1: \documentclass[doublecol]{epl2}
2: \title{Stresses in smooth flows of dense granular media}
3: \shorttitle{Stresses in smooth granular flows }
4: \author{Martin Depken\inst{1} \and Jeremy B. Lechman\inst{2} \and Martin van Hecke\inst{3} \and Wim van Saarloos\inst{1} \and Gary S. Grest\inst{2} }
5: \shortauthor{Martin Depken \etal}
6: \institute{
7: \inst{1} Instituut--Lorentz, Universiteit Leiden, Postbus
8: 9506, 2300 RA Leiden, The Netherlands\\
9: \inst{2} Sandia National Laboratories, Albuquerque, NM 87185,
10: USA\\
11: \inst{3} Kamerlingh Onnes Lab, Universiteit Leiden, Postbus 9504,
12: 2300 RA Leiden, The Netherlands
13: }
14:
15: \pacs{83.80.Fg}{Granular rheology}
16: \pacs{45.70.Mg}{Granular flow :classical mechanics of discrete systems}
17: \pacs{45.70.-n}{Granular systems}
18:
19:
20: \abstract{The form of the stress tensor is investigated in smooth,
21: dense granular flows which are generated in split-bottom shear
22: geometries. We find that, within a fluctuation fluidized spatial
23: region, the form of the stress tensor is directly dictated by the
24: flow field: The stress and strain-rate tensors are co-linear. The
25: effective friction, defined as the ratio between shear and normal
26: stresses acting on a shearing plane, is found not to be constant
27: but to vary throughout the flowing zone. This variation can not be
28: explained by inertial effects, but appears to be set by the local
29: geometry of the flow field. This is in agreement with a recent
30: prediction, but in contrast with most models for slow grain flows,
31: and points to there being a subtle mechanism that selects the flow
32: profiles.
33: }
34:
35:
36: \begin{document}
37:
38: \maketitle
39:
40:
41: \section{Introduction}
42: %MvH Rewrite
43: Granular media are amorphous and athermal materials which can jam
44: into stationary states, but which can also yield and flow
45: under sufficiently strong external forcing \cite{jamming,gm}.
46: Slowly flowing granulates, for which momentum transfer by enduring
47: contacts dominates over collisional transfer, are characterized by
48: a yielding criterion and rate independence. The former expresses
49: that granulates only start to flow when the applied shear stresses
50: exceed a critical yielding threshold~\cite{jamming,gm,nederman},
51: while the latter signifies that a change in the driving rate
52: leaves both the spatial structure of the flow and the stresses
53: essentially unaltered~\cite{bob,TCmueth,bonn,Fenistein,SFS}.
54:
55: Solid friction exhibits a similar combination of yielding and
56: rate-independence: According to the Coulomb friction law, a block
57: of material resting on an inclined plane starts to slide when its
58: ratio of shear to normal forces exceeds the static friction
59: coefficient. And, once the block slides, the same ratio is given
60: by a lower dynamical friction coefficient, which is essentially
61: rate independent.
62:
63: There is no unique manner in which these friction laws can be
64: translated into a continuum theory, and there exists a plethora of
65: approaches describing slow granular flows
66: \cite{nederman,SFS,tl,aranson,bazant,gdr,pouliquen_nature,Unger}.
67: To test these theories, one would like to determine the stresses
68: and strain rates within the material. However, experiments can not
69: easily access the flow in the bulk of the material, nor probe the
70: stress tensor in sufficient detail. In addition, slow grain flows
71: often exhibit sharp gradients, thus casting doubt on the validity
72: of continuum theories \cite{nederman,bob,TCmueth,bonn,tl}.
73: Finally, granular flows are notoriously sensitive to subtle
74: microscopic features \cite{TCmueth}, which often translates into a
75: substantial number of tunable parameters in the models
76: \cite{aranson}. As far as we are aware, no direct comparison
77: between the full stress and strain rate tensor has been undertaken
78: for slow granular flows.
79:
80: In this Letter, we numerically study grain flows in split-bottom
81: geometries as shown in fig.~\ref{fig:1}. Recently, these flows
82: were shown to exhibit robust and continuum-like flow profiles that
83: are numerically tractable and are governed by a number of
84: universal, i.e. grain-independent, scaling relations, making them
85: eminently suitable for our purpose. We relate the stress tensor to
86: the strain-rate tensor in these flows, thus providing a benchmark
87: for the testing and development of theoretical models for smooth
88: and dense grain flows. Experiments and numerics so far have
89: focussed on the flow in a cylindrical geometry
90: (fig.~\ref{fig:1}c), where a wide shear zone is generated by
91: rotating a centre bottom disc with respect to the cylindrical
92: container \cite{Fenistein,Unger,xiang}. We present some data for
93: this cylindrical case, but focus on the linear version of this
94: geometry (fig.~\ref{fig:1}a), where we find a wide shear zone to
95: emanate from the relative motion of two bottom plates along their
96: ``fault line''. In this system, the physics behind the stresses is
97: easier to disentangle because the stream lines are not curved.
98:
99: \begin{figure}
100: \includegraphics[width=\columnwidth]{fig1.png}
101: \vspace{0cm} \caption{\label{fig:1}(a) Linear shear geometry where
102: a split along the middle of the system generates a wide shear zone
103: in a layer of grains. The curves indicate sheets of constant
104: velocity. (b) Cuboid element of material showing the definition of
105: the angle $\theta$, the stresses in the SFS framework, and the
106: labelling of the axis --- the grey objects in a) and c) are
107: examples of such elements. (c) Cylindrical split-bottom geometry,
108: where the grains are driven by the rotation of a bottom disc. The
109: two surfaces of rotation indicate sheets of constant angular
110: velocity. Note that in the limit $R_s \!\to\! \infty$ one obtains
111: the linear geometry (a).}
112: \end{figure}
113:
114: %\section{Main findings}
115: Our main finding is that, throughout the flowing zone, the stress
116: and strain tensor are co-linear, meaning that their
117: eigen-directions, or equivalently, their principle directions,
118: coincide. Moreover, we find that the ratios of the non-zero stress
119: components, such as the effective friction coefficient, which is
120: the ratio between shear and normal stresses acting on a shearing
121: plane, are not constants but vary throughout the flowing zone.
122: This variation is crucial to understand the finite width of the
123: shear zones, and is not due to the variation in the magnitude of
124: the local strain rate. Both of these findings are in accord with
125: the main features of theory developed in \cite{SFS}, and
126: constitute an important step forward in establishing a general
127: framework for the modelling of grains flows.
128:
129: \section{SFS framework}
130: We formulate our results in the context of the theoretical
131: framework recently developed by Depken {\em et al.}~\cite{SFS}.
132: The central assumption of this so-called SFS theory is that, once
133: the material is flowing, strong fluctuations in the contact forces
134: enable otherwise jammed states to relax within a spatial region
135: which we refer to as the fluctuation fluidized region. In this
136: region there can not be a shear stress without a corresponding
137: shear flow. This assumption can be interpreted as stating that the
138: yielding threshold, which determines the onset of flow, is no
139: longer relevant once part of the material flows, since this
140: induces strong non-local fluctuations in the contact forces.
141: Further one observes that the flows can be locally (and in the
142: present cases also globally) seen as comprised of material sheets,
143: with no internal average strain rate, sliding past each other (see
144: fig.~\ref{fig:1}).
145:
146: Combining these two ingredients, it follows that both the shear
147: strains and shear stresses in these material sheets are zero, and
148: we refer to them as a Shear Free Sheet (SFS). It also follows that
149: the stress and strain-rate tensors are co-linear. The major and
150: minor principle directions of the strain-rate tensor are at an
151: angle of 45$^\circ$ with respect to the SFSs, and in the more
152: intuitive basis specified by these sheets (see
153: fig.~\ref{fig:1}b) the stress tensor takes the form:
154: \begin{equation}
155: \sigma_{{\rm SFS}}=\left(\begin{array}{ccc}P'
156: &0&0\\0&P&\tau\\0&\tau&P\end{array}\right)~. \label{eq:stressr}
157: \end{equation}
158:
159: To test this prediction, we check whether the numerically
160: obtained stresses are co-linear with the strain rate tensor and
161: thus are of the form (\ref{eq:stressr}). Moreover, when no further
162: assumptions are made, the three components $P$, $P'$, and $\tau$
163: will be different, and in general vary throughout the sample. In
164: fact, if the stress is of this form, a simple stress balance
165: argument shows that $\mu_{\rm eff}:=\tau/P$ has to vary throughout
166: the shear zones \cite{SFS}: A constant $\mu_{\rm eff}$ would
167: correspond to a shear zone of zero width, clearly inconsistent
168: with the available data \cite{Fenistein,xiang}.
169:
170: To put these predictions in perspective, let us briefly consider
171: the case of faster flows, where collisions play a role. The
172: arguments for the form of the stress tensor can be extended to
173: apply also for such systems, and Pouliquen and co-workers
174: \cite{pouliquen_nature} have suggested that the stress is of the
175: form eq.~(\ref{eq:stressr}). However, they introduce the following
176: restriction: $P'\!=\!P$ and $\tau=\mu_{\rm eff}(I)P$, where the
177: effective friction is a material dependent function of the
178: so-called inertial number $I=\dot{\gamma} d /\sqrt{P/\rho}$
179: \cite{Inote}, and $d$ and $\rho$ are the particles diameter and
180: density, respectively~\cite{gdr,pouliquen_nature}. For the slow
181: flows under consideration here, we should consider the limit
182: $I\to0$. If we only consider $\mu_{\rm eff}$ to depend on $I$,
183: $\mu_{\rm eff}$ becomes a material constant, which is, as we
184: explained above, incompatible with the finite width of the shear
185: zones \cite{SFS,Unger,jop_private}. Our study will thus illuminate
186: how subtle details of the form of the stress tensor have
187: significant consequences for the grain flow.
188:
189: \section{Method} The simulations are carried out with a discrete
190: element method (DEM) for $80\!-\!100$k mono-disperse Hertzian
191: spheres satisfying the Coulomb friction laws. The relevant
192: parameters describing the material properties of the spheres are
193: the normal stiffness $k_n=2 \times 10^5mg/d$, the tangential
194: stiffness $k_t=2/7k_n$, the normal and the tangential viscous
195: damping coefficients $\gamma _n=50 \sqrt{g/d}$ , $\gamma_t=0$, and
196: the microscopic coefficient of friction $\mu_m=0.5$. Here $d$ and
197: $m$ are the diameter and the mass of spheres, and $g$ is the
198: gravitational acceleration. The characteristic timescale $t_0$ is
199: given by $\sqrt{d/g}$ (e.g., $t_0 = 0.0101 sec$ if $d = 1 mm$). We
200: have studied a range of driving rates varying from from $\pm0.05$
201: to $\pm0.005~d/t_0$ and $0.015$ to $0.005~rad/t_0$ for the linear
202: and circular geometries, respectively. Stresses and velocities are
203: averaged over the symmetry direction (along split) and are
204: resolved with a resolution of $0.9d$ in the cross section. The
205: stress tensor within this volume is the sum of contact and
206: collisional stresses \cite{silbert_chute}, where the latter is
207: three orders of magnitude smaller than the former. The linear
208: setup has dimensions $20d$ in the shearing direction (periodic
209: boundary conditions), a width of $80d$, and a height of $50d$. The
210: details of the specific implementation can be found elsewhere
211: \cite{silbert_chute}.
212:
213:
214: \begin{figure}
215: \includegraphics[width=\columnwidth]{fig2.png}
216: \caption{\label{fig:2} Stress ratios $\sigma_{ij}/\sigma_{33}$ in
217: the linear geometry. The data was taken for a run where the
218: velocity difference across the sliding bottom plates was 0.05
219: $d/t_0$, and was averaged over the time interval ranging from 7150
220: $t_0$ to 9150 $t_0$ --- similar results where obtained for a
221: velocity difference that was ten times smaller. We plot the data
222: as a function of $\theta$, the angle between the ''1'' direction
223: of the SFS basis (see fig.~\ref{fig:1}) and the vertical. Black (red) dots
224: refer to points inside (outside) the fluctuation fluidized zone
225: (see text), and the curves are quadratic fits to the data in this
226: zone. The stress tensor follows eq.~(\ref{eq:stressr}): The ratios
227: $\sigma_{12}/\sigma_{33}$, $\sigma_{13}/\sigma_{33}$ and
228: $\sigma_{22}/\sigma_{33}$ are very close to zero, zero and one,
229: while the ratios $\sigma_{11}/\sigma_{33}=P'/P$ and
230: $\sigma_{23}/\sigma_{33}=\mu_{\rm eff}$ vary with $\theta$ and do
231: not attain any special value. }
232: \end{figure}
233:
234: \section{The form of the stress tensor} We first study the
235: relation between stresses and strain rates in the linear geometry.
236: Through a cross section of the flow we record the time-averaged
237: stress and velocity fields, and from the latter we extract the
238: orientation of the SFS basis. In the region far away from the
239: shear zone, these fields fluctuate strongly, and we limit the
240: analysis to a ``fluctuation fluidized zone''. For this particular
241: data set we take the boundary of the fluctuation fluidized zone to
242: be defined by where the inertial number $I$ attains the value
243: $I_{\rm cut}=4\times10^{-5}$
244: --- why this is reasonable is detailed below.
245:
246: Within this zone, we express the stresses in the SFS basis, and
247: compare our numerically obtained stresses to the SFS form
248: (\ref{eq:stressr}). We find that, due to gravity, all stress
249: components grow roughly proportional with depth. Since the SFS
250: theory makes no prediction on the absolute values of the various
251: stress components, we focus on the stress ratios
252: $\sigma_{ij}/\sigma_{33}$ (note that $\sigma_{23}/\sigma_{33}$
253: directly yields the effective friction coefficient, $\mu_{\rm
254: eff}$).
255:
256: In fig.~\ref{fig:2} we plot the stress ratios as a function of the
257: angle $\theta$, which parameterizes the orientation of the SFS
258: basis with respect to gravity (see fig.~\ref{fig:1}b). Even though the
259: stress ratios could vary arbitrarily with position throughout the
260: cross section, we find that their main variation is with $\theta$
261: --- the relevance of this angle
262: will be discussed below. Figure~\ref{fig:2} illustrates that in
263: the fluctuation fluidized zone the stress tensor takes the SFS
264: form (\ref{eq:stressr}). First, all stress ratios within this
265: region appear to collapse on single curves when plotted as
266: function of $\theta$, while data outside the region is scattered
267: more strongly. Second, the values for the ratio's
268: $\sigma_{12}/\sigma_{33}$ and $\sigma_{13}/\sigma_{33}$ scatter
269: around zero. Third, the ratio $\sigma_{22}/\sigma_{33}$ is close
270: to one and does not vary with $\theta$. Together these points show
271: the validity of the SFS picture within the fluctuation fluidized
272: region (see below for a more precise definition). Finally, the
273: stress ratios $\sigma_{11}/\sigma_{33} = P'/P$ and
274: $\sigma_{23}/\sigma_{33} = \tau/P$ are not constant and do not
275: attain any special values. The data does not suggest that it is
276: possible to simplify the form of the stress tensor
277: (\ref{eq:stressr}) any further.
278:
279: \section{Angle dependence of stress} The
280: variation of the effective friction $\mu_{\rm eff}$ with angle
281: $\theta$ takes on a special significance in the linear geometry.
282: In \cite{SFS} it was shown that, given a stress tensor of the SFS
283: form, force balance dictates that $\mu_{\rm eff}$ attains its
284: maximum in the middle of the shear zone, where $\theta\!=\!0$. It
285: was further shown that the curvature of $\mu_{\rm eff}(\theta)$
286: could be directly related to the scaling of the width of the shear
287: zone with vertical position in the sample; $W\sim z^\alpha$,
288: $\alpha=1/(1+\partial_{\theta\theta}\mu_{\rm eff}|_{\theta=0})$.
289: For constant $\mu_{\rm eff}$, $\alpha\!=\!1$ and the shear zones
290: become of zero width \cite{SFS,Unger,jop_private}.
291:
292: As fig.~\ref{fig:2}a shows, $\mu_{\rm eff}$ varies by roughly 10\%
293: throughout the fluctuation fluidized region and indeed attains a
294: maximum in the middle. A quadratic fit to $\mu_{\rm eff}$ yields
295: that $\partial_{\theta\theta}\mu_{\rm eff}|_{\theta=0} \approx
296: 2.5[5]$, which suggests the scaling exponent $\alpha=0.35[5]$.
297: From the numerical data presented here, and from the data in
298: \cite{Fenistein} and \cite{xiang}, the value of this width
299: exponent can be estimated to be somewhere in the range $0.25-0.4$,
300: consistent with our estimate \cite{footnote_errorbar}. We
301: interpret this coincidence as a strong check on the validity of
302: the SFS form
303: --- the variation of $\mu_{\rm eff}$ is clearly a subtle effect, and one
304: could imagine that small and systematic deviations of the stresses
305: from the SFS form could destroy the relation between $\alpha$ and
306: $\partial_{\theta\theta}\mu_{\rm eff}|_{\theta=0}$.
307:
308:
309: \section{Spatial variation of the stress} In fig.~\ref{fig:3}
310: we plot the variations of the stress ratios $\sigma_{ij}/
311: \sigma_{33}$ throughout a cross section of the linear cell,
312: including data from outside the fluctuation fluidized zone. We
313: will now provide support for our assertion that the dominant
314: variations of the stress ratios are with $\theta$. We first
315: checked that the correlation between $\mu_{\rm eff}$ and
316: dimensionless quantities, such as the density and the curvatures
317: of the SFS basis, are unconvincing. Other potential candidates are
318: $\theta$, $\dot{\gamma}$, and $I$, and these are also shown
319: fig.~\ref{fig:3}. Figure~\ref{fig:3} shows that the spatial
320: variation of $\mu_{\rm eff}$ is closer to $\theta$ than it is to
321: $\dot{\gamma}$ or $I$.
322:
323: Moreover, if the variation of $\mu_{\rm eff}$ was dominated by the
324: variation of $I$ or $\dot{\gamma}$, one would expect the width of
325: the shear zones to strongly depend on the shear rate
326: --- in stark contrast to both experimental \cite{Fenistein,xiang}
327: and numerical data \cite{xiang}. In fact, in runs done for a
328: driving rate which is a factor 10 smaller than shown here, the
329: stresses, flow profiles and $\mu_{\rm eff}(\theta)$ are
330: indistinguishable from those reported here --- the system is truly
331: rate independent. Finally, for the small inertial numbers here, $d
332: \mu_{\rm eff}/dI \sim {\cal O}(1)$ (based on the data presented
333: in~\cite{gdr}), while variations in $I$ over the shear zone are
334: ${\cal O}(10^{-3})$ --- far too small to explain the 10\%
335: variation in $\mu_{\rm eff}$.
336:
337: \begin{figure*}[bht]
338: \includegraphics[width=\textwidth]{fig3_alt2.png}
339: \caption{Components of the stress tensor in the SFS basis for the
340: linear shearing geometry, compared to the angle $\theta$, local
341: strain rate $\dot{\gamma}$ and inertial number $I$. This data is
342: based on the same run as shown in fig.~\ref{fig:2}. A qualitative change of
343: behavior of the stress fields is clearly visible around the dashed
344: lines, which indicates the boundary of what is taken as the
345: fluctuation fluidized zone, $I=I_{\rm cut} = 4\times10^{-5}$ (see
346: text). \label{fig:3}}
347: \end{figure*}
348:
349:
350: \section{The fluctuation fluidized region} Central to the SFS
351: picture is that shear flows induce force fluctuations that spread
352: through and fluidize the material, thus eliminating the yield
353: stress. How far do these fluidized zones spread? As can be seen in
354: Figs.~\ref{fig:2} and~\ref{fig:3}, there is a clear region within
355: which the stresses satisfy the SFS form of
356: eq.~(\ref{eq:stressr}), while outside this region the
357: stresses become noisy. We initially expected the total local
358: strain experienced since the start of the numerical experiment,
359: $\dot{\gamma} t$, to distinguish regions where fluctuations have
360: allowed the stresses to relax to the SFS form. But, when
361: attempting to maximize the spatial region of co-linearity, we
362: found that the inertial number, $I$, leads to a better estimate of
363: the fluctuation fluidized region: For the same required accuracy
364: in co-linearity, a larger region is selected~\cite{Inote}. In
365: Figs.~\ref{fig:2} and ~\ref{fig:3}, the region is defined as $I >
366: I_{\rm cut}= 4\times10^{-5}$. It should be noted that this cut-off
367: does not define a region of visible shear (as seen from the
368: $\dot{\gamma}$ plot in fig.~\ref{fig:3}), but rather a region
369: within which the microscopic fluctuations, created mainly in the
370: region of relatively large strain rates, remove any static shear
371: resistance.
372:
373: Can we understand the numerical value of $I_{\rm cut}$? The total
374: strain experienced after $t=8000~ t_0$ (taken in the middle of the
375: total time-interval over which the stresses are averaged) equals
376: $8000 ~t_0~ \dot{\gamma} = 8000 ~\sqrt{d/g}~ \dot{\gamma}$. At
377: the edge of the fluctuation fluidized zone at a certain height
378: $z$, the local strain rate equals $I_{\rm cut}/d~
379: \sqrt{P/\rho}=I_{\rm cut}\sqrt{g(h-z)}/d$; hence the total strain
380: experienced at this edge equals $8000 ~I_{\rm
381: cut}~\sqrt{(h-z)/d}=0.32~ \sqrt{(h-z)/d}$. Near the bottom the
382: total strain thus approximates five, while near the surface it
383: becomes of order 0.3. Even though the fluctuation fluidized region
384: is not directly given by $\gamma$, these numbers nevertheless set
385: a reasonable scale for the amount of strain the material needs to
386: experience before it is fluidized, in particular if one realizes
387: that due to the pressure gradient, the strain near the bottom
388: couples more strongly to the fluctuations of the forces near the
389: surface than vice versa. It should also be noted that we do not
390: expect the numerical value of $I_{\rm cut}$ to be universal --- in
391: particular, for longer runs we expect the fluctuation fluidized
392: region to spread slowly, with $I_{\rm cut} \sim 1/t $.
393:
394: \section{Results in cylindrical geometry} In fig.~\ref{fig:4} we show simulation
395: results for a cylindrical geometry with $H/R_s = 0.675$ ---
396: similar results are reached for a number of other filling heights
397: not shown here \cite{the_future_is_bright}. Figure~4 shows that also
398: for the curved geometry, the stresses are in the SFS form: The
399: values for the ratio's $\sigma_{12}/\sigma_{33}$,
400: $\sigma_{13}/\sigma_{33}$ and $\sigma_{22}/\sigma_{33}$ scatter
401: around zero, zero, and one respectively, with the ratio's
402: $\sigma_{11}/\sigma_{33} = P'/P$ and $\sigma_{23}/\sigma_{33} =
403: \tau/P$ varying throughout the fluctuation fluidized zone.
404:
405: Note that due to the more complex curved geometry, we have no
406: a-priori theoretical reason for expecting the stress ratios to
407: vary with $\theta$ alone. Moreover, there is no reason that
408: $\mu_{\rm eff}$ should be maximal in the middle, nor is it known
409: how $\mu_{\rm eff}(\theta)$ would be related to the width of the
410: shear zones --- if at all.
411:
412: \begin{figure}[tbh]
413: \begin{center}
414: \includegraphics[width=\columnwidth]{fig4.png}
415: \caption{ \label{fig:4} Stress ratios $\sigma_{ij}/\sigma_{33}$ as
416: a function of $\theta$, for a circular geometry. The driving rate
417: $\Omega$ was equal to 0.015 rad/$t_0$, and averages where taken
418: for time ranging from 8000 $t_0$ to $10^4$ $t_0$, corresponding to
419: the interval from approximately 19 to 24 turns. We have checked
420: that similar results where obtained for a run with $\Omega$ equal
421: to 0.005 rad/$t_0$. As before, black (red) dots referring to
422: points inside (outside) the fluctuation fluidized zone. For
423: details, see text.}
424: \end{center}
425: \end{figure}
426:
427: \section{Conclusion and Outlook} Based on the single,
428: straightforward and minimal assumption that fluctuations on the
429: grain scale forbid the occurrence of shear stresses without an
430: associated shear flow, it was in \cite{SFS} predicted that the
431: stress tensor in slow grain flows should take the form
432: (\ref{eq:stressr}), with the stress ratios varying throughout the
433: sample. The data presented here fully confirms this prediction:
434: {\em{(i)}} In the flowing zones, the stresses indeed take the form
435: (\ref{eq:stressr}). The three different components $P'$, $P$ and
436: $\tau$ are sufficiently different that no further simplifications
437: are consistent with the data. {\em{(ii)}} The ratio $P'/P$ and the
438: effective friction $\mu_{\rm eff}=\tau/P$ are not constant.
439: {\em{(iii)}} The variation of $\mu_{\rm eff}$ can be directly
440: related to the width of the shear zones. {\em{(iv)}} For the
441: cylindrical geometry, the stress tensor also satisfies the SFS
442: criteria, with $P'/P$ and $\mu_{\rm eff}=\tau/P$ varying over the
443: shear zone, but due to the more complex geometry we can not relate
444: this variation directly to the width of the shear zones.
445:
446: The SFS approach thus provides a powerful framework for unraveling
447: the relations between flow and stresses in granular media in
448: general, and the crucial but subtle spatial variation of the
449: effective friction $\mu_{\rm eff}$ and the unexpected variation of
450: $P'/P$ in particular.
451:
452: The range of validity of the SFS approach is not yet clearly
453: mapped out, and additional studies to answer the following key
454: questions are called for. {\em{(i)}} How does the stress tensor
455: evolve when the flow rate is increased? The stress tensor in the
456: Pouliquen approach for fast flows is similar to ours, but with the
457: restrictions that $P'\!=\!P$ and that $\mu_{\rm eff}(I)$ is a
458: function of the local strain rate only \cite{pouliquen_nature}.
459: Here $\mu_{\rm eff}$ apparently depends on geometry, and the
460: crossover from geometry ($\theta$) to inertial number ($I$)
461: dependence needs to be explored. {\em{(ii)}} We have seen here
462: that $P$ and $P'$ are systematically different, as was also seen
463: in simulations of chute flows \cite{silbert_chute}, and moreover,
464: that $P'/P$ is not a constant. Though we do not understand the
465: cause, nor the precise relevance, of this, it can not be a priori
466: ignored given the crucial role played by such variation of
467: $\mu_{\rm eff}$ in the formation of the wide shear zones in the
468: linear geometry. {\em{(iii)}} What distinguishes the zone where
469: the stresses are in the SFS form from the region where they are
470: not? Underlying the SFS picture is the assumption that the
471: fluctuations are sufficiently strong and fast, and one imagines
472: that far away from the shear zones this no longer holds true, thus
473: leading to a breakdown of co-linearity. Preliminary data suggest,
474: however, that the fluctuation fluidized region, most of which is
475: established after a short transient, very slowly expands as a
476: function of time \cite{the_future_is_bright}. Possibly, after
477: sufficiently long time, all the material has experienced flow and
478: the stress tensor takes the SFS form everywhere, but this may be
479: hard to verify numerically. Similar questions on the validity of
480: the SFS framework can also be asked when the driving rate is made
481: excessively slow. Ultimately, these questions are related to the
482: puzzling nature of the transition between the static and flowing
483: state of granular media \cite{silbert_chute,corwin}. {\em{(iv)}}
484: Is the variation of the effective friction the cause or effect of
485: the smoothness of our shear profiles? We suggest that the
486: spreading of contact force {\em fluctuations}, from the rapidly
487: shearing center to the tails of the shear zones, may elucidate the
488: microscopic mechanism by which the width of the shear zones are
489: selected. In this picture, the spread of fluctuations would also
490: drive the subtle variations of the coarse grained and time
491: averaged stresses, which thus serve to signal an underlying, but
492: unknown, fluctuation driven mechanism \cite{unger2006}.
493:
494: \acknowledgments
495: We thank M. Cates for discussions. MD
496: acknowledges support from the physics foundation FOM and the EU
497: network PHYNECS, and MvH from the science foundation NWO through a
498: VIDI grant. Sandia is a multiprogram laboratory operated by Sandia
499: Corporation, a Lockheed Martin Company, for the United States
500: Department of Energy's National Nuclear Security Administration
501: under contract DE-AC04-
502: 94AL85000.
503:
504:
505: \begin{thebibliography}{0}
506:
507:
508: \bibitem{jamming} A.J. Liu and S.R. Nagel, Nature {\bf 396}, 21 (1998).
509:
510: \bibitem{gm} H. M. Jaeger, S. R. Nagel and R. P. Behringer, Rev. Mod. Phys.
511: {\bf 68}, 1259 (1996).
512:
513: \bibitem{nederman} R. M. Nedderman, {\it Statics and Kinematics of
514: Granular Materials},
515: Cambridge University Press (Cambridge), 1992
516:
517: \bibitem{bob} C.T. Veje, D.W. Howell and R.P. Behringer, Phys.
518: Rev. E {\bf59}, 739 (1999).
519:
520: \bibitem{TCmueth} D. M. Mueth, G. F. Debregeas, G. S. Karczmar {\em et
521: al.}, Nature {\bf406} 385 (2000).
522:
523: \bibitem{bonn} F. Da Cruz, F. Chevoir, D. Bonn {\em et al.},
524: Phys. Rev E {\bf 66} 051305 (2002).
525:
526: \bibitem{Fenistein} D. Fenistein and M. van Hecke, Nature {\bf 425}, 256
527: (2003); D. Fenistein, J. W. van de Meent and M. van Hecke, Phys.
528: Rev. Lett. {\bf 92}, 094301, (2004); {\em ibid.} Phys. Rev. Lett.
529: {\bf 96} 118001 (2006).
530:
531: \bibitem{SFS} M. Depken, W. {van Saarloos}, and M. {van Hecke},
532: Phys. Rev. E {\bf 73}, 031302 (2006).
533:
534:
535:
536: \bibitem{tl} L. Bocquet, J. Errami and T. C. Lubensky, Phys. Rev. Lett.
537: {\bf89}, 184301 (2002).
538:
539: \bibitem{aranson} I. S. Aranson, L. S. Tsimring, Phys. Rev. E {\bf 64},
540: 020301 (2001)
541:
542: \bibitem{bazant}C. H. Rycroft, M. Z.
543: Bazant, G. S. Grest and J. W. Landry, Phys. Rev. E {\bf 73} 051306
544: (2006).
545:
546: \bibitem{gdr} GDR MiDi, Eur. Phys. J. E {\bf 14}, 367, (2004).
547:
548: \bibitem{pouliquen_nature} P. Jop, Y. Forterre, O. Pouliquen,
549: Nature {\bf 441}, 727 (2006)
550:
551: \bibitem{Unger} T. Unger, J. Kert\'esz and
552: D.E. Wolf, Phys. Rev. Lett. {\bf 94}, 178001 (2005).
553:
554: \bibitem{xiang} X. Cheng, J. B. Lechman, A. Fernandez-Barbero {\em et al.}
555: Phys. Rev. Lett. {\bf 96}, 038001 (2006).
556:
557: \bibitem{Inote} The inertial number can be seen as the ratio of
558: the characteristic timescale for a grain to relax back from a
559: local dilation event ($d\sqrt{\rho/P}$) to the strainrate
560: ($1/\dot\gamma$).
561:
562: \bibitem{jop_private} P. Jop, private communications.
563:
564: \bibitem{silbert_chute}
565: L. E. Silbert, D. Ertas, G. S. Grest, T. C. Halsey, D. Levine, and
566: S. J. Plimpton, Phys Rev E {\bf 64} 051302 (2001).
567:
568: \bibitem{footnote_errorbar}
569: The error bar on $\partial_{\theta\theta}\mu_{\rm eff}|_{0}$ is
570: based on its variation with $I_{\rm cut}$, when this cutoff ranges
571: from $1.5 \times 10^{-5}$ to $10^{-4}$.
572:
573: \bibitem{the_future_is_bright} J. B. Lechman, M. Depken, M. van
574: Hecke, W. van Saarloos and G. S. Grest, in preparation.
575:
576: \bibitem{corwin} E. I. Corwin, H. M. Jaeger and S. R. Nagel,
577: Nature {\bf 435}, 1075 (2005).
578:
579: \bibitem{unger2006} J. T\"or\"ok, T. Unger, J. Kertesz and D. E. Wolf,
580: Phys. Rev. E {\bf 75}, 011305 (2007)
581:
582: \end{thebibliography}
583:
584:
585: \end{document}
586:
587: