Huisstijltemplate voor overheadsheets prints

1
Membrane emulsification the influence of pore
geometry, wall contact angle and membrane
morphology on process performance
A.J. Gijsbertsen-Abrahamsea, A. van der Padtb,
R.M. Booma

• This can be done by increasing the number of
  active pores. However, it is not yet known on
  which parameters the pore activity depends.
  Therefore, based on the results in Fig. 1, we
  studied whether
• pore size differences
• differences in wetting properties of the pores
• membrane morphology
• can explain the gradual increase of pore activity
  upon increasing pressure.

Introduction Observations Membrane
emulsification has large potential for the
preparation of very well-defined emulsions. We
have studied the process with micro-engineered
membranes, that feature absolutely uniform pores,
an extremely flat and well defined membrane
surface and a membrane thickness of only a
micrometer. In Fig. 1 the observed of active
pores is shown 1. To make the cross-flow
membrane process commercially feasible, higher
disperse phase fluxes must be reached. Pore
size Because of the uniformity of the pores, it
was expected that above a certain critical
pressure (pcrit calculated with the Laplace
equation (eq. 1)), all the pores would form
droplets. eq. 1 g interfacial tension q
wall contact angle Pore geometry and
wetting In case the pore edges are rounded, the
wall contact angle does affect the critical
pressure (eq. 1). We did two-dimensional, two
phase computational fluid dynamic (CFD)
simulations to quantify this effect. In CFX 4.3
we modeled a pore with rounded edges (Fig. 3) and
studied whether a droplet formed (indicated with
o) at various wall contact angles and pressures
(Fig. 3). Membrane morphology The results above
show that pore size, pore shape and wetting
properties can only have a small effect on the
critical pressure and thus affect the activation
of individual pores only slightly. Therefore, we
developed a model that not only considers the
flow through the pores (resistance Rp ), but also
describes the flow under the membrane toplayer
(resistance Rs). The model predicts a linear
increase of the number of active pores (N) with
increasing transmembrane pressure (ptrm, eq. 2,
Fig. 4). eq. 2


Figure 1 of pores at which droplets are formed
as a function of transmembrane pressure 1.
Fig. 2 shows that the critical pressure (stage 3)
is only determined by the interfacial tension and
the pore radius, when the pore has sharp edges.
In the experiment (Fig. 1) the 1st pore became
active at 4.9 kPa, the 16th pore at ptrm 14
kPa. Because it is given that the standard
deviation of the pore size is less than 1, it is
not possible that the 16th active pore had a
radius of approximately 3 (14/4.9) times the
radius of the 1st active pore. From Fig. 3 it
is clear that wetting properties of the pores
change the critical pressure of the pores only
very slightly the decrease in critical pressure
is 20 at maximum in the modeled geometry. It is
not very plausible that either the pore opening
was much wider than in the modeled geometry and
that the differences in wall contact angles
between the individual pores was so
large. Conclusion To obtain a high throughput
of the to-be-dispersed phase through the
membrane, the fraction of pores at which droplets
are formed, should be as high as possible. In
this study we show that the number of active
pores is mainly determined by the membrane
morphology. Hence, the number of active pores can
be increased by increasing the ratio of pore flow
resistance and flow resistance in the membrane
substructure. When designing a membrane for
emulsification, these factors should be taken
into account to obtain a system with commercially
feasible disperse phase fluxes.
1
Figure 2 stages of droplet formation. Stage 3
critical pressure
2
7
µm
7
µm
1
µm
1
µm
4
µm
4
µm
5
µm
5
µm
Figure 3 droplet formation (o) at various
pressures and wall contact angles. Calculated
with a geometry with rounded edges (inset) with
CFD.
3
?
Figure 4 experimental data and model fit. pcrit
2.9 kPa, Ntotal 100
a Food and Bioprocess Engineering Group b
Friesland Coberco Dairy Foods, P.O. Box 8129,
6700 EV Wageningen, The Netherlands Corporate
Research, P.O. Box 87, Anneke.Gijsbertsen_at_algemeen
.pk.wau.nl 7400 AB Deventer, The Netherlands ?
31 317 482240 September 2002
1. Abrahamse, A.J., R. van Lierop, R.G.M. van der
Sman, A. van der Padt, R.M. Boom. Journal of
Membrane Science 204 (1-2), 125 137, 2002.