwards-Anderson parameter

942: of the age of the sample has been observed experimentally by Bonn and coworkers

943: on a colloidal glass at very low concentrations.~\cite{bonn99}

944: 

945: \subsubsection{Short time response}

946: 

947: At short time, if free-volume can be assumed large or constant,

948: the exponential term $\exp\left[-{1/\chi}\right]$ does not vary very much,

949: and stress undergoes an exponential relaxation:

950: $$

951: G(t) = 

952: \exp\left[-\mu\,E_0\,\exp\left[-{1\over\chi_0}\right]\;\left(t-t_w\right)\right]

953: \quad.

954: $$

955: This relaxation occurs on a time scale of order, 

956: $$

957: \tau_\sigma = {1\over\mu\,E_0}\,\exp\left[{1\over\chi_0}\right]

958: \quad,

959: $$

960: which should be compared with the time scale of free-volume relaxation, $\tau_\chi$.

961: If $\tau_\sigma>>\tau_\chi$, free-volume relaxes faster than stress, 

962: and the long time, free-volume controlled, stress relaxation can be observed.

963: If $\tau_\sigma<<\tau_\chi$, the stress relaxes exponentially before free-volume

964: achieves sufficiently low values to slow down stress relaxation.

965: The crossover between exponential and KWW relaxation is determined by,

966: $\tau_\sigma\sim\tau_\chi$ or,

967: $$

968: {1\over\mu\,E_0}\,\exp\left[{1\over\chi_0}\right] \sim 

969: {\chi_0\over E_1}\;\exp\left[{\kappa\over\chi_0}\right]

970: \quad.

971: $$

972: For a large initial free-volume, this reduces to,

973: $$

974: E_1 \sim \mu\,E_0\,\chi_0

975: \quad;

976: $$

977: KWW relaxation can be observed for a larger value of $E_1$, 

978: or for sufficiently small initial value of the free-volume, $\chi(t_w)$,

979: (that is for sufficiently large waiting time $t_w$).

980: 

981: In order to look at the short and long time response of the system, 

982: equations~(\ref{eqn:sigma:0}),~(\ref{eqn:stzdil:iso:1}) and~(\ref{eqn:stzdil:iso:2})

983: are integrated numerically with a Runge-Kutta algorithm with adaptive timestep.

984: The system is prepared in a dilute state, $\chi=\chi_0$, which is allowed to relax 

985: during time $t_w$, in the absence of any forcing ($\sigma([0,t_w])=0$).

986: At time $t_w$, a strain $\epsilon=10^{-3}$ is applied and produces an 

987: elastic response, 

988: $\sigma(t_w^+)=\mu\epsilon$. The dynamics of free-volume and stress are then monitored.

989: 

990: \paragraph{Crossover stress.}

991: In a first series of tests, $t_w$ is set to 0, 

992: and the relaxation spectra are displayed for different values of the parameter $E_1$.

993: Parameters $E_0$, $\mu$ and $\alpha$ are set to 1; $\kappa=2$ and $\chi_0=20$.

994: The results are displayed figure~\ref{fig:relax1}.

995: For the first value $E_1=1$, only exponential relaxation is observed

996: because all the stress has been released before $\chi$ entered the logarithmic

997: relaxation.

998: The value $E_1=20$ corresponds to the crossover $\mu\,E_0\,\chi_0$:

999: KWW relaxation is observed at later stages of the relaxation,

1000: with a non-vanishing values of $G$.

1001: %Figure~\ref{fig:relax1} shows that the plateau in the relaxation spectrum

1002: %is enhanced by a high value of $E_1$. 

1003: %In fact, since one of the constants $E_1$ and $\mu E_0$ can be eliminated

1004: %by a rescaling of time, 

1005: %the qualitative dynamics depends on the ratio $\mu E_0/E_1$ only.

1006: %The time to crossover is determined by free-volume relaxation.

1007: For a given $\chi_0$,

1008: the stress release during the exponential part of the relaxation is

1009: proportional to $\mu\,E_0$, hence inversely proportional to $E_1$.

1010: The plateau in the relaxation spectrum can be observed only 

1011: for sufficiently large values of $E_1/(\mu\,E_0)$ so that $G$ is still observable 

1012: by the time $\chi$ enters the logarithmic relaxation.

1013: 

1014: These relaxation curves are very similar to those observed in 

1015: MD simulations,~\cite{kob95,angell00,debenedetti01} 

1016: or in experiments;~\cite{megen94}

1017: the general appearance of those curves depends sensibly

1018: on the parameters considered. 

1019: From a practical point of view, it means that it depends

1020: on the time window that is available, numerically or experimentally.

1021: \begin{figure}

1022: \narrowtext

1023: \begin{center}

1024: \unitlength = 0.005\textwidth

1025: \begin{picture}(100,110)(5,0)

1026: \put(10,5){\resizebox{90\unitlength}{!}{\includegraphics{relax1.eps}}}

1027: %\put(0,0){\line(1,0){100}}

1028: %\put(0,0){\line(0,1){100}}

1029: \put(5,95){\makebox(0,0){\large $1/\chi$}}

1030: \put(5,50){\makebox(0,0){\large $G$}}

1031: \put(90,2){\makebox(0,0){\large $\log_{10}(t)$}}

1032: \end{picture}

1033: \end{center}

1034: \caption{\label{fig:relax1}

1035: Numerical integration of equations~(\ref{eqn:sigma:0}),~(\ref{eqn:stzdil:iso:1})

1036: and~(\ref{eqn:stzdil:iso:2}) for a fixed strain $\epsilon=10^{-3}$.

1037: Parameters are $E_0=\alpha=\mu=1$, and $\kappa=2$.

1038: The initial stress is set to $\sigma(t_w)=\mu\epsilon=10^{-3}$, 

1039: and $\chi_0=20$, $t_w=0$.

1040: Top: $1/\chi$ as a function of $\log_{10}(t)$

1041: for $E_1 = 1, 20, 40, 80, 100$ from right to left.

1042: Bottom: relaxation spectra with $E_1$ increasing from left to right.

1043: }

1044: \end{figure}

1045: 

1046: %Such curves are reminiscent of intermediate scattering functions as 

1047: %displayed in many works on colloidal suspensions~\cite{vanmegenm}

1048: %or in recent experiments on Laponite.~\cite{bonn98}

1049: 

1050: \paragraph{Aging}

1051: The crossover between exponential relaxation can also be studied by varying

1052: the initial free-volume. This amounts to varying the waiting time $t_w$

1053: for a given $\chi(t=0)$.

1054: In order to observe a plateau in the relaxation spectrum, parameter $E_1$ is set to 20,

1055: and parameters $E_0$, $\mu$ and $\alpha$ are set to 1; $\kappa=2$.

1056: In order to obtain an exponential relaxation for short waiting times, 

1057: and to observe a wide variety of behavior, the initial value of $\chi$ is set to 100.

1058: The results are displayed figure~\ref{fig:relax2}. With this choice of parameter,

1059: the curves resemble experimental data.~\cite{angell00,gopal99,hebraud97,megen98,cipelletti00} 

1060: 

1061: Such curves strongly depend on the available experimental window, and

1062: on the width of the crossover region which is unobservable in a log-lin plot for 

1063: smaller values of $E_1/(\mu\,E_0)$. 

1064: To emphasize this point,

1065: another set of relaxation spectra is presented figure~\ref{fig:relax}

1066: for values of the parameters: $E_0=\alpha=\mu=1$, $E_1=8$, and $\kappa=2$.

1067: The initial condition is $\chi(0)=100$ and waiting times are:

1068: $t_w = 0, 5, 10, 15, 50, 100$. The relaxation spectra are presented in log-lin,

1069: log-log, and log-log(log) plot; the crossover between exponential and KWW relaxation 

1070: is observable around values of the relaxation modulus 

1071: of order $10^{-5}$, which is featureless in a log-lin plot.

1072: \begin{figure}

1073: \narrowtext

1074: \begin{center}

1075: \unitlength = 0.005\textwidth

1076: \begin{picture}(100,110)(5,0)

1077: \put(10,5){\resizebox{90\unitlength}{!}{\includegraphics{relax2.eps}}}

1078: %\put(0,0){\line(1,0){100}}

1079: %\put(0,0){\line(0,1){100}}

1080: %\put(100,0){\line(1,0){100}}

1081: \put(5,95){\makebox(0,0){\large $1/\chi$}}

1082: \put(5,50){\makebox(0,0){\large $G$}}

1083: \put(90,2){\makebox(0,0){\large $\log_{10}(t-t_w)$}}

1084: \end{picture}

1085: \end{center}

1086: \caption{\label{fig:relax2}

1087: Numerical integration of equations~(\ref{eqn:sigma:0}),~(\ref{eqn:stzdil:iso:1}) 

1088: and~(\ref{eqn:stzdil:iso:2}) for a fixed strain $\epsilon=10^{-3}$.

1089: Parameters are $E_0=\alpha=\mu=1$, $E_1=20$, and $\kappa=2$.

1090: Initial conditions are obtained by a quench at time $t=0$ from a dilute state

1091: ($\sigma(0)=0$ and $\chi(0)=100$) followed by density relaxation without forcing during 

1092: time $t_w$; at time $t_w$ the stress is set to $\sigma(t_w)=\mu\epsilon=10^{-3}$.

1093: Top: $1/\chi$ as a function of $\log_{10}(t-t_w)$

1094: with $t_w = 0, 5, 10, 15, 50, 100$ from bottom to top.

1095: Bottom: relaxation spectra for increasing $t_w$ from left to right.

1096: }

1097: \end{figure}

1098: \begin{figure}

1099: \narrowtext

1100: \begin{center}

1101: \unitlength = 0.005\textwidth

1102: \begin{picture}(100,110)(5,0)

1103: \put(10,5){\resizebox{90\unitlength}{!}{\includegraphics{relax.eps}}}

1104: %\put(0,0){\line(1,0){100}}

1105: \put(5,95){\makebox(0,0){\large $G$}}

1106: \put(5,50){\makebox(0,0){\large $\log_{10}(G)$}}

1107: \put(90,2){\makebox(0,0){\large $\log_{10}(t-t_w)$}}

1108: \end{picture}

1109: \end{center}

1110: \caption{\label{fig:relax}

1111: Numerical integration of equations~(\ref{eqn:sigma:0}),~(\ref{eqn:stzdil:iso:1}) 

1112: and~(\ref{eqn:stzdil:iso:2}) for a fixed strain $\epsilon=10^{-3}$.

1113: Parameters are $E_0=E_1=\alpha=\mu=1$, and $\kappa=1.5$.

1114: Initial conditions are obtained by a quench at time $t=0$ from a dilute state

1115: ($\sigma(0)=0$ and $\chi(0)=10$) followed by density relaxation without forcing during 

1116: time $t_w$; at time $t_w$ the stress is set to $\sigma(t_w)=\mu\epsilon=10^{-3}$.

1117: Top: $1/\chi$ as a function of $\log_{10}(t-t_w)$

1118: with $t_w = 0, 5, 10, 50, 100$ from bottom to top.

1119: Bottom: relaxation spectra for increasing $t_w$ from bottom to top.

1120: There is no inconsistency in the large logarithmic scale for $\log_{10}(G)$: 

1121: the scale is determined up to the reduced parameter $A$.

1122: Inset: $\log_{10}(-\log_{10}(G))$ as a function of $\log_{10}(t-t_w)$,

1123: with $t_w$ increasing from top to bottom.

1124: The spectra with the longest waiting time are already in the KWW regime.

1125: }

1126: \end{figure}

1127: 

1128: 

1129: \paragraph{Ergodic to non-ergodic transition.}

1130: I now consider the dependency of relaxation spectra on the parameter $\kappa$.

1131: This parameter is of primary importance in the current approach and

1132: is directly related to the existence of entropy barriers of unequal heights 

1133: in the phase space.

1134: The parameters are taken to be, $E_0=\alpha=\mu=1$, $E_1=10$, 

1135: the initial condition is $\chi(0)=100$, and the waiting

1136: is set to $t_w=10$. This choice is primarily motivated to obtain easily

1137: observable features on the relaxation spectrum.

1138: The results are displayed figure~\ref{fig:relax3}.

1139: When $\kappa$ goes to 1, the system presents a transition to non-ergodic behavior.

1140: For $\kappa<1$, the asymptotic value of the stress is no longer vanishing.

1141: At the point $\kappa=1$, the long time relaxation of $\sigma$ verifies:

1142: $$

1143: \dot\sigma = -\mu\,E_0\,\exp\left[-{1\over\chi}\right]\sigma

1144: \simeq -\mu\,{E_0\over E_1\,t}\sigma

1145: \quad,

1146: $$

1147: whence an asymptotic power law relaxation:

1148: $$

1149: G(t) \simeq \left({t\over t_w}\right)^{-{\mu\,E_0/ E_1}}

1150: \quad.

1151: $$

1152: 

1153:  

1154: \begin{figure}

1155: \narrowtext

1156: \begin{center}

1157: \unitlength = 0.005\textwidth

1158: \begin{picture}(100,110)(5,0)

1159: \put(10,5){\resizebox{90\unitlength}{!}{\includegraphics{relax3.eps}}}

1160: %\put(0,0){\line(1,0){100}}

1161: %\put(0,0){\line(0,1){100}}

1162: \put(5,95){\makebox(0,0){\large $1/\chi$}}

1163: \put(5,50){\makebox(0,0){\large $G$}}

1164: \put(90,2){\makebox(0,0){\large $\log_{10}(t-t_w)$}}

1165: \end{picture}

1166: \end{center}

1167: \caption{\label{fig:relax3}

1168: Numerical integration of equations~(\ref{eqn:sigma:0}),~(\ref{eqn:stzdil:iso:1}) 

1169: and~(\ref{eqn:stzdil:iso:2}) for a fixed strain $\epsilon=10^{-3}$.

1170: Parameters are $E_0=\alpha=\mu=1$ and $E_1=10$; $\kappa$ varies.

1171: Initial conditions are obtained by a quench at time $t=0$ from a dilute state

1172: ($\sigma(0)=0$ and $\chi(0)=100$) followed by density relaxation without 

1173: forcing during time $t_w$; 

1174: at time $t_w$ the stress is set to $\sigma(t_w)=\mu\epsilon=10^{-3}$.

1175: Top: $1/\chi$ as a function of $\log_{10}(t-t_w)$

1176: with $\kappa = 1, 1.2, 1.4, 1.6, 1.8, 2$ from top to bottom.

1177: Bottom: relaxation spectra for increasing $\kappa$ from top to bottom.

1178: }

1179: \end{figure}

1180: 

1181: 

1182: \subsection{General case}

1183: 

1184: KWW relaxation is also achieved with the complete set of 

1185: equations~(\ref{eqn:stzdil:lin:1}-\ref{eqn:stzdil:lin:3}).

1186: At time $t=0$, $\Delta(0)=0$, and no constraint is applied until time $t_w$:

1187: $\sigma([0,t_w]) = 0$.

1188: During the ramp, the deformation of the material is essentially elastic, 

1189: no rearrangement occurs, hence, $\Delta(t_w^+)=0$, while the imposed 

1190: strain determines the initial value the stress: $\sigma(t_w^+)=\mu\epsilon$.

1191: 

1192: During the ensuing relaxation, $\dot\sigma=-\mu\dot\gamma$ and,

1193: for a small deformation, $\dot\Delta\simeq\dot\gamma/\epsilon_0$. 

1194: It results that, at all times, the quantity $\sigma+\mu\,\epsilon_0\,\Delta$ 

1195: is constant:

1196: $$

1197: \sigma+\mu\,\epsilon_0\,\Delta = \epsilon\, U_0

1198: \quad.

1199: $$

1200: Here, $U_0=\mu$,

1201: but this expression may vary, depending on the preparation of a sample.

1202: To treat initial values of $\sigma$ and $\Delta$ in the greatest generality,

1203: the initial value for the difference $\sigma-\Delta$ is denoted $\epsilon\,S_0$

1204: (and here, $S_0=\mu$).

1205: 

1206: At long times, the dynamics of $\Delta$ and $\sigma$ verify,

1207: \begin{eqnarray*}

1208: \dot\sigma-\dot\Delta &\simeq& -\left(\mu+{1\over\epsilon_0}\right)\,E_0\,

1209: \exp\left[-{1\over\chi}\right]\left(\sigma-\Delta\right)\\

1210: &\simeq& -\left(\mu+{1\over\epsilon_0}\right) 

1211: \,E_0\,\left({E_1\,t\over\kappa}\right)^{-1/\kappa}\left(\sigma-\Delta\right)

1212: \end{eqnarray*}

1213: whence,

1214: $$

1215: \sigma-\Delta \simeq \epsilon\,S_0\,\exp\left[{A'\,(t_w^\beta-t^\beta)}\right]

1216: $$

1217: with,

1218: $$

1219: \beta = 1-{1\over\kappa}\quad {\rm and}\quad 

1220: A'=\left(\mu+{1\over\epsilon_0}\right)\,{E_0\over\beta}\,

1221: \left({E_1\over\kappa}\right)^{-1/\kappa}

1222: \quad.

1223: $$

1224: The evolution in time of the quantity $\sigma-\Delta$ 

1225: is governed by the same equation as the dynamics of $\sigma$ in the isotropic limit.

1226: Therefore, all the previous discussion presented in the isotropic limit pertains

1227: to the general case, up to a constant term, which is fixed by the conservation

1228: of the quantity $\sigma+\mu\,\epsilon_0\,\Delta$.

1229: The dynamics of $\Delta$ and $\sigma$ follow.

1230: 

1231: In particular, the evolution of $\Delta$ reads,

1232: $$

1233: \Delta(t) \propto \epsilon\;{U_0-S_0\exp\left[{A'\,(t_w^\beta-t^\beta)}\right]\over 1+\mu\epsilon_0}

1234: \quad,

1235: $$

1236: and the relaxation modulus verifies,

1237: $$

1238: G(t) \propto {U_0+\mu\epsilon_0\,S_0\,\exp\left[{A'\,(t_w^\beta-t^\beta)}\right]\over 1+\mu\epsilon_0}

1239: \quad.

1240: $$

1241: The isotropic expression for $G(t)$ is recovered by taking $\epsilon_0\to\infty$.

1242: Note that the relaxation of the stress could lead to a negative value if $U_0<0$,

1243: which depends on the preparation of the sample.

1244: 

1245: For $\beta>0$ ({\it i.e.} $\kappa>1$), the stress undergoes 

1246: a KWW relaxation towards a non-vanishing value, which

1247: depends on the properties of the material and on the preparation of the sample.

1248: For $\beta<0$ power law relaxation is observed towards a non-vanishing value

1249: of the stress.

1250: %Such relaxation towards a negative value of the stress has been observed in 

1251: %recent experiments....

1252: 

1253: In the current work, two origins are therefore identified for 

1254: the existence of a non-vanishing stress at long times: 

1255: the creation of mechanical anisotropy of the contact network (jamming), 

1256: or the existence of a ergodic to non-ergodic transition

1257: for the hopping motion in phase space (freezing). 

1258: Recent observations by Bonn {\it et al} indicate that for a given material, 

1259: jamming could not be identified only to a loss of ergodicity, 

1260: and that some amount of jamming could coexist 

1261: with liquid-like behavior.~\cite{bonn02a,bonn02b} 

1262: Such features are allowed in the current theory and show the importance 

1263: of the STZ mechanism for soft glassy materials.

1264: 

1265: The validity of the linear approximation is checked by considering the 

1266: long times relaxation of the rate of plastic deformation:

1267: $$

1268: \dot\gamma \propto t^{-{1/\kappa}}\;\exp\left[{-A'\,t^\beta}\right]

1269: $$

1270: and the work of plastic deformation decays like $\dot\gamma$ since $\sigma$ goes to a constant;

1271: the non-linear term in equation~(\ref{eqn:stzdil:lin:3}) 

1272: is negligible at all times.

1273: 

1274: % ???:  \subsection{Response to a large strain}

1275: \section{Response to a constant strain rate}

1276: 

1277: \subsection{\label{sec:steady} Steady plastic flow}

1278: When a constant strain rate $\dot\epsilon$ is applied,

1279: the material is driven strongly out-of-equilibrium;

1280: the response of the system involves non-linear terms.

1281: I study here the stationary states of a sheared material, in which, $\dot\gamma=\dot\epsilon$.

1282: In the first sections, I make a very primitive stability analysis of those stationary

1283: states, by assuming that $\dot\gamma=\dot\epsilon$ is fixed at all times, and that the

1284: stress is given by the constitutive relation~(\ref{eqn:stzdil:lin:1}) 

1285: or~(\ref{eqn:stzdil:iso:1}); this allows to decouple equation~(\ref{eqn:sigma:0}).

1286: 

1287: The existence of an unstability in the complete set of equations accounts 

1288: for the emergence of stick-slip motion in boundary lubrication.~\cite{lemaitre01a}

1289: This question will be briefly discussed in section~\ref{sec:hopf},

1290: but raises so many issues that it will be tackled thoroughly in another publication;

1291: for the sake of completeness, 

1292: the calculation of the Hopf criterion is given in appendix~\ref{app:hopf}.

1293: 

1294: \subsubsection{Isotropic limit}

1295: In the limit $\mu_0\to\infty$ (no yield stress), the relation between

1296: shear rate and shear stress reads (from~(\ref{eqn:stzdil:iso:1})):

1297: $$

1298: \sigma = {\dot\epsilon \over E_0}\, \exp\left[{1\over\chi}\right]

1299: \quad.

1300: $$

1301: Eliminating $\sigma$ between equations~(\ref{eqn:stzdil:iso:1}) and~(\ref{eqn:stzdil:iso:2}), 

1302: yields:

1303: \begin{equation}

1304: \dot\chi = -E_1 \exp\left[{-{\kappa\over\chi}}\right] + \alpha {\dot\epsilon^2\over E_0}\, \exp\left[{1\over\chi}\right]

1305: \label{eqn:chidyn}

1306: \end{equation}

1307: which is positive iff,

1308: $$

1309: {\alpha\dot\epsilon^2\over E_0\,E_1} > e^{-{\kappa+1\over\chi}}

1310: \quad.

1311: $$

1312: In the stationary flow, the relation between shear strain and volume reads,

1313: $$

1314: \dot\epsilon = \sqrt{E_0\,E_1\over\alpha}

1315: \,\exp\left[{-{\kappa+1\over2\chi}}\right]

1316: \quad.

1317: $$

1318: For future use, I introduce the critical rate of deformation,

1319: $$

1320: \dot\epsilon^* = \sqrt{E_0\,E_1\over\alpha}

1321: \quad.

1322: $$

1323: \begin{figure}

1324: \narrowtext

1325: \begin{center}

1326: \unitlength=0.005\textwidth

1327: \begin{picture}(110,60)(8,-8)

1328: \put(13,40){\makebox(0,0){\large$\dot\epsilon^*$}}

1329: \put(10,48){\makebox(0,0){\large$\dot\epsilon$}}

1330: \put(85,-13){\makebox(0,0){\large$\chi$}}

1331: \put(10,-13){\resizebox{90\unitlength}{!}{\includegraphics{graph.ps}}}

1332: \end{picture}

1333: \end{center}

1334: \caption{\label{fig:vf}

1335: Bifurcation diagram for equation~(\protect\ref{eqn:chidyn}).

1336: $\dot\epsilon$ is drawn as a function of $\chi$,

1337: and the strain rate $\dot\epsilon$ is fixed.

1338: For $\dot\epsilon> \dot\epsilon^*$,

1339: no solution exists, $\dot\chi>0$, the system blows up;

1340: for $\dot\epsilon < \dot\epsilon^*$ the system admits one stable solution.}

1341: \end{figure}

1342: The bifurcation diagram for $\chi$ is presented figure~\ref{fig:vf}.

1343: If $\dot\epsilon> \dot\epsilon^*$, then $\dot\chi>0$ at all times: 

1344: free-volume diverges, the system blows up.

1345: If $\dot\epsilon < \dot\epsilon^*$, the equation for $\chi$ admits one solution:

1346: $$

1347: \chi = -{\kappa+1\over 2\,\left(\log{\dot\epsilon}-\log{\dot\epsilon^*}\right)}

1348: \quad.

1349: $$

1350: The existence of a stationary value for the free-volume indicates 

1351: that for any finite rate of deformation, aging stops.

1352: %This has already been observed in experiments.

1353: In fact, this is not completely exact if a vanishingly small shear rate

1354: was applied, such that the associated steady free-volume is so small, 

1355: that it is not accessible at reasonable timescales. 

1356: As a consequence, for very small $\dot\gamma$,

1357: the apparent viscosity saturates to a constant value. This is a common feature

1358: of complex liquids (see {\it e.g.}~\cite{larson99}).

1359: 

1360: In the steady state the value of the shear stress is,

1361: $$

1362: \sigma= {\dot\epsilon^{n}\over E_0}\,\left({\alpha\over E_0 E_1}\right)^{{n-1\over2}}

1363: \quad,

1364: $$

1365: with,

1366: $$

1367: n={\kappa-1\over\kappa+1}

1368: \quad.

1369: $$

1370: The system behaves like a power law fluid;

1371: the exponent $n$ of the power law viscosity is

1372: related to the exponent $\beta$ of KWW relaxation by,

1373: $$

1374: n = {\beta\over 2-\beta}

1375: \quad;

1376: $$

1377: those two exponents are directly related to the slope $\kappa$ of $1/\chi$

1378: as a function of $\ln t$.

1379: 

1380: The assumptions underlying the current approach 

1381: do not permit to evaluate the activation volumes associated with density relaxation

1382: and shear motion. However, the single parameter $\kappa$ determines two 

1383: exponents associated with the linear and non-linear response of a strained material.

1384: 

1385: The existence of such a relation constitutes a important result of the current theory,

1386: although the form given to this relation depends on non-generic features:

1387: how non-linear dilatancy is introduced in the equation of motion for free-volume; 

1388: how the update frequencies (here taken constant) are written.

1389: In a hard-sphere material, for example, collision frequencies

1390: depend on free-volume as $1/\chi$, and the relation between $\sigma$ and 

1391: $\dot\epsilon$ is expected to be modified by logarithmic corrections.

1392: 

1393: Molecular dynamics simulations of sheared fluid are consistent with power law rheologies

1394: for $n$ equal to 0.1~\cite{berthier00,berthier01} or 0.2~\cite{yamamoto97,yamamoto98}

1395: Recent measurement of power law rheology have also been 

1396: obtained on a colloidal glass,

1397: leading to values of $n$ ranging from 0.1 to 0.35.~\cite{bonn02a,bonn02b}

1398: A value of $\beta$ of order $0.4$ corresponds to $n=0.25$, which seems 

1399: in reasonable agreement with the currently available data.

1400: 

1401: The theory of structural rearrangement described here, 

1402: is expected to present some degree of universality, 

1403: and hold, in a given form,  for wide classes of materials.

1404: What this model shows is that there is an intimate relation between 

1405: time-logarithmic density relaxation, KWW relaxation,

1406: and power law rheology of a dense material, and that this relation may depend

1407: only on generic features for a given class of materials.

1408: Therefore the existence of a relation between $\beta$ and $n$

1409: could be tested by considering families of materials. 

1410: The experimental or numerical evaluation of such a relation 

1411: appears as an important mean of characterization of the coupling between 

1412: two major modes of deformation.

1413: 

1414: \subsubsection{General case}

1415: 

1416: The constitutive equations are now~(\ref{eqn:stzdil:lin:1}-\ref{eqn:stzdil:lin:3}).

1417: In the steady state,

1418: $$

1419: \Delta = {1\over\mu_0\sigma}

1420: \quad;

1421: $$

1422: plugging expression~(\ref{eqn:stzdil:lin:1}) for the plastic deformation 

1423: in equation~(\ref{eqn:stzdil:lin:3}) leads to the following relation between $\chi$

1424: and $\sigma$ in the steady state:

1425: $$

1426: \exp\left[{1-\kappa\over\chi}\right]={\alpha\,E_0\over E_1}\,\left(\sigma^2-{1\over\mu_0}\right)

1427: %\sigma = \sqrt{{1\over\mu_0}+{E_1\over\alpha E_0}\,\exp\left[{1-\kappa\over\chi}\right]}}

1428: \quad.

1429: $$

1430: Using~(\ref{eqn:stzdil:lin:3}) again, 

1431: $\dot\epsilon$ can then be written as a function of $\sigma$:

1432: $$

1433: \dot\epsilon = 

1434: {E_1\over\alpha}\;

1435: \left(\alpha\,E_0\over E_1\right)^{{\kappa\over\kappa-1}}

1436: \;{1\over\sigma}\;\left(\sigma^2-{1\over\mu_0}\right)^{{\kappa\over\kappa-1}}

1437: \quad.

1438: $$

1439: For large $\sigma$ or $\dot\epsilon$, the system behaves as a power law fluid, 

1440: with the same exponent $n$ as in the isotropic limit.

1441: However, $\dot\epsilon$ vanishes for a non-zero value of $\sigma$, 

1442: which determines the yield stress:

1443: $$

1444: \sigma_y = {1\over\sqrt{\mu_0}}

1445: \quad.

1446: $$

1447: 

1448: The following expression can also be obtained for $\dot\epsilon$ as a function of $\chi$:

1449: $$

1450: \dot\epsilon = {

1451: \sqrt{E_0\,\mu_0}\,E_1\;\exp\left[-{\kappa/ 2\chi}\right] 

1452: \over

1453: \sqrt{\alpha}\;

1454: \sqrt{E_0\alpha\,\exp\left[{\kappa/\chi}\right] 

1455: 		    + E_1\mu_0\,\exp\left[{1/\chi}\right]}

1456: }

1457: \quad,

1458: $$

1459: which is an increasing function of $\chi$, and which has a maximum for $\chi\to\infty$:

1460: $$

1461: \dot\epsilon^*=

1462: {\sqrt{E_0\,\mu_0}\,E_1

1463: \over

1464: \sqrt{\alpha}\;

1465: \sqrt{E_0\alpha+ E_1\mu_0}

1466: }

1467: $$

1468: For $\dot\epsilon>\dot\epsilon^*$, like in the isotropic case,

1469: the shear deformation leads to a constant increase of free-volume, 

1470: and the system blows up.

1471: 

1472: \subsection{Transients}

1473: 

1474: When a shear rate is suddenly imposed on a material, 

1475: the transient dynamics often lead to a first rise of the stress before 

1476: a dynamical yield stress is reached at which plastic deformations begin;

1477: once plastic deformation occur, the stress relaxes towards its strain-rate

1478: dependent asymptotic value.~\cite{larson99}

1479: These features are observed in sheared bulk solids~\cite{hassan95}

1480: but also in sheared thin films,~\cite{drummond00,drummond01,persson00}

1481: granular materials~\cite{losert00}

1482: or in solid friction.~\cite{bureau02}

1483: Another generic property of transient regimes is that 

1484: the dynamical yield stress depends on the age of the material: 

1485: during aging, a material strengthens.

1486: 

1487: These phenomena are captured by the proposed equations.

1488: %The interplay of the dynamics of free-volume and of arrangements during

1489: %transient regimes leads to a rather diverse phenomenology.

1490: Strain softening and strengthening result essentially from the equation of motion 

1491: for free-volume and are captured by the isotropic equations. 

1492: 

1493: Figure~\ref{fig:bump} present some examples of transient dynamics

1494: obtained by numerical integration of

1495: equations~(\ref{eqn:sigma:0}),~(\ref{eqn:stzdil:iso:1}-\ref{eqn:stzdil:iso:2}).

1496: Parameters are set to $E_0=E_1=\mu=\alpha=1$.

1497: The initial value for the free-volume (at time $t=0$) is set to $\chi(0)=10$, 

1498: and the system relaxes without forcing during a time $t_w$.

1499: At time $t_w$, a constant strain rate is applied.

1500: To prevent the blow-up of the system,

1501: $\dot\epsilon$ must be smaller than $\dot\epsilon^*$,

1502: which equals $1$ with the chosen parameters;

1503: for this reason, the strain rate is set to $\dot\epsilon=0.9$.

1504: Two different values of the parameter $\kappa$ have been used to explore

1505: the two important cases when $\kappa$ is larger or smaller than 1.

1506: The stress is displayed as a function of the total time $t$ 

1507: since preparation of the sample.

1508: \begin{figure}

1509: \narrowtext

1510: \begin{center}

1511: \unitlength = 0.005\textwidth

1512: \begin{picture}(100,110)(5,0)

1513: \put(10,5){\resizebox{90\unitlength}{!}{\includegraphics{bumps.eps}}}

1514: %\put(0,0){\line(1,0){100}}

1515: %\put(0,0){\line(0,1){100}}

1516: %\put(100,0){\line(1,0){100}}

1517: \put(5,95){\makebox(0,0){\large $1/\chi$}}

1518: \put(5,50){\makebox(0,0){\large $\sigma$}}

1519: \put(90,2){\makebox(0,0){\large $t$}}

1520: \end{picture}

1521: \end{center}

1522: \caption{\label{fig:bump}

1523: Numerical integration of equations~(\ref{eqn:sigma:0}),~(\ref{eqn:stzdil:iso:1}) 

1524: and~(\ref{eqn:stzdil:iso:2}) for a fixed strain rate $\dot\epsilon=0.9$.

1525: Parameters are $E_0=E_1=\alpha=\mu=1$, and values of $\kappa$:

1526: $\kappa=0.8$ (solid lines) and $\kappa=1.2$ (dashed lines).

1527: No stress is applied during the initial density relaxation of the material 

1528: from, $\chi(0)=10$.

1529: At various times, $t_w$, the strain rate is suddenly applied and the ensuing

1530: dynamics of $1/\chi$ (top) and  $\sigma$ (bottom) are displayed.

1531: The strong glass $\kappa<1$ presents larger of the dynamical yield, 

1532: which result from the smaller values of $\chi$ reached.

1533: }

1534: \end{figure}

1535: 

1536: A dynamical yield stress clearly emerges for a sufficiently aged material, 

1537: and increases with time. The yielding of the material is accompanied by a sudden

1538: rise of the free-volume allowing transformation rates to display non-vanishing values.

1539: 

1540: %When a constant strain rate is imposed on the material, 

1541: %the stress starts to increase linearly with time,

1542: %so long as the plastic deformation rate is negligible

1543: %compared to the elastic deformation, or $\dot\epsilon>>\dot\gamma$.

1544: %This vanishing value of $\dot\gamma$ is due to the very small 

1545: %free-volume, hence the small prefactor $\exp[-1/\chi]$ 

1546: %in the constitutive relation. During the ramp, free-volume continues

1547: %to decay, so long as the relaxation is not balance by dilatancy.

1548: %Free-volume starts to increase when,

1549: %$$

1550: %\sigma^2(t)={E_1\over \alpha\,E_0}\exp\left[{1-\kappa\over\chi}\right]

1551: %\quad.

1552: %$$

1553: %Since, $\sigma(t) = \mu\,(\epsilon-\gamma)\simeq \mu\,\dot\epsilon\,(t-t_w)$,

1554: %this occurs at a time $t_y$ such that

1555: %$$

1556: %t_y - t_w \simeq \sqrt{E_1^{1/\kappa}\over \alpha\,E_0} \, t^{-\beta/2}

1557: %\simeq \sqrt{E_1^{1/\kappa}\over \alpha\,E_0} \, t_w^{-\beta/2}

1558: %\quad.

1559: %$$

1560: %The increase of free-volume triggers the plastic flow, further increase of 

1561: %free-volume,\dots, and the process of acceleration occurs of a very small

1562: %time interval. Therefore, the value of the stress that inverts free-volume

1563: %dynamics is almost equal to the yield stress itself:

1564: %$$

1565: %\sigma_y^d\simeq \sqrt{E_1\over \alpha\,E_0} \, t^{-\beta/2}

1566: %\quad.

1567: %$$

1568: 

1569: These transients result from a rather complex interplay between the dynamics

1570: of free-volume and shear transformation. The complete study of this process

1571: may lead to many developments that would divert this work from the main questions

1572: that I would like to address, and I prefer to leave the discussion here.

1573: 

1574:  

1575: \subsection{\label{sec:hopf} Stick-slip instability}

1576: 

1577: The stability analysis of equations~(\ref{eqn:stzdil:lin:1}-\ref{eqn:stzdil:lin:3}) 

1578: and~(\ref{eqn:sigma:0}) shows the existence of a Hopf bifurcation which is responsible 

1579: for oscillatory response to a given strain rate.

1580: This leads to some understanding of stick-slip motion in sheared 

1581: lubricants~\cite{thompson92,yoshizawa93,demirel96a,demirel96b,drummond00,drummond01}

1582: or in granular materials.~\cite{thompson91,nasuno97,nasuno98}

1583: In this work, I mention this instability for the sake of completeness,

1584: but it raises many questions, and will require a more thorough consideration.

1585: %Since this question raises many issues, I prefer to stop the discussion here, 

1586: %and discuss stick-slip motion in another publication.%~\cite{lemaitre02b}

1587: The calculation of the Hopf criterion of bifurcation is given 

1588: in appendix~\ref{app:hopf}.

1589: 

1590: In the isotropic limit, 

1591: the steady plastic deformation is stable iff $\mu>\mu_{\rm hopf}$, with:

1592: $$

1593: \mu_{\rm hopf} = {E_1\over E_0}\,{1-\kappa\over(\kappa+1)^2}\;

1594: \left({\alpha\,\dot\epsilon^2\over E_0\,E_1}\right)^{{\kappa-1\over\kappa+1}}\;

1595: \ln\left[{\alpha\,\dot\epsilon^2\over E_0\,E_1}\right]^2

1596: \quad.

1597: $$

1598: For $\kappa>1$, $\mu_{\rm hopf}<0$: there is no Hopf bifurcation.

1599: For $\kappa<1$, $\mu_{\rm hopf}>0$ on the interval $[0,\dot\epsilon^*]$,

1600: and the steady state becomes unstable iff $\mu<\mu_{\rm hopf}$.

1601: This criterion is represented figure~\ref{fig:hopf:iso}.

1602: For driving strain rates above  $\dot\epsilon^*$, the system looses stability,

1603: because there is no stationary solution for $\chi$.

1604: Below $\dot\epsilon^*$, the Hopf criterion separates a domain of stability,

1605: above $\mu_{\rm hopf}$ where the system reaches steady plastic deformation,

1606: and and domain below this curve where stick-slip motion is achieved.

1607: \begin{figure}

1608: \narrowtext

1609: \begin{center}

1610: \unitlength=0.005\textwidth

1611: \begin{picture}(110,60)(8,-8)

1612: \put(6,40){\makebox(0,0){\large$\mu_{\rm hopf}$}}

1613: \put(68,-13){\makebox(0,0){\large$\dot\epsilon^*$}}

1614: \put(85,-13){\makebox(0,0){\large$\dot\epsilon$}}

1615: \put(10,-13){\resizebox{90\unitlength}{!}{\includegraphics{hopfiso.ps}}}

1616: \end{picture}

1617: \end{center}

1618: \caption{\label{fig:hopf:iso} 

1619: Phase diagram for equations~(\ref{eqn:sigma:0}),~(\ref{eqn:stzdil:iso:1}) 

1620: and~(\ref{eqn:stzdil:iso:2}), and $\kappa<1$.

1621: The solid line denotes the curve $\mu_{\rm hopf}$ below 

1622: which stick-slip motion occurs. $\mu_{\rm hopf}$ vanishes for

1623: $\dot\epsilon=\dot\epsilon^*$ (dashed line)

1624: where steady plastic regime disappears, leading

1625: the system to break-up.

1626: }

1627: \end{figure}

1628: 

1629: The introduction of the variable $\Delta$ in the set of equations~(\ref{eqn:sigma:0}),~(\ref{eqn:stzdil:lin:1}-\ref{eqn:stzdil:lin:3}) 

1630: slightly modifies this picture (see appendix~\ref{app:hopf}).

1631: Two limit curves $\mu_{\rm hopf}^\pm$ emerge, which delimit the unstable 

1632: domain in the plane $(\dot\epsilon,\mu)$. 

1633: These curves are displayed figure~\ref{fig:hopf:lin} along with the previous curves $\mu_{\rm hopf}$ of the isotropic limit.

1634: The unstable domain is a subspace of what it is in the isotropic limit,

1635: and the system present a creeping zone, close to $\dot\epsilon=0$,

1636: when steady sliding is stable for very small values of the strain rate.

1637: \begin{figure}

1638: \narrowtext

1639: \begin{center}

1640: \unitlength = 0.005\textwidth

1641: \begin{picture}(90,90)(0,0)

1642: \put(10,5){\resizebox{70\unitlength}{!}{\includegraphics{hopf.eps}}}

1643: \put(5,75){\makebox(0,0){\large $\mu$}}

1644: \put(70,2){\makebox(0,0){\large $\dot\epsilon$}}

1645: \end{picture}

1646: \end{center}

1647: \caption{\label{fig:hopf:lin}

1648: Phase diagram for equations~(\ref{eqn:sigma:0}),~(\ref{eqn:stzdil:lin:1}-\ref{eqn:stzdil:lin:3}) and $\kappa<1$.

1649: Parameters used are, $E_0=E_1=\alpha=\epsilon_0=\mu_0=1$, and $\kappa=0.8$.

1650: The solid line denotes the Hopf criterion $\mu_{\rm hopf}$ in the isotropic

1651: limit. The dashed lines indicates the Hopf criterion $\mu_{\rm hopf}^\pm$

1652: with the complete set of equations.

1653: }

1654: \end{figure}

1655: 

1656: Some examples of stick-slip motion are displayed figure~\ref{fig:stickslip},

1657: with the set of parameters used in the phase diagram of figure~\ref{fig:hopf:lin}:

1658: $E_0=E_1=\alpha=1$, and $\kappa=0.8$ in the isotropic limit (solid lines)

1659: and $\epsilon_0=\mu_0=1$ for the complete set of equations (dashed lines).

1660: The fast bursts of plastic deformation are accompanied by sudden 

1661: jumps in the free-volume.

1662: $\Delta$ does not play a major role for the definition of the dynamical

1663: yield stress, as seen from the almost identical values of the maximum stress;

1664: $\Delta$, however, strongly influences smaller values of the stress as it shifts 

1665: upwards the unstable fixed point; 

1666: this also modifies the periodicity of stick-slip motion. 

1667: \begin{figure}

1668: \narrowtext

1669: \begin{center}

1670: \unitlength = 0.005\textwidth

1671: \begin{picture}(100,110)(5,0)

1672: \put(10,5){\resizebox{90\unitlength}{!}{\includegraphics{stickslip.eps}}}

1673: \put(5,95){\makebox(0,0){\large $1/\chi$}}

1674: \put(5,50){\makebox(0,0){\large $\sigma$}}

1675: \put(90,2){\makebox(0,0){\large $t$}}

1676: \end{picture}

1677: \end{center}

1678: \caption{ \label{fig:stickslip}

1679: Solid lines: 

1680: Numerical integration of equations~(\ref{eqn:sigma:0}),~(\ref{eqn:stzdil:iso:1}) 

1681: and~(\ref{eqn:stzdil:iso:2}) for a fixed strain rate $\dot\epsilon=0.1$.

1682: Parameters are $E_0=E_1=\alpha=\mu=1$, and $\kappa=0.8$.

1683: Initial value of the free-volume is $\chi=10$.

1684: The regime of steady plastic deformation is unstable and leads to stick-slip motion.

1685: Fast relaxations of the stress result from sudden dilatancy of the material.

1686: Dashed lines: Same for equations~(\ref{eqn:sigma:0}),~(\ref{eqn:stzdil:lin:1}-\ref{eqn:stzdil:lin:3}) with the same parameters and, $\epsilon_0=\mu_0=1$.

1687: }

1688: \end{figure}

1689: 

1690: % ???: bifurcation diagram

1691: %\subsection{Stick-slip motion}

1692: %\subsubsection{Over-damped stick-slip}

1693: %\subsubsection{Hopf analysis}

1694: %The introduction of the equation of motion for $\Delta$ slightly changes this picture.

1695: 

1696: \section{Response to a constant stress}

1697: In a creep test, a material is tested by forcing a constant stress.

1698: Since $\sigma$ is constant, the elastic deformation is $\epsilon=\sigma/\mu$

1699: at all times, and the plastic deformation is determined by the constitutive equations.

1700: 

1701: %The  preparation of the sample is the same as previously:

1702: %until time $t_w$, no constraint is applied to the material

1703: %which undergoes density relaxation. 

1704: %At time $t_w$ a constant stress is suddenly applied, and the elastic deformation

1705: %adapts in a very small time interval while no plastic deformation has yet occurred.

1706: 

1707: \subsection{Linear response}

1708: 

1709: \subsubsection{\label{sec:lin:iso} Isotropic limit}

1710: A small constant stress $\sigma$ is applied at time $t=t_w$.

1711: The plastic deformation of the material is determined 

1712: by equations~(\ref{eqn:stzdil:iso:1}-\ref{eqn:stzdil:iso:2}),

1713: for a fixed $\sigma$.

1714: At large times, the rate of plastic deformation reads

1715: $$

1716: \dot\epsilon=\dot\gamma \simeq E_0\, \left({E_1\,t\over \kappa}\right)^{-1/\kappa}\sigma

1717: \quad.

1718: $$

1719: This lead to a power law variation of the compliance, $J=\gamma/\sigma$:

1720: $$

1721: J(t) \propto A\;\left(t^\beta-t_w^\beta\right)

1722: \quad.

1723: $$

1724: If $\kappa>1$, or $\beta>0$, the compliance diverges at long times;

1725: if $\kappa<1$, or $\beta<0$, the compliance saturates at long times: 

1726: $$

1727: J(\infty) = |A|\,t_w^{-|\beta|}

1728: \quad,

1729: $$

1730: the system jams. This jamming is purely entropic, and does not result

1731: from a structural anisotropy of the force network.

1732: Note that the asymptotic value of the compliance 

1733: decays with the increasing age of the sample.

1734: 

1735: The validity of the linear approximation must be checked

1736: by considering the evolution of the work of plastic deformation:

1737: $$

1738: \sigma\,\dot\gamma\propto t^{-{1/\kappa}}

1739: \quad;

1740: $$

1741: the linear approximation is justified so long as this term is negligible,

1742: when compared to the first term in free-volume relaxation~(\ref{eqn:stzdil:iso:2}),

1743: which decays as $\kappa/t$.

1744: When $\kappa<1$, the work $\sigma\,\dot\gamma$ decays faster than $\kappa/t$,

1745: and therefore, the linear approximation is valid at all times;

1746: when $\kappa>1$, however, this approximation fails when:

1747: $$

1748: \alpha\,E_0\,\left({E_1\,t\over\kappa}\right)^{-1/\kappa}\sigma^2\sim \kappa/t

1749: \quad,

1750: $$

1751: or,

1752: $$

1753: t \sim t^{\rm n.l.} = \kappa\,\left({E_1^{1/\kappa}\over \alpha\,E_0\,\sigma^2}\right)^{\kappa\over\kappa-1}

1754: \quad.

1755: $$

1756: The time $t^{\rm n.l.}$ is a typical time after which the linear 

1757: approximation fails; this time diverges when $\sigma\to0$.

1758: It will appear clearer in the following that, when $\kappa>1$,

1759: free-volume presents a non-vanishing stationary value for any small applied stress.

1760: This asymptotic value is determined by the balance between density relaxation

1761: and shear induced dilatancy. The failure of the linear approximation, for $\kappa>1$,

1762: is due to the 

1763: existence of a non-zero stationary value of $\chi$ for any applied stress.

1764: 

1765: \subsubsection{\label{sec:lin:lin} General case}

1766: 

1767: 

1768: In the general case, the constitutive equations 

1769: are~(\ref{eqn:stzdil:lin:1}-\ref{eqn:stzdil:lin:3}).

1770: The constant applied stress deforms the material elastically,

1771: and at time $t_w^+$, no plastic rearrangement has yet occurred. 

1772: Note that the initial value of the state variable $\Delta$ 

1773: depends on the preparation of the sample. 

1774: In the experiments by Cloitre {\it et al},

1775: for example, the sample is prepared by applying a strong shear stress,

1776: above the yield stress during a rather long time.~\cite{cloitre00}

1777: In this case, the initial value of $\Delta$ is expected

1778: to be non-vanishing, and this leads to their observation of strain recovery.

1779: For the sake of generality, I will consider that the initial value of $\Delta$

1780: is non-vanishing: $\Delta(t_w^+)=\Delta_0$.

1781: In this section, I will assume that this value of $\Delta$ is small enough

1782: so that the linear approximation is still valid.

1783: When plastic rearrangements occur in the material,

1784: they induce a structuration of the material, and 

1785: the variable $\Delta$ starts to evolve as $\dot\Delta\simeq\dot\gamma/\epsilon_0$.

1786: 

1787: The plastic rate of deformation evolves as,

1788: $$

1789: \dot\gamma \simeq E_0\, \left({E_1\,t\over\kappa}\right)^{-1/\kappa}(\sigma-\Delta)

1790: \quad,

1791: $$

1792: and $\Delta$ as, 

1793: $$

1794: \dot\Delta \simeq {E_0\over\epsilon_0}\,\left({E_1\,t\over\kappa}\right)^{-1/\kappa}

1795: \,(\sigma-\Delta)

1796: \quad.

1797: $$

1798: 

1799: From this equation, the relaxation of the variable $\sigma-\Delta$ is easily

1800: obtained:

1801: $$

1802: \sigma-\Delta(t) \simeq (\sigma-\Delta_0)\,\exp\left[{A''\,(t_w^\beta-t^\beta)}\right]

1803: \quad,

1804: $$

1805: with

1806: $$

1807: A'' = {E_0\over\beta \epsilon_0}{\left({E_1\over\kappa}\right)^{-1/\kappa}}

1808: \quad.

1809: $$

1810: When $\kappa>1$, $\sigma-\Delta$ undergoes a KWW relaxation towards 0:

1811: the material relaxes towards a jammed state in which the stress is supported

1812: by structural anisotropy.

1813: When $\kappa<1$, $\sigma-\Delta$ saturates to a non-vanishing value: 

1814: at long times, a fraction of the stress is supported by 

1815: the bias between arrangements, measured by $\Delta$, but another fraction 

1816: is supported by the loss of ergodicity resulting from the

1817: freezing of elementary shear processes.

1818: 

1819: From equation $\dot\Delta\simeq\dot\gamma/\epsilon_0$, the compliance reads,

1820: $$

1821: J(t) \simeq \epsilon_0\,\left(1-{\Delta_0\over\sigma}\right)\,

1822: (1 - \exp\left[{A''\,(t_w^\beta-t^\beta)}\right])

1823: \quad.

1824: $$

1825: The deformation relaxes to a constant value, which depends on the preparation of

1826: the sample; the relaxation is either a KWW or a power law

1827: relaxation depending on $\kappa$.

1828: Note that the power law variation of the compliance that was found 

1829: in the isotropic limit is recovered when $A''\to0$;

1830: for finite $A''$, this power law behavior shows up at short times, 

1831: when $\Delta$ is still far from saturation.

1832: %The crossover between those regimes occurs when $t^\beta\simeq t_w^\beta-1/A''$.

1833: 

1834: \subsection{Jamming}

1835: 

1836: In this section I start by considering the complete set of equations~(\ref{eqn:stzdil:lin:1}-\ref{eqn:stzdil:lin:3}), 

1837: and in particular, I recall how the dynamics of $\Delta$ lead to jamming.

1838: Then I show how this leads to an effective equation for free-volume which

1839: is valid in the isotropic limit or in the general case.

1840: 

1841: A constant stress $\sigma$ is applied. 

1842: The dynamics of $\Delta$ is governed by the following equation:

1843: \begin{equation}

1844: \dot\Delta = {E_0\over\epsilon_0}\,\exp\left[-{1\over\chi}\right]\,

1845: \left(\sigma-\Delta\right)\;\left(1-\mu_0\,\sigma\,\Delta\right)

1846: \quad.

1847: \label{eqn:deltadyn}

1848: \end{equation}

1849: Since the normalized free-volume $\chi$ enters this equation only through

1850: a positive factor, the time-dependent value of $\chi$ does not change 

1851: the stability analysis performed at any fixed free-volume.

1852: The bifurcation diagram for  $\Delta$ is drawn figure~\ref{fig:delta}.

1853: \begin{figure}

1854: \narrowtext

1855: \begin{center}

1856: \unitlength=0.005\textwidth

1857: \begin{picture}(110,60)(8,-8)

1858: \put(12,40){\makebox(0,0){\large$\Delta$}}

1859: \put(85,-13){\makebox(0,0){\large$\sigma$}}

1860: \put(10,-13){\resizebox{90\unitlength}{!}{\includegraphics{delta.ps}}}

1861: \end{picture}

1862: \end{center}

1863: \caption{\label{fig:delta}

1864: Bifurcation for the dynamics of the variable $\Delta$ as defined by 

1865: equation~(\protect\ref{eqn:deltadyn}). The intersection between the two branches 

1866: of solutions define the yield stress $\sigma_y$. 

1867: For $\sigma<\sigma_y$, the stable solution is the jammed state, while it is the 

1868: plastic flow for $\sigma>\sigma_y$.~\protect\cite{falk98}

1869: }

1870: \end{figure}

1871: 

1872: The dynamics of the variable $\Delta$ leads to jamming for any stress smaller

1873: than the yield stress~\cite{falk98}

1874: $$

1875: \sigma_y = {1\over\sqrt{\mu_0}}

1876: \quad,

1877: $$

1878: in which case, $\Delta = \sigma$ and $\dot\epsilon=0$.

1879: Above the yield stress, the steady shear flow becomes a stable solution, 

1880: in which case,

1881: $$

1882: \Delta = {1\over\mu_0\,\sigma}

1883: \quad,

1884: $$

1885: and

1886: \begin{equation}

1887: \dot\epsilon = 

1888: {E_0}\,\exp\left[-{1\over\chi}\right]\,\left(\sigma-{1\over\mu_0\,\sigma}\right)

1889: \quad.

1890: \label{eqn:eps:creep}

1891: \end{equation}

1892: 

1893: If $\sigma<\sigma_y$, the systems jams, 

1894: and the second term in the equation of motion for the free-volume vanishes.

1895: Free-volume continues to relax logarithmically in time while the system does not deform.

1896: 

1897: When $\sigma>\sigma_y$, the system presents a non-zero plastic flow,

1898: and the evolution of free-volume is governed by the following asymptotic equation:

1899: \begin{equation}

1900: \dot\chi = -E_1 \exp\left[{-{\kappa\over\chi}}\right]

1901: +\alpha {E_0}\,\exp\left[-{1\over\chi}\right]\,\left(\sigma^2-{1\over\mu_0}\right)

1902: \quad.

1903: \label{eqn:chi:creep}

1904: \end{equation}

1905: At the level of this equation, the jamming resulting from STZ's 

1906: shows up only through the term $1/\mu_0$ in equation~(\ref{eqn:chi:creep}), 

1907: which shifts the square of the applied stress.

1908: The stability analysis pertains essentially to the isotropic limit.

1909: From this point on, I will consider the response to a constant stress 

1910: as governed by equations~(\ref{eqn:eps:creep}) and (\ref{eqn:chi:creep})

1911: (which amounts to neglect the influence of $\Delta$ during transients).

1912: 

1913: \subsection{Brittleness and ductility}

1914: I now study the free-volume dynamics which result from equation~(\ref{eqn:chi:creep}).

1915: The material dilates iff $\sigma>\sigma_y^*$, with:

1916: $$

1917: \sigma_y^* = \sqrt{{1\over\mu_0} +

1918: {E_1\over\alpha \,E_0} \exp\left[{{1-\kappa\over\chi}}\right]}

1919: $$

1920: and contracts otherwise. The resulting bifurcation diagram for $\chi$

1921: is drawn figure~\ref{fig:chi:1} and~\ref{fig:chi:2} for $\kappa>1$ and $\kappa<1$

1922: respectively.

1923: 

1924: \subsubsection{Soft glasses}

1925: 

1926: For $\kappa>1$, free-volume dynamics admit a stationary solution for

1927: any $\sigma\in[\sigma_y,\sigma^*]$, with,

1928: $$

1929: \sigma^* = \sqrt{{1\over\mu_0} + {E_1\over\alpha \,E_0}}

1930: \quad.

1931: $$

1932: For values of the applied stress in the interval $[\sigma_y,\sigma^*]$, 

1933: the material undergoes steady plastic deformation: it is ductile.

1934: The shear deformation reads,

1935: $$

1936: \dot\epsilon = 

1937: {E_1\over\alpha}\;

1938: \left(\alpha\,E_0\over E_1\right)^{{\kappa\over\kappa-1}}

1939: \;{1\over\sigma}\;\left(\sigma^2-{1\over\mu_0}\right)^{{\kappa\over\kappa-1}}

1940: \quad,

1941: $$

1942: and the strain rate behaves as a power of the stress for large stresses.

1943: When the applied stress varies in the interval, $\sigma\in[\sigma_y,\sigma^*]$,

1944: the strain rate can display a whole interval of values

1945: $$

1946: \dot\epsilon\in\left[0,E_0\,\left(\sigma^*-{1\over\mu_0\,\sigma^*}\right)\right]

1947: \quad.

1948: $$

1949: For larger values of $\sigma$, above $\sigma^*$, free-volume diverges, leading

1950: to the break-up of the material.

1951: 

1952: \begin{figure}

1953: \narrowtext

1954: \begin{center}

1955: \unitlength=0.005\textwidth

1956: \begin{picture}(110,60)(8,-8)

1957: \put(10,48){\makebox(0,0){\large$\sigma$}}

1958: \put(13,40){\makebox(0,0){\large$\sigma^*$}}

1959: \put(13,-1){\makebox(0,0){\large${1\over\mu_0}$}}

1960: \put(85,-13){\makebox(0,0){\large$\chi$}}

1961: \put(10,-13){\resizebox{90\unitlength}{!}{\includegraphics{chi1.ps}}}

1962: \end{picture}

1963: \end{center}

1964: \caption{Bifurcation diagram for the free-volume dynamics defined by 

1965: equation~(\protect\ref{eqn:chi:creep}), for $\kappa>1$. 

1966: For any value of the stress

1967: in the interval $[\sigma_y,\sigma^*]$, there exist a stationary value of $\chi$. 

1968: The applied stress stops aging, and the material undergoes steady 

1969: plastic deformation.

1970: Above $\sigma^*$, free-volume diverges leading to the break-up of the material.

1971: }

1972: \label{fig:chi:1}

1973: \end{figure}

1974: 

1975: The resulting dynamics for the compliance and for the free-volume are presented figure~\ref{fig:compl:soft} for different values of the driving stress $\sigma$.

1976: The first part of the dynamics is determined by the initial value of 

1977: free-volume and the corresponding timescale.

1978: As easily seen on the bifurcation diagram~(\ref{fig:chi:1}), for small stresses, 

1979: an initially large free-volume decays, and slows down the plastic deformation. 

1980: This results in a plateau for the dynamics of the compliance; a larger times, free-volume converges 

1981: towards a constant value, and the dynamics of $J$ resumes a steady increase.

1982: 

1983: These curves are to compare with measurements of particle diffusion:

1984: the particle diffusion in a dense material is expected to be dominated 

1985: by collective processes; the diffusive constant is then $\propto\exp[-1/\chi]$, 

1986: and the average motion of a particle in a glassy medium is essentially

1987: governed by the same equation as $J$. 

1988: The features displayed of figure~\ref{fig:compl:soft}

1989: are amazingly similar to the curves found in 

1990: several recent experiments.~\cite{megen98,cloitre00}

1991: \begin{figure}

1992: \narrowtext

1993: \begin{center}

1994: \unitlength = 0.005\textwidth

1995: \begin{picture}(100,110)(5,0)

1996: \put(10,5){\resizebox{90\unitlength}{!}{\includegraphics{compl-soft.eps}}}

1997: \put(5,95){\makebox(0,0){\large $1/\chi$}}

1998: \put(5,50){\makebox(0,0){\large $\log_{10}(J)$}}

1999: \put(90,2){\makebox(0,0){\large $\log_{10}(t)$}}

2000: \end{picture}

2001: \end{center}

2002: \caption{

2003: Numerical integration of equations~(\ref{eqn:stzdil:iso:1}) 

2004: and~(\ref{eqn:stzdil:iso:2}) for constant stresses $\sigma$, and for $\kappa>1$.

2005: Parameters are $E_0=E_1=\alpha=\mu=1$, and $\kappa=2$;

2006: for these parameter, $\sigma^*=1$.

2007: Initial value of the free-volume is $\chi=100$ and $t_w=0$.

2008: Top: $1/\chi$ as a function of time for 

2009: $\sigma = 0.2, 0.4, 0.6, 0.8, 1$ from top to bottom; for $\sigma=1$, and for

2010: all values above, free-volume diverges.

2011: Bottom: $\log_{10}(J)$ as a function of time for $\sigma$ 

2012: increasing from bottom to top.

2013: }

2014: \label{fig:compl:soft}

2015: \end{figure}

2016: 

2017: If free-volume diverges, just before break-up, 

2018: the current equations indicate

2019: that the strain rate should display a behavior of the form,

2020: $$

2021: \dot\epsilon = E_0\,\left(\sigma-{1\over\mu_0\,\sigma}\right)

2022: \quad.

2023: $$

2024: However, this expression must be taken with caution,

2025: since the strongly out-of-equilibrium behavior of a material during break-up, 

2026: when $\chi$ is large, is beyond the scope of the current theory.

2027: In particular, assumption has been made that the distribution 

2028: of voids is close to an equilibrium state.

2029: During break-up, the distribution of free-volume is expected to display strong

2030: heterogeneities leading to the nucleation of fractures.

2031: 

2032: \subsubsection{Hard glasses}

2033: 

2034: For $\kappa<1$, and for any $\sigma<\sigma^*$, $\dot\chi<0$: 

2035: free-volume relaxes to 0. In this case, the applied stress does not stop aging,

2036: shear deformation is vanishingly small: the material creeps.

2037: At long time, the linear term dominates free-volume relaxation: the calculation

2038: of linear response that has been presented in section~\ref{sec:lin:iso} is valid,

2039: and indicates that compliance saturates: the system jams.

2040: In general, part of this jamming is supported by the structuration of the material,

2041: which lowers the value of the stress that enters equation~(\ref{eqn:chi:creep});

2042: the remaining stress is supported by entropic freezing of the rearrangement dynamics.

2043: 

2044: The sign of free-volume relaxation can be inverted only for larger stresses, 

2045: above $\sigma_y^*(\chi)$; it this case, the response depends sensitively on 

2046: the value of $\sigma$ and on the state of the material at the time $t_w$

2047: when stress is applied:

2048: a larger stress is required to break an older material.

2049: The stationary state is unstable: when $\sigma$ is large enough 

2050: to invert the sign of free-volume dynamics, the material breaks up.

2051: In this case, there are two types of response: creep for low stress, and break-up

2052: for large values of the stress. The material is brittle.

2053: \begin{figure}

2054: \narrowtext

2055: \begin{center}

2056: \unitlength=0.005\textwidth

2057: \begin{picture}(110,60)(8,-8)

2058: \put(10,40){\makebox(0,0){\large$\sigma$}}

2059: \put(13,20){\makebox(0,0){\large$\sigma^*$}}

2060: \put(85,-13){\makebox(0,0){\large$\chi$}}

2061: \put(10,-13){\resizebox{90\unitlength}{!}{\includegraphics{chi2.ps}}}

2062: \end{picture}

2063: \end{center}

2064: \caption{Bifurcation diagram for the free-volume dynamics defined by 

2065: equation~(\protect\ref{eqn:chi:creep}), for $\kappa<1$. 

2066: For all $\sigma<\sigma^*$, free-volume continues relaxation towards 0;

2067: shear deformation occurs at a vanishingly small rate; the material creeps.

2068: The sign of free-volume relaxation can be changed only by applying a stress at least

2069: larger than $\sigma^*$; the required stress depends on the age of the material.

2070: The steady solution is always unstable: large stresses lead to break up.}

2071: \label{fig:chi:2}

2072: \end{figure}

2073: 

2074: The resulting dynamics for free-volume and for the compliance are 

2075: displayed figure~\ref{fig:compl:hard}. The saturation of the compliance

2076: is clearly seen for small stresses while the divergence occurs at 

2077: a value at $\sigma=\sigma_y^*(\chi)$.

2078: The shape of these curves changes drastically in a very small interval of $\sigma$

2079: around $\sigma_y^*(\chi)$ because of long transient to escape from 

2080: the unstable fixed point. These features are commonly observed in creep tests.~\cite{hassan95,larson99}

2081: \begin{figure}

2082: \narrowtext

2083: \begin{center}

2084: \unitlength = 0.005\textwidth

2085: \begin{picture}(100,110)(5,0)

2086: \put(10,5){\resizebox{90\unitlength}{!}{\includegraphics{compl-hard.eps}}}

2087: \put(5,95){\makebox(0,0){\large $1/\chi$}}

2088: \put(5,50){\makebox(0,0){\large $J$}}

2089: \put(90,2){\makebox(0,0){\large $t$}}

2090: \end{picture}

2091: \end{center}

2092: \caption{

2093: Numerical integration of equations~(\ref{eqn:stzdil:iso:1}) 

2094: and~(\ref{eqn:stzdil:iso:2}) for constant stresses $\sigma$.

2095: Parameters are $E_0=E_1=\alpha=\mu=1$, and $\kappa=0.8$.

2096: Initial value of the free-volume is $\chi=1$ and $t_w=0$.

2097: For these values of the parameters, and of initial conditions, 

2098: the critical value of $\sigma$ where the free-volume diverges is, 

2099: $\sigma_y^*\simeq 1.105$.

2100: Top: $1/\chi$ as a function of time for 

2101: $\sigma = 0.5, 1, 1.05, 1.075, 1.1, 1.15, 1.2$ from top to bottom.

2102: Bottom: $J$ as a function of time for $\sigma$ increasing from bottom to top.

2103: }

2104: \label{fig:compl:hard}

2105: \end{figure}

2106: 

2107: 

2108: \subsubsection{Jamming and freezing}

2109: 

2110: Materials with $\kappa>1$ are liable to display ductile behavior,

2111: but can break; those with $\kappa<1$ either creep or break-up, 

2112: and could be classified as unconditionally brittle. 

2113: The study of STZ equations one a one-dimensional

2114: elasto-plastic decohesion has shown the importance of dynamics in ductile to brittle

2115: transition.~\cite{lobkovsky98}

2116: The current analysis slightly contradicts those results,

2117: and indicates that two major types of materials

2118: can be identified, which display different qualitative behavior: 

2119: this is in agreement

2120: with the common idea that some materials are ductile, some other brittle.

2121: 

2122: In fact, for hard materials ($\kappa<1$), the existence of a strong entropic crisis

2123: precludes dynamics of rearrangements to come into play: intrinsic

2124: properties of the material completely dominate the dynamical process of deformation.

2125: The existence of a dynamically driven transition between ductile and brittle

2126: behavior is still expected for soft materials ($\kappa>1$) for which the dynamics

2127: or arrangements can fully play its role.

2128: The interplay between dynamics and intrinsic glassy properties of material, 

2129: as displayed in the current approach, seems a promising perspective.

2130: 

2131: As far as fracture mechanics is concerned, this picture indicates that

2132: during the breaking of a hard glass, the motion of molecules in a material

2133: should be essentially localized around the surface of a crack,

2134: where there is some free-volume,

2135: whereas in the bulk, plastic deformation can occur only in places where a very large

2136: stress in concentrated. This is apparently related to a 

2137: theoretical approach of fracture mechanics,

2138: recently introduced by Brener and Spatschek,~\cite{brener02} 

2139: where the dynamics of the fracture is controlled by diffusion 

2140: of molecules along the surface.

2141: For soft glassy materials, on the contrary, the propagation

2142: of a fracture is expected to induce more easily observable 

2143: plastic deformation in the bulk of the material.

2144: 

2145: %In the case of hard materials, ($\kappa<1$), the current model predicts that

2146: %there is either creep motion with a very small $\dot\epsilon$, depending

2147: %on the age of the system, or the system breaks up, and the rate of deformation achieves

2148: %the preceding value.

2149: 

2150: %\subsubsection{Creep test}

2151: %I am now in position to draw a complete picture of a creep test.

2152: %This is to compare {\it e.g.} with the experimental results obtained

2153: %by Hassan and Boyce on PMMA.~\cite{hassan95}

2154: 

2155: \section{Conclusion}

2156: 

2157: The current work relies on the introduction of dynamical equations for 

2158: the intensive quantity associated with entropic fluctuations 

2159: of molecular configurations.

2160: In this work, this quantity is identified to free-volume,

2161: but other physical pictures could be given.

2162: For example, free-volume is related to the average number of contacts per molecule;

2163: in a foam, the deformation of the bubbles can be accounted for a effective 

2164: free-volume that is determined by the packing of effective overlapping

2165: particles as those used in the simulations by Durian,~\cite{durian95,durian97} 

2166: even though the relative proportions

2167: of a fluid mixture is kept constant; in a colloidal suspension, the existence

2168: of electrostatic screening contributes to an effective size of the particle

2169: that depends on the packing structure. 

2170: In fact, the assumption made in this work is rather general,

2171: and many situation can be expected to lead to a dynamically 

2172: driven effective free-volume; this may explain the ubiquity of glassy behavior.

2173: 

2174: %The current approach might be seen as the opposite point of view 

2175: %of Berthier {\it et al}~\cite{berthier00}

2176: %who introduced two temperatures to account for multiple timescales.

2177: %But here, the analysis comes out from a real-space picture of rearrangements:

2178: %there is no reasons whatsoever to assume that different types of rearrangements

2179: %should be governed by a unique barrier height and, 

2180: %the multiplicity of timescales is directly provided by this simple remark,

2181: %and quite minimal assumption.

2182: 

2183: The theoretical approach to glassy materials can be summarized as:

2184: how ergodicity breaks? or what ergodicity breaking is sufficient to account

2185: for slow modes of relaxation. The introduction of a whole distribution of timescales

2186: is a very general way to account for non-ergodicity.

2187: Here I assume that a much weaker ergodicity breaking is sufficient,

2188: and this is directly supported by the observation that fluctuation-dissipation

2189: theorem holds in a weak (but, not so weak) form.

2190: This indicates that the underlying

2191: distributions of state variables should not (in fact) be too weird,

2192: and that an effective width accounts for its essential features.

2193: However, since the system is out-of-equilibrium, 

2194: this width must, at least, become a dynamical quantity.

2195: The equation proposed appear as a very minimal first order closure 

2196: of the dynamics of an underlying distribution of voids.

2197: 

2198: The set of curves displayed in figure~\ref{fig:relax3} 

2199: is very reminiscent of the temperature

2200: dependency of relaxation spectra.~\cite{larson99} Obviously, the parameters

2201: of the theory should depend on temperature and other thermodynamical parameters, 

2202: but here I purposely do not intend to address this question. 

2203: There are (at least) two reasons for that: firstly, because

2204: glassy behavior is observed in many systems, 

2205: in particular in suspensions or in granular materials, 

2206: where thermodynamical temperature is not a relevant parameter.

2207: Secondly, because it seems to me more interesting to extract all possible conclusions

2208: from the simple equations proposed without trying to incorporate them into a more

2209: specific setup. I do think that an effective dynamically driven free-volume

2210: should emerge from microscopic dynamics under minimal assumptions. 

2211: Therefore, it seems to me important to try to get

2212: some fundamental understanding of the consequences of such assumptions for a fixed

2213: set of parameters, without complicating the discussion.

2214: The dependency of the parameters of this theory on thermodynamic quantities

2215: is definitely an important issue to address in the future, but it first requires

2216: to have some intuition about the outcome of those equations. 

2217: This work is intended to provide a grid of possible behavior displayed

2218: by simple equations, and I hope it can be helpful for further developments.

2219: 

2220: Finally, one major interest of the current work is, in my opinion, that is gives 

2221: a unified picture where different mechanisms for jamming coexist.

2222: This coexistence of jamming with non-linear rheology has been observed 

2223: experimentally,

2224: but could not be captured by earlier model of glassy rheology.

2225: Moreover, it opens the way towards a description of glassy materials 

2226: where thermodynamics (that determine the parameters of constitutive equations)

2227: and dynamics coexists. 

2228: %Here, jamming results either from a phase transition,

2229: %or from the dynamics of a structural arrangements.

2230: %These simple equations also permit to clarify 

2231: %the relative importance of solid-like or liquid-like features 

2232: %and the importance of force and time scales. 

2233: %Soft glassy materials are thus expected

2234: %to present some features of a yield stress and some of a power law fluid.

2235: %The observed behavior of a given material is expected to depend drastically 

2236: %on the position of the experimental accessible window of forcing parameters 

2237: %relatively to the internal characteristics of the material.

2238: 

2239: \acknowledgements

2240: 

2241: This work has primarily benefited from the support and encouragements of

2242: Jean Carlson and Jim Langer who received me among their people

2243: in the busy heights of UCSB;

2244: I acknowledge the opportunity they gave me to pursue my ideas,

2245: and their continuous interest for my work.

2246: I have enjoyed useful discussions with them, and also with others:

2247: Ralph Archuleta, Pascal Favreau, Anthony Foglia, Delphine Gourdon,

2248: Jacob Israelachvili, Daniel Lavall\'ee, Jean-Christophe Nave.

2249: I also would like to thank Daniel Bonn for giving me early versions of his papers,

2250: which were helpful in the writing of this manuscript.

2251: 

2252: This work was supported by the W. M. Keck Foundation,

2253: and the NSF Grant No. DMR-9813752,

2254: and EPRI/DoD through the Program on Interactive Complex Networks.

2255: 

2256: \thanks

2257: 

2258: \appendix

2259: \section{\label{app:hopf} Hopf bifurcation at constant strain rate}

2260: 

2261: \subsection{Isotropic limit}

2262: In order to carry out the stability analysis of the 

2263: system~(\ref{eqn:sigma:0}),~(\ref{eqn:stzdil:iso:1}),~(\ref{eqn:stzdil:iso:2}),

2264: it is convenient to introduce

2265: the variable,

2266: $$

2267: \phi = \exp\left[-{1\over\chi}\right]

2268: \quad.

2269: $$

2270: 

2271: The system~(\ref{eqn:sigma:0}),~(\ref{eqn:stzdil:iso:1}),~(\ref{eqn:stzdil:iso:2})

2272: reduces to two equations for $\sigma$ and $\chi$,

2273: \begin{eqnarray*}

2274: \dot\sigma &=& \mu\,\dot\epsilon-\mu\,E_0\,\exp\left[-{1\over\chi}\right]\sigma\\

2275: \dot\chi &=& -E_1 \exp\left[-{\kappa\over\chi}\right] + \alpha\,E_0\,\exp\left[-{1\over\chi}\right]\,\sigma^2

2276: \end{eqnarray*}

2277: and the jacobian of this dynamical system reads:

2278: $$

2279: \left(

2280: \matrix{

2281: -\mu\,E_0\,\phi  & 

2282: -\mu\,E_0\,\sigma\,\phi\,\log[\phi]^2\cr

2283: 2\,E_0\,\alpha\,\sigma\,\phi& 

2284: \left(-\kappa\,E_1\phi^\kappa + E_0\,\alpha \,{\sigma }^2\,\phi\right)\log[\phi]^2\cr

2285: }

2286: \right)

2287: \quad.

2288: $$

2289: 

2290: In the steady state, the variable $\sigma$ verifies,

2291: $$

2292: \sigma^2 = {E_1\over\alpha\,E_0}\,\phi^{\kappa-1}

2293: $$

2294: and $\phi$ takes on the value,

2295: $$

2296: \phi_{\rm s}=\left({\alpha\,\dot\epsilon^2\over E_0\,E_1}\right)^{1\over\kappa+1}

2297: \quad.

2298: $$

2299: For all $\kappa>0$, $\phi_{\rm s}$ is a strictly increasing function of $\dot\epsilon$.

2300: In terms of $\phi_{\rm s}(\dot\epsilon)$, the eigenvalues $\lambda$ of the jacobian verify:

2301: \begin{eqnarray*}

2302: \lambda^2 +

2303: \lambda\;\left(

2304: \mu\,E_0\,\phi_{\rm s}

2305: + {E_1\,(\kappa-1)}\,\phi_{\rm s}^\kappa

2306: \;\ln\left[\phi_{\rm s}\right]^2

2307: \right)

2308: &+&\\

2309: {\mu\,E_0\,E_1\,(\kappa+1)}\;\phi_{\rm s}^{\kappa+1}\,\ln\left[\phi_{\rm s}\right]^2

2310: &=&0

2311: \quad,

2312: \end{eqnarray*}

2313: and the Hopf bifurcation is determined by a critical value of $\mu$ as a function

2314: of $\phi_{\rm s}$:

2315: $$

2316: \mu_{\rm hopf} = {E_1\over E_0}\,(1-\kappa)\,\phi_{\rm s}^{\kappa-1}\,\ln[\phi_{\rm s}]^2

2317: $$

2318: or,

2319: $$

2320: \mu_{\rm hopf} = {E_1\over E_0}\,{1-\kappa\over(\kappa+1)^2}\;

2321: \left({\alpha\,\dot\epsilon^2\over E_0\,E_1}\right)^{{\kappa-1\over\kappa+1}}\;

2322: \ln\left[{\alpha\,\dot\epsilon^2\over E_0\,E_1}\right]^2

2323: \quad.

2324: $$

2325: 

2326: \subsection{General case}

2327: In the general case, the system~(\ref{eqn:sigma:0}),~(\ref{eqn:stzdil:lin:1}-\ref{eqn:stzdil:lin:3})

2328: involves one additional variable ($\Delta$), and the bifurcation analysis is somewhat more complicated.

2329: In the steady state, the relation between strain rate and free-volume can be rewritten as,

2330: $$

2331: \dot\epsilon = {

2332: \sqrt{E_0\,\mu_0}\,E_1\;\phi^{{\kappa/2}}

2333: \over

2334: \sqrt{\alpha}\;

2335: \sqrt{E_0\alpha\,\phi^{-\kappa} 

2336: 		    + E_1\mu_0\,\phi^{-1}}

2337: }

2338: \quad.

2339: $$

2340: Since this expression is an strictly increasing function of $\phi$, 

2341: it is possible to carry out the analysis in terms of $\phi=\phi_{\rm s}$ as an effective

2342: parameter. When the strain rate varies between $0$ and $\dot\epsilon^*$,

2343: $\phi$ varies on the interval $[0,1]$. The values above $1$ are meaningless

2344: because there is no solution to the equations of motion.

2345: 

2346: The Hopf criterion can be written in the form, 

2347: $$

2348: {\cal H}=A\,\mu^2+B\,\mu+C=0

2349: $$

2350: %\begin{widetext}

2351: with:

2352: \begin{eqnarray*}

2353: A &=&  {E_0}^2\,\alpha \,\epsilon_0\,{\phi }^2\,

2354: \Big( \mu_0\,\left( 2\,E_0\,\alpha \,\phi  + 

2355:         E_1\,\mu_0\,{\phi }^{\kappa } \right)  \Big.\\

2356: &+& \Big.

2357:      \alpha \,\epsilon_0\,\left( E_0\,\alpha \,\kappa \,\phi  + 

2358:         E_1\,\left( 1 + \kappa  \right) \,\mu_0\,{\phi }^{\kappa } \right) \,

2359:       {\log (\phi )}^2 \Big)\\

2360: B &=& E_0\,E_1\,{\phi }^{1 + \kappa }\,

2361:    \Big( {\mu_0}^2\,

2362: \left( 2\,E_0\,\alpha \,\phi+ E_1\,\mu_0\,{\phi }^{\kappa } \right)

2363: + \alpha \,\epsilon_0\,\mu_0\,\Big.\\

2364: \Big.&&\times \left( 

2365: 	E_0\,\alpha \,\left( 2\,\kappa-3  \right) \,\phi +

2366:  	2\,E_1\,\left( \kappa -1 \right) \,\mu_0\,{\phi }^{\kappa } 

2367: 	\right)\, {\log (\phi )}^2 \Big.\\

2368: \Big.&+&

2369: {\alpha }^2\,{\epsilon_0}^2\,\left( \kappa -1  \right) \,

2370:       	\left( E_0\,\alpha \,\kappa \,\phi  + 

2371:         E_1\,\left( 1 + \kappa  \right) \,\mu_0\,{\phi }^{\kappa } \right) \,

2372:       {\log (\phi )}^4 \Big)\\

2373: C &=& {E_1}^2\,\left(\kappa  -1  \right) \,\mu_0\,{\phi }^{2\,\kappa }\,

2374:    \left( E_0\,\alpha \,\phi  + E_1\,\mu_0\,{\phi }^{\kappa } \right) \,

2375:    {\log (\phi )}^2\, \\

2376: &&\times\left( \mu_0 +  \alpha \,\epsilon_0\,\left(\kappa -1 \right) \,{\log (\phi )}^2 \right)

2377: \end{eqnarray*}

2378: %\end{widetext}

2379: The quantity ${\cal H}$ has the opposite sign of the real part 

2380: of the complex eigenvalues of the jacobian: the steady state is stable,

2381: iff ${\cal H}>0$.

2382: 

2383: For $\kappa>1$, and for any $\phi\in[0,1]$, $C$ is positive: the two values of $\mu$, 

2384: solutions of ${\cal H}=0$ have the same sign. This sign can change only

2385: at a point where $C$ vanishes, that is $\phi=0$ or $1$. The sign shared by these

2386: solutions on the interval $\phi\in[0,1]$ can then be evaluated by taking the limit

2387: value of the coefficient $B$ close to 0 or 1: it is positive.

2388: Therefore, the equation ${\cal H}=0$ has no real positive solution for $\kappa>1$:

2389: there is no Hopf instability. It can be checked that there is no other instability,

2390: therefore, for $\kappa>1$, the system is stable.

2391: 

2392: 

2393: I am now considering the case when $\kappa<1$.

2394: To get a first idea of the phase diagram, in the parameter space $\{\mu,\phi\}$,

2395: the quantity ${\cal H}$ can be evaluated on the line $\mu=0$: ${\cal H}(\mu=0,\phi)=C$.

2396: On this line, the positiveness of ${\cal H}$ is then equivalent to,

2397: $$

2398: {\log (\phi )}^2 > {\mu_0\over\alpha \,\epsilon_0\,\left(1-\kappa\right) }

2399: \quad,

2400: $$

2401: which requires that:

2402: $$

2403: \phi < \phi_{\rm hopf}^-(\mu=0) = 

2404: \exp\left[-\sqrt{\mu_0\over\alpha \,\epsilon_0\,\left( 1-\kappa\right) }\right]

2405: \quad.

2406: $$

2407: 

2408: For $\phi>\phi_{\rm hopf}^-$ the coefficient $C$ is negative, hence, equation ${\cal H}=0$ 

2409: has a single positive solution.

2410: On the interval $[0,\phi_{\rm hopf}^-]$ this equation admits two solutions which are positive

2411: because $B$ is negative close to $\phi=0$: the solution $\mu_{\rm hopf}^-$ which vanishes

2412: at the point $\phi_{\rm hopf}^-$ and the solution $\mu_{\rm hopf}^+$ which exists on the whole

2413: interval $[0,1]$. The unstability occurs for $\mu\in[\mu_{\rm hopf}^-,\mu_{\rm hopf}^+]$.

2414: At small shear rates, or small $\phi$, the criterion ${\cal H}$ is dominated by:

2415: %\begin{eqnarray*}

2416: %A &\simeq&  (E_0\,\alpha \,\epsilon_0)^2\,{\phi }^2\,

2417: %    \left( E_0\,\alpha \,\kappa \,\phi  + 

2418: %        E_1\,\left( 1 + \kappa  \right) \,\mu_0\,{\phi }^{\kappa } \right) \,

2419: %      {\log (\phi )}^2\\

2420: %B &\simeq& E_0\,E_1\,{\alpha }^2\,{\epsilon_0}^2\,{\phi }^{1 + \kappa }\,

2421: %\left(\kappa-1\right)\,

2422: %      	\left( E_0\,\alpha \,\kappa \,\phi+

2423: %        E_1\,\left( 1 + \kappa  \right) \,\mu_0\,{\phi }^{\kappa } \right) \,

2424: %      {\log (\phi )}^4\\

2425: %C &\simeq& {E_1}^2\,\left(\kappa-1\right)^2\,\alpha \,\epsilon_0\,\mu_0\,{\phi }^{2\,\kappa }\,

2426: %   \left( E_0\,\alpha \,\phi  + E_1\,\mu_0\,{\phi }^{\kappa } \right) \,

2427: %   {\log (\phi )}^4

2428: %\end{eqnarray*}

2429: \begin{eqnarray*}

2430: A &\simeq&  E_0^2\,E_1\,\alpha^2\,\epsilon_0^2\,\mu_0\,

2431: \left( 1 + \kappa  \right) \,{\phi }^{2+\kappa }\,{\log (\phi )}^2\\

2432: B &\simeq& E_0\,E_1^2\,{\alpha }^2\,{\epsilon_0}^2\,\mu_0\,\left(\kappa^2-1\right)\,

2433: {\phi }^{1+2\kappa } \, {\log (\phi )}^4\\

2434: C &\simeq& {E_1}^3\,\alpha \,\epsilon_0\,\mu_0^2\,\left(\kappa-1\right)^2\,

2435: {\phi }^{3\,\kappa }\,{\log (\phi )}^4

2436: \end{eqnarray*}

2437: whence,

2438: \begin{eqnarray*}

2439: {B\over A} &\simeq& {E_1\over E_0}\,(\kappa-1)

2440: {\phi }^{\kappa-1} \, {\log (\phi )}^2\\

2441: {C\over A} &\simeq& {E_1\,\mu_0\over E_0^2\,\alpha\,\epsilon_0}\,

2442: {\left(\kappa-1\right)^2\over\kappa+1}\,

2443: {\phi }^{2\,\kappa-2}\,{\log (\phi )}^2\\

2444: {C\over B} &\simeq& {\mu_0\over E_0\,\alpha\,\epsilon_0}\,

2445: {\kappa-1\over\kappa+1}\,{\phi }^{\kappa-1}

2446: \end{eqnarray*}

2447: The solution $\mu_{\rm hopf}^+$ can then be identified to:

2448: $$

2449: \mu_{\rm hopf}^+ \simeq -{B\over A}\simeq 

2450: {E_1\over E_0}\,(1-\kappa) {\phi }^{\kappa-1} \, {\log (\phi )}^2

2451: \quad,

2452: $$

2453: close to $\phi=0$.

2454: The same expression was found in the isotropic limit, 

2455: but here, the relation between $\phi$ and the shear rate, 

2456: $\dot\epsilon$ is different.

2457: The solution $\mu_{\rm hopf}^-$ can be expanded as,

2458: $$

2459: \mu_{\rm hopf}^- \simeq -{C\over B}\simeq  

2460: {\mu_0\over E_0\,\alpha\,\epsilon_0}\,

2461: {1-\kappa\over\kappa+1}\,{\phi }^{\kappa-1}

2462: $$

2463: %\bibliography{thermo}

2464: 

2465: \begin{thebibliography}{100}

2466: 

2467: \bibitem{goetze92}

2468: W. G\"otze and L. Sj\"ogren, Rep. Prog. Phys. {\bf 55},  241  (1992).

2469: 

2470: \bibitem{ediger96}

2471: M.~D. Ediger, C.~A. Angell, and S.~R. Nagel, J. Phys.: Condens. Matter {\bf

2472:   100},  13200  (1996).

2473: 

2474: \bibitem{larson99}

2475: R. Larson, {\em The structure and rheology of complex fluids} (Oxford

2476:   University Press, New York, 1999).

2477: 

2478: \bibitem{angell00}

2479: C.~A. Angell {\it et~al.}, J. Appl. Phys. {\bf 88},  3113  (2000).

2480: 

2481: \bibitem{berthier01}

2482: L. Berthier and J.-L. Barrat, Phys. Rev. E {\bf 63},  012503  (2001).

2483: 

2484: \bibitem{yamamoto97}

2485: R. Yamamoto and A. Onuki, Europhysics Letters {\bf 40},  61  (1997).

2486: 

2487: \bibitem{yamamoto98}

2488: R. Yamamoto and A. Onuki, Phys. Rev. E {\bf 58},  3515  (1998).

2489: 

2490: \bibitem{bonn02a}

2491: D. Bonn {\it et~al.}, Laponite: aging and shear rejuvenation of a colloidal

2492:   glass, in preparation, 2002.

2493: 

2494: \bibitem{bonn02b}

2495: D. Bonn {\it et~al.}, Rheology of soft glassy materials, in preparation, 2002.

2496: 

2497: \bibitem{sollich97}

2498: P. Sollich, F. Lequeux, P. H\'ebraud, and M.~E. Cates, Phys. Rev. Lett. {\bf

2499:   78},  2020  (1997).

2500: 

2501: \bibitem{sollich98}

2502: P. Sollich, Phys. Rev. E {\bf 58},  738  (1998).

2503: 

2504: \bibitem{hebraux98}

2505: P. H\'ebraux and F. Lequeux, Phys. Rev. Lett. {\bf 81},  2934  (1998).

2506: 

2507: \bibitem{berthier00}

2508: L. Berthier, J.-L. Barrat, and J. Kurchan, Phys. Rev. E {\bf 61},  5464

2509:   (2000).

2510: 

2511: \bibitem{fielding00}

2512: S. Fielding, P. Sollich, and M. Cates, J. Rheol. {\bf 44},  323  (2000).

2513: 

2514: \bibitem{bouchaud96}

2515: J. Bouchaud, L. Cugliandolo, J. Kurchan, and M. M\'ezard, Physica A {\bf 226},

2516:   243  (1996).

2517: 

2518: \bibitem{monthus96}

2519: C. Monthus and J.-P. Bouchaud, J. Phys. A {\bf 29},  3847  (1996).

2520: 

2521: \bibitem{derec01}

2522: C. Derec, A. Ajdari, and F. Lequeux, Eur. Phys. J. E {\bf 4},  355  (2001).

2523: 

2524: \bibitem{liu98}

2525: A.~J. Liu and S.~R. Nagel, Nature {\bf 396},  21  (1998).

2526: 

2527: \bibitem{ohern01}

2528: C.~S. O'Hern, S.~A. Langer, A.~J. Liu, and S.~R. Nagel, Phys. Rev. Lett. {\bf

2529:   86},  111  (2001).

2530: 

2531: \bibitem{knight95}

2532: J.~B. Knight {\it et~al.}, Phys. Rev. E {\bf 51},  3957  (1995).

2533: 

2534: \bibitem{boutreux97}

2535: T. Boutreux and P.~G. de~Gennes, Physica A {\bf 244},  59  (1997).

2536: 

2537: \bibitem{nowak98}

2538: E.~R. Nowak {\it et~al.}, Phys. Rev. E {\bf 57},  1971  (1998).

2539: 

2540: \bibitem{falk98}

2541: M.~L. Falk and J.~S. Langer, Phys. Rev. E {\bf 57},  7192  (1998).

2542: 

2543: \bibitem{falk00}

2544: M.~L. Falk and J.~S. Langer, M.R.S. Bulletin {\bf 25},  40  (2000).

2545: 

2546: \bibitem{spaepen77}

2547: F. Spaepen, Acta Metall. {\bf 25},  407  (1977).

2548: 

2549: \bibitem{argon79a}

2550: A. Argon, Acta Metall. {\bf 27},  47  (1979).

2551: 

2552: \bibitem{argon79b}

2553: A. Argon and H. Kuo, Mater. Sc. Eng. {\bf 39},  101  (1979).

2554: 

2555: \bibitem{spaepen81}

2556: F. Spaepen and A. Taub,  in {\em Physics of Defects}, {\em Les Houches

2557:   Lectures, Session XXXV}, edited by J.-P. Balian and M. Kleman (North-Holland,

2558:   Amsterdam, 1981), p.\ 133.

2559: 

2560: \bibitem{argon83}

2561: A. Argon and L. Shi, Acta Metall. {\bf 31},  499  (1983).

2562: 

2563: \bibitem{lemaitre01b}

2564: A. Lema\^{\i}tre, cond-mat/0108442, 2001.

2565: 

2566: \bibitem{dieterich78}

2567: J.~H. Dieterich, Pure Appl. Geophys. {\bf 116},  790  (1978).

2568: 

2569: \bibitem{dieterich79}

2570: J.~H. Dieterich, J. Geophys. Res. {\bf 84},  2161  (1979).

2571: 

2572: \bibitem{rice83}

2573: J. Rice and A. Ruina, J. Appl. Mech. {\bf 105},  343  (1983).

2574: 

2575: \bibitem{radjai98a}

2576: F. Radjai, D.~E. Wolf, M. Jean, and J.-J. Moreau, Phys. Rev. Lett. {\bf 80},

2577:   61  (1998).

2578: 

2579: \bibitem{radjai98b}

2580: F. Radjai and D.~E. Wolf, Granular Matter {\bf 1},  3  (1998).

2581: 

2582: \bibitem{lemaitre01a}

2583: A. Lema\^{\i}tre, cond-mat/0107422, 2001.

2584: 

2585: \bibitem{schlesinger84}

2586: M.~F. Schlesinger and E. Montroll, Proc. Nat. Acad. Sc. US {\bf 81},  1280

2587:   (1984).

2588: 

2589: \bibitem{langer99}

2590: J.~S. Langer, Phys. Rev. E {\bf 62},  1351  (1999).

2591: 

2592: \bibitem{langer01}

2593: J.~S. Langer, Phys. Rev. E {\bf 64},  011504  (2001).

2594: 

2595: \bibitem{saulnier02}

2596: F. Saulnier, E. Raphael, and P.-G. de~Gennes, Dewetting of thin polymer films

2597:   near the glass transition, cond-mat/0201528, 2002.

2598: 

2599: \bibitem{lubliner90}

2600: J. Lubliner, {\em Plasticity Theory} (Macmillan, New York, 1990).

2601: 

2602: \bibitem{stillinger84}

2603: F.~H. Stillinger and T.~A. Weber, Science {\bf 225},  983  (1984).

2604: 

2605: \bibitem{cohen59}

2606: M.~H. Cohen and D. Turnbull, J. Chem. Phys. {\bf 31},  1164  (1959).

2607: 

2608: \bibitem{turnbull61}

2609: D. Turnbull and M.~H. Cohen, J. Chem. Phys. {\bf 34},  120  (1961).

2610: 

2611: \bibitem{turnbull70}

2612: D. Turnbull and M.~H. Cohen, J. Chem. Phys. {\bf 52},  3038  (1970).

2613: 

2614: \bibitem{edwards89}

2615: S.~F. Edwards and R.~B.~S. Oakeshott, Physica A {\bf 157},  1080  (1989).

2616: 

2617: \bibitem{mehta89}

2618: A. Mehta and S.~F. Edwards, Physica A {\bf 157},  1091  (1989).

2619: 

2620: \bibitem{edwards94}

2621: S. Edwards,  in {\em Granular Matter: An Interdisciplinary Approach}, edited by

2622:   A. Mehta (Springer-Verlag, New-York, 1994), pp.\ 121--140.

2623: 

2624: \bibitem{savage79}

2625: S.~B. Savage, J. Fluid. Mech. {\bf 92},  53  (1979).

2626: 

2627: \bibitem{coppersmith96}

2628: S. Coppersmith {\it et~al.}, Phys. Rev. E {\bf 53},  4673  (1996).

2629: 

2630: \bibitem{cloitre00}

2631: M. Cloitre, R. Borrega, and L. Leibler, Phys. Rev. Lett. {\bf 85},  4819

2632:   (2000).

2633: 

2634: \bibitem{bureau02}

2635: L. Bureau, T. Baumberger, and C. Caroli, Rheological aging and rejuvenation in

2636:   solid friction contacts, cond-mat/0202245, 2002.

2637: 

2638: \bibitem{yoshino98}

2639: H. Yoshino, Phys. Rev. Lett. {\bf 81},  1493  (1998).

2640: 

2641: \bibitem{boon91}

2642: J.-P. Boon and S. Yip,  in {\em Molecular Hydrodynamics}, edited by McGraw-Hill

2643:   (Dover, New York, 1991).

2644: 

2645: \bibitem{debenedetti01}

2646: P.~G. Debenedetti and F.~H. Stillinger, Nature {\bf 400},  259  (2001).

2647: 

2648: \bibitem{bonn99}

2649: D. Bonn {\it et~al.}, langmuir {\bf 15},  7534  (1999).

2650: 

2651: \bibitem{kob95}

2652: W. Kob and H.~C. Andersen, Phys. Rev. E {\bf 52},  4134  (1995).

2653: 

2654: \bibitem{megen94}

2655: W. van Megen and S. Underwood, Phys. Rev. E {\bf 49},  4206  (1994).

2656: 

2657: \bibitem{gopal99}

2658: A. Gopal and D. Durian, J. Coll. Int. Sc. {\bf 213},  169  (1999).

2659: 

2660: \bibitem{hebraud97}

2661: P. H\'ebraud, F. Lequeux, J. Munch, and D. Pine, Phys. Rev. Lett. {\bf 78},

2662:   4657  (1997).

2663: 

2664: \bibitem{megen98}

2665: W. van Megen, T. Mortensen, S. Williams, and J. {M\"uller}, Phys. Rev. E {\bf

2666:   58},  6073  (1998).

2667: 

2668: \bibitem{cipelletti00}

2669: L. Cipelletti, S. Manley, R. Ball, and D. Weitz, Phys. Rev. Lett. {\bf 84},

2670:   2275  (2000).

2671: 

2672: \bibitem{hassan95}

2673: O.~A. Hassan and M.~C. Boyce, Pol. Engineer. and Science {\bf 35},  331

2674:   (1995).

2675: 

2676: \bibitem{drummond00}

2677: C. Drummond and J. Israelachvili, Macromolecules {\bf 33},  4910  (2000).

2678: 

2679: \bibitem{drummond01}

2680: C. Drummond and J. Israelachvili, Phys. Rev. E {\bf 63},  041506  (2001).

2681: 

2682: \bibitem{persson00}

2683: Persson,  in {\em "Sliding Friction : Physical Principles and Applications

2684:   (Nanoscience and Technology)"}, edited by "" ("Springer Verlag", "", 2000).

2685: 

2686: \bibitem{thompson92}

2687: P.~A. Thompson, G.~S. Grest, and M.~O. Robbins, Phys. Rev. Lett. {\bf 68},

2688:   3448  (1992).

2689: 

2690: \bibitem{yoshizawa93}

2691: H. Yoshizawa and J. Israechvili, J. Phys.: Condens. Matter {\bf 97},  11300

2692:   (1993).

2693: 

2694: \bibitem{demirel96a}

2695: A.~L. Demirel and S. Granick, Phys. Rev. Lett. {\bf 77},  2261  (1996).

2696: 

2697: \bibitem{demirel96b}

2698: A.~L. Demirel and S. Granick, Phys. Rev. Lett. {\bf 77},  4330  (1996).

2699: 

2700: \bibitem{thompson91}

2701: P.~A. Thompson and G.~S. Grest, Phys. Rev. Lett. {\bf 67},  1751  (1991).

2702: 

2703: \bibitem{nasuno97}

2704: S. Nasuno, A. Kudrolli, and J. Gollub, Phys. Rev. Lett. {\bf 79},  949  (1997).

2705: 

2706: \bibitem{losert00}

2707: S.~N. W.~Losert, J.-C.~G\'eminard and J. Gollub, Phys. Rev. E {\bf 61},  4060

2708:   (2000).

2709: 

2710: \bibitem{lobkovsky98}

2711: A.~E. Lobkovsky and J.~S. Langer, Phys. Rev. E {\bf 58},  1568  (1998).

2712: 

2713: \bibitem{brener02}

2714: E.~A. Brener and R. Spatschek, Fast crack propagation by surface diffusion,

2715:   submitted, 2002.

2716: 

2717: \bibitem{durian95}

2718: D. Durian, Phys. Rev. Lett. {\bf 75},  4780  (1995).

2719: 

2720: \bibitem{durian97}

2721: D. Durian, Phys. Rev. E {\bf 55},  1739  (1997).

2722: 

2723: \end{thebibliography}

2724: 

2725: \end{document}

2726: