Powder Technology 214 (2011) 491499

Contents lists available at SciVerse ScienceDirect

Powder Technology
journal homepage: www.elsevier.com/locate/powtec

Experimental analysis of the dynamic properties of wet granular matter in a

rotating drum
S.H. Chou, S.S. Hsiau
Department of Mechanical Engineering, National Central University, Jhongli, 32001, Taiwan, ROC

a r t i c l e

i n f o

Article history:
Received 14 June 2011
Received in revised form 30 August 2011
Accepted 8 September 2011
Available online 16 September 2011
Keywords:
Rotating drum
Liquid bridge force
Velocity
Granular temperature
Energy dissipation

a b s t r a c t
We performed experiments to measure the dynamic properties of wet granular matter in a rotating drum device. Four different amounts of liquid and rotation speeds were used in the experiments. The purpose was to
quantify the effect of the cohesive force in the granular system. The results show that when only very small
amounts of liquid were added, no liquid bridges formed. This is because the liquid was rst trapped on the
surface of the particles due to the particle roughness. When the volume fraction of the uid became larger,
liquid bridges formed on almost every particle. The results showed that the addition of liquid contents, and
changing the rotation speed, both had a signicant effect on the dynamic properties of granular matter.
This was due to the hysteretic formation and rupturing of liquid bridges and the introduction of inertial
force to the device. After the liquid bridges formed between all particles the average energy dissipation
due to the hysteretic formation and rupturing of the liquid bridges increased with an increase in the liquid
content.
2011 Elsevier B.V. All rights reserved.

1. Introduction
The handling and processing of granular materials is important in
a variety of different industries, such as pharmaceuticals, foodstuffs,
ceramics, detergents, chemicals, plastics, etc. The ubiquity of granular
materials in the processes involved makes how to handle them efciently very important economically. Furthermore, their special properties (they can behave like solids, liquids or gasses without being any
of them) have attracted the attention of the scientic community
over the last two decades [1,2]. Granular solid material can be made
to ow like a uid, for example, in avalanches, hopper ows, and uidized beds [3,4]. In such ows, the, granular kinetic energy may be
transformed into thermal energy due to inelastic collisions between
particles. Actually, the particles will stop moving if no extra energy
is added to the system. The most common sources of input energy
are from gravity, rotating the system, producing shear force at the
boundary of the system, or vibrating the system. The rotating drum
has become a commonly used experimental device used to investigate the physics of granular ow, partly because of its simple closed
geometry. It is also in practical use in many industrial processes for
the drying, segregation and mixing of granular materials [57].
When a horizontal circular drum partially lled with granular material is rotated about its axis, the material rotates as a solid body until
it reaches its dynamic angle of repose. The angle of repose is an

Corresponding author. Tel.: + 886 3 426 7341; fax: + 886 3 425 4501.
E-mail address: sshsiau@cc.ncu.edu.tw (S.S. Hsiau).
0032-5910/$ see front matter 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.powtec.2011.09.010

important parameter affecting the behavior of the granular ow in a

rotating drum. Liu et al. [8] found that the upper and lower angles
of repose could be inuenced by the size of the drum, the fraction
lled with particles, the speed of the rotation and the roughness of
the particles. In the rolling regime, when the rotation speed is low
enough, two distinct ow regions can be observed in the drum (see
Fig. 1): a bulk solid-body rotation undergoing slow plastic deformation [9], and a thin owing layer. The physical mechanisms mainly
occur in the owing layer.
There are a variety of industrial processes (wet granulation, granular condensation, drying, cake ltration, etc.) which involve wet
granular matter [1012]. However, in the past few decades, the behavior of wet granular matter has received less attention than that
of dry granular systems. There are dramatic changes observed in the
behavior of the granular ow when some amount of liquid is added
to the material, due to the formation of liquid bridges between the
particles which in turn cause an increase in interparticle cohesion
[1316]. Nase et al. [17] dened the Granular Bond Number Bog as
the ratio between the maximum capillary force and the weight of
the particle. Li and McCarthy [18] strove to extend this dimensionless
group for nonuniform particle properties by rewriting the Granular
Bond Number (Bog) for the binary-mixture system. They found that
the cohesive effect was important when the Granular Bond Number
is greater than 1. Kohonen et al. [19] showed that the number of liquid bridges increases as the liquid content increases, reaching a stable
value when the liquid content becomes greater than a critical value.
Samadani and Kudrolli [20] showed that segregation could be mitigated by adding a little liquid to a silo. They argued that in a wet

492

S.H. Chou, S.S. Hsiau / Powder Technology 214 (2011) 491499

Table 1
Experimental parameters: the particles are glass beads (b = 2476 kg/m3) with diameters of d = 4 mm.

Filling degree (f):0.4

Flowing Layer

Operating parameters

Interface
Fixed Bed

L =0.015 m

Fig. 1. Diagram of the tank rotated by the rotating drum showing the ow inside the
drum which can be divided in two parts: a bulk solid-body rotation undergoing a
slow plastic deformation; and a thin owing layer.

granular system the number of liquid bridges inuences the ow behavior. Yang and Hsiau [21,22] found that the self-diffusion coefcients and the uctuation velocities were smaller in a wet sheared
granular ow than in a dry sheared granular ow. Chou et al. [23]
showed that the segregation condition could be predicted based on
the angle of repose of the granular materials, regardless of the liquid
content or viscosity, in the case of wet granular ows.
It is well known that the liquid bridge force has a signicant inuence on the static properties of wet granular matter. In this study, we
employed a particle tracking method to measure the dynamic properties of wet granular matter. We studied the effect of the liquid bridge
force on the dynamic properties of wet granular matter by adding silicone oils to obtain different liquid contents and utilizing different rotation speeds.
2. Experimental procedure
A schematic representation of the circular drum used in the quasitwo-dimensional experiments is shown in Fig. 1. The diameter of the
drum is 0.3 m and the axial length L is 0.015 m. The back surface of
the drum was constructed of a black anodized aluminum plate to
minimize electrostatic effects on the particles and optical noise effects
in the digital images. The front faceplate was made of clear acrylic to
permit optical access. Before each experimental test, the glass faceplates were cleaned and coated with an inorganic lm (water repellent silicone) to prevent the formation of liquid bridges between the
particles and the walls. Mellmann [24] indicated that the rolling regime exists when 10 4 b Fr b 10 2. In this study, we investigated the
dynamic properties of wet granular matter in the rolling regime. A
stepper motor and micro series driver combination were used to rotate the drum at speeds of 1 rpm, 2 rpm, 3 rpm, and 4 rpm, corresponding to the Froude number of Fr = R 2 / g of 1.67 10 4,
6.71 10 4, 1.51 10 3, and 2.68 10 3, where is the angular velocity of the drum (=2 / T, T = rotation period); R is the radius of the
drum; and g is the acceleration of gravity. The dimensionless axial
thickness of the drum, dened as the ratio between the drum's axial
length and the particle diameter, was set to 3.75 in this study.
Mono-sized glass beads were used as the granular material in all experiments. These glass beads were 4 mm in diameter with a standard
deviation of 0.09 mm and their density was 2.476 g/cm 3. Details of
the experimental conditions are provided in Table 1.
In this study, the lling degree is dened as the ratio of volume occupied by the granular material to the total drum volume. The experiment is dened as a quasi-two-dimensional experiment. The lling
degree is dened as the ratio of the area occupied by the granular material to the total drum area, formulated as

Aparticles

sin

2
2
Adrum

Drum speed ():0.1047 rad/s, 0.2093 rad/s,

0.3140 rad/s 0.4187 rad/s
Liquid content (V):0, 1.996 10 3, 9.9 10 3,
14.778 10 3 19.6 10 3

where is the segmental angle formed between the area occupied by

the particles and the center of a circle.
The surface tension and density of the silicone oil added to
the granular materials in this study were 19.7 mN/m and
0.953 g/cm 3, respectively. Prior to the experiments, specic amounts
of silicone oil and glass beads were placed into a sealed jar, which
was then shaken sufciently to mix the silicon oil with the beads.
The wet glass beads were then poured carefully into the drum. The
weight of the sealed jar with the residual silicone oil was measured
to regulate the error. The dimensionless liquid volume V can be
given by V = Vl / (Vl + Vs), where Vl is the volume of liquid; and Vs
is the total volume of particles [22,23].
3. Image processing
The granular ow in the test section is assumed to be quasi-twodimensional, in the streamwise direction along the x-axis and in the
transverse direction along the y-axis. A high-speed CMOS camera
(IDT X-3 plus, monochrome, capable of shooting 5002000 frames per
second (fps), with a resolution 1280 1024 pixels), was used to record
the sequential motion of the granular ow during the experiments.
The ow was illuminated by four halogen lamps: DEDO DLH650 650W
3200K.
Fig. 2(a) shows a portion of a raw grayscale image captured during
the experiments. The raw images were rst ltered with a Gaussian
lter to diminish any high-frequency noise that might arise due to
variations in light intensity. A Laplacian lter was then applied to enhance the image contrast which transformed the bright pixels into
separate white patches as seen in Fig. 2(b). The light source caused
bright spots to appear on the slightly reective surfaces of the glass
beads, which were employed to locate the individual spheres. The
particle centers were identied by using a maximum brightness lter
as can be seen in Fig. 2(c). The velocity eld of the granular materials
in the selected region is shown in Fig. 2(d). The velocity vectors between two image pairs were calculated with the autocorrelation technique. It is found that the velocity is higher at the free surface of the
owing layer than between the owing layer and the xed bed.
The autocorrelation technique was employed to process the
stored images and to determine the shift of each tracer particle between two consecutive images [25,26]. A set of consecutive motion
images is shown in Fig. 3. The small window (n1x n1y pixels) in the
rst image outlines an identied tracer. The larger window
(n2x n2y pixels) in the second image outlines the area of possible locations of the tracer particle. In the rst window we can see a shift in
both directions (streamwise and transverse) by si and sj pixels each
time. This, multiplied by the corresponding pixel values for the second window, can be used to obtain the autocorrelation value c(si,
sj)


 n1x n1y



c si ; sj P1 i; jP2 i si ; j sj ;

i1 j1

where i and j indicate the pixel coordinates in the images; and P1 and
P2 are the pixel values in the two image windows. Natarajan et al.
[26] and Hsiau and Jamg [25] used shifts of si and sj when the maximum autocorrelation value c occurred for the movement of the

S.H. Chou, S.S. Hsiau / Powder Technology 214 (2011) 491499

493

Fig. 2. Image manipulation process: (a) the raw grayscale image; (b) the image after Gaussian and Laplacian ltering; (c) the obtained sphere locators; (d) the velocity eld.

tracer particle. In order to improve the accuracy, the correction of the

gray level derivatives in the x and y directions is also calculated as
follows:
(
)




 n1x n1y




;
cx si ; sj P1 i 1; jP1 i; j P2 i 1 si ; j sj P2 i si ; j sj
i1 j1

The ensemble average velocities in the streamwise and transverse

directions buN and bvN are dened as follows:
N

ui
buN

i1

and

vi

 n1x n1y n
h




io
cy si ; sj P1 i 1; jP1 i; j P2 i si ; j 1 sj P2 i si ; j sj
:

bvN

i1

i1 j1

4
The total correlation value R(si, sj) is decided by multiplying the sum
of these three correlation values c(si, sj), cx(si, sj) and cy(si, sj) by the
corresponding weighting factors k, kx and ky












R si ; sj kc si ; sj kx cx si ; sj ky cy si ; sj

where i represents the ith tracer particle; N is the total number of velocities used for averaging the mean values; and ui and vi are the velocities of the ith tracer particle measured from the pair of
consecutive images containing the ith tracer particle. The uctuation
velocities in the two directions are calculated by

The tracer displacements in two consecutive images are then determined from si and sj given the maximum value of R.

2 1=2

bu N

v
uN
u
u ui b u N2
t
i1
;
N

494

S.H. Chou, S.S. Hsiau / Powder Technology 214 (2011) 491499

y
j

n1x

n1y

n1 x

n1 y

Fig. 3. Images taken before and after a time delay.

i1

The granular temperature T is dened as the specic uctuation

kinetic energy of the granular ow and can be used to quantify the kinetic uctuation energy of the granular ow. This is a key property for
studying the dynamic behavior of granular ows [2729]. A granular
system behaves more like a liquid or a gas when it has a relatively
higher granular temperature. The granular temperature in a quasitwo-dimensional system can be calculated by
2

b u v N
:
2

10

4. Results and discussion

The experiment results are presented and discussed in this section.
The two operational parameters of this system, namely the amount of
liquid and the rotation speed, both have an impact on the dynamic processes; the other system parameters remain xed; see Table 1.
In order to investigate the effects of the liquid bridges on the dynamic properties of wet granular matter, some important forces
need to be introduced. The total dynamic liquid bridge force can be
obtained by looking at the superposition of the capillary force Fc and
the viscous force Fv. Based on Reynolds lubrication theory, an expression for the viscous force can be written as follows [20,30,31]:
Fv



3
h 2 1 dh
2
d 1
;
8
H b h dt

11

where b is the radius of the wetted area of the particle; d is the diameter
of a bead; h is the separation distance of the liquid bridge; is the liquid
dynamic viscosity; and dh/dt is the rate of approach of particles toward
each other. Note that in this study, the viscous forces in different wet systems are almost the same, because the viscosity of the uid remains constant. The capillary force Fc can be written as follows [30,32]:




4Vb 1=2
;
Fc d cos 1 1
2
dh

12

where is the surface tension; Vb =d[H2(b)h2 /4] is the liquid bridge

volume; and is the contact angle between the particle and liquid.

The angle of repose is an important parameter for understanding

the physical mechanism of granular materials. Several studies have
been carried out to explain the segregation phenomenon by investigating the angle of repose between the particles [33,34]. The dynamic
angle of repose m is the angle between the horizontal and the free
surface of a granular pile after a land slide has restored the pile to a
metastable equilibrium slope, as shown in Fig. 1. We rst want to understand the relationship between the liquid content and the dynamic
angle of repose. In this study, we measured the dynamic angle of repose in the rotating drum after the ow in the drum reached a steady
state by superimposing 15 images, then drawing a line along the topmost particles. Fig. 4 displays results showing the relationship between the dynamic angle of repose and liquid content in monodispersed wet ( = 19 cP) and dry granular systems. Results show
that the dynamic angle of repose rst decreases, but then increases
as the liquid content continues to increase. It is also found that the dynamic angle of repose increases with increased rotation speed, and the
inuence of the liquid content becomes greater when the dynamic
angle of repose is greater at a larger rotation speed.

= 0.1047 rad/s
= 0.2093 rad/s
= 0.3140 rad/s
= 0.4187 rad/s

42

40

r (deg.)

2 1=2

bv N

v
uN
u
u vi b v N2
t

38

36

34

32

10

15

20

V*(10-3)
Fig. 4. Relationship between the dynamic repose angle and different liquid contents for
= 19 cP for both with mono-dispersed systems.

S.H. Chou, S.S. Hsiau / Powder Technology 214 (2011) 491499

This phenomenon is explained below. Xu et al. [35] found that a

suitable parameter (Ra) could be determined by considering the
time scale tx required for two particles in neighboring streamlines to
travel past each other and the time scale ty for the same two particles
to approach each other while squeezing the liquid from in between
them. Thus,
Ra

tx
"
ty

0:13=
#
2b2 dhs b
1 2b2 =hs d

ln
1 2b=d
2b2 hs d 2b2 bd

13

where hs is the surface roughness of the particles. When Ra 1 the

two particles pass each other without coming into contact; when
Ra 1 the kinetic friction and viscous forces between the particles
dominate the overall owing behavior. In this study, we found that
for a small liquid content, such as when the particles rst become uniformly coated with the liquid and liquid bridges are continuously
formed between the particles, then b~ hs [36]. Within this limit, both
terms in the denominator of Eq. (15) tend to be 0 and therefore

(a)

5 mm

495

Ra 1. Since particles touch as they ow past each other, the kinetic

friction forces and viscous forces become important. We use a visualization method to examine this phenomenon. After the addition of
silicone oil (V = 1.996 10 3) to the dry glass beads, the mixture
is shaken, and then a small quantity of the wet glass beads is removed. The uorescent intensity of the silicone oil uid on the surface
of the spheres, is measured rst, as shown in Fig. 5(a). When only
very small amounts of liquid were added, no liquid bridges formed.
This is because the liquid was rst trapped on the surface of the
beads due to the particle roughness [3739]. When enough liquid
was added, V = 9.9 10 3, we found that these bridges formed between almost all the particles. Fig. 5(b) illustrates the case for fully
developed bridges. According to the explanation above, when the
amount of liquid added to the particles is insufcient to form liquid
bridges but instead is trapped by the surface roughness, the frictional
force between the particles will be smaller than the gliding force
resulting from gravity. This causes an initial decrease in the angle of
repose. When the amount of liquid added becomes large enough,
the number of liquid bridges in the system will increase with the liquid content, causing an increase in the liquid bridge force between
the particles. The angle of repose increases when the liquid bridge
force is greater than the gliding force resulting from gravity. This continues until the attractive force and the gliding force are in equilibrium. In addition, the inertial force increases as the rotation speed
increases, and the force used to resist the gliding force resulting
from gravity also increases. Therefore, we found that the angle of repose increases with increasing rotation speed.
To obtain the velocity prole in the owing layer look at Fig. 6,
which shows a typical ow eld measured in a specic region (dotted
rectangle in Fig. 1). The velocities of the particles in a narrow strip
about 19 particles wide in the streamwise x-direction and extending
into the xed bed in the transverse y-direction near the center of
the rotating drum are collected from all 2166 image pairs. Fig. 7(a)
shows the proles of the velocities in the streamwise and transverse
directions after the addition of silicone oil to obtain different liquid
contents, but at a xed rotation speed of 0.4187 rad/s. In this study,
two distinct ow regions can be identied: the owing layer and
the xed bed. In the owing layer, the streamwise particle velocity
decreases with the depth from the free surface, and with increasing
liquid content. In the xed bed the streamwise particle velocities observed at different radial positions are almost equal to the tangential
velocity of the boundary of the rotation drum. The ow in the xed-

(b)
-0.01

velocity scale
1 m/s

y-axis

0.01
0.02
0.03
0.04
0.05

5 mm
Fig. 5. Fluorescence images for different liquid contents, with 4 mm diameter glass
beads: (a) V = 1.996 10 3: no bridges have yet formed; (b) V = 9.9 10 3: liquid
bridges between the beads have formed.

-0.02

0.02

x-axis
Fig. 6. Typical ow eld obtained for a dry system in a region about 19 particles wide in
the streamwise x-direction and extending into the xed bed in the transverse y-direction
near the center of the rotating drum shown in Fig. 1.

496

S.H. Chou, S.S. Hsiau / Powder Technology 214 (2011) 491499

(a)

0.01

Y (m)

0.02

<u>
air
<u> V* = 1.996 x 10 -3
*
<u> V =
9.9 x 10 -3
<u> V* = 14.778 x 10 -3
<u> V* =
19.6 x 10 -3
<v>
air
<v> V* = 1.996 x 10 -3
<v> V* =
9.9 x 10 -3
<v> V* = 14.778 x 10 -3
<v> V* =
19.6 x 10 -3

0.03

0.04

0.05

0.06

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Velocity (m/s)

(b)

0.01

Y (m)

0.02

0.03

<u' 2 >1/2
air
<u' 2 >1/2 V* = 1.996 x 10 -3
2 1/2
*
-3
<u' > V =
9.9 x 10
<u' 2 >1/2 V* = 14.778 x 10 -3
<u' 2 >1/2 V* =
19.6 x 10 -3
<v' 2 >1/2
air
<v' 2 >1/2 V* = 1.996 x 10 -3
<v' 2 >1/2 V* =
9.9 x 10 -3
<v' 2 >1/2 V* = 14.778 x 10 -3
<v' 2 >1/2 V* =
19.6 x 10 -3

0.04

0.05

0.06

0.1

0.2

0.3

Fluctuation Velocity (m/s)

(c)

0.01

Y (m)

0.02

0.03

0.04

air
V* = 1.996 x 10 -3
*
V =
9.9 x 10 -3
*
-3
V = 14.778 x 10
*
-3
V =
19.6 x 10

0.05

0.06
0

0.01

0.02

0.03

Granular Temperature (m2/s2)

Fig. 7. Distribution of: (a) velocities in the streamwise and transverse directions;
(b) uctuation velocities; (c) granular temperature, with the depth from the free surface for different liquid contents.

bed is the so-called plug-ow. No matter how large the amount of liquid, the transverse velocities are close to 0, because there is no ux in
the y-direction at the center of the owing layer (particles neither
enter nor leave the owing layer).
When there is no formation of liquid bridges at V = 1.996 10 3,
liquid is trapped on the surface due to the roughness of the bead,
which would cause the lower dynamic angle of repose. The overall
potential energy of the particles in the rotating drum would be
lower than that of dry granular matter. This is why the streamwise
velocity decreases rst in a wet granular system. In Fig. 4, it can also
be seen that the dynamic angle of repose increases with the increase
of liquid content. However, the streamwise velocity decreases continuously with the increase of liquid content as shown in Fig. 7(a). The
reason is that when the amount of liquid is large enough, bridges
will form between almost every particle pair. A higher liquid content
leads to the formation of greater liquid bridge forces, resulting in the
smaller streamwise velocity. As can be seen in Fig. 7(a), the velocity
gradient is the largest in the dry case, because no liquid bridges
have formed between particles. The shear rate in the granular ow
is weaker in wet granular matter due to the dynamic liquid bridge
force [40,41]. The addition of a higher liquid content leads to the formation of greater liquid bridge forces in the system. The rupturing energy also increases, resulting in the smaller velocity gradient.
The components of the uctuation velocities in the streamwise
and transverse directions for different liquid contents but at a xed
speed of rotation of 0.4187 rad/s are shown in Fig. 7(b). The uctuation velocities in the streamwise direction are greater than those in
the transverse direction. The uctuation component is not isotropic
as assumed in the dense-gas kinetic theory [28]. An explanation for
this anisotropic phenomenon can be found in Hsiau and Shieh's
study [42]. In Fig. 7(b), it can be seen that the uctuation velocities
decrease with the depth from the free surface due to the decrease
of the shear rate (dbuN/dy). The uctuation velocity is the greatest
in the dry system where no liquid bridges form and particle motions
and interactive collisions are stronger. An examination of Fig. 7(b)
also shows that the uctuation velocity decreases with the increase
of liquid content. Accordingly, the lower angle of repose (lower overall potential energy) of particles in a rotating drum means that particle motions become weaker when no liquid bridges form. As the
amount of liquid become larger, bridges form between particles that
come into contact. Particle motions and interactive collisions will be
mitigated due to the greater liquid bridge forces.
Granular temperature is dened as the mean square of the uctuation velocity. This physically denotes the specic kinetic energy of a
granular system and plays a similar role as temperature in general
gasses. The change in granular temperature with the depth from the
free surface for the ve cases (different liquid contents but a xed rotation speed of 0.4187 rad/s) is shown in Fig. 7(c). It can be seen that
the granular temperature is the highest near the free surface, and it
gradually decreases along the y direction. The above discussion
makes it clear that the uctuation velocities of particles and interactive collisions are stronger in the owing layer and weaker in the
xed bed. This means that the granular temperature is higher in the
owing layer but lower in the xed bed. We also nd that the granular temperature decreases as the amount of the liquid added increases. Accordingly, when there is no formation of liquid bridges,
the lower overall potential energy of particles in a rotating drum
would be conducive to the weakening of particle motions, in turn
causing a decrease in the number of collisions. The addition of higher
liquid contents to granular materials leads to the formation of stronger liquid bridge forces, and decreases the number of interactive collisions. As a consequence, the granular temperature is the highest in
the dry system where no liquid bridge force exists. The potential energy is higher than V = 1.996 10 3, and is smallest for the wet system with the addition of V = 19.6 10 3 of silicone oil. This is the
liquid content that would lead to the largest liquid bridge force and

S.H. Chou, S.S. Hsiau / Powder Technology 214 (2011) 491499

(a)

0.01

Y (m)

0.02

0.03
<u>
<u>
<u>
<u>
<v>
<v>
<v>
<v>

0.04

0.05

0.06

0.1

0.2

= 0.1047 rad/s
= 0.2093 rad/s
= 0.3410 rad/s
= 0.4187 rad/s
= 0.1047 rad/s
= 0.2093 rad/s
= 0.3410 rad/s
= 0.4187 rad/s

0.3

0.4

Velocity (m/s)

(b)

0.01

Y (m)

0.02

0.03
<u' >
2 1/2
<u' >
<u' 2 >1/2
<u' 2 >1/2
2 1/2
<v' >
2 1/2
<v' >
<v' 2 >1/2
<v' 2 >1/2
2

0.04

0.05

0.06

0.05

1/2

= 0.1047 rad/s
= 0.2093 rad/s
= 0.3410 rad/s
= 0.4187 rad/s
= 0.1047 rad/s
= 0.2093 rad/s
= 0.3410 rad/s
= 0.4187 rad/s

0.1

0.15

FluctuationVelocity(m/s)

(c)

0.01

0.02

Y (m)

the weakest particle motions. From Fig. 7(a)(c) it can be seen that
the liquid content has a signicant impact and an important inuence
on the dynamic properties of the granular ow system.
The introduction of greater external force to a granular matter system causes stronger particle motion, but the cohesive force between
the particles may result in additional energy dissipation. Therefore, we
discuss the effect of the liquid bridge force on the dynamic properties
of wet granular matter at four different rotation speeds. Fig. 8(a)
shows the granular velocities in the streamwise and transverse directions, with four different rotation speeds, but using the same addition
of silicone oil (V = 9.9 10 3). As expected, in each case, the streamwise velocity gradually decreases with the depth from the free surface;
the transverse velocity is close to 0. The trend is similar to that mentioned above for Fig. 8(a). It is also demonstrated that the velocity gradient increases in the streamwise direction as the rotation speed
increases. A higher rotation speed introduces more energy into the
granular system and generates stronger interactions between particles,
resulting in a higher velocity gradient. This leads to an increase in the
thickness of the owing layer, as observed in Fig. 8(a).
The components of the uctuation velocities in both directions are
shown in Fig. 8(b). The shear rate is higher in the owing layer, due to
the stronger interactive collisions and uctuations. Thus we found
that the uctuation velocity is greater in the owing layer. It is also
shown that the uctuation velocity increases with the increase of rotation speed which in turn is conducive to the particle motions becoming more random and stronger in the high shear region where
the velocity uctuations are also greater.
As can be seen in Fig. 8(c) the granular temperature decreases
with the depth from the free surface. The particle motions and uctuation velocities become weaker with the depth from the free surface
due to smaller shear rate, which in turn generates a smaller granular
temperature. However, we found that the granular temperature increases as the rotation speed increases. This is because more energy
is introduced into the granular system, generating a higher shear
rate. The increased rotation speed causes stronger particle motion
and more interactive collisions due to the higher shear rate, which
leads to the higher granular temperature, as shown in Fig. 8(c).
The physical mechanisms occur mostly in the owing layer in the
rotating drum. Understanding the effect of adding a small amount of liquid to the granular materials in the granular owing layer is important.
Therefore, the owing layer was subdivided into 10 bins to determine
the owing thickness. At the interface between the owing layer and
the xed bed, the streamwise velocity had a zero or slightly negative
value, while the values were positive for all other bins. The thickness
was measured based on when the value of the streamwise velocity becomes zero or has a negative value [43]. Fig. 9(a) shows the owing
layer thickness plotted against the liquid content V. It can be seen
that the thickness of the owing layer in wet granular matter increases
with increased liquid content. The particles start to roll downward
along the owing layer when they obtain enough potential energy. Liquid bridge forces exist between wet particles. When the liquid bridge
force is large enough, the rolling particles may bring other particles,
which would not otherwise have enough potential energy to roll,
downward. We also found that the thickness of the owing layer rst
decreases when the dry granular matter becomes wet. According to
the above, no liquid bridges form at V = 1.996 10 3. Although at
V = 1.996 10 3, the force of lubrication is produced, the potential energy introduced into the wet granular system would also be reduced, so
they there is not enough gliding force to balance the drag force between
the particles. The relationship between the owing layer thickness and
rotation speed is shown in Fig. 9(b). It can be seen that the owing layer
thickness increases with increasing rotation speed. The reason is that
the potential energy of the particles increases when the rotation
speed increases, strengthening the gliding force resulting from gravity
which is conducive to more uidized particle motion. This result has
been observed in several studies [44,45].

497

0.03

0.04

0.05

0.06

0.005

= 0.1047
= 0.2093
= 0.3410
= 0.4187

0.01

rad/s
rad/s
rad/s
rad/s

0.015

Granular Temperature (m2/s2)

Fig. 8. Distribution of: (a) velocities in the streamwise and transverse directions;
(b) uctuation velocities; (c) granular temperature, with the depth from the free surface for different rotation speeds.

498

S.H. Chou, S.S. Hsiau / Powder Technology 214 (2011) 491499

(a)

0.016
0.04

a ir
*
V =
*
V =
*
V =
*
V =

Average Granular Temperature (m2/s2)

0.014

Flowing thickness (m)

0.035

0.03

0.025

= 0.1047
= 0.2093
= 0.3140
= 0.4187

rad/s
rad/s
rad/s
rad/s

0.012
0.01
0.008
0.006
0.004
0.002

0.1
0.02

10

15

Flowing layer thickness (m)

0.035

0.03

air
V* = 1.996 x 10 -3
V* =
9.9 x 10 -3
*
-3
V = 14.778 x 10
*
-3
V =
19.6 x 10

0.025

0.1

0.2

0.3

0.4

0.5

Fig. 10. Relationship between the average granular temperature and rotation speed.

(b)0.04

0.02

0.2

Rotation speed (rad/s)

20

V*(10-3)

-3

1 .9 9 6 x 1 0
-3
9 .9 x 1 0
-3
1 4 .7 7 8 x 1 0
-3
1 9 .6 x 1 0

0.3

0.4

Rotation speed (rad/s)

Fig. 9. (a) Relationship between the owing layer and the different liquid contents;
(b) relationship between the owing layer and the different rotation speeds.

The average granular temperature, averaged from the granular

temperature taken from the chosen region for all 20 tests, is plotted
as a function of the rotation speed in Fig. 10. The capillary interaction
of wet granular matter has a well-dened binding energy [41]. It has
been demonstrated experimentally under realistic dynamical conditions, with impact velocities typical for strongly uidized wet granular matter, that the hysteretic character of the interaction is essential.
The dominant mechanism for dissipation is the hysteretic formation
and rupture of capillary bridges, the binding energy of which is irreversibly taken from the kinetic energy of the granular motion whenever a liquid bridge ruptures [41]. The liquid bridge energy has been
quantied [39,41,46,47], and is proportional to the radius of the
beads, the surface tension of the liquid, and the volume fraction of
the added liquid with respect to the total volume of all glass spheres.
It can be seen in Fig. 10 that, in the wet granular system, the average
granular temperature decreases with increased liquid content. This is
also consistent with the physical explanation. We found that liquid
bridges formed between almost all the particles, and the energy dissipation due to the hysteretic formation and rupture of capillary bridges was
proportional to the liquid content. Although no liquid bridges formed at
V = 1.996 10 3, the particle motions and interactive collisions still

became weaker due to the lower angle of repose (lower potential energy) of particles in the rotating drum. The particle motions and interactive collisions were mitigated due to the greater liquid bridge force
resulting from the addition of more liquid [21,48]. The granular temperature is indicative of the energy of the kinetic uctuation of the particles
due to the interactive collisions between them. More energy is introduced into the granular system with the higher rotation speed, resulting
in the stronger particle motions and interactive collisions. Therefore,
the average granular temperature increases with the increase of the rotation speed, as shown in Fig. 10. It can also be seen from this gure that
the average granular temperature increases linearly with increased rotation speed but the same liquid content. The slopes are 3.73 10 3 for
a dry system, 1.36 10 3 for V = 1.996 10 3, 1.19 10 3 for
V = 9.9 10 3, 7.88 10 4 for V = 14.778 10 3 and 7.70 10 4
for V = 19.6 10 3, respectively. This could be explained as follows:
liquid bridges form when two particles come into contact, while rupturing of these bridges occurs when two particles separate after exceeding
a critical distance. At larger rotation speeds, the owing layer is thicker,
so that the frequency and number of interactive collisions between particles becomes greater. Therefore the frequency of hysteretic formation
and rupturing of liquid bridges is also higher, resulting in more energy
being dissipated into the surroundings.

5. Conclusions
In this study we performed experiments to examine the effect of
the liquid content on the dynamic properties of wet granular materials in a quasi two-dimensional rotating drum. Using image processing technology and the particle tracking method, we measured the
dynamic angle of repose, average velocities, uctuation velocities,
granular temperature and owing layer thickness. We demonstrated
that the liquid content has a signicant inuence on the dynamic
properties of wet granular matter.
The dynamic angle of repose is the angle between the horizontal and
the free surface of a granular pile, after a land-slide has restored the pile
to metastable equilibrium. The results show that when the amount of
liquid added to the particles is not enough to allow the formation of liquid bridges and the liquid is trapped by the rough surface, the force of
friction between the particles is smaller than the gliding force resulting
from gravity. This causes a decrease in the angle of repose. When the
amount of liquid added becomes large enough, liquid bridges form at almost every particle contact, increasing the liquid bridge force between
the particles. The angle of repose increases continuously.

S.H. Chou, S.S. Hsiau / Powder Technology 214 (2011) 491499

The dynamic behaviors of granular matter decreased with an increase

in the liquid content and increased with an increase in the rotation speed.
The average energy dissipation due to the hysteretic formation and rupture of capillary bridges increased with the increase in liquid content
when liquid bridges formed between all particles. The thicker owing
layer caused an increase in the overall energy dissipation with the increased rotation speed.
Acknowledgment
The authors would like to acknowledge the support from the National
Science Council of the ROC for this work through grants NSC 98-3114-E008-004.
References
[1] H.J. Herrmann, Physics of granular media, Chaos solitons & fractals 6 (1995)
203212.
[2] H.M. Jaeger, S.R. Nagel, Physics of the granular state, Science 255 (1992) 1523.
[3] A. Datta, B.K. Mishra, S.P. Das, A. Sahu, A DEM analysis of ow characteristics of noncohesive particles in hopper, Materials and Manufacturing Processes 23 (2008)
196203.
[4] M. Shearer, J.M.N.T. Gray, A.R. Thornton, Stable solutions of a scalar conservation
law for particle-size segregation in dense granular avalanches, European Journal
of Applied Mathematics 19 (2008) 6186.
[5] J.M. Ottino, D.V. Khakhar, Mixing and segregation of granular materials, Annual
Reviews Fluid Mechanics 32 (2000) 5591.
[6] J.M. Ottino, R.M. Lueptow, Materials Science: on mixing and demixing, Science
319 (2008) 912913.
[7] G.H. Ristow, Pattern Formation in Granular Materials, Springer, Belin, 2000.
[8] X.Y. Liu, E. Specht, J. Mellmann, Experimental study of the lower and upper angles
of repose of granular materials in rotating drums, Powder Technology 154 (2005)
125131.
[9] T.S. Komatsu, S. Inagaki, N. Nakagawa, S. Nasuno, Creep motion in a granular pile
exhibiting steady state surface ows, Physical Review Letters 86 (2001) 17571760.
[10] C. Bacher, P.M. Olsen, P. Bertelsen, J.M. Sonnergaard, Compressibility and compactibility of granules produced by wet and dry granulation, International Journal
of Pharmaceutics 358 (2008) 6974.
[11] D. Geromichalos, M.M. Kohonen, F. Mugele, S. Herminghaus, Mixing and condensation
in a wet granular medium, Physical Review Letters 90 (2003) 168702.
[12] W.K. Ng, R.B.H. Tan, Case study: optimization of an industrial uidized bed drying
process for large Geldart Type D nylon particles, Powder Technology 180 (2008)
289295.
[13] P. Tegzes, T. Vicsek, P. Schiffer, Development of correlations in the dynamics of
wet granular avalanches, Physical Review E 67 (2003) 051303.
[14] P. Tegzes, T. Vicsek, P. Schiffer, Avalanche dynamics in wet granular materials,
Physical Review Letters 89 (2002) 094301.
[15] S. Nowak, A. Samadani, A. Kudrolli, Maximum angle of stability of a wet granular
pile, Nature Physics 1 (2005) 5052.
[16] R. Brewster, G.S. Grest, J.W. Landry, A.J. Levine, Plug ow and the breakdown of
Bagnold scaling in cohesive granular ows, Physics Review E 72 (2005) 061301.
[17] S.T. Nase, W.L. Vargas, A.A. Abatan, J.J. McCarthy, Discrete characterization tools
for cohesive granular material, Powder Technology 116 (2001) 214223.
[18] H.M. Li, J.J. McCarthy, Cohesive particle mixing and segregation under shear, Powder
Technology 164 (2006) 5864.
[19] M.M. Kohonen, D. Geromichalos, M. Scheel, C. Schierb, S. Herminghaus, On capillary
bridges in wet granular materials, Physica A 339 (2004) 715.
[20] A. Samadani, A. Kudrolli, Segregation transitions in wet granular matter, Physical
Review Letters 85 (2000) 51025105.

499

[21] W.L. Yang, S.S. Hsiau, Wet granular materials in sheared ows, Chemical Engineering
Science 60 (2005) 42654274.
[22] W.L. Yang, S.S. Hsiau, The effect of liquid viscosity on sheared granular ows,
Chemical Engineering Science 61 (2006) 60856905.
[23] S.H. Chou, C.C. Liao, S.S. Hsiau, An experimental study on the effect of liquid content
and viscosity on particle segregation in a rotating drum, Powder Technology 201
(2010) 266272.
[24] J. Mellmann, The transverse motion of solids in rotating cylinders forms of motion
and transition behavior, Powder Technology 118 (2001) 251270.
[25] S.S. Hsiau, H.W. Jang, Measurements of velocity uctuations of granular materials
in a shear cell, Experimental Thermal and Fluid Science 17 (1998) 202209.
[26] V.V.R. Natarajan, M.L. Hunt, E.D. Taylor, Local measurements of velocity uctuations
and diffusion coefcients for a granular material ow, Journal of Fluid Mechanics
304 (1995) 125.
[27] G. Bunin, Y. Shokef, D. Levine, Frequency-dependent uctuationdissipation relations
in granular gases, Physical Review E 77 (2008) 051301.
[28] C.S. Campbell, Rapid granular ows, Annual Reviews Fluid Mechanics 22 (1990)
5792.
[29] S.S. Hsiau, W.L. Yang, Stresses and transport phenomena in sheared granular
ows with different wall conditions, Physics of Fluids 14 (2002) 612621.
[30] O. Pitois, P. Moucheront, X. Chateau, Liquid bridge between two moving spheres:
an experimental study of viscosity effects, Journal of Colloid and Interface Science
231 (2000) 2631.
[31] C.C. Liao, S.S. Hsiau, T.H. Tsai, C.H. Tai, Segregation to mixing in wet granular matter
under vibration, Chemical Engineering Science 65 (2010) 11091116.
[32] O. Pitois, P. Moucheront, X. Chateau, Rupture energy of a pendular liquid bridge,
European Physical Journal B: Condensed Matter and Complex Systems 23 (2001)
7986.
[33] K.M. Hill, J. Kakalios, Reversible axial segregation of rotating granular media,
Physical Review E 52 (1995) 43934400.
[34] N.A. Pohlman, B.L. Severson, J.M. Ottino, R.M. Lueptow, Surface roughness effects
in granular matter: inuence on angle of repose and the absence of segregation,
Physical Review E. 73 (2006) 031304.
[35] Q. Xu, A.V. Orpe, A. Kudrolli, Lubrication effects on the ow of wet granular materials,
Physical Review E 76 (2007) 031302.
[36] T.G. Mason, A.J. Levine, D. Ertas, T.C. Halsey, Critical angle of wet sandpiles, Phyical
Review E 60 (1999) R5044R5047.
[37] D.J. Hornbaker, R. Albert, I. Albert, A.L. Barabasi, P. Schiffer, What keeps sandcastles
standing, Nature 387 (1997) 765-765.
[38] T.C. Halsey, A.J. Levine, How sandcastle fall, Physical Review Letters 80 (1998)
31413144.
[39] Ch.D. Willett, M.J. Adams, S.A. Johnson, J.P.K. Seville, Capillary bridges between
two spherical bodies, Langmuir 16 (2000) 93969405.
[40] A. Fingerle, K. Roeller, K. Huang, S. Herminghaus, Phase transitions far from equilibrium
in wet granular matter, New Journal of Physics 10 (2008) 053020.
[41] S. Herminghaus, Dynamics of wet granular matter, Advances in Physics 54 (2005)
221261.
[42] S.S. Hsiau, Y.H. Shieh, Fluctuations and self-diffusion of sheared granular material
ows, Journal of Rheology 43 (1999) 10491066.
[43] N. Jain, J.M. Ottino, R.M. Lueptow, Effect of interstitial uid on a granular owing
layer, Journal of Fluid Mechanics 508 (2004) 2344.
[44] G. Felix, V. Falk, U. Ortona, Segregation of dry granular material in rotating
drum: experimental study of the owing zone thickness, Powder Technology
128 (2002) 314319.
[45] A.V. Orpe, D.V. Khakhar, Scaling relations for granular ow in quasi-twodimensional rotating cylinders, Physical Review E 64 (2001) 031302.
[46] G.P. Lian, C. Thornton, M.J. Adams, Discrete particle simulation of agglomerate impact
coalescence, Chemical Engineering Science 53 (1998) 33813391.
[47] S.J.R. Simons, R.J. Fairbrother, Direct observations of liquid binder-particle interactions:
the role of wetting behavior in agglomerate growth, Powder Technology 110 (2000)
4458.
[48] S.S. Hsiau, L.S. Lu, C.Y. Chou, W.L. Yang, Mixing of cohesive particles in a shear cell,
International Journal of Multiphase Flow 34 (2008) 352362.