You are on page 1of 15

Mixing and Segregation in Powders:

Evaluation, Mechanisms and Processes†

R. Hogg
Department of Energy and Mineral Engineering,
The Pennsylvania State University1

 Mixing in powders generally results from relative motion of groups of particles convective
mixing or of individuals diffusive mixing. Segregation or demixing occurs when the motion of
individual particles is biased according to their particular characteristics size, shape, composition
etc. In the absence of such bias, individual motion invariably leads to homogenization of the mixture.
Relationships between mixing/segregation processes and the external and interpar ticle forces
responsible for causing or opposing relative motion are reviewed. Specific examples of mixing and
segregation in flow over surfaces, in rotating cylinders and other applications are described.
Keywords: Homogeneity, macromixing, micromixing, diffusive mixing, convective mixing, demixing, role of agita-
tion and shear, percolation, mixing and segregation in flow systems

erations commonly involve movement of a layer of

vibrated particles over the screen surface. Segrega-
 Mixing of particulate solids is an important process tion of coarse (oversize) particles towards the top of
in its own right, as in blending of the components the layer can improve screening efficiency by reduc-
of a mixed powder; it can also play a critical role in ing blockage of the screen apertures. Agglomeration
the performance of other particle processing opera- processes take advantage of segregation of large ag-
tions such as grinding, granulation, classification and glomerates to promote preferential discharge of the
chemical treatment. In single-phase systems such as product from balling drums or pans.
gases or miscible liquids, mixing proceeds spontane-  It is important to distinguish between the results
ously and irreversibly so that actions such as stirring of active segregation processes and those due simply
are needed only to speed up the process. In contrast, to inadequate mixing. If some of the components of
solids mixing processes are neither spontaneous nor a mixture are added in agglomerated form, complete
irreversible some kind of mechanical agitation is mixing requires break-up and redistribution of this
required for mixing to occur and a tendency for dif- material. Low-shear devices such as tumblers may
ferent kinds of particles to segregate is commonly be incapable of performing this action. Indeed, they
observed. may actually enhance segregation by encouraging
 Segregation is typically described as a problem. agglomerate formation and growth.
While this is clearly appropriate for simple blend-  Another distinction can be made in the context of
ing operations, there are cases where it is actually mixing as an element of some other process such as
desirable by providing some degree of classification. a continuous grinding operation. In liquid-phase reac-
Segregation due to agitation or flow of a powder is tors, “perfect” mixing refers to uniform distribution
rarely an effective means of classification on its own of the components throughout the vessel ­ intimate
but it can be a useful adjunct to other classification mixing at the molecular level usually follows auto-
processes. For example, continuous screening op- matically. The latter is often not true for particle sys-
tems. Spatial uniformity may be achieved while the

components remain in the form of clumps. However,
Accepted: July 22nd, 2009
1 uniformity of this kind may be quite sufficient to re-
213 Hosler Building, University Park, PA 16802
  TEL: (814) 865-3802, FAX: (814) 865-3248 gard the reactor as fully mixed for estimating process
  E-mail: parameters such as residence time distributions.

ⓒ 2009 Hosokawa Powder Technology Foundation

KONA Powder and Particle Journal No.27 (2009) 3
 This paper is intended as an overview of the sub- homogeneity, i.e., at the individual particle level, and
ject rather than a comprehensive review of published long-range variations within a batch of material. The
work in the area. expected, randomly mixed variance described above
refers to micromixing. Since the observed variance
also includes the effects of errors in sample assay, it
Assessment of Mixture Homogeneity
is necessary increase the expected variance by using
 A generally accepted procedure for evaluating the small samples to ensure that incomplete micromixing
relative homogeneity of a particulate mixture is to is not masked by such errors,. At the same time, how-
take a number of samples, at random, and determine ever, the increase in the expected variance can mask
the variance of their composition: the more homoge- the effects of small, but significant long-range varia-
neous the mixture, the lower the expected variance. tions. For example, in a study of axial mixing in a ro-
Because of the discrete nature of the particles, the tating cylinder9) it was found, using samples of about
expected variance of a random mixture is ultimately 500 particles, that the observed sample variance be-
limited by the chance of different kinds of particles, came equal to the expected value after about 70,000
especially those of minor components, being includ- revolutions of the drum. At that stage, however, there
ed in a sample. Based on simple statistics, it has been remained a small but detectable and significant varia-
shown1) that the expected variance, σ2 of the relative tion in composition along the length of the cylinder.
composition of samples containing N particles each To detect both long-range and short-range variations,
taken from a completely random binar y mixture is it is necessar y to sample at more than one level
given by: small samples to identify local clustering of compo-
nents and larger samples to test for the long-range
p(1 − p)
σ2 = (1) variations. Typically, the latter can be minimized by
further mixing while the former may require alterna-
where p is the overall number fraction of one of the tive procedures such as increased shear. Procedures
components. The simple treatment has been extend- for characterizing particulate mixtures have been de-
ed to the more practical case of the expected variance scribed in greater detail in a recent publication by the
in the mass fraction of a given component in samples author10).
of equal mass taken from a multi-component mix-
ture2-6). The equivalent to Equation 1 then becomes
Mixing Mechanisms
fi (1 − fi ) wi + fi2 (w̄ − wi )
(σm )i =  Lacey11) proposed the following basic mechanisms
for solids mixing: convective mixing which involves
where (σm )i is the expected variance of the mass
the collective transfer of groups of particles from one
concentration of component i in samples taken from location to another; diffusive mixing defined as the
a random mixture of all of the components; f i is the distribution of particles over a freshly developed sur-
overall mass fraction of component i in the mixture; face and shear mixing due to slipping planes set up
wi is the mass of a single particle of that component within the mixture. Further consideration, however,
and w̄ is the overall mean particle mass, defined by: suggests that, for free-flowing powders, the funda-
 mental mechanisms are simply convection essentially
w̄ = fj wj (3) as defined by Lacey and dif fusion defined more
j broadly as the random motion of individual particles.
An observed sample variance that is greater than the Shear mixing can be regarded as a combination of
expected value as determined using Equations 1 or convection and diffusion resulting from the redistri-
2 implies that the mixture is heterogeneous (non- bution of material across slipping planes. Shear is an
random). It is possible, however, for such a result important feature of the mixing process by providing
to be obtained, by chance, for a random mixture. the driving force for convection and diffusion. A third
Comparison of the ratio of the observed to expected mechanism transfer between agglomerates is im-
variance using, for example, the simple F-test7) gives portant for cohesive particles and is discussed below.
an indication of whether or not the difference is sig-  Typically, diffusion is a rather slow process while
nificant statistically. convection proceeds rapidly. Practical mixing sys-
 Danckwer ts 8) used the terms micromixing and tems employ combinations of these two primar y
macromixing to distinguish between shor t-range mechanisms. Convection generally provides rapid

4 KONA Powder and Particle Journal No.27 (2009)

macromixing and substantially reduces the distances generally work in opposition in segregation.
over which diffusion must occur to ensure complete  Segregation is caused by differences in the basic
micromixing. Agitation and shear are necessary for particle characteristics which can be defined in terms
both mechanisms to occur. Convection is controlled of size, shape, composition and structure10). As far as
primarily by the geometr y of the system including segregation is concerned, structure the distribu-
the presence of elements such as paddles and baffles. tion of phases within a single particle is unlikely
that induce collective motion of the particles. This to cause differences in mobility and can probably be
process can lead to a kind of diffusive macromix- discounted. Particle size and shape affect mobility
ing, akin to turbulent diffusion in fluids. Diffusive largely through their effect on the resistance to mo-
micromixing is determined by the mobility of indi- tion whereas density (determined by particle compo-
vidual particles in response to agitation and shear. sition) is more likely to influence the actual motion
In free-flowing powders, mobility is generally high, resulting from applied forces.
promoting relatively rapid micromixing. In the case
of cohesive powders, on the other hand, individual
Size Segregation
mobility is low and sometimes virtually absent. Mix-
ing of these materials must rely heavily on shear to  Variations in individual particle mobility with size
break up the material into smaller and smaller ag- typically lead to segregation in a bed of particles sub-
glomerates. Since cohesive powders normally con- jected to vibration or flow. However, the form of the
tain very fine particles and agglomerate breakage is resulting segregation can vary widely depending on
a reversible process, liberation of individual particles the actual conditions. This has lead to a number of
is fleeting at best. Micromixing by simple diffusion proposed mechanisms which are actually different
is essentially eliminated and must occur by random manifestations of the same basic phenomenon: the
transfer of particles between breaking and re-forming ability of small particles to penetrate the bed more
agglomerates. readily than coarser material.
 Imposed vibration or particle-particle collisions in
a flow stream cause some dilation of the bed and in-
Segregation Mechanisms
creased particle mobility. Gravitational forces lead to
 Micromixing by diffusion in a free-flowing powder a net downward motion of the smaller particles which
requires that the motion of individual particles in can readily pass through the voids in the expanded
response to agitation etc. be truly random and inde- bed. This simple mechanism is especially important
pendent of the type of particle concerned. In other when the fine material is a minor component of the
words, all particles in the system should have the mixture and is commonly referred to as percolation
same mobility. When there are differences in mobil- or sieving12,13,14). Segregation can also occur, by the
ity among the components of a mixture, segregation same basic mechanism, regardless of the overall
becomes a possibility. Particle motion results from composition of the mixture. If the coarse particles
the effects of mechanical actions such as tumbling or are a minority each surrounded by the finer compo-
stirring, and external forces, primarily gravity. Par- nents their ability to move upwards is constrained
ticle mobility is determined by the response to these only by the weight of the overlying material while
forces and the resistance to motion due to other par- downward motion requires either compaction of the
ticles, their environment and the mixing device itself. material below or the unlikely possibility of encoun-
In many ways, segregation in powders is analogous tering a sufficiently large void. At the same time, the
to differential settling of particles in a fluid. void created by upward motion of the particle is eas-
 Convective mixing does not normally depend on ily filled by percolation of adjacent fines, which pre-
individual particle mobility; the mixing effect tends to vents a return to its original position15). The net result
be more or less random and is unlikely to contribute is a ratcheting upwards of the coarser particles. This
directly to segregation. At the same time, however, effect has been described as displacement segrega-
the actions that lead to convective mixing generally tion12,13) or the “Brazil nut effect”16).
promote individual motion which can cause segrega-  Size segregation is also widely obser ved when
tion. Because of restricted mobility, segregation is particles are poured onto a heap or in the filling of a
usually less prevalent in cohesive powders than in container. Again the basic mechanism involved is the
free-flowing material. Whereas convection and diffu- difference in the ability of small and large particles
sion are complementar y in mixing processes, they to penetrate a bed of material, although the result ap-

KONA Powder and Particle Journal No.27 (2009) 5

pears to be quite different. In contrast to a vibrated
(a) Smooth
Smooth Surface
or agitated bed, a heap is essentially static with flow 50
50 o
34 o
34 ,, 72.7
72.7 cm
occurring only in the surface layers. Particles flow

over the surface until they become trapped in a sur- 40 Total

40 Coarse
face void. Since the finer material is more likely to
encounter a sufficiently large void, the result is that

coarse particles travel further down the slope and ac- 20
cumulate around the lower edges of the heap leaving
excess fines in the core. This effect has often been 10

called rolling segregation17). 0

 The term trajectory segregation18) has been used 0

to describe separation of coarse and fine particles

projected from the end of a chute or fast-moving belt. 40

Retardation of the lateral motion by aerodynamic (b)

(b) Rough
Rough Surface
drag is greater for small than for large particles. Pro- 30
40 ,, 80.4
80.4 cm
jection of the particles individually would certainly Total

lead to segregation. In practice, however, material
is normally projected as a fairly dense, continuous 20

stream carr ying the adjacent air along with it. As
a result, aerodynamic drag acts on the stream as a 10
whole rather than on individual particles, thereby
minimizing the segregation effect. Obser vation of
segregation in such systems can probably be attrib- 0
uted to existing segregation, prior to projection. This Horizontal
Horizontal Displacement,
Displacement, cm
effect was demonstrated by Augenstein19,20) and is Fig. 1 Distribution of trajectories for mixtures of coarse (30×40 US
illustrated in Fig. 1. Mixtures of coarse and fine par- mesh) and fine (70×100 US mesh) particles projected from in-
ticles projected from chute surfaces were collected clined chute surfaces. Data of Augenstein 19).

in a sampling grid to determine the distributions of

horizontal displacement. For smooth chute surfaces, for example21,22). In part, this may be an impression
the range of displacements was found to be narrow, gained due to the difficulties involved in quantifying
indicating a narrow range of velocities, and the com- small dif ferences in shape. Tang and Puri 13) have
position was essentially uniform (Fig. 1a). When noted that combinations of size and shape differences
the surface was rough, on the other hand, a much may lead to enhanced segregation.
broader range of displacements was observed and a
progressive increase in the concentration of coarse
Density Segregation
particles with displacement was found (Fig. 1b).
 The extent of size segregation is generally limited  Differences in particle density are known to cause
by packing considerations. In the case of a binar y segregation, which might be anticipated from poten-
mixture, segregation can be expected to lead to the tial energy considerations. However, it is generally
development of two regions: one consisting essen- agreed that the effects are usually substantially less
tially of the major component alone and the other than those due to size dif ferences 23).This may be
containing both components under roughly optimum partly due to the smaller range of densities found
packing conditions. in particle systems a 5:1 range in density is con-
sidered to be large while 1000:1 is a quite common
size range. The analogy to sedimentation is again
Shape Segregation
appropriate here. The (gravitational) force acting on
 Particle shape also affects mobility and differences a particle is affected by both size and density while
are, therefore, a potential cause of segregation. It is the fluid drag (resistance to motion) is independent
generally agreed, however, that segregation by shape of density. It is likely that the same is largely true for
becomes significant only when the differences are segregation in powders. As noted above, a primary
substantial smooth spheres from irregular particles, cause of segregation is the relative ability to pen-
flakes or needles from roughly isometric particles, etrate a bed of particles; increased density provides

6 KONA Powder and Particle Journal No.27 (2009)

little advantage in this respect, especially in fairly tion. Apparently interaction forces, introduced by the
compact systems such as vibrated or agitated beds or surface coating, were sufficient to cause segregation
flow streams. A “push-away” mechanism24) has been of these relatively coarse (∼0.35 mm) particles. In
proposed in which a heavy particle is able to force this particular study, the problem was circumvented
its way between lighter components. Such an effect by switching to a different color.
is unlikely to be significant in relatively compact
beds since the weight of a single particle is generally
Role of Cohesion
much smaller than the force needed to shear the bed.
For highly dilated beds subject to vigorous agitation  By restricting the mobility of individual particles,
or vibration, and especially for fully fluidized beds, a cohesion generally serves to reduce the tendency for
mechanism of this kind may indeed be viable, leading segregation to occur. However, cohesion in dry pow-
to an accumulation of the denser component(s) in the ders is normally associated with fine (e.g., < 20μm)
lower regions of the bed25). These conditions could particles. Other than perhaps acquiring a coating of
also have a significant effect on size segregation. By fines, coarser particles in a mixture can retain their
reducing the resistance to downward motion, a more mobility and be subject to segregation. Furthermore,
open bed could actually reverse the segregation ef- the additional restriction of fine par ticle mobility
fect by allowing settling of the large particles. This may actually compound the problem. Agitation of a
could, in fact, account for some seemingly contradic- cohesive powder commonly leads to the formation of
tory observations noted in the literature. agglomerates which readily become segregated from
the bulk (fine) powder. This effect is widely exploited
in granulation processes.
Segregation Due to Particle Interactions
 Cohesion due to the presence of liquid phases gen-
 In some circumstances interaction forces, either erally extends over a broader range of particle sizes
attractive or repulsive, can promote segregation. than that due to fine components and, in some situ-
Selective agglomeration of certain components is an ations, may provide a means of minimizing segrega-
obvious example and could occur if particles of one tion problems.
component specifically attract particles of the same
kind or repel other kinds. Typically, agglomeration is
Mixing and Segregation Processes
most prevalent among the fine, cohesive components
of a mixture. Selective agglomeration is quite com-  Mixing of free-flowing powders can generally be
mon in liquid suspensions but the close confinement regarded as proceeding through some combination
of particles in dry or semi-dry powders is not gener- of the diffusion and convection mechanisms, convec-
ally conducive to selectivity. Most cases of segrega- tion being primarily responsible for macromixing
tion by clumping of fine components probably reflect while intimate micromixing relies on diffusion. The
incomplete micromixing due to low particle mobility, relative contributions of these two depend mostly on
rather than a spontaneous segregation process. the geometry and operating conditions of the mixing
 Segregation due to interaction forces is possible device but may also be influenced by powder charac-
when the forces are strong enough, or of sufficiently teristics.
long range to affect relatively large, mobile particles.
Magnetic particles are an obvious example. Other
Mixing in Rotating Cylinders
kinds of forces can produce similar effects as was
obser ved by this author. For a mixing study using  The simple horizontal drum mixer has been stud-
particles differing only in color, a narrow sieve frac- ied extensively. For non-segregating powders (equal
tion (40×50 US mesh) of crushed dolomite was dyed size, shape and density), it has been established that
using a commercial food color. Attempts to prepare axial mixing proceeds through essentially pure dif-
uniform mixtures of these particles with the same fusion9,11) while transverse mixing is dominated by
size fraction of the same dolomite, but not dyed, for convection with superimposed diffusion26-29). A bed
use as calibration standards proved to be remarkably of powder in a horizontal, rotating cylinder exists as
unsuccessful. Even when a reasonably homogeneous two, fairly distinct regions: a “static” region in which
mixture was obtained by careful stirring of small particles follow the rotation with no relative motion
quantities, any slight disturbance, by tapping or shak- and a “shear zone” where particles flow down the
ing the container, led to an immediate visible segrega- free surface of the bed. Conditions in the shear zone

KONA Powder and Particle Journal No.27 (2009) 7

depend on the Froude number: Fr = Rω2/g where R of revolutions of the cylinder and L is its length. Us-
is the cylinder radius, ω is the angular velocity and ing the above value of D, Equation 4 predicts that
g is the acceleration due to gravity. At low rotational more than 20,000 revolutions would be required to
speeds (Froude numbers less than about 0.01), ve- reduce the variance from an initial value of 0.25 to
locities in the shear zone are much higher than those 0.0025 for a cylinder 10 cm long. This required time
in the static region so that the shear zone is thin and would increase with the square of mixer length.
the free surface remains essentially planar. As the  Experimental studies by Rao et al.33) indicated that
speed is increased, velocities in the static region in- the diffusion coefficient generally decreases with in-
crease relative to those in the shear zone, which are creased filling of the cylinder and increased particle
mostly determined by gravity; the zone increases in size and increases with increased rotational speed.
thickness and assumes a characteristic S-shape. With Higher diffusion coefficients at low filling levels were
further increase in speed, cateracting occurs where attributed to an increase in the number of passes
particles are projected from the bed and follow a free- through the shear zone per revolution, partially off-
flight trajector y before re-entering the bed. At the set by the corresponding reduction in the path length
critical speed, the centrifugal force acting on the lay- in the zone. The particle size effect was thought to
er of particles adjacent to the shell just balances that result from higher velocities in the shear zone due
due to gravity (Rω2 = g) and centrifuging occurs. to higher dynamic angle of repose obser ved for
The critical speed ωc is defined by: finer particles. The increase with speed, even on a
 per revolution basis was also attributed to increases
c = (4) in the dynamic angle of repose. The approximate
analysis presented by Rao et al. suggests that the dif-
And corresponds to Fr = 1. Under steady flow condi- fusion coefficient (per revolution) should increase
tions, continuity requires that a particle entering the with the square of the cylinder diameter d for the
shear zone at some axial location will, on the average, same rotational speed. This would also be consistent
re-enter on the same transverse plane. Inter-particle with models used to describe axial transport through
collisions during flow down the surface cause the horizontal cylinders34-36). For conditions of dynamic
random lateral displacements that lead to axial diffu- similarity (constant Froude number) the diffusion
sion30,31). Convection in the transverse plane results coefficient (per time) should vary with d3/2. Unfortu-
primarily from differences in the time taken for par- nately, systematic investigations of the relationship
ticles at different radial locations to complete one between axial diffusion and cylinder diameter appear
pass through the static region. Random displacement not to have been carried out.
of particles during flow through the shear zone pro-  The addition of mixing aids in the form of rela-
vides the diffusive component. tively large balls (usually plastic or rubber) has been
shown to increase axial mixing rates significantly35-37).
Sawahata37) and Chaudhuri and Fuerstenau38) found
Axial Mixing
that the axial diffusion coefficient was increased by a
 Because of the absence of any contribution from factor of about 3 for mixing aid additions comprising
convection, axial mixing in a simple drum proceeds about 10% of the mixer volume but became constant
very slowly. Experimental values of the diffusion co- after further addition. Shoji et al.39) attributed the lev-
efficient for particles of about 0.2 mm in size rotated eling off to segregation of the balls towards the cen-
at relatively low speeds in cylinders with diameters tral region of the cylinder and correspondingly high
in the 5-10 cm range are typically found to be about 1 diffusion rates in that region. As the ball filling was
mm2/revolution9,32,33). increased further, their concentration in the central
 It has been shown9) that, for a simple cylinder load- region became more or less constant while their dis-
ed end-to-end with equal quantities of two, essentially tribution along the cylinder became more uniform.
identical components, the expected sample variance When the concentration of balls had become essen-
due to incomplete macromixing can be estimated tially uniform along the length of the cylinder, further
from: ball additions again led to enhanced diffusion of the
  particles up to the condition where the volume of
2 2π 2 DN
σ = 2 exp −
(5) particles was just sufficient to fill the voids in the bed
π L2
of balls. Abouzeid and Fuerstenau40) demonstrated
where D is the diffusion coefficient, N is the number that the presence of mixing aids led to substantial

8 KONA Powder and Particle Journal No.27 (2009)

reduction of size segregation of particles in a rotating

Transverse Mixing T
To r

 The convection process that dominates mixing


in the transverse plane of a rotating cylinder results

from the different rates at which particles at differ-
ent radial locations circulate the particle bed26,27). As Fig. 2 Geometric arrangement used in the analysis of convective mixing
Figure 2. Geometric arrangement used in the analysis of convective mixing in the
in the transverse plane of a rotating cylinder.
shown in Fig. 2, a particle at a radial distance r from transverse plane of a rotating cylinder.

the center of the cylinder travels through an angle

2θ in each pass through the static region. For a
cylinder rotating at an angular velocity ω, the time 0 Revolutions 1/4 Revolutions

to pass though the static region is equal to 2θ/ω.

If the filling of the cylinder with powder is less than
50%, the angle and the circulation time decrease with
decreasing radial distance from the center. Thus,
particles close to the center circulate several times in
the time taken for a particle adjacent to the shell to
circulate once.
 The result of the difference in circulation time is
that particles in the bed undergo a progressive angu- 1/2 Revolutions
1 Revolution

lar displacement leading to the transformation of an

initial side-by-side arrangement of two components
into a series of spiral bands as illustrated in Fig. 3.
The pattern shown in the figure was determined us-
ing the assumption that particles entering the shear
zone on the upper part of the free surface reappear Fig. 3 Schematic representation of transverse mixing in a horizontal
rotating cylinder at 30% filling.
instantaneously at the same radial position on the
lower part, i.e., the time spent in the shear zone is
negligible27). Modifying the calculations by allowing a through the shear zone (diffusion) causes a blurring
finite time for passage through the shear zone leads of the boundaries and a rapid approach to complete
to a retardation of the development of the pattern but mixing. Hogg and Fuerstenau27) presented a simpli-
to only minor changes in its form. fied model for the overall process, treating the dif-
 As the filling of the cylinder approaches 50%, the fusion component as equivalent to simple diffusion
difference in circulation times is reduced until, ideal- between the alternate bands produced by convection.
ly, at exactly 50%, the free surface passes through the More recently, the analysis has been extended to
center of rotation, the time for passage through the include a description of the actual mixing process in
static region becomes the same for all radial locations the shear zone28,29). Prigozhin and Kalman29) present-
and no convective mixing occurs. The time taken for ed a model which included an implicit assumption of
particles to pass through the shear zone may actu- perfect mixing in the shear zone which is essentially
ally assist the mixing process for this case. Due to equivalent to the treatment by Inoue et al. 26) who
the greater distances involved, particles entering the assumed that particles entering the shear zone are
zone close to the cylinder wall require a longer time redistributed at random as they reenter the rotating
than those near the center. This effect again leads bed. Kharkhar et al,28) developed a simplified model
to an angular displacement of particles at different for flow in the shear zone and described mixing in
radial positions, resulting in a spiral pattern similar to the zone as a process of diffusion between layers.
those shown in Fig. 3 but taking substantially longer Their results suggest that particles are not complete-
to develop. ly mixed in the shear zone and that description as a
 The patterns shown in the Fig. 3 are the result of diffusion process is appropriate. Each of the models
convection alone. In practice, random interchange agrees qualitatively with experimental obser vation
between particles in different layers as they pass but, because of the many approximations involved

KONA Powder and Particle Journal No.27 (2009) 9

and the problem of obtaining precise measurements, simple binary mixtures and has shown that even size
quantitative comparisons are difficult to obtain. differences as small as 20% can cause segregation.
For such small differences, the concept of percolation
seems inappropriate. Nevertheless, the differences in
Segregation in Rotating Cylinders
resistance to motion noted above still apply and lead
 Both axial and transverse segregation are com- to segregation. As might be expected, larger size dif-
monly obser ved in rotating cylinder systems 41) . ferences promote faster segregation45).
Transverse segregation usually proceeds rapidly and
appears to be a necessary precursor to axial segrega-
Axial Segregation
 Separation of dissimilar components in the form of
alternate bands along the axis of the cylinder is com-
Transverse Segregation
monly observed41,47). It appears that the process is a
 Segregation of particles in the transverse plane consequence of axial transport of particles caused
of a horizontal rotating cylinder occurs during flow by small variations in the inclination of the rotating
through the shear zone. Segregation by size, den- bed due to friction at the end-walls and differences in
sity and shape has been observed. Size segregation internal friction of the mixture41,48-50). Axial transport
results from differences in resistance to flow. Large due to the end-effect is illustrated in Fig. 4 which
particles flow readily over the surface but encounter shows schematically the expected trajectories of
considerable resistance to penetration into the bed particles from different locations in the bed as they
of flowing material. For smaller par ticles, on the pass through the shear zone. The curved paths are
other hand, trapping effects restrict their ability to the result of motion in the direction of steepest slope
flow over the surface but they are able to penetrate and are estimated using modification of the models
into and through voids in the bed. The combination developed for continuous flow through rotating cyl-
of these two mechanisms leads to accumulation of inders34,35,51). As expected, particles originating from
the coarser components around the periphery of the close to the outer shell of the cylinder are subject
rotating material and concentration of fines in a cen- to greater axial displacement than those closer to
tral core. Segregation by shape may also occur due the center of the bed. Since each of the trajectories
to differences in resistance to flow. Again, it might shown returns to its original axial location, these
be expected that particles that exhibit the greater estimates, based on independence of the individual
resistance would accumulate in the central core, but paths, would predict no net axial displacement.
experimental data in this area are ver y limited in- However, the figure also shows that, especially for
deed. In the case of density segregation, the greater
force of gravity acting on the denser particles causes
a bias in the relative motion leading to accumulation
of these particles in the central core43). The density
effect may be enhanced at high rotational speeds
Transverse Position

where the shear zone becomes partially fluidized.

 Transverse segregation occurs rapidly, approaching
a steady state in a few revolutions of the cylinder44-46).
In practice, however, variations in the distribution of
particles across the transverse plane may continue
for substantially longer times due to the effects of si-
multaneous, but much slower, axial segregation. The
limiting segregation pattern seems to be relatively
independent of rotational speed and volumetric filling
of the cylinder, except when a “dead zone” exists for
fillings greater than 50%. Some attenuation of the seg-
0.00 0.10 0.20
regation may be expected at higher rotational speeds
due to enhanced diffusive mixing. The size distribu- Axial Position
tion of the powder is obviously an important factor. Fig. 4 Schematic illustration of particle trajectories in the surface layers
Most research in the area has been carried out on adjacent to the end-walls of a rotating cylinder.

10 KONA Powder and Particle Journal No.27 (2009)

the outer layers, the paths of particles starting from from the ends of the cylinder may result from statis-
different distances from the end-wall converge dur- tical fluctuations in bed depth or inclination48,49,53,54).
ing passage through the shear zone. Obviously, However, it seems unlikely that such perturbations
therefore, the individual particle paths cannot really would be of sufficient magnitude and persistence to
be independent, but there must nevertheless be a cause significant axial segregation. Propagation of
tendency for accumulation of those particles in that disturbances from the end-walls appears to offer a
axial location. In order to maintain uniform loading more a more realistic explanation.
of particles along the cylinder, migration of material
away from that region must occur from all layers in
Mixing in Simple Flow
the shear zone. Since there is less replenishment into
the inner layers, the result is a reduction in the rela-  Both mixing and segregation in powders are
tive amount of that material. If transverse segrega- generally promoted by flow of the material. Flow in
tion has led to concentration of the finer (or heavier) powders can be considered to fall into one or other
particles in the central core and of the coarser (or of two basic types: rapid flow as in the shear zone of
lighter) material in the outer layers, the net effect of a rotating cylinder or in the formation of a heap and
the axial transport described above is to form a band slow flow as in internal motion during slow discharge
enriched in the outwardly segregating material close of a hopper. Slow flow appears to be characterized by
to but slightly removed from the end-wall, with cor- shearing at specific slip planes between other wise
responding regions of depletion on both sides. static layers of powder. In rapid flow, on the other
 Axial variations in the composition of the mixture hand, the process involves the existence of a continu-
due to the end-effects or even random fluctuations ous velocity gradient, akin to that in flow of a viscous
may lead to corresponding changes in inclination of fluid.
the rotating charge and consequent axial migration  Flow over an inclined sur face such as a chute
by similar mechanisms to those described above and is a simple example of rapid flow. If the surface is
the formation of alternating bands along the length of smooth, motion occurs by sliding and there is little
the cylinder. or no shearing within the powder. Since there is little
 It is clear that the mechanisms of axial segregation relative motion of individual particles under such
are complex and the convective processes described conditions, mixing effects are minimal. However, if
here can only represent part of the picture. Individual the surface is roughened, to an extent equal to or
particle mobility can be expected to var y with size greater than the surface of the powder bed itself,
and shape and perhaps with density. It follows that shear occurs within the powder and liquid-like flow
dif fusivity should be similarly variable and could occurs, provided of course the inclination of the sur-
either promote or oppose segregation. In the case face exceeds the angle of friction between powder
of segregation due to variations in inclination of the and surface. Particle-particle collisions due to shear
bed, Savage52) has shown that both migratory convec- provide a driving force for diffusional mixing.
tion and random diffusion fluxes could be roughly If diffusion is promoted by shear, it is reasonable to
proportional to the concentration gradient. Math- expect that the rates should increase with increasing
ematically, the process could then be represented by shear rate55,56). A simple linear relationship has been
a simple diffusion equation with an effective diffusion proposed56) of the form:
coefficient given by the difference between contribu-
tions from convection and true diffusion. Under some D = D0 (1 + α ) (6)
circumstances, this could result in a negative effec-
tive diffusion coefficient meaning that the process where D is the diffusion coefficient, ∂v / ∂y is the
would lead to enhanced rather than reduced con- velocity gradient perpendicular to the flow direction,
centration gradients, i.e., to segregation. Simulation Do and α are constants. Experiments were carried
of the process using a cellular automaton showed out to evaluate the mixing of sand with garnet pow-
that combinations of random (diffusion) and biased der as a tracer fed as initially separate layers, one on
(convection) interchange of particles in an initially top of the other onto the surface of an inclined chute
uniform mixture can indeed result in band formation whose sur face was roughened using an attached
and exhibit the kinds of instabilities observed in real layer of the same sand. Concentration profiles were
systems. determined for several distances of flow down the
 It has been suggested that band formation away chute by a splitting procedure followed by magnetic

KONA Powder and Particle Journal No.27 (2009) 11

separation of the components 56). Velocity profiles 1.0

were estimated using the trajector y procedure de- Distance Along Surface, cm

Cumulative Mass Fraction of Tracer

0.8 0
scribed by Augenstein and Hogg57). An example of 51.4
the distribution of tracer through the depth of the 0.6 106.3
bed is shown in Fig. 5 for increasing distance along g
the inclined surface. The solid lines are model predic- e Mi
0.4 et
tions based on Equation 6 with fixed values of the pa- Co

rameters D0 and α of 1.15×10-2 cm2/sec and 7.03× 0.2

10-3 sec. respectively. The agreement between model

and experiment is very good. 0.0
0.0 0.2 0.4 0.6 0.8 1.0
 The results shown in Fig. 5 are for the case where Fraction of Bed Thickness
the tracer was added above the layer of sand. Be- Fig. 5 Mixing of tracer particles during flow over an inclined surface.
cause of the dependence of the diffusion coefficient Tracer fed as a thin layer at the bottom of the steam. Points are
on the velocity gradient, more extensive mixing experimental; lines are calculated based on Equation 6. Data of
Hwang 58).
should be expected if the tracer was added below the
sand, due to the higher gradient closer to the chute 1.0

surface. A comparison of the results for tracer addi-

tion above and below the sand layer is shown in Fig.

Cumulative Fraction of Tracer

51.4 cm
6. It is clear that placing the tracer in the high-shear 0 cm
region adjacent to the chute surface does indeed lead 0.6 ix i
eM 51.4 cm
to enhanced mixing. Furthermore, the fact that the Co

results could be fitted using the same values of the
parameters D0 and α provides strong support for the Tracer Above
0.2 Tracer Below
applicability of Equation 6 to this system. It appears 0 cm

that agitation of the powder due to flow is sufficient 0.0

to promote diffusive mixing and is further enhanced 0.0 0.2 0.4 0.6 0.8 1.0

Fraction of Bed Thickness

by velocity gradients within the flow stream.
Fig. 6 Comparison of mixing behavior for tracer fed above and below the
main flow stream. Calculated lines are based on the same values
Segregation in Simple Flow of the diffusion parameters. Data of Hwang 58).

 The conditions that promote mixing during simple 1.0

flow of non-cohesive powders also favor segrega- Distance Along Surface, cm

tion of particles with different characteristics. Some 0
results of experiments carried out on initially homo- 0.8
Cumulative Mass Fraction of Coarse

geneous mixtures consisting of 10% coarse (20×30

US mesh) and 90% fine (50×70 US mesh) sand are
shown in Fig. 7. It appears that segregation occurs
rapidly during the initial stages of the flow, leading to
significant depletion of the coarse particles close to
the chute surface and corresponding enhancement
near the upper, free surface of the stream. Further 0.4
downstream, the process continues more slowly.
Hwang20,58) proposed that the data were consistent
with a simple extension to the diffusive mixing model
described above with segregation assumed to occur
through a constant upward convective velocity of the
coarse particles. The solid lines shown in Fig. 7 were
obtained using this model. The slowing down of the 0.0
0.0 0.2 0.4 0.6 0.8 1.0
segregation process following the initially high rate
Fraction of Bed Thickness
is attributed to the effects of diffusive mixing which
Fig. 7 Segregation of coarse (20×30 US mesh) and fine (50×70 US
assume greater importance as the concentration gra- mesh) sand particles during flow of an initially homogeneous
dient develops. mixture over a rough surface. Data of Hwang 58).

12 KONA Powder and Particle Journal No.27 (2009)

 Savage and Lun59) conducted similar experiments  It is generally agreed 22) that large par ticles en-
using mixtures of polystyrene beads but with the counter less resistance to flow over the surface than
finer particles as the minor component. These au- does finer material which is more easily trapped in
thors proposed two segregation mechanisms: a “ran- surface voids. As a result the coarser components
dom fluctuating sieve” mechanism and a “squeeze in a mixture can move further down the surface and
expulsion” mechanism. The former corresponds to tend to accumulate around the bottom of the heap.
size-dependent percolation while the latter refers to At the same time, build-up of the finer fractions in
the ejection of particles, by mechanical forces, from the upper regions of the surface can lead to periodic
one layer to another and was assumed to be size- avalanching producing patches or streaks of fines
independent with no preferential direction. This com- within the accumulated coarse material. In the case
bination represents a more realistic treatment of the of particles of similar size but different density, the
segregation mechanisms than the simple convection heaver particles tend to be concentrated in the center
process described above, in particular by providing of the heap with the lighter components towards the
the means to maintain mass conservation across the periphery22,24). The segregation patterns formed in bi-
bed. Analysis of the two processes provides some nary mixtures have been shown to depend primarily
predictive capability for the effects of variables such on the relative sizes and densities of the components,
as particle size ratio on segregation rates. Diffusive their concentration and the velocity of impingement
mixing was not included in the model, leading to the on the surface of the heap12,22).
rather unlikely conclusion that the process ultimately  Studies of the formation of two-dimensional heaps
results in complete separation of the coarse and fine from binar y mixtures with relatively small dif fer-
components. ences in size have shown the development of a regu-
 Dolgunin et al.60,61) conducted experiments simi- lar pattern of striations over the entire heap, except
lar to those described by Hwang et al., 20,58). They at the central core formed during the initial stages of
concluded that the bulk density was highest in the the process65-67). However, there is reason to suspect
central region of the flowing stream as were the con- that the highly regular pattern may be an artifact of
centrations of the larger or denser particles in the the materials used and the nature of the test proce-
mixture. It should be noted that these experiments dure.
were restricted to relatively thin layers typically  In practice, pouring generally leads to the forma-
around 4 particle diameters which might account tion of three-dimensional, roughly conical heaps. This
for the form of the bulk density profiles. The authors adds further complication to the formation process
correlated the observed segregation patterns to the and any resulting segregation12). Avalanching, due to
bulk density profile which may also explain the dif- local variations in slope, is unlikely to occur simulta-
ferences in their obser vations from those typically neously around the heap leading to asymmetric radial
seen in other studies: migration of coarse particles distributions of the different particles. Flow over the
towards the upper region of the flowing bed. surface is divergent, causing downstream thinning
 Theoretical analyses have been used 62-66) to predict of the stream and possibly increased opportunity for
the effects of particle and system variables size, trapping of particles.
density, elasticity, chute inclination, etc., on segre-
gation in flow over inclined surfaces. Different pro-
Segregation in Bins and Hoppers
cedures appear to give equivalent results and to be in
qualitative agreement with experimental data.  Segregation can be a serious problem in the stor-
age of granular materials in bins or silos. While some
segregation may occur across failure planes during
Segregation in Heap Formation
flow through a bin68), such effects are probably small
 Flow over the surface of a heap has much in com- relative to surface segregation during filling and/or
mon with flow down a chute. However, the process is discharge. The problem of segregation is most pro-
more complex due to the finite extent of the surface nounced in funnel-flow bins in which flow takes place
in the flow direction and the continuous development through a channel formed within the mass of powder.
of the surface itself. Because the surface is inclined Mass-flow bins, where material flows essentially en
at essentially the angle of repose of the powder, flow masse by sliding at the walls, provide little opportu-
can readily become unstable and subject to tempo- nity for segregation69).
rary build-up of particles followed by avalanching.  Segregation can occur for both types if the material

KONA Powder and Particle Journal No.27 (2009) 13

is fed into the bin so as to form a heap leading, for through combinations of two basic mechanisms: Con-
example, to accumulation of coarse particles close to vection and diffusion. Other proposed mechanisms
the walls, with a central core of finer particles. Seg- such as “shear mixing” represent special cases or
regation can also result from air entrainment of very specific combinations of the two. Convection pro-
fine particles during feeding12). This effect can cause vides for rapid macromixing but is considerably less
deposition of fines close to the walls leaving excess effective in promoting micromixing. Diffusion, on the
coarse material in the central region. In a mass-flow other hand, is the primary mechanism for producing
bin, the segregation pattern remains more or less intimate micromixing. However, because the process
intact during passage through the bin but becomes occurs at the individual particle scale, mixing by dif-
recombined in the converging hopper section at the fusion alone proceeds very slowly. Convective mixing
discharge. The situation is more complicated for is controlled largely by the geometry and mechanical
funnel-flow bins. Since flow is confined to an internal action of a mixing system while diffusion is more de-
channel, the initial discharge comes from the fines- pendent on particle mobility and is severely hindered
rich central core. As flow continues, a funnel-shape by cohesion of particles.
depression develops in the upper region of the bin  Segregation, i.e., de-mixing, is possible whenever
contents. Coarse particles from the region close to different kinds of particles var y in mobility or re-
the walls then begin to flow preferentially towards the sponse to external forces. Segregation by size can
center of the bin, reversing the segregation process. usually be attributed to differences in individual parti-
As a result, the discharge can switch from excess cle mobility fines have higher mobility within a bed
fines to excess coarse material. In many applications, of particles while coarser material is more mobile on
bins are used to provide surge capacity with alternate free surfaces. Shape effects can be expected to fol-
or simultaneous feeding or discharge. The segrega- low a similar pattern, with isometric particles having
tion patterns will then change according to whether greater mobility than plates or needles, for example.
feeding or discharge dominates and cycling of the Density differences can promote segregation due to
discharge composition can occur. the different gravitational force experienced by heavy
 Johanson70) and Carson et al.,71) have described pro- or light particles. The actual form of segregation
cedures for minimizing segregation in storage bins. A top to bottom or side to side, for example depends
principal recommendation is to ensure that mass-flow on the specific conditions of mechanical agitation of
conditions prevail. This should certainly be a priority the system. Variations in both size and density can
in the design of new installations. However, despite lead either to enhancement or reduction in segrega-
more than 40 years of awareness of the advantages tion tendency depending on specific conditions.
of such designs, the widespread use of funnel-flow  Because segregation is normally a result of differ-
bins remains. Conversion to or replacement by mass- ences in individual particle mobility, its occurrence
flow systems is often considered to be economically is most common in free-flowing powders. However,
impractical, especially for large-scale systems. Partial segregation of coarse particles is possible even in the
conversion can be achieved through the use of prop- presence of cohesive fine material. Aggregation of
erly designed and placed inserts70). It should be noted fines due to cohesion can also lead to segregation.
that complete drainage should be avoided, even for  Procedures for minimizing segregation tend to be
mass-flow bins; a minimum filling level, somewhat highly system-specific; there is no universal solution.
above the converging hopper, should be maintained Avoidance of mechanical agitation is generally appro-
to ensure uniform flow of the material. priate the use of completely, rather than partially,
 Segregation during bin filling, due to heap forma- filled containers, for example. Introducing cohesion
tion, can be reduced by distributing the feed over by liquid addition, where practical, can be effective
the bin cross section. Air entrainment effects can but will, of course, exacerbate problems in mixing
be largely eliminated by minimizing free-fall height and storage.
during filling. A useful summary of procedures for
suppressing segregation in storage systems has been
presented by Tang and Puri13).
D diffusion coefficient [L2/T]
Do diffusion parameter (Equation 6) [L2/T]
Concluding Remarks
fi mass fraction of component i in a mixture[ ]
 Mixing of solid par ticles generally proceeds g acceleration due to gravity [L/T2]

14 KONA Powder and Particle Journal No.27 (2009)

L length of a mixing drum [L] regation of Particulate Materials Mechanisms and
M sample weight [M] Testers, KONA Powder and Particle Journal, No.14.
Np number of particles in a sample [ ] pp. 31-42.
13) Tang, P. and Puri, V.M. (2004): Methods for Minimiz-
r radial coordinate [L]
ing Segregation: a Review, Part. Sci. Technol., 22. pp.
ro distance between powder bed and the 321-337.
 center of a horizontal cylinder [L] 14) Bridgewater, J. (1972): The Mixing of Cohesionless
R radius of a cylinder [L] Powders, Powder Technol., 5. pp. 257-260.
v velocity [L/T] 15) Williams, J.C. (1963): The Segregation of Powders and
wi mean mass of a particle of component i [M] Granular Materials, Fuel Soc. J., 14. pp. 29-34.
w overall mean particle mass in a mixture [M] 16) Rosato, A.D., Prinz, F., Standburg, K.J. and Svendsen,
R. (1987): Phys. Rev. Lett. 58. pp. 1038-1040.
y vertical location in a flow stream [L]
17) Matthée, H. (1968): Segregation Phenomena Relating
α diffusion parameter (Equation 6) [T]
to Bunkering of Bulk Materials: Theoretical Consid-
θ angular coordinate [ ] erations and Experimental Investigations,” Powder
θo angle defining the filling of a cylinder [ ] Technol., 1. pp. 265-271.
σ2 expected variance (by number) 18) Williams, J.C. (1976): The Segregation of particulate
 of mixture composition [ ] Materials. A Review, Powder Technol., 15. pp. 245-251.
σ2m expected variance (by mass) 19) Augenstein, D.A. (1974): A Study of the Flow of Dry
 of mixture composition [ ] Par ticulate Solids, PhD Thesis, The Pennsylvania
State University,.
ωc critical angular velocity of
20) Hogg, R., Augenstein, D.A. and Hwang, C.L.( 1975):
 a rotating cylinder [T 1]
Segregation In Flowing Powders, Paper No. 98b,
presented at the 68th Annual Meeting of AIChE, Los
Angeles, CA.
21) Lawrence, L.R. and Beddow, J.K. (1969): Powder Seg-
1) Lacey, P.M.C. (1945): The Mixing of Solid Particles, regation during Die Filling, Powder Technol., 2. pp.
The Chemical Age, 53. Aug. 11/18, pp. 119-124. 253-259.
2) Buslik, D. (1950): Mixing and Sampling with Special 22) Drahun, J.A. and Bridgwater, J. (1983): The Mecha-
Reference to Multi-Sized Granular Material, ASTM nism of Free Surface Segregation, Powder Technol.,
Bull., 165. pp. 66-73. 36. pp. 39-53.
3) Stange, K. (1954): Degree of Mixing in a Random Mix- 23) Vallance, J.W. and Savage, S.B. (2000): Particle Seg-
ture as a Basis for Evaluating Mixing Experiments, regation in Granular Flows down Chutes, in IUTAM
Chemie-Ing. Tech., 26. pp. 331-337. Symposium on Segregation in Granular Flows, A.D.
4) Stange, K. (1963): Mixing Quality in a Random Mix- Rosato and D.L. Blackmore, eds., Kluwer Academic
ture of Three or More Components, Chemie-Ing. Publishers, Boston, pp. 31-51.
Tech., 35. pp. 580-582. 24) Tanaka, T. (1971): Segregation Models of Solid Mix-
5) Poole, K.R., Taylor, R.F. and Wall, G.P. (1964): Mixing tures Composed of Different Densities and Particle
Powders to Fine-Scale Homogeneity: Studies of Batch Sizes, Ind. Eng. Chem., Proc. Des. and Dev., 10. pp.
Mixing, Trans. Inst. Chem. Eng., pp. T305-T315. 332-340.
6) Harnby, N. (1967): The Estimation of the Variance of 25) Geldart, D., Pope, D.J., Baeyens, J. and van de Wijer,
Samples Withdrawn from a Random Mixture of Multi- P. (1981): Segregation in Beds of Large Particles at
Sized Particles, Chem. Eng., 214. pp. CE270-CE271. High Velocities, Powder Technol., 30. pp.195-205.
7) Walpole, R.E. and Myers, R.H. (1985): “Probability and 26) Inoue, I., Yamaguchi, K. and Sato, K. (1970): Motion
Statistics for Engineers and Scientists, 3rd Ed.” , Mac- of Particle and Mixing Process in a Horizontal Drum
millan, New York. Mixer, Kagaku Kogaku, 34. pp. 1323-1330.
8) Danckwerts, P.V. (1952): The Definition and Measure- 27) Hogg, R. and Fuerstenau, D.W. (1972): Transverse
ment of some Characteristics of Mixtures, Appl. Sci. Mixing in Rotating Cylinders, Powder Technol., 6. pp.
Res., 3 (A). pp. 279-296. 139-148.
9) Hogg, R., Cahn, D.S., Healy, T.W. and Fuerstenau, 28) Khakhar, D.V., McCarthy, J.J. and Ottino, J.M. (1997):
D.W. (1966): Diffusional Mixing in an Ideal System, Transverse Flow and Mixing of Granular Materials in
Chem. Eng. Sci., 21. pp. 1025-1037. a Rotating cylinder, Phys. Fluids, 9. pp.31-43.
10) Hogg, R. (2003): “Characterization of Relative Homo- 29) Prigozhin, L. and Kalman, H. (1998): Radial Mixing
geneity in Particulate Solids,” Int. J. Miner. Process, and Segregation of a Binar y Mixture in a Rotating
72. pp. 477-487. Drum: Model and Experiment, Phys. Rev., 57. pp.
11) Lacey, P.M.C. (1954): Developments in the Theory of 2073-2080.
Particle Mixing, J. App. Chem., 4. pp. 257- 268. 30) Kaye, B.H., and Sparrow, D.B. (1964): Role of Surface
12) Mosby, J., de Silva, S.R. and Enstad, G.G. (1996): Seg- Diffusion as a Mixing Mechanism in a Barrel-Mixer,

KONA Powder and Particle Journal No.27 (2009) 15

Indust. Chemist, 5. pp. 246-250. Cylinder, Chem. Eng. Sci., 46, pp. 1513-1517.
31) Cahn. D.S., Fuerstenau, D.W., Healy, T.W., Hogg, R. 50) Hill, K.M., Caprihan, A. and Kakalios, J. (1997):
and Rose, H.E. (1966): Diffusional Mechanisms of Axial Segregation of Granular Media Rotated in a
Solid-Solid Mixing, Nature, 209. pp. 194-496. Drum Mixer: Pattern Evolution, Phys. Rev. E, 56, pp.
32) Abouzeid, A-Z.M., and Fuerstenau, D.W. (1972): Ef- 4386-4393.
fects of Humidity on Mixing of Particulate Solids, Ind. 51) Saeman, W.C. (1951): Passage of Solids through Ro-
Eng. Chem., Proc. Des. and Dev., 11. pp. 296-301. tary Kilns,” Chem. Eng. Progr., 47, pp. 508-514.
33) Rao, S.J., Bhatia, S.K. and Khakhar, D.V. (1991): Axial 52) Savage, S.B. (1993): Disorder, Diffusion and Structure
Transport of Granular Solids in Rotating Cylinders. Formation in Granular Flow, in "Disorder and Granu-
Part 2. Experiments in a Non-Flow System, Powder lar Media", D. Bideau and A. Hansen, eds., Elsevier,
Technol., 67. pp. 153-162. Amsterdam, pp. 255-285.
34) Vahl, L. and Kingma, W. G. (1951): Transport of Solids 53) Bridgwater, J. (1976): Fundamental Powder Mixing
through Horizontal Rotar y Cylinders, Chem. Eng. Mechanisms, Powder Technol., 15, pp. 215-236.
Sci., 1. pp 253-258. 54) Zik, O., Levine, D., Lipson, S.G. Shtrickman, S. and
35) Hogg, R., Shoji. K. and Austin, L.G. (1974): Axial Stavans, J. (1994). Rotationally Induced Segregation of
Transport of Dry Powders in Horizontal Rotating Cyl- Granular Materials, Phys. Rev Lett., 73, pp. 644-647.
inders,” Powder Technol., 9. pp. 99-106. 55) Hogg, R. and Hwang, C.L. (1977): Mixing in Simple
36) Karra, V.K. and Fuerstenau, D.W. (1978): Scale-up of Flow Systems, J. Powder and Bulk Solids Technol., 1,
the Axial Profile of Material Hold-up in Horizontal Ro- pp.72-75.
tating Cylinders, Powder Technol., 19. pp. 265-269. 56) Hwang, C.L. and Hogg, R. (1980): Diffusive Mixing in
37) Sawahata, Y. (1966): On the Mixing of Solids with Mix- Flowing Powders, Powder Technol., 26, pp. 93-101.
ing Aids, Kagaku Kogaku, 30. pp. 178-179. 57) Augenstein, D.A. and Hogg, R. (1978): An Experimen-
38) Chaudhuri, P.K. and Fuerstenau, D.W. (1971): The tal Study of the Flow of Dr y Powders over Inclined
Effect of Mixing Aids on the Kinetics of Mixing in a Surfaces, Powder Technol., 19. pp. 205-215.
Rotating Drum, Powder Technol., 4. pp. 146-150. 58) Hwang, C.L.( 1978): Mixing and Segregation in Flow-
39) Shoji, K., Hogg, R. and Austin, L.G. (1973): Axial Mix- ing Powders, PhD Thesis, The Pennsylvania State
ing of Particles in Batch Ball Mills, Powder Technol., University,.
7. pp. 331-336. 59) Savage, S.B. and Lun, C.K.K. (1988): Particle Size Seg-
40) Abouzeid, A-Z.M. and Fuerstenau, D.W. (1979): Effect regation in Inclined Chute Flow of Dry Cohesionless
of Mixing Aids on the Transport Behavior of Particu- Granular Solids, J. Fluid Mech., 189, pp. 311-335.
late Solids, Powder Technol., 23. pp. 261-277. 60) Dolgunin, V.N. and Ukolov, A.A. (1995): Segregation
41) Donald, M.B. and Roseman, B. (1962): Mechanisms Modeling of Particle Rapid Flow, Powder Technol., 83,
in a Horizontal Dr um Mixer, Br. Chem. Eng., 7, pp. 95-103.
pp.749-752. 61) Dolgunin, V.N., Kudy, A.N. and Ukolov, A.A. (1998) :
42) Ottino, J.M. and Khakhar, D.V. (2000): Mixing and Development of the Model of Segregation of Particles
Segregation of Granular Materials, Ann. Rev. Fluid. undergoing Granular Flow down an Inclined Plane,
Mech., 32. pp. 55-91. Powder Technol., 96, pp. 211-218.
43) Ristow, G.H. (1994): Particle Mass Segregation in a 62) Hsiau, S.S. and Hunt, M.L. (1996): Granular Thermal
Two-Dimensional Rotating Drum, Europhys. Lett., 28. Dif fusion in Flows of Binar y-Sized Mixtures, Acta
pp. 97-101. Mech., 114, pp. 121-137.
44) Rogers, A.R. and Clements, J.A. (1972): The Examina- 63) Hirschfeld, D. and Rapaport, D.C. (1997): Molecular
tion of Segregation of Granular Materials in a Tum- Dynamics Studies of Grain Segregation in a Sheared
bling Mixer, Powder Technol., 5. pp. 167-178. Flow, Phys. Rev. E, 56, pp. 2012-2018.
45) Nityanand, N., Manley, B. and Henein, H. (1986): An 64) Khakhar, D.V., McCarthy, J.J. and Ottino, J.M. (1999):
Analysis of Radial Segregation for Dif ferent Sized Mixing and Segregation of Granular Materials in
Spherical Solids in Rotary Cylinders, Metall. Trans. B, Chute Flow, CHAOS, 9, pp. 594-610.
17B, pp. 247-257. 65) Makse, H.N., Ciseau, P. and Stanley, H.E. (1997): Pos-
46) Cantelaube, F., Bideau, D. and Roux, S. (1997): Kinet- sible Stratification Mechanism in Granular Mixtures,
ics of Segregation of Granular Media in a Two-Dimen- Phys. Rev. Lett., 78, pp. 3298-3301.
sional Rotating Drum, Powder Technol., 93, pp. 1-11. 66) Koeppe, J.P., Enz, M. and Kakalios, J. (1998): Phase
47) Oyama, Y. (1940): Mixing of Binar y System of Two Diagram for Avalanche Stratification of Granular Me-
Sizes by Ball-Mill Motion, Sci. Papers Inst. Phys. dia, Phys. Rev. E, 58, pp. R4104-R4107.
Chem. Research (Tokyo) 37, pp. 17-29. 67) Boutreux, T. (1998): Surface Flows of Granular Mix-
48) Bridgwater, J., Sharpe, N.W. and Stocker, D.C. (1969): tures: II. Segregation with Grains of Different Size,
Particle Mixing by Percolation, Trans. Inst. Chem. Eur. Phys. J. B, 6, pp. 419-424.
Eng., 47, pp. T114-119. 68) Stephens, D.J. and Bridgwater, J. (1978): The mixing
49) Das Gupta, S. Bhatia, S.K. and Khakhar, D.V. (1991): and Segregation of Cohesionless Particulate Materials
Axial Segregation of Particles in a Horizontal Rotating Part II. Microscopic Mechanisms for Particles Differ-

16 KONA Powder and Particle Journal No.27 (2009)

ing in Size, Powder Technol., 21, pp.29-44. Solutions, Powder and Bulk Engineering, 2, No. 8, pp.
69) Jenike, A.W. (1964): Storage and Flow of Solids,” Bull. 13-19.
No. 123, Utah Eng. Expt. Stn. 71) Carson, J.W., Royal, T.A. and Goodwin, D.A. (1986):
70) Johanson, J.R. (1988): Solids Segregation: Causes and Understanding and Eliminating Particle Segregation
Problems, Bulk Solids Handling, 6, pp. 139-144.

Author’s short biography

Richard Hogg
Richard Hogg is Professor Emeritus of Mineral Processing and GeoEnvironmental
Engineering at the Pennsylvania State University. He received a B.Sc. from the Uni-
versity of Leeds and the M.S. and PhD degrees from the University of California
at Berkeley. Dr Hogg s research interests include fine particle processing, particle
characterization, and colloid and surface chemistry.

KONA Powder and Particle Journal No.27 (2009) 17