Q IWA Publishing 2010 Water Science & Technology—WST | 61.7 | 2010
1733
Influence of buoyant media on particle layer dynamics in microfiltration membranes R. K. Aryal, S. Vigneswaran and Jaya Kandasamy
ABSTRACT This study forms a part of the physical study of the membrane bioreactor in presence of buoyancy media. Kaolin clay suspension with buoyancy media (anthracite) was used as a suspension and the particle layer development on membrane surface with evolution of time was studied. Presence of buoyancy media reduced the pressure development by almost two folds compared to in absence of the media. The particles deposition on membrane surface was size selective. The mean particle diameter (0.45 mm) deposited on the membrane surface remained almost similar in presence of the media after 7 hrs run where as in its absence the mean diameter finer particles deposition occurred at the beginning followed by coarser particles. Key words
R. K. Aryal S. Vigneswaran (corresponding author) Jaya Kandasamy Faculty of Engineering and Information Technology, School of Civil and Environmental Engineering, University of Technology, Broadway, Sydney NSW 2007, Australia E-mail:
[email protected];
[email protected];
[email protected]
| buoyancy media, kaolin clay, microfiltration membrane, particle sizes
INTRODUCTION Microfiltration (MF) is one of the most popular methods of
reversible fouling is generally controlled through the shear
colloidal suspension separation. The uplifting air bubble
force of air bubbles against the membrane surface and/or
plays an important role for cake removal. Although it has
back-pulsing the membrane module. Selective deposition of
many advantages compared to the tradition techniques
the particles on the membrane surface and the external
such as sedimentation, centrifuging etc, membrane fouling
forces exerted on the particles deposited on the membrane
has become one of the major hurdles in the membrane
surface determines whether the particle deposit perma-
purification that slowly inhibits the filtration process and
nently or not.
the need for membrane cleaning for further filtration
Many research works have been carried out in cross
( Judd 2006). The mechanism of cake formation and its
flow filtration about the particle deposition and cake layer
further growth to an equilibrium cake thickness are very
formation. Lu & Ju (1989) analysed the particle size
important aspects of understanding the fouling dynamics
distribution of different layers of cake formed from a rotary
and control of MF.
filter press and found that the finer particles appeared in the
Since the filtration resistance in microfiltration is
upper layer of the cake. Foley et al. (1992) also found that
mainly determined by the amount and the structure of the
smaller particles in the suspension are preferentially
filter cake when the particles are deposited on the
deposited on the microfiltration membrane operated in
membrane surface, to understand how these factors affected
cross flow mode. Fradin & Field (1999) carried out cross
by operating conditions are the essential steps in grasping
flow microfiltration with suspension of same particle size
the problems of filtration and its optimization. It is
distribution and zeta potential but with different rheological
proposed that the hydrodynamic forces and viscosity acting
and settling properties. They observed different filtration
on the suspended colloids determine the rate of cake
behaviour. Hwang & Lin (2002) carried out the microfiltra-
build up process ( Jio & Sharma 1994; Altman & Ripperger
tion with dual sized submicron particles. They reported
1997). With the given set of operating conditions, easily
increase of presence of large particle fraction resulted in a
doi: 10.2166/wst.2010.112
1734
R. K. Aryal et al. | Influence of buoyant media on particle layer dynamics in microfiltration
Water Science & Technology—WST | 61.7 | 2010
decrease in specific filtration resistance of cake but in an increase in the cake mass which caused increase of pseudo-steady state filtration rate. Hwang & Chen (2008) reported that increase in aeration intensity led to an increase in the filtration flux due to the reduction of cake formation on the membrane surface. They added that higher filtration pressure caused more severe internal blocking of membrane which in turn led to a lower filtration flux. Few researchers (Basu & Huck 2005; Krause et al. 2008) discussed impact of support media in transmembrane pressure development and alteration of organic substances in their microbiological reactor (MBR) system. No experimental discussion has been found so far on the use of buoyant medium in the suspension and its role in cake deposition. Authors believe that the media may potentially enhance the scouring of the membrane surface and change the cake amount and structure by changing size of the
Figure 1
|
Schematic diagram of force acting for particle deposition.
particles deposited. In order to understand the effect of buoyancy media on microfiltration, it is essential to analyse how the flux rate influences the deposition mechanism of
The drag force can be represented by the following Equation (Foley et al. 1992)
the particle on the membrane surface. In this article, the effect of particle size distribution on the cake properties is studied under different flux in
F D ¼ KD Jdf
ð3Þ
presence and absence of buoyant medium. where, KD is a constant and f is a parameter which increases with increasing membrane resistance and J is Theory
the flux.
Figure 1 shows the pattern of the deposition of particles near the membrane surface. As the particle of diameter d approaches the membrane surface, it will be deposited if the membrane drag force exceeds the tangential force exerted
If a particle in contact with membrane is considered, and if it is assumed that the membrane has coefficient of friction, m, then the particle will not remain on the membrane surface if
on the particle. Let FD be the drag force and FT be the tangential force, then the particle deposition occurs if
d.
ð4Þ
ð1Þ
F T # fF D where, f is the frictional factor.
The Equation (4) shows that increase of particle
Assuming that the retentate stream is laminar flow then (Lu & Ju 1989), F T ¼ KT Ud
KD J mf KT U
2
diameter increases the flux and increase of laminar flow rate decreases the particle sizes on the membrane surface.
ð2Þ
Furthermore, the particle deposition is also size selective. Similar selective size deposition has been discussed
where, KT is a constant, d is the diameter of the particle and
elsewhere ( Jio & Sharma 1994; Altman & Ripperger 1997;
U is the flow velocity respectively.
Zhang et al. 2005).
R. K. Aryal et al. | Influence of buoyant media on particle layer dynamics in microfiltration
1735
Water Science & Technology—WST | 61.7 | 2010
The membrane was finally submerged in sodium hypo-
MATERIALS AND METHODS
chlorite solution (200 ppm) for 3 hrs. The membrane
A flat sheet membrane module containing 8 flat sheets
hydraulic resistance (clean membrane flux) was measured
was submerged in 12 L reactor tank in (Figure 2). The
prior to each run and compared with that of a virgin
membrane module is made of polyvinyledene (PVDF) with
membrane. In all runs the flux of the cleaned membrane
pore size 0.14 mm and the total membrane area is 0.2 m2.
was found to be virtually the same as that of virgin membrane.
Membrane filtration was performed from outside to inside
Membrane fouling was studied by analyzing the
in a continuous mode. Bubbles of size 2 –4 cm in diameter
transmembrane pressure (TMP) development, particles in
was injected from the bottom of the reactor at an aeration
suspension with evolution of time and turbidity at two
rate of 2 – 4 m3 h21 m22 (200 – 400 l/h) to create shear
permeate flux rate 30 and 50 L/m2/h at an aeration rate
stresses on the membrane surface to minimise fouling.
3 m3/h/m2. The deposition amount was calculated by
Two processes were evaluated. Process ‘A’ contained kaolin
analysing the turbidity and hence the solid concentration
clay suspension (10 g/L) while Process ‘B’ contained 0.1%
of the suspension and particles layer formation was
fill fraction of buoyant medium in addition to the kaolin
investigated by measuring the particle size in the suspension
clay suspension. The permeate was extracted by a peristaltic
using particle size analyser (Malvern 2600).
pump at a constant flux and discharged to the tank again to maintain a constant volume. The transmembrane pressure (TMP) was recorded at an intervals of 5 minutes with online data acquisition. Each experiment was run for 7 hrs and
RESULTS AND DISCUSSION
samples were collected from the tank (suspension) at an The kaolin clay contained particles with a median particle
hourly interval.
diameter of 4.6 mm with D[v. 0.1] and D[v, 0.9] are 3.96 and
Fill fraction is defined as
5.93 mm respectively. Figure 3 shows the particle size volume occupied by media Fill Fraction ¼ 100 £ reactor volume
ð5Þ
distribution of kaolin clay. The nature of the particles are almost unimodal. During the membrane filtration, the force driving a
The applied buoyant media (anthracite) had density of 1,200 kg/m3. 10 g of anthracite (640 – 2,000 mm) was added in 10 L of water containing kaolin clay to make 0.1% fill fraction. After each run, the membrane was cleaned with a sponge followed by ultrasonification for 3 hrs.
particle of the filter cake surface is a normal hydrodynamic drag force caused by filtration flow into the filter medium. Figure 4 shows the temporal variation of filtration flux in submerged membrane filtration under two filtration flux 30 l/m2/h and 50 l/m2/h at an aeration of 3 m3/h/m2. At the beginning of filtration the filtration flux declines very
Figure 2
|
Experimental set up.
Figure 3
|
Particle size distribution of kaolin clay suspension.
R. K. Aryal et al. | Influence of buoyant media on particle layer dynamics in microfiltration
1736
Figure 4
|
Water Science & Technology—WST | 61.7 | 2010
Time course of filtration rate under flux (a) 30 l/m2/h and (b) 50 l/m2/h at an aeration of 3 m3/h/m2. Figure 5
|
TMP rise (a) at 30 l/m2/h and (b) 50 l/m2/h with evolution of time in absence and presence of buoyant medium at an aeration of 3 m3/h/m2.
quickly due to the particle fouling; however the flux gradually approaches a pseudo-steady state after certain
increase is steep and tend to be linear for higher permeate
time. This is similar to the previous results (Wisniewski et al.
flux rate. For the second zone (II) (t ¼ 1 – 3 hrs), the rate of
2000; Hwang & Chen 2008). The filtration flux attenuation
rise of TMP decreased and the TMP tends to plateau out.
found more rapidly in the beginning at 50 l/m2/h compared
The last zone (III), the plateau extends and almost
to 30 l/m2/h. However, at the later stage the attenuation
saturation exists throughout the experiment. Presence of
rate was almost similar. The attenuation rate in presence of
buoyant medium did not show sharp TMP rise at the
buoyant medium at 30 l/m2/h and 50 l/m2/h was almost
beginning and attained the plateau region quickly. Equation
similar. Further the flux reduction was much lower.
(3) supports the experimental result. The influence of
This indicated that the presence of buoyant medium
permeate flux did not appear on TMP rise in presence
changed the cake thickness, structure as well as porosity.
of buoyant medium.
Figure 5 shows the transmembrane pressure develop-
Figure 6 shows an average particle diameter in suspen-
ment with evolution of time at two flux rates in presence
sion and particle deposition on the membrane surface with
and absence of buoyancy media. Clear difference in TMP
evolution of time at two flux rates 30 and 50 l/m2/h.
rise in presence and absence of buoyancy media observed.
The particle diameter results showed that at the beginning,
Incorporation of buoyant medium reduced the TMP
fine particles deposited on the membrane surface followed
development by two to three folds. For the experiment in
by the deposition of coarser particles. After 7 hours of
absence of buoyant medium, the transmembrane pressure
operation, the mean diameters of the particles in suspension
development can be divided into three parts. During the
were 5.14 mm and 5.63 mm at 30 l/m2/h and 50 l/m2/h,
first one (I), t ¼ 0 to t ¼ 1 hr, the TMP increases. This initial
respectively. The mean particle diameter in the suspension
R. K. Aryal et al. | Influence of buoyant media on particle layer dynamics in microfiltration
1737
Water Science & Technology—WST | 61.7 | 2010
The particle deposition on the membrane surface was significantly different in the presence and in the absence of buoyant medium. At flux rates of 30 and 50 l/m2/h, the particle deposition was 40.7 and 48.3 g/m2 in absence of medium whereas in the presence of buoyant medium, the deposition was 13.9 and 18.1 g/m2, respectively. The result indicated that the presence of the medium reduced the membrane fouling significantly. Random deposition of the particles with different size seemed to make the cake layer less compact that allows the flux to pass though the membrane and thus reduces the transmembrane pressure. The presence of buoyant medium reduced the membrane fouling significantly. This seemed to have caused low TMP rise during operation and low flux decline. Similar particle diameter appeared in the suspension at an early stage of membrane operation was mainly due to higher concentration of kaolin clay in the suspension. Decrease of concentration differentiated the particle sizes. This was supported by the experiment at 50 l/m2/h where rapid deposition occurred leaving a less concentrated Kaolin clay suspension. Figure 6
|
2
2
Mean particle diameter in suspension at (a) at 30 l/m /h and (b) 50 l/m /h at an aeration of 3 m3/h/m2.
increased with evolution of time and after 7 hrs (420 minutes). This results shows that increase of permeate flux
CONCLUSIONS
increased the drag force resulting more fine particles to be
Influence of buoyant medium in submerged microfiltration
deposited on the membrane surface quickly. This is
process was studied using kaolin clay suspension of particle
supported by the Equation (4). However, application of
size with median diameter 4.6 mm. Buoyant medium of
buoyant medium did not alter the mean particle diameter
anthracite of particle size 640 –2,000 mm was used with a fill
even after 7 hrs operation. The particle deposition on the
fraction of 0.1%. Flux decline was observed with time and
membrane surface in presence of buoyant medium indi-
attenuation rate was higher at higher permeate flux in
cated that particles of different sizes are deposited randomly
absence of the medium. TMP development was higher at
on the membrane surface.
higher flux. Presence of the buoyant medium reduced the
Since, higher degree of finer particles deposition was
flux decline as well as TMP by two to three folds. Increase in
noticed at the beginning of membrane filtration operation,
mean particle diameter in suspension was observed with
one can assume that compact cake layer is formed due to
time indicated the deposition of finer fraction on the
finer particles deposition. This resulted in the initial decline
membrane surface. Higher flux caused finer particles
of permeate flux with time. This effect is more pronounced
deposition rapidly. Incorporation of the medium did not
at higher permeate rate which is further supported by TMP
alter the mean particle diameter in suspension during
data. It is seen that the TMP development at the beginning
the experimental run time showing a well scouring and
is higher which is possibly due to finer particles deposition
mixing. The deposition was not preferential in terms of
on the membrane surface.
particle sizes.
1738
R. K. Aryal et al. | Influence of buoyant media on particle layer dynamics in microfiltration
REFERENCES Altman, J. & Ripperger, S. 1997 Particle deposition and layer formation at the crossflow microfiltraiton. J. Membr. Sci. 124, 119– 128. Basu, O. D. & Huck, P. M. 2005 Impact of support media in an integrated biofilter-submerged membrane system. Water Res. 39, 4220 –4228. Foley, G., MacLoughlin, P. F. & Malone, D. M. 1992 Preferential deposition of smaller cells during cross-flow microfiltration of yeast suspension. Biotechnol. Tech. 6(2), 115– 120. Fradin, B. & Field, R. W. 1999 Crossflow microfiltration of magnesium hydroxide suspensions: determination of critical fluxes, measurement and modelling of fouling. Sep. Purif. Technol. 16, 25 –45. Hwang, K. J. & Chen, F. F. 2008 Modeling of particle fouling and membrane blocking in submerged membrane filtration. Sep. Sci. Technol. 42, 2595 – 2614.
Water Science & Technology—WST | 61.7 | 2010
Hwang, K. J. & Lin, K. P. 2002 Crossflow microfiltration of dual sized submicron particles. Sep. Sci. Technol. 37(10), 2231 –2249. Jio, D. & Sharma, M. M. 1994 Mechanism of cake buildup in crossflow filtration of colloidal suspensions. J. Colloid Interface Sci. 162, 454 –462. Judd, S. J. 2006 Permeable permutations - membrane options for MBRs. Water 21 (June 2006), 20 –21. Krause, S, Tournier, R., Cornel, P. & Siembida, B. 2008 Granulatedriven fouling control in a submerged membrane module for MBR application. World Water Congress, 2008 Vienna. Lu, W. M. & Ju, S. C. 1989 Selective particle deposition in cross flow filtration. Sep. Sci. Technol. 517(7– 8), 517– 540. Wisniewski, C., Grasmick, A. & Cruz, A. L. 2000 Critical particle size in membrane bioreactors. Case of denitrifying bacterial suspension. J. Membr. Sci. 178, 141 –150. Zhang, J., Zhang, J. P. & Fan, L. S. 2005 Effect of particle size ratio on the drag force of an interactive particle. Chem. Eng. Res. Des. 83(A4), 339 –343.
