Fouling monitoring during microfiltration of humic acid by

streaming potential measurement

On-line Number 464
Nakamura Kazuho, Liu Hui-yuan, Orime Takako, Matsumoto Kanji
Yokohama National University Graduate School of Engineering, Department of Chemical Engineering,
79-5, Tokiwadai, Hodogaya-ku, Yokohama, 240-8501 Japan E-mail: k_naka@chemeng.bsk.ynu.ac.jp

ABSTRACT
Fouling by natural organic matter like humic substances is one of the major factors limiting the
application of microporous membrane for water production. The charge of the pore surface of membrane,
which can be characterized by the streaming potential(SP) measurement as apparent zeta potential ζobs,
has a large effect on its fouling performance. However the relationship between the fouling condition and
the apparent zeta potential is not clear. The objective of this study was to develop a fouling monitoring
technique during microfiltration of humic acid by streaming potential measurement. The apparent zeta
potential was measured continuously during the filtration. It was confirmed that the fouling by Humic
acid was occurred at the filtration surface in the filtration conditions. The apparent zeta potential
increased with the increase in filtration resistance during filtration. The apparent zeta potential of fouled
membrane was depended on the degree of the fouling however the absolute value of the apparent zeta
potential was not same with the zeta potential of Humic acid, –33mV, which was obtained by
electrophoresis method. And it was demonstrated that the apparent zeta potential is useful for the
confirmation of the effect of the cleaning of the fouled membrane by the cleaning experiments using pure
water and alkali detergent.

KEYWORDS
Microfiltration membrane, Streaming potential, Fouling

INTRODUCTION
The membrane fouling is one of the major factors limiting the application of microporous membrane. As
membrane filtration of natural organic matter(NOM), such as humic substance, for drinking water supply
is becoming increasingly widespread(T.Thorsen, 1999), the fouling mechanism of humic acid has been of
interest. Humic acid fouling mechanisms was reported by Wei, et al. from the view point of flux decline
in UF(Wei, et al., 2000) and in MF(Wei, et al., 2002). C.Combe et al., studied adsorptive fouling by
humic acid using various surface modified CA membranes(C.Combe et al., 1999). Seong-Hoon et al.
showed the effect of calcium ion on the humic acid fouling in NF membrane(Seong-hoon et al. 1998).
The charge of the pore surface of membrane, which can be characterized by the streaming
potential(SP) measurement as the apparent zeta potential, has a large effect on its fouling performance.
We have developed the continuous streaming potential measurement method(Hui-yuan et al. 2003) and
studied about the effect of the pore structure of membrane on the apparent zeta potential(K.Nakamura et
al. 2003). However the relationship between the fouling and the apparent zeta potential is not clear.
The objective of this work is to investigate the relationship between the fouling and the apparent zeta
potential in the filtration of humic acid and show the ability of the streaming potential measurement for
the fouling monitoring technique. The apparent zeta potential was measured continuously during filtration
by pulse pressurizing method. The fouling mechanism was studied using the two-stage filtration
apparatus. The effect of cleaning condition using pure water and alkali detergent for the fouled membrane
on the apparent zeta potential was studied.

1
EXPERIMENTAL ⑦
Cellulose acetate membrane(Fuji Film, FM
⑧
series) with pore sizes of 0.22 were used. These ⑨ ⑥ ②
membranes have asymmetric pore structure and
was normally used in the direction from the ⑦
top(shiny) surface to the bottom(dull) surface. ⑨ ⑩
③
⑧
1mM KCl solution filtered with a 0.5µm ④
membrane filter was used as the electrolyte for ⑪
streaming potential measurement. Fouling studies
⑫
were performed with humic acid(Aldrich
Chemical). 0.1g/L Humic acids solutions were ⑬ ⑤
prepared by dissolving 0.1g of the powdered ①
Humic acid in 100mL of 0.01N NaOH with
stirring for 3hrs and stirring for 24hrs after ⑭
neutralization with 1N HCl and diluting with
pure water to 1L and filtering with a 0.5µm Fig.1 Experimental apparatus for streaming
membrane filter before use. Feed humic acid potential measurement: 1:N2 gas cylinder, 2:KCl
solutions were prepared by diluting the 0.1g/L solution reservoir, 3:Pressure controller,
Humic acids with the 1mM KCl solution. The 4:Pressure holding tank, 5:Temperature control
zeta potential of the humic acid in the solution bath, 6:Membrane filter for particle removal,
was –33.6mV measured by a electrophoretic light 7:Pressure transducer, 8: Electrodes, 9:Digital
scattering spectrophotometer(Otsuka Electronics, multi meter, 10:Measuring membrane holder,
ELS-8000). The concentration of humic acid was 11:Conductivity flow cell, 12:Filtrate reservoir,
measured on a UV spectrophotometer(Shimadzu, 13:Electric balance, 14:Personal computer.
UV1200) at 260nm.
Fig.1 shows experiment apparatus. Two
membrane housings(25mm diameter) were
∆PTM [Pa]
98

connected in series. These membrane housings

had platinum electrodes to measure the electrical
potential difference generated across the
membrane. The permeate flux, trans-membrane
pressure, electrical potential difference across
membrane and conductivity of permeate solution
were measured with a electric balance(Shimadzu, 0 15 60 75
EB3300SW), pressure transducer(Kyowa,
PVL-5KB), digital multi meter(Advantest, time [s]
R6441) and conductivity meter(TOA, CM-40G),
respectively. The measuring data were recorded Fig.2 Pulse pattern of applied pressure.
by a personal computer. Temperature in feed tank
was kept at 298K. Applied pressure was controlled at the constant value of 98kPa or the pulse pattern
(0-98kPa) as shown in Fig.2.
A new membrane was set into the each membrane holders. After the measurements of the pure water
flux at 98kPa, the streaming potential and permeate flux with 1mM KCl solution, humic acid solution was
filtered. During the filtration the permeate flux, absorbances at 260nm of the permeate solution and the
streaming potential were measured. After the filtration the streaming potential with 1mM KCl solution
was measured again.
Membrane cleaning effects were examined by measuring the streaming potential and the permeate flux
with 1mM KCl solution. The fouled membrane was rinsed twice with 50ml pure water and then cleaned

2
with 50ml alkali detergent (0.1% Ultrasil 11) for 30minuts. The humic acid filtration was repeated 3times
with the cleaned membrane.

RESULT AND DISCUSSION 0.03

Streaming Potential Measurement R2 = 0.9994
0.02
Fig.3 shows a typical result for streaming
potential measurement. The relationships between 0.01

ETM 　[V]
the electrical potential difference, ETM, and the

ΔETM
0
pressure difference, PTM, showed good linearity,
which shows the validity of the measurement. The -0.01 ΔPTM
streaming potential(SP) was defined as the slope of
-0.02
the line,
∆ETM -0.03
Streaming Potential (SP) = . (1)
∆PTM 0 0.2 0.4 0.6
SP was determined linear regression for the plots of P TM 　[kgf/cm2]
which the coefficient of determination R2 is above
Fig.3 The relationship between PTM and
0.998.
ETM for the membrane of pore size
The observed zeta potential, ζobs, was estimated
0.22µm.
using Helmholtz- smoluchouski equation,
µ λ ∆ETM µλ
ζ obs = = SP (2)
ε 0 ε r ∆PTM ε 0 ε r 1.2 0.5
FM22

Rejection Cefficient [-]

where µ is the solution viscosity, λ is the solution 1 Two-stage filtration 0.4
conductivity, ε0 is the permittivity of vacuum, εr is Plus pressure mode
0.8
J/Jo [-]

the dielectric constant of the medium. C0: 5mg/L 0.3

0.6
Humic Acid Fouling 0.2
0.4
Fig.4 shows typical results for the flux and the 0.1
humic acid rejection coefficient based on E260, 0.2
E 0 0
Rejection coefficient = 1 − 260 filtrate , (3)
E260 feed 0 0.5 1 1.5
as functions of filtrate volume per unit membrane V/A [ｍ3/ｍ2]
area, V/A, during the two-stage serial filtration Fig.4 Filtrate flux and humic acid rejection
with 0.22µm membranes(FM22) of 5mg/L humic coefficients for the two-stage serial
acid solution in the pulse pressure mode. The flux filtration with 0.22µm membranes(FM22) of
was normalized by J0, the pure water flux 5mg/L humic acid solution of the pulse
evaluated before the filtration. Flux declined with pressure mode.
increase in filtrate volume and became about 0.5
at V/A of 1m. Rejection coefficient was lower than 0.1 throughout the filtration. Almost same rejection
coefficient was also obtained in the case of one-stage filtration with the same membrane. The color of the
filtration surface of the 1st stage membrane changed from white into brawn during the filtration while that
of the 2nd stage membrane remained white after the filtration. From these observations it was found that
the 0.22µm membrane retained a part of the humic acid, which might be humic acid aggregate, on its
filtration surface and passed the humic acid in dissolution state.
The place where the fouling take place was confirmed by the comparing the filtration resistances, that
reflects pressure drop, of the 1st and 2nd stage membranes,

3
 PTM ,Total PTM ,1 + PTM , 2 PTM ,1 PTM , 2
J= = = = (4)
µ RTotal µ ( R1 + R2 ) µ R1 µ R2
where R is filtration resistance and subscripts 1
and 2 mean 1st and 2nd stage membrane, 5
FM22

R 1/R 1(t=0) or R 2 /R2(t=0) [-]

respectively. Fig.5 shows the comparison of the 1st stage
st 4 Two-stage filtration membrane
filtration resistance between the 1 stage Pulse pressure mode
nd C : 5mg/L
membrane and 2 stage one. The filtration 3 0

resistance was normalized by the filtration

2 2nd stage
resistance measured with pure water before the
st membrane
filtration. The filtration resistance of 1 stage 1
membrane increased with increase in filtration
volume while that of 2nd stage membrane showed 0
almost 1 throughout the filtration. This shows that 0 0.5 1 1.5
the fouling took place at the filtration surface of 3 2
V/A [ｍ /ｍ ]
1st stage membrane and did not take place inside
of the pore. And it is expected that the fouling Fig.5 Comparison of filtration resistance
would be caused by the humic acid aggregates between 1st stage membrane and 2nd stage
and would not be caused by the dissolved humic membrane during the two-stage filtration
acid. with 0.22µm membranes(FM22) of 5mg/L
Fig.6 shows the ζobs estimated from SP humic acid solution of the pulse pressure
measured for both 1st and 2nd membranes during mode.
the filtration. The ζobs of the FM22 membrane in
clean condition showed negative value of 0
–20±2mV. ζobs of the 1st stage membrane FM22
1st stage
increased with increase in filtration volume from -5 Two-stage filtration membrane
Pulse pressure mode
the initial value of –18.4mV to –7mV while that
ζ obs　 [mV]

C : 5mg/L
of 2nd stage membrane showed almost constant -10 0

value of –20mV throughout the filtration. The

-15 2nd stage
tendencies of the change in ζobs during filtration
membrane
of the each membrane were almost the same as
-20
those in the filtration resistance as was shown in
Fig.5. It is seemed that the change in ζobs during -25
the filtration would reflect the change in the
0 0.5 1 1.5
fouling condition. It was expected that ζobs of
fouled membrane would approach to the zeta V/A [ｍ3/ｍ2]
potential of the humic acid of –33.4mV, but ζobs
changed to the opposite direction from the initial Fig.6 Change in ζobs during the two-stage
ζobs value of –18.4mV. The reason is not clear. filtration with 0.22µm membranes(FM22)
From these observation it was shown that the of 5mg/L humic acid solution in the pulse
change in ζobs during the filtration could indicate pressure mode.
the degree of the fouling but the meaning of the
absolute value of ζobs is not clear.

The Effect of feed concentration

Fig.7 shows the effect of the humic acid concentration on the filtration resistance as a function of the
filtrate volume per unit filtration area for the filtration of 1, 5 and 10 mg/L humic acid solutions by the

4
 5 0
Closed symbols : 1st stage membrane Cfeed [mg/L]
R 1/R 1(t=0) or R 2 /R2(t=0) [-]

Open symbols : 2nd stage membrane -5 ▲,△ : 1

4 ●,○ : 5
-10 ■,□ : 10

ζ obs [mV]
3 Cfeed [mg/L]
▲,△ : 1 -15
2 ●,○ : 5
■,□ : 10 -20
1
-25 Closed symbols : 1st stage membrane
Open symbols : 2nd stage membrane
0 -30
0 0.5 1 1.5 0 0.5 1 1.5
V/A [m3/m2]
V/A [m3/m2]

Fig.7 Effect of feed humic acid concentration on filtration resistance and ζobs during
the two-stage filtration with 0.22µm membranes(FM22) in the pulse pressure mode.

two-stage serial filtration with 0.22µm membranes(FM22) in the pulse pressure mode. The increase rate
of the filtration resistance of the 1st stage membrane increased with increasing humic acid concentration
while the filtration resistance of the 2nd membrane showed almost constant value of 1 throughout the
filtration. The rejection coefficient was smaller than 0.11 under these conditions(data are not shown).
From these observations it was considered that the fouling was only occurred at the filtration surface of
the 1st membrane regardless of feed humic acid concentration. As the curves of the filtration resistance
overlapped almost one line regardless of the feed concentration when they were plotted as a function of
the amount of filtered humic acid per unit filtration area, C0V/A, the filtration resistance would be
determined by the total amount of the filtered humic acid in this filtration condition. The increase rate of
ζobs of 1st stage increased with increasing humic acid concentration like as the change in the filtration
resistance. The scattering of the ζobs of the 2nd membrane at feed humic acid concentration of 10mg/L
might be caused by the increase in the experimental error of SP with decreasing PTM,2. It was seemed that
the change of ζobs roughly reflected the increase in the filtration resistance regardless of feed humic acid
concentration.

Membrane Cleaning
In order to clarify if the ζobs is applicable for the effect of membrane cleaning the ζobs and permeate flux
were measured with 1mM KCl solution before and after the fouling and cleaning experiments. Table 1
shows the change in ζobs estimated at the fouling and the cleaning steps.

Table 1 The change in ζobs and flux measured with 1mM KCl solution before and after fouling and
cleaning.
Clean condition Fouled condition After the rinsing After the cleaning with
Membrane
(before filtration) (after filtration) with pure water the alkali detergent
st
ζobs 1 stage -21.1 -10.1 -15.8 -19.1
[mV] 2nd stage -20.3 -21.0 -21.5 -20.7
3
Flux *10 [m/s] 1.89 0.88 1.40 1.94

The ζobs of the 1st stage membrane changed from –10.1mV of the fouled condition to –15.8mV by the
rinsing with pure water and then recovered to –19.1mV, which was almost the same value of the clean

5
membrane, after the cleaning with the alkali detergent. The other hand, the flux improved depending on
the each cleaning step and recovered to the almost same value of the clean membrane after the cleaning
with the alkali detergent. The ζobs of the 2nd stage membrane, which didn’t show the increase in the
filtration resistance during the humic acid filtration, was almost constant value of 21±1mV independently
of the fouling and the cleaning steps. It was demonstrated that the ζobs is a useful for the effect of
membrane cleaning.

Conclusion
The following conclusion can be drown from the present study:
(1) The fouling during the microfiltration of humic acid solution with 0.22µm membranes(FM22) was
caused by humic acid aggregates at the filtration surface of the membrane and the dissolved humic acid
did not cause the fouling.
(2) The change in ζobs during the filtration had a relation to the change in the filtration resistance and
could indicate the degree of the fouling. However the meaning of the absolute value of ζobs was not clear.
(3) It was demonstrated that the effect of cleaning condition for the fouled membrane could be
confirmed by the change in the apparent zeta potential.

REFERENCES
C.Combe, E.Molis, P.Lucas, R.Riley and M.M. Clark, “The effect of CA membrane properties on
adsorptive fouling by humic acid”, J. Membrane Sci. 154, 73-87(1999)
Hui-yuan Liu, Kazuho Nakamura, Kanji Matsumoto, ”Measurement of streaming potential for
microfiltration membrane by continuous pressure increasing method”, Maku(Membrane)
28,278-287(2003)
Kazuho Nakamura, Hui-yuan Liu and Kanji Matsumoto, “Effects of pore size and conductivity of
electrolyte solution on the measurement of streaming potential and zeta potential of microporous
membrane”, Maku(Membrane), in printing
T.Thorsen; “Membrane filtration of humic substances -State of the art-“, Wat. Sci. Tech. 40,
105-112(1999)
Seong-hoon Yoon, Chung-Hak Lee, Kyu-Jin Kim and Anthony G. Fane; “Effect of calcium ion on the
fouling of nanofilter by humic acid in drinking water production”, Wat. Res. 32, 218-2186(1998)
Wei Yuan and Andrew L. Zydney; “Humic acid fouling during ultrafiltration”, Envirn. Sci. Technol., 34,
5043-5050(2000)
Wei Yuan, Aleksandra Kocic and Andrew L.Zydney, “Analysis of humic acid fouling during
microfiltration using a pore blockage-cake filtration model”, J. Membrane Sci. 198, 51-62(2002)