1: \documentclass[prl,twocolumn,amsmath,amssymb]{revtex4}
2: \usepackage{graphicx}
3: \begin{document}
4: \title{Aging after shear rejuvenation in a soft glassy colloidal suspension:\\ evidence for two different regimes}
5: \author{F. Ianni$^{1,2}$}
6: \email{francesca.ianni@phys.uniroma1.it}
7: \author{R. Di Leonardo$^{2}$}
8: \author{S. Gentilini$^1$}
9: \author{G. Ruocco$^{1,2}$}
10: \affiliation{ $^{1}$ Dipartimento di Fisica, Universit\'a di Roma
11: ``La
12: Sapienza'', I-00185, Roma, Italy.\\
13: $^{2}$ SOFT-INFM-CNR c/o Universit\'a di Roma ``La Sapienza'',
14: I-00185, Roma, Italy. }
15: \date{\today}
16:
17: \pacs{83.80.Hj,83.85.Ei,42.25.Fx}
18: \begin{abstract}
19: The aging dynamics after shear rejuvenation in a glassy, charged
20: clay suspension have been investigated through dynamic light
21: scattering (DLS). Two different aging regimes are observed: one is
22: attained if the sample is rejuvenated before its gelation and one
23: after the rejuvenation of the gelled sample. In the first regime,
24: the application of shear fully rejuvenates the sample, as the
25: system dynamics soon after shear cessation follow the same aging
26: evolution characteristic of normal aging. In the second regime,
27: aging proceeds very fast after shear rejuvenation, and classical
28: DLS cannot be used. An original protocol to measure an ensemble
29: averaged intensity correlation function is proposed and its
30: consistency with classical DLS is verified. The fast aging
31: dynamics of rejuvenated gelled samples exhibit a power law
32: dependence of the slow relaxation time on the waiting time.
33: \end{abstract}
34: \maketitle
35: \section{Introduction}
36: Soft glassy colloidal suspensions are widespread in nature and are
37: of broad technological importance. Macroscopically, they are
38: characterized by a non-Newtonian rheology and, often, by a
39: viscosity that increases many orders of magnitude as time evolves
40: and the gelation process proceeds \cite{sollich}. These complex
41: fluids also exhibit a strong sensitivity to external forces and
42: have mechanical properties typical of soft solids, such as
43: solid-like behavior below a finite yield stress and thixotropic
44: response to applied deformation \cite{larson}. In particular, a
45: shear flow may induce a thinning effect and reduce the fluid
46: viscosity. At the microscopic level, the gelation process
47: corresponds to a slowing down of the structural relaxation with
48: the elapsed time: such a behavior is called aging. Moreover, the
49: system structural dynamics are accelerated by the shear flow. This
50: phenomenon is called shear rejuvenation. The competition between
51: the thickening, produced by the aging, and the thinning, induced
52: by shear flow, gives rise to
53: interesting phenomena.\\
54: On the theoretical and numerical side, evidences of the
55: competition between these two effects come from the extension of
56: slow dynamics theories in complex systems, such as mode coupling
57: theory (MCT) \cite{fuchs}, mean-field models \cite{bbk}, trap
58: models \cite{sollich} and molecular dynamics simulations
59: \cite{yamamoto, berthier, angelani}, to the presence of external
60: driving forces. These studies show that the structural dynamics
61: are very sensitive to even moderate shear rates, providing that
62: the characteristic timescale for structural rearrangements becomes
63: of the same order of the inverse shear rate. In particular, even
64: starting from non equilibrium, aging states, the presence of shear
65: ensures the existence of a stationary state whose correlation
66: function decays to zero on a
67: timescale given by $\dot{\gamma}^{-1}$.\\
68: On the experimental side, the investigation of the
69: shear-influenced slow dynamics at the microscopic level is still
70: relatively poor. Diffusing Wave Spectroscopy (DWS) \cite{bonnreju,
71: viasnoff, pascal} and Light Scattering Echo (LSE) experiments
72: \cite{petekidis, kaloun} gave indirect evidences for a shear
73: dependent structural relaxation time and for rejuvenation of aged
74: samples. Unfortunately, the statistical properties of multiple
75: scattered light (probed in DWS and LSE experiments) are not easily
76: represented in terms of the particles' correlation functions,
77: which, on the contrary, is the outcome of the dynamic light
78: scattering technique in the single scattering regime (DLS). The
79: DLS technique probes the intermediate scattering function of the
80: colloidal particles $F_q(t)$ \cite{berne}, which plays a central
81: role in both theoretical and numerical approaches. The
82: intermediate scattering function of a colloidal suspension
83: exhibits two decays: a fast relaxation, which accounts for single
84: particle diffusion, and a slow relaxation, whose characteristic
85: timescale $\tau_s$ grows many orders of magnitude with the waiting
86: time $t_w$ and accounts for the structural rearrangements of the
87: system. At long $t_w$ the system enters a glass or gel phase; only
88: the fast relaxation remains and is observed as a decay towards a
89: plateau, reflecting the
90: onset of structural arrest.\\
91: Unfortunately, also DLS suffers of limitations. Indeed, it is not
92: the proper technique to investigate the system dynamics when the
93: characteristic slow relaxation time $\tau_s$ becomes very large or
94: the aging dynamics proceed very fast. The correlation function is
95: usually measured as an average on the time origin and the
96: experimental acquisition time needed to get a good signal to noise
97: ratio must be kept longer than $\tau_s$ and shorter than the time
98: one should wait before changes in $\tau_s$ are significant. On the
99: contrary, Multi-Speckle Dynamic Light Scattering (MS-DLS) or
100: Multi-Speckle Diffusive Wave Spectroscopy (MS-DWS) perform an
101: ensemble average over the speckle pattern of the light scattered
102: intensity and acquisition time is strongly reduced. This enables
103: the investigation of the system dynamics for much longer waiting
104: times $t_w$ or for faster aging processes. However, in both
105: techniques, the time resolution is much smaller than in DLS, as it
106: is limited by the camera device used as detector. Overall, the
107: experimental studies of aging and rejuvenation under shear of
108: colloidal suspensions still suffer of technical limitations and
109: the "ideal" technique has not yet been introduced.\\
110: On the samples' side, among others, proper candidates for the
111: study of the aging and rejuvenation phenomena in soft glassy
112: colloids are suspensions of charged anisotropic colloidal
113: particles such as clay, which have been widely investigated both
114: for their important industrial application \cite{olphen} and as a
115: prototype of soft glassy materials \cite{kroon, bonnaging,
116: barbara0, barbara1}. Screened charged interactions between these
117: anisotropic particles induce the formation of disordered arrested
118: phase in the suspension even at very low volume fractions
119: \cite{barbara0}. In particular, for sufficiently concentrated
120: samples, a two stage aging process has been observed through DLS
121: experiments in a Laponite sample, a suspension of $25$ nm diameter
122: discoidal charged particles \cite{lequeux}: for small $t_w$, the
123: structural relaxation time $\tau$ increases exponentially, or more
124: than exponentially, with $t_w$, while for long $t_w$, when
125: $\tau>t_w$, a power law behavior shows up. The rejuvenation effect
126: of a shear flow on such systems has been investigated using DWS
127: \cite{bonnreju, pascal}, LSE \cite{kaloun}, and DLS \cite{noi}.
128: The particles' dimension, the sample's history and the kind of
129: shear applied vary among these works; consequently, the effects of
130: the rejuvenation on the aging dynamics changes and a comprehensive
131: picture of the phenomenology has not yet been reached. In Ref.
132: \cite{noi}, a Laponite suspension is used and the waiting time is
133: set to zero after the sample filtration. Normal aging in such a
134: sample is characterized by an intermediate scattering function
135: having a stretched exponential slow relaxation, whose
136: characteristic time $\tau_s$ grows exponentially with waiting time
137: $t_w$ while becoming strongly stretched. Moreover, an exponential
138: relaxation process is observed at short times and its
139: characteristic time $\tau_f$ slightly increases with $t_w$
140: \cite{Lille}. To investigate the evolution of the aging dynamics
141: during a steady shear, measurements are acquired while the shear
142: is stopped for a few minutes between different shearing periods.
143: The system dynamics soon after shear cessation are affected by the
144: shear flow just applied. In particular, they are characterized by
145: a strong reduction of aging as soon as $\tau_s$ enters the
146: timescale set by the inverse shear rate. In \cite{kaloun, pascal},
147: aged samples of Saponite particles, $125$ nm diameter charged
148: platelets, are rejuvenated by a strong oscillating shear, after
149: which the counting of $t_w$ starts. The sample is then submitted
150: to various periodic shear protocols and dynamics of tracer
151: particles are investigated during shear and after shear cessation.
152: Under these conditions, aging is much faster and the structural
153: relaxation time exhibits a power law dependence on $t_w$.\\
154: In this paper, we aim to elucidate the apparently contrasting
155: behavior observed in the two previously described systems
156: \cite{noi, Lille} and \cite{kaloun, pascal}. We investigate an
157: aqueous suspension of Laponite, a synthetic layered silicate
158: composed of monodisperse discoidal particles, which get charged in
159: water. Aging dynamics following the shear rejuvenation before the
160: sample gelation occurs are investigated through DLS technique. On
161: the contrary, aging after rejuvenation of gelled samples is very
162: fast and classical DLS could not be used, as the system wouldn't
163: be stationary during the acquisition time. Thus, a novel protocol
164: of DLS acquisition, consisting in an ensemble average over many
165: rejuvenating experiments, is proposed to measure the intensity
166: correlation function and its coherence with classical DLS is
167: checked. This new method has the double advantage of probing the
168: intermediate scattering function of the colloidal particles as in
169: classical DLS, with the "ensemble" average enabling the
170: investigation of a dynamics changing very rapidly in the system,
171: while a time resolution much higher than in MS-DLS technique is
172: achieved. As a result, we find the existence of two different
173: regimes of aging after shear rejuvenation as a common feature of
174: charged discoidal clay suspensions. If the shear is applied before
175: sample gelation is completed, the rejuvenation is followed by a
176: slow aging regime, behaving just like normal aging after sample
177: preparation. On the contrary, when a gelled sample is rejuvenated
178: by shear, we observe a fast aging regime after the shear
179: cessation, characterized by a power law dependence of the slow
180: relaxation time on the waiting time.
181: \section{Experiments and results}
182: The investigated system consists of an aqueous suspension of
183: Laponite RD, a synthetic layered silicate provided by Laporte Ltd.
184: Particles are disk shaped with a diameter of $25$ nm and $1$ nm
185: thickness. Laponite powder is dispersed in ultrapure water at
186: $3\%$ wt concentration and stirred for $\sim30$ min. The obtained
187: suspension, which is optically transparent and initially "liquid",
188: is loaded into a home made, cone and plate shear cell having a
189: flat optical window as the static plate. Cell loading (through a
190: $0.45$ $\mu$m filter) is taken as the origin of waiting times for
191: normal aging. Incident laser beam (diode pumped solid-state laser,
192: $\lambda=532$ nm, $P=150$ mW) and scattered light pass through the
193: same optical window. The scattered light is collected by a
194: mono-mode optical fiber and detected by a photomultiplier, after
195: being optionally mixed with a coherent local oscillator field.
196: Photocounts are acquired through a general purpose, counter/timer
197: PCI board (National Instruments PCI 6602). We developed a set of
198: software classes (implemented as extension modules of the object
199: oriented language Python) designed to perform basic tasks for the
200: statistical analysis of digital pulse trains. A typical
201: application is real time multi-tau photon correlation. However, a
202: software approach, having access to the full photocounts train,
203: allows to efficiently prototype different analysis protocols,
204: going far beyond the simple autocorrelation function
205: \cite{photonlab}. The scattering geometry is fixed (scattering
206: vector $q=22\;\mu\textrm{m}^{-1}$). The shift of the cell allows
207: us to select the position of the scattering volume in the cell gap
208: and the possibility of choosing an heterodyne correlation scheme
209: \cite{hetero} enables direct access to the detailed velocity
210: profile in the shear cell.
211: \subsection{Rejuvenation before gelation}
212: We first investigate the aging dynamics that follows a shear
213: rejuvenation induced before the full gelation of the sample takes
214: place. We let the system age for a time $t_w^0$ after cell
215: loading, then we apply a shear and finally we follow the dynamics
216: after shear cessation through DLS. We defining $t_w$ the time
217: elapsed since shear cessation. Aging of the system is monitored
218: through the normalized intensity autocorrelation function
219: $g^{(2)}(t_w, t)=\langle I(q,t_w)I(q,t_w+t) \rangle_T/\langle I(q,
220: t_w)\rangle_T^2$, where $\langle..\rangle_T$ indicates temporal
221: average over the acquisition time $T$. In the single scattering
222: regime and within the Gaussian approximation
223: $g^{(2)}(t_w,t)=1+|F_q(t_w,t)|^2$ \cite{berne}, where
224: $F_q(t_w,t)=\langle \rho_{-q}(t_w)\rho_q(t_w+t) \rangle/\langle
225: \rho_{-q}(t_w)\rho_q(t_w)\rangle$ is the intermediate scattering
226: function of the colloidal particles. An example of the results of
227: this procedure is reported in Fig. \ref{corrs}, left panel. In
228: this example, at $t_w^0=13.4\; h$ a shear rate of $100\;s^{-1}$ is
229: applied for two minutes; once the shear is stopped, aging is
230: followed through the intensity autocorrelation function
231: $g^{(2)}(t_w,t)$ for a set of $t_w$ between 0.2 and 3 hours.
232: \begin{figure}
233: \begin{center}
234: \includegraphics[height=6.5 cm]{1.eps}
235: \caption{Normalized intensity autocorrelation functions for two
236: different aging regimes after shear rejuvenation. \textit{Left}: a
237: Laponite sample rejuvenated before entering the arrested phase.
238: Correlation functions are obtained from an average on the time
239: origin for 16 equally spaced waiting times between 0.2 and 3 hours
240: (from left to right) after shear cessation. \textit{Right}: a
241: gelled Laponite sample for a set of waiting times between $0.3$ s
242: and $40$ s (from left to right) after shear cessation. Correlation
243: functions are obtained from an ensemble average over bunches of
244: counts, all acquired after the application of the same shear rate
245: for a duration of two minutes.}\label{corrs}
246: \end{center}
247: \end{figure}
248: \begin{figure}
249: \begin{center}
250: \includegraphics[height=6.5 cm]{2.eps}
251: \caption{Difference of the two aging regimes after shear
252: rejuvenation as evidenced from the evolution of the $\beta$
253: parameter, deduced from a stretched exponential fit
254: $exp[(-t/\tau_s)^{\beta}]$ of the intensity autocorrelation
255: function. \textit{Left}: aging of a sample rejuvenated before
256: entering the arrested phase. \textit{Right}: aging of a
257: rejuvenated gelled sample.}\label{twobeta}
258: \end{center}
259: \end{figure}
260: \begin{figure}
261: \begin{center}
262: \includegraphics[height=6.5 cm]{3.eps}
263: \caption{Complete rejuvenation of the Laponite sample in the slow
264: aging regime, \textit{i.e.} before gelation. Average slow
265: relaxation time $\langle \tau_s\rangle$ is plotted as a function
266: of waiting time $t_w$ for a sample aging soon after filtration
267: (empty circles) and for the same sample aging after the
268: application of a shear rate of $100\:\: s^{-1}$ for 2 minutes,
269: 13.4 hours after filtration (full circles). $t_w$ is shifted in
270: order to superimpose the first $\langle \tau_s\rangle$ measured
271: after the application of shear with the first aging curve (empty
272: circles). The black line is a guide for the eye. In the
273: \textit{inset} the fast relaxation time $\tau_f$ is plotted as a
274: function of $t_w$ for the two aging evolution with the same shift
275: in $t_w$. The slow aging regime is also compared to the fast aging
276: regime, which is observed after the shear rejuvenation of a gelled
277: sample: white diamonds represent the evolution of the slow
278: relaxation time after a shear rate $\dot{\gamma_1}=0.5\:\:
279: s^{-1}$, while grey diamonds are for $\dot{\gamma_1}=100\:\:
280: s^{-1}$. Both curves are shifted in $t_w$ through the procedure
281: previously described.}\label{reju1}
282: \end{center}
283: \end{figure}
284: As for normal aging, the two step decay for $F_q(t_w,t)$:
285: \begin{equation}
286: F_q(t_w,t)=f\exp\left[-(t/\tau_s)^\beta\right]+(1-f)\exp\left[-t/\tau_f\right]
287: \end{equation} where all parameters ($f$, $\tau_s$, $\beta$, $\tau_f$) depend
288: on $t_w$, provides a very good fit for all the correlations. As
289: the waiting time $t_w$ evolves, the slow relaxation decay becomes
290: more stretched: the stretching parameter $\beta$ is smaller than
291: one and decreases with $t_w$, as shown in the left panel of Fig.
292: \ref{twobeta}. In order to quantify the slow decay timescale, we
293: calculate the average slow relaxation time $\langle\tau_s\rangle$
294: ($\langle\tau_s\rangle=\int_0^\infty\exp[-(t/\tau_s)^\beta]\;dt=\tau_s/\beta\;\Gamma(1/\beta)$,
295: where $\Gamma$ is the Euler Gamma function) and we plot it as a
296: function of $t_w$ in Fig. \ref{reju1} (full circles). The
297: evolution of $\langle\tau_s\rangle$ as a function of $t_w^0$ for
298: the normal aging is also plotted (open circles). In order to show
299: that shear application completely rejuvenates the system, the
300: $t_w$ axis for aging after shear rejuvenation has been shifted by
301: superimposing the first point of the curve on the normal aging
302: curve. The same behavior can be observed in the evolution of the
303: fast relaxation timescale $\tau_f$, which is plotted in the inset
304: of Fig. \ref{reju1}. In fact, the dependence of the fast
305: relaxation time on the waiting time is in disagreement with
306: previous observations \cite{bonnaging, lequeux}, that assigned the
307: fast dynamics to the single particle diffusion.\\
308: Summarizing, once a shear is applied before gelation, a reduced
309: timescale for relaxation is observed soon after shear stops. This
310: starting value depends on the applied shear rate, as evidenced in
311: Ref. \cite{noi}. The following evolution of the relaxation times
312: (both fast and slow) traces the same aging curves (once the
313: waiting time axis is shifted) exhibited during normal aging,
314: showing that complete rejuvenation is achieved through shear flow.
315: \subsection{Rejuvenation after gelation}
316: We now turn to the investigation of the aging evolution after a
317: shear is applied to gelled samples. Before describing the
318: measurements in such a regime, we want to point out that when the
319: gelled Laponite suspension is put under shear, drastic wall slip
320: takes place and all the fluid rotates as a solid body leaving a
321: null shear in the core. Measurements of the velocity profile are
322: performed through the heterodyne dynamic light scattering setup
323: \cite{hetero}. In order to apply a controlled shear to rejuvenate
324: the gelled sample, the solid band is first broken through the
325: application of an high shear rate ($\dot{\gamma}_0>100\;s^{-1}$)
326: of the duration of two minutes. We checked that this strong shear
327: doesn't alter the nature of aging dynamics after subsequent shear
328: applications. In such a regime, the sample is found to quickly age
329: back to a gel state. By looking at the intensity scattered by the
330: sample after shear cessation, the signal fluctuations exhibit a
331: very rapid slowing down, as can be observed in Fig. \ref{counts}.
332: Now, being $I(t)$ not stationary for a time period long enough to
333: collect the data, the intensity correlation function cannot be
334: obtained by time averaging and an ensemble averaging over many
335: rejuvenating experiments is used: $g^{(2)}(t_w, t)=\langle
336: I(q,t_w)I(q,t_w+t) \rangle_e/\langle I(q, t_w)\rangle_e^2$, where
337: $\langle..\rangle_e$ indicates the ensemble average over several
338: intensity evolutions acquired after cessation of a repeated shear
339: application. In particular, after the system has been left aging
340: for 72 hours in order to reach gelation, a strong shear rate
341: $\dot{\gamma}_0$ is first applied in order to eliminate wall slip
342: in the following shear application, then the system is sheared at
343: a given $\dot{\gamma}_1<\dot{\gamma}_0$ for two minutes. After
344: shear cessation the intensity fluctuations are collected for a
345: time interval of two minutes with a time resolution of $.01\:\:s$;
346: then a shear rate of the same value $\dot{\gamma}_1$ is applied
347: again for two minutes and the cycle starts again. The whole
348: measurement lasts several hours and we obtain an ensemble of
349: hundreds of acquisitions after the same initial conditions are
350: imposed. In any acquisition, the intensity fluctuations look
351: stationary in a log-linear plot of the intensity evolution, as
352: shown in the bottom panel of Fig. \ref{counts}. To speed up the
353: computation of the correlation function, the acquired counts are
354: thus logarithmically binned: a logarithmic binning of $t_w$ is
355: performed and the counts are averaged in each bin. The intensity
356: autocorrelation function is then calculated as an ensemble average
357: over all the bunches of counts, in the time window $10^{-2}-10^2$
358: for a set of waiting times between $1$ and $40$ s (Fig.
359: \ref{corrs}, right frame). The time interval $t_w<1\;s$ is not
360: considered, as the correlation functions may be influenced by
361: inertial effects due to flow stop. In order to show that an aging
362: of the sample under shear \cite{noi} is negligible during the
363: whole experiment, we also calculate the correlation functions by
364: taking only the first or last group of acquired counts and obtain
365: the same results. The ensemble averaged correlation functions are
366: first fitted by a stretched exponential decay, the fast component
367: of dynamics being below the present time window. Results of the
368: data analysis are checked to be invariant under different binning
369: and fitting procedures. Compared to normal aging, the new aging
370: regime entered by the rejuvenated gelled sample is characterized
371: by correlation functions of a different form and by a much more
372: rapid evolution of the slow relaxation time with $t_w$. In
373: particular, the stretching parameter of the correlation function
374: statistically fluctuates around one and doesn't seem to depend on
375: the waiting time (Fig. \ref{twobeta}, right panel). The
376: correlation function can thus be well fitted by a single
377: exponential decay. Besides, through the effect of the shear, the
378: structural relaxation time soon after shear cessation is reduced
379: by many orders of magnitude from the value characterizing the
380: arrested phase, which must be higher than $100\;s$ (Fig.
381: \ref{reju1}). Then, $\tau_s$ shows a power law dependence on
382: $t_w$: $\tau_s=A\;t_w^c$ (Fig. \ref{tauvstw}). To compare the fast
383: aging evolution in this regime to the slow aging evolution
384: characterizing the regime before gelation, we added in Fig.
385: \ref{reju1} the plot of the slow relaxation time evolution in this
386: second regime, for the applied stresses
387: $\dot{\gamma}_1=0.5,100\;s^{-1}$. A power law behavior of the
388: structural relaxation time versus $t_w$, with exponent close to
389: one, has already been observed after rejuvenation of aged samples
390: through MS-DWS experiments on clay suspensions \cite{pascal,
391: kaloun} and on other colloidal glasses \cite{viasnoff}. Also in
392: normal aging, clay suspensions have shown such a power law
393: behavior, but only for very long waiting times
394: \cite{lequeux}.\\
395: \begin{figure}
396: \begin{center}
397: \includegraphics[height=6.5 cm]{4.eps}
398: \caption{Fast slowing down of the scattered intensity fluctuations
399: with the waiting time for a rejuvenated gelled Laponite sample.
400: \textit{Top}: the evolution of the counts revealed by the
401: photomultiplier in $10^{-2}\:\:s$ soon after shear cessation is
402: plotted, $t_w=0$ corresponds to the shear stop. \textit{Bottom}:
403: log-linear plot of the counts evolution; a logarithmic binning of
404: $t_w$ has been performed and the counts have been averaged in each
405: bin.}\label{counts}
406: \end{center}
407: \end{figure}
408: \begin{figure}
409: \begin{center}
410: \includegraphics[height=6.5 cm]{5.eps}
411: \caption{Aging after rejuvenation of a gelled Laponite sample. The
412: evolution with $t_w$ of the slow relaxation time $\tau_s$,
413: obtained from a single exponential fit of the intensity
414: auto-correlation function, is plotted for various applied shear
415: rates $\dot{\gamma}_1$. Circles: $\dot{\gamma}_1=0.5\;s^{-1}$,
416: squares: $\dot{\gamma}_1=3\;s^{-1}$, diamonds:
417: $\dot{\gamma}_1=16\;s^{-1}$, down-triangles:
418: $\dot{\gamma}_1=70\;s^{-1}$ and up-triangles:
419: $\dot{\gamma}_1=100\;s^{-1}$. All the curves are well fitted by
420: the power law $\tau_s=A\;t_w^c$. When small enough shear rates are
421: applied ($\dot{\gamma}_1<<100\;s^{-1}$) the $c$ exponent lies in
422: the interval $1.23\pm 0.08$. For higher shear rates
423: ($\dot{\gamma}_1=70,100\;s^{-1}$), the $c$ exponent drops to a
424: value smaller than one, lying in the interval $0.56\pm 0.08$. On
425: the contrary, the $A$ parameter increases with the shear rate for
426: all the curves. In the inset, $\tau_s/[\tau_s(t_w=1\;s)]$ is
427: plotted as a function of $t_w$ in order to better distinguish
428: among the two groups of curves.}\label{tauvstw}
429: \end{center}
430: \end{figure}
431: In order to investigate the power law behavior of the aging
432: dynamics as a function of the shear rate value applied for
433: rejuvenation, we repeated the same experiment by varying
434: $\dot{\gamma}_1$. For small applied shear rates, Laponite aged
435: samples exhibit a nonlinear velocity profile \cite{bonnrheo}.
436: Heterodyne photocorrelation is thus used to measure the detailed
437: velocity profile, in order to calculated $\dot{\gamma}_1$ as the
438: local shear rate in the scattering volume \cite{berne}. On the
439: contrary, for higher applied shear, the shear rate is constant
440: along the gap and $\dot{\gamma}_1$ can simply be calculated as the
441: cone velocity over the gap width. For
442: $\dot{\gamma}_1<<\dot{\gamma}_0$, no significant changes are found
443: in the $c$ exponent of the power law. Indeed, for different values
444: of $\dot{\gamma}_1$, spanning more than one decade:
445: $\dot{\gamma}_1=0.5,3,16\;s^{-1}$, all the aging curves are well
446: fitted by the power law with $c$ lying in the interval $1.23\pm
447: 0.08$ (Fig. \ref{tauvstw}). When $\dot{\gamma}_1=70,100\;s^{-1}$
448: instead, the $c$ parameter drops to a value smaller than one and
449: lies in the interval $0.56\pm 0.08$. On the contrary, for all the
450: shear rates here investigated, the $A$ parameter increases with
451: $\dot{\gamma}_1$, but it is not clear why the relaxation time soon
452: after shear cessation is longer for an higher applied shear rate.
453: In particular, one would expect that when the shear rejuvenation
454: effect dominates aging, the inverse shear rate sets the slow
455: relaxation timescale. However, in our data, we don't find this
456: scaling for the slow relaxation time soon after shear cessation.
457: For example, at small shear rates, the shear does accelerate the
458: system dynamics, but the slow relaxation time gets much smaller
459: than the inverse shear rate.\\
460: Finally, consistency with classical DLS has been checked for this
461: original method of DLS employed to measure the intensity
462: correlation function for a rapidly aging sample. In order to
463: compare the correlation functions measured through both methods,
464: we monitored the aging dynamics of a Laponite sample under shear
465: at small $t_w^0$. In this regime, the system dynamics after shear
466: cessation are stationary over the timescale of the tens of seconds
467: and can thus be investigated through classical DLS also. Soon
468: after cell loading, the sample is left aging under shear at
469: $\dot{\gamma}=100\;s^{-1}$ and shear is stopped each $60$ s, a
470: pause of $5$ s is taken in order to avoid inertial effects due to
471: flow stop and then the scattered intensity is acquired for $1$ s,
472: with a time resolution of $10^{-5}$ s. Each hour, the
473: time-averaged intensity correlation function is also measured
474: through classical DLS by stopping the shear for $80$ s. The
475: experiment lasts 24 h and the ensemble averaged intensity
476: correlation function is calculated from all the acquired bunches
477: of counts. Such a long acquisition time is necessary in order to
478: reach a good statistics in the correlation function. In comparison
479: to MS-DLS technique, where the intensity correlation function is
480: calculated as an ensemble average over the speckle pattern, the
481: acquisition time is extremely large, but the time resolution is
482: much higher and enables the investigation of faster dynamics. As
483: expected for this regime, the intensity correlation function is
484: independent on $t_w$, the time elapsed since shear stop. On the
485: contrary, as the system under shear keeps on aging with $t_w^0$
486: \cite{noi}, the dynamics change during the experiment and the
487: ensemble averaged correlation function will provide an average
488: value of the slow relaxation timescale. The ensemble averaged
489: intensity correlation function is thus compared to the average of
490: all the intensity correlation functions measured during the
491: experiment through classical DLS (Fig. \ref{confronto}). In
492: conclusion, intensity correlation function measured through an
493: ensemble average over many acquisitions of the intensity evolution
494: after shear cessation shows a good agreement with the one measured
495: through classical DLS, where an average over the time origin is
496: performed.
497: \begin{figure}
498: \begin{center}
499: \includegraphics[height=6.5 cm]{6.eps}
500: \caption{Intensity correlation function measured as an ensemble
501: average over many intensity acquisitions after shear cessation
502: (thin line) compared to the time averaged correlation function
503: measured through classical DLS (bold line). Measurements are
504: performed on a Laponite sample aging under shear at
505: $\dot{\gamma}=100\;s^{-1}$ for 24 h. As the sample ages during
506: this long acquisition time, the correlation function plotted in
507: bold line is an average of all the correlation functions acquired
508: during the experiment through classical DLS.}\label{confronto}
509: \end{center}
510: \end{figure}
511: \section{Conclusions}
512: We show the existence of two different regimes of aging for a
513: suspension of charged discoidal clay particles after shear
514: rejuvenation. Before gelation, classical DLS measurements have
515: shown that a shear flow can completely rejuvenate the sample. The
516: system dynamics after shear cessation depend on the applied shear
517: rate and the following aging evolution is identical to the one
518: after sample preparation: the fast relaxation time increases
519: slightly with the waiting time $t_w$, while the slow relaxation
520: time grows exponentially with $t_w$ and becomes strongly
521: stretched. On the contrary, in a gelled sample, aging dynamics
522: after shear rejuvenation proceed very rapidly, so classical DLS
523: measurements cannot be used and a novel dynamic light scattering
524: method to measure the intensity correlation function has been
525: proposed. An ensemble average over many rejuvenating experiments
526: is used and the validity of this method is checked through
527: comparison with classical DLS measurements. With this method, we
528: investigated the aging dynamics of a gelled sample after shear
529: rejuvenation by measuring the ensemble averaged intensity
530: correlation function. Its time evolution is different from the one
531: typical of the other regime, as its decay is characterized by a
532: single exponential. Moreover, the slow relaxation time soon after
533: shear stop is reduced by many orders of magnitude with respect to
534: the large relaxation time typical of the arrested phase, while a
535: very rapid evolution of the aging follows. In particular, we
536: observed a power law behavior of the slow relaxation time with
537: $t_w$, whose
538: exponent shows two different values for small or high applied shear rates.\\
539: Further studies are needed to clarify the nature of these
540: different aging evolutions after shear rejuvenation and how they
541: are linked to the gelation process.\\
542: The authors wish to thank Gianni Bolle and Md Islam Deen for
543: technical assistance.
544: \begin{references}
545: \bibitem{sollich} P.~Sollich, F.~Lequeux, P.~H{\'e}braud, and M. E.~Cates, Phys. Rev. Lett., {\bf 78}, 2020, (1997).
546: \bibitem{larson} R. G.~Larson, \textit{The Structure and Rheology of Complex Fluids}, Oxford University Press (1999).
547: \bibitem{fuchs} M.~Fuchs and M. E.~Cates, Phys. Rev. Lett. {\bf 89}, 248304
548: (2002).
549: \bibitem{bbk} L.~Berthier, J.-L.~Barrat, and J.~Kurchan, Phys. Rev. E {\bf 61}, 5464-5472
550: (2000).
551: \bibitem{yamamoto} R. Yamamoto and A. Onuki J. Chem. Phys. {\bf 117}, 2359,
552: (2002).
553: \bibitem{berthier} L. Berthier and J.-L. Barrat J. Chem. Phys. {\bf 116}, 6228,
554: (2002).
555: \bibitem{angelani} L. Angelani, G. Ruocco, F. Sciortino, P. Tartaglia, and F. Zamponi, Phys. Rev. E {\bf 66}, 061505
556: (2002).
557: \bibitem{bonnreju} D. Bonn, S. Tanase, B. Abou, H. Tanaka, and J. Meunier, Phys. Rev. Lett., {\bf 89}, 015701,
558: (2002).
559: \bibitem{viasnoff} V.~Viasnoff and F.~Lequeux, Phys. Rev. Lett. {\bf 89}, 065701
560: (2002).
561: \bibitem{pascal} F. Ozon, T. Narita, A. Knaebel, G. Debr{\'e}geas, P. H{\'e}braud, J.-P. Munch, Phys. Rev. E, {\bf 68}, 032401, (2003).
562: \bibitem{petekidis}G.~Petekidis, A.~Moussa$\ddot{\textrm{i}}$d, and P. N.~Pusey, Phys. Rev. E {\bf 66}, 051402
563: (2002).
564: \bibitem{kaloun}S.~Kaloun, M.~Skouri, A.~Knaebel, J.-P.~M$\ddot{\textrm{u}}$nch, P.~H{\'e}braud, Phys. Rev. E {\bf 72}, 011401
565: (2005).
566: \bibitem{berne} B.~Berne and R.~Pecora, Dynamic Light Scattering, Wiley, New York,
567: (1976).
568: \bibitem{Goetze}W.~G$\ddot{\textrm{o}}$tze and L.~Sj$\ddot{\textrm{o}}$gren, Phys. Rev. A {\bf 43}, 5442
569: (1991).
570: \bibitem{olphen} H.~Van Olphen, \textit{An Introduction to Clay Colloid Chemistry}, 2nd ed. (Wiley, New York, 1977).
571: \bibitem{kroon} M. Kroon, G. H. Wegdam, and R. Sprik, Phys. Rev. E, {\bf 54}, 6541, (1996).
572: \bibitem{barbara0} B.~Ruzicka, L.~Zulian, G.~Ruocco, Phys. Rev. Lett., {\bf 93}, 258301, (2004).
573: \bibitem{barbara1} B.~Ruzicka, L.~Zulian, G.~Ruocco, Langmuir, {\bf 22}, 1106, (2006).
574: \bibitem{bonnaging} B. Abou, D. Bonn, and J. Meunier, Phys. Rev. E, {\bf 64}, 021510, (2001).
575: \bibitem{lequeux} M. Bellour, A. Knaebel, J. L. Harden, F. Lequeux, J.-P. M$\ddot{\textrm{u}}$nch, Phys. Rev. E, {\bf 67}, 031405, (2003).
576: \bibitem{noi} R. Di Leonardo, F. Ianni, and G. Ruocco, Phys. Rev. E, {\bf 71}, 011505, (2005).
577: \bibitem{Lille} R. Di Leonardo, S. Gentilini, F. Ianni, G. Ruocco J. non-Crystalline Solids, \textit{to be published}.
578: \bibitem{photonlab} The PhotonLab package is available for download at: http://glass.phys.uniroma1.it/dileonardo/photonlab.html.
579: \bibitem{hetero} R.~Richert, J. Phys.: Condens. Matter {\bf 14}, R703 (2002).
580: \bibitem{bonnrheo} D. Bonn, P. Coussot, H. T. Huynh, F. Bertrand, and G. Debr{\'e}geas, Europhys. Lett., {\bf 59}, 786,
581: (2002).
582: \end{references}
583: \end{document}
584: