cond-mat0002102/fc.tex
1: %\documentstyle[aps,multicol,epsf,rotate]{revtex}
2: %\documentstyle[prl,twocolumn,aps]{revtex}
3: \documentstyle[aps,multicol,epsf]{revtex}
4: %\documentstyle[aps]{revtex}
5: %\documentstyle[aps,preprint,epsf]{revtex}
6: 
7: \begin{document}
8: \draft
9: 
10: \title{Packing of Compressible Granular Materials} 
11: 
12: \author{Hern\'an A. Makse, David L.
13: Johnson, and Lawrence M. Schwartz}
14: 
15: \address{ Schlumberger-Doll Research, Old Quarry Road, Ridgefield, CT
16: 06877} 
17: \date{\today} 
18: \maketitle
19: \begin{abstract}
20: 
21: 3D Computer simulations and experiments are employed to study random
22: packings of compressible spherical grains under external confining
23: stress.  Of particular interest is the rigid ball limit, which we
24: describe as a continuous transition in which the applied stress
25: vanishes as $(\phi-\phi_c)^\beta$, where $\phi$ is the (solid phase)
26: volume density.  This transition coincides with the onset of shear
27: rigidity.  The value of $\phi_c$ depends, for example, on whether the
28: grains interact via only normal forces (giving rise to random close
29: packings) or by a combination of normal and friction generated
30: transverse forces (producing random loose packings).  In both cases,
31: near the transition, the system's response is controlled by localized
32: force chains.  As the stress increases, we characterize the system's
33: evolution in terms of (1) the participation number, (2) the average
34: force distribution, and (3) visualization techniques.
35: \end{abstract}
36: \pacs{PACS: 81.06.Rm}
37: %81.06.Rm: porous materials, granular materials
38: 
39: %\newpage
40: 
41: \begin{multicols}{2}
42: 
43: Dense packings of spherical particles are an important starting point
44: for the study of physical systems as diverse as simple liquids,
45: metallic glasses, colloidal suspensions, biological systems, and
46: granular matter \cite{bernal,finney,berryman,ibm,torquato}.  In the
47: case of liquids and glasses, finite temperature molecular dynamics
48: (MD) studies of hard sphere models have been particularly important.
49: Here one finds a first order liquid-solid phase transition as the
50: solid phase volume fraction, $\phi$, increases. Above the freezing
51: point, a metastable disordered state can persist until
52: $\phi\to\phi_{\mbox{\scriptsize RCP}}^-$ \cite{torquato}, where
53: $\phi_{\mbox{\scriptsize RCP}}$ is the density of random close packing
54: (RCP)--- the densest possible random packing of hard spheres.
55: 
56: 
57: 
58: This Letter is concerned with the non-linear elastic properties of
59: granular packings.  Unlike glasses and amorphous solids, this is a
60: zero temperature system in which the interparticle forces are both
61: non-linear, and path (i.e., history) dependent.  [Because these forces
62: are purely repulsive, mechanical stability is achieved only by imposing
63: external stress.]  The structure of packing depends in detail on the
64: forces acting between the grains during rearrangement of grains;
65: indeed, different rearrangement protocols can lead to either RCP or
66: random loose packed (RLP) systems.
67: 
68: 
69: In the conventional continuum approach to this problem, the granular
70: material is treated as an elasto-plastic medium \cite{nedderman}.  
71: However, this
72: approach has been challenged by recent authors \cite{bouchaud}
73: who argue that granular
74: packings represents a new kind of {\it fragile} matter and that more
75: exotic methods, e.g., the fixed principal axis ansatz, are required to
76: describe their internal stress distributions.  These new continuum
77: methods are complemented by microscopic studies based on either 
78: contact dynamics simulations of {\it rigid} spheres or statistical
79: models, such as the q-model, which makes no attempt to take account of
80: the character of the inter-grain forces \cite{chicago,radjai}.
81: 
82: In our view, a proper description of the stress state in granular
83: systems must take account of the fact that the individual grains are
84: deformable.  We report here on a 3D study of deformable spheres
85: interacting via Hertz-Mindlin contact forces.  Our simulations cover
86: four decades in the applied pressure and allow us to understand the
87: regimes in which the different theoretical approaches described above
88: are valid.  Since the grains in our simulations are deformable, the
89: volume fraction can be increased above the hard sphere limit and we
90: are able to study the approach to the RCP and RLP
91: states from this realistic perspective.
92:  Within this framework, the rigid
93: grain limit is described as a continuous phase transition where the
94: order parameter is the applied stress, $\sigma$, which vanishes
95: continuously as $(\phi-\phi_c)^\beta$.  Here $\phi_c$ is the critical
96: volume density, and $\beta$ is the corresponding critical exponent.
97: We emphasize that the fragile state corresponding
98: to rigid grains is reached by looking at the limit $\phi\to\phi^+_c$ from 
99: above.
100: 
101: %We describe the rigid ball limit, as a continuous phase transition
102: %where the order parameter is the stress, $\sigma$, which vanishes
103: %continuously as $(\phi-\phi_c)^\beta$, where $\phi_c$ is the critical
104: %volume density, and $\beta$ is the corresponding critical exponent.
105: %Thus, we characterize the hard sphere limit
106: %%state where deformations are negligible by
107: %by looking at the limit $\phi\to\phi^+_c$ from above.
108: 
109: 
110: Of particular importance is the fact that $\phi_c$ depends on the type
111: of interaction between the grains.  If the grains interact via normal
112: forces only \cite{bubbles}, they slide and rotate freely mimicking the
113: rearrangements of grains during shaking in experiments
114: \cite{bernal,finney,berryman}.  We then obtain the RCP value $\phi_c=
115: 0.634(4)(\approx\phi_{\mbox{\scriptsize RCP}})$.  By contrast, if the
116: grains interact by combined normal and friction generated transverse
117: forces, we get RLP \cite{ibm} at the critical point with
118: $\phi_c=0.6284(2)<\phi_{\mbox{\scriptsize RCP}}$.  The power-law
119: exponents characterizing the approach to $\phi_c$ are not universal
120: and depend on the strength of friction generated shear forces.
121: 
122: %In both cases we show that the system at the critical density is at
123: %the minimal average coordination number, $Z_c$, required for
124: %mechanical stability.  
125: Our results indicate that the transitions at both RCP or RLP are
126: driven by localized force chains. Near the critical
127: density there is a percolative fragile structure which we characterize
128: by the participation number (which measures localization of force
129: chains), the probability distribution of forces, and also by
130: visualization techniques.  
131: A subset of our results are experimentally
132: verified using carbon paper measurements to study force
133: distributions in the granular assembly.   We also consider in some
134: detail the relationship between our work and recent experiments in
135: 2D Couette geometries \cite{behringer}.
136: 
137: %Part of these results are experimentally
138: %verified using standard carbon paper experiments to study force
139: %distributions in the granular assembly, and we also compare with
140: %recent experiments in 2D Couette geometries relevant to the present
141: %study \cite{behringer}.  Our results provide further test to which competing
142: %theories of granular materials could be confronted to.
143: 
144: 
145: 
146: \begin{figure}
147: \centerline{
148: \hbox{ 
149: \epsfxsize=4.cm
150: \epsfbox{fig1a.ps}
151: \epsfxsize=4.cm
152: \epsfbox{fig1b.ps}
153: }}
154: %\vspace{1.cm}
155: \narrowtext
156: \caption{(a) Confining stress and (b)
157: average coordination number
158: as a function of volume fraction for friction and frictionless balls.}
159: \label{pressure-Z-pi}
160: \end{figure}
161: 
162: {\it Numerical Simulations}: To better understand the behavior of real
163: granular materials, we perform granular dynamics simulations of
164: unconsolidated packings of deformable spherical glass beads using the
165: Discrete Element Method developed by Cundall and Strack
166: \cite{cundall}.  Unlike previous work on rigid grains, we work with a
167: system of deformable elastic grains interacting via normal and
168: tangential Hertz-Mindlin forces plus viscous dissipative forces
169: \cite{johnson}.  The grains have shear modulus 29 GPa,
170: Poisson's ratio 0.2 and radius 0.1 mm.
171: %At each grain contact
172: %the normal force is $f_n = \frac{2}{3} ~ C_n R^{1/2}w^{3/2}$, and 
173: %tangential 
174: %force
175: %is $\Delta f_t = C_t (R w)^{1/2} \Delta s$ \cite{johnson}, where
176: %the normal overlap is $2 w= (R_1+R_2) - |\vec{x}_1 - \vec{x}_2|
177: %>0$, and $\vec{x}_1$, $\vec{x}_2$ are the center positions of the
178: %contacting spheres.  The shear displacement is
179: %$s(t)=\int_{t_0}^t v_s(t') dt'$, where $v_s$ is the relative shear
180: %velocity between the spheres, and $t_0$ is the initial time when the
181: %contact is established. $R_1$ and $R_2$ are the radii of the 
182: %two spheres and $R= 2 R_1 R_2/(R_1+R_2)$.  The normal and
183: %tangential prefactors, $C_n=4 G / (1-\nu)$ and $C_t=8 G / (2-\nu)$,
184: %are specified in terms of the shear modulus $G$ and the Poisson's
185: %ratio, $\nu$, of the material from which the grains are made.  We take
186: %$G=29$ GPa and $\nu = 0.2$, typical values for glass.  In our simulations
187: %half the grains had $R_1=0.105$ mm and the other half had 
188: %$R_2=0.095$ mm.  An extra condition of sliding is also added. When the
189: %transverse force exceeds the Coulomb threshold ($\mu_c f_n$) the
190: %contacting grains slide and $f_t = \mu_c f_n$, where $\mu_c$ is the
191: %coefficient of friction between the spheres (typically $\mu_c = 0.3$).
192: %The normal force acts only in traction, $f_n \equiv 0$ when $w<0$.
193: 
194: 
195: Our simulations employ periodic boundary conditions and begin with a
196: gas of 10000 non-overlapping spheres located at random positions in a
197: cube 4 mm on a side.  Generating a mechanically stable packing is not
198: a trivial task \cite{torquato}.  At the outset, a series of
199: strain-controlled isotropic compressions and expansions are applied
200: until a volume fraction slightly below the critical density.  At this
201: point the system is at zero pressure and zero coordination number.  We
202: then compress along the $z$ direction,
203: %holding the lateral system size fixed (uniaxial compression test),
204: until the system equilibrates at a desired vertical stress $\sigma$
205:  and a non-zero average coordination number $\langle Z \rangle$.
206: %During the compression many grains (roughly 
207: %$20\%$ of the contacts) are sliding.  However, once the system reaches 
208: %static equilibrium, all the contacts satisfy $f_t<\mu_c f_n$.  
209: 
210: Figure \ref{pressure-Z-pi}a shows the behavior of the stress as a
211: function of the volume fraction. We find that the pressure vanishes at
212: a critical $\phi_c=0.6284(2)$. Although we cannot rule out a
213: discontinuity in the pressure at $\phi_c$--- as we could expect for a
214: system of hard spheres--- our results indicates that the transition is
215: continuous and the behavior of the pressure can be fitted to a power
216: law form
217: \begin{equation}
218: \sigma \sim (\phi - \phi_c)^{\beta},
219: \label{p}
220: \end{equation}
221: where $\beta = 1.6(2)$.  Our 3D results contrast with recent
222: experiments of slowly sheared grains in 2D Couette geometries
223: \cite{behringer} where a faster than exponential approach to $\phi_c$
224: was found, while they agree qualitative with similar continuous
225: transition found in compressed emulsions and foams \cite{bubbles}.
226: 
227: Figure \ref{pressure-Z-pi}b shows the behavior of the mean
228: coordination number, $ \langle Z \rangle$, as a function of $\phi$.
229: We find
230: \begin{equation}
231:  \langle Z \rangle - Z_c \sim (\phi - \phi_c)^{\theta},
232: \label{z}
233: \end{equation}
234: where $Z_c=4$ is a minimal coordination number, and $\theta=0.29(5)$
235: is a critical exponent.  At criticality the system is very loose and
236: fragile with a very low coordination number. The value of $Z_c$ can be
237: understood in term of constraint arguments as discussed in
238: \cite{edwards}; in the rigid ball limit, for a disordered system with
239: both normal and transverse forces, we find $Z_c = D+1 = 4$
240: \cite{edwards}.  As we compress the system more contacts are created,
241: providing more constraints so that the forces become overdetermined.
242: %The grains start
243: %to deform, and the fragile initial RCP  structure gives way to a more
244: %compa
245: 
246: We notice that $\phi_c$ obtained for this system is considerably lower
247: than the best estimated value at RCP \cite{berryman}, 
248: $\phi_{\mbox{\scriptsize
249: RCP}}=0.6366(4)$ obtained by Finney 
250: \cite{finney} using ball bearings.  This latter value is obtained by
251: carefully vibrating the system and letting the grains settle into the
252: most compact packing. Numerically, this is achieved by allowing the
253: grains reach the state of mechanical equilibrium interacting only via
254: normal forces. By removing the transverse forces, grains can slide
255: freely and find most compact packings than with transverse forces.
256: Numerically we confirm this by equilibrating the system at zero
257: transverse force.  The critical packing fraction found in
258: this way is $\phi_c=0.634(4)$($ \approx \phi_{\mbox{\scriptsize RCP}}$
259: within error bars).  The stress behaves as in Eq. (\ref{p}) but with a
260: different exponent $\beta=2.0(2)$ (Fig. \ref{pressure-Z-pi}a).  At the
261: critical volume fraction the average coordination number is now $
262: Z_c=6$ [and $\theta=0.94(5)$, Fig. \ref{pressure-Z-pi}b], which again
263: can be understood using constraint arguments which would give a
264: minimal coordination number equal to 2D for frictionless rigid balls
265: \cite{edwards}.
266: 
267: 
268: \begin{figure}
269: %\centerline{ \vbox{ \hbox{\epsfxsize=12.cm \epsfbox{fig1a.ps} }
270: %\epsfxsize=12.cm \epsfbox{fig1b.ps} }} 
271: \centerline{ 
272: \hbox{
273: \epsfxsize=4.cm \epsfbox{fig2a.ps} 
274: \epsfxsize=4.cm \epsfbox{fig2b.ps} } }
275: \centerline{  
276: \epsfxsize=5.cm  \epsfbox{fig2c.ps} }
277: \narrowtext
278: %\vspace{1.cm}
279: \caption{Distribution of forces for different confining  stresses $\sigma$
280: obtained (a) in
281: the numerical simulations of friction balls, and (b) in the carbon paper
282: experiments.  The straight solid lines are fittings to
283: exponential forms and the dashed lines are fittings to Gaussian forms.
284: In both graphs we shift down the distributions corresponding to the
285: two larger stresses for clarity.
286: (c) Participation number versus external stress for the same system as in 
287: (a).}
288: \label{distri}
289: \end{figure}
290: \vspace{-.5cm}
291: We conclude that the value $\phi_c \approx
292: 0.6288<\phi_{\mbox{\scriptsize RCP}}$ found with transverse forces
293: corresponds to the RLP limit, experimentally achieved by pouring balls
294: in a container but without allowing for further rearrangements \cite{ibm}.
295: Experimentally, stable loose 
296: packings with $\phi$ as low as 0.60 have been
297: found \cite{ibm}.  In our simulations,  $\phi_c$ lower than $0.6288$ can
298: be obtained by increasing the strength of the tangential forces.  This
299: is in agreement with experiments of Scott and Kilgour \cite{scott2}
300: who found that the maximum packing density of spheres decreases with
301: the surface roughness (friction) of the balls.
302: 
303: While previous studies characterized RCP's and RLP's by using radial
304: distribution functions and Voronoi constructions \cite{finney}, we
305: take a different approach which allow us to compare our results
306: directly with recent work on force transmissions in granular matter.
307: %Dense random packings have been characterized by their radial
308: %distribution functions and also by using Voronoi constructions 
309: %\cite{finney}. Here we take a different approach and study the 
310: %changes in the spatial distribution of interparticle
311: %forces in the packing as $\phi_c$  is approached.
312: Previous studies of granular media indicate that, for forces greater
313: than the average value, the distribution of inter-grain contact forces
314: is exponential \cite{chicago,radjai}.
315: %, a result verified using scalar
316: %models of force transmission in granular materials, such as the
317: %``q-model'' \cite{chicago}, and through numerical simulations of rigid
318: %particles packs \cite{radjai}.  
319: In addition, photo-elastic
320: visualization experiments and simulations \cite{force_chains,chicago}
321: show that contact forces are strongly localized along ``force chains''
322: which carry most of the applied stress.
323: %The continuum limit of 
324: %a vector version of the q-model,
325: %with a particular constitutive relation 
326: % \cite{bouchaud} (see
327: %\cite{savage} for alternative formulations)
328: %between the stress components, gives rise to hyperbolic equations 
329: %for stress propagation whose characteristics are believed to describe
330: %force chains.  
331: The existence of force chains and exponential force distributions are
332: thought to be intimately related.
333: 
334: Here we analyze this scenario in the entire range of pressures: from
335: the $\phi_c$ limit and up.  Figure \ref{distri}a shows the force
336: distribution obtained in the simulations with friction balls.  At low
337: stress, the distribution is exponential in agreement with previous
338: experiments and models.
339: %Our results 
340: %indicate that this behavior extends to the full range of forces.  There 
341: %is no indication of a power-law distribution for forces smaller than
342: %$\langle f \rangle$ as suggested by other models 
343: %\cite{chicago,radjai}.  
344: When the system is compressed further, we find a gradual transition to
345: a Gaussian force distribution. We observe a similar transition in our
346: simulations involving frictionless grains under isotropic compression.
347: This suggests that our results are generic, and do not depend,
348: qualitatively, on the preparation history or on the existence of
349: friction generated transverse forces between the grains.
350: %Of course, the precise location of the transition may depend on the system's 
351: %preparation.  
352: %For example, we see in Figs. \ref{distri} that our simulations 
353: %show a crossover to Gaussian behavior for 
354: %stresses that are, in general, larger 
355: %than those found in the experiments. This may be due to differences 
356: %in the experimental and numerical set up protocols.  [In the experiments 
357: %Janssen's effect \cite{pgg} affects the total pressure at the bottom 
358: %of the bin while in the simulations, where we have periodic boundary
359: %conditions and no rigid lateral walls, there is no corresponding effect.]
360: %Moreover, we will see below indications that the transition to a Gaussian 
361: %behavior starts at $\approx 2.1$MPa, in better agreement with the 
362: %experimental data.
363: 
364: 
365: Physically, we find that the transition from Gaussian to exponential
366: force distribution is driven by the localization of force chains as
367: the applied stress is decreased.  In granular materials, with
368: particles of similar size, localization is induced by the disorder of
369: the packing arrangement.  To quantify the degree of localization, we
370: consider the participation number $\Pi$:
371: \begin{equation}
372: \Pi\equiv\left(M\sum_{i=1}^M q_i^2\right)^{-1} \: \: .
373: \label{pi}
374: \end{equation}
375: Here $M$ is the number of contacts between the spheres, $ \langle
376: Z\rangle =2 M / N$ is the average coordination number, and $N$ is the
377: number of spheres.  $q_i\equiv {f_i}/{\sum_{j=1}^{M}{f_j}}$, where
378: $f_i$ is the magnitude of the total force at every contact. From the
379: definition (\ref{pi}), $\Pi=1$ indicates a limiting state with a
380: spatially homogeneous force distribution (${q_i}=1/M$, $\forall i$).
381: On the other hand, in the limit of complete localization, $\Pi\approx
382: 1/M\to 0$ and $M\rightarrow\infty$.
383: 
384: 
385: Figure \ref{distri}c shows our results for $\Pi$ versus
386:  $\sigma$. Clearly, the system is more localized at low stress than at
387:  high stress.  Initially, the growth of $\Pi$ is logarithmic,
388:  indicating a smooth delocalization transition.  This behavior is seen
389:  up to $\sigma\approx$ 2.1 MPa, after which the participation number
390:  saturates to a higher value:
391: \begin{equation}
392: \begin{array}{rcll}
393: \Pi(\sigma) & \propto & \log (\sigma) & \qquad[ \mbox{$\sigma<$2.1
394: MPa}]\\ \Pi(\sigma) & \approx & 0.62 & \qquad[ \mbox{$\sigma>$2.1
395: MPa}]\\
396: \end{array}
397: \label{62}
398: \end{equation}
399: This behavior suggests that, near the critical density, the forces are
400: localized in force chains sparsely distributed in space.  As the
401: applied stress is increased, the force chains become more dense, and
402: are thus distributed more homogeneously.
403: 
404: 
405: How might we expect the participation number to depend upon other
406: system parameters when the forces are transmitted principally by force
407: chains?  In an idealized situation, the system has $N_{FC}$ force
408: chains, each of which has $N_z$ spheres. Each sphere in a force chain
409: has two major load bearing contacts, which loads must be approximately
410: equal.  In the lateral directions, roughly four weak contacts are
411: required for stability. These contacts carry a fraction $\alpha<1$ of
412: the major vertical load.  All other contacts have $f_i \approx 0$.
413: Under these assumptions,
414: 
415: \begin{equation}
416: \Pi = \frac{2}{ \langle Z \rangle}\frac{(1+2\alpha)^2}{(1+2\alpha^2)}
417: \frac{ N_{FC} N_z}{ N} \leq \frac{2}{ \langle Z \rangle
418: }\frac{(1+2\alpha)^2}{(1+2\alpha^2)} \:\:\:.
419: \label{pifc}
420: \end{equation}
421: The last inequality becomes an equality {\it iff} all the balls are in
422: force chains.  From our simulations at large pressure $\alpha\approx
423: 2/5$, so at $\langle Z \rangle \approx 8$, $\Pi \approx 0.62$, which
424: implies that the system has been completely homogenized.
425: %that at $\sigma$ = 100 MPa, $\Pi = 0.62 = 4/ \langle Z \rangle$,
426: Although Eq. (\ref{pifc}) is oversimplified, we believe that the
427: change in slope in Fig. \ref{distri}c is emblematic of the complete
428: disappearance of well-separated chains.
429: 
430: \begin{figure}
431: \centerline{
432: \hbox{
433: \epsfxsize=5.cm
434: %\epsfxsize=5cm
435: \epsfbox{fig3a.ps}
436: \epsfxsize=5.cm
437: %\epsfxsize=5cm
438: \epsfbox{fig3b.ps}
439: }
440: }
441: %\vspace{.2cm}
442: \narrowtext
443: \caption{Example of percolating
444: force chains for the same system as in Fig. \protect\ref{distri}a: 
445: (a) near $\phi_c$ and (b) away from $\phi_c$ at large
446: confining
447: stress. The color code of the chains is according to the total force
448: in N carried by the
449: chains.}
450: \label{chains}
451: \end{figure}
452: 
453: 
454: 
455: The localization transition can be understood by studying the behavior
456: of the forces during the loading of the sample.  Clearly, visualizing
457: forces in 3D systems is a complicated task.  In order to exhibit the
458: rigid structure from the system we visually examine all the forces
459: larger than the average force; these carry most of the stress of the
460: system.  The forces smaller than the average are thought to act as an
461: interstitial subset of the system providing stability to the buckling
462: of force chains \cite{force_chains,radjai}.  We look for force chains
463: by starting from a sphere at the top of the system, and following the
464: path of maximum contact force at every grain.  We look only for the
465: paths which percolate, i.e., stress paths spanning the sample from the
466: top to the bottom.  In Fig. \ref{chains} we show the evolution of the
467: force chains thus obtained for two extreme cases of confining stress.
468: We clearly see localization at low confining stress: the force-bearing
469: network is concentrated in a few percolating chains.  At this point
470: the grains are weakly deformed but still well connected. We expect a
471: broad force distribution, as found in this and previous studies.  As
472: we compress further, new contacts are created and the density of force
473: chains increases. This in turn gives rise to a more homogeneous
474: spatial distribution of forces, which is consistent with the crossover
475: to a narrow Gaussian distribution.
476: 
477: {\it Experiments}: Some of the predictions of our numerical study can
478: be tested using standard carbon paper experiments \cite{chicago},
479: which have been used successfully in the past to study the force
480: fluctuations in granular packings.  A Plexiglas cylinder, 5 cm
481: diameter and varying height (from 3 cm to 5 cm), is filled with
482: spherical glass beads of diameter $0.8\pm0.05$ mm.  At the bottom of
483: the container we place a carbon paper with white glossy paper
484: underneath.  We close the system with two pistons and we allow the top
485: piston to slide freely in the vertical direction, while the bottom
486: piston is held fixed to the cylinder.  The system is compressed in the
487: vertical direction with an Inktron$^{TM}$ press and the beads at the
488: bottom of the cylinder left marks on the glossy paper. We digitize
489: this pattern and calculate the ``darkness'' \cite{chicago} of every
490: mark on the paper.  To calibrate the relationship between the marks
491: and the force, a known weight is placed on top of a single bead; for
492: the forces of interest in this study (i.e., from $\approx 0.05$N to 6
493: N), there is a roughly linear relation between the darkness of the dot
494: and the force on the bead.
495: 
496: We perform the experiment for different external forces, ranging from
497: 2000 N to 9000 N, and different cylinder heights.  The corresponding
498: vertical stress, $\sigma$, at the bottom of the cylinder ranges
499: between 100 KPa and 2.3 MPa (as measured from the darkness of the
500: dots).
501: %The stress at the bottom of the cylinder is smaller than the
502: %stress applied at the top due to the screening of the walls 
503: %(Janssen's
504: %effect \cite{pgg}).  
505: The results of four different measurements are shown in
506: Fig. \ref{distri}b. For $\sigma$ smaller than $\approx 750$ KPa, the
507: distribution of forces, $f$, at the bottom piston decays
508: exponentially:
509: \begin{equation}
510: P(f) = \langle f \rangle ^{-1}\exp{[- f/\langle f \rangle ]}, \qquad[
511: \mbox{ $ \sigma<$ 750 KPa}],
512: \end{equation}
513: where $ \langle f \rangle$ is the average force.  When the stress is
514: increased above 750 KPa there is a gradual crossover to a Gaussian
515: force distribution as we find in the simulations.  For example, at 2.3
516: MPa we have
517: \begin{equation}
518: P(f) \propto \exp{\left[- k^2 \left(f-f_o\right)^2\right]}, \qquad[
519: \mbox{$\sigma=$ 2.3 MPa}].
520: \end{equation}
521: %Because the distribution is cut off for negative values of $f$, one
522: %has $ \langle f\rangle 
523: %= f_o + \exp[-k^2 f_o^2]/\sqrt{\pi}k[1 + {\rm erf}(kf_o)]$.  We
524: %find 
525: where $k f_o \approx 1$, and therefore $ \langle f \rangle \approx
526: f_o$. Similar results have been found in 2D geometries
527: \cite{behringer}.
528: 
529: 
530: {\it Discussion: } In summary, using both numerical simulations and
531: experiments, we have studied unconsolidated compressible granular
532: media in a range of pressures spanning almost four decades.  In the
533: limit of weak compression, the stress vanishes continuously as
534: $(\phi-\phi_{c})^\beta$, where $\phi_c$ corresponds to RLP or RCP
535: according to the existence or not of transverse forces between the
536: grains, respectively. At criticality, the coordination number
537: approaches a minimal value $Z_c$ (=4 for friction and 6 for
538: frictionless grains) also as a power law.  Our result $Z_c=6$ agrees
539: with experimental analysis of Bernal packings for close contacts
540: between spheres fixed by means of wax \cite{bernal}, and our own
541: analysis of the Finney packings \cite{finney} using the actual sphere
542: center coordinates of 8000 steel balls.  However, no similar
543: experimental study exists for RLP which could be able to confirm
544: $Z_c=4$.  A critical slowing down--- the time to equilibrate the
545: system increases near $\phi_c$--- and the emergency of shear rigidity
546: (to be discussed elsewhere) is also found at criticality.  The
547: distribution of forces is found to decay exponentially.  The system is
548: dominated by a fragile network of 
549: relatively few force chains which span the system.
550: 
551: When the stress is increased away from $\phi_c$ to the point that the
552: number of contacts has significantly increased from its initial value
553: $Z_c$ we find: (1) the distribution of forces crosses over to a
554: Gaussian (2) the participation number increases, and then abruptly
555: saturates and (3) the density of force chains increases to the point
556: where it no longer makes sense to describe the system in those terms.
557: Our simulations indicate that the crossover is associated with a loss
558: of localization and the ensuing homogenization of the force-bearing
559: stress paths.
560: 
561: %The system has become homogeneous down to a scale comparable to the
562: %grain size.
563: %Our results could be interpreted as a crossover 
564: %between a fragile state at  $\langle Z \rangle \approx Z_c$ and low 
565: %confining pressure,
566: %to an elastic state at larger pressures.
567: %In this regime the balls are considerably deformed, and an analysis of
568: %stress distributions independent of the state of strain of the sample,
569: %as assumed in the q-model, is not valid.  Thus deviations from the
570: %prediction of the q-model are due to the deformability of particles in
571: %realistic compressible media.
572: %\cite{socolar2}.   
573: 
574: 
575: %Interestingly, it has been recently conjectured that cohesionless 
576: %rigid balls at the minimal coordination $Z_c$ are 
577: %in a ``fragile state'' \cite{fragile}, 
578: %i.e.,
579: %that they are able to support only certain loads without severe 
580: %rearrangements.  Further,
581: %it is conjectured that finite deformability of the grains could restore the 
582: %elastic 
583: %response of the system \cite{savage}. 
584: % Thus, our results could be interpreted as a crossover 
585: %between a fragile state at  $\langle Z \rangle \approx Z_c$ and low 
586: %confining pressure,
587: %to an elastic state at larger pressures.
588: 
589: \vspace{-.7cm}
590: 
591: \begin{references}
592: 
593: \vspace{-1.8cm}
594: 
595: \bibitem{bernal} 
596: J. D. Bernal, Nature {\bf 188}, 910 (1960);
597: %Proc. Roy. Soc. London, Ser. A {\bf 280}, 299 (1964);
598: {\it Disorder and Granular Media}, edited by D. Bideau and A. Hansen 
599: (Elsevier, Amsterdam, 1993).
600: 
601: \bibitem{finney}
602: J. L. Finney, Proc. Roy. Soc. London, Ser. A {\bf 319}, 479 (1970).
603: 
604: \bibitem{berryman}
605: J. G. Berryman, Phys. Rev. A {\bf 27}, 1053 (1983).
606: 
607: \bibitem{ibm}
608:  G. D. Scott, Nature {\bf 188}, 908 (1960);
609: G. Y. Onoda and E. G. Liniger, Phys. Rev. Lett. {\bf 64}, 2727 (1990);
610: 
611: \bibitem{torquato}
612: M. D. Rintoul and S. Torquato, Phys. Rev. Lett. {\bf 77}, 4198 (1996).
613: 
614: \bibitem{nedderman}
615: {\it Statics and Kinematics of Granular Materials}, by R. M.
616: Nedderman (Cambridge University Press, 1992).
617: 
618: 
619: \bibitem{bouchaud}
620: J. P. Bouchaud, {\it et al.},
621: %P. Claudin, M. E. Cates, and J. P. Wittmer, 
622: in 
623: {\it Physics of Dry Granular Media}, H. J. Herrmann, J. P. Hovi,
624: and S. Luding (eds) (Kluwer, Dordrecht, 1998).
625: 
626: 
627: \bibitem{chicago}
628: C.-H. Liu, {\it et al.},
629: % S. R. Nagel, D. A. Schecter, S. N. Coppersmith, S. N. Majumdar,
630: %O. Narayan, T. A. Witten, 
631: Science {\bf 269}, 513 (1995); 
632: % S. N. Coppersmith, {\it et al.},
633: %C.-H. Liu,  S. N. Majumdar,  O. Narayan, T. A. Witten,
634: %Phys. Rev. E {\bf 53}, 4673 (1996);
635: D. M. Mueth,
636: H. M. Jaeger, and S. R.  Nagel, Phys. Rev. E {\bf 57}, 3164 (1998).
637: 
638: 
639: 
640: \bibitem{radjai}
641: %F. Radjai, D. E. Wolf, S. Roux, M. Jean, and J. J. Moreau,
642: %Phys. Rev. Lett. (1997).
643: F. Radjai, {\it et al.},
644: %M. Jean, and J. J. Moreau,  S. Roux,
645: Phys. Rev. Lett. {\bf 77}, 3110 (1997).
646: 
647: 
648: %\bibitem{jenkins}
649: %P. A. Cundall, J. T.  Jenkins, and I. Ishibashi, in {\it Powder and Grains
650: %1989}, (Biarez and Gourv\`es, eds) (Balkema, Rotterdam, 1989).
651: %G. W. Baxter, in {\it Powders and Grains 97}, R. P. Behringer and 
652: %J. T. Jenkins (eds)
653: %(Balkema, Rotterdam, 1997).
654: 
655: 
656: \bibitem{bubbles}
657: Another experimental realization of frictionless balls is the packing of 
658: bubbles and compressed microemulsions. 
659:  D. J. Durian, Phys. Rev. Lett. {\bf 75}, 4780 (1995);
660:  M.-D. Lacasse, {\it et al.},
661: %G. S. Grest, D. Levine, T. G. Mason,
662: %and D. A. Weitz, 
663: Phys. Rev. Lett. {\bf 76}, 3448 (1996)
664: 
665: 
666: \bibitem{behringer}
667: D. Howell, R. P. Behringer, and C. Veje, Phys. Rev. Lett. {\bf 82}, 5241 
668: (1999).
669: 
670: 
671: \bibitem{cundall}
672: P. A. Cundall and O. D. L. Strack, G\'eotechnique {\bf 29}, 47 (1979).
673: 
674: 
675: \bibitem{johnson}
676: {\it Contact Mechanics}, by
677: K. L. Johnson (Cambridge University Press, 1985).
678: 
679: %\bibitem{makse}
680: %H. A. Makse, N. Gland, D. L. Johnson, L. Schwartz, 
681: %Phys. Rev. Lett. {\bf 83}, 5070 (1999).
682: 
683: 
684: 
685: 
686: \bibitem{edwards}
687: S. Alexander, Phys. Rep. {\bf 296}, 65 (1998);
688: S. F. Edwards and D. V. Grinev, Phys. Rev. Lett. {\bf 82}, 5397 (1999);
689: A. Tkachenko and T. A. Witten, Phys. Rev. E {\bf 60},
690: 687 (1999).
691: 
692: 
693: \bibitem{scott2}
694: G. D. Scott and D. M. Kilgour, Br. J. Appl. Phys. {\bf 2}, 863 (1969).
695: 
696: 
697: \bibitem{force_chains}
698: %P. Dantu, in
699: %{\it Proc. 4th Int. Conf. on Soil Mechanics Foundation Engineering}, p. 144
700: % (Butterworths, London, 1957);
701: P. Dantu, Ann. Ponts Chauss. {\bf IV}, 193 (1967).
702: %A. Drescher and G. De Josselin De Jong, J. Mech. Phys. Solids {\bf 20}, 337
703: %(1972).
704: %\bibitem{cundall2}
705: %P. A. Cundall and O. D. L. Strack, in {\it Mechanics of Granular Materials:
706: %New Models and Constitutive Relations},
707: %(J. T. Jenkins and M. Satake, eds) (1983);
708: %C. Thornton, and D. J. Barnes, Acta Mechanica {\bf 64}, 45 (1986).
709: 
710: 
711: %\bibitem{pgg}
712: %P.-G. de Gennes, Physica A {\bf 261}, 267 (1998);
713: 
714: 
715: 
716: %\bibitem{socolar}
717: %J. E. S. Socolar, Phys. Rev. E {\bf 57}, 3204 (1998).
718: 
719: 
720: 
721: 
722: 
723: %\bibitem{savage}
724: %S. B. Savage, in {\it Powders and Grains 97}, R. P. Behringer and 
725: %J. T. Jenkins (eds)
726: %(Balkema, Rotterdam, 1997).
727: 
728: 
729: %\bibitem{socolar2}
730: %Other scalar lattice models of deformable springs
731: %give rise to Gaussian distribution of forces. See, 
732: %M. G. Sexton, J. E. S. Socolar, and D. G. Schaeffer, Phys. Rev. E {\bf 60} 1999
733: %(1999).
734: 
735: 
736: %\bibitem{fragile}
737: %M. E. Cates, {\it et al}., Phys. Rev. Lett. {\bf 81}, 1841 (1998).
738: 
739: %\bibitem{wolf}
740: %J. Sch\"afer, S. Dippel, and D. E. Wolf, J. Phys. I France {\bf 6}, 5 (1996).
741: 
742: 
743: 
744: 
745: %\bibitem{ball}
746: %R. C. Ball and D. V. Grinev, (unpublished).
747: 
748: 
749: \end{references}
750: 
751: 
752: %ACKNOWLEDGMENTS.
753: %We would like to thank J. St. Germain and B. Andrews for their valuable help
754: %in the experiments and visualization, and N. Gland
755: %and D. Pissarenko
756: %for discussions.
757: 
758: %\newpage
759: 
760: %FIG. \ref{pressure-Z-pi}.   (a) Confining stress and (b)
761: %average coordination number
762: %as a function of volume fraction for friction and frictionless balls.
763: 
764: 
765: %FIG. \ref{distri}. 
766: %Distribution of forces for different confining  stresses $\sigma$
767: %obtained (a) in
768: %the numerical simulations of friction balls, and (b) in the carbon paper
769: %experiments.  The straight solid lines are fittings to
770: %exponential forms and the dashed lines are fittings to Gaussian forms.
771: %In both graphs we shift down the distributions corresponding to the
772: %two larger stresses for clarity.
773: %(c) Participation number versus external stress for the same system as in 
774: %(a).
775: 
776: %FIG. \ref{chains}. 
777: %Example of percolating
778: %force chains for the same system as in Fig. \protect\ref{distri}a: 
779: %(a) near $\phi_c$ and (b) away from $\phi_c$ at large
780: %confining
781: %stress. The color code of the chains is according to the total force
782: %in N carried by the
783: %chains.
784: 
785: 
786: %\newpage
787: 
788: 
789: %\newpage
790: 
791: 
792: %\newpage
793: 
794: 
795: 
796: 
797: \end{multicols}
798: 
799: 
800: 
801: \end{document}
802: 
803: 
804: %We study random packings of compressible spherical grains subjected to
805: %an external confining stress near and above the densest packing
806: %fraction in 3D.  The rigid ball limit is described as a continuous
807: %transition where the stress of the system vanishes as
808: %$(\phi-\phi_c)^\beta$, where $\phi_c$ is the critical volume density;
809: %a transition which is coincident with the appearance of shear
810: %rigidity.  If the grains interact via normal forces only, the value of
811: %$\phi_c$ corresponds to the random close packing (RCP) fraction.  By
812: %contrast, if the grains interact by combined normal and friction
813: %generated transverse forces, we find random loose packing (RLP) at the
814: %critical point.  Our results indicate that the transition is driven by
815: %localized force chains. Near $\phi_c$ there is a percolative fragile
816: %structure which we characterize by the participation number, the
817: %probability distribution of forces, and visualization techniques.
818: %Moreover, we find a smooth transition from strong localization of the
819: %force distribution along ``force chains'' at low pressures
820: %to a delocalized and more
821: %homogeneous state as the grains deform with increased confining
822: %stress; a crossover evidenced in the  change of the probability
823: %distribution of forces.
824: %%we find that the exponential distribution of intergranular forces
825: %%observed by recent authors is valid only at low confining stresses
826: %%where the grains are, in effect, rigid.  At larger stresses, above
827: %%RCP, we find a crossover to a narrower, Gaussian distribution.  This
828: %%crossover reflects 
829: %%In real granular materials the grains are compressible, a fact that
830: %%strongly influences the distribution of contact forces.  
831: 
832: 
833: %%%%%%%%%%%%%%%%%%%%%%%%%%%
834: %%%%%%%{\bf old}
835: %%%%%%%%%%%%%%%%%%%%%%%%%%%
836: %Dense random packing of spheres \cite{bernal,berryman} are useful
837: %models for many physical systems such as the arrangements of atoms in
838: %simple fluids, amorphous metals, or glasses as well as colloidal
839: %suspensions, or biological systems \cite{torquato}.
840: 
841: %Random packings have been studied experimentally and in molecular
842: %dynamics and Monte Carlo computer simulations of hard spheres at a
843: %finite temperature.  An equilibrium hard sphere system presents a
844: %first order liquid-solid phase transition as the volume fraction
845: %$\phi$ (defined as the ratio between the volume occupied by the
846: %spheres and the volume of the system) is increased up to the freezing
847: %point, above which the disordered state can persist in a metastable
848: %branch until $\phi\to\phi_{\mbox{\scriptsize RCP}}^-$ \cite{torquato},
849: %where $\phi_{\mbox{\scriptsize RCP}}$ is the density of random close
850: %packing (RCP)--- the densest possible packing of random hard spheres.
851: 
852: Another realization of the problem is related to macroscopic granular
853: materials. In a typical experiment spherical balls are poured in a
854: container which is squeezed and shaken to achieve the best possible
855: packing.  The best estimation of the volume fraction at RCP was
856: obtained by Finney \cite{finney} for ball bearings:
857: $\phi_{\mbox{\scriptsize RCP}}=0.6366(4)$ \cite{berryman}.  Looser
858: packings of mechanically stable spheres, with packing fractions
859: ranging from $\approx 0.60$ to RCP are much less reproducible and are
860: generated by rolling spheres into a container without shaking--- this
861: limit is known as the random loose packing (RLP) \cite{ibm}.
862: 
863: Unlike glasses and amorphous solids, granular materials are zero
864: temperature systems whose interparticle forces are exclusively
865: repulsive: the rigidity of the system arises due to the applied
866: external stress.
867: %Thus, the only way to
868: %achieve a mechanically stable packing is by imposing an external
869: %pressure.  
870: In this paper we study the approach to RCP and RLP for realistic
871: compressible granular materials interacting via non-linear
872: (Hertz-Mindlin) normal and friction generated transverse contact
873: forces.  Since the balls are deformable, the volume fraction of the
874: system can be increased above the hard sphere limit by the action of
875: the external pressure.
876: % and we are interested in the limit 
877: %$\phi\to\phi_{\mbox{\scriptsize RCP}}^+$.
878: 
879: 
880: % We show that the RCP
881: %state is achieved in our system of ``soft'' spheres
882: %as a limit where the pressure of
883: %the system vanishes as a power-law as $\phi$ approaches a critical density 
884: %$\phi_c=0.634(4)(\approx\phi_{\mbox{\scriptsize RCP}}) $.
885: %This critical value is obtained 
886: % when the grains are allowed to rearrange, sliding  and  
887: %rotating freely (interacting 
888: %via normal forces only), mimicking the rearrangements of grains 
889: %during  shaking.
890: We describe the rigid ball limit, as a continuous phase transition
891: where the order parameter is the stress, $\sigma$, which vanishes
892: continuously as $(\phi-\phi_c)^\beta$, where $\phi_c$ is the critical
893: volume density, and $\beta$ is the corresponding critical exponent.
894: Thus, we characterize the hard sphere limit
895: %state where deformations are negligible by
896: by looking at the limit $\phi\to\phi^+_c$ from above.
897: 
898: 
899: Of particular importance is the fact that $\phi_c$ depends on the type
900: of interaction between the grains.  If the grains interact via normal
901: forces only (so that they slide and rotate freely mimicking the
902: rearrangements of grains during shaking \cite{bubbles}), we obtain the
903: value of $\phi_c= 0.634(4)(\approx\phi_{\mbox{\scriptsize RCP}})$
904: corresponding to RCP.  By contrast, if the grains interact by combined
905: normal and friction generated transverse forces, we find random loose
906: packing (RLP) at the critical point with
907: $\phi_c=0.6284(2)<\phi_{\mbox{\scriptsize RCP}}$.  The power-law
908: exponents characterizing the approach to $\phi_c$ are not universal
909: and depend on the strength of friction generated shear forces as well.
910: 
911: %When the balls are allowed to interact 
912: %with friction generated transverse forces we find that the 
913: %pressure vanishes but  at a lower packing fraction 
914: %$\phi_c=0.6284(2)<\phi_{\mbox{\scriptsize RCP}}$. 
915: %We argue that this corresponds to the limit 
916: %experimentally achieved by
917: %pouring balls in a container but without allowing for 
918: %further rearrangements.  
919: %The power-law exponents characterizing the approach to $\phi_c$ are not
920: %universal and depend
921: %on the strength of friction generated shear forces.
922: 
923: In both cases we show that the system at the critical density is at
924: the minimal average coordination number, $Z_c$, required for
925: mechanical stability.  Our results indicate that the transition at the
926: RCP or RLP is driven by localized force chains. Near the critical
927: density there is a percolative fragile structure which we characterize
928: by the participation number (which measures localization of force
929: chains), the probability distribution of forces, and also by
930: visualization techniques.  Part of these results are experimentally
931: verified using standard carbon paper experiments to study force
932: distributions in the granular assembly, and we also compare with
933: recent experiments in 2D Couette geometries relevant to the present
934: study \cite{behringer}.
935: 
936: %For grains
937: %interacting via normal and tangential forces this number is D+1 (D
938: %is the dimension), and for frictionless 
939: %balls it is
940: %2D as conjectured recently using constraint counting arguments
941: %\cite{edwards}, and we confirm these results with our numerical
942: %simulations, for D=3.  
943: %Concomitant with the approach to the critical
944: %volume fraction  we find a
945: %change in the probability distribution and in the degree of spatial
946: %correlation of the interparticle forces: 
947: %localized force chains with an exponential probability
948: %distribution--- as observed in recent studies 
949: %\cite{force_chains,chicago}---
950: %are found at low confining stress
951: %giving way to a considerably less localized and homogeneous
952: %arrangement of the forces with a Gaussian probability distribution at
953: %higher stress away from the critical density.  
954: 
955: 
956: %The densest possible acking of equal sized spheres is the ordered
957: %close packing (fcc) with a volume fraction $\phi=\pi/18^2\approx 0.7405$ in 
958: %three dimensions.
959: 
960: %Specifically, we consider the force distribution 
961: %when the packing is subjected to uniaxial strains which can be large 
962: %enough to
963: %deform the individual beads and significantly change the average 
964: %coordination 
965: %number.  For low confining stresses (i.e., in the rigid grain limit) we 
966: %confirm previous results; both experiment and simulation indicate that 
967: %the force distribution is exponential.  
968: %%For grains interacting via a combination 
969: %%of normal and tangential forces, constraint counting arguments predict
970: %%$\langle Z \rangle = D+1$, where $D$ is the spatial dimension 
971: %%\cite{edwards}.  
972: %However, when the system is compressed at large 
973: %confining stresses we find a transition to a Gaussian force distribution.  
974: %At this point the grains are considerable
975: %deformed and the system is like 
976: % an amorphous elastic solid \cite{edwards} with a correspondingly
977: %more homogeneous force distribution. 
978: 
979: 
980: 
981: 
982: 
983: 
984: