12th European Conference on Mixing

Bologna, 27-30 June 2006

DE-AGGLOMERATION OF PRE-WETTED NANO-PARTICLES USING

MIXED FLOW AND HIGH SHEAR IMPELLERS
L. Xie a, C. D. Rielly a*, W. Eagles b, G. Özcan-Taşkin b
a
Department of Chemical Engineering, Loughborough University, Loughborough, Leics.,
LE11 3TU, UK; email: C.D.Rielly@lboro.ac.uk
b
BHR Group Ltd, Cranfield, Bedfordshire, MK43 0AJ, UK

Abstract. When nano-particulate Aerosil® 200V powders are dispersed into water, agglomerates as
large as hundreds of microns are formed initially. To break down these agglomerates, a number of
impellers which generate different power inputs and flow patters were studied. The low energy
dissipation impeller configurations, e.g. single LE-20, PBT, dual PBTs, dual LE-20s and 6DT,
provided only a small degree of deagglomeration and their flow patterns did not affect the results. At
high rotational speeds, the higher local energy dissipation created by the EkatoMizer impeller could
break the Aerosil agglomerates into submicron particles, but the rate of the de-agglomeration was
slow. Rotor-stators which impart high maximum dissipations rates to the dispersion could break the
agglomerates quickly into submicron sizes. Several combinations of low and high energy dissipation
impellers were studied and these results showed that the de-agglomeration process is controlled by the
maximum energy dissipation rates produced by the mixing device.

Key words: nano-particle; Aerosil; mixing; energy dissipation; EkatoMizer; rotor-stator.

1. INTRODUCTION
Dispersion of nano-particle powders into a liquid typically involves three parallel processing
operations: (i) wetting and incorporation of the dry powders, which often form large
agglomerates; (ii) de-agglomeration to meet the product specifications in terms of the particle
size distribution and (iii) stabilization using surfactants and thickeners and thereby the
formation of a micro-structure. The study described here is part of a large EC funded project,
PROFORM, which examines a number of different chemical, hydrodynamic and mechanical
mechanisms, and a range of processing equipment, for processing of nano-particles. In
practice, although the primary particle size may be of O(10) nm, some of the larger
agglomerates sizes formed are within the inertial sub-range, O(1-100) µm and thus
hydrodynamics may be effective at breaking these agglomerates. To create sub-micron sizes
requires a combination of chemical effects to destabilise the agglomerates and very high local
energy dissipation rates, εmax, which cannot be achieved in stirred tank reactors (STRs).
However, the controlled wetting of powders is difficult to achieve in high shear devices, such
as in-line rotor-stators or valve homogenizers. Hence the STR is often used to perform
wetting and preparation of a crude suspension, which is then fed to more intensive mixers to
achieve the final product specifications. Blockage of valve homogenisers can occur if the
crude dispersion contains large agglomerates and thus the STR must reduce particle sizes to
an acceptable level before more intensive processing takes place.
The application of hydrodynamic forces to deagglomerate nano-particles has a number
of similarities to liquid-liquid break-up processes. The flows generated by industrial scale
equipment are generally turbulent and so unsteady stress fluctuations act upon the droplets,
causing them to deform. If the forces causing deformation exceed the resisting forces due to
surface tension and viscous stresses within the drop, then the drop may break-up into smaller
fragments. The classical analyses of Kolmogorov [1] and Hinze [2], lead to an expression for
the maximum stable drop size, which is related to ε the energy dissipation rate. Zhou and
Kresta [3] discussed the importance of relating droplet sizes to the local impeller
hydrodynamics and therefore it is reasonable to suppose that the maximum drop size should
be related to the maximum local energy dissipation rate, εmax. Such analyses are for the special
case where viscous stresses within the drop are negligible and the drop sizes are in the inertial
sub-range of the turbulence energy spectrum, i.e. where the drops are larger than the
Kolmogorov scale, λk = (ν3/εmax)1/4. Smaller droplets may also be broken by hydrodynamic
effects: Baldyga and Bourne [4] describe how the Kolmogorov shear rate, (εmax/ν)1/2, produces
a viscous shear stress which is capable of breaking drops.
Similarly, impeller driven flows may be used to break down, or de-agglomerate, the
nano-particle structures in a STR. Thus, depending on the range of agglomerate diameters and
the minimum Kolmogorov length scale, the breakage in turbulent flow might be dominated by
inertial and / or viscous mechanisms, but in both cases the maximum size and the kinetics of
breakage are expected to depend on εmax. In this work, a model system of nano-particulate
Aerosil 200V dispersion in deionised water, without stabilisers was prepared. In the case of
liquid-liquid dispersions, the classical analyses apply only to dilute systems, where
coalescence may be ignored, but this provides a good analogy with the work here, where re-
agglomeration of Aerosil dispersions has not been observed. The de-agglomeration
performance of several types of impellers, which generate different pumping capacity, flow
patterns and specific power inputs, were studied. In addition, high shear rotor-stators placed in
the STR were also studied. The de-agglomeration process was then characterized using
estimates of the local energy dissipation rate.

2. EXPERIMENTAL
2.1 Experimental set-up
The experimental data reported here were obtained using a model nano-particle system,
Aerosil® 200V (Degussa) which is a densified hydrophilic fumed silica (> 99.8% SiO2 by
weight). The solid nano-powder has a specific surface area of 200 m2/g. The primary particle
has an average size ~12 nm and its tapped density is approximately 120 kg/m3. The liquid
used was deionised water, which had an electrical conductivity less than 1 µS cm–1. (Similar
work is being conducted on the effects of pH, temperature and chemical additives on
agglomerate breakage [5].) Zeta potential measurements indicated that the stability of Aerosil
agglomerates is approximately the same for ultra-pure, deionised and distilled water.
Preliminary experiments showed that the size of agglomerates formed did not depend on the
quality of the water used. Moreover, the dispersions formed were stable and re-agglomeration
did not occur during storage for several days.
The de-agglomeration experiments were
performed in a 305 mm diameter cylindrical tank with a
torispherical base, as shown in Figure 1. Four equally
R=T spaced baffles were mounted on the walls of the tank. In
H=T the experiments discussed here the mass fraction of solids
was 1%, although higher concentrations were also studied.
Özcan-Taşkin et al. [6] showed that the wetting
r =T/10
T/5
C and draw-down of initially dry powders from the top
surface of a stirred vessel is not an instantaneous process
T/60
T/10 and can take between 10 and 100 s, depending on the
T solids mass fraction and impeller speed. In such a
Figure 1 The mixing tank geometry situation, the initial agglomerates enter the flow at
(T = 305 mm) used in this study. different times and with a broad range of sizes. To avoid
inconsistencies with the dynamics of the initial wetting
stage, a procedure was developed to add hand-mixed wetted Aerosil® 200V into the liquid in
the STR at the 1% concentration. The very crude dispersion produced by hand-mixing was
shown to have a very reproducible size distribution and thus provided a consistent starting
point for the deagglomeration studies.

2.2 Impellers configurations investigation in the de-agglomeration experiments

The geometric details and dimensions of the four types of impellers used in the experiments
are presented in Table 1 and Figure 2. The LE-20 and PBT impellers are axial and mixed
flows, whereas the 6DT impeller generates a radial discharge flow; the LE-20, PBT and 6DT
impellers were all categorised as low εmax devices (< O(0.1) kW/kg). The EkatoMizer, a saw-
tooth impeller run on a separate shaft, was able to generate a higher local energy dissipation
rate, but only a weak discharge flow. The EkatoMizer also individually or combined with
other axial/radial flow impellers. A series of de-agglomeration experiments were carried using
a variety of impeller configurations and rotational speeds covering a range mean specific
power inputs from about 0.05 to 3 kW/m3 and a range of εmax values up to about 0.3 kW/kg.
In addition, two sizes of IKA Ultra-Turrax® rotor-stators were used to provide very high
energy dissipation rates with tank average values up to 6 kW/m3 and εmax up to about 9
kW/kg. The method for estimating εmax is discussed in §3.4.

Table 1. The impeller specification and power number measured at C/T=1/3 and T=0.305 m
Impeller Description D [m] L [m] W [m] t [m] Np
LE-20 Hydrofoil impeller (down pumping) 0.152 - - - 0.28
PBT 45◦ pitched 4-bladed turbines 0.150 0.061 0.030 0.0021 1.39
6DT 6-bladed radial disc turbine 0.152 0.038 0.031 0.0017 5.02
EkatoMizer® Saw-tooth impeller 0.095 - 0.025 0.0030 0.095

t t
α
W

W
L L
Dd 6DT PBT
D
D

LE-20 EkatoMizer

W2
δ1 δ2
γ1
δ3

γ2
W1 α1 α2
D
Figure 2. Impellers used in the de-agglomeration experiments

2.3 Sampling and data analysis

Samples of the dispersion were withdrawn, at specific times from the stirred tank using a
head-chopped 50 cm3 volumetric pipette, to follow the dynamics of the particle size
distribution (PSD) evolution. The samples were analysed by laser diffraction, using a Malvern
Mastersizer® S, to yield the PSD which was expressed as a volume probability density
function, q3. A ‘polydisperse’ analysis model was chosen in the calculation of the PSD and a
complex refractive index 1.46+0.1i was applied in all data analysis presentations. Preliminary
tests showed that the sampling position and the sampling velocity had negligible effects on the
PSD measured. Moreover, the samples were found to be stable for several days and the
sample-holder pumping speed for the Mastersizer S was found to have no effect on the PSD.

3. RESULTS AND DISCUSSION

3.1 Low energy dissipation impellers (LE-20, PBT and 6DT)
Figure 3 shows that for the low εmax devices, the LE-20, PBT and 6DT impellers, the
agglomerate mean sizes were rapidly reduced from the initial sizes generated by hand-mixing
(initially D[3,2] ≈ 60 µm and D[max] ≈ 800 µm) and then slowly decreased for a further 10-
15 min; at longer times D[3,2] remains approximately constant. (Similar effects were
obtained with D[4,3].) Typically the D[3,2] of the agglomerates was quickly reduced to a size
of the order of the Kolmogorov length scale (in these experiments between 5 and 60 µm) and
thereafter no further size reduction took place.

60 8
pre-mixed
Diameter, D[3,2] (µm)

50 10 min
60 min
6 120 min
40
Volume %

180 min
240 min
30 4 300 min
2 x LE-20 180 rpm
20 2 xPBT 130 rpm
2
6DT 165 rpm
10
PBT 320 rpm
0 0
0 100 200 300 400 500 600 1 10 100 1000
Deagglomeration time (min) Particle size (mm)

Figure 3. Reduction in the agglomerate diameter Figure 4. Measurement of the agglomerate PSD with
D[3,2] with dual LE-20s, dual PBTs, 6DT and time: 6DT impeller at 165 rpm. The curve on the right
single PBT. The mean specific power inputs were is the pre-mixed sample used as the starting point for
0.064, 0.107, 0.395 and 0.801 kW/m3 respectively. the de-agglomeration runs.

As the specific power input of the single and double impeller systems increased, the
agglomerate mean sizes were found to decrease (see Figure 3). In these cases mono-modal
PSDs were obtained (e.g. see Figure 4 for the 6DT) and at long times the agglomerate PSDs
remained almost constant; for example, the differences between the PSDs at 180, 240 and 300
min in Figure 4 are within the experimental reproducibility. Furthermore, the agglomerates
produced have a wide size distribution from a few µm to hundreds of µm. With these low εmax
devices, sub-micron agglomerates were not observed even after 10 hours of agitation.

3.2 Higher energy dissipation impellers: EkatoMizer and rotor-stators

The EkatoMizer saw-tooth impeller is designed to be operated at very high rotational speed
for operations such as droplet breakage and de-agglomeration. This device will produce
higher local εmax values, compared to the LE-20, PBT and 6DT, even though the mean P/V is
comparable. Later this will be confirmed using particle image velocimetry techniques to
estimate the local energy dissipation rate [7]. Estimates of εmax are discussed in §3.4.
Figures 5 and 6 show that for the Ekatomizer the mean particle sizes continued to
decrease for one hour and had not reached a steady-state or equilibrium PSD. The
development of a second mode in the sub-micron range of the PSD in Figure 5, after about
15-30 minutes, indicates that this impeller is capable of generating finely dispersed materials.
In contrast, no sub-micron fines were generated using the LE-20, PBT and 6DT impellers.
 0.10 40
Probability density q 3 (µm-1)

pre-mixed 750 rpm

Diameter D [3,2] (µm)

1 min 1500 rpm
0.08 10 min 2200 rpm
30
15 min 2600 rpm
0.06 30 min
60 min
20
0.04
10
0.02

0.00 0
0.01 0.1 1 10 100 1000 0 10 20 30 40 50 60 70 80 90
Particle size (µm) De-agglomeration time (min)

Figure 5. Agglomerate size probability density function Figure 6. Agglomerate sizes D[3,2] vs. de-
for different de-agglomeration times: EkatoMizer agglomeration time at EkatoMizer speeds of 750, 1500,
impeller at 2600 rpm and C=T/3. Fines (left peak) can 2200 and 2600 rpm corresponding to mean P/V of
be found after 15 – 30 mins. 0.072, 0.574, 1.812 and 2.991 kW/m3 respectively.

Figure 6 presents the evolution of the agglomerate D[3,2] at four EkatoMizer speeds,
all starting from the same hand pre-mixed dispersion, which had D[3,2] of about 60 µm. At
the lowest speed of 750 rpm, the agglomerate mean sizes are 30–40 µm, which is comparable
to the dispersions produced by the low εmax impellers (LE-20, PBT and 6DT). At the higher
Ekatomizer speeds, these large agglomerates were slowly broken down and the PSD moved to
the left, as shown in figure 5. Above 22 000 rpm, a second mode appeared in the PSD,
indicating that sub-micron particles have been produced by the high local energy dissipations
rates generated by the Ekatomizer. The rate at which these fines are generated is slow and
continued beyond the duration of the experiment. This low rate of particle breakage may be
attributed to (i) the intermittent nature of the energy dissipation; (ii) the low volume fraction
of the regions containing high dissipation rates and (iii) the frequency at which agglomerates
are circulated back to the impeller. Podgórska and Bałdyga [8] suggest that in liquid-liquid
dispersion, the smallest drops are only formed by the most extreme turbulent events which
occur in small regions of space and only intermittently. Nano-particle dispersion appears to
occur by an analogous process: the slow drift of the agglomerates sizes towards the stable
PSD is likely to be caused by turbulence intermittency.
When the high speed, high shear rate rotor-stator (see figure 7) was introduced into the
stirred tank, the agglomerates were broken down more rapidly. A second mode of sub-micron
fines appeared in the PSD after about 1-5 minutes (see Figure 8). Compared with the
EkatoMizer at similar times, the volume fractions contained in the fine particle modes of the
PSDs generated by the rotor-stator were
large, i.e. the breakage rates are much
greater. In Figure 9 the agglomerate mean
size D[3,2] decreases very quickly from its
initial value of about 60 µm and falls to
less than 2 µm after 60 minutes; the
average agglomerate sizes are sub-micron
after three hours agitation. As shown in
Figures 6 and 9, the mean P/V for the T25
(a) IKA T25 (b) IKA T50 rotor-stator is greater than for the
EkatoMizer impeller. Moreover, if it is
Figure 7. Ultra-Turrax rotor-stator dispersion tool
(a) IKA T25 and (b) IKA T50; the outer diameters of assumed that dissipation predominantly
the rotor and stator and gap width between them are occurs within the swept volume of the
(a) 18, 25 and 0.5 mm and (b) 36, 45 and 1 mm. rotor-stator, then the IKA T25 and T50
Impellers were mounted at C/ T = 1/2 should give much higher local εmax values
than the Ekatomizer (see §3.4).
-1
Probability density q 3 (µm ) 1.2 25.0
Pre-mixed T25 11 000 rpm

Diameter D [3,2] (µm)

1.0 5 min 20.0 T25 13 000 rpm
T25 16 000 rpm
0.8 10 min
20 min 15.0
0.6 30 min
60 min 10.0
0.4
120 min
0.2 5.0

0.0 0.0
0.01 0.1 1 10 100 0 60 120 180 240 300
Particle size (mm) De-agglomeration time (min)

Figure 8. Agglomerate size volume probability density Figure 9. Agglomerate sizes D[3,2] vs. de-
function for different de-agglomeration times: T25 agglomeration time at three T25 rotor-stator speeds of
rotor-stator at 16 000 rpm and C=T/2. 11 000, 13 000 and 16 000 rpm corresponding to mean
P/V of 2.21, 3.17 and 6.01 kW/m3, respectively.

3.3 Combination of low and high energy dissipation impellers

Further de-agglomeration experiments were carried out using a combination of low and high
energy dissipation impellers. Figure 10 presents three experiments run at the same total power
input, with different EkatoMizer and dual LE-20s speeds; the speed of the EkatoMizer
controls the maximum dissipation rate εmax. Although the data are not completely consistent,
the mean sizes D[3,2] are usually greater for the lower EkatoMizer speeds, even though the
total specific power input remains constant. These results support the conclusion that the
maximum local energy dissipation controls the de-agglomeration process, rather than the
mean P/V. Although the PBT helps in generating the bulk flows in the tank, its effect is
limited, because the EkatoMizer itself generates sufficient discharge flow for macro-mixing at
high rotational speeds. This finding is confirmed by the results shown in Figure 11, which
shows a series of experiments run at a fixed EkatoMizer speed (and hence constant εmax) with
various PBT power inputs. Adding a PBT impeller and increasing its power input up to 16.5%
of the total energy input, results in no change in D[3,2] produced by the EkatoMizer alone.
Similar results were found with a rotor-stator and PBT combination, as shown in Figure 12:
maintaining a constant εmax (controlled by the rotor-stator speed) showed no change in D[3,2]
when the PBT power input was increased. Conversely, keeping the PBT power input constant
and increasing εmax produced by the rotor-stator resulted in the production of smaller
agglomerates. Thus in both cases the high energy dissipation devices, the EkatoMizer and
rotor-stator, control the value of εmax and hence determine kinetics of the de-agglomeration
process and the final sizes of agglomerates produced.

40 40
P PBT / P Total
Diameter D [3,2] (µm)

Diameter D [3,2] (µm)

Mizer, 2200 rpm 0

30 30 + PBT 145 rpm 3.94%
+ PBT 205 rpm 10.4%
+ PBT 245 rpm 16.5%
20 20
3
P /V kW/m
10 Ekato 1500 rpm 0.580 10
Ekato+2 LE20s 1300+270 rpm 0.588
Ekato+2 LE20s 1425+200 rpm 0.574
0 0
0 20 40 60 80 100 120 0 20 40 60 80 100 120
De-agglomeration time (min) De-agglomeration time (min)

Figure 10. Single EkatoMizer combined with Figure 11. De-agglomeration experiment using a
two LE-20 impellers. The EkatoMizer speed was single EkatoMizer in combination with a PBT.
changed from 1300 rpm to 1500 rpm. The total The EkatoMizer power input was fixed, whereas
power input in all three experiments was the PBT speed was varied from 145 to 245 rpm,
approximately constant. changing its power input.
 30 40
ε max ≈ 8.6 kW/kg T25, 16 000 rpm

Vol fraction of fines Vf (%)

T25 11 000 rpm + PBT 230 rpm
Diameter D [3,2] (mm)
T25, 13 000 rpm
T25 11 000 rpm + PBT 104 rpm T25, 11 000 rpm
T25 13 000 rpm + PBT 104 rpm 30 T50, 7500 rpm
20 4.6 kW/kg T50, 6000 rpm
T50, 4000 rpm
2.2 kW/kg
20
2.8 kW/kg
10
10
1.1 kW/kg
0.3 kW/kg
0 0
0 20 40 60 80 100 120 0 100 200 300 400
De-agglomeration time (min) De-agglomeration time (min)

Figure 12. Deagglomeration using a T25 rotor- Figure 13. The rate of generation of fines by re-analysis
stator plus a single PBT at three speed of PSD curves. Estimates of the εmax from eq.(2) for
combinations. each rotor-stator operation are also shown.
3.4 Analysis of the rate of production of fine particles
In most of the experimental PSDs, the lower modes (fine agglomerates) were in the sub-
micron scale, whereas the upper modes (coarse agglomerates) were usually above 1 µm. This
suggests an erosional mechanism, by which loosely-attached sub-micron particles are broken
away slowly from larger agglomerates, at a rate determined by the extremes of intermittent
dissipation rates and the circulation frequency. With non-overlapping modes, it is
straightforward to obtain the volume fraction of fine particles in the PSD. Figure 13 shows a
comparison of the rate of fines generation by the rotor-stators, T25 and T50. The power
inputs of these impellers over a range of rotational speeds were measured using a well-
insulated calorimetric experiment,: the power numbers, Np, were constant with values of 3.9
and 1.0 for the T25 and T50 rotor-stators (the rotor diameter has been used to calculate Np and
the rotor-stators are not geometrically similar). Wu and Patterson [9] found that about 30% of
the power of the 6DT was dissipated in the impeller region, whereas Zhou & Kresta [10]
estimated this to be about 50% for the PBT. More recently, using particle image velocimetry
methods, Khan et al. [7] showed that for the PBT, about 40% of the dissipation occurred
inside the swept volume. Thus an crude estimate of the maximum dissipation rate may be
obtained assuming that about 40% of the power input is dissipated close to all of the impeller
used here, leading to the following equation.
0 .4 N p N 3 D 5
ε max = (1)
Vswept
In practice, intermittency and gradients in the dissipation rate within the swept volume, mean
eq.(1) is likely to provide an underestimate of εmax. These estimates of εmax are shown in
figure 13; the corresponding Kolmogorov scales are about 3-7 µm, which are smaller than the
largest agglomerate sizes in most cases, indicating that inertia could be an effective breakage
mechanism. In all the de-agglomeration experiments where fine particles were produced, the
smallest agglomerates had approximately constant sizes of around 0.3 µm; thus these may
have been produced by a viscous shear breakage mechanism. The small swept volume and
larger power number gave larger values of εmax for the T25 than the T50. For both rotor-
stators, decreasing the rotational speed (and hence εmax) resulted in a decrease in the rate at
which fines are generated. Operation at 11 000 rpm with the T25 and 7500 rpm with the T50
rotor-stator gave about the same value of εmax and similar rates of fines generation (see figure
13) were obtained, which tends to confirm that εmax controls the kinetics of de-agglomeration.
Equation (1) may also be used to estimate εmax values for the other impellers. For
example, at 2200 rpm with the Ekatomizer (the lowest speed at which submicron fines were
generated), eq.(2) yields that ε max ≈ 0.2 kW/kg . In that case, fines were extremely slowly
produced and were not detected until after 60 min, which is broadly consistent with the rotor-
stator results of figure 13. Estimates of the maximum dissipation rate for the single PBT at
320 rpm and the 6DT at 165 rpm yielded values of 0.02 and 0.04 kW/kg and as a result no
fine agglomerates were detected, even after prolonged times of agitation.

4. CONCLUSIONS
All of the impellers and combinations investigated were successful in producing a degree of
size reduction of the nano-particulate agglomerates of Aerosil® 200V dispersions. For the low
maximum energy dissipation impellers, e.g. the single and dual LE-20s, PBTs and 6DT, the
agglomerate sizes were rapidly reduced initially and then slowly decreased after about 10-15
minutes, but not to submicron sizes. Similar results were obtained for the EkatoMizer impeller
at low speed. At higher speeds, the EkatoMizer generated a higher εmax and the Aerosil
agglomerates were broken into sub-micron fragments. A study of various impeller
combinations showed that the de-agglomeration process is controlled by the value of εmax
produced by the high shear devices and was unaffected by the power input from the LE-20 or
PBT. The most successful devices were the rotor-stators, which were able to generate very
high εmax values and hence sub-micron agglomerates. Order of magnitude type estimates of
εmax values were obtained by assuming that a fixed fraction of the power was dissipated within
the impeller volume. Using these estimates it was possible to show that the rate at which fines
were generated depended on the maximum dissipation rate. The largest agglomerate sizes
produced were in the inertial subrange, but the smallest were sub-Kolmogorov in size and
hence it is likely that a combination of inertial and viscous shear mechanisms are responsible
for breakage.

Acknowledgements: This study was carried out within the project PROFORM (“Transforming Nano-particles
into Sustainable Consumer Products Through Advanced Product and Process Formulation” EC Reference
NMP4-CT-2004-505645) which was partially funded by the 6th Framework Programme of EC. The contents of
this paper reflects only the authors’ view. The authors gratefully acknowledge the useful discussions held with
other partners of the Consortium: Karlsruhe University, Inst. of Food Process Eng.; Bayer Technology Services
GmbH; Unilever UK Central Resources Ltd.; Birmingham University School of Eng.; Warsaw University of
Technology, Dept. of Chemical and Process Eng.; Poznan University of Technology, Inst. of Chemical Tech. and
Eng.; Rockfield Software Limited; Centre for Continuum Mechanics.

REFERENCE
1. Kolmogorov A.N., 1941. “The local structure of turbulence in compressible viscous fluid for very
large Reynolds numbers”, Doklady Akademii Nauk SSSR, 30, 301-305.
2. Hinze J.O., 1955. “Fundamentals of the hydrodynamics mechanism of splitting in a dispersion
process”, A.I.Ch.E. Journal, 1, 289-295.
3. Zhou G., Kresta S.M., 1998. “Correlation of mean drop size and minimum drop size with the
turbulence energy dissipation and the flow in an agitated tank”, Chem. Eng. Sci., 53, 2063-2079.
4. Bałdyga J., Bourne J.R., 1994, “Drop breakup in the viscous subrange: a source of possible
confusion”, Chem. Eng. Sci., 49, 1077-1078.
5. Pacek A.W., Ding P., Utomo, A.T., 2006, “Effects of energy input , pH and temperature on de-
aggregation in nano-particles suspensions” 12th Eur. Conf. on Mixing (Bologna, 27-30 June).
6. Özcan-Taşkin G., Eagles W., Clements P., Xie L., Rielly C.D., 2006. “Suspension of fine
particles in a liquid using stirred tanks”, 12th Eur. Conf. on Mixing (Bologna, 27-30 June).
7. Khan F.R., Rielly C.D., Brown D.A.R., 2006. “Angle-resolved stereo-PIV measurements close to
a down-pumping pitched-blade turbine”, Chem.Eng.Sci., 61, 2799 – 2806.
8. Podgórska W., Bałdyga J., 2002. “Effect of structure of turbulence on drop breakage”, Chemical
Papers-Chemicke Zvesti, 56, 412-417.
9. Wu H., Patterson G.K., 1989. “Laser-Doppler Measurements of turbulent flow parameters in a
stirred mixer”, Chem. Eng. Sci., 44, 2207-2221.
10. Zhou G., Kresta S.M., 1996. “Distribution of energy between convective and turbulent flow for
three frequently used impellers”, Trans IChemE, 74, Part A, 379-389.