Dispersing of gases or immiscible liquids within another continuous liquid fluid phase
is a standard operation for the processing of multiphase systems in the industrial
areas like food, cosmetics pharmaceuticals and fine chemicals. The commonly used
dispersing tools are rotor-stator devices with tooth-/pin geometries arranged in
circular, concentric or axial, stacked disc configurations. Depending on the gap width
between the rotor-stator pins or tooth elements and the viscosity function of the twophase
system, laminar, turbulent or transitional flow conditions act in the dispersive
mixing procedure. – A new generation of dispersing devices are the static or dynamic
membrane devices in which the disperse drops or bubbles are formed once and
detached from the membrane surface by cross- or co-flowing fluid streams. This
procedure means significantly reduced mechanical treatment of the multiphase fluid
system and allows to tailor narrow disperse size distributions.
Rotor-Stator Dispersing Processes
Our recent work concerning Rotor-Stator (R/S) dispersing process developments
has mainly focused on laminar to transition domain flow conditions. We have
investigated bubble and drop break-up in some detail under steady and transient
shear or elongation as well as mixed flow conditions. Different types of R/S (model)
flow apparatus were applied comprising concentric  /eccentric  cylinder-, four
roller- , single-  and multi-toothed  geometries. From respective experiments
expanded maps of critical bubble-/drop break-up characteristics (critical Weber
(WeC) or critical Capillary (CaC) numbers as a function of viscosity ratio
λ, deformation rate G, total deformation D and flow type α have been derived.
Figure 1 exemplarily demonstrates CaC(α) for drop break-up under pure shear (α =
0), equibiaxial (α = -1), or planar (α = 1) elongation as well as for mixtures of these
flow types. Steady and transient drop break-up were investigated experimentally (a),
by numerical flow simulation using CFD (b) and by a non- equilibrium
thermodynamics modeling approach (c) /1-3/. Consistent results from these three
approaches (a-c) were received for surfactant free as well as for surfactant covered
drop interfaces.

Fig.1: Critical Ca- number for different flows

Fig.2: Surfactant distribution at drop
interface in shear flow

For surfactant covered drop interfaces a criterion to distinguish between diffusion
and convection driven interfacial coverage with surfactant molecules was defined as
the ratio of Peclet number (Pe) / Capillary number (Ca), denoted as α and
implemented into a convection diffusion equation which forms the bases for
respective CFD calculations. As a result of these calculations surfactant
concentration distributions along the interfacial contour of drops deformed in shear,
elongation and mixed flows were received and satisfying comparability with
experimental drop deformation data was found. Figure 2 shows such calculated
concentration distributions of surfactant at the deformed drop interface for different
Capillary numbers and a viscosity ratio λ of 4.
The impact of transient shear and elongation flows has been investigated within an
eccentric cylinder gap and transferred to a complex multi toothed rotor-stator
dispersing geometry. CFD based simulations applying a particle tracking procedure
along distinct particle flow tracks allowed us to quantify the transient drop
deformation history of selected drops along their paths through the dispersing
apparatus. Comparisons with respective experimental results demonstrated again
good agreement as demonstrated in Figure 3.

Fig. 3: left: Transient deformation and Ca-number; right: multi-toothed R/S geometry
Membrane / Micro-channel Dispersing Processes
In addition to rotor-stator flow devices we considered also channel / nozzle / pore
flows with respect to their dispersive mixing performance. New microfluidics devices
have been developed in our Laboratory at ETH Zürich in close collaboration with the
University of Queensland in Brisbane (Australia); Prof. J. Cooper-White. Within the
lasts two years we investigated drop formation in co-flow and cross-flow micro- and
macro channels. By means of micro particle imaging the velocimetry (Micro-PIV) we
accessed velocity fields around respective drops and used this information for
optimizing the dispersing channel flow geometries and to derive scale up criteria
(e.g. We = f (Re) characteristics) over several orders of magnitude like demonstrated
in figure 4 .
As a scaleable solution with application relevance, derived from micro channel cross
flow results, a Rotating Membrane Device (ROME) with Controlled Pore Distance
(CPD) was developed. The cross flow is generated by the rotational motion of a
membrane cylinder within a surrounding concentric housing through which the
continuous fluid phase is axially pumped. The disperse fluid or gas-phase enters

Fig 4: Micro-/macro channel co-flow dispersing; experimental data / M. Duxenneuner
through a hollow shaft into the rotating cylinder membrane body and forms disperse
liquid droplets or bubbles at the membrane surface, from which the cross flowing
continuous fluid phase flow detaches them as soon as a critical shear stress is
exceeded .

There are a number of reasons to be interested in the writing of linux device computer drivers. The rate at which new hardware becomes available (and obsolete!) alone guarantees  that driver writers will be busy for the foreseeable future. Individuals may need to know about drivers such like windows drivers in order to gain access to a particular device that is of interest to them. Hardware vendors, by making a linux driver available for their products, can add the large and growing linux user base to their potential markets. And the open source nature of the linux system means that if the driver writer wishes, the source to a driver / driver download can be quickly disseminated to millions of users.

When writing drivers, a programmer should pay particular attention to this fundamental concept / write kernel code to access the hardware, but don’t force particular policies on the user, since different users have different needs. The driver should deal with making the hardware available, leaving all the issues about how to use the hardware to the applications. A driver, then, is flexible if it offers access to the hardware capabilities without adding constraints. Sometimes, however, some policy decisions must be made. For example, a digital I/O driver may only offer byte-wide access to the hardware in order to avoid the extra code needed to handle individual bits.

PVC is in terms of revenue one of the most important
products of the chemical industry. Globally over 50% of
PVC manufactured is used in construction. Worldwide
80 % Percent of PVC is produced by suspension
polymerisation. In such processes mechanical agitation is
used to mix the monomer droplets into an aqueous liquid
phase. Growing markets and growing economies lead to
higher PVC production rates. Limits and demands in
space and transportation are changing the outfit of the
used mixing reactors. The height (H) is increasing with
constant diameter (D). Did most of the apparatuses start
with a ratio of height vs. diameter of one, ratios of two to
three are normal today and ratios of over four are expected
for the nearer future. Such unique geometries need to
fulfil the still growing exigencies in economy and
ecology. Therefore the analysis and optimisation of such
liquid/liquid systems is of major interest for the chemical
industry.
The step of scaling up a reactor from pilot plant to
industrial scale is an issue where much empiricism is still
used and where expensive and time-consuming
experimental programs are usually required [VivaldoLima
et al. 1997] and only accurate prediction of system
behaviour will change that situation. To develop such
prediction methods cooperation was set up between the
Vinnolit GmbH & Co. KG and the TU-Berlin.
From the different tasks for scale up and for the
production process of PVC the dispersion of the two
immiscible liquids is of major interest for this work. So
the drop size distributions of two model systems,
chlorobutanol/water and toluene/water, were analysed.
Here parameter variation for reactor height vs. diameter
(1.0 to 4.5), stirrer type (Rushton turbine, Retreat Curve
Impeller, single and multi-stage stirrer systems),
dissipation rate, dispersed phase fraction (5 to 50 Percent)
and influence of colloids were carried out for the named
systems. For the mathematical description of such drop
size distributions (DSD) a quantitative understanding of
drop breakage and coalescence mechanisms is essential to
develop predictive models. The mathematical model used
here is the Population Balance Equation (PBE).
After adaptation and enhancements of classical models
from the literature (Coulaloglou & Tavlarides 1977;
Kumar & Ramkrishna 1996, Alopaeus et al. 2002)
simulations for the presented system were carried out. The
use of colloids is inevitable for the suspension
polymerisation and resulting in a strong inhibition of
coalescence. So a major focus on breakage submodels of
the PBE was set. Therefore single drop breakage events

were carried out to analyze crucial influence parameters of
the breakage rate like breakage time and energy
dissipation rate. These results were used to validate and
enhance the breakage submodels of the PBE. Then the
simulation results from different models were compared
with the experimental data and each other.
METHODS
A special in-situ endoscope technique has been developed
[Ritter & Kraume 2000; Maaß et al. 2007b]. With this
technique, drop size distributions for all phase fractions
even under transient conditions can be determined with
high time resolution . The Population Balance
Equation is applied with the intention to calculate these
transient drop size distributions in the stirred system. In
order to solve the transient space averaged PBE, the
commercial numerical solver PARSIVAL® (Particle Size
Evaluation) [Wulkow et al. 2001] is applied. For the
parameter estimation the experimental data of the stirred
vessel are used. The fitted parameters had to be
significant, i.e. the confidence interval was required to be
small compared to the value of the parameter, and they
had to be independent from each other.
The single drop experiments are carried out in an in house
developed breakage cell (Maaß et al. 2007a).