1: \documentclass[twocolumn,showpacs,aps,pre,

2: groupedaddress,amssymb,amsmath,nobalancelastpage]{revtex4}

3: 

4: \usepackage{graphicx}% Include figure files

5: \usepackage{longtable}% Include figure files

6: 

7: \begin{document}

8: 

9: %\preprint{APS/123-QED}

10: 

11: \title{Statistics of Bubble Rearrangements in a Slowly Sheared Two-dimensional Foam}

12: 

13: \author{Michael Dennin}

14: \affiliation{Department of Physics and Astronomy, University of

15: California at Irvine, Irvine, California 92697-4575}

16: 

17: \date{\today}

18: 

19: \begin{abstract}

20: 

21: Many physical systems exhibit plastic flow when subjected to slow

22: steady shear. A unified picture of plastic flow is still lacking;

23: however, there is an emerging theoretical understanding of such

24: flows based on irreversible motions of the constituent

25: ``particles'' of the material. Depending on the specific system,

26: various irreversible events have been studied, such as T1 events

27: in foam and shear transformation zones (STZ's) in amorphous

28: solids. This paper presents an experimental study of the T1 events

29: in a model, two-dimensional foam: bubble rafts. In particular, I

30: report on the connection between the distribution of T1 events and

31: the behavior of the average stress and average velocity profiles

32: during both the initial elastic response of the bubble raft and

33: the subsequent plastic flow at sufficiently high strains.

34: 

35: \end{abstract}

36: 

37: % insert suggested PACS numbers in braces on next line

38: \pacs{82.70.-y,83.60.La,62.20.Fe}

39: % insert suggested keywords - APS authors don't need to do this

40: %\keywords{}

41: 

42: %\maketitle must follow title, authors, abstract, \pacs, and \keywords

43: \maketitle

44: 

45: \section{Introduction}

46: 

47: Bubble rafts have been used as a model experimental system for the

48: study of crystalline and amorphous solids \cite{BL49,AK79} and for

49: the study of two-dimensional foam \cite{KE99,LTD02}. This overlap

50: is just one of many examples that points to an important question

51: in the study of the mechanical response of materials. Under

52: conditions of slow steady shear, what, if any, is the connection

53: between the response of ``mesoscopic'' materials, such as foams,

54: emulsions, pastes, and slurries, and plastic flow of ``molecular''

55: systems, such as amorphous solids? Based on macroscopic

56: measurements, the systems are similar. There is an initial elastic

57: response for small strains (or stresses) and a yield stress, above

58: which irreversible deformations, or plastic deformations, occur.

59: Eventually, above some critical stress (or strain), the system

60: enters a ``flowing'' state that is characterized by irregular

61: periods of stress increase and decrease. This is often referred to

62: as unbounded plastic flow. For the purposes of this paper, this

63: will simply be referred to as plastic flow. As one reduces the

64: shear rate, the critical stress approaches the yield stress in

65: such a way that for sufficiently slow shear rates the behavior of

66: the system is essentially shear-rate independent. This is often

67: referred to as the {\it quasistatic regime}. A complete

68: ``microscopic'' picture of plastic flow still does not exist,

69: where microscopic refers to the fundamental length scale relevant

70: to the system in question. For example, in bubble rafts, it would

71: be the dynamics of individual bubbles. Open questions include: the

72: microscopic source of the stress release events; the spatial and

73: temporal distribution of such events; and the nature of such

74: events during periods of stress increase. Experimentally, the

75: challenge is identifying systems for which the microscopic events

76: are directly observable. This is one of the main advantages of

77: mesoscopic systems, such as the bubble raft, and the reason for

78: the interest in making connections between mesoscopic systems and

79: molecular systems, such as amorphous solids.

80: 

81: Models and simulations of diverse systems, ranging from solids

82: \cite{SVE83,BA94a,BA94b,BA94c,FL98,BVR02,O03,PALB04} to foam

83: \cite{KNN89,KOKN95,OK95,D95,D97,WBHA92,HWB95,JSSAG99,TSDKLL99,KD03},

84: have provided a number of insights into these questions. The focus

85: of this paper is on the role of T1 events in foam. Aqueous foam

86: consists of gas bubbles separated by liquid walls

87: \cite{S93,K88,WH99}. A T1 event is an irreversible neighbor

88: switching event. Both a schematic representation and an actual T1

89: event are presented in Fig.~\ref{defT1}. For the purposes of this

90: paper, we will focus on the role of T1 events during the steady

91: shear of foam. However, it should be mentioned that understanding

92: the nonlinear events that are {\it not} shear induced may be

93: important when comparing foam and amorphous solids. In the absence

94: of external stress, foam coarsens, and T1 events occur due to

95: geometric changes in the foam structure. These T1 events are not

96: necessarily distinguishable from those caused by flow. In

97: contrast, most amorphous solids do not exhibit the equivalent of

98: coarsening. However, thermally activated events may be important.

99: Possible differences between thermally activated and coarsening

100: events is an interesting question, but one beyond the scope of

101: this paper.

102: 

103: Simulations of flowing foam have characterized different aspects

104: of T1 events. Often, one separately considers T1 events and

105: ``avalanches'', i.e. sudden releases of stress (or energy) in the

106: foam. One issue is whether or not the probability of the number of

107: T1 events in a given avalanche exhibits power-law behavior. A

108: common model to study dry foam (where the bubbles are essentially

109: polygonal) is the vertex model \cite{KNN89,KOKN95,OK95}.  In the

110: vertex model, a T1 event is defined to occur when the distance

111: between two vertices (i.e. the wall between two bubbles) is below

112: a threshold value. One then eliminates this wall, creating a T1

113: event. Therefore, within this model, all T1 events are essentially

114: instantaneous. In this case, simulations found evidence for

115: power-law behavior of the probability of T1 events. Another

116: characterization of the T1 events is the number of T1 events per

117: bubble per unit strain, $R_{T1}$. For the vertex model, this

118: quantity is $R_{T1} = 0.5$. A modified version of the vertex model

119: has been used to study the issue of flow localization under shear

120: \cite{KD03}. These simulations report a correlation between the

121: spatial localization of T1 events to the neighborhood of a system

122: boundary with the localization of shear in the same region. The

123: issue of shear localization will be discussed in more detail

124: later.

125: 

126: Another class of models focuses on wet foam (foam in which the

127: bubbles are essentially spherical, or in two-dimensions,

128: circular), using a quasistatic simulation \cite{WBHA92,HWB95}.

129: These simulations involve making a small step strain and then

130: allowing the system to relax to an energy minimum before applying

131: the next step strain. Anytime the energy decreases after a step

132: strain, one declares this an ``avalanche'' or ``event''. By

133: comparing neighbors in the initial and final state, one can count

134: the number of T1 events for a particular avalanche. In these

135: simulations, avalanches consisting of a large number of T1 events

136: were observed, suggesting the possibility of power-law behavior

137: \cite{WBHA92,HWB95}. For this model, $R_{T1}$ was not reported.

138: 

139: Wet foam under steady shear has also been simulated using a

140: q-Potts model \cite{JSSAG99}. In this case, different bubbles are

141: identified by different spin orientations. Simulations of the

142: q-Potts model find that the distribution of topological

143: rearrangements are not power-law-like. However, the distribution

144: of energy drops may be consistent with power-law behavior

145: \cite{JSSAG99}.

146: 

147: Another important set of simulations for wet foam involved

148: studying the steady shear of the bubble model

149: \cite{D95,D97,TSDKLL99}. This model treats bubbles as spheres (or

150: circles) that interact via a spring force proportional to their

151: overlap and a viscous drag proportional to velocity differences.

152: As this model directly simulates dynamics, the duration of T1

153: events exhibits a distribution of duration times. Simulations

154: focused on small shear-rates in the quasistatic limit, i.e. the

155: flow properties were independent of the shear rate. Under these

156: conditions, no evidence of power-law behavior is observed in the

157: bubble model at high bubble density, and $R_{T1} = 0.15$

158: \cite{D95,D97,TSDKLL99}. If one decreases the density of the

159: bubbles, it appears that the distribution of events approaches a

160: power-law as one approaches the critical density for the

161: ``melting'' of the foam \cite{TSDKLL99}.

162: 

163: Before discussing the current state of experiments in foam, it is

164: useful to put the theoretical work on T1 events in foam in the

165: context of two points of view of plasticity in amorphous

166: materials. First, the idea of shear transformation zones (STZ's),

167: as developed by Falk and Langer \cite{FL98}, has received

168: significant attention. STZ are a way of describing local,

169: irreversible rearrangements of particles during shear.  STZ are

170: based on previous work by Spaepen and Argon on activated

171: transitions and Turnball, Cohen, and others on free-volume

172: fluctuations. As the STZ refers to a small region of the material

173: with certain properties \cite{A79}, there is only a loose

174: connection between the STZ and T1 events. However, it is

175: reasonable to identify as an STZ regions in which a few T1 events

176: combine to form a local slip (see, for instance,

177: Fig.~\ref{t1elastic}e). It is expected that the local

178: rearrangements identified as STZ are associated with quadrupolar

179: energy fluctuations. In fact, the expected quadrupolar energy

180: fluctuations have been observed associated with T1 events in a

181: simulation of foam \cite{KD03}, but not, as of yet, in simulations

182: of molecular systems.

183: 

184: Another view of plasticity is based on shear induced changes in

185: the potential energy landscape, as proposed by Malandro and Lacks

186: \cite{ML99}. This picture derives from an inherent structure

187: formalism and focuses on changes in the macroscopic mechanical

188: response of a material due to shear induced changes of the

189: potential energy. This formulism has been used to study

190: simulations of a quasistatic version of the bubble model

191: \cite{ML04}. In this case, system wide rearrangement events are

192: observed. This is not seen in bubble model simulations of the

193: quasistatic limit, but it is seen in other quasistatic simulations

194: of foam. The work in Ref.~\cite{ML04} suggests the need to

195: carefully define the concept of an ``event'', especially for

196: steady-state experiments where the time scale for events to occur

197: relative to the applied shear can be important. For example, a

198: shear-rate regime may exist that is quasistatic as defined by the

199: behavior of quantities such as the average stress, but not in a

200: quasistatic with regard to the duration of stress releases. Hence,

201: large events get ``broken up'' by the steady shear, changing the

202: nature of the distribution of events.

203: 

204: A number of experimental studies of bubble rearrangements in model

205: foam have been carried out. As mentioned, some of the earliest

206: work was done using bubble rafts \cite{BL49,AK79,MGC89}, i.e.

207: layers of gas bubbles floating on a liquid surface, as a model

208: molecular system, both for crystalline and amorphous solids

209: \cite{BL49,AK79}. One major advantage of bubble rafts is that

210: their two dimensional nature allows for easy imaging and tracking

211: of all of the ``particles'' in the system. Another reason that

212: bubble rafts have been so useful in the study of molecular systems

213: is that there exists detailed calculations of the bubble

214: interactions \cite{SA82}. More recently, bubble rafts were used to

215: study rearrangements after a step strain in order to make

216: comparisons with the quasistatic simulations of foam \cite{KE99}.

217: This work did not directly measure T1 events, but it did look at

218: changes in the number of neighbors for bubbles. The results

219: suggested that large scale events were possible.

220: 

221: Experiments have also been carried out using monolayer foam

222: \cite{DK97}. Langmuir monolayers consist of a single layer of

223: molecules confined to the air-water interface. They exhibit a

224: large number of two-dimensional phases, including gas-liquid

225: coexistence. This allows for the formation of a foam of gas

226: bubbles with liquid walls. For a monolayer foam under steady

227: shear, only a small number of simultaneous T1 events were

228: observed, with $R_{T1} = 0.12 \pm 0.03$. These results are

229: consistent with the bubble model.

230: 

231: As mentioned, the other aspect of T1 dynamics is their potential

232: role in explaining shear localization in yield-stress materials

233: \cite{KD03}, such as foam and granular systems. It has long been

234: known that a yield stress and/or nonlinear viscosity can lead to

235: inhomogeneous flows \cite{BOOKS}. However, it is only recently

236: that experimental techniques have allowed for detailed

237: measurements of such behavior. A number of such studies have been

238: carried out in granular materials, where exponential velocity

239: profiles (or other strongly localized velocity profiles) are

240: generally observed \cite{HBV99,LBLG00,MDKENJ00}. In contrast,

241: measurements in various three-dimensional pastes, slurries, and

242: foams show a different type of inhomogeneous flow. In this case,

243: the flow is not strongly localized, and there is a

244: shear-discontinuity at the boundary between flow and no flow

245: \cite{CRBMGH02,DCBC02}.

246: 

247: For two-dimensional foams, the situation is ambiguous.

248: Three-dimensional foam that is confined between plates to form a

249: model two-dimensional system exhibits shear localization analogous

250: to granular systems \cite{DTM01}. This work motivated simulations

251: of the modified vertex model discussed earlier that showed a

252: connection between T1 events and shear localization \cite{KD03}.

253: In this case, it appears that the spatial distribution of stress

254: released by the T1 event results in the subsequent localization of

255: the events. The localization of T1 events is correlated with the

256: localization of flow. In contrast, experiments with a bubble raft

257: exhibit a shear-discontinuity \cite{LCD04} similiar to that

258: reported in Refs.~\cite{CRBMGH02,DCBC02}. In Ref.~\cite{LCD04}, T1

259: events were not measured.

260: 

261: The work reported in this paper addresses the general question of

262: the temporal and spatial distribution of T1 events during the

263: slow, steady shear of a bubble raft. Also, connections between the

264: T1 events and the velocity profiles reported in Ref.~\cite{LCD04}

265: are made. The rest of the paper is organized as follows.

266: Section~\ref{ExpMeth} provides the details of the experimental

267: setup. The results are presented in two parts.

268: Section~\ref{elasticsec} presents the initial response of the

269: system. Section~\ref{plasticsec} presents the behavior during

270: plastic flow. Finally, the results are summarized and discussed in

271: Sec.~\ref{summary}.

272: 

273: \section{Experimental Methods}

274: \label{ExpMeth}

275: 

276: The experimental system consisted of a standard bubble raft

277: \cite{AK79} in a Couette geometry. The bubble raft was produced by

278: flowing regulated nitrogen gas through a hypodermic needle into a

279: homogeneous solution of 80\% by volume deionized water, 15\% by

280: volume glycerine, and 5.0\% by volume Miracle Bubbles (Imperial

281: Toy Corp.). The bubble size was dependent on the nitrogen flow

282: rate, which was varied using a needle valve.  A random

283: distribution of bubble sizes was used, with an average radius of

284: $1\ {\rm mm}$. The resulting bubbles were spooned into a

285: cylindrical Couette viscometer. This produced a two-dimensional

286: model of a wet foam on a homogeneous liquid substrate.

287: Figure~\ref{schematic} presents a schematic side view of the

288: bubbles in the apparatus and an image of a top view of the bubble

289: raft.

290: 

291: Due to the nature of the bubble raft, no measurable coarsening was

292: observed. However, after approximately two hours significant

293: numbers of bubbles would pop, presumably due to loss of fluid in

294: the walls. This set the upper limit on the total time of the

295: measurements. In contrast, during the initial two hour period,

296: only two out of approximately 400 bubbles in the field of view

297: were observed to pop.

298: 

299: An important feature of the bubble raft is the gas area fraction.

300: To achieve a desired gas area fraction, the bubble raft was

301: constructed by placing the approximate number of desired bubbles

302: in the trough with the outer barrier set to a large radius. It is

303: important to note that the bubbles exhibited a strong attraction

304: to each other. The outer barrier was compressed until the desired

305: radius was reached. The gas area fraction was determined by

306: thresholding images of the bubbles and counting the area inside of

307: the bubbles. Because of the three-dimensional nature of the

308: bubbles, this represents an operational definition of gas-area

309: fraction based on the details of the image analysis. However, the

310: choice of threshold was consistent with an estimate of the gas

311: area fraction based on the area of trough and expected

312: distribution of bubble sizes. For all of the data reported here,

313: the gas area fraction was approximately 0.95.

314: 

315: The Couette viscometer is described in detail in Ref.~\cite{app}

316: and shown schematically in Fig.~\ref{schematic}(a). It consists of

317: a shallow dish that contains the liquid substrate. Two concentric

318: Teflon barriers are placed vertically in the dish. Sections of

319: both of these barriers are visible in Fig.~\ref{schematic}(b). The

320: outer barrier is a ring consisting of twelve segmented pieces. It

321: has an adjustable radius. For the experiments discussed here, the

322: outer radius was fixed at $r_o = 7.43\ {\rm cm}$. The inner

323: barrier, or rotor, is a Teflon disk with a radius $r_i = 3.84\

324: {\rm cm}$. The outer edge of the disk is a knife edge that is just

325: in contact with the water surface. It was suspended by a wire to

326: form a torsion pendulum.

327: 

328: To shear the bubble raft, the outer Teflon barrier was rotated at

329: a constant angular velocity $\Omega = 8 \times 10^{-4}\ {\rm

330: rad/s}$. The first layer of bubbles at either boundary did not

331: slip relative to the boundary. Due to the finite size of the

332: bubbles, this results in an effective inner radius of $r = 4.4\

333: {\rm cm}$. Due to the cylindrical geometry, the shear rate is not

334: uniform across the system and is given by $\dot{\gamma}(r) = r

335: \frac{d}{dr}\frac{v(r)}{r}$. Here $v(r)$ is the azimuthal velocity

336: of the bubbles. During plastic flow, the average azimuthal

337: velocity of the inner cylinder is zero. Measurements of the

338: average azimuthal velocity profile allows for calculations of the

339: shear rate. As measured at $r = 4.4\ {\rm cm}$, $\dot{\gamma} = 4

340: \times 10^{-3}\ {\rm s^{-1}}$. In this regime, where reported, the

341: strain ($\gamma$) is taken to be the strain at this radius, and is

342: computed from $\gamma = \dot{\gamma}t + \gamma_o$, where $t$ is

343: the time since the initiation of plastic flow and $\gamma_o$ is

344: the amount of strain developed during the initial period. During

345: the initial period, the inner barrier has a finite angular speed.

346: However, one can still compute the effective shear rate at $r =

347: 4.4\ {\rm cm}$. In this regime, $\dot{\gamma} = 3 \times 10^{-4}\

348: {\rm s^{-1}}$. Again, where reported, the strain is the strain at

349: the inner cylinder: $\gamma = \dot{\gamma}t$, where in this case

350: $t$ is measured from the initiation of shear.

351: 

352: The details of the velocity measurements are given in

353: Ref~\cite{LCD04}. Video tape of roughly one third of the trough

354: was recorded. Images from this tape were taken every 3.2 s and

355: digitized. An image processing routine based on standard

356: Labwindows functions was developed that detected and tracked

357: individual bubbles. This tracking software was also used to

358: compute the average bubble displacements. This is used to compute

359: the deviation of the bubble motion from ideal elastic behavior, as

360: discussed in Sec.~\ref{elasticsec}.

361: 

362: The T1 events were measured by stepping the digitized images one

363: frame at a time and visually searching for the location and time

364: at which T1 events occurred. A T1 event was defined to occur when

365: two bubbles were observed to lose contact, and two other bubbles

366: moved into the resulting space (see Fig.~\ref{defT1}). Due to the

367: associated motions of the other neighboring bubbles, T1 events are

368: relatively easy to detect by hand \cite{test}. For automatic

369: tracking of T1 events, it is critical to accurately detect

370: essentially all of the bubbles, as one is interested in

371: determining neighbor switching events. This is in contrast with

372: the displacement and velocity measurements where the requirement

373: is that one tracks enough bubbles to have sufficient statistics.

374: These are the reasons that automatic methods were used for

375: displacement and velocity measurements, but the detection of T1

376: events was by hand.

377: 

378: The stress on the inner rotor was determined using two different

379: methods. In both cases, the torque, $T = \kappa \theta$, on the

380: inner rotor is determined by measuring the angular displacement,

381: $\theta$. (For the experiments presented here, the torsion

382: constant $\kappa = 5.7 \times 10^{-7}\ {\rm N m}$.) The stress is

383: then determined from $\sigma = T/2\pi r^2$. The difference in the

384: two methods is the determination of $\theta$. The first method

385: uses a magnetic flux technique, and the details of this technique

386: are in Ref.~\cite{app}. This is the more precise of the two

387: methods, with a stress resolution of $3 \times 10^{-3}\ {\rm

388: mN/m}$. The second method uses the video images of the inner

389: cylinder and tracks fixed features on the disk. This method has a

390: resolution of $0.043\ {\rm mN/m}$. The second method is used to

391: correlate the video analysis of bubble motions (displacements and

392: T1 events) with the detailed stress fluctuations determined from

393: the magnetic flux measurements that are reported in

394: Ref.~\cite{PD03}.

395: 

396: As mentioned, foams are inherently nonequilibrium systems. One

397: complication that arises from this is the definition of the yield

398: stress. For sufficiently low shear rates, foam will spontaneously

399: release stress, usually as part of coarsening process as bubble

400: sizes change. This complication is minimized in the bubble raft

401: given that no substantial coarsening was observed in the absence

402: of applied shear. In either case, a useful operational definition

403: of the yield stress is the zero shear-rate limit of the stress.

404: For the bubble raft of interest here, the average stress as a

405: function of shear rate is well described by a Herschel-Bulkley

406: model ($\sigma(\dot{\gamma}) = \tau_o + \mu \dot{\gamma}^n$)

407: \cite{BAH77}.  From these results, one can determine a yield

408: stress: $\tau_o = 0.8 \pm 0.1\ {\rm mN/m}$ \cite{PD03,LCD04}. For

409: the particular shear rate of interest here, this is different from

410: the ``critical'' stress at which the system begins to undergo

411: ``steady'' flow.

412: 

413: \section{Experimental Results}

414: \subsection{Elastic Regime} \label{elasticsec}

415: 

416: The initial stress response of the system is given in

417: Fig.~\ref{initstress}. There are a number of interesting features

418: of this regime. First, there are three separate regions of the

419: initial response, which is essentially set by the slope of the

420: stress versus strain curve. These regions are indicated by the

421: vertical dashed lines and are separated by isolated stress-drops.

422: 

423: The first region is the linear, elastic response of the material.

424: During this period no T1 events are observed. The second two

425: regions represent plastic deformations in the sense that T1 events

426: occur. These events are too small to produce stress drops. But,

427: they modify the slope of the stress-strain curve and generate

428: irreversible deformation. Hence, the identification of these

429: regions with plastic deformations. The onset of plastic response

430: is another useful definition of the yield stress. However, there

431: is always ambiguity associated with the definition of the onset of

432: T1 events due to the possibility of T1 events that are the result

433: of coarsening and not shear. As discussed for the bubble raft,

434: this difficulty is minimized as no coarsening was observed.

435: However, for the measurements in Fig.~\ref{initstress}, only a

436: fraction of the trough was in view. This limits the degree to

437: which the yield stress can be measured by this method. However, it

438: is useful to note that the onset of T1 events for the single set

439: of data studied here is consistent with the value of yield stress

440: as determined by fits to the behavior of stress as a function of

441: rate of strain.

442: 

443: The fact that any deviation of the stress-strain curve from linear

444: behavior is small allows for the definition of an effective shear

445: modulus of the bubble raft, $G$, during periods of stress

446: increase. For the initial region in Fig.~\ref{initstress}, $G$ is

447: the elastic shear modulus. The calculation of $G$ assumes that the

448: stress is proportional to the strain. The boundary conditions

449: consist of a fixed rotation rate at the outer boundary and an

450: inner boundary that is free to rotate, but supported by a torsion

451: wire. Because of the symmetry of the Couette geometry, the

452: azimuthal velocity, $v(r)$ is only a function of the radial

453: position $r$. This is a standard problem; however, given the

454: slightly unorthodox boundary conditions of this experiment, the

455: solution is repeated here. The relevant constitutive equation is

456: \begin{equation}

457: \label{elastic}

458: \sigma(r) = G \gamma(r).

459: \end{equation}

460: Here $\gamma(r)$ is the shear strain, and $\sigma(r)$ is the

461: resulting shear stress. In the cylindrical geometry, $\gamma(r) =

462: r\frac{d\theta(r)}{dr}$, where $\theta(r)$ is the angular

463: displacement of the bubble raft. For a material confined between

464: two cylinders, the shear stress is given by $\sigma (r) = T/(2\pi

465: r^2)$. This follows directly from balancing torques on each

466: material element. If the bubble raft was a perfectly rigid solid,

467: one would simply have $v(r) = \Omega r$. This is due to the fact

468: that the inner boundary is supported by a torsion wire and rotates

469: as it measures the torque on the inner cylinder. However, for a

470: finite value of $G$, plugging into Eq.~\ref{elastic}, we get

471: \begin{equation}

472: \frac{T}{2\pi r^2} = G\left( r\frac{d\theta(r)}{dr} \right).

473: \end{equation}

474: Integrating this equation, and using the fact that the bubble raft

475: does not slip at either boundary, gives for $G$,

476: \begin{equation}

477: \label{Gexp}

478: G = \frac{\omega}{\Omega - \omega}\left(

479: \frac{\kappa}{4\pi}\right) \left(\frac{1}{r_i^2} - \frac{1}{R^2}

480: \right),

481: \end{equation}

482: and for $v$,

483: \begin{equation}

484: \label{velastic}

485:  v(r) = \Omega r - \left[

486: \frac{\kappa\omega r}{4\pi G} \left(\frac{1}{r^2} - \frac{1}{R^2}

487: \right) \right],

488: \end{equation}

489: This equation for $v$ can be rewritten by plugging in for $G$,

490: \begin{equation}

491: vr) = \Omega r + \left[ \frac{(\Omega - \omega)r_i^2}{R^2 - r_i^2}

492: \left(r - \frac{R^2}{r} \right) \right].

493: \end{equation}

494: The second piece in the expressions for $v(r)$ is due to the

495: elastic nature of the bubble raft and the motion of the inner

496: cylinder. When $\kappa/G$ is small, the system behaves as a rigid

497: body ($\omega = \Omega$), as expected (large $G$ limit).

498: 

499: Using the above results, one can find $G$ from measurements of the

500: average velocity using Eq.~\ref{velastic} or from $\omega$ using

501: Eq.~\ref{Gexp}. For example, the results for $v(r)$ in region (A)

502: of Fig.~\ref{initstress} are given in Fig.~\ref{elflow}. The solid

503: line is a fit to Eq.~\ref{velastic}, with $v(r)/r = 8.2 \times

504: 10^{-4}\ {\rm rad/s} - 0.003\ {\rm rad/cm^2s}(1/r^2 - 0.0193\ {\rm

505: cm^{-2}})$. The T1 events result in a reduction in the effective

506: elastic modulus of the bubble raft. The calculated values of $G$

507: for the three regions are: (A) $G = 11.2 \pm 0.1\ {\rm mN/m}$; (B)

508: $G = 5.4 \pm 0.1\ {\rm mN/m}$; and (C) $G = 8.9 \pm 0.1\ {\rm

509: mN/m}$.

510: 

511: We use the elastic regime to provide a characterization of the

512: local deviation from elastic flow. First, we take the fit of

513: $v(r)$ in region (A) as the definition of ``ideal'' elastic

514: motion. Knowing this velocity curve, we can compute the expected

515: displacement of a bubble during a strain interval. From this, we

516: define $\Delta$ to be a measure of the deviation from elastic

517: behavior: $\Delta = \sqrt{(x - x_e)^2 + (y - y_e)^2}$, where $x$

518: and $y$ are the actual displacements of the bubble and $x_e$ and

519: $y_e$ are the expected displacements if the motion was ``ideal''

520: elastic behavior. As can be seen from Fig.~\ref{elflow}, even in

521: the ``pure elastic'' regime there is a significant non-zero

522: variation to the bubble motions. (The error bars represent the

523: standard error based on the standard deviation of measured

524: velocities in each radial bin.) The variation in bubble velocity

525: is due to a combination of effects, including the obvious fact

526: that one expects a distribution of displacements due to the finite

527: size of the bubbles. Therefore, the variation in displacements

528: from the ideal elastic behavior in region (A) is used to set a

529: minimum threshold value for $\Delta$ of $0.05\ {\rm cm}$. Bubbles

530: with a value of $\Delta$ below this threshold are considered to

531: have undergone ``elastic'' motion. Even with this cutoff, there

532: are a small number of bubbles in the tails of the distribution

533: that are classified as deviating from elastic behavior even in

534: region A. This is illustrated in Fig.~\ref{t1elastic}a-c. Each of

535: Fig.~\ref{t1elastic}a-c, represent a period of strain of 0.064 in

536: which no T1 events occur. The periods were selected from the

537: corresponding region (A - C) of Fig.~\ref{initstress}. The circles

538: indicate the location of tracked bubbles (so only a fraction of

539: the total bubbles are shown). The color of the bubbles indicates

540: the deviation from elastic behavior, with white bubbles having a

541: value of $\Delta < 0.05\ {\rm cm}$. The color code is indicated in

542: the figure.

543: 

544: Figure~\ref{t1elastic}d and e illustrate two classes of T1 events

545: that do not result in a stress drop. Fig.~\ref{t1elastic}d is from

546: region (B) of Fig.~\ref{initstress}. This illustrates an isolated

547: pair of T1 events that have an associated region in which the

548: bubbles deviate from elastic behavior. Fig.~\ref{t1elastic}e is

549: from region (C) of Fig.~\ref{initstress}. This illustrates the

550: slippage of two, short rows of bubbles due to simultaneous T1

551: events. Again, there is a relatively localized region of deviation

552: from elastic behavior associated with these T1 events.

553: 

554: \subsection{Plastic Flow Regime} \label{plasticsec}

555: 

556: In the plastic flow regime, there are two main questions regarding

557: the T1 events. First, what is the correlation between T1 events

558: and stress? Second, what is the correlation between T1 events and

559: bubble motion, as measured by either the average velocity or the

560: deviation from elastic behavior?

561: 

562: Regarding the correlation between T1 events and stress, it is

563: interesting to consider the periods of stress increase. As with

564: region B and C in Fig.~\ref{initstress}, there are often T1 events

565: during these periods of stress increase. Therefore, in general,

566: these are periods of plastic deformation, though preliminary

567: observations suggest that occasional increases exist during which

568: no T1 events occur. One way to characterize stress increases is to

569: use Eq.~\ref{Gexp} to calculate an effective shear modulus, $G$,

570: for each separate period of stress increase. This can than be

571: correlated with the number of T1 events that occur in that period.

572: A preliminary measurement of this is shown in Fig.~\ref{GvT1}.

573: This result is preliminary because only a fraction of the sample

574: was viewed. Therefore, the results for the number of T1 events

575: represent a lower bound. However, it is interesting that the data

576: all falls below the straight line, suggesting a correlation

577: between $G$ and the number of T1 events, as expected.

578: 

579: It is natural to expect that one necessary condition for a T1

580: event to occur is that the local stress exceeds some critical

581: value. This would suggest a correlation between the stress and the

582: location of the T1 events. Figure~\ref{t1pos} illustrates that no

583: correlation exists between the stress on the {\it inner cylinder}

584: and the radial positions of T1 events. One would expect such a

585: correlation if the stress field was given by the continuum limit,

586: which in a Couette geometry is $\sigma(r) = (\sigma(r_i)

587: r_i^2)/r^2$, and the critical stress for a T1 event was spatially

588: uniform. Under these conditions, for each $\sigma(r_i)$, there is

589: a maximum $r$ at which T1 events can occur. This is set by the

590: critical stress required for the generation of a T1 event. The

591: lack of a correlation suggests that at least one, if not both, of

592: these assumptions is false. In fact, work in other systems

593: suggests that both assumptions are false. Given the direct

594: measurement of stress chains in granular matter \cite{HBV99}, it

595: is reasonable to expect such chains in the bubble raft. This would

596: represent a breakdown of the continuum assumption for the stress

597: distribution. Also, simulations of amorphous metal have shown the

598: existence of localized, high stress regions (referred to as

599: $\tau$-defects) that are the sources of local flow \cite{SVE83}.

600: In other contexts, models and simulations have suggested the

601: existence of ``weak'' zones in complex fluids

602: \cite{LRMGW96,SLHC97,LL97} that are the source of viscous-like

603: behavior. A more detailed study of these issues will required

604: improved images, but the current work is very suggestive.

605: 

606: To summarize the average properties of T1 events as a function of

607: strain, Fig.~\ref{t1rate} plots the number of T1 events per bubble

608: versus strain. Again, this is shown simultaneously with the stress

609: versus strain curve to illustrate the general correlation between

610: the size of the stress drops and the total number of T1 events.

611: One observes that most stress drops consist of a cascade of events

612: throughout the stress drop. As with the stress increases, detailed

613: correlations between the size of a stress drop and the number of

614: T1 events will require images of the entire sample. However, one

615: can compute $R_{T1}$. For this shear rate, $R_{T1} = 0.18 \pm

616: 0.01$, in reasonable agreement with both the bubble model and

617: Langmuir monolayer foam.

618: 

619: The next question is the connection between velocity profiles and

620: T1 events. Based on the results of Ref~\cite{LCD04}, it is known

621: that there exists a shear discontinuity at $r_c = 6.7\ {\rm cm}$

622: for the system reported on here. Therefore, there is no

623: expectation of strong localization of the T1 events are reported

624: in Ref.~\cite{KD03} because there is no shear localization.

625: However, one might expect a connection between the radial

626: distribution of T1 events and the shear discontinuity.

627: 

628: The shear discontinuity divides the system into two regimes. Below

629: $r_c$, the average velocity is consistent with that of a power-law

630: fluid. Above $r_c$, the systems acts like an elastic solid.

631: Figure~\ref{t1vel} illustrates the connection between the average

632: velocity profile and the spatial distribution of T1 events. The

633: vertical line indicates the spatial location of the shear

634: discontinuity \cite{LCD04}. The basic shape of the distribution is

635: similar to that found in the simulations reported in

636: Ref.~\cite{KD03}. There is a ``peak'' at smaller radii, with the

637: distribution tailing off as one goes to larger radii. The main

638: difference is the location of the cutoff in the T1 distribution.

639: As reported in Ref.~\cite{KD03}, the cutoff in velocity and T1

640: events is at essentially the same radius. In contrast, for the

641: system reported on here, there is no obvious signature in the

642: distribution of T1 events at the shear discontinuity (see

643: Fig.~\ref{t1vel}). This may be due to the fact that even though

644: the shear-rate is zero, the bubbles are still moving near the

645: outer wall, and differences in bubble size may lead to T1 events.

646: Also, it may be an artifact of how close the shear discontinuity

647: is to the outer wall. Future work in larger systems is needed to

648: better understand this issue.

649: 

650: In order to better understand the detailed connection between T1

651: events and stress drops, two short periods of strain are

652: highlighted, as indicated in Fig.~\ref{t1highlight}. These are

653: segments of the data presented in Fig.~\ref{t1pos}. The period of

654: strain illustrate in Fig.~\ref{t1highlight}a was selected to

655: highlight the nature of potential correlations between the stress

656: behavior and the number of T1 events. First, the initial elastic

657: rise A is included for comparison with the stress increase in the

658: interval labelled C. During the initial rise, there is only one

659: observed T1 event, and the effective elastic modulus is $G = 5.6

660: \pm 0.1\ {\rm mN/m}$. In contrast, during region C, there are 9

661: observed T1 events, and the effective elastic modulus is $G = 2.1

662: \pm 0.1\ {\rm mN/m}$. These results reinforce the general

663: connection between number of T1 events and effective elastic

664: modulus discussed with respect to Fig.~\ref{GvT1}. In contrast,

665: the regions labelled B - E all have roughly the same number of T1

666: events. Yet, region B and E are stress drops. Region C is a period

667: of stress increase, and region D is a slight decrease in stress.

668: One difficulty in drawing definitive conclusions from this data is

669: the fact that only part of the system is being viewed. However,

670: this strongly suggests that the additional bubble motions, not

671: just the T1 events, play an important role in determining the

672: overall stress evolution.

673: 

674: The interval illustrated by Fig.~\ref{t1highlight}b was selected

675: to make connections with the velocity profiles reported in

676: Ref.~\cite{LCD04} in an attempt to better understand the shear

677: discontinuity that occurs at $r_c$. Here, the average bubble

678: displacements are measured, but these are directly related to

679: average velocities. This sequence is particularly interesting

680: because there are three stress drops that occur at different

681: average stress values, and the drop at the lowest average stress

682: (region E) exhibits the larger value of $r_c$ \cite{LCD04}. (We

683: are considering the behavior in region B and C as two separate

684: stress drops because of the short plateau between them. However,

685: this points out the issue regarding the definition of ``events''

686: as discussed earlier in the context of Ref.~\cite{ML04}.)

687: 

688: The spatial distribution of T1 events and bubble deviations from

689: elastic behavior are given in Fig.~\ref{t1flow} using the same

690: criteria as described for Fig.~\ref{t1elastic} in

691: Sec.~\ref{plasticsec}. White bubbles represent essentially elastic

692: behavior, and the color of the other bubbles is the degree to

693: which their motion deviates from elastic. The letters in

694: Fig.~\ref{t1highlight}b correspond to the labelling of

695: Fig.~\ref{t1flow}.  One observes very similar distributions of T1

696: events for all three stress drops (see Fig.~\ref{t1flow}b,c, and

697: e). If one looks carefully, the distinguishing factor appears to

698: be the number of bubbles deviating from elastic behavior at any

699: given radius. This is made clearer by considering the average

700: bubble displacement as a function of radial position, as

701: illustrated in Fig.~\ref{bubdispl}.

702: 

703: Figure~\ref{bubdispl} is a plot of $\Delta \theta/ \Omega \Delta

704: t$ versus radial position. The values of $\Delta \theta$ are

705: computed by dividing the system into 20 equally spaced radial bins

706: and averaging the angular displacements over all bubbles in a

707: given bin over the time interval of interest. The time intervals

708: are selected so that they match the strain intervals indicated in

709: Fig.~\ref{t1highlight}b. For comparison, the displacements during

710: the essentially flat regions in stress are given as open symbols,

711: and the displacements during the stress drops are given as closed

712: symbols. The angular displacement is normalized by $\Omega\Delta

713: t$. The solid line is the ``ideal'' elastic behavior given by the

714: fit to the data in Fig.~\ref{elflow}. One can see that for

715: interval E (solid triangles), the deviation from the expected

716: elastic displacement occurs at the largest value of $r_c$. This is

717: consistent with the results reported in Ref.~\cite{LCD04} for the

718: velocity profiles. What is new in these results is the ability to

719: correlate the location of T1 events during a stress drop and the

720: location of the deviation from elastic behavior.

721: 

722: For event E, there are two clear deviations from elastic behavior,

723: as shown in Fig.~\ref{bubdispl}: (a) at $r = 6.48\ {\rm cm}$ there

724: is a positive deviation; and (b) at $r = 6.16\ {\rm cm}$ there is

725: a negative deviation. During the drop, the maximal radial position

726: of a T1 event is $r = 6.24\ {\rm cm}$. The existence of positive

727: and negative deviations is consistent with the bubbles associated

728: with a T1 event moving both forward and backward relative to the

729: average flow. The average displacements during B and C are

730: essentially identical. However, for C more then B there is some

731: evidence for a positive and negative deviation at $r = 5.84\ {\rm

732: cm}$ and at $r = 5.44\ {\rm cm}$, respectively. For these drops,

733: the maximal radial position of a T1 event is $r = 6.29\ {\rm cm}$.

734: Comparing these numbers strongly suggests that the location of T1

735: events is not the main contribution to the determination of the

736: deviation from elastic behavior, and hence, the determination of

737: $r_c$. Instead, it is the detailed motion of the surrounding

738: bubbles. Interestingly, the greatest difference between the two

739: events in terms of T1 position is in the precursor to the drops;

740: yet the precursors have very similar angular displacements (open

741: symbols in Fig.~\ref{bubdispl}). During the interval labelled D in

742: Fig.~\ref{t1highlight}b, one observes T1 events as far out as $r =

743: 6.98\ {\rm cm}$. In contrast, during the interval labelled A in

744: Fig.~\ref{t1highlight}b, one only observes T1 events as far out as

745: $r = 5.87\ {\rm cm}$. Presumably these events play an important

746: role in establishing the local stress fields that govern the

747: bubble motions during the subsequent stress drop.

748: 

749: \section{Discussion} \label{summary}

750: 

751: Even though various aspects of the work presented here are

752: preliminary in the sense that only a portion of the entire raft

753: was studied, a number of questions regarding the role of T1 events

754: in the macroscopic response of a bubble raft to flow have been

755: addressed. First, the contribution of T1 events to the effective

756: shear modulus was considered. T1 events during periods of stress

757: increase effectively lower the shear modulus of the bubble raft.

758: Similarly, during stress drops, there is a correlation between the

759: size of the drop and the total number of T1 events. Future work is

760: required to establish a more detailed correlation between the

761: number of T1 events and the effective shear modulus and size of

762: stress drops.

763: 

764: Correlations between positions of T1 events, average stress,

765: individual bubble displacements, and average angular displacements

766: of bubbles were considered. A general picture that emerges from

767: these measurements is the importance of understanding the local

768: stress field and the local geometry of bubbles. For example, an

769: investigation of individual bubble motions before and during a

770: stress drop (see discussion of Figs.~\ref{t1highlight} and

771: \ref{t1flow}) shows that the radial distribution of T1 events can

772: not be understood in terms of a simple continuum model and single

773: stress threshold. The T1 events in the strain period immediately

774: prior to a stress drop play an important role in establishing the

775: local stress field and geometric relations between bubbles that

776: sets the subsequent motions. For example, the two different stress

777: drops highlighted in Fig.~\ref{t1highlight} exhibited similar

778: distributions of T1 events, but the deviations from elastic

779: behavior and the average displacements were very different. The

780: main differences between the events was in the distribution of

781: precursor T1 events, not in the average bubble motions.

782: 

783: The connection between T1 events and the position of the shear

784: discontinuity was also considered. Both in terms of the average

785: properties (see Fig.~\ref{t1vel}) and the short time motions (see

786: Fig.~\ref{bubdispl}). There is no clear evidence for a connection

787: between the positions of T1 events and the shear discontinuity,

788: but larger system sizes need to be studied. However, there may be

789: an indirect connection through the stress relaxation and

790: subsequent motion of surrounding bubbles. Indirectly, these

791: measurements have some potential implications for the simulations

792: of the modified vertex model \cite{KD03}. These simulations

793: illustrate that a localization of T1 events can lead to a shear

794: localization. This system does not exhibit localization of either

795: the T1 events or the shear. This indirectly supports the

796: connection between T1 event localization and shear localization.

797: What remains an important questions is why would T1 events

798: localize in one case and not the other? An obvious difference

799: between the T1 events in the bubble raft and in the simulation is

800: the duration of the T1 events. In the model, the T1 events all

801: occur on a very short time scale, by construction. For the bubble

802: raft, there is a distribution of times for the duration of T1

803: events. Some events occur very slowly (over 10 - 20 seconds). It

804: is these difference in duration that may modify the impact on the

805: local stress. Again, this point to the importance of understanding

806: the local stress fields generated by the T1 events, and not just

807: the distribution of the events themselves. Furthermore, as part of

808: the future work that focuses on local stress fields, it will be

809: important to correlate the changes in local stress with the

810: duration of the T1 events.

811: 

812: The measurements reported here focused on bubble displacements and

813: T1 events. Where possible, comparisons with the bubble model show

814: quantitative agreement, such as for $R_{T1}$. This adds support to

815: previous results with the bubble raft that were also in general

816: agreement with the bubble model \cite{LTD02,PD03}. Having strong

817: agreement between the experiments and a theoretical model is

818: useful for the next stage of experimental studies. Essentially all

819: of the results point to the need for measurements of the local

820: stress field. Future experimental work is planned that will use

821: the bubble distortion as a direct measure of local stress, as has

822: been done with other foam systems \cite{RK00,BB02b,AJGG03,JG03}.

823: Close contact with simulations that focus on the stress released

824: by T1 events and STZ's, as well as experimental studies of

825: granular material, will be important for understanding this future

826: work.

827: 

828: \begin{acknowledgments}

829: 

830: This work was supported by the Department of Energy grant

831: DE-FG02-03ED46071, the Research Corporation, and the Alfred P.

832: Sloan Foundation. I thank John Lauridsen for use of his video data

833: of bubble rafts that was taken while an undergraduate at the

834: University of California, Irvine. I thank Craig Maloney, Corey

835: O`Hern, Michael Falk, and Georges Debr\'{e}geas for fruitful

836: discussions.

837: 

838: \end{acknowledgments}

839: 

840: %\bibliography{d04}

841: \begin{thebibliography}{50}

842: \expandafter\ifx\csname

843: natexlab\endcsname\relax\def\natexlab#1{#1}\fi

844: \expandafter\ifx\csname bibnamefont\endcsname\relax

845:   \def\bibnamefont#1{#1}\fi

846: \expandafter\ifx\csname bibfnamefont\endcsname\relax

847:   \def\bibfnamefont#1{#1}\fi

848: \expandafter\ifx\csname citenamefont\endcsname\relax

849:   \def\citenamefont#1{#1}\fi

850: \expandafter\ifx\csname url\endcsname\relax

851:   \def\url#1{\texttt{#1}}\fi

852: \expandafter\ifx\csname

853: urlprefix\endcsname\relax\def\urlprefix{URL }\fi

854: \providecommand{\bibinfo}[2]{#2}

855: \providecommand{\eprint}[2][]{\url{#2}}

856: 

857: \bibitem[{\citenamefont{Argon and Kuo}(1979)}]{AK79}

858: \bibinfo{author}{\bibfnamefont{A.~S.} \bibnamefont{Argon}} \bibnamefont{and}

859:   \bibinfo{author}{\bibfnamefont{H.~Y.} \bibnamefont{Kuo}},

860:   \bibinfo{journal}{Mat. Sci. and Eng.} \textbf{\bibinfo{volume}{39}},

861:   \bibinfo{pages}{101} (\bibinfo{year}{1979}).

862: 

863: \bibitem[{\citenamefont{Bragg and Lomer}(1949)}]{BL49}

864: \bibinfo{author}{\bibfnamefont{L.}~\bibnamefont{Bragg}} \bibnamefont{and}

865:   \bibinfo{author}{\bibfnamefont{W.~M.} \bibnamefont{Lomer}},

866:   \bibinfo{journal}{Proc. R. Soc. London, Ser. A}

867:   \textbf{\bibinfo{volume}{196}}, \bibinfo{pages}{171} (\bibinfo{year}{1949}).

868: 

869: \bibitem[{\citenamefont{Kader and Earnshaw}(1999)}]{KE99}

870: \bibinfo{author}{\bibfnamefont{A.~A.} \bibnamefont{Kader}} \bibnamefont{and}

871:   \bibinfo{author}{\bibfnamefont{J.~C.} \bibnamefont{Earnshaw}},

872:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{82}},

873:   \bibinfo{pages}{2610} (\bibinfo{year}{1999}).

874: 

875: \bibitem[{\citenamefont{Lauridsen et~al.}(2002)\citenamefont{Lauridsen,

876:   Twardos, and Dennin}}]{LTD02}

877: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Lauridsen}},

878:   \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Twardos}}, \bibnamefont{and}

879:   \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Dennin}},

880:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{89}},

881:   \bibinfo{pages}{098303} (\bibinfo{year}{2002}).

882: 

883: \bibitem[{\citenamefont{Falk and Langer}(1998)}]{FL98}

884: \bibinfo{author}{\bibfnamefont{M.~L.} \bibnamefont{Falk}} \bibnamefont{and}

885:   \bibinfo{author}{\bibfnamefont{J.~S.} \bibnamefont{Langer}},

886:   \bibinfo{journal}{Phys. Rev. E} \textbf{\bibinfo{volume}{57}},

887:   \bibinfo{pages}{7192} (\bibinfo{year}{1998}).

888: 

889: \bibitem[{\citenamefont{Onuki}(2003)}]{O03}

890: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Onuki}},

891:   \bibinfo{journal}{Phys. Rev. E} \textbf{\bibinfo{volume}{68}},

892:   \bibinfo{pages}{061502} (\bibinfo{year}{2003}).

893: 

894: \bibitem[{\citenamefont{Picard et~al.}(2004)\citenamefont{Picard, Ajdari,

895:   Lequeux, and Bocquet}}]{PALB04}

896: \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Picard}},

897:   \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Ajdari}},

898:   \bibinfo{author}{\bibfnamefont{F.}~\bibnamefont{Lequeux}}, \bibnamefont{and}

899:   \bibinfo{author}{\bibfnamefont{L.}~\bibnamefont{Bocquet}},

900:   \bibinfo{journal}{cond-mat/0403647}  (\bibinfo{year}{2004}).

901: 

902: \bibitem[{\citenamefont{Bulatov and Argon}(1994{\natexlab{a}})}]{BA94a}

903: \bibinfo{author}{\bibfnamefont{V.~V.} \bibnamefont{Bulatov}} \bibnamefont{and}

904:   \bibinfo{author}{\bibfnamefont{A.~S.} \bibnamefont{Argon}},

905:   \bibinfo{journal}{Modell. Simul. Mater. Sci. Eng.}

906:   \textbf{\bibinfo{volume}{2}}, \bibinfo{pages}{167}

907:   (\bibinfo{year}{1994}{\natexlab{a}}).

908: 

909: \bibitem[{\citenamefont{Bulatov and Argon}(1994{\natexlab{b}})}]{BA94b}

910: \bibinfo{author}{\bibfnamefont{V.~V.} \bibnamefont{Bulatov}} \bibnamefont{and}

911:   \bibinfo{author}{\bibfnamefont{A.~S.} \bibnamefont{Argon}},

912:   \bibinfo{journal}{Modell. Simul. Mater. Sci. Eng.}

913:   \textbf{\bibinfo{volume}{2}}, \bibinfo{pages}{185}

914:   (\bibinfo{year}{1994}{\natexlab{b}}).

915: 

916: \bibitem[{\citenamefont{Bulatov and Argon}(1994{\natexlab{c}})}]{BA94c}

917: \bibinfo{author}{\bibfnamefont{V.~V.} \bibnamefont{Bulatov}} \bibnamefont{and}

918:   \bibinfo{author}{\bibfnamefont{A.~S.} \bibnamefont{Argon}},

919:   \bibinfo{journal}{Modell. Simul. Mater. Sci. Eng.}

920:   \textbf{\bibinfo{volume}{2}}, \bibinfo{pages}{203}

921:   (\bibinfo{year}{1994}{\natexlab{c}}).

922: 

923: \bibitem[{\citenamefont{Baret et~al.}(2002)\citenamefont{Baret, Vandembroucq,

924:   and Roux}}]{BVR02}

925: \bibinfo{author}{\bibfnamefont{J.~C.} \bibnamefont{Baret}},

926:   \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{Vandembroucq}},

927:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Roux}},

928:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{89}},

929:   \bibinfo{pages}{195506} (\bibinfo{year}{2002}).

930: 

931: \bibitem[{\citenamefont{Srolovitz et~al.}(1983)\citenamefont{Srolovitz, Vitek,

932:   and Egami}}]{SVE83}

933: \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{Srolovitz}},

934:   \bibinfo{author}{\bibfnamefont{V.}~\bibnamefont{Vitek}}, \bibnamefont{and}

935:   \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Egami}},

936:   \bibinfo{journal}{Acta. Met.} \textbf{\bibinfo{volume}{31}},

937:   \bibinfo{pages}{335} (\bibinfo{year}{1983}).

938: 

939: \bibitem[{\citenamefont{Kawasaki et~al.}(1989)\citenamefont{Kawasaki, Nagai,

940:   and Nakashima}}]{KNN89}

941: \bibinfo{author}{\bibfnamefont{K.}~\bibnamefont{Kawasaki}},

942:   \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Nagai}}, \bibnamefont{and}

943:   \bibinfo{author}{\bibfnamefont{K.}~\bibnamefont{Nakashima}},

944:   \bibinfo{journal}{Phil. Mag. B} \textbf{\bibinfo{volume}{60}},

945:   \bibinfo{pages}{399} (\bibinfo{year}{1989}).

946: 

947: \bibitem[{\citenamefont{Kawasaki et~al.}(1992)\citenamefont{Kawasaki, Okuzono,

948:   Kawakatsu, and Nagai}}]{KOKN95}

949: \bibinfo{author}{\bibfnamefont{K.}~\bibnamefont{Kawasaki}},

950:   \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Okuzono}},

951:   \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Kawakatsu}},

952:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Nagai}}, in

953:   \emph{\bibinfo{booktitle}{Proc. Int. Workshop of Physics of Pattern

954:   Formation}}, edited by \bibinfo{editor}{\bibfnamefont{S.}~\bibnamefont{Kai}}

955:   (\bibinfo{publisher}{Singapore: World Scientific}, \bibinfo{year}{1992}).

956: 

957: \bibitem[{\citenamefont{Okuzono and Kawasaki}(1995)}]{OK95}

958: \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Okuzono}} \bibnamefont{and}

959:   \bibinfo{author}{\bibfnamefont{K.}~\bibnamefont{Kawasaki}},

960:   \bibinfo{journal}{Phys. Rev. E} \textbf{\bibinfo{volume}{51}},

961:   \bibinfo{pages}{1246} (\bibinfo{year}{1995}).

962: 

963: \bibitem[{\citenamefont{Durian}(1995)}]{D95}

964: \bibinfo{author}{\bibfnamefont{D.~J.} \bibnamefont{Durian}},

965:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{75}},

966:   \bibinfo{pages}{4780} (\bibinfo{year}{1995}).

967: 

968: \bibitem[{\citenamefont{Durian}(1997)}]{D97}

969: \bibinfo{author}{\bibfnamefont{D.~J.} \bibnamefont{Durian}},

970:   \bibinfo{journal}{Phys. Rev. E} \textbf{\bibinfo{volume}{55}},

971:   \bibinfo{pages}{1739} (\bibinfo{year}{1997}).

972: 

973: \bibitem[{\citenamefont{Weaire et~al.}(1992)\citenamefont{Weaire, Bolton,

974:   Herdtle, and Aref}}]{WBHA92}

975: \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{Weaire}},

976:   \bibinfo{author}{\bibfnamefont{F.}~\bibnamefont{Bolton}},

977:   \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Herdtle}}, \bibnamefont{and}

978:   \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Aref}},

979:   \bibinfo{journal}{Phil. Mag. Lett.} \textbf{\bibinfo{volume}{66}},

980:   \bibinfo{pages}{293} (\bibinfo{year}{1992}).

981: 

982: \bibitem[{\citenamefont{Hutzler et~al.}(1995)\citenamefont{Hutzler, Weaire, and

983:   Bolton}}]{HWB95}

984: \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Hutzler}},

985:   \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{Weaire}}, \bibnamefont{and}

986:   \bibinfo{author}{\bibfnamefont{F.}~\bibnamefont{Bolton}},

987:   \bibinfo{journal}{Phil. Mag. B} \textbf{\bibinfo{volume}{71}},

988:   \bibinfo{pages}{277} (\bibinfo{year}{1995}).

989: 

990: \bibitem[{\citenamefont{Jiang et~al.}(1999)\citenamefont{Jiang, Swart, Saxena,

991:   Asipauskas, and Glazier}}]{JSSAG99}

992: \bibinfo{author}{\bibfnamefont{Y.}~\bibnamefont{Jiang}},

993:   \bibinfo{author}{\bibfnamefont{P.~J.} \bibnamefont{Swart}},

994:   \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Saxena}},

995:   \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Asipauskas}},

996:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{J.~A.}

997:   \bibnamefont{Glazier}}, \bibinfo{journal}{Phys. Rev. E}

998:   \textbf{\bibinfo{volume}{59}}, \bibinfo{pages}{5819} (\bibinfo{year}{1999}).

999: 

1000: \bibitem[{\citenamefont{Tewari et~al.}(1999)\citenamefont{Tewari, Schiemann,

1001:   Durian, Knobler, Langer, and Liu}}]{TSDKLL99}

1002: \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Tewari}},

1003:   \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{Schiemann}},

1004:   \bibinfo{author}{\bibfnamefont{D.~J.} \bibnamefont{Durian}},

1005:   \bibinfo{author}{\bibfnamefont{C.~M.} \bibnamefont{Knobler}},

1006:   \bibinfo{author}{\bibfnamefont{S.~A.} \bibnamefont{Langer}},

1007:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{A.~J.} \bibnamefont{Liu}},

1008:   \bibinfo{journal}{Phys. Rev. E} \textbf{\bibinfo{volume}{60}},

1009:   \bibinfo{pages}{4385} (\bibinfo{year}{1999}).

1010: 

1011: \bibitem[{\citenamefont{Kabla and Debr\'{e}geas}(2003)}]{KD03}

1012: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Kabla}} \bibnamefont{and}

1013:   \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Debr\'{e}geas}},

1014:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{90}},

1015:   \bibinfo{pages}{258303} (\bibinfo{year}{2003}).

1016: 

1017: \bibitem[{\citenamefont{Stavans}(1993)}]{S93}

1018: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Stavans}},

1019:   \bibinfo{journal}{Rep. Prog. Phys.} \textbf{\bibinfo{volume}{56}},

1020:   \bibinfo{pages}{733} (\bibinfo{year}{1993}).

1021: 

1022: \bibitem[{\citenamefont{Kraynik}(1988)}]{K88}

1023: \bibinfo{author}{\bibfnamefont{A.~M.} \bibnamefont{Kraynik}},

1024:   \bibinfo{journal}{Ann. Rev. Fluid Mech.} \textbf{\bibinfo{volume}{20}},

1025:   \bibinfo{pages}{325} (\bibinfo{year}{1988}).

1026: 

1027: \bibitem[{\citenamefont{Weaire and Hutzler}(1999)}]{WH99}

1028: \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{Weaire}} \bibnamefont{and}

1029:   \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Hutzler}},

1030:   \emph{\bibinfo{title}{The Physics of Foams}} (\bibinfo{publisher}{Claredon

1031:   Press, Oxford}, \bibinfo{year}{1999}).

1032: 

1033: \bibitem[{\citenamefont{Argon}(1979)}]{A79}

1034: \bibinfo{author}{\bibfnamefont{A.~S.} \bibnamefont{Argon}},

1035:   \bibinfo{journal}{Acta. Met.} \textbf{\bibinfo{volume}{27}},

1036:   \bibinfo{pages}{47} (\bibinfo{year}{1979}).

1037: 

1038: \bibitem[{\citenamefont{Malandro and Lacks}(1999)}]{ML99}

1039: \bibinfo{author}{\bibfnamefont{D.~L.} \bibnamefont{Malandro}} \bibnamefont{and}

1040:   \bibinfo{author}{\bibfnamefont{D.~J.} \bibnamefont{Lacks}},

1041:   \bibinfo{journal}{J. Chem. Phys} \textbf{\bibinfo{volume}{110}},

1042:   \bibinfo{pages}{4593} (\bibinfo{year}{1999}).

1043: 

1044: \bibitem[{\citenamefont{Maloney and Lema\^{i}tre}(2004)}]{ML04}

1045: \bibinfo{author}{\bibfnamefont{C.}~\bibnamefont{Maloney}} \bibnamefont{and}

1046:   \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Lema\^{i}tre}},

1047:   \bibinfo{journal}{cond-mat/042148}  (\bibinfo{year}{2004}).

1048: 

1049: \bibitem[{\citenamefont{Mazuyer et~al.}(1989)\citenamefont{Mazuyer, Georges,

1050:   and Cambou}}]{MGC89}

1051: \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{Mazuyer}},

1052:   \bibinfo{author}{\bibfnamefont{J.~M.} \bibnamefont{Georges}},

1053:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{B.}~\bibnamefont{Cambou}},

1054:   \bibinfo{journal}{J. Phys. France} \textbf{\bibinfo{volume}{49}},

1055:   \bibinfo{pages}{1057} (\bibinfo{year}{1989}).

1056: 

1057: \bibitem[{\citenamefont{Shi and Argon}(1982)}]{SA82}

1058: \bibinfo{author}{\bibfnamefont{L.~T.} \bibnamefont{Shi}} \bibnamefont{and}

1059:   \bibinfo{author}{\bibfnamefont{A.~S.} \bibnamefont{Argon}},

1060:   \bibinfo{journal}{Philo. Mag. A} \textbf{\bibinfo{volume}{46}},

1061:   \bibinfo{pages}{255} (\bibinfo{year}{1982}).

1062: 

1063: \bibitem[{\citenamefont{Dennin and Knobler}(1997)}]{DK97}

1064: \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Dennin}} \bibnamefont{and}

1065:   \bibinfo{author}{\bibfnamefont{C.~M.} \bibnamefont{Knobler}},

1066:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{78}},

1067:   \bibinfo{pages}{2485} (\bibinfo{year}{1997}).

1068: 

1069: \bibitem[{BOO()}]{BOOKS}

1070: \bibinfo{note}{Various books cover both the modelling and experimental

1071:   measurement of yield stress materials, and complex fluids in general. Two

1072:   examples are R. B. Bird, R. C. Armstrong, and O. Hassage, {\it Dynamics of

1073:   Polymer Liquids} (Wiley, New York, 1977) and C. Macosko, {\it Rheology

1074:   Principles, Measurements, and Applications} (VCH Publishers, New York,

1075:   1994).}

1076: 

1077: \bibitem[{\citenamefont{Losert et~al.}(2000)\citenamefont{Losert, Bocquet,

1078:   Lubensky, and Gollub}}]{LBLG00}

1079: \bibinfo{author}{\bibfnamefont{W.}~\bibnamefont{Losert}},

1080:   \bibinfo{author}{\bibfnamefont{L.}~\bibnamefont{Bocquet}},

1081:   \bibinfo{author}{\bibfnamefont{T.~C.} \bibnamefont{Lubensky}},

1082:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{J.~P.}

1083:   \bibnamefont{Gollub}}, \bibinfo{journal}{Phys. Rev. Lett.}

1084:   \textbf{\bibinfo{volume}{85}}, \bibinfo{pages}{1428} (\bibinfo{year}{2000}).

1085: 

1086: \bibitem[{\citenamefont{Mueth et~al.}(2000)\citenamefont{Mueth, Debregeas,

1087:   Karczmar, Eng, Nagel, and Jaeger}}]{MDKENJ00}

1088: \bibinfo{author}{\bibfnamefont{D.~M.} \bibnamefont{Mueth}},

1089:   \bibinfo{author}{\bibfnamefont{G.~F.} \bibnamefont{Debregeas}},

1090:   \bibinfo{author}{\bibfnamefont{G.~S.} \bibnamefont{Karczmar}},

1091:   \bibinfo{author}{\bibfnamefont{P.~J.} \bibnamefont{Eng}},

1092:   \bibinfo{author}{\bibfnamefont{S.~R.} \bibnamefont{Nagel}}, \bibnamefont{and}

1093:   \bibinfo{author}{\bibfnamefont{H.~M.} \bibnamefont{Jaeger}},

1094:   \bibinfo{journal}{Nature} \textbf{\bibinfo{volume}{406}},

1095:   \bibinfo{pages}{385} (\bibinfo{year}{2000}).

1096: 

1097: \bibitem[{\citenamefont{Howell et~al.}(1999)\citenamefont{Howell, Behringer,

1098:   and Veje}}]{HBV99}

1099: \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{Howell}},

1100:   \bibinfo{author}{\bibfnamefont{R.~P.} \bibnamefont{Behringer}},

1101:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{C.}~\bibnamefont{Veje}},

1102:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{82}},

1103:   \bibinfo{pages}{5241} (\bibinfo{year}{1999}).

1104: 

1105: \bibitem[{\citenamefont{Coussot et~al.}(2002)\citenamefont{Coussot, Raynaud,

1106:   Bertrand, Moucheront, Guilbaud, and Huynh}}]{CRBMGH02}

1107: \bibinfo{author}{\bibfnamefont{P.}~\bibnamefont{Coussot}},

1108:   \bibinfo{author}{\bibfnamefont{J.~S.} \bibnamefont{Raynaud}},

1109:   \bibinfo{author}{\bibfnamefont{F.}~\bibnamefont{Bertrand}},

1110:   \bibinfo{author}{\bibfnamefont{P.}~\bibnamefont{Moucheront}},

1111:   \bibinfo{author}{\bibfnamefont{J.~P.} \bibnamefont{Guilbaud}},

1112:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{H.~T.} \bibnamefont{Huynh}},

1113:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{88}},

1114:   \bibinfo{pages}{218301} (\bibinfo{year}{2002}).

1115: 

1116: \bibitem[{\citenamefont{Cruz et~al.}(2002)\citenamefont{Cruz, Chevoir, Bonn,

1117:   and Coussot}}]{DCBC02}

1118: \bibinfo{author}{\bibfnamefont{F.~D.} \bibnamefont{Cruz}},

1119:   \bibinfo{author}{\bibfnamefont{F.}~\bibnamefont{Chevoir}},

1120:   \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{Bonn}}, \bibnamefont{and}

1121:   \bibinfo{author}{\bibfnamefont{P.}~\bibnamefont{Coussot}},

1122:   \bibinfo{journal}{Phys. Rev. E} \textbf{\bibinfo{volume}{66}},

1123:   \bibinfo{pages}{051305} (\bibinfo{year}{2002}).

1124: 

1125: \bibitem[{\citenamefont{Debr\'{e}geas et~al.}(2001)\citenamefont{Debr\'{e}geas,

1126:   Tabuteau, and di~Meglio}}]{DTM01}

1127: \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Debr\'{e}geas}},

1128:   \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Tabuteau}}, \bibnamefont{and}

1129:   \bibinfo{author}{\bibfnamefont{J.~M.} \bibnamefont{di~Meglio}},

1130:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{87}},

1131:   \bibinfo{pages}{178305} (\bibinfo{year}{2001}).

1132: 

1133: \bibitem[{\citenamefont{Lauridsen et~al.}(2004)\citenamefont{Lauridsen, Chanan,

1134:   and Dennin}}]{LCD04}

1135: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Lauridsen}},

1136:   \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Chanan}}, \bibnamefont{and}

1137:   \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Dennin}},

1138:   \bibinfo{journal}{Phys. Rev. Lett.} p. \bibinfo{pages}{accepted}

1139:   (\bibinfo{year}{2004}).

1140: 

1141: \bibitem[{\citenamefont{Ghaskadvi and Dennin}(1998)}]{app}

1142: \bibinfo{author}{\bibfnamefont{R.~S.} \bibnamefont{Ghaskadvi}}

1143:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Dennin}},

1144:   \bibinfo{journal}{Rev. Sci. Instr.} \textbf{\bibinfo{volume}{69}},

1145:   \bibinfo{pages}{3568} (\bibinfo{year}{1998}).

1146: 

1147: \bibitem[{tes()}]{test}

1148: \bibinfo{note}{The detection of T1 events by hand was tested using data

1149:   generated based on the bubble model. The number of T1 events detected by hand

1150:   and with the computer agreed well within expected statistical errors.}

1151: 

1152: \bibitem[{\citenamefont{Pratt and Dennin}(2003)}]{PD03}

1153: \bibinfo{author}{\bibfnamefont{E.}~\bibnamefont{Pratt}} \bibnamefont{and}

1154:   \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Dennin}},

1155:   \bibinfo{journal}{Phys. Rev. E} \textbf{\bibinfo{volume}{67}},

1156:   \bibinfo{pages}{051402} (\bibinfo{year}{2003}).

1157: 

1158: \bibitem[{\citenamefont{Bird et~al.}(1977)\citenamefont{Bird, Armstrong, and

1159:   Hassuage}}]{BAH77}

1160: \bibinfo{author}{\bibfnamefont{R.~B.} \bibnamefont{Bird}},

1161:   \bibinfo{author}{\bibfnamefont{R.~C.} \bibnamefont{Armstrong}},

1162:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{O.}~\bibnamefont{Hassuage}},

1163:   \emph{\bibinfo{title}{Dynamics of Polymer Liquids}}

1164:   (\bibinfo{publisher}{Wiley, New York}, \bibinfo{year}{1977}).

1165: 

1166: \bibitem[{\citenamefont{Liu et~al.}(1996)\citenamefont{Liu, Ramaswamy, Mason,

1167:   Gang, and Weitz}}]{LRMGW96}

1168: \bibinfo{author}{\bibfnamefont{A.~J.} \bibnamefont{Liu}},

1169:   \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Ramaswamy}},

1170:   \bibinfo{author}{\bibfnamefont{T.~G.} \bibnamefont{Mason}},

1171:   \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Gang}}, \bibnamefont{and}

1172:   \bibinfo{author}{\bibfnamefont{D.~A.} \bibnamefont{Weitz}},

1173:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{76}},

1174:   \bibinfo{pages}{3017} (\bibinfo{year}{1996}).

1175: 

1176: \bibitem[{\citenamefont{Sollich et~al.}(1997)\citenamefont{Sollich, Lequeux,

1177:   H\'{e}braud, and Cates}}]{SLHC97}

1178: \bibinfo{author}{\bibfnamefont{P.}~\bibnamefont{Sollich}},

1179:   \bibinfo{author}{\bibfnamefont{F.}~\bibnamefont{Lequeux}},

1180:   \bibinfo{author}{\bibfnamefont{P.}~\bibnamefont{H\'{e}braud}},

1181:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{M.~E.} \bibnamefont{Cates}},

1182:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{78}},

1183:   \bibinfo{pages}{2020} (\bibinfo{year}{1997}).

1184: 

1185: \bibitem[{\citenamefont{Langer and Liu}(1997)}]{LL97}

1186: \bibinfo{author}{\bibfnamefont{S.~A.} \bibnamefont{Langer}} \bibnamefont{and}

1187:   \bibinfo{author}{\bibfnamefont{A.~J.} \bibnamefont{Liu}},

1188:   \bibinfo{journal}{J. Phys. Chem. B} \textbf{\bibinfo{volume}{101}},

1189:   \bibinfo{pages}{8667} (\bibinfo{year}{1997}).

1190: 

1191: \bibitem[{\citenamefont{Reinelt and Kraynik}(2000)}]{RK00}

1192: \bibinfo{author}{\bibfnamefont{D.~A.} \bibnamefont{Reinelt}} \bibnamefont{and}

1193:   \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Kraynik}},

1194:   \bibinfo{journal}{J. Rheology} \textbf{\bibinfo{volume}{44}},

1195:   \bibinfo{pages}{453} (\bibinfo{year}{2000}).

1196: 

1197: \bibitem[{\citenamefont{Aubout et~al.}(2003)\citenamefont{Aubout, Jiang,

1198:   Glazier, and Graner}}]{AJGG03}

1199: \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Aubout}},

1200:   \bibinfo{author}{\bibfnamefont{Y.}~\bibnamefont{Jiang}},

1201:   \bibinfo{author}{\bibfnamefont{J.~A.} \bibnamefont{Glazier}},

1202:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{F.}~\bibnamefont{Graner}},

1203:   \bibinfo{journal}{Granular Matt.} \textbf{\bibinfo{volume}{5}},

1204:   \bibinfo{pages}{67} (\bibinfo{year}{2003}).

1205: 

1206: \bibitem[{\citenamefont{Janiaud and Graner}(2003)}]{JG03}

1207: \bibinfo{author}{\bibfnamefont{E.}~\bibnamefont{Janiaud}} \bibnamefont{and}

1208:   \bibinfo{author}{\bibfnamefont{F.}~\bibnamefont{Graner}},

1209:   \bibinfo{journal}{cond-mat/0306590}  (\bibinfo{year}{2003}).

1210: 

1211: \bibitem[{\citenamefont{Ball and Blumenfeld}(2002)}]{BB02b}

1212: \bibinfo{author}{\bibfnamefont{R.}~\bibnamefont{Ball}} \bibnamefont{and}

1213:   \bibinfo{author}{\bibfnamefont{R.}~\bibnamefont{Blumenfeld}},

1214:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{88}},

1215:   \bibinfo{pages}{115505} (\bibinfo{year}{2002}).

1216: 

1217: \end{thebibliography}

1218: 

1219: \clearpage

1220: 

1221: \begin{figure}

1222: \includegraphics[width=8cm]{figure1.eps}

1223: \caption{\label{defT1} (a) Schematic representation of the three

1224: main steps in a T1 event. The bubbles labelled 1 and 2 are

1225: initially neighbors. As the bubbles are sheared, all four bubbles

1226: meet at a vertex. After the event, the bubbles labelled 3 and 4

1227: are neighbors. (b) - (d) are three images illustrating an actual

1228: T1 event in the bubble raft. The location of one T1 event is

1229: highlighted by artificially coloring the bubbles involved white.

1230: The images are taken 3.2~s apart and the white scale bar in (b) is

1231: 2~mm long.}

1232: \end{figure}

1233: 

1234: \begin{figure}

1235: \includegraphics[width=8cm]{figure2.eps}

1236: \caption{\label{schematic} (a) A schematic drawing of the

1237: apparatus showing a side view. The main elements of the apparatus

1238: are the knife edge disk that is supported by a torsion wire and

1239: that serves as the inner cylinder for the bubbles. There is a

1240: separate fixed inner cylinder in the fluid (in gray). There is a

1241: segmented outer cylinder for generating flow, and there is a fixed

1242: dish that holds the fluid. The bubbles sit on top of the fluid, as

1243: indicated by the circles. (b) An image from the top of the bubbles

1244: in the apparatus that shows a portion of both the outer and inner

1245: cylinder. The black scale bar in the lower left corner is 5~mm.}

1246: \end{figure}

1247: 

1248: \begin{figure}

1249: \includegraphics[width=8cm]{figure3.eps}

1250: \caption{\label{initstress} Stress on the inner cylinder versus

1251: strain for the initial period of shear. The curve is divided into

1252: three regions labelled A, B, and C. Region A is the only period

1253: for which no T1 events are observed.}

1254: \end{figure}

1255: 

1256: \begin{figure}

1257: \includegraphics[width=8cm]{figure4.eps}

1258: \caption{\label{elflow} A plot of $v(r)/r$ versus radial position

1259: for the bubble motion during interval A in Fig.~\ref{initstress}.

1260: The solid squares are data averaged over all bubbles at a given

1261: radial position. The solid line is a fit to Eq.~\ref{velastic}.}

1262: \end{figure}

1263: 

1264: \begin{figure}

1265: \includegraphics[width=8cm]{figure5.eps}

1266: \caption{\label{t1elastic} Five images representing typical bubble

1267: deviations from elastic flow during the initial stress rise.

1268: Images (a) - (c) are taken from the corresponding three strain

1269: intervals in Fig.~\ref{initstress} and show typical strain

1270: intervals in which no T1 events occur. Image (d) shows a typical

1271: localized T1 event from region (B) in Fig.~\ref{initstress}. Image

1272: (e) shows an event composed of multiple T1 events in which two

1273: rows of bubbles slip pass each other. This event is taken from

1274: region (C) in Fig.~\ref{initstress}. The circles indicate the

1275: location of a subset of bubbles that have been tracked. The sizes

1276: of all the circles are the same, independent of actual bubble

1277: size, for clarity. They are color coded based on the deviation

1278: from elastic displacements, as defined in the text. White

1279: represents deviations less then $0.05\ {\rm cm}$. The color bar

1280: gives the scale for deviations greater than $0.05\ {\rm cm}$. The

1281: squares represent the location of T1 events, where the color

1282: equals the time relative to the start of the interval. The scale

1283: bar in image (a) is 0.5~mm.}

1284: \end{figure}

1285: 

1286: \begin{figure}

1287: \includegraphics[width=8cm]{figure6.eps}

1288: \caption{\label{GvT1}

1289: Scatter plot of the number of T1 events during a period of stress

1290: increase versus the effective elastic modulus $G$ for the same

1291: period of strain. The solid line is a guide to the eye.}

1292: \end{figure}

1293: 

1294: \begin{figure}

1295: \includegraphics[width=8cm]{figure7.eps}

1296: \caption{\label{t1pos} The individual points are the radial

1297: position of T1 events as a function of strain. The solid curve is

1298: the stress as a function of strain (measured using images of the

1299: inner cylinder).}

1300: \end{figure}

1301: 

1302: \begin{figure}

1303: \includegraphics[width=8cm]{figure8.eps}

1304: \caption{\label{t1rate} The solid line is the same stress versus

1305: strain curve as shown in Fig.~\ref{t1pos}. The bars summarize the

1306: data in Fig.~\ref{t1pos} by plotting only the number of T1

1307: events/bubble in a strain interval of $0.013$.}

1308: \end{figure}

1309: 

1310: \begin{figure}

1311: \includegraphics[width=8cm]{figure9.eps}

1312: \caption{\label{t1vel} The squares are the average velocity of the

1313: bubbles normalized by $\Omega r$ as a function of radial position.

1314: The bars give the total number of T1 events/bubble as a function

1315: of radial position. The solid lines are guides to the eye. The

1316: horizontal line is $v(r)/\Omega r = 1$, which corresponds to the

1317: motion of a rigid body. The vertical line is the location of the

1318: shear discontinuity, as reported in Ref.~\cite{LCD04}.}

1319: \end{figure}

1320: 

1321: \begin{figure}

1322: \includegraphics[width=8cm]{figure10.eps}

1323: \caption{\label{t1highlight} This is two strain intervals from

1324: Fig.~\ref{t1pos}, showing both the location of T1 events (solid

1325: squares) and the stress (solid line, as measured by the magnetic

1326: method) as a function of strain. Each interval is further divided

1327: into smaller strain intervals by the vertical lines. The labels in

1328: (b) correspond to the images in Fig.~\ref{t1flow}.}

1329: \end{figure}

1330: 

1331: \begin{figure}

1332: \includegraphics[width=8cm]{figure11.eps}

1333: \caption{\label{t1flow} Five images representing typical bubble

1334: deviations from elastic flow during the corresponding strain

1335: intervals as indicated in Fig.~\ref{t1highlight}(b). The circles

1336: indicate the location of a subset of bubbles that have been

1337: tracked. The size of the circle is the same for all bubbles for

1338: clarity. The circles are color coded based on the deviation from

1339: elastic displacements, as defined in the text. White represents

1340: deviations less then $0.05\ {\rm cm}$. The color bar gives the

1341: scale for deviations greater than $0.05\ {\rm cm}$. The squares

1342: represent the location of T1 events, where the color equals the

1343: time relative to the start of the interval. The scale bar in image

1344: (a) is 0.5~mm.}

1345: \end{figure}

1346: 

1347: \begin{figure}

1348: \includegraphics[width=8cm]{figure12.eps}

1349: \caption{\label{bubdispl} The average angular displacement of the

1350: bubbles normalized by $\Omega \Delta t$ as a function of radial

1351: position. The different symbols are for the different strain

1352: intervals in Fig.~\ref{t1highlight}b: (A) open circles; (B) solid

1353: squares; (C) solid circles; (D) open squares; and (E) solid

1354: triangles. Here $\Delta t$ is the time interval for each strain

1355: interval. The solid line is the fit to elastic behavior from

1356: Fig.~\ref{elflow}.}

1357: \end{figure}

1358: 

1359: \end{document}

1360: