Introduction 1

1-1 Definition and Purpose

The behavior of systems involving the motion of aggregates of small par-
ticles relative to fluids in which they are immersed covers a wide range of
phenomena of interest to both scientists and engineers. Broadly speaking
we may assign these processes to several classes. Particles may move to-
gether in bulk through a fluid, as in sedimentation. In turn, the particles may
remain more or less stationary as in a packed bed. The relative particle-fluid
motions may be more complex, as in fluidized systems. Finally, the phe-
nomenon of suspension viscosity or resistance to shear is encountered when
solid particles move relative to each other owing to shearing motion of
the suspending fluid, as contrasted with situations where the fluid moves
relative to the entire particle system. Many processes involving these types
of motion are found in nature and technology. It is the basic purpose of
this book to develop an understanding of such behavior of multiparticle
systems, starting with the dynamics of single particles.
One might suppose that all the basic problems and important applica-
tions had been solved long ago. For it is fashionable for much of present-day
science to probe the behavior of objects of extremely small size by means
of cyclotrons or those of much larger size by radio telescopes. Only recently,
however, have we begun to establish a satisfactory synthesis of the basic
principles of slow viscous flow. Before proceeding to other matters of general
interest, it is desirable to form an idea of the size range of objects with which
2 Introduction

we shall be concerned. The term particle is a loose one, but for convenience
we shall assume that any solid object less than 10 cm in diameter (2.54 cm
= 1 in.) falls into this category. Particles may vary downward from this
size until molecular dimensions are reached, the lower limit of size of interest
to us here being that in which the fluid surrounding our particles can be
regarded as continuous. This will vary depending on the system being stud-
ied, but interesting results within the purview of the present treatment are
obtained down to sizes of 10- 7 cm (10 Angstrom units). This range of size
may be visualized by noting that the diameter of Earth is about 12 x 108
cm (8000 miles). Thus, Earth is just about as many times as big as the largest
particle we shall consider, as this particle in turn is larger than the largest
organic molecules. Figure 1-1.1 gives an idea of the different types of
materials which may be encountered. Note that the scale is plotted
logarithmically.

Diameter
(em)
Earth
109

108

107

106
105

104

103

10 2
101 Baseball
lOa

10-1 Bead catalyst } Sieve

Sand
Limit of 10- 2
Fluid cracking catalyst
visibility - 10- 3 Size
Cloud droplets } range
1 micron, p 10- 4 Microscopic of
(1/25,OOOin.)- Paint pigment "particles ..
10- 5
Tobacco smoke }
10-6 Plant viruses Colloidal

10- 7 Organic molecules

1 angstrom __ 10-8
Hydrogen atom
Unit, A
10- 9

Fllur. 1-1.1. Scale of sizes of various objects.

1-1 Definition and Purpose 3

In order to construct tractable mathematical models of the flow systems

involving particles, it is necessary to resort to a number of simplifications.
In this book it is assumed that the flow is laminar and, further, that it is
sufficiently "slow" that inertial effects need not be considered in arriving
at a solution of the equations of motion which describe the passage of fluid
relative to particles in these systems. This simplification is justified, since
many multiparticle systems do involve sufficiently slow motions for this
assumption to be valid. Often, systems of interest will consist of very small
particles, and even when the particles move rapidly with respect to container
walls, they will move slowly with respect to fluid passing through them or
with respect to each other. The systems which we shall treat are those which
exhibit "strongly viscous" behavior, to use an expression coined by Shapir0 40 •
Their behavior is governed by the so-called creeping motion or Stokes
equations.
A dimensionless criterion which determines the relative importance of
inertial and viscous effects is the Reynolds number:
R ld b fluid density x speed x size
eyno s num er = viscosity

Situations in which the Reynolds number is small are called slow viscous
flows, because viscous forces arising from shearing motions of the fluid
predominate over inertial forces associated with acceleration or deceleration
of fluid particles. But the Reynolds number may be small for reasons other
than the slowness of the motion or the high viscosity of the fluid. Thus flight
of an object through rarefied air high above Earth's surface may represent
a very viscous flow even though the air through which the object passes has
a very low viscosity, because its density is correspondingly much lower.
Of course, in this case, the dimensions of the object must be large compared
with the mean-free path of the air molecules. Otherwise the continuum
hypothesis is invalid. A speck of dust or a mist settling slowly through nor-
mal air, if sufficiently small, may represent a more viscous situation than
a steel ball falling in molasses. In many practical situations involving sedi-
mentation and fluidization, the Reynolds number (based on particle diam-
eter) will be less than about 5. These phenomena are then amenable to
treatment by the creeping motion equations.
Complications often arise because of the complex geometry encountered
in assemblages composed of particles of arbitrary shape. Though the basic
differential equations of motion may be well understood, it is still very diffi-
cult to arrive at exact or even approximate solutions in all but the simplest
cases. Essentially, two techniques have been employed for handling bound-
ary value problems involving a number of particles, namely, the method of
reflections and the unit cell technique. In the method of reflections the
boundary conditions are satisfied successively on each of the separate bound-
4 Introduction

ing surfaces involved, including the container walls confining the suspension
when the fluid-particle system is bounded in extent. If the system is sufficient-
ly dilute, that is, the number of particles per unit volume is small, rapid
convergence is possible. This technique is thus especially adapted to deter-
mining the interaction effects among a few particles or between single
particles and container walls. The unit cell technique, on the other hand,
involves the concept that an assemblage can be divided into a number of
identical cells, one particle (usually a sphere) occupying each cell. The
boundary value problem is thus reduced to consideration of a single particle
and its bounding envelope. This technique applies strictly only to periodic
arrays. It can, however, also be applied in some stochastic sense to random
particle arrays. The cell model is of greatest applicability in concentrated
assemblages, where the effect of container walls can be neglected.
The cell technique may be employed to show schematically, and in a
highly idealized form, the nature of the flow pattern involved in the basic
types of motion just discussed. Various investigators have employed different
shapes of cells, but the assumption of a spherical shape both for each par-
ticle and for a fictitious envelope of fluid surrounding it is of great conven-
ience. The spherical form is of interest mathematically because it enables
a surface to be described in terms of a single parameter. It is also of unusual

Uniform flow
Figure 1·1.2. Sedimentation. The solid Figure 1·1.3. Permeability. The solid
sphere at the center of the envelope sphere is stationary. The undisturbed
moves downward. field entering the fluid envelope is
a uniform flow.
 1-1 Definition and Purpose 5

practical interest because many

particles approximate the spherical
form. For illustration, we describe
briefly the fluid patterns associated
with the assumption of a concentric
spherical cell model.
Figures 1-l.2-4 schematically
illustrate the spherical "free sur-
face" 23 cell model for the cases of
sedimentation, flow through porous
media, and suspension viscosity. In
the case of sedimentation a number
of particles are assumed to be
settling with equal velocity under
the influence of gravity through a
fluid. Attention is focused on one
particle, which is surrounded by the ~ Shearing flow
dotted line which constitutes the
fluid envelope surrounding it. The Flgur. 1-1.4. Viscosity. The solid
radius of this fluid envelope is estab- sphere rotates. The undisturbed pat-
lished by assuming that the cell tern entering the fluid envelope is a
contains the same volumetric pro- shearing field.
portion of solid to fluid as exists in
the entire assemblage. Naturally, the hypothetical envelopes or cells
surrounding each particle in an actual assemblage will be distorted, and some
"leakage" of fluid from one cell to another will occur, but it is assumed that
on the average a spherical cell may be employed, owing to the random ar-
rangement. The entire disturbance due to each particle is thus confined within
the cell of fluid with which the particle is associated. The outside surface,
represented by the dotted line, is frictionless (that is, the tangential stresses
are zero), so that fluid is free to pass over the surface, as if the cell were a
raindrop freely suspended in the atmosphere. Instantaneous streamlines
representing the path of flow of fluid within this envelope are shown
schematically. They depict a circulatory type of motion. The direction and
magnitude of motion at the surface of the fluid envelope is such that, if a
similar fluid envelope surrounding another sphere approached the one
depicted at the point where they touched, it would be found that the motion
would be equal in magnitude and direction for the two spheres. Thus no
friction is supposed to occur between adjacent fluid envelopes.
From the fluid velocity pattern shown in Fig. 1-l.2 it is possible to
compute the drag force exerted by the fluid on the particle contained in the
cell. As the particles come closer together the fluid surrounding each particle
will be confined to a smaller envelope and resistance to its motion will in-
6 Introduction

crease. In the case of sedimentation, this means that, when a constant gra-
vitational force acts on a particle, as concentration increases, the cells and
the particles within them settle at a slower rate than in a system which was
sufficiently dilute that the particles did not influence each other.
In exactly the same proportion as the settling velocity of an assemblage
of given porosity decreases, resistance to flow of a fluid through a fluidized
or packed bed of equal porosity increases. Figure 1-1.3 shows an ideal-
ized fluid velocity pattern for flow through a stationary cell. Knowledge
of this field enables one to estimate, with considerable precision, the pressure
drop resulti~g from passage of fluid through a dense bed of particles.
Figure 1-1.4 depicts the flow pattern which results when an otherwise
uniform shearing field is disturbed by the presence of particles. The flow
pattern shown permits us to calculate the energy dissipated by fluid friction
in the envelope surrounding the particle. A spherical particle carried along
by a shearing field will rotate to accommodate itself to the fluid motion as
shown. The rate of energy dissipation in the cell occupied by the particle
may be compared with that experienced by the undisturbed fluid when
sheared at the same rate in the absence of the particle. This comparison
provides a reliable estimate of the relative viscosity of a suspension of par-
ticles in fairly concentrated systems.
It should be emphasized that although cell models of this type apparently
give a satisfactory approximation of the actual average flow pattern close
to the particles in a real physical system, they cannot be expected to be valid
for points close to the imaginary cell boundaries. Thus, such models will
not predict effects such as convective transfer of fluid from one cell to others.
For this purpose an approximate solution of the boundary value problem
involved, using a technique like the reflection procedure, leads to more
satisfactory results.
In addition to the significant simplification arising from the linearization
of the equations of motion, and from further simplification of the geometry
involved, it is necessary in many cases to make other additional assumptions
in order to bring the analysis to fruition. These include such items as the
assumption of uniformly sized particles, neglect of Brownian motion of
the particles, and except for gravity, the absence of extraneous forces, such
as electrostatic or interfacial forces. Good agreement of theoretical results
with available experimental data is usually found in situations where these
assumptions are warranted.
This book begins with a discussion of the basic equations of motion and
their limitations. The behavior of single particles is then considered. Sub-
sequently, the effect of hydrodynamic interactions among individual par-
ticles and between particles and walls is examined. Finally, these effects
are superposed to treat the behavior of particle assemblages. The basic
principles seem clearly established. They serve to illustrate the versatility
1-1 Definition and Purpose 7

of the creeping motion equations for handling problems where the viscous
effects in particulate motion are dominant.
The treatment in the following chapters is largely mathematical.
Experimental data are studied mostly in order to verify that the solutions
obtained apply to real physical situations. In all cases treated, however,
emphasis is on the hydrodynamic factors involved.
Familiarity with basic vector calculus is assumed. To a lesser extent,
some knowledge of tensor and polyadic analysis is also required. Those
not familiar with these topics and with the other aspects of engineering
mathematics involved, should consult standard treatises. Bird, Stewart,
and Lightfoot5 provide a good introduction to the field of transport phe-
nomena. A brief exposition of the vector and polyadic concepts useful in
studying the theory of diffusion, fluid dynamics, and related topics is given
in Drew'sl3 handbook. Special mention should also be made of Gibbs'
text* on vector and polyadic analysis. Aris 1 has published a useful text
which devotes special attention to the application of vectors and tensors
to problems in fluid mechanics.
At the end of this book, Appendix A provides a useful tabulation of the
properties of several important curvilinear coordinate systems. Appendix
B briefly summarizes the vector and tensor notation employed in this book.
The following chapter contains references to ,a number of texts in the
general field of hydrodynamics. Closest to the present treatment in objective
is the first section of the classic treatise by Oseen 35 , which considers, among
other things, situations involving the slow motion of particles in the presence
of bounding walls. t
As for flow through assemblages of particles, numerous references are
cited in Chapter 8. Scheidegger's41 monograph on the physics of flow
through porous media offers a good survey of both English and foreign
research papers and texts, covering fundamental subjects and especially
systems where the particulate medium is not dispersed.
*J. W. Gibbs, and E. B. Wilson, Vector Analysis, Dover reprint (New York: Dover,
1960).
tOther books devoted exclusively or almost exclusively to low Reynolds number flows
are H. Villat, Lerons sur les Fluides Visquex (Paris: Gauthier-Villars, 1943); W. E.,
Langlois, Slow Viscous Flow (New York: Macmillan, 1964); O. A. Ladyzhenskaya, The
Mathematical Theory of Viscous Incompressible Flow (New York: 'Gordon and Breach,
1964). In addition to these books, the following chapters in other books contain extensive
summaries of various facets of low Reynolds number flows: I-Dee Chang in "Handbook
of Engineering Mechanics" (W. Flugge, ed.). New York: McGraw Hill, 1962; C. R.
Iin~worth in "Laminary Boundary Layers" (L. Rosenhead, ed.). London: Oxford,
1963; R. Berker in "Encyclopedia of Physics: Fluid Dynamics II" (S. Flugge and C.
Truesdell, eds.) Vol. 82. Berlin: Springer-Verlag, 1963; p, A. Lagerstrom in "Theory of
Laminar Flows" (F. K. Moore, ed.). Princeton, N. J.: Princeton University Press, 1964;
H. Brenner in "Advances in Chemical Engineering" Vol. 6 (T. B. Drew, J. W. Hoopes,
Jr., and T. Vermeulen, eds.), New York: Academic Press, 1966.
8 Introduction

The remainder of the present chapter is devoted to subjects of general

interest in the study of particulate matter; namely, a brief historical review
of how the subject has developed and an outline of the various applications
in science and technology which supply motivation for research in this area.

1-2 Historical Review

The general history of experimental hydraulics and of theoretical hydrody-

namics is interestingly treated by Rouse and Ince 39 • The treatises by Basset 4
and Dryden, Murnaghan, and Bateman 14 also include numerous references
to the work of early investigators, especially with regard to laminar flows.
Although practical hydraulics had its origins in antiquity, scientific attention
to flow relative to particulate media began scarcely one hundred years ago.
We do not intend to repeat the names, or summarize the work of all who
have contributed to this field, but simply to call attention to those scientists
and engineers who have been especially identified with fluid flow relative
to particles and porous systems. Thus, we shall not enter into detail regard-
ing the work of such leaders as Navier (1785-1836) who, together with
Stokes, is credited with the formulation of the equations of motion used
in hydrodynamics; Poiseuille (1799-1869), whose careful experiments on
laminar flow established the law which bears his name; Lord Rayleigh (1842-
1919), who investigated many phases of hydrodynamics, including energy
dissipation; and Boussinesq (1842-1929), who made comprehensive
mathematical treatments of laminar flow in pipes and channels. References
to pertinent investigations of early contributors not discussed here do appear
at appropriate sections of later chapters.
Studies of flow through porous media first engaged the attention of
several engineers of the famous Corps des Ponts et Chaussees during the
second half of the nineteenth century. Henry P. G. Darcy (1803-1858),
a native of Dijon, after being educated in Paris, returned to the city of his
birth where he became director of public works. There his major accom-
plishment was the design and execution of a municipal water supply system.
This system not only functioned admirably, but also gave rise to a series of
researches which he conducted on the flow of water through sand bed filters.
He published the results of these studies along with much other information
on the development of water supply systems in 1856 12 • (See Fig. 1-2.1.)
The law which Darcy discovered, namely, that the rate of flow is propor-
tional to pressure drop through a bed of fine particles, bears his name. It is
widely employed for investigating the behavior of all types of water flow
through porous media, such as underground flow to wells, flo'w in soils
being irrigated, and the permeability of dam foundations. In addition, the
flow of oil in underground substructures has been found to follow Darcy's
1-2 Historical Review 9

LES

FONTAINES PUBLIQUES
DE LA VILLE DE DIJON
EXPOSITION ET APPLICATION

DES PRINCIPES A SmYRE ET DES FORMULES A EMPLOYER

DANS LES QUESTIONS

DISTRIBUTION D'EAU
IGYbCllIUllI

PM! UN APPENDICE REL1TIF lUX FOURNITURES D'L\U DE PWSlIlJII ¥ILLES

AU FlLTRAGE DES EAUX

... IA FABRIU.TJON DIS TVTlUl DE 101fT!. DE 'LOlli. PI TOLE n DI alTUM.

P. .

HENRY DARCY
INiPIECTEUR GEJIIEIIoAL DES POIITS F.T CHAUSSW.

1.1 na.. qllililli dl!'~ .,.,; NDIIlIIe del I'~

plu:.ilil NOll! dt.1riLoJftls"u.ae 'l'iiR, II
Iralllintl pIIIl.·i.D....... q ...•• _trelaii''' ...
ill. boil_ toIa.lQte .... _ _ _ et_UIi••"I, rtll .............
.'Ju"". ....._'n,.i.rYaI
qllli .....' It
ri_ ....... lei . . . . •

It<:ideall .., InqUil. rei

dft 1oa1.l.1_. del ri\"itNl," , . . . . . . . . . . . . .
htuloia_t_r.... rrl1el . .·OIIft ~iYe
...a.l.., ••,...11'
e.......'r. . .IItn ................
• ..,il . . . . . . . .,.puiu
~

d'uu 1<iIl.Hllt dn _ _
PI Jr"lUt. Bill . • '·.t~lI; ,..,al, . . . ._ _ • nIl. p. Ill.

-.-
PARIS
VICTOR DALMONT, EDITEUR,
SurffSSf'ur de r.lfi"-I~J' y'r 0,1.011,

• • •1 .e.
LIIRAIRF. DES CORPS .M.'tRIAUl DF.S rONTS IT CRA.USStlS IT DES liNE!",
A ••••• I •• , •••

1856

Figur. 1-2.1. Cover of Darcy's book

law, and a unit of permeability designated as the darcy is quite generally

used in the oil industry today, Literally hundreds of studies have been devot-
ed to experimental determination of the Darcy permeability for different
types of porous media. Darcy's health had been failing for some time when
he published his treatise. He died in 1858 while engaged on a project involv-
10 Introduction

ing the movement of running water in open canals. This work was later
completed by his capable pupil and colleague, Bazin.
Another member of the Corps, Arsene J. E. J. Dupuit, continued Darcy's
studies on flow through porous media, and published his researches in a
second edition 15 of a book on the transportation and distribution of water.
In that edition, Dupuit made the first comments on the capacity of a stream
to transport sedimentary particles in suspension. Other successors in the
Corps carried forward studies in transport of sediment by rivers.
The first study of geometrical arrangements of spheres was made in
1899 42 by Slichter, who was especially interested in the flow of water through
soil. He was led to a study of sphere arrangements in order to reduce the
hydraulics of a complex soil to an idealized system. He also made the first
attempt to derive a porosity function for beds of uniform spheres, by making
the simplifying assumption that the average cross-sectional area for flow
would be triangular. By applying the equivalent of Poiseuille's law for flow
through a tube of triangular cross section, he obtained an appropriate per-
meability equation. Slichter's basic formula was proved inadequate because
his generalized model was oversimplified. Nevertheless, his treatment pro-
vided the starting point of many subsequent studies.
One of the most fruitful ideas leading to the type of permeability equa-
tions currently employed for predicting resistance to flow through porous
media was the utilization of the empirical hydraulic radius concept 6 by
Blake (1922), an American chemical engineer. It had been found earlier that
resistance to flow through passages of various noncircular conduits could
be brought into agreement with that for circular pipes by employing the
hydraulic radius to characterize the respective cross sections. This quantity
is defined by the expression,

Hydra ul ic radi us = -=:.c~;'­

cross-sectional area normal to flow
wetted perimeter

An equivalent form is
volume filled with liquid
Hydraulic radius = wetted surface

Blake was apparently the first to realize that a packed bed might be regarded
as a single pipe with a very complicated cross section, and that the inter-
stitial volume could be divided by the wetted packing area to obtain a
hydraulic radius. Kozeny (1927)26 extended this treatment. Subsequently,
Carman (1937)8 suggested modifications which resulted in a semiempirical
equation, now widely used to correlate data for flow through packed beds,
see Eq. (8-5.10).
During the same period that the initial experimental studies were being
conducted on flow through porous media, attention was first devoted to
1-2 Historical Review 11

the theoretical aspects of flow in particulate systems. The earliest study on

resistance of a solid body moving relative to a fluid, in which viscosity was
taken into account, was published by Sir George G. Stokes (1819-1903).
Born in Skreen, Ireland, Stokes received his education at Cambridge. He
became a professor and remained in Cambridge the remainder of his life,
making many important contributions to theoretical physics. His early
interests centered around the subject of hydrodynamics, which was then
rather neglected in the general researches being undertaken. His early papers,
mostly in theoretical hydrodynamics, appeared in the Cambridge Philo-
sophical Transactions. Of greatest interest to us is the paper in which he
linearized the general equations of motion of a viscous incompressible fluid
and thus obtained a time-dependent form of the creeping motion equations.
He applied these linearized equations to estimate the frictional damping
of the motion of a spherical pendulum bob due to air resistance (1851)41.
As the frequency of oscillation of the pendulum approaches zero, it moves
with essentially constant velocity through the air. The resistance to fall
of such a body of spherical shape was developed in this paper, and the rela-
tionship he discovered is known as Stokes' law, see Eq. (2-6.3.) It has been
found to apply to the sedimentation of all manner of small particles which
are moving slowly. The derivation he presented is an elegant one mathe-
matically and appears in many text books on hydrodynamics. It applies
in the situation where the particles are far enough apart so that the motion
of each one of them is not affected by the motion of its neighbors. Stokes
made many other important contributions to the Cambridge school of
mathematical physics before his death at the age of eighty-four.
Lamb's3o classic treatise on hydrodynamics which first appeared in 1879,
and subsequently passed through six editions, contains much historical
and technical information on the development of solutions of the creeping
motion equations, though it is devoted mainly to potential flows. Worth
special mention also is the solution of the steady translation of an ellipsoid
parallel to a principal axis in a viscous. fluid by Oberbeck (1876)34.
As for the influence of bounding surfaces on the motion of a single body,
H. A. Lorentz (1896)31, following the method developed by Stokes (1845)46,
determined the motion of a sphere in the presence of a plane wall. The tech-
nique used involves "reflection" of the original motion produced by the
body from the surface of the wall and back again to the body. Soon after,
R. Ladenburg (1907)29 exploited the same technique to determine the effect
of a cylindrical tube on the axial motion of a centrally positioned sphere.
The same method was employed by the Polish mathematician M. von
Smoluchowski to determine the effects of hydrodynamic interaction be-
tween two spheres moving in a viscous fluid (1911)43. Shortly afterwards, he
again employed the "method of reflections" to study the sedimentation of
an assemblage of spheres (1912)44. E. Cunningham (1910)10 considered the
12 Introduction

sedimentation of a cloud of particles in a closed vessel, employing a cell

model. His estimate of the decrease in terminal settling velocity due to par-
ticle interaction was based on the approximate supposition that each particle
moves, on the average, as if it were contained in a rigid spherical envelope,
of radius equal to half the distance to its nearest neighbors.
These basic methods are still of considerable utility. For use of the meth-
od of reflections see Chapters 6 and 7; various cell models for sedimentation
problems are developed in Section 8-4.
Many of the early contributions to low Reynolds number hydrodynamics
are summarized in the book by the Swedish physicist, Carl W. Oseen (1927)35.
Of special interest are the contributions of Hilding Faxen, his coworker,
whose early papers are discussed by Oseen. Later researches of Faxen are
mentioned in various chapters of the present book. Faxen's work has been
of signal importance in advancing the theory of particulate systems.
The problem of finding the disturbance caused by the presence of a par-
ticle suspended in an otherwise uniform shearing flow was undertaken
somewhat later than that for uniform motion. It is interesting that this
problem was first solved as the doctoral dissertation of Albert Einstein
(1879-1955). Einstein was born in Germany, but studied physics at the
Zurich Polytechnic Institute. When he obtained his doctorate in 1.905, he
had become a Swiss subject. His thesis was concerned, among other things,
with a new method for determining the size of molecules of chemical sub-
stances. In order to accomplish this, he developed a theory for the resistance
to shear of a suspension of small spherical particles immersed in a continu-
ous fluid, as a model for large molecules in solution. He showed theoreti-
cally that the apparent increase in viscosity of the suspending liquid could
be related to the volumetric concentration of solid particles (or solute mole-
cules) by a simple proportionality constant (1906, 1911)16. The Einstein
law for suspension viscosity has been used since as the basis for almost all
theories of the behavior of suspensions in shearing fields of flow. (See Section
9-6 for a discussion of his findings on the size of the sugar molecule.) Like
Stokes' law, Einstein's applies to the case where the suspended particles
are far enough apart on the average that their motion is not influenced by
mutual interaction of the disturbances produced by individual particles.
As is well known, Einstein's interests soon turned to relativity and quantum
theory. He spent the last years of his life in the United States, of which
he became a citizen.
Einstein's viscosity theory for suspensions of spherical particles and
Jeffery's25 extension of it to particles of ellipsoidal shape were further
extended in several directions in a series of papers by Guth and his coworkers
at the University of Vienna in 1936. Guth 21 has given a brief account of
these studies. Especially interesting is Guth and Simha's paper22 which
1-3 Application in Science and Technology 13

considers wall effects and interaction between particles on the apparent

viscosity of suspensions.
Developments during the last thirty years have been more varied and,
in common with other branches of science, have become more numerous.
Many of these developments were motivated either by other scientific or
technological problems or because they found application in applied fields.
It is believed that these advances are of interest not only to those interested
in engineering applications, but also, hopefully, to the research worker,
who may find in these advances sources of ideas from apparently unrelated
disciplines.

1-3 Application in Science and Technology

The impact of the science of small particles in many directions has been
well documented by DallaValle l1 , whose book Micromeritics covers many
aspects of particle technology other than those of a purely hydrodynamic
nature. It includes such topics as the geometry of packing of particles, size
measurement, sieving and grading, as well as electrical, optical, sonic, and
surface properties of particles. In the survey which follows, we ~onsider
mostly those applications which relate in some way to the basic hydrody-
namic theme of this book. We consider engineering applications first, in
keeping with the idea that hydrodynamics may be approached as an engin-
eering science in its own right, rather than as merely furnishing peripheral
information to the physical sciences.

Chemical engineering
Perhaps the most obvious application of particle dynamics occurs in
situations where mutual interaction of particles can be neglected, so that
the fundamentals of single particle motion apply. Dust and mist collection
from dilute suspensions of finely divided solids, and from liquids in gases,
furnish simple practical examples of such idealizations 36 • One common
application is in the elimination of atmospheric pollution, as in the cleaning
of ventilating air or in the reduction of industrial health hazards, such as
toxic fumes from chemical operations. The recovery of valuable by-products
from the dusts leaving dryers and smelters represents another important
application. Finally, many industrial operations involve the production
of a powdered product which must be separated from the gas or liquid in
which it is suspended, as, for example, the spray drying of milk and soap
and the manufacture of zinc oxide and carbon black. Often, very fine par-
ticles are involved, permitting the laws of slow viscous flow for bodies falling
in infinite media to be applied.
14 Introduction

Gravitational and centrifugal sedimentation of dense suspensions of

fine particles, in the form of slurries in liquids, constitutes a closely allied
application, which arises in many industries. Concentrating and thickening
devices are usually still designed on a fairly empirical basis. But it seems
reasonable to suppose that their design can be made more rational as we
acquire further knowledge of the laws of slow viscous flow.
In the more concentrated range, systems involving towers packed with
special shapes or crushed solids have been widely used in chemical process-
ing for many years. Such towers serve as contacting devices: to bring to-
gether gases and liquids for the purpose of absorption or desorption; to
contact liquids with liquids for extraction; and to transfer heat to or from
gases, as in kilns and gas producers. In some of these processes a moving
bed of solids is employed, as in a blast furnace.
The past twenty-five years have seen numerous developments of fluid-
particle operations in the· chemical process industries51, 28,37, stemming
from the commercial exploitation of continuous fluids-solids processing
in the catalytic cracking of petroleum to produce high octane-number gaso-
line. An obvious development employed a moving bed in which a pelleted
catalyst flowed downward through a reactor against an ascending stream of
vapor. The spent catalyst was then subsequently conveyed either by elevator
or pneumatically to a regenerator. These endeavors stimulated further
developments in moving bed techniques 50. Among the more recent applica-
tions of interest may be mentioned the moving bed oil-shale retort and the
use of moving bed reactors for the production of uranium tetrafluoride.
A more striking application followed further developmental work, in
which the catalyst was not only conveyed pneumatically, but suspended by
vapor in the reactor and regeneration chambers. This led to the fluidized-bed
technique. In this contacting method, catalyst is used in powder form. Small
particles suspended by the proper velocity of vapors passing through them
behave much like true liquids and can be transported or contacted with
vapors, as if they were homogeneous fluids.
Figure 1-3.1 illustrates a modern fluid-solid catalytic cracking unit.
In these processes, oil-feed vapors carry hot regenerated catalyst particles,
varying from 20 to 180 microns, into a fluidized bed in the reactor. The
gasified feed is cracked to produce gasoline. The cracked components pass
to a separation system and the carbon-coated catalyst passes to the regen-
erator, flowing by gravity, just as if it were a liquid. The regenerator bed
is kept in a fluid condition by means of an air stream which burns off the
deposit on the catalyst, and the hot regenerated catalyst is again ready for
contact with oil vapors. In a modern unit of the type depicted, 5000 tons/hr
of catalyst particles contact the oil fractions being cracked. This is a solid
moving process comparable to a major mining operation, such as that con-
ducted by the Kennecott Copper Co. in Utah (300,000 tons/day). Yet it is
1.3 Application in Science and Technology 15

Flgur. 1-3.1. A fluid bed

catalytic cracking unit.
(Courtesy of M. W. Kel-
log Co.)

handled by only a handful of men simply watching control instruments,

since all the particles are transported as suspensions and maintained in the
fluidized state in the reactor and combustion zones.
Many other applications have followed the development of fluidization
in the petroleum industry. Among the more interesting are metallurgical
applications, such as the use of a fluidized iron ore bed to prepare iron by
direct reduction with hydrogen, and the fluid-bed roasting of pyritic ores
and limestone. The application of the fluid-bed technique to coal gasification
in the production of water-gas appeared in the patent literature as early
as 1922 as a development by Winkler. A number of plants were built using
this process. It seems likely that this technique will find further wide
industrial application.
16 Introduction

The dynamics of particles suspended in fluidized beds is very complicat-

ed Sl • It is too much, as yet, to expect theoretical relationships to predict their
behavior in detail. As noted by Zenz and Othmer 5 1, however, Reynolds
numbers of particles encountered in catalytic cracking systems normally
lie in the range N Re = 0.01-5.0. The equations for slow viscous flow ought
therefore to provide a meaningful basis for interpretation. In this range,
pressure drop relationships of the type developed later in this book (see
Chapter 8) are in very good agreement with data obtained for the case of
particulate fluidization, where smooth uniform expansion of the bed of
particles occurs. This type of fluidization normally occurs only when the
continuous fluid is a liquid. When gases are employed, aggregative fluidi-
zation usually results. Here, nonuniform bed expansion accompanied by
bubble formation occurs, making quantitative correlation difficult. This
area presents a continuing challenge for both theoretical and experimental
work.
The flow properties of many plastic materials and paint products involve
the hydrodynamics of shearing flows of suspensions, as do also the flow
and aggregating properties of paper-making fibers 45 •

Civil engineering
There are many fluid-solids dynamical processes controlling man's
environment, which consists, roughly speaking, of earth, water, and air.
Earth is to a large extent composed of soil formed from sedimentary de-
posits of minerals like sand and clay. Soil has many of the properties of
packed beds, but is more complicated owing to variations in particle size
and shape as well as to nonisotropic distribution in strata.
The infiltration of water and oil through soil finds application in many
fields, typical of which are petroleum engineering, soil mechanics, ground
water hydrology, and sanitary engineering. Collins 9 presents a unified treat-
ment of flow through porous media, starting with the basic physical char-
acteristics of porous materials and leading ultimately to a treatment of some
of the more complex flow problems, among these being simultaneous laminar
flow of miscible fluids and flow accompanied by a phase transition.
More dilute systems, where the particles do not touch each other, occur
in the case of silt which is carried by rivers. This silt may end by being dis-
tributed over wide areas of ocean bottom or confined to the formation of
deltas as in the case of the Mississippi and the Nile. Such silt deposits often
impart important agricultural advantages, but may, on the other hand,
present problems when they accumulate in reservoirs behind dams or in
navigation channels. The basic hydrodynamic principles involved are much
the same as those encountered in the problems of the process industries.
A different category of problems is that involving dust, smoke, and
mist 20 , where the particles are smaller and are suspended in air. Though
the basic hydrodynamic laws are the same for aerosols and hydrosols, the
1·3 Application in Science and Technology 17

former have many special properties due to their very fine state of subdivi-
sion. Many problems of industrial hygiene derive from contamination of
the atmosphere by smoke, bacteria, and other pollutants.
Mining engineering
Knowledge of the behavior and characteristics of suspensions of small
particles in liquids is important in the separation of minerals. After ores
are crushed to prescribed sizes, various flotation and settling processes are
often employed for separation. Indeed, the movement of particles through
fluids plays a role in all mineral-dressing processes 19 •
Flow and seepage of water and solutions of minerals through permeable
soil structures was important in early prehistory in forming deposits of
water-soluble materials, and knowledge of such phenomena is important
to the mining engineer. Control of the rheological properties of oil well
drilling muds is important in petroleum production.
Problems of mine ventilation and prevention of dust explosions from
combustible materials, such as coal dust, depend upon a knowledge of the
dynamics of dilute suspensions of particles in the atmosphere.
Physical sciences
The flow properties of many disperse systems (suspensions, emulsions,
gels, liquid sprays, foams, and powders) are important in their scientific
study and characterization 24 , 17. Innumerable rheological investigations
on suspensions of colloidal materials and macromolecules have been carried
out with the objective of obtaining information as to the physicochemical
nature of the particles themselves. Disperse systems, either as emulsions
or dispersions, comprise a large group of materials of industrial importance 32 ,
including cement, plaster, paint, ink, and paper coatings.
The falling-ball viscometer, which consists of a sphere falling in a circu-
lar cylinder, represents one of the standard fundamental methods 18 for
determination of fluid viscosity. The exact theory of this apparatus is a
direct application of the appropriate particle-wall interaction dynamics.
Biology
Normal blood is a suspension of particulate matter (red cells, white cells,
and platelets) in a continuous medium, the plasma. Of the particulate matter
present in whole blood, the red cells dominate in volume, occupying on the
average 40 volume per cent. A comprehensive review of the rheological
properties of blood, blood plasma, and certain mucous fluids has been given
by Merrill and Wells 33 • A more detailed treatment is given in the symposium
Flow Properties of Blood, edited by Copley and Stainsby.* The biologic
functions of blood, plasma, and body fluids are shown to be intimately
related to their rheological behavior, which appears far more complicated
·A. L. Copley and G. Stainsby (eds.), Flow Properties of Blood (New York: Pergamon
Press, 1960).
18 introduction

than man-engineered systems, exhibiting non-newtonian behavior when

tested· in a special viscometer operating on whole blood samples before
clotting occurred. Suspensions of red blood corpuscles in brine, however,
appear to exhibit newtonian behavior up to higher concentrations (see Ta-
ble 9-6.3).
As Merrill and Wells point out, this is but one example of increasingly
effective collaboration between engineers, physicians, and scientists on
medically important problems.

Earth sciences

Soil permeability has already been discussed under engineering appli-

cations. It also has an important bearing on various agricultural problems
involving such matters as crop yield, soil accumulation, and vegetation as
related to drainage of water supplied by natural and irrigation sources.
Geologists and geographers have always been concerned with the
formation of sediments and various types of rock formation produced by
them. Twenhofel's49 treatise provides a source of basic information of many
aspects of historical sedimentology. Sedimentologists have been much con-
cerned in recent years with the role of turbidity currents.
Such currents consist of suspensions of sediment, flowing along the
inclined floor of a quiescent body of clear water, propelled by gravitational
forces resulting from the greater density of the sedimentary suspension.
Workers on recent marine sediments are convinced of the important role
of turbidity currents in accounting for many topographical features of the
continental shelves, such as submarine canyons. One of the largest of these
canyons is that formed by the Hudson River. This canyon extends for more
than 100 miles across the continental shelf, originating at the entrance to
New York Harbor. Its topological structure probably presents an under-
water panorama as impressive and spectacular as the familiar Palisades,
a few miles up the river. Many geologists believe that these canyons were
created by heavy mud flows during periods when the glacial epic lowered
the ocean level by several hundred feet. It is generally believed that the fine
earth sediments of particles on the continental shelf were churned by storm
waves into suspensions which then flowed down the continental slopes at
great velocities without mixing with the water above. It is clear that, though
inertial and turbulent effects are involved, the slow motion and shear pro-
perties of these suspensions will also be important 21 •
These turbidity currents display similarities to other natural phenomena.
Thus, Kuenen likens the part played by the silt on the shelf to that of a ring-
ing voice bringing down an avalanche which gains in force as it advances.
Another interesting example is presented by the clouds which sometimes
1-3 Application in Science and Technology 19

emerge from volcanic lavas_ These are so heavily laden with ashes that they
flow down the slopes of the mountain with great velocities 7 _ Suspensions
of volcanic ash in water can also move rapidly down mountain sides, thus
behaving similarly to fluidized beds_
Quicksand provides a dramatic illustration of a fluidization process
found in nature_ Here, finely divided particles of sand are suspended in the
fluidized state by the passage of an underground spring of water through
them_ The density of the suspension is not very much greater than water,
but the suspended material vastly increases the viscosity of the water, making
the resistance of the suspension to shear very high_ Hence the difficulty in
swimming or other movement experienced by those trapped in such
a deposit_
Another interesting form of fluid particle interaction in nature is the
transportation of particles of solid by saltation-a series of leaps and bounds,
alternating with impacts on the ground surface_ This type of action is respon-
sible for the formation of snow drifts, sand dunes, and various forms
of beach erosion. Bagnold3,2 has described and analyzed many problems
related to the formation of ripples and sand ridges and the growth and
movement of sand dunes. In the case of sand storms, particles are not lifted
high in the air and carried over great distances, as normally supposed, but
move in a relatively dense suspension with saltation.
Meteorologists are concerned with more dilute suspensions, such as are
involved in raindrop formation 38 and in the presence of fine particles in the
atmosphere. Here, problems are encountered which involve the motion
of different sizes of droplets and the possibilities of their agglomeration.
Hydrodynamic forces play an important role in these phenomena.
To summarize, there is a whole world of fine particles whose size ranges
from those of small molecules to ordinary dust and sand visible with the
naked eye. Protein molecules, viruses, synthetic polymers, colloidal particles,
cosmic dust, soot, and fly ash particles from atomic tests are part of a do-
main in which these materials have characteristic size, shape, and properties.
Observation of the particles themselves 48 by modern tools, such as the elec-
tron microscope, is one phase of their study. Another is their mutual
interaction and their movement in the presence of boundaries. In this area
the study of hydrodynamics presents a remarkably unified understanding
of much of their behavior. True, their hydrodynamic behavior often in-
volves inertial and turbulent flow; other factors, such as particle collisions,
electrical charges, and surface properties, may complicate matters further.
A study of the effects produced by viscous forces, such as are exerted in
slow flows, will, however, usually be advantageous and will very often
account to a major extent for the phenomena observed.
20 Introduction

BIBLIOGRAPHY

Aris, R., Vectors, Tensors, and the Basic Equations of Fluid Mechanics. Engle-
wood Cliffs, N. J.: Prentice-Hall, Inc., 1962.
2 Bagnold, R. A., Phil. Trans. Roy. Soc. (London) 249A (1956), 235.
3 - - , The Physics of Blown Sand and Desert Dunes. London: Methuen,
1941.
4 Basset, A. B., A Treatise On Hydrodynamics, 2 vols. (1888); New York: Dover,
1961.
5 Bird, R. B., W. E. Stewart, and E. N. Lightfoot, Transport Phenomena, New
York: Wiley, 1960.
6 Blake, F. C., Trans. Amer. Inst. Chern. Engrs. 14 (1922),415.
7 Burgers, J. M., Proc. Koningl. Akad. Wetenschap. 45 (1942), 9.
8 Carman, P. C., Trans. Inst. Chern. Engrs. (London) 15 (1937), 150. See also his
book Flow of Gases through Porous Media. New York: Academic Press, 1956.
9 Collins, E. R., Flow of Fluids through Porous Materials. New York: Reinhold,
1961.
10 Cunningham, E., Proc. Roy. Soc. (London) A83 (1910),357.
11 DallaValle, J. M., Micromeritics-The Technology of Fine Particles, 2nd ed.
New York: Pitman, 1948.
12 Darcy, H. P. G., Les Fontaines Publiques de la Ville de Dijon. Paris: Victor
Dalmont, 1856..
13 Drew, T. B., Handbook of Vector and Polyadic Analysis. New York: Reinhold,
1961.
14 Dryden, H. L., F. P. Murnaghan, and H. Bateman, Hydrodynamics, reprint,
Bull. No. 84 Nat. Res. Counc. New York: Dover, 1956.
15 Dupuit, A. J., Traite theoretique et practique de fa conduite et de la distribution
des eaux. Paris: 1865.
16 Einstein, A., Ann. Physik 19 (1906), 289; with correction 34 (1911) 591. For
English translation of these and related papers, see Einstein, A., The Theory
of Brownian Movement, New York: Dover, 1956.
17 Everett, D. H., and F. S. Stone, The Structure and Properties of Porous Ma-
terials. New York: Academic Press; Butterworth's, London, 1958.
18 Fox, T. G., S. Qratch, and S. Roshaek, in Rheology-Theory and Applications,
Vol. I, ed. F. R. Eirich. New York: Academic Press, 1956, Chap. 12.
19 Gaudin, A. M., Principles of Mineral Dressing. New York: McGraw-Hill, 1939.
20 Green, H. L., and W. R. Lane, Particulate Clouds, Dusts, Smokes, and Mists.
London: Spon, 1957.
Bibliography 21

21 Guth, E., Proc. International Congr. Appl. Mech. (1938), pp. 448-55.
22 - - , and R. Simha, Kolloid. Z. 74 (1936), 266.
23 Happel, J., Trans. N. Y. Acad. Sci. 10 (1958), 404.
24 Hermans, J. J., Flow Properties of Disperse Systems. Amsterdam: North-
Holland Publishing Co.; New York: Interscience, 1953.
25 Jeffery, G. B., Proc. Roy. Soc. (London) AI02 (1922),161.
26 Kozeny, J., Sitz-Ber. Wiener Akad., Abt. IIa, 136 (1927), 271; see also
Hydraulik. Wien: Springer, 1953.
27 Kuenen, P. H., Society of Economic Paleontologists and Mineralogists, Special
Publication No.2, pp. 14-33, November, 1951.
28 Leva, M., Fluidization. New York: McGraw-Hill, 1959.
29 Ladenburg, R., Ann. d. Phys. 23 (1907), 447.
30 Lamb, H., Hydrodynamics, 6th ed. Cambridge: Cambridge Univ. Press, 1932;
New York: Dover, 1945.
31 Lorentz, H. A., Zittingsverl. Akad. van Wet. 5 (1896),168; neu bearb.: Abhandl.
theoret. phys. 1 (1907), 23.
32 Mill, C. C., Rheology of Disperse Systems. New York: Pergamon, 1~59.
33 Merrill, E. W., and R. E. Wells, Jr., Appl. Mech. Reviews 14 (1961), 663.
34 Oberbeck, H. A, Crelle, 81 (1876), 62.
35 Oseen, C. W., Hydrodynamik. Leipzig: Akademische Verlag, 1~27.
36 Perry, J. H., Chemical Engineers' Handbook, 3rd ed. New York: McGraw-Hill,
1950.
37 Reboux, P., Phenomenes de Fluidization. Paris: Association Franc;aise de
Fluidization, 28, rue St. Dominique, 1954.
38 Richardson, E. G., Aerodynamic Capture of Particles. New York: Pergamon,
1960.
39 Rouse, H., and S. Ince, "History of Hydraulics," State Univ. Iowa Inst. Hy-
draulic Res. 1957.
40 Shapiro, A H., Shape and Flow-The Fluid Dynamics of Drag. AnChor Books:
Science Study Series. New York: Doubleday, 1961.
41 Scheidegger, A E., The Physics of Flow through Porous Media, 2nd ed. New
York: Macmillan, 1960.
42 Slichter, C. S., "Theoretical Investigation of the Motion of Ground Waters,"
U. S. Geological Survey, 19th Ann. Rep., Part 2 (1899), pp. 301-384.
43 Smoluchowski, M., Bull. Intern. acado polonaise sci. lettres, lA (1911), 28.
44 --,5th Intern. Congr. Math. 2 (1912),192.
45 Stccnberg, B., and B. Johansson, Svensk Papperstidning 61 (1958), 696.
22 Int,oduction

46 Stokes, G. G., Trans. Cambro Phil. Soc. 8 (1845), 287.

47 - - , Trans. Cambro Phil. Soc. 9, pt. II (1851), 8.
48 Turkevich, J., Amer. Scientist, 47 (1959),97.
49 Twenhofel, W. R., Treatise on Sedimentation. New York: Dover, 1960.
50 Vener, R. E., Chern. Eng. 62 (1955), 175.
51 Zenz, F. A, and D. F. Othmer, Fluidization and Fluid Particle Systems. New
York: Reinhold, 1960.