Separation and Purification Technology xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Separation and Purification Technology

journal homepage: www.elsevier.com/locate/seppur

Compression and expansion properties of filter cake accompanied with

step change in applied pressure in membrane filtration
Eiji Iritani ⇑, Nobuyuki Katagiri, Genki Inagaki
Department of Chemical Engineering, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8603, Japan

a r t i c l e i n f o a b s t r a c t

Article history: In order to well elucidate the true nature of filter cake formed in membrane filtration of aqueous colloids,
Received 31 August 2016 the transient flux decline behaviors accompanied with step-up and step-down changes in the applied fil-
Received in revised form 26 September tration pressure during the course of constant pressure filtration were investigated by carrying out both
2016
microfiltration experiments of kaolin slurry and ultrafiltration experiments of nanosilica sol under vari-
Accepted 25 November 2016
Available online xxxx
ous experimental conditions. When the applied pressure was increased stepwise, for each colloid the plot
of the reciprocal filtration rate vs. the filtrate volume per unit membrane area referred to as the Ruth plot
tends to approach a linear relationship obtained in constant pressure filtration conducted under the
Keywords:
Membrane filtration
increased pressure condition from the beginning of filtration, after a transient period when the preformed
Filter cake cake under the initial pressure was compressed. In contrast, in the case of the step-down change in the
Irreversibility applied pressure, when the volume of the preformed cake is adequate, the plot after a transient period
Compressive pressure tends to approach a higher straight line parallel to a linear relationship for constant pressure filtration
Expansion conducted under the decreased pressure condition from the beginning. Such flux decay behaviors were
well described by the expression having two constants: one is the constant representing the final filter-
ability for the preformed filter cake, and the other is the rate constant describing the change rate in the
structure of preformed cake resulting from a pressure jump. Moreover, the influence of experimental
conditions on the cake reversibility and the rate constant was revealed both for kaolin slurry and for
nanosilica sol.
Ó 2016 Elsevier B.V. All rights reserved.

1. Introduction The expansion properties of filter cake are of significance in the

Membrane filtration of aqueous colloids has a broad variety of
applications in widely diversified fields ranging from industry to
environmental protection. One of the critical issues obstructing
more widely spread use of membrane filtration may be a signifi-
cant flux decline occurring with the progress of filtration, arising
from the membrane fouling. Especially, when a highly compress-
ible filter cake forms on the membrane surface during filtration,
the increase in the applied filtration pressure does not necessarily
lead to a noticeable increase in the overall performance of mem-
brane filtration [1]. Therefore, an understanding of properties of fil-
ter cake such as the cake compressibility is of paramount
significance to controlling filtration behaviors. Until now, the prop-
erties of compressible filter cakes have been evaluated by a num-
ber of researchers through pressure filtration tests [2–9],
compression-permeability (C-P) cell tests [10–14], and centrifuga-
tion tests [15–19], etc.
⇑ Corresponding author.
E-mail address: iritani@nuce.nagoya-u.ac.jp (E. Iritani).
practical applications such as the rebound of water content of cake
at the time of pressure release seen in the cake discharge [20] and
variable pressure filtration in which the applied pressure varies
rapidly with time, in addition to the basic understanding of cake
properties. However, the expansion properties of filter cake have
been less examined compared to the work on the compression
behaviors [21,22], with the exception of the studies on the expan-
sion behaviors of compressed cakes produced by expression oper-
ation [23,24]. Okamura and Shirato [25] investigated the temporal
variation of the filtration rate in stepwise pressure filtration of
slurry for ignition plug and Gairome clay when the applied filtra-
tion pressure stepwise increased or decreased during the course
of cake filtration operation conducted under a constant pressure
condition. As a result, they found that the constant pressure filtra-
tion coefficient defined by Ruth [26,27] after the pressure increase
was closely accorded with that in constant pressure filtration con-
ducted at the increased pressure from the beginning of filtration,
but that the coefficient after pressure decrease was in disagree-
ment with that in constant pressure filtration conducted at the
decreased pressure from the beginning. Subsequently, Shirato

http://dx.doi.org/10.1016/j.seppur.2016.11.067
1383-5866/Ó 2016 Elsevier B.V. All rights reserved.

Please cite this article in press as: E. Iritani et al., Compression and expansion properties of filter cake accompanied with step change in applied pressure in
membrane filtration, Separ. Purif. Technol. (2016), http://dx.doi.org/10.1016/j.seppur.2016.11.067
2 E. Iritani et al. / Separation and Purification Technology xxx (2016) xxx–xxx

Nomenclature

ds mean specific surface area size of particles (m) Greek letters

K Ruth coefficient of constant pressure filtration (m2/s) aav average specific cake resistance (m/kg)
Ke value of K corresponding to cake structure finally a1 empirical constant in Eq. (1) (m/kg)
exhibiting under later pressure for filer cake preformed b empirical constant in Eq. (2) ()
at initial pressure (m2/s) Dpc pressure drop across filter cake (Pa)
m ratio of mass of wet to mass of dry cake () eav average cake porosity ()
n empirical constant in Eq. (1) () e1 empirical constant in Eq. (2) ()
p applied filtration pressure (Pa) g rate constant in Eq. (7) (s1)
pa empirical constant in Eqs. (1) and (2) (Pa) h filtration time (s)
Rc resistance of filter cake (m1) ht filtration time when filtration pressure changes (s)
Rt total filtration resistance to filtrate flow (m1) l viscosity of filtrate (Pa s)
s mass fraction of particles in colloid () q density of filtrate (kg/m3)
t elapsed filtration time after pressure change (s) qs true density of particles (kg/m3)
v cumulative filtrate volume per unit effective membrane
area (m3/m2) Subscripts
vt cumulative filtrate volume per unit effective membrane 1 initial stage
area when filtration pressure changes (m3/m2) 2 later stage

et al. [28] obtained a similar result for dead-end ultrafiltration of
nanosilica sol. Reihanian et al. [29] examined the flux decline
behaviors when the filtrate flow was interrupted due to filtration
cell depressurization during the course of filtration in ultrafiltra-
tion of bovine serum albumin (BSA) solutions.
The focus of the present study is to delineate the transient flux
decline behaviors in dead-end membrane filtration under step-up
pressure and step-down pressure conditions, in order to obtain
an in-depth understanding of the nature of filter cake formed on
the membrane surface in membrane filtration of colloids. Mem-
brane filtration experiments are performed in the dead-end mode
using kaolin slurry and nanosilica sol forming compressible filter
cakes. The empirical filtration rate equation is proposed in this
paper, and two adjustable fitting parameters in the equation are
examined by carrying out the membrane filtration experiments
under various experimental conditions.
2. Materials and methods
2.1. Materials
The materials used in the experiments were kaolin (Sigma-
Aldrich) and nanosilica sol (ST-XS in Snowtex (ST)-series, Nissan
Chemical Industries). Test colloids were prepared by dispersing
preweighed quantities of the powder or solution in ultrapure,
deionized water (resistivity of at least 18 MX cm) produced by
purifying tap water through ultrapure water systems equipped
both with Elix-UV20 and with Milli-Q Advantage (Bio-POD) for lab-
oratory use (Millipore). The particle concentration by weight frac-
tion was prepared at 0.15 and 0.30 for kaolin slurry and at 0.03 for
nanosilica sol. The concentration of kaolin slurry was set to much
higher values than that of nanosilica sol in order to avoid precipi-
tation of kaolin particles. Table 1 lists the mean specific surface
area size ds and true density qs of colloidal particles, and the prop-
erties of filter cake. The mean specific surface area sizes of kaolin
and nanosilica particles shown in the table were calculated from
the particle size distributions measured by a laser diffraction par-
ticle size analyzer (SALD-2200, Shimadzu Corp.) and a dynamic
light scattering (DLS) photometer (DLS-8000, Otsuka Electronics),
respectively. The true density of particles was measured by a pyc-
nometer. The average specific cake resistance aav and the average
cake porosity eav are, respectively, related to the pressure drop
Dpc across the filter cake as [9,30]
 n
Dpc
aav ¼ a1 1 þ ð1Þ
pa
 
Dp b
1  eav ¼ ð1  e1 Þ 1 þ c ð2Þ
pa
where a1, pa, n, e1, and b are the empirical constants. The values of
these constants were determined with reference to dead-end mem-
brane filtration methods outlined elsewhere [9,31] and listed in the
table. It is found that the filter cakes of both kaolin and nanosilica
sol exhibit compressible behaviors.
The membranes employed are mixed cellulose ester microfiltra-
tion membranes (Advantec Toyo Corp.) with a nominal pore diam-
eter of 0.1 lm for microfiltration of kaolin slurry and regenerated
cellulose ultrafiltration membranes (Millipore) with a nominal
molecular weight cutoff of 10 kDa for ultrafiltration of nanosilica
sol. The complete rejection of colloidal particles was assured
throughout the course of membrane filtration with the use of these
membranes.
2.2. Experimental apparatus and technique
Membrane filtration experiments were conducted in the dead-
end mode by using unstirred batch filtration cells with effective
membrane areas of 19.40 and 24.61 cm2 for microfiltration and
ultrafiltration, respectively. The filtration cells were fabricated in
our laboratory. In the filtration cell for microfiltration experiments,

Table 1
Properties of test colloids.

Material ds (lm) qs (g/cm3) a1 (m/kg) pa (kPa) n (–) e1 (–) b (–)

Kaolin 3.09 2.60 7.57  1011 28.4 0.753 0.772 0.231
Nanosilica sol 0.0048 2.27 2.07  1014 15.9 0.629 0.977 0.524

Please cite this article in press as: E. Iritani et al., Compression and expansion properties of filter cake accompanied with step change in applied pressure in
membrane filtration, Separ. Purif. Technol. (2016), http://dx.doi.org/10.1016/j.seppur.2016.11.067
 E. Iritani et al. / Separation and Purification Technology xxx (2016) xxx–xxx 3

a little ingenuity for a perforated plate supporting the microfiltra- x103

tion membrane was used so as to reduce the resistance to the fil- 4
(a)
trate flow through the very thin membrane. The membrane Kaolin Step-up pressure
filtration experiments were performed by adjusting the applied fil- s = 0.15 p1 = 49 kPa, p2 = 490 kPa
tration pressure automatically using by a computer-driven elec- vt = 1.45 cm
3
vt = 2.69 cm
tronic pressure regulator with the application of compressed
vt = 4.76 cm
nitrogen gas. The step-up and step-down pressure filtration exper-

d /dv (s/cm)
Cal.
iments were carried out under initial constant pressure condition
followed by a step change in the applied filtration pressure and 2
then continued under the changed constant pressure condition.
The filtrate was accumulated in a reservoir placed on an elec-
tronic balance (Shimadzu Corp.) connected to a personal computer 1
to collect and record filtrate mass vs. filtration time data during a Constant pressure
p = 49 kPa
filtration run. The weights were converted to volumes with the p = 490 kPa
vt vt vt
use of density correlations to obtain the temporal variation of the
cumulative filtrate volume per unit effective membrane area and 0
0 2 4 6 8 10
thus the dependence of the reciprocal filtration rate on the filtrate v (cm)
volume per unit membrane area. For comparison, the conventional
constant pressure membrane filtration experiments in which the
applied filtration pressure remained constant throughout the x103
14
course of filtration were also conducted. Step-down pressure (b)
p1 = 490 kPa, p2 = 49 kPa
12 vt = 0.65 cm
vt = 1.43 cm
3. Results and discussion vt = 2.59 cm
10
vt = 4.27 cm

3.1. Transient flux decline behaviors in microfiltration of kaolin slurry
Fig. 1 illustrates the flux decline behaviors represented as the
form of reciprocal filtration rate (dh/dv) against the filtrate volume
v per unit effective membrane area referred to as the Ruth plot
[26,27] in dead-end filtration of kaolin slurry accompanied with
a step change in the applied filtration pressure at a specified filtrate
volume vt per unit effective membrane area in the filtration pro-
cess. Fig. 1(a) and (b) corresponds to transient filtration behaviors
in step-up and step-down changes in the applied pressure, respec-
tively. Also included is the plot for the conventional constant pres-
sure filtration conducted under each constant pressure condition
throughout the course of filtration. The plot of dh/dv vs. v for con-
stant pressure filtration shows a linear relationship with a slope
corresponding to each filtration pressure of 49 and 490 kPa in
accordance with the Ruth filtration rate equation on the assump-
tion of a negligible membrane resistance compared to the cake
resistance, as represented by [26,27]
dh 2
¼ v
dv K
where K is the Ruth coefficient of constant pressure filtration
defined by
lqsaav
K¼
2pð1  msÞ
where l is the viscosity of filtrate, q is the density of filtrate, s is the
mass fraction of particles in colloid, p is the applied filtration pres-
sure, and m is the average ratio of the mass of wet to the mass of dry
cake related to the average cake porosity eav as
qeav
m¼1þ
qs ð1  eav Þ
The slope of the straight line decreases as the filtration pressure
increases.
In step-up pressure filtration, the applied pressure is suddenly
increased from 49 to 490 kPa at the filtration times ht of 15, 45,
and 140 min corresponding to vt of 1.45, 2.69, and 4.76 cm, respec-
tively, as shown in Fig. 1(a). For instance, when the applied pres-
sure increases stepwise from the initial pressure of 49 kPa to the
later pressure of 490 kPa at vt of 4.76 cm, the reciprocal filtration
d /dv (s/cm)
Cal.
8
Constant pressure
6 p = 49 kPa
p = 490 kPa
4
Kaolin
s = 0.15
vt
2 vt
vt vt
0
0 2 4 6 8 10
v (cm)
Fig. 1. Relation between reciprocal filtration rate and filtrate volume per unit
effective membrane area in dead-end microfiltration of kaolin slurry: (a) step-up
pressure filtration, (b) step-down pressure filtration.
rate (dh/dv) declines from 3.96  103 to 4.0  102 s/cm instanta-
neously in response to the increase in the applied pressure. The
ratio of the reciprocal filtration rate immediately after the pressure
ð3Þ
jump to that immediately before the jump is exactly the same as
the ratio of the initial pressure to the later pressure, i.e., 0.10. This
suggests that the filtration rate changes immediately after the
pressure jump only due to the contribution of the increase in the
driving force since the structure of preformed cake does not change
ð4Þ immediately. However, dh/dv increases readily, thereafter the
increase in dh/dv becomes gradual, and finally the plot is in almost
agreement with a linear relationship obtained in constant pressure
filtration conducted at the pressure of 490 kPa from the beginning
of filtration. This fact indicates that the filter cake formed at the
pressure of 49 kPa is compressed under the pressure of 490 kPa
as well as the cake growth after the pressure jump because the fil-
ð5Þ ter cake comprised of kaolin particles exhibits a compressible
behavior. Consequently, the final cake structure in this step-up
pressure filtration appears nearly identical to that in constant pres-
sure filtration conducted at the increased pressure from the
beginning.
However, in step-down pressure filtration in which the applied
pressure decreases stepwise during the course of constant pressure
filtration, the flux decline behaviors are quite different from those
seen in step-up pressure filtration. Fig. 1(b) illustrates the results
for step-down pressure filtration in which the applied pressure is

Please cite this article in press as: E. Iritani et al., Compression and expansion properties of filter cake accompanied with step change in applied pressure in
membrane filtration, Separ. Purif. Technol. (2016), http://dx.doi.org/10.1016/j.seppur.2016.11.067
4 E. Iritani et al. / Separation and Purification Technology xxx (2016) xxx–xxx

suddenly decreased from 490 to 49 kPa inversely with the condi- x1012
tion shown in Fig. 1(a) at ht of 1, 5, 15, and 40 min corresponding 1.0
(a)
to vt of 0.65, 1.43, 2.59, and 4.27 cm, respectively. For example, Step-up pressure
p1 = 49 kPa, p2 = 490 kPa
at the moment the applied pressure decreases stepwise from vt = 1.45 cm
0.8
490 kPa to 49 kPa at vt of 4.27 cm, dh/dv readily increases from vt = 2.69 cm Constant pressure
1.17  103 to 1.17  104 s/cm. In this case, the value of the recipro- vt = 4.76 cm p = 49 kPa
Cal. p = 490 kPa
cal filtration rate immediately after the pressure jump divided by 0.6

Rt (cm-1)
that immediately before the jump is exactly equal to the ratio
of the initial pressure to the later pressure, i.e., 10.0. Thereafter,
dh/dv slightly increases with the progress of filtration. In sharp 0.4
contrast to the behaviors appeared in step-up pressure filtration,
the relation between dh/dv and v is in disagreement with a linear
0.2
relationship obtained in constant pressure filtration carried out Kaolin
at the pressure of 49 kPa. The value of dh/dv is much larger than vt vt s = 0.15
vt
that in constant pressure filtration at the same value of v and even- 0
tually tends to increase with v parallel to the straight line for con- 0 2 4 6 8 10
stant pressure filtration at 49 kPa. This result clearly indicates that v (cm)
the cake structure formed in step-down pressure filtration in
x1012
which the pressure was reduced from 490 to 49 kPa is considerably 1.0
different from that obtained in constant pressure filtration at the Step-down pressure (b)
p1 = 490 kPa, p2 = 49 kPa Kaolin
pressure of 49 kPa. The filter cake preformed at the pressure of vt = 0.65 cm s = 0.15
490 kPa in step-down pressure filtration swells and expands due 0.8 vt = 1.43 cm
to the substantial decrease in the solid compressive pressure vt = 2.59 cm
vt = 4.27 cm
caused by the sudden decrease in the filtration pressure because Cal.
0.6

the filter cake exhibits a compressible behavior. However, it should
be stressed that the structure of the filter cake is not fully reversi-
ble for the dramatic decrease in the filtration pressure, in sharp
contrast to the case of the dramatic increase in the filtration
pressure.
The reciprocal filtration rate (dh/dv) shown in Fig. 1 is influ-
enced not only by the filter cake resistance controlled by the cake
properties such as the specific cake resistance and cake porosity
but also by the applied filtration pressure p as the driving force
in filtration, as expressed by the Darcy’s law in the form [32]:
dh lRt
¼ ð6Þ
dv p membrane area in dead-end microfiltration of kaolin slurry: (a) step-up pressure
where Rt is the total filtration resistance. In the case of negligible
membrane resistance assumed in this study, the resistance Rt
equals the resistance Rc of filter cake. As a result, the temporal
behaviors of cake resistance can be evaluated by replotting the rela-
tion between Rt and v from the data shown in Fig. 1 based on Eq. (6).
In the conversion from dh/dv to Rt, p1 and p2 before and after a pres-
sure jump, respectively, are utilized as the value of p in Eq. (6). In
Fig. 2, the total filtration resistance Rt is plotted against v based
on the data shown in Fig. 1. In this figure, the properties of filter
cake may be most obvious. In step-up pressure filtration shown in
Fig. 2(a), the resistance Rt just after a pressure increase very rapidly
increases and tends to approach a straight line for constant pressure
filtration of 490 kPa, whereas in step-down pressure filtration seen
in Fig. 2(b), the resistance Rt begins departing from the linear rela-
tionship for constant pressure filtration of 490 kPa, but Rt remains
high compared to a linear relationship for constant pressure filtra-
tion of 49 kPa, indicating the irreversibility of structure of filter cake
observed in the sudden decrease in the applied pressure.
3.2. Description of transient flux decline behaviors
Based on the above-mentioned transient flux decline behaviors
in step-up and step-down pressure filtration of kaolin slurry, the
relation between dh/dv and v after a pressure jump may be empir-
ically described by
 
dh dt p1 2 2 2 2 dh
¼ ¼   expðgtÞv t þ v t þ ðv  v t Þ ð7Þ
dv dv p2 K 1 K e Ke K2
Rt (cm-1)
0.4
0.2 Constant pressure
p = 49 kPa
vt vt
vt p = 490 kPa
vt
0
0 2 4 6 8 10
v (cm)
Fig. 2. Relation between total filtration resistance and filtrate volume per unit
filtration, (b) step-down pressure filtration.
where t is the elapsed filtration time after a pressure change, K1 and
K2 are the Ruth coefficients of constant pressure filtration at the
pressures of p1 and p2 defined as Eq. (4), respectively. The constant
Ke denotes K-value corresponding to the cake structure finally
exhibiting under the later pressure p2 for the filter cake preformed
at the initial pressure p1 and is a measure of the irreversibility of fil-
ter cake when the pressure change occurs. The constant g is the rate
constant describing the change rate in the structure of the pre-
formed cake from the state at the initial pressure p1 to that at the
later pressure p2. Therefore, the first, second, and third terms in
Eq. (7) represent the contribution to dh/dv corresponding to the
temporal variation of the resistance of filter cake preformed at the
initial pressure, the contribution to dh/dv corresponding to the final
resistance of filter cake preformed at the initial pressure, the contri-
bution to dh/dv corresponding to the resistance of filter cake depos-
ited after the pressure change, respectively. It is interesting to note
that Eq. (7) is applicable both to step-up pressure filtration and to
step-down pressure filtration. When vt equals zero, Eq. (7) reduces
to the conventional constant pressure filtration rate Eq. (3). The
coefficient Ke is equal to K2 for the compressible cake with the prop-
erties reversible for a pressure change. Moreover, when the
response to a jump in the applied pressure is instantaneous, i.e.,
g = 1, Eq. (7) reduces to
2
¼ v ð8Þ
dv K 2

Please cite this article in press as: E. Iritani et al., Compression and expansion properties of filter cake accompanied with step change in applied pressure in
membrane filtration, Separ. Purif. Technol. (2016), http://dx.doi.org/10.1016/j.seppur.2016.11.067
 E. Iritani et al. / Separation and Purification Technology xxx (2016) xxx–xxx 5

The variations with time t of v can be obtained by solving the
ordinary differential Eq. (7) using the Runge-Kutta-Gill scheme
under the initial condition of v = vt at t = 0. As a result, the relation
between dt/dv (= dh/dv) and v can be also obtained. The two con-
stants Ke and g are determined by fitting the numerical calcula-
tions of Eq. (7) to the experimental flux decline data shown in
Fig. 1. The solid curves in Fig. 1 are the calculations based on
Eq. (7) and in reasonably agreement with the plots for both step-
up and step-down pressure filtration operations. In a similar
way, the solid curves shown in Fig. 2 are calculated by combining
Eq. (6) with Eq. (7) and well describes the experimental data.
Based on Eq. (7), the reversibility of the preformed filter cake
resulting from a sudden change in the applied filtration pressure
may be more rigorously evaluated by
p1
 K1e
Reversibility ½% ¼
p2 K 1
 100
p1
p2 K 1
 K12
The cake reversibility defined by the above equation represents the
ratio of the ultimate variation of the resistance of the preformed
cake due to a pressure jump to that of the resistance in the reversi-
ble cake in which the cake structure corresponding to the later pres-
sure is attained ultimately. Fig. 3 illustrates the plots of the cake
reversibility defined by Eq. (9) and the rate constant g in Eq. (7)
against the product of the mass fraction s of particles in slurry
and vt. The results for s of 0.30 are also included in the figure, in
addition to the values calculated from the results shown in Fig. 1.
It is shown that the structure of filter cake in step-up pressure filtra-
tion is perfectly reversible in a pressure jump, as shown in Fig. 3(a).
In contrast, in step-down pressure filtration, the cake reversibility
becomes lower with increasing svt. The cake reversibility decreases
as low as 7.2% at svt of 0.92. It should be stressed that the plots show
a similar tendency irrespective of the difference of the slurry con-
centration s when plotted in this manner. It is concluded that the
cake reversibility of the preformed cake decreases with increasing
cake thickness since the cake thickness may vary linearly with svt.
As can be seen in the semi-logarithmic plots in Fig. 3(b), the rate
constant g decreases with increasing svt, indicating slower expan-
sion/compression of the preformed cake due to the additional cake
ð9Þ formation. Also in this case, the change in g is represented by func-
tions of svt. The figure clearly indicates that the rate constant g for
step-up pressure filtration is much higher than that for step-down
pressure filtration, implying that the structure of filter cake changes
more promptly in step-up pressure filtration. This result is greatly
similar to that reported previously for compression and expansion
behaviors of a granular bed comprised of superabsorbent hydrogels
[24].
3.3. Effect of diminution in pressure on expansion of cake

(a)
In order to examine the effect of diminution in the applied pres-
100
sure on the expansion behaviors of the preformed filter cake after
the pressure decrease, the resistance Rt is plotted in Fig. 4 against v
80 for step-down pressure filtration in which the pressure changes
from 490 to 294 kPa and from 490 to 98 kPa. In step-down pres-
sure filtration in which the diminution rate is insignificant, the plot
Reversibility (%)

60 after a pressure decrease accords closely with that of constant

pressure filtration conducted at the later pressure from the begin-
ning of a run. In contrast, in step-down pressure filtration accom-
40 Kaolin
s = 0.15, s = 0.30 panied with a significant pressure decrease from 490 to 98 kPa, the
Step-up , plot after a pressure decrease is in disagreement with that of con-
p1 = 49 kPa
p2 = 490 kPa stant pressure filtration conducted at the pressure of 98 kPa up to
20
Step-down , the end of test conducted in this study. The solid curves are the cal-
p1 = 490 kPa
p2 = 49 kPa culations and then the values of Ke and g can be determined. Fig. 5
0 plots the cake reversibility and the rate constant g in step-down
0 0.2 0.4 0.6 0.8 1.0 pressure filtration against the pressure ratio (p1/p2) defined as
svt (cm) the ratio of the initial pressure p1 to the later pressure p2. As can
be seen in Fig. 5(a), the semi-logarithmic plot of the reversibility
Kaolin (b)
-1
s = 0.15, s = 0.30
10 Step-up , x1012
p1 = 49 kPa 1.0
p2 = 490 kPa Step-down pressure
Step-down , p1 = 490 kPa
p1 = 490 kPa
10 -2
p2 = 49 kPa 0.8 p2 = 98 kPa
p2 = 294 kPa
Cal.
Constant pressure
(s-1)

0.6
Rt (cm-1)

10-3 p = 98 kPa
p = 294 kPa
p = 490 kPa
0.4
10-4

0.2
Kaolin
-5 vt s = 0.15
10
0 0.2 0.4 0.6 0.8 1.0 0
svt (cm) 0 1 2 3 4 5 6
v (cm)
Fig. 3. Dependence of parameters in Eq. (7) on product of mass fraction of particles
in slurry and filtrate volume per unit membrane area when filtration pressure Fig. 4. Relation between total filtration resistance and filtrate volume per unit
changes stepwise in dead-end microfiltration of kaolin slurry: (a) cake reversibility, membrane area in dead-end microfiltration of kaolin slurry in which pressure
(b) rate constant. changes from 490 to 294 kPa and from 490 to 98 kPa.

Please cite this article in press as: E. Iritani et al., Compression and expansion properties of filter cake accompanied with step change in applied pressure in
membrane filtration, Separ. Purif. Technol. (2016), http://dx.doi.org/10.1016/j.seppur.2016.11.067
6 E. Iritani et al. / Separation and Purification Technology xxx (2016) xxx–xxx

100 x105
(a) 1.2
(a)
Nanosilica sol Step-up pressure
80 s = 0.03 p1 = 49 kPa, p2 = 490 kPa
1.0 vt = 0.84 cm
vt = 1.40 cm
Reversibility (%)

vt = 1.78 cm
60 0.8 vt = 2.26 cm

dθ/dv (s/cm)
Cal.
0.6
40
Kaolin
s = 0.15 0.4
20 vt = 2.56 2.86 cm
Step-down pressure Constant pressure
p1 = 490 kPa 0.2 p = 49 kPa
0 0 p = 490 kPa
vt vt vt vt
10 101 102
0
p1/p2 ( ) 0 1 2 3 4 5
v (cm)
10-3
x105
(b)
Step-down pressure (b)
p = 490 kPa, p2 = 49 kPa
3 1 v = 0.82 cm
t
vt = 1.42 cm
vt = 2.16 cm
vt = 2.60 cm
(s-1)

dθ/dv (s/cm)
vt = 3.58 cm
10-4 2 Cal.

Kaolin
s = 0.15
vt = 2.56 2.86 cm
Step-down pressure
p1 = 490 kPa
10-5 0
10 10 1
10
p1/p2 ( )
Fig. 5. Dependence of parameters in Eq. (7) on pressure ratio defined as ratio of
initial pressure to later pressure: (a) cake reversibility, (b) rate constant.
vs. the pressure ratio (p1/p2) can be approximated by a straight line
with a negative slope, indicating that a larger pressure decrease
leads to a more severe irreversibility of the cake structure. The log-
arithmic plots shown in Fig. 5(b) indicates that the rate constant g
decreases with increasing pressure ratio (p1/p2) and that the rate of
decrease becomes less noticeable as the ratio (p1/p2) increases.
3.4. Transient flux decline behaviors in ultrafiltration of silica sol
The flux decline behaviors of step-up and step-down pressure
ultrafiltration of nanosilica sol are shown in the form of dh/dv vs.
v in Fig. 6(a) and (b), respectively. The solid curves are the calcula-
tions based on Eq. (7) and describe well the experimental data.
Also in nanosilica sol, the flux decline behaviors are similar to
those for microfiltration of kaolin slurry. It is more clear from
Fig. 6(b) that the plots after a pressure decrease from 490 to
49 kPa show the straight lines with the same slope as that of a
straight line obtained in constant pressure filtration. Fig. 7 shows
the plots of the cake reversibility and the rate constant g vs. svt
in step-up and step-down pressure ultrafiltration of nanosilica
sol. In step-down pressure filtration, the filter cake exhibits a com-
plete reversibility at the values of svt smaller than about 0.05, as
seen in Fig. 7(a). As the value of svt further increases, the cake irre-
versibility resulting from the step-down pressure becomes more
marked. However, the degree of the cake irreversibility is insignif-
icant compared to the case of kaolin slurry. It is found from Fig. 7
(b) that the rate constant g substantially decreases with increasing
svt. It is interesting to note that the semi-logarithmic plots of g vs.
Nanosilica sol
s = 0.03 Constant pressure
p = 49 kPa
1 p = 490 kPa
vt
vt
2 vt vt vt
0
0 1 2 3 4 5
v (cm)
Fig. 6. Relation between reciprocal filtration rate and filtrate volume per unit
membrane area in dead-end ultrafiltration of nanosilica sol: (a) step-up pressure
filtration, (b) step-down pressure filtration.
svt can be represented by a unique curve irrespective of step-up or
step-down pressure ultrafiltration. The difference in the behaviors
of cake reversibility and creeping action between kaolin slurry and
silica sol may depend largely on the difference in the rheological
properties of the filter cake, which are closely related to the net-
worked structure of the compressed cake [20,33]. More detailed
study using a variety of slurries which exhibit different rheological
behaviors may provide additional insights into the mechanisms
which govern the dynamic response of filtration behaviors accom-
panied with the step change in the pressure. The dependence of fil-
tration behaviors on the rheological properties of the compressed
cake is the subject of an ongoing investigation.
4. Conclusions
The dynamic response of filtration rate to stepwise changes in
the applied filtration pressure during the course of filtration were
investigated for microfiltration of kaolin slurry and ultrafiltration
of nanosilica sol. When the pressure was increased stepwise, the
structure of the filter cake formed under the initial pressure tended
to approach that of the cake formed in filtration conducted under
the later pressure from the beginning of filtration. However, in
the many cases of step-down pressure conditions, the compressed
cake formed under the initial high pressure was not completely
expanded to the state corresponding to the later low pressure.
The filtration rate equation with two constants describing the final

Please cite this article in press as: E. Iritani et al., Compression and expansion properties of filter cake accompanied with step change in applied pressure in
membrane filtration, Separ. Purif. Technol. (2016), http://dx.doi.org/10.1016/j.seppur.2016.11.067
 E. Iritani et al. / Separation and Purification Technology xxx (2016) xxx–xxx 7

[4] M. Shirato, M. Sambuichi, H. Kato, T. Aragaki, Internal flow mechanism in filter

(a) cakes, AIChE J. 15 (1969) 405–409.
100 [5] F.M. Tiller, T.C. Green, Role of porosity in filtration IX Skin effect with highly
compressible materials, AIChE J. 19 (1973) 1266–1269.
[6] W.M. Lu, K.L. Tung, S.M. Hung, J.S. Shiau, K.J. Hwang, Constant pressure
80 filtration of mono-dispersed deformable particle slurry, Sep. Sci. Technol. 36
(2001) 2355–2383.
[7] E. Vorobiev, Derivation of filtration equations incorporating the effect of
Reversibility (%)

pressure redistribution on the cake-medium interface: A constant pressure

60 filtration, Chem. Eng. Sci. 61 (2006) 3686–3697.
[8] J. Mendret, C. Guigui, P. Schmitz, C. Cabassud, In situ dynamic characterization
of fouling under different pressure conditions during dead-end filtration:
40 compressibility properties of particle cakes, J. Membr. Sci. 333 (2009) 20–29.
[9] E. Iritani, N. Katagiri, R. Nakajima, K.J. Hwang, T.W. Cheng, Nanocolloid cake
Nanosilica sol properties determined from step-up pressure filtration with single-stage
Step-up reduction in filtration area, AIChE J. 61 (2015) 4426–4436.
20 p1 = 49 kPa, p2 = 490 kPa [10] H.P. Grace, Resistance and compressibility of filter cakes, Part I, Chem. Eng.
Step-down Prog. 49 (1953) 303–318.
p1 = 490 kPa, p2 = 49 kPa
[11] M. Shirato, T. Aragaki, R. Mori, K. Sawamoto, Predictions of constant pressure
0 and constant rate filtrations based upon an approximate correction for side
0 0.05 0.10 0.15 wall friction in compression permeability cell data, J. Chem. Eng. Jpn. 1 (1968)
svt (cm) 86–90.
[12] F.M. Tiller, S. Haynes, W.M. Lu, The role of porosity in filtration VII Effect of
side-wall friction in compression-permeability cells, AIChE J. 18 (1972) 13–20.
10-2 [13] F.M. Tiller, W.M. Lu, The role of porosity in filtration VIII. Cake nonuniformity
(b) in compression-permeability cells, AIChE J. 18 (1972) 569–572.
[14] E. Iritani, N. Katagiri, K. Yamaguchi, J.H. Cho, Compression-permeability
properties of compressed bed of superabsorbent hydrogel particles, Drying
Technol. 24 (2006) 1243–1249.
[15] K. Kawasaki, A. Matsuda, Y. Mizukawa, Compression characteristics of excess
10-3 activated sludges treated by freezing-and-thawing process, J. Chem. Eng. Jpn.
24 (1991) 743–748.
(s-1)

[16] E. Iritani, N. Katagiri, K. Aoki, M. Shimamoto, K.M. Yoo, Determination of

10-4
Nanosilica sol
Step-up
p1 = 49 kPa, p2 = 490 kPa
Step-down
p1 = 490 kPa, p2 = 49 kPa
10-5
0 0.05 0.10
svt (cm)
Fig. 7. Dependence of parameters in Eq. (7) on product of mass fraction of particles
in colloid and filtrate volume per unit membrane area when filtration pressure
changes stepwise in dead-end ultrafiltration of nanosilica sol: (a) cake reversibility,
(b) rate constant.
state and transient behavior of preformed cake was proposed in
this paper and well described the flux decline behaviors in both
step-up and step-down pressure membrane filtration experiments.
We believe that the results obtained in this investigation provide
important new insights into an understanding of the nature of
compressible filter cake which is crucially important in membrane
filtration controlled by cake formation mechanism.
Acknowledgements
This work has been partially supported by Grant-in-Aid for Sci-
entific Research from The Ministry of Education, Culture, Sports,
Science and Technology, Japan. The authors wish to acknowledge
with sincere gratitude the financial support leading to the publica-
tion of this article.
References
[1] F.M. Tiller, T. Cleveland, R. Lu, Pumping slurries forming highly compactible
cakes, Ind. Eng. Chem. Res. 38 (1999) 590–595.
[2] F.M. Tiller, H.R. Cooper, The role of porosity in filtration: IV. Constant pressure
filtration, AIChE J. 6 (1960) 595–601.
[3] F.M. Tiller, M. Shirato, The role of porosity in filtration: VI. New definition of
filtration resistance, AIChE J. 10 (1964) 61–67.
permeability characteristics from centrifugal flotation velocity of deformable
oil droplets in O/W emulsions, Sep. Purif. Technol. 58 (2007) 247–255.
[17] E. Iritani, N. Katagiri, K. Aoki, M. Shimamoto, J.H. Cho, Compression
characteristics of consolidated deposits formed in centrifugation of soft and
hard colloids, Kagaku Kogaku Ronbunshu 33 (2007) 553–560.
[18] M. Loginov, E. Vorobiev, N. Lebovka, O. Larue, Compression-permeability
characteristics of mineral sediments evaluated with analytical
photocentrifuge, Chem. Eng. Sci. 66 (2011) 1296–1305.
[19] M. Loginov, N. Lebovka, E. Vorobiev, Multistage centrifugation method for
0.15 determination of filtration and consolidation properties of mineral and
biological suspensions using the analytical photocentrifuge, Chem. Eng. Sci.
107 (2014) 277–289.
[20] P.S. Yen, D.J. Lee, Errors in bound water measurements using centrifugal
settling method, Water Res. 35 (2001) 4004–4009.
[21] S.G. Sveegaard, K. Keiding, M.L. Christensen, Compression and swelling of
activated sludge cakes during dewatering, Water Res. 46 (2012) 4999–5008.
[22] M.K. Jørgensen, K. Keiding, M.L. Christensen, On the reversibility of cake
buildup and compression in a membrane bioreactor, J. Membr. Sci. 455 (2014)
152–161.
[23] T. Murase, M. Iwata, M. Wakita, T. Adachi, N. Hayashi, M. Shirato, Expansion of
consolidated material after release of load, J. Chem. Eng. Jpn. 22 (1989) 195–
199.
[24] E. Iritani, N. Katagiri, K.M. Yoo, H. Hayashi, Consolidation and expansion of a
granular bed of superabsorbent hydrogels, AIChE J. 53 (2007) 129–137.
[25] S. Okamura, M. Shirato, Comparison between processes of filtration at
constant and altered pressures, Kagaku Kogaku 18 (1954) 358–363.
[26] B.F. Ruth, Studies in filtration. III. Derivation of general filtration equations,
Ind. Eng. Chem. 27 (1935) 708–723.
[27] B.F. Ruth, Correlating filtration theory with industrial practice, Ind. Eng. Chem.
38 (1946) 564–571.
[28] M. Shirato, T. Murase, E. Iritani, S. Nakatsuka, Experimental analysis of flux
decline mechanism of batch ultrafiltration (Filtration characteristics of gel
layer), Filtr. Sep. 28 (2) (1991) 104–109.
[29] H. Reihanian, C.R. Robertson, A.S. Michaels, Mechanisms of polarization and
fouling of ultrafiltration membranes by proteins, J. Membr. Sci. 16 (1983) 237–
258.
[30] E. Iritani, N. Katagiri, R. Nakajima, K.J. Hwang, T.W. Cheng, Cake properties of
nanocolloid evaluated by variable pressure filtration associated with reduction
in cake surface area, AIChE J. 60 (2014) 3869–3877.
[31] T. Murase, E. Iritani, J.H. Cho, S. Nakanomori, M. Shirato, Determination of
filtration characteristics due to sudden reduction in filtration area of filter cake
surface, J. Chem. Eng. Jpn. 20 (1987) 246–251.
[32] E. Iritani, N. Katagiri, Y. Ishikawa, D.Q. Cao, Cake formation and particle
rejection in microfiltration of binary mixtures of particles with two different
sizes, Sep. Purif. Technol. 123 (2014) 214–220.
[33] J. Zhao, C.H. Wang, D.J. Lee, C. Tien, Plastic deformation in cake consolidation, J.
Colloid Interface Sci. 261 (2003) 133–145.

Please cite this article in press as: E. Iritani et al., Compression and expansion properties of filter cake accompanied with step change in applied pressure in
membrane filtration, Separ. Purif. Technol. (2016), http://dx.doi.org/10.1016/j.seppur.2016.11.067