2.
3 Particle Properties 25
Counter flow fractionation
Sieve fractionation
Cross flow fractionation
Figure 2.12 Sieving, counterflow, and cross-flow fractionation.
Separation means the separation according to different phases, such as solid
particles and liquid. In general, all three aggregate states can be separated from
each other. In each case, one phase is defined as the disperse phase (e.g. parti-
cles), which is distributed in a continuous phase (e.g. liquid). During the cake
filtration process, the solid and liquid phases are changing their character. Ini-
tially, when the cake is formed, the particles are dispersed in the liquid. After the
end of cake formation, the solid particles represent the continuous phase in the
form of an interconnected porous network and the liquid to be removed from
this cake structure is distributed as a disperse phase in the cakes pores.
Sorting means the subdividing of a particle collective according to other phys-
ical characteristics, such as a kind of material, surface properties, magnetic or
electric properties, color, shape, density, or others. If density is the criterion, the
particle collective is divided in a heavy component and a light component.
Fractionation means the subdividing of a particle collective according to the
particle size into partial fractions smaller and greater than a cut size xt , as demon-
strated in Figure 2.12 on the examples of sieving, counterflow fractionation, and
cross-flow fractionation.
Sieving separates particles directly according to their size. Counter and cross
flow fractionation is based on different particle settling velocities in gas or liquid
because of different particle mass and thus size. The density of all particles is pre-
supposed as the same, and between solids and fluid, there must exist a density
difference. Constant density of all solids is essential because different solid den-
sities would lead to the situation that a great and light particle would settle with
the same velocity than a small and heavy particle.
A further example for an excellent fractionation device, which is based on cen-
trifugal sedimentation, is given by the hydrocyclone in Figure 2.13.
The tangential feed of the fixed cyclone first leads to a potential vortex of the
slurry with increasing centrifugal acceleration toward the center. This is due to
the condition of continuity and the tangential velocity of the spiral flow and thus
the centrifugal acceleration toward the center must increase. The maximal accel-
eration is reached at the radius of the vortex finder. At this position, the cut size
xt is determined. The coarse particle fraction, which is separated at the cyclones
wall, is discharged in the concentrated form downward through the underflow
nozzle. The main part of the liquid including the fine particle fraction is not able
to leave the cyclone through the throttled underflow nozzle but must change its
26 2 Slurry Characterization
Overflow pipe
(vortex finder)
fine material “F”
Feed pipe
(evolute)
feed material “A”
Secondary vortex
(rigid body vortex) Primary vortex
(potential vortex)
Underflow nozzle
(apex)
coarse material “G”
Figure 2.13 Hydrocyclone.
flow direction and leaves the cyclone upward through the overflow pipe. Here,
the liquid rotates in the form of a rigid body.
The coarse fraction in principle contains all particles greater than a cut size
xt and the fine fraction contains all particles smaller than xt . Any particle size
between xmin and xmax can be defined as cut size. One can distinguish between sta-
tistically defined cut sizes from particle size distributions and physically defined
cut sizes from physical separation conditions. To characterize the results of a frac-
tionation or a more or less complete separation, the information is needed, which
share of the particles in the feed is found after the process in the coarse and fine
fraction. A first answer to this question is given by the formulation of an integral
material balance as depicted in Figure 2.14.
A feed material MA (index A) with a cumulative particle distribution QA (x) and
a frequency function qA (x) enters a separator with a certain separation function
T(x), which leads to a split into a coarse fraction (index G) and a fine fraction
(index F) as indicated in Eq. (2.17)
M A = MG + MF (2.17)
Feed Figure 2.14 Integral material balance.
MA
QA(x)
qA(x)
T(x)
MG MF
QG(x) QF(x)
qG(x) qF(x)
Coarse fraction Fine fraction
2.3 Particle Properties 27
Q(x) q(x)
1
QA(x)
qA(x)
[Link](x)
[Link](x)
0.5
g
f
x x
xmin xt xmax xmin xt xmax
Figure 2.15 Ideal sharp fractionation.
Normally, not the absolute amount of material is interesting, but according to
Eq. (2.18), the percentage of fines f and the percentage of coarse material g in
relation to 100% of the feed material
MG MF
=g =f (2.18)
MA MA
The sum of the relative portions g and f of course add up to 1 or 100% (Eq.
(2.19))
g+f =1 (2.19)
Equation (2.17) describes the integral material balance for the fractionation
process. In many cases, it is interesting to know the separation efficiency for each
particle size. For this purpose, in Eq. (2.20), not an integral but a differential mate-
rial balance is formulated
qA (x) = g ⋅ qG (x) + f ⋅ qF (x) (2.20)
Equation (2.20) can also be written in the integral form
QA (x) = g ⋅ QG (x) + f ⋅ QF (x) (2.21)
An ideal sharp fractionation at the cut size xt is shown in Figure 2.15.
In this case, no particle smaller than xt gets into the coarse fraction and no par-
ticle greater than xt gets into the fine fraction. The amount of misplaced particles
amounts zero. The share of coarse and fine particles is calculated in Eq. (2.22) by
integration of the frequency function
xmax xt
qA (x) ⋅ dx = g qA (x) ⋅ dx = f (2.22)
∫xt ∫xmin
As depicted in Figure 2.16, in the technical reality, the fractionation will never
be ideally sharp, but more or less misplaced particles can be found in both frac-
tions.
The particle size distributions of the fine and the coarse fraction are overlap-
ping now. In the fine fraction, particles of x > xt can be found, and in the coarse
fraction, particles of x < xt . These portions of misled particles can be calculated
according to Eq. (2.23)
xmax xF,max xmax
qA (x) ⋅ dx = f ⋅ qF (x) ⋅ dx + g ⋅ qG (x) ⋅ dx (2.23)
∫xmin ∫xmin ∫xG,min
