laporan akhir projek penyelidikan - core.ac.uk · laporanakhirprojek penyelidikan jangka pendek...
Post on 11-Aug-2019
228 Views
Preview:
TRANSCRIPT
Laporan Akhir Projek PenyelidikanJangka Pendek
Modulated Nanofiltration Process for~
Pesticides Treatment
byProf. Abdul Latif Ahmad.,
Dr. Syamsul Rizal Abd. Shukor
IlI.11IUNIVERSITI SAINS MALAYSIA
LAPORAN AKHIR PROJEK PENYELIDIKAN JANGKAPENDEKFINAL REPORT OF SHORT TERM RESEARCHPROJECTSila kemukakan laporan akhir ini melalui Jawatankuasa Penyelidikan di PusatPengajian dan Dekan/PengarahlKetua Jabatan kepada Pejabat Pelantar Penyelidikan
4. Tajuk Projek:Title ofProject
Modulated Nanofiltration Process for Pesticides Treatment
i) Pencapaian objektif projek:Achievement ojproject objectives
ii) Kualiti output:Quality ojoutputs
iii) Kualiti impak:Quality ojimpacts
iv) Pemindahan teknologi/potensi pengkomersialan:Technology transjer/commercialization potential
v) Kualiti dan usahasama :Quality and intensity ojcollaboration
vi) Penilaian kepentingan secara keseluruhan:Overall assessment ojbenefits
DD
DD
DD
DD
DD
DD
D
D
DD
0D
DD
EJD
DD
DD
Laporan Akhir Projek Penyelidikan Jangka PendekFinal Report O/Short Term Research Project
7. Sila sediakan laporan teknikallengkap yang menerangkan keseluruhan projek ini.ISila gunakan kertas berasingan]Applicant are required to prepare a Comprehensive Technical Report explaning the project.(This report must be appended separately)Please refer to Appendix 2
Senaraikan kata kunci yang mencerminkan peny~lidikan anda:List the key words that reflects your research: •
Bahasa Malaysia
Membrane penurasan nanD
Racun perosak
Teknologi membran
Bahasa lnggeris
Nanofiltration membrane
Pesticides
Membrane technology
2
Laporan Akhir Projek Penyelidikan Jangka PendekFinal Report O/Short Term Research Project
9. Peralatan yang Telah DibeIi:Equipment that has been purchased
I. HPLC column
2. Membrane test cell
;,
Tandatangan PenyelidikSignature ofResearcher
3
~~"/~""1--Tarikh
Date
Komen Jawatankuasa Penyelidikan Pusat Pengajian/PusatComments by the Research Committees ofSchools/Centres
Lllporan Akhir Projek Penyelidikan Jangka PendekFinal Report OfShort Term Research Project
TANDATANGAN PENGERUSIJAWATANKUASA PENYELIDIKAN
PUSAT PENGAJIAN/PUSATSignature ofChairman
[Research Committee ofSchool/CentreJ
4
I,
TarikhDate
APPENDIXl
Abstract of Research
•
,.
I,
Abstract
In Malaysia, little attention has been given to the presence of pesticides in the
water source and its adverse effects on human health. These huge amount of pesticides
used, especially in agriculture practice, are the emerging contaminants in drinking water
supplies. This is because pesticides applied directly to the soil can be washed off by rain
into nearby bodies of surface water or percolate through the soil to lower soil layers and
groundwater. This project focuses on the removal of pesticide from aqueous solution
using nanofiltration membrane. Two pesticides, atrazine and dimethoate, were selected
for study in this research. Four nanofiltration membranes (NF90, NF200, NF270 and DK)
were subjected to a stirred dead-end filtration of the pesticide solution. It was found that
NF90 showed the best rejection performance, followed by NF200 and DK. Meanwhile,
although NF270 showed the highest permeate flux out of the four membranes tested, it
showed the poorest rejection. In overall, for the four membranes tested, atrazine was
consistently better rejected than dimethoate.
\,
Abstrak
Di Malaysia, terlalu sedikit perhatian telah diberikan kepada kehadiran racun
serangga dalam sumber air dan kesan buruknya terhadap kesihatan manusia. Amaun
besar racun serangga yang digunakan, terutamanya dalam bidang pertanian, telah menjadi
bahan pencemar yang muncul dalam bekalan air minuman. Ini adalah kerana racun
serangga yang disembur ke tanah boleh dialirkan oleh air hujan sumber air yang
berdekatan atau meresap melalui tanah ke lapisan tanah yang lebih dalam. Projek ini
memfokuskan kepada penyingkiran racun serangga dari larutan akues menggunakan
membran penurasan nano. Dua jenis racun serangga, atrazin dan dimetoat, telah dipilih
untuk diuji. Empat jenis membran penurasan nano (NF90, NF200, NF270 and DK) telah
diuji dengan menjalankan penurasan hujung mati teraduk menggunakan larutan racun
serangga. Melalui kajian ini, ia didapati bahawa NF90 menunjukkan prestasi penolakan
yang terbaik, diikuti dengan NF200 dan DK. Sementara itu, walaupun NF270
menunjukkan prestasi hasil telapan yang tertinggi di antara empat membran yang diuji, ia~
menunjukkan prestasi penolakan yang p~ling lemah. Pada keseluruhannya, keempat-
empat membran yang diuji dapat menolak atrazin dengan lebih baik daripada dimethoate
secara konsisten.
APPENDIX 2
Comprehensive Technical Report
•
I,
Comprehensive Technical Report
In overall, the project has successfully achieved its objectives in application of different
type ofnanofiltration membranes in separation ofpesticides from water. The following sections
report on the results obtained from this research.
Type of pesticides used in Malaysia
Malaysia is actively involved in agriculture practice, planting oil palm, paddy, fruit,
vegetables and others for both local consumption and export purposes. In order to achieve the
objectives such as to maintain the quantity and quality of agriculture productions, pesticides are
used in agriculture sector as a mean ofpest control for sustainability ofthe industry.
In Malaysia, the annual pesticides sales figure exceeds RM 300 million. It is estimated
that annual crop losses in our country could exceed 30% without pesticides (MCPA, 2005).
Approximately 70% ofpesticides sold goes to the plantations. The commonly used pesticides in
the plantations are herbicides. Glyphosate and atrazine are among the most commonly used
herbicides. Meanwhile, insecticides accounts for approximately 16% of pesticides sold in our
country. The rest of the pesticides sold are~fungicides and rodenticides (Chooi, 2005). The
commonly used insecticides are organophosphorus pesticides which are widely used to replace
the organochlorine pesticides due to its short half-life.
Comparison between Membranes
In order to explore the application of different type of nanofiltration membranes in separation of
pesticides from water, four nanofiltration membranes, namely, NF90, NF200, NF270 and DK
were used as subject of study. The study the adsorption process of pesticides on membrane and
its effect on rejection of pestici~es could not be carried out as the analytical equipment available
was not sensitive enough to detect the very minute changes in concentration of pesticides. This
matter has been reported in the 6th month progress report previously.
Plate 1 and Figure 1 show the dead end filtration rig used in this project and its
schematic diagram respectively. The operatihg pressure for the filtration was supplied by
pressurized nitrogen gas. The nitrog~n gas outlet pressure was regulated using a single stage
pressure regulator of Concoa brand whereby the equilibrium pressure was shown on the
pressure gauge.
Plate I: Dead end filtration rig.
StainlessSteel~edTank
Legend:o :Pressure regulator
[:)i<] : Valve
Figure I: Diagram of experimental set-up.
\,
MembraneStirred Cell
Magnetic Stirrer
Permeate
AnalyticalBalance
A pressure relieve valve was installed between the nitrogen gas and the reservoir to
relief the pressure build up in the reservoir. The pressurized reservoir forced the water or
solution in the solution chamber to flow into the membrane cell. All pipelines, valves and
reservoir were made of stainless steel material. The reservoir could hold volume up to 1.8 Litre.
Filtration experiments were carried out under stirred dead-end filtration. The cell was
pressurized with compressed high purity nitrogen gas. The solution in the cell was stirred by a
Teflon-coated magnetic bar using Heidolph MR3000D (Germany). The membrane was
immersed for 24 hours in deionized water before being used in any experimental work. In order
to determine the pure water permeability, the membrane was first filtered using deionized water
at 15 x 105 Pa for a minimum of seven hours for compaction. This was done to avoid
compression effect in the later stage of experiment.
Pure water permeability for all four membranes was first determined to obtain the water
flux of the membranes without the presence ofpesticides. After that, the removal of atrazine and
dimethoate from aqueous solution were examined for the four nanofiltration membranes.
For pesticide flitration process, the same compaction process was carried out before the
test cell was emptied and 1.8 Litre of feedsolution was filled into test cell and solution reservoir.
Feed solution for each experiment was prepared by weighing specific amounts of pesticides
before dissolving them in approximately 5 mL of methanol. After that, the solution was added
into deionized water and shaken well. The cell filled with pesticides solution was then
pressurized at operating pressure indicated by a pressure regulator.
.!
The duration of the filtration experiments was approximately 80 minutes for each run.
Permeate from the bottom ctll was collected every 20 min, whereby the cumulative weight was
continuously measured with an ~ectronic balance with an accuracy of ± 0.01 g. The cumulative
weight were converted to cumulative volume and the value of permeate flux were calculated
from there on.
Meanwhile, the permeate concentrations which contained pesticides were measured by
using HPLC (Perkin Elmer Series 200) at wavelength of 200nm. After each run, the cell and the
membrane were washed thoroughly with deionized water. The membrane permeability was
checked and it was observed that the permeability varied within ±2% of the initial measured
value. All experiments were conducted at room temperature (28 ± 2°C).
Pure Water Permeability
Figure 2 shows the pure water permeability graph for the NF90, NF200, NF270 and DK
membranes. Pure water permeability for the membranes was determined by obtaining the flux
for deionized water for each membrane at different operating pressure. As shown in the figure,
pure water flux for all membranes increased linearly with the increasing applied pressure. All
the curves passed through origin in accordance to the null value of the operating pressure. The
pure water permeability for these membranes was calculated from the slope of the water flux
against the operating pressure.
5.00E-05
4.50E-05
4.00E-05
3.50E-05
Ul 3.00E-05N'
E.... 2.50E-05g><:::J 2.00E-05ii:
1.50E-05
1.00E-05
5.00E-06
O.OOE+OO
0 2 4 6 8 10 12 14
• NF90
• NF200
.. NF270
xDK
16
Figure 2: Pure water permeabilip'.
Pressure x 105 (Pa)
NF270 had the highest water permeability under all pressure conditions among the four
membranes studied, followed by NF90 and NF200 while the DK had the lowest water
permeability compared to other membranes. The pure water permeability value for NF270,
NF90, NF200 and DK were 3.46 x 10-11 m3/(p12.s.Pa), 2.36 x 10-11 m3/(m2.s.Pa), 1.86 x 10-11
,m3/(m2.s.Pa) and 1.12 x 10-11 m3/(m2.s.Pa) respectively.
The finding that NF270 had the highest permeability corresponded to results from Hilal
et at. (2005) that NF270 had the largest average pore size, which was 0.71 nm, followed by
NF90 with 0.55 nm of average pore size while NF200 had average pore size of 0.38 nm
(Lefebvre et al., 2003). However, DK showed the lowest water permeability although it had
similar average pore size with NF90 (Santos et al., 2006). This suggests that membrane
formulation produced by DowlFilmtec (USA) for NF90, NF200 and NF270 produces high
permeate flux while different membrane formulation by Osmonic (USA) for DK produces low
permeate flux. Formulation of the membrane is very important because water flux through a
membrane is greatly dependent on its ability to form hydrogen bonds with the hydrophilic
groups of the membrane polymer (Williams et al., 1999).
Pesticide Solution Performance
In order to monitor the performance of membranes as a function of time, deionized
water with pesticide of 10 mglL was used as feed. The experiments were conducted at operating
pressure of 6 x 105 Pa and the stirring rate was set at 1000 rpm. This operating pressure was
chosen as to strive for low operating cost by using low operating pressure. High stirring rate was
chosen in order to create high turbulence to the filtration system.
The rejection performance was monitored at every 20 minutes over a period of 120
minutes. This was done to identify the stabl: trend of the rejection as Braeken et al. (2005)
observed that retention would decrease itrongly during the first 15 to 30 minutes and then it
would reach a stable value. The rejection 'performance for NF90, NF200, NF270 and DK is
shown in Figure 3 for atrazine while rejection performance for dimethoate is shown in Figure 4.
I,
.... -._ ..... _._ ..... _.- ..... -._ ..... _._ ..
..........................................)( )<----.)( ~( )( --)(.... - ....... - ........ _.......... - ......... _.. -.
100
90
80
70
c 600;JUQl 50'iii'0:::~ 40
30
20
10
0
0 20 40 60 80 100 120 140
-·.·-NF90...•... NF200
-'4-' NF270
~DK
Time (min)
Figure 3: Comparison between membranes on atrazine rejection with time at
operating pressure of6 x 105 Pa and stirring rate of 1000 rpm. Feed pesticide.concentration was 10 mg/L,
.... - ....... - ...... - ....... - ....... -........•.....•.....•.....•....
J.: . j( )( - )( --.:...~. )( ~.... - ..- '"-. ., -''- .. - ...... - ',- .. - ..
100
90
80
70c
600;JUQl 50'iii'
0::: 40~
30
20
10
00 20 40 60 80
;
Time (min)'
100 120 140
- +-. NF90..•.. NF200_ .... - NF270~DK
Figure 4: Comparison between membranes on dimethoate rejection with time at
operating pressure of 6 x 105 Pa and stirring rate of 1000 rpm. Feed pesticide
concentration was 10 mg/L.
-
It could be observed from Figure 3 and Figure 4 that all membranes tested showed a
stable trend of rejection for atrazine and dimethoate with minor fluctuation. Therefore, as to
counter the possibility of the minor fluctuation of rejection, the rejection value obtained from
next section onwards would be an average of the rejection performance over a period of 80
minutes.
NF90 produced the best rejection performance for both pesticides tested, which was
more than 95% rejection for atrazine and approximately 85% for dimethoate. The performance
ofNF200 was the second highest of all four membranes tested while DK showed slightly lower
rejection than NF200. NF270 showed the lowest rejection performance out of the four
membranes tested. This corresponded to the finding that NF270 had the largest average pore
size, which was 0.71 nm (Hilal et al., 2005), and this caused more pesticides to be able to pass
through the membrane.
In overall, all four membranes tested showed better rejection for atrazine than
dimethoate. This observation was obtained despite the fact that dimethoate has slightly higher
molecular weight than atrazine (229.28 g/mol for dimethoate and 215.69 g/mol for atrazine).
Kiso et al. (2001) and Bellona et al. (2004) suggested that although molecular sieving effect
played important role in determining the retention performance by membrane, hydrophobicity.of the solutes must not be neglected. The vl:U!1e of log octanol/water partition coefficient, Ko/w, is
usually used to gauge the hydrophobicity of a particular organic solute. The higher the value of
log Ko/w, the better the rejection would be. This behaviour can be observed in this study since
atrazine has higher hydrophobicity than dimethoate. The value for log Ko/w for atrazine and
dimethoate are 2.34 and 0.70, respectively (Kamrin, 1997). However, it must be noted that the
influence of hydrophobicity of the solutes on rejection would decrease as the molecular size of
the solutes increased (Braeken et al., 2005). This is because molecular sieving effect would beI
very much prominent when ·the molecular size of the solutes was much bigger compared to the
pore size of the membrane. "'f
The dipole moment of the organic solutes also affected the rejection performance of the
membrane. Dimethoate has dipole moment of 5.164 debye while atrazine has dipole moment of
1.763 debye (Kim, 2006). Van der Bruggen et al. (2001) described that molecules with larger
dipole moment approaching the charged membrane were, by electrostatic attraction, properly
oriented towards the pores and entered more easily into the membrane structure. Consequently,
they were less rejected. This situation contributed to the lower rejection of dimethoate as the
large dipole moment of its molecules facilitated its passing through the membrane.
In addition, dimethoate has aliphatic molecular structure compared to the heterocyclic
aromatic structure of atrazine. Van Gauwbergen and Baeyens (1998) reported that branching of
the molecular structure would improve rejection. Meanwhile, according to Kiso et al. (2000),
non-phenylic structured pesticides were rejected at a relatively lower degree than phenylic
structured pesticides. Hence, although atrazine has slightly lower weight than dimethoate, it was
better rejected compared to dimethoate.
The variation of the permeate flux by the four nanofiltration membranes tested for
filtration ofatrazine and dimethoate is presented in Figure 5 and Figure 6. The filtration process
was started after a minimum of seven hours of membrane compaction using pure deionized
water. A slight decrease of flux with the increase oftime could be observed in both figures. This
could be attributed to slight fouling due to the adsorption of organic pesticide molecules on the
membrane (Williams et al., 1999; Kimura et al., 2003). Therefore, in order to avoid
overestimation of flux due to the adsorption effect, the permeate flux value obtained from next
section onwards would also be an average of flux performance over a period of80 minutes.
Jr- •• _ -Jr- •• _._...... _ .'"-. -- •• - -Jr- • - - -&
+--_.-.-.-.""""'---....--_.-.-.- ...... -_ ....-_ ...•.....•.....•. __ ..•.....•
2.50E-05
2.00E-05
Uf 1.50E-05N
...~E......)( 1.00E-05~
ii:
5.00E-06
O.OOE+OO
0 20
)(
40
)(
60
)(
80
)(
100
)(
120 140
- ...... NF90••••. NF200
- -&- - NF270
~DK
Time (min)
Figure 2: Comparison between membranes on permeate flux during atrazine
rejection with time at operating pressure of6 x 105 Pa and stirring rate of 1000 rpm.
Feed pesticide concentration was 10 mgIL.,
+- . - ....... - ....... - . --.....~--._ ....... _..........•.....•.....•.....•.....•
.- .. - ..- .. - ..- .. - ..- .. - ..- .. - ..2.50E-05
2.00E-05
-II! 1.50E-05N
C",-
.§.>< 1.00E-05~
u::
5.00E-06
O.OOE+OO
0
)(
20
)(
40
)(
60
)(
80
)(
100
)(
120 140
- +-. NF90••••. NF200
_ •• - NF270
~DK
ft, . ~
Time (min)
Figure 3:Comparison between membranes on permeate flux during dimethoate
rejection with time at operating pressure of6 x 105 Pa and stirring rate of 1000 rpm.
Feed concentration was 10 mglL.
This observation was also in agr~ement with Chang et al. (2002) that retention of
organic molecules in nanofiltration membranes was not only caused by molecular sieving, but
also by electrostatic effects and by adsorption on the membrane surface. Thus, convective
transport through the membrane pores and diffusive transport of adsorbed compounds through
the membrane matrix can be responsible for solute permeation.
In overall, the permeate flux for all membranes is still in agreement with the earlier
study for the pure water pefmeability whereby NF270 had the highest flux followed by NF90
and NF200 while DK had the lpwest permeate flux. This also showed that 0.55nm of average
pore size for NF90 was sufficient to retain dimethoate and atrazine with high percentage of
rejection. Nevertheless, solute-membrane interaction factor was also important (Kosutic at aI.,
2005; Kim et al., 2005) as DK and NF200 could not sustain as much rejection as NF90 although
DK had similar average pore size with NF90 (Santos et al., 2006) while NF200 had smaller
average pore size. It is believed that the intl\raction between pesticides tested and material
formulation for DK and NF200 contributed to the crossing of solutes; just as how interaction
between different membrane formulations and water could contribute to different water flux
performances as mentioned earlier. However, the detailed structural interaction ofthe membrane
and solutes was not studied in this research.
Effect of Water Quality
The effect of different types of water quality on rejection of pesticides and permeate
flux is investigated in this section. Four types of water were used as feed water in experiment.
They were deionized water, distillated water, tap water and river water which were spiked with
pesticide. The performance of the nanofiltration membranes when filtering river water were
gauged against the performance of the membranes when filtering the three other water matrices.
Fixed parameters including pesticide concentration of 10 mg/L, operating pressure of 6 x 105 Pa
and stirring rate of 1000 rpm were chosen for the experiment. This experiment was conducted in
view of the possibility of water treatment from river water.
Water Composition
The result of water composition analysis is given in Table 1. Deionized water was used
as a reference to the other type of water matrices in this experiment. The deionized water and
distilled water was obtained from the Chemical Laboratory while the tap water was obtained
from Research Laboratory 1 at School of Chemical Engineering, Universiti Sains Malaysia. The
source of river water was from the Kerian River at Nibong Tebal, Pulau Pinang. The river water
had been pre-filtered with a regenerated cenulose (RC)-membrane filters with pore size of.0.45/-lm prior to analysis and preparation as' feed solution.
Table 1: Composition offeed water.
Distilled waterParameter Tap water (mg/L) River water (mgIL)
(mgIL)
COD Not detected Not detected 20.3
Aluminum, (AI) 0.013 0.075 0.011
Barium, (Ba) 12 15 21
Calcium, (Ca).,
0.11 2.94 1.22
Chloride, (Cn 0.4 7.6 4.5
Chromium, (Cr) 0.006 0.015 0.020
Copper, (Cu) 0 4 3
Magnesium, (Mg) 0.01 2.24 3.20;,
Nitrate, (N03-) 0.1 0.2 0.3
Sulfate, (SOl-) 1 18 12
Zinc, (Zn) 0.04 0.05 0.10
Lead, (Pb) 0.001 0.005 0.009
Effect of Water Quality on Pesticide Rejection
Rejection performance of atrazine and dimethoate by NF90, NF200, NF270 and DK at
different types of water quality is presented in Figure 7 and Figure 8 respectively. It can be seen
from these figures that the rejection trend of the atrazine and dimethoate was generally higher in
the tap water and the river water than the rejection from distilled water and deionized water. As
for the performance of each individual membrane, NF90 showed the highest rejection followed
by NF200, DK and NF270 for both atrazine and dimethoate in all types of water quality
involved.
100
90..
80
70c 600
:;:lu.!!!, 50Gl
D::40
~
30
20
10
0NF90 NF200 NF270 OK
D Deionized water
[J] Distilled water
8 Tap water
mRiver water
Membrane
Figure 4: Effect ofwater quality on rejection of atrazine at operating pressure of 6 x
105 Pa and stirring rate of 1000 rpm. Feed concentration was 10 mgIL.
1
In the case of rejection of atrazine, the rejection from tap water and river water was
slightly higher than rejection f~m distilled water and deionized water for NF90. Similar trend
was observed for NF200, NF270 and DK. However, they experienced a more obvious increase
of rejection i.e. approximately 10% increment from the lowest to the highest atrazine rejection
performance for each individual membrane.
\,
100
90
80
70c
60..
0:;:l(.)Gl 50'Q)'
0::40';je.
30
20
10
0NF90 NF200 NF270 DK
o Deionized water
IDJ Distilled water
STap water
~ River water
Membrane
Figure 5:Effect ofwater quality on rejection of dimethoate at operating pressure of 6 x
105 Pa and stirring rate of 1000 rpm. Feed concentration was 10 mg/L.
Meanwhile, for the rejection of dimethoate, all four membranes showed apparently
higher rejection in tap water and river water compared to deionized water and distilled water.
The four nanofiltration membranes showed an increase of dimethoate rejection between 7 to..14% for the two groups of water quality (i.e. deionized water with distilled water and tap water
with river water).
The observation of higher rejection occured for river water compared to the deionized
water and distillated water can be explained by the influence of natural organic matter in the
source of water. Naturally, the concentration of natural organic matter was higher in the river
water than in the deionized lYater or distilled water as the latter received further water treatments.
Pesticide could associate with the functional group present on the natural organic matter and..,.form macromolecular complex (Agbekodo et al., 1996). This phenomenon enhanced the effect
of molecular sieving. In addition, there was adsorption of pesticide onto the outer surface or
inside pores of membrane due to the hydrophobicity of that natural organic matter itself (Zhang
et al., 2004). The complex that formed from association of pesticide and natural organic matter
could also compete with the pesticide, result~d in steric congestion whereby more pesticide
being retained during the transportation through the membranes (Agbekodo et al., 1996).
Interestingly, the rejection of pesticide in tap water observed in Figure 7 and Figure 8
was comparable to those in river water, although tap water had undetectable natural organic
matter content. Therefore, it is believed that besides the influence of natural organic matter, the
effect of ion adsorption also played a significant role in nanofiltration process. The high
concentration of divalent ion such as Ca2+, Mg2+, and SO/- and other monovalent ion in the tap
water and river water, as shown in Table 4.2, also contributed to higher rejection ofthe pesticide.
Ion adsorption could happen due to the interaction between the ion and the membrane (Schaep
et al., 1998; Thanuttamavong et al., 2002). This phenomenon narrowed the membranes pores
and thus, decreased the transportation of pesticide through the membrane. The pore structure of
membranes would be so tight that together with pesticide molecules, the major part of ions
would be removed (Zhang et al., 2004).
Effect of Water Quality on Permeate Flux
Figure 9 and Figure 10 show the flux performance for the rejection of both atrazine and
dimethoate from the different types of water quality tested. Generally, the permeate flux
declined from the deionized water to the river water for all type ofmembranes studied. It can be
observed that while both rejections atrazine and dimethoate in tap water and river water were
higher than rejections in deionized water and distilled water, as shown in Figure 7 and Figure 8,
the flux for both tap water and river water was consistently lower than the flux for deionized
water and distillated water. Based on Figure '9 and Figure 10, it can be observed that the flux
decline was up to 19% for NF90 and NF2aq, 24% for NF270 and 18% for DK for both atrazine
and dimethoate rejection involved.
2.50E-05
2.00E-05
--III 1.50E-05'",§..§.>< 1.00E-05:::Iu:
5.00E-06
O.OOE+OO
NF90 NF200 NF270 DK
(] Deionized water
IIlI Distilled water
EI Tap water
~ River water
Membrane
Figure 6: Effect of water quality on flux performance during rejection ofatrazine at
operating pressure of 6 x 105 Pa and stirring rate of 1000 rpm. Feed concentration
was 10 mglL.
2.50E-05
2.00E-05
Ul 1.50E-05",'
E;;-§.>< 1.00E-05::Ju:
5.00E-06
O.OOE+OO
NF90 NF200 NF270 DK
o Deionized water
IT] Distilled water
8 Tap water
l.'lI River water
Membrane
Figure 7:Effect ofwater quality on flux performance during rejection ofdimethoate at
operating pressure of 6 x 105 Pa and stirring rate of 1000 rpm. Feed concentration was
10 mg/L.
The addition of pesticides to water was not the cause of flux decline as the relative.water flux in water with the addition of .pesticides only produced minor decline of flux
performance. Therefore, the major influence in the flux decline between the different types of
water quality could be attributed to the concentration of ions and natural organic matter in the
water. Adsorption of ions inside the membrane pores caused a decrease of the effective pore
size and consequently decreased the water flux (Zhang et al., 2004).
Besides, pore blocking happened when molecules with a size that corresponded with thet
size of an important fraction'ofthe pores blocked the membrane pores and caused fouling (Van
der Bruggen et al., 1998). This;rwould cause a decrease in the number of both pesticide and
water molecules passing through the membrane. Consequently, the water flux decreased, but the
rejection was increased. Hence, this fmding is in agreement with the increased rejection of
atrazine and dimethoate described in earlier section.
i,
Parameter Study for Nanofiltration
In order to understand the effect of operating conditions to the rejection and permeate
flux, operating parameters such as operating pressure, pesticide concentration and stirring rate
were varied one at a time to investigate its impact to the performance of the nanofiltration
membranes. The effect of pH of solution and binary solute mixture were also investigated as to
observe its influence to the rejection and flux performance should such conditions occur.
Effect of Operating Pressure
The effect of operating pressure on pesticide rejection and permeate flux were
investigated at various operating pressure (5 x 105 Pa to IS x 105 Pa), A fixed parameter
including pesticide concentration of 10 mglL and stirring rate of 1000 rpm were chosen for
these experiments.
Effect of Operating Pressure on Pesticide Rejection
Rejection performance of atrazine and dimethoate by NF90, NF200, NF270 and DK at
different operating pressure is presented in Figure 11 and Figure 12 respectively. From these
figures, it can be observed that the rejection ofboth atrazine and dimethoate inclined to be better
when the operating pressure was increased.
............ - ...... -: ..... _- ..... - ...
- +-. NF90- - •. 'NF200- -J. - NF270~DK
1715
••
13119
<Of..
7
_·A_ -J.- •. - -J.-"·r· .
~.-" ok' .
5
•.... -.. ,
••••••
100
90
80
c0 70~uQ)
'iii'0::: 60~
50
40
303
Pressure x 105 (Pa)
Figure 8: Effect of operating pressure on rejection ofatrazine at feed concentration
of 10 mglL and stirring rate 1000 rpm.
Transport of solute through nanofiltration membrane can be explained in terms of
diffusion and convection (Hilal et al., 2004). Higher rejection was observed at higher pressure
due to the increased water flux. The concentration of permeate became diluted with the
increased water flux as the solute molecule was rejected by molecular sieving effect.
-+-.-.-+-._--...... -.-+-.--......... -
..--...•. -....•... --. -.. ... . . -. ....•..•.. _.ok- ..
. -r··-_.ok--.... _..k" ..
100
90
80c0
70:;:;uC1)
'Qj'0::: 60';f.
50
40
303 5 7 9 11 13 15 17
- +- - NF90. -. -. NF200- .. - NF270~DK
Pressure x 105 (Pa)
Figure 9: Effect of operating pressure on rejection of dimethoate at feed
concentration of 10 mgIL and stirring rate 1000 rpm.
In other words, at higher pressures, "the flux of water increased, but the transport of.solute did not increase since the driving force of solute was not pressure but concentration
dependent. This caused the actual solute concentration in the permeate to be lower and therefore
the rejection was higher (Krieg et al., 2004).
However, the rejection of atrazine by NF90 increased for only approximately 2% from
between the lowest and highest operating pressure applied, suggesting that the contribution of
diffusion for atrazine throu&h NF90 was still relevant even at high operating pressure.
<If
Effect of Operating Pressure on Permeate Flux
Figure 13 and Figure 14 show flux performance of the membranes tested during
atrazine and dimethoate rejection respectively. Based on these figures, it was obvious that the
increase in pressure had positive effect on permeate flux for both atrazine and dimethoate;
filtration. 't
....."
-- .-- NF90
----.---- NF200
---Il--- NF270
--*-DK
171513119
..... .---................
75
5.00E-05
4.50E-05
4.00E-05
3.50E-05
Uf 3.00E-05N
E-.. 2.50E-05.§.><:::l 2.00E-05
u:::
1.50E-05
1.00E-05
5.00E-06
O.OOE+OO
3
Pressure x 105 (Pa)
Figure 1O:Effect ofoperating pressure on permeate flux during rejection ofatrazine at
feed concentration of 10 mglL and stirring rate 1000 rpm.
5.00E-05
4.50E-05• .A...
It'4.00E-05 ",
K3.50E-05
",
..... K .....3.00E-05 .-II! ......- .. - +- - NF90N ...
E ",..- ...... - -. -. NF200E 2.50E-05 'if. ¥ ... .. '
........It' ...... _ •• - NF270
>< •••:::l 2.00E-05 It" ... ---*"-DKu::~ •••
1.50E-05 K... ... ••1.00E-05 ••5.00E-06 I,
O.OOE+OO
3 5 7 9 11 13 15 17
Pressure x 105 (Pa)
Figure 11: Effect ofoperating pressure on permeate flux during rejection of
dimethoate at feed concentration of 10 mg/L and stirring rate 1000 rpm.
All membranes tested experienced a steady increment of permeate flux with the
increase of pressure applied to the system. While the increase of permeate flux against pressure
formed a perfectly straight correlation in pure water permeability in Figure 1, the deviation of
permeate flux from the direct correlation could be attributed to the adsorption of pesticide
molecules on the membrane (Kimura et al., 2003).
In an osmotic pressure controlled nanofiltration system, the flux can simply be related
to the pressure as follows,
(4.1)
where Vw is permeate flux (m3/m2.s), Lp is the membrane permeability (m3/(m2.s.Pa), Mis
the pressure differential (Pa), a is the reflection coefficient (dimensionless) and !:J.1C is the
osmotic pressure differential (Pa) (Spiegler and Kedem, 1966). For a fixed solute concentration,
the effective driving force for the solvent transport would be higher with an increment in
operating pressure (Syamal et al., 1997). Furthermore, the increase in driving force would
overcome the membrane hydraulic resistance. Therefore, increasing the pressure would force
more water to pass through the membrane and resulted in a higher permeate flux.
.Effect of Feed Pesticide Concentratio~
The effect of feed pesticide concentration on pesticide rejection and permeate flux were
investigated at various feed pesticide concentration (2 to 22 mgIL). A fixed parameter including
operating pressure of 6 x lOS Pa and stirring rate of 1000 rpm were chosen for these experiments.
Effect of Feed Pesticide Concentration on Pesticide RejectionI .
Rejection performance of atrazine and dimethoate by NF90, NF200, NF270 and DK at
different pesticide concentratio1;' is presented in Figure 15 and Figure 16. It can be seen from the
figures that there is a small decrement on pesticide rejection for all membranes tested with an
increase in feed pesticide concentration. This is because the effect ofpesticides concentration on
pesticides rejection is less pronounced with dilute feed concentration. This finding is in
agreement with work by Zhang et al. (2004) and Causserand et al. (2005). This shows that in;
practical terms, the membranes have almost tIle same efficiency level for pesticide rejection in
water even though the feed concentration varies from time to time.
100
90
80
c0 70:;:;uCIl'0)0::: 60~
50
+-._.- ..... _.- .... . -.- ............ ..... _.- ..
•.......•.......•.......•* .......•.......•)( ------)(.. )( )(.. -....... -....... - ~
" ..... - ....." -...
- +-. NF90
" •• ,NF200
_ ... - NF270
--*-DK
40
8 10 12 14 16 18 20 22 24
Concentration (mg/L)
642
30 +-----,,---,----,---,-----,---,----,--r---r--.--,---,
o
Figure 12: Effect ofpesticide concentration on rejection ofatrazine at operating
pressure of6 x 105 Pa and stirring r.ate 1000 rpm,
100
90
80
c0 70:;:;uCIl'0)0::: 60?f!.
50
40
.... -.- ..... _.- ..... -._ ..... -._ ..... -._ ..
•.......•.......•.......•.......•>E ••••••••
)(---.--)*(----))oE-(---7<)«(----4lX
..... - •• ...L--- .. -...... ..-........ _........ _....;,
- +-. NF90••••. NF200
_ •• - NF270
--7E-DK
8 10 12 14 16 18 20 22 24
Concentration (mg/L)
642
30 -f---.---,---r----,----r---,----,--,--,----,----...------,
o
Figure 13: Effect ofpesticide concentration on rejection of dimethoate at operating
pressure of6 x 105 Pa and stirring rate 1000 rpm.
However, concentration gradient had been acknowledged in various nanofiltration
models such as solution-diffusion model and Spiegler-Kedem model as the driving force to
solute transport (Bhattacharya and Ghosh, 2004). Decrease of pesticides rejection was observed
with the increase of feed concentration. This was due to the increased competitive sorption on
membrane sites between feed solute and water (Williams et al., 1999). The solute activity at the
influent side became higher, which resulted in higher diffusion of the solute through the
membrane. Another possible explanation for this trend of result was a reduction of the net
driving force due to the increase of osmotic difference between the retentate and permeate when
the feed concentration increased (Freger et al., 2000). This resulted in lower water flux, which
translated into more concentrated permeate.
Effect of Feed Pesticide Concentration on Permeate Flux
Figure 17 and Figure 18 show flux performance of the membranes tested during
atrazine and dimethoate rejection respectively. The same pattern of permeate flux was observed
in this experiment as there was also small decrement on permeate flux with an increase in feed
pesticide concentration.
2.50E-05 ..... - ......... _ ..... -.--- ...... _ ....... _.-..6.
2.00E-05
......III~ 1.50E-05",-E......~ 1.00E-05ii:
..... -• . . -+- . - - -+- . - . -+- . - . -+- - - . -+
- ....•..... - •.... _-.- .....•.... _-.
- ....... NF90_•• - ·NF200
- -. - NF270
--*-DK
5.00E-06)( )( )( )( )( )(
6 8 10 12 14 16 18 20 22 24
Concentration (mg/L)
O.OOE+OO +----,.--,---,-------,--...,--r---.---,----,.--,---,----,
024
Figure 14: Effect ofpesticide concentration on permeate flux during rejection of
atrazine at operating pressure of6 x 10\Pa and stirring rate 1000 rpm.
.... - •• -jr .. -'. -jr .. -., -jr .. - .. -jr •. -"-4
~'-.-.~._.-.~._.-.~.-._.~._._.~
..............................................
2.50E-05
2.00E-05
'iii 1.50E-05N
E......§.>< 1.00E-05:::li!
5.00E-06)( )( )( )( )( )(
- .•. - NF90
...•... NF200
-'4-' NF270
~DK
O.OOE+OO +--,---,---,--.,.----,,.------,--.---,---,---,-----.---.
o 2 4 6 8 10 12 14 16 18 20 22 24
Concentration (mg/L)
Figure 15: Effect ofpesticide concentration on permeate flux during rejection of
dimethoate at operating pressure of 6 x 105 Pa and stirring rate 1000 rpm.
The slight decrease ofpermeate flux was due to the increase of concentration difference
between the two sides of the membrane and tl1.e subsequent increase in the osmotic pressure that
opposed the permeate flow (Ahn et al., 1999). Moreover, the increased competitive sorption on
membrane sites between feed solute and water also caused the permeation of water through
membrane to decrease, thus resulting in flux decline (Williams et al., 1999). There was also
reduction of the net driving force due to the increase of osmotic difference between the retentate
and permeate when the feed concentration increased, causing lower water flux (Freger et al.,
2000).
Effect of Stirring Rate .,.The effect of stirring rate on pesticide rejection and permeate flux were investigated at
various stirring rate (300 to 1000 rpm). This range of stirring rate was chosen because the
motion of stirrer beyond it was no longer smooth. A fixed parameter including operating
pressure of 6 x 105 Pa and pesticide concentration of 10 mg/L were chosen for these
experiments. \,
Effect of Stirring Rate on Pesticide Rejection
Figure 19 and Figure 20 show the rejection performance of atrazine and dimethoate by
NF90, NF200, NF270 and DK at different stirring rate. An increase of pesticide rejection was
observed for both atrazine and dimethoate with the increase ofstirring rate.
This observation was obtained because increasing the water turbulence on the
membrane surface disturbed the onset of the mass transfer boundary layer near the membrane
wall and reduced the accumulation of solutes on the membrane surface, thus, reducing the effect
of concentration polarization (AI-Bastaki and Abbas, 2001; Bhattacharjee and Datta, 2003).
Usually, in the case of cross-flow filtration, the most direct technique to promote mixing is to
increase the fluid velocity past the membrane surface or integrating membrane spacers (Baker,
~._._._.~._._._.~._._._._._.-+
•...........•...........•.................;)( )( )(
2000).
100
90
80
r::.2 70...(J
.S!,Gl0:: 60~
50
40
30100 300
..
500
..
700 900 1100
_ .•. - NF90
. ..•... NF200
-...-NF270
---*-DK
Stirring rate (rpm)
Figure 16: Effect ofstirring rate on rejection ofatrazine at operating pressure of6 x
lOs Pa and feed concentration of 10 mg/L.
,,
"'--.'_'._'.-'.A",-- .. - .. -.-.- ..- .. -.- '
.........................•• • • • • • • ' )()( X
~
_.~._._.~._.-.-.- ..+-. -'
100
90
80
c0 70;lCJQ)
'Qi'n:: 60';/!.
50
40
30100 300 500 700 900 1100
- ..... NF90..•. ·NF200- .• - NF270~DK
Stirring rate (rpm)
Figure 17: Effect of stirring rate on rejection of dimethoate at operating pressure of6 x
105 Pa and feed concentration of 10 mgIL.,.
In the case of dead-end filtration, increasing the stirring rate would reduce the boundary
layer thickness by increasing turbulent mixing at the membrane surface. The reduced membrane
wall concentration occurring on the membrane surface with the increased stirring rate resulted in
lower concentration gradient. Therefore, as shown in Figure 19 and Figure 20, better rejection
was obtained.
IEffect of Stirring Rate o'n Permeate Flux
.;Flux performance ofNF90, NF200, NF270 and DK tested during atrazine and
dimethoate rejection at different stirring rate is presented in Figure 21 and Figure 22
respectively, Only slight increase was observed in the flux performance for all membrane tested,
This is because the main driving force of the flux was pressure gradient and in this experiment,
constant pressure was applied to the membra~e. The slight improvement of flux was obtained,
due to the reduced boundary layer caused by the increasing turbulent mixing at the membrane
surface.
2.50E-05Jr •. _ ......... _ ......... _ •• _ .......
2.00E-05
.......III
",' 1.50E-05E,;;-E.....~ 1.00E-05ii:
... .-.,-,+-.-.-. '-'_.+-'-'-........•.......•........... .. - +-. NF90
•••• ·NF200_ •., - NF270
---*-DK
5.00E-06 )( )( )( )(
1100900700500300
O.OOE+OO +----,--------r-----,------.---------,
100
Stirring rate (rpm)
Figure 18: Effect of stirring rate on permeate flux during rejection ofatrazine at
operating pressure of6 x 105 Pa and feed concentration of 10 mglL.
2.50E-05
Jr •• _ .......... _ ......... _ •• _ .......
2.00E-05
'iii",' 1.50E-05E
",-.§.
~ 1.00E-05ii:
.-._.-+.... _._ .... _._ .... -•.......•.......•...... - .
- +-. NF90
•••• ·NF200_ .., - NF270
---*-DK
)( )( )( )(
I
of
300 500 700 900 1100
Stirring rate (rpm)
O.OOE+OO +----,-------,------,-----,-----,
100
5.00E-06
Figure 19: Effect ofstirring rate on permeate flux during rejection of dimethoate at
operating pressure of6 x 105 Pa and f~ed concentration of 10 mgIL.,
Having similar phenomenon as the earlier section, membrane wall concentration
occurring on the membrane surface decreased with the increased stirring rate. This caused the
subsequent decrease in the osmotic pressure that less opposed to the permeate flow (Ahn et al.,
1999) while at the same time, competitive sorption on membrane sites between feed solute and
water decreased (Williams et al., 1999). Consequently, the increased permeate flux was
observed.
Effect of Initial pH of Solution
The effect of initial pH of solution on pesticide rejection and permeate flux were
investigated at several pH of solution (4 to 9). A fixed parameter including operating pressure of
6 x 105 Pa, pesticide concentration of 10 mgIL and stirring rate of 1000 rpm were chosen for
these experiments.
Effect of Initial pH of Solution on Pesticide Rejection
The effect of initial pH of solution on the atrazine and dimethoate rejection at fixed
operating pressure, pesticide concentration and stirring rate are presented in Figure 23 and
Figure 24. From the figures, it can be seen that the rejection performance for atrazine and
dimethoate by NF200, NF270 and DK increased as the pH was increased while the rejection
trend for NF90 was almost constant regardless ofthe pH condition.
~._-_.-.-._.-.-._._~._._.-.-.-.~
..........•",.-"..•..... '
~.. -" .. -·.·-NF90
...•.•• NF200
--,&-' NF270
--*-DK
10987;,
pH
6
.'.. '.. ,.'
54
~.,
k" •
100
90
80
70c0
60:o:lUCII'Cjj'a:: 50~
40
30
20
10
3
Figure 20: Effect of initial pH of solution on rejection of atrazine at operating pressure
of6 x 105 Pa, feed concentration of 10 mg/L and stirring rate 1000 rpm.
Polyamide thin-film composite membranes have charge characteristics that influence
the separation capabilities. This can be altered by pH of solution. The isoelectric point of
polyamide membrane is generally between 4 to 5 (Puasa, 2006). The occurrence of an
isoelectric point means that at lower pH than the isoelectric point, the membrane is positively
charged and vice-versa. Hence, in the case of polymeric membranes, surface membrane charge
is typically negative at high pH values, it increases as the pH decreases and switches to positive
values at low pH (Bandini and Mazzoni, 2005).
~'-'-'---'-'---'-'-~'-'---'-'-'-'
••.'
--.·-NF90...•... NF200
_.JL_. NF270
~DK
10987654
-'_.,k'"' •
.'.'
_.,_.,
.. '.•..~ ........ ,.. '.......
......
-..:.:':'~'.:..'-'-------~~'lJ •• -_ .. ...c.'
100
90
80
70c0
60+:llJQ).....Q)
c::: 50'#.
40
30
20
103
pH
1000 rpm.
Figure 21: Effect of initial pH ofsolution on rejection ofdimethoate at operating
pressure of6 x 105 Pa, feed concentration of 10 mgIL and stirring rateI
"f
However, in contrary to the usual phenomenon which occurs for ionic species whereby
at isoelectric point, the flux is usually at the highest while the rejection is at the lowest (Ooi,
2005), the trend observed for the uncharged pesticides molecules is somewhat different. In the
case of uncharged molecules, instead of being influenced by the changes in membrane surface
charge, it is believed that it was the changes of\the membrane structures and/or formation of
high molar mass complexes which significantly affected the performance of solute rejection and
permeate flux.
According to Nystrom et ai. (1995), the retention of vanillin was very low at low pH but
the retention was high at pH 10. It was deduced that vanillin formed high molar mass complexes
at this pH. Nevertheless, the possibility of formation of high molar mass complexes at high pH
was sidelined in this research since the rejection of atrazine and dimethoate only increased at
high pH for NF200, NF270 and DK while NF90 showed a slight decrease of rejection at high
pH.
Hence, it is deduced that the trend of atrazine and dimethoate rejection obtained for
NF200, NF270 and DK in this experiment was due to the changes of the membrane structures
caused by the pH of solution. The results obtained were in agreement with observation done by
Freger et al. (2000) whereby the rejection of lactate decreased with the decrease of pH. It was
concluded that acidic hydrolysis disrupted the chemical links, which reduced the degree of
crosslinking (Le., rigidity) of the polymer matrix and caused the polyamide polymer to become
more hydrophilic (Freger et al., 2005).
On the other hand, the increase ofatrazine and dimethoate rejection at high pH observed
for NF200, NF270 and DK could be caused by the hydration swelling of the membrane skin
layer. This could result in shrinking of membrane pore size, and thus, reduced the permeation of
solute through the pores of the membrane (Freger et al., 2000). However, it is believed that
NF90 was rather chemical-resistant as it showed somewhat consistent performance regardless of
the pH of solution. There was only a &op of about 3% of rejection performance for NF90
compared to the obvious increase or reduction of rejection performance shown by the rest of the
nanofiltration membranes tested.
Effect of Initial pH of Solution on Permeate Flux
The effect of initial pH of solution on the permeate flux during rejection of atrazine and
dimethoate at fixed operaAng pressure, feed pesticide concentration and stirring rate are
presented in Figure 25 and FiglJfe 26, respectively. Since the acid hydrolysis (Freger et al., 2005)
and swelling of membrane skin layer (Freger et al., 2000) is believed to be responsible for the
increase or decrease in pesticide rejection for NF200, NF270 and DK, it is expected that the
permeate flux would be as much affected by pH of solution as the pesticide rejection
performance.;,
2.50E-05
2.00E-05
...._.. _.. _.. - .. _....... -........... - ..-..
J! 1.50E-05E.,-g~ 1.00E-05
u:::
~._.-._._._._._.~._.-._.-.-.
•.......................•...............•
-·.·-NF90
...•... NF200
-':A-' NF270
-x--DK
5.00E-06)( )(
10987654
O.OOE+OO +---,----r------,-----,---....,----,------,
3
pH
Figure 22: Effect of initial pH of solution on permeate flux during rejection of atrazine at
operating pressure of 6 x 105 Pa, feed concentration of 10 mglL and stirring rate
1000 rpm.
2.50E-05
2.00E-05
..... - ........ -. "'_ ... - ........ - .. - .. - ....
J! 1.50E-05.§g~ 1.00E-05
~._._.-.-._._._.~._._._._.-....~ ..,
-·.·-NF90
...•... NF200
-':A-' NF270
~DK
5.00E-06)( )( )(
1098; 7,654
O.OOE+OO -j----,----,.-----,-----,-----,---,--------,
3
pH
Figure 23: Effect of initial pH of solution on permeate flux during rejection of dimethoate at
operating pressure of 6 x 105 Pa, feed concentration of 10 mg/L and stirring rate
1000 rpm.
However, based on Figure 25 and Figure 26, it seemed that except for NF270, the effect
of pH of solution seemed not be to as much on permeate flux if compared to the degree of
changes seen in the rejection performance. Thus, it is deduced that the difference in permeate
flux was not that obvious because the changes at the polymer was little, but it was sufficient to
efficiently retain or allow more solutes through the membrane. Again, NF90 showed that it was
somewhat resistant to the changes of pH of solution as it showed almost constant flux
performance regardless ofthe pH condition.
Effect of Binary Solnte Mixture
The effect of binary solute mixture on pesticide rejection were investigated at several
ratio of atrazine:dimethoate. The ratio of atrazine:dimethoate was set at 0:100, 20:80, 50:50,
80:20 and 100:0 for a total of 10 mgIL pesticides (i.e. 0:100 means 0% of atrazine and 100%
dimethoate in the solution, 20:80 means 20% of atrazine and 80% of dimethoate in the solution
and so on). Fixed parameters of operating pressure at 6 x 105 Pa and stirring rate at 1000 rpm
were chosen for these experiments.
Effect of Binary Solute Mixture on Pesticide Rejection.Rejection of binary atrazine-dimethoate mixture was tested at a fixed applied pressure,
total pesticides concentration and stirring rate to examine if the membrane would have the same
good performance when the two pesticides co-exist. Figure 27 (a) to (d) shows that all four
nanofiltration membranes tested had slightly lower retention for both atrazine and dimethoate in
the presence of binary solute mixture compared to the single solute condition.
This observationV(as in line with observation made by Plakas et al. (2006) which
suggested that simultaneous filtration of more than one pesticide resulted in a kind of"f
competitive adsorption on the membrane surface and, thus, created a greater passage to the
permeate side. These results were also in agreement with the report by Kiso et al. (2000) which
found that herbicides displaying higher rejection in single ,solute solutions may permeate more
in mixed solute systems.
;,
100 100
• ......•90 90
80 80c c0 0
70:;:; 70 :;:;u u .----.CIl CIl'Gj' 'Gj' • •a::: 60 a::: 60C>l! C>l!
50 50
40 40
30 30
0 20 40 60 80 100 0 20 40 60 80 100
%Atrazine %Atrazine
(a) (b)
100 100
90 90
80 80
c c,2 70
,2 70... tlu CIlCIl'Gj'
60~ 'l 60II::
~ ~
50 .--.. 50
•40 • 40
30 300 20 40 60 80 100 0 20 40 60 80 100
% Atrazine % Atrazine(c) (d)
I-+--Atrazine --Dimethoate II
Figure 24: Effect ofbinary solute mixture on rejection ofatrazine and dimetboate on (a) NF90,
(b) NF200, (c) NFflo and (d) DK at operating pressure of6 x 105 Pa, total
pesticides concentration of 10 mgIL and stirring rate 1000 rpm.
Effect of Binary Solute Mixture on Permeate Flux
The effect of binary solute mixture on"permeate flux is presented in Figure 28. The
effect of binary solute mixture on permeate flux was minor since the pressure, total
concentration and stirring rate in this experiment were set constant and the only changes of
parameter was the ratio ofpesticides. Therefore, this posed little influence to the performance of
permeate flux. This observation provides some insight that in the real case of water treatment,
binary or multiple solute mixture would have little significance in the permeate flux
performance.
2.50E-05
2.00E-05.. - .. -... .. - .. - .. - .... . - .. - .. - ... .. - ........
Iii'N' 1.50E-05
...~§.><~ 1.00E-05iL
. _. _. +-. _. _. _. -+-. -' _. _. -+-. _. - .•........•............•............•........•
- +-. NF90••••. NF200
_ •• - NF270
~DK
5.00E-06
10080604020
O.OOE+OO +----,...-----,-----.-----,-----,o
% Atrazine
Figure 25: Effect of binary solute mixture o~ permeate flux during rejection of mixture of.atrazine and dimethoate at operating pressure of 6 x 105 Pa, total pesticides
concentration of 10 mg/L and stirring rate 1000 rpm.
Statistical Analysis using Design.of Experiment method
Since operating pressure, feed pesticide concentration and stirring rate were the main
operating conditions for napofiltration system tested, a statistical approach using general
factorial design was employed to compare the significance of their contribution in affecting the
rejection and flux performance:The factor of membrane selection was also included in the
analysis to compare its significance in affecting the final results compared to the contributions
by the operating conditions.
General Factorial Designi,
Experiments based on the general factorial experimental design in Table 2 were carried
out and relevant results, which lists the rejection and flux performance for both atrazine and
dimethoate, is presented accordingly. The results were further analyzed using the Design Expert
6.0.6. software. Since membrane (NF90, NF200, NF270 and DK) was categorical factor while
operating pressure, feed pesticide concentration and stirring rate were numerical factors, the
relationship between those numerical factors and the two dependent variables (% rejection and
flux) was analyzed for each membrane.
Table 2: Experimental design for atrazine and dimethoate.
Run no. A: Membrane B: Pressure c: Pesticide D: Stirring rate
x lOs concentration
(Pa) (mg/L) (rpm)
I NF90 6 2 3002 6 2 10003 6 20 3004 6 20 10005 12 2 3006 12 2 10007 12 20 3008 12 20 10009 NF200 6 2 30010 6 2 1000II 6 20 30012 6 20 100013 12 2 30014 12· 2 1000IS 12 20 30016 12 20 100017 NF270 6 2 30018 6 2 100019 6 20 30020 6 20 100021 12 2 30022 12 2 100023 12 20 30024 12 20 100025 DK -I 6 2 30026 6 2 100027 6 20 30028 6 20 100029 12 2 30030 12 2 100031 12 i 20 300,32 12 20 1000
ANOVA Analysis
Table 3 shows the experimental results based on the general factorial experimental
design. The results in Table 3 allowed the development of mathematical equations where each
response (% rejection and flux) is assessed as a function of pressure (B), pesticide concentration
(C) and stirring rate (D) and the 2-factors interaction among the independent variables (AB, AC,
AD, BD, BC, and CD) for each membrane (A).
Table 1: Experimental results based on the general factorial experimental design.
Atrazine DimethoateRUl.,lno.
12345678910111213141516171819202122232425262728293031
32
% Rejection
98.2998.7397.2397.7598.5298.9398.5897.9366,4367.5463.0168.9878,4274.8377.8175.0048.3555.5540.9748!9664.38
"I"6l.65 .
56,4259.7354.6765.4352.6859.3970.0472.9567.14
63.89
1.28E-051.2lE-051.17E-051,45E-052.56E-052.79E-052.50E-052.55E-05l.OlE-05
"l.20E-05'9.67E-061.08E-051.93E-052.23E-052.l2E-052.07E-051.96E-052.03E-051.99E-052.00E-053.76E-053.97E-053.6lE-053.8lE-054.32E-065.79E-064. 11E-065,400-068.93E-061.l0E-058.34E-06
1.2lE-05
% Rejection
85.2287.7783.0285.9391.299l.7889,7388.1151.4951.9254.6262.1465.7462.7359.0655.0448.6135.8336.3334.8353.4158,4847.7961.6644.5850.0143.1646.2864.8574.0162.51
65.31
1.13E-051.37E-051.3lE-051.37E-052,43E-052.83E-052.25E-053.02E-05l.OlE-051.00E-059.55E-069.89E-061.98E-052.08E-051.90E-051.97E-051.89E-051.93E-051.86E-051.94E-053.62E-054.13E-053.74E-053.83E-054.8lE-065.10E-064.84E-064.99E-069.66E-069.33E-069.84E-06
9.53E-06
The statistical significance of the factors were evaluated by F-test and "Prob>F" using
analysis of variance (ANOVA) and is presented in Table 4 and Table 5 for atrazine and
dimethoate, respectively. As shown in the tables, there are certain model terms with
"Prob>F>O.5" which indicated that those terms were insignificant to the rejection or flux
performance. Those insignificant terms can be reduced by reselecting only the significant terms
to be included in the model in order to improve the accuracy ofthe model equation.
Table 2: ANOVA for the significance ofmodel and model terms for atrazine.
Source Model Sum of Mean F Value Prob> F Remarks
terms Squares Square
% Rejection
Model 9543.82 530.21 129.53 < 0.0001 significant
A 8570.81 2856.94 697.96 < 0.0001 significant
B 546.51 546.51 133.51 < 0.0001 significant
C 75.82 75.82 18.52 0.0009 significant
D 36.77 36.77 8.98 0.0103 significant
AB 166.05 55.35 13.52 0.0003 significant
AC 47.14 "15.71 3.84 0.0361 significant
AD 31.09 10.36 2.53 0.1024
BC 0.24 0.24 0.06 0.8114
BD 69.34 69.34 16.94 0.0012 significant
CD 0.05 0.05 0.01 0.9135
Flux
Model 3.17E-09 1.76E-1O 182.71 < 0.0001 significant
A 1.89E-09 6.30E-1O 653.67 < 0.0001 significant
B 1'!08E-09 1.08E-09 1125.63 < 0.0001 significant
C 1.16E-12 1.16E-12 1.20 0.2926
D 1.80E-II 1.80E-II 18.67 0.0008 significant
AB I.72E-IO 5.73E-II 59.47 < 0.0001 significant
AC 5.85E-13 1.95E-13 0.20 0.8929
AD 1.21E-12 4.d2E-13 0.42 0.7437
BC 6.13E-13 6.13E-13 0.64 0.4393
BD 1.28E-12 1.28E-12 1.33 0.2691
CD 9.79E-14 9.79E-14 0.10 0.7550
Table 3: ANOVA for the significance ofmodel and model terms for dimethoate.
Source Model Sum of Mean F Value Prob> F Remarks
terms Squares Square
% Rejection
Model 9178.98 509.94 19.98 < 0.0001 significant
A 7512.23 2504.08 98.11 < 0.0001 significant
B 1125.21 1125.21 44.08 < 0.0001 significant
C 55.62 55.62 2.18 0.1637
D 28.96 28.96 1.13 0.3062
AB 376.44 125.48 4.92 0.0169 significant
AC 18.88 6.29 0.25 0.8623
AD 28.84 9.61 0.38 0.7714
BC 17.95 17.95 0.70 0.4168
BD 7.07 7.07 0.28 0.6074
CD 7.76 7.76 0.30 0.5907
Flux
Model 3.31E-09 1.84E-10 149.43 < 0.0001 significant
A 1.93E-09 6.4'3E-10 522.84 < 0.0001 significant
B l.llE-09 '1.11E-09 903.94 < 0.0001 significant
C l.77E-13 1.77E-13 0.14 0.7106
D l.77E-ll 1.77E-ll 14.40 0.0022 significant
AB 2.25E-10 7.49E-1l 60.91 < 0.0001 significant
AC 1.67E-12 5.55E-13 0.45 0.7206
AD 1.63E-11 5.42E-12 4.41 0.024 significant
BC 5.16E-13 5.16E-13 0.42 0.5285I
BD . 6.1lE-12 6.11E-12 4.96 0.0442 significant
CD l.!2E-13 1.22E-13 0.10 0.7576
The statistical parameters obtained from the ANOVA for the reduced model of the
responses are given in Table 6. ANOVA results of the reduced models showed Prob>F<O.OOOl
for all the responses, indicating that the model equation adequately described the rejection and,
flux performance in the interval of investigation. The effect of each variable on the response
was the combination of coefficients and variable values as well as contribution of interaction of
variables that cannot be observed by conventional experimental methods. A high R2 value of
close to 1, is desirable to indicate that 1ack-of-fit is insignificant (Noordin et al., 2004). In this
statistical analysis, the values of R2 for % rejection and flux for atrazine and dimethoate were
in Table 7.
0.9912,0.9945,0.9478 and 0.9944, respectively. This indicates that only 0.56-5.22% ofthe total
variation was not explained by the model.
Table 4: Statistical parameters obtained from the ANOVA for the reduced models.
Variable Atrazine Dimethoate
% Rejection Flux % Rejection Flux
Significant terms A,B,C,D, A,B,D,AB A,B,AB A,B,D,AB,
AB,AC,BD AD,BD
Prob>F <0.0001 <0.0001 <0.0001 <0.0001
R2 0.9912 0.9945 0.9478 0.9944
Adequate precision 40.44 74.56 22.56 55.95
Standard deviation, SD 2.17 8.72E-07 4.5502 9.86E-07
Coefficient ofvariance, CV 3.02 4.87 7.31 5.60
The coefficient of variance, CV, is the ratio of the standard error to the mean value of
the observed response (as a percentage). It is a measure of reproducibility of the model. As a
general rule of thumb, a model can be considered reasonably reproducible if its CV is not
greater than 10% (Wong, 2006). None of the tv values obtained for the responses studied in.this experiment exceeded 10%, as shown in Table 6, indicating the good reproducibility of the
model.
Meanwhile, adequate precision shown in Table 6 compares the range of the predicted
values at the design points to the average prediction error. Ratio greater than 4 is considered to
be adequate model discrimination (Noordin et al., 2004). All the adequate precision values
obtained in this analysis (40.44, 74.56, 22.56, 55.95) were well above 4, indicating the modelsI
were significant. Empirical models in terms of coded values for the two responses are tabulatedwi!
,,
Table 5: Empirical models in terms ofcoded values.
Membrane Empirical models
NF90 Atrazine:Rejection = 87.78 + 0.99*B - 0.04*C + 0.02*D -0.014*B*D
Flux = -1.86E-06 + 2.2IE-06*B + 2.14E-09*D
Dimethoate:Rejection = 80.74 + 0.79*B
Flux = -1.33E-06 + 1.95E-06*B + 1.49E-09*D + 4.l6E-IO*B*D
NF200
NF270
Atrazine:Rejection = 46.65 + 2.58*B - 0.03*C + 0.02*D -0.014*B*D
Flux = -9.82E-07 + 1.7lE-06*B + 2.14E-09*D
Dimethoate:Rejection = 49.44 + 0.93*B
Flux = 1.9lE-06 + 1.38E-06*B - 3.00E-09*D + 4.16E-IO*B*D
Atrazine:Rejection = 29.82 + 2.93*B -·0.33*C + 0.02*D -0.014*B*D
Flux = 6.25E-07 +2.99E-06*B +2.14E-09*D
Dimethoate:Rejection = 22.46 + 2.74*B
Flux = 5.54E-07 + 2.94E-06*B - 1.15E-09*D + 4.l6E-IO*B*D
DK Atrazine:Rejection = 40.41f. + 2.65*B - 0.28*C + 0.02*D -0.014*B*D
Flux = -1.64E-06 + 8.62E-07*B + 2.l4E-09*D
Dimethoate:Rejection = 25.35 + 3.44*B
;,Flux = 2.76E-06 + 5.05E-07*B - 3.8lE-09*D + 4.l6E-IO*B*D
Usually, it is necessary to check the fitted model in order to ensure that it provides an
adequate approximation to the real system. Unless the model shows an adequate fit, proceeding
with investigation of the fitted model is likely to give poor or misleading results. By applying
the diagnostic plots such as the predicted against actual values plot, normal probability plot and
studentized residuals against predicted plots, the model adequacy can be judged. The predicted
against actual value plot of % rejection and flux for atrazine and dimethoate is presented in
Figure 29. The model for % rejection and flux for both pesticides were satisfactory as the actual
values were distributed relatively near to the straight line as shown in the figures.
1011.9/
nIl
31.33
.-__Predlctedvs. Actual -.
Actual
(a)
Predicted vs, Actualr----- ------,
III
.1 m
/UActual
3.07£-01
l::Ii11;.1J5
WE·O'
3MOI
;,
".--_............:Predictedvs. Actual. -.
/
Actual
(b)
r--............ PredictedVs. Actual -.
..
Actual
(c) (d)
Figure 26: DESIGN-EXPERT plots for atrazine and dimethoate. Predicted versus actual values
plot for (a) % rejection; (b) flux for atrazine and (c) %rejection; (d) flux for dimethoate.
Figure 30 shows the normal probability plots for atrazine and dimethoate which
indicates whether the residuals follow a normal distribution. If the model is adequate, the points
on the normal probability plots of the residuals should form a straight line (Idris et al., 2006). A
check on points on normal probability plots of the residuals reveal that the residuals fall on a
straight line implying that the errors are distributed normally for all the responses.
III
-1,01
Normal.PlotofResiduals....--- ----I
II-l
~95.::t
3
:1
;1f01r~
};f]
I
•
r-__Normal.PlotofResiduals._.-..-..........,
91
i 90
.D !II,ro.De
Q.III
'f.i 311
E 200Z 10
Sludentized Residuals(a)
,.--__Normal Plot of Residuals__...,
Stude~dRe~duaJs
(b)
•..--.-- ........NormaFPlotofResiduals. -,
99
95
~90
1l 80III.D 10eQ.
~o
~
1 30
20
Z 10
III
III
\,
StudenliZedResiduals Sludentized Residuals
(c) (d)
Figure 27: DESIGN-EXPERT plots. Normal probability of residual for (a) % rejection; (b) flux
for atrazine and (c)% rejection; (b) flux for dimethoate.
Figure 31 shows the plots of studentized residual versus predicted values for atrazine
and dimethoate. The plots of the residuals against the predicted response should be structureless,
that is, they should contain no obvious patterns and unusual structure (Idris et al., 2006). The
figures show a random spread of the studentized residuals across almost all levels of the
predicted values. This implies that the models proposed are adequate.
.--__Residuals vs. Predicted__---..,300-1--------------_--1
...Residuals V$. Predicted
r-----~---..--,
HIl-III
IIIII III • " III II IIIiI. ... • iIII ::l .11 III::l III :!Z III'l1'iii II l1li " II IIIII II .111 41 l1lia: a: III ...
III 'tJ ,-'tJ 0,00 II .. II.Ill. 1lI
~: III. 18~ III III IIIIII i IIc •III 18 'tJ II'l1 II III. II
~III ~ •
811 •
1/III
III
·3.00 .-I I I 1 I
1.11,,06 1:~ me·os 3.06 3m,
.PredictadPradieted
(a) (b)
Residuals.vs.••Predicted Residualsvs. Predicted300f 300
MIl.
llII II II•II!I II
•!!. II !!.II II::l • ::l
III II~ III ~. II II..
II..
I IIIa: a: III
" 0.00. • ~0,(11)..
~ I III III .:!t II IIIII "f" ..c III III c IIIII • III III 111 II IB" " II IIIJl III III JlCIl ·1.10 III CIl -1.50
• IIIII
-3.00 ·300
3tHIII 5l.1J 64.$ mo 1Il.13 I mE·1II! fJ3&O; 131&0; l.OIE·o; 31l6E-$,
Predicted Predictad
(c) (d)
Figure 28: DESIGN-EXPERT plots. Studentized residuals versus predicted values plot for (a)
% rejection; (b) flux for atrazine and (c) %rejection; (d) flux for dimethoate.
The interaction graph between membrane and pressure as well as membrane and
pesticide concentration for each membrane in the case of rejection of atrazine are presented in
Figure 32. It can be seen from Figure 32 (a) and (b) that while the rejection by NF90 was not
much affected by the increase of pressure or pesticide concentration, rejection was generally
higher with higher pressure and rejection was lower when the pesticide concentration was high.
-_.-_..--~
• 6 X 105 Pa
".
'.
B: Pressure
z.,'<:<>,
.. ~".. .. '"
""'" ,'" I",
""'~.,70.53
55.75
100..10
NFSO NF200 NF270
A: Membrane
(a)C: Cone
85A1
70.60
56,]8
.2 mg/L
.&20 mg/L
NF90 NF200 NF270
A: Membrane(b)
Figure 29: DESIGN-EXPERT plots. Interaction graph between (a) membranes and pressure; (b)I,
membranes and pesticide concentration.
Meanwhile, the interaction plot between operating pressure and stirring rate is presented
in Figure 33. The interaction trend for all membranes had a common point whereby they
showed the lowest rejection at low pressure and low stirring rate.
8'PresSUfeWOOO 6.00
D'stirringfate
(a) (b)
(c) (d)
Figure 30: DESIGN-EXPERT plots. Interaction\graph between pressure and stirring rate for (a)
NF90; (b) NF200; (c) NF270; (d) DK for rejection ofatrazine.
Figure 34 shows the interaction graph between membrane and pressure for flux during
rejection of atrazine. The flux was generally higher when the pressure was higher. This trend
corresponded to the finding in the earlier section that pressure highly influenced flux
performance.
3.97E·05
2.9-BE·05
1.9BE'05
B: Pressure
.6 X 105 Pa
.... 12 X 105 Pa
.•.:III:•.....................
9.93E-OG
OOOE+OO
::IE: _._ - ;]IE' •• '
"
A: Membrane
NF270 DK
Figure 31: DESIGN-EXPERT plots. Interaction graph between membrane and
pressure for flux during rejection ofatrazine.
.Figure 35 shows the interaction graph between membrane and pressure for rejection of
dimethoate. The results observed from the interaction graph are in good agreement with the
model for rejection ofdimethoate obtained earlier. A positive sign for the factor B in the model
indicates that the ability ofthe system to achieve higher rejection with the increase in value of
pressure (B).
93.55 -
-
B: Pressure
.6x105 Pa
.&12 X 105 Pa
64 .. 19 -
49.51 -
34.33 -
NF90
I., ". ....~, 1 ······ I
INF270
A: Membrane
Figure 32: DESIGN-EXPERT plots. Interaction graph between membrane and pressure for
rejection ofdimethoate.
Figure 36 shows the interaction graph between membrane and pressure as well as
membrane and stirring rate for flux performance during rejection of dimethoate. The effect of
pressure obviously had high influence on the flux performance as a double increment of
pressure seemed to have much more positive effect on the flux compared to the more than triple
increment ofstirring rate.
4.13E·05
3.20E-OS
~ 22SE-05
4.21E·06
B: Pressure
x----------------.--I
.6 X 105 Pa
... 12 X 105 Pa
4.13E-05 -
3.20E-05 -
2.2BE-05 -
1.35E-05 -
4.21E-Q(s -
I
INF90
NF200
" A: Membrane(a)
D: Stirring rate
INF2QQ\,
NF270
INF270
OK
.300 rpm
... 1000 rpm
IDK
I
A; .••Membranelb)
Figure 33: DESIGN-EXPERT plots. Interaction between (a) membranes and pressure; (b)
membranes and stirring rate for flux during rejection ofdimethoate.
Meanwhile, interaction graph between pressure and stirring rate for each membrane is
presented in Figure 37. All membranes showed the highest flux performance at high pressure
and stirring rate as shown in Figure 37 (a) - (d).
3.11tJ.00 600
(a) (b)
\, 10000 600
~ ~
Figure 34: DESIGN-EXPERT plots. Interaction between pressure and stirring rate for (a) NF90;
(b) NF200; (c) NF270; (d) DK for flux during rejection ofdimethoate.
Confirmation Test
In order to validate the adequacy of the model, four confirmation runs were performed
for each pesticide. The conditions are listed in Table 6 and Table 7 for atrazine and dimethoate,
respectively. The test conditions for the confirmation experiments were obtained from Design
Expert 6.0.6. software. The predicted values were also obtained from the software based on the
models developed previously. The predicted values and the actual experimental values were
compared and the percentage error was calculated. All these values are listed in Table 6 and
Table 7 as well. The percentage error between the actual and predicted value for rejection and
flux rate are within 10% range. Thus, the empirical models developed were reasonably accurate,
for both the rejection and permeate flux.
Modeling of Spiegler-Kedem
Although the Design Expert 6.0.6. software was able to produce regression model for
the performance of the nanofiltration membranes tested, it is important to validate performance
of the membrane using mathematical model. Since NF90 prevailed as the best-performed
nanofiltration membrane in this study, it was subjected to validation using Spiegler-Kedem (SK)
model. Spiegler-Kedem model is a mathema~cal model for reverse osmosis and nanofiltration
process based on irreversible thermodynamics (Ahmad et al., 2005). The rejection and permeate
flux data used for this section was obtained from earlier section on effect ofoperating pressure.
\,
Table 6: Conftrmation experiments for atrazine.
Membrane Pressure Pesticide Stirring % Rejection Flux (m3/m2.s)
x 105 concentration rate Actual Predicted Error Actual Predicted Error
(Pa) (mg/L) (rpm) (%) (%)
NF90 12 15.78 1000 95.02 97.89 3.02 2.49E-05 2.67E-05 7.37
NF200 12 5.07 -- 1000 72.32 76.31 5.52 2.34E-05 2.16E-05 7.61
NF270 12 6.39"; 1000 58.03 61.67 6.27 3.94E-05 3.86E-05 2.02
OK 12 13.99 1000 65.98 67.27 1.96 9.86E-06 1.08E-05 9.50
...-Table 7: Conftrmation experiments for dimethoate.
Membrane Pressure Pesticide Stirring % Rejection Flux (m3/m2.s)
x 105 concentration rate Actual Predicted Error Actual Predicted Error
(Pa) (mg/L) (rpm) (%) (%)
NF90 12 4.42 1000 88.34 90.23 2.14 2.53E-05 2.75E-05 8.77
NF200 12 8.98 1000 61.68 60.64 1.69 2.35E-05 2.12E-05 9.96
NF270 12 6.03 1000 50.41 55.33 9.76 3.99E-05 3.96E-05 0.79
OK 12 10.05 955 59.72 64.67 8.28 9.88E-06 9.95E-06 0.75
Parameter Estimation
The Spiegler-Kedem model was characterized by the hydraulic permeability of the
membrane, Pw ' reflection coefficient, (j', solute permeability, Ps and mass transfer coefficient,
ks ' Results of parameter estimation for NF90 using Levenberg-Marquardt method is shown in
Table 10.
Table 8: Parameters estimated using Lavenberg Marquardt Method on the
experimental results.
Parameter Value
Atrazine Dimethoate
Hydraulic permeability, Pw
Reflection coefficient, (j'
Solute permeability, Ps
Mass transfer coefficient, ks
2.361lE-ll
0.9835
3,4317E-07
1.2894E-05
2.3611E-ll
0.9560
2,4142E-06
1.2524E-05
The reflection coefficient, a, was in good agreement with the results obtained in the
experimental work as it showed that NF90 had the value of an almost ideal membrane. This is
because the value close to 1 meant that it had high ability to pass solvent in preference to solute
(Spiegler and Kedem, 1966), resulting in high rejection of solute by NF90. Meanwhile, atrazine
had slightly higher mass transfer coefficient than dimethoate due to its slightly lower molecular
weight (Schwarzenbach et al., 1993). However, atrazine had obviously lower solute permeability,I
Ps ' compared to dimethoate. This lower solute permeability value possessed by atrazine explains.,.its higher rejection compared to dimethoate.
Comparison between Experimental and Modeling Data
The rejection of pesticide against permeat<v flux curve is presented in Figure 35. It can be,seen from the comparison between the experimental data and predicted data that the Spiegler-
Kedem model provided a good prediction in representing the experimental value. In fact, the
coefficient of determination (R2) obtained for the fitted data was 0.9871 and 0.9692 for atrazine
and dimethoate, respectively. This shows that as confirmed by the irreversible thermodynamics
model, the rejection of solute increased with the increasing permeate flux.
The solute rejection and permeate flux are plotted against applied pressure in Figure 36
and Figure 37, respectively. It is observed that while the solute rejection against pressure curve
for predicted value by Spiegler-Kedem model fitted well with the experimental data, the model
was unable to match the slope of the experimental flux against pressure curve as good as it did in
the case for rejection. However, the trend as experimental data was still in agreement with the
predicted data by the model and did not deviate far from each other.
1.00
0.90
0.80
0.70
c 0.600
+:30.50CJ
CIl'ara::: 0.40
0.30
0.20
0.10
.. . - --. - - - -.............................
0.00 ----,------,----..,----,----,-------,------,
O.OOE+OO 5.00E-06 1.00E-05 1.50E-05 2.00E-05 2.50E-05 3.00E-05 3.50E-05
Flux (m3/m2.s)
• Experim~t atrazine•••• - . SK model atrazine
• Experiment dimethoate
---SK model dimethoate
Figure 35: Solute rejection against permeate flux curve from experimental data and the predicted
results from Spiegler-Kedem model.
\,
.........................................1.00
0.90
0.80
0.70
c 0.600:t:lu 0.50ell'(j)'a::: 0.40
0.30
0.20
0.10
0.000 2
.................
4 6 8 10 12 14 16
Pressure x 105 (Pa)
• Experiment atrazine•••••. SK model atrazine
• Experiment dimethoate
---SK model dimethoate
Figure 36: Solute rejection plotted against preSSlffe using the experimental data and predicted
results from Spiegler-Kedem mod;!:
4.00E-05
3.50E-05
3.00E-05
U! 2.50E-05...E
",-
2.00E-05.§.><:::J 1.50E-05ii:
1.00E-05
5.00E-06
O.OOE+OO0 2 4 6 8 10
I
Pressure x 105 (Pa)
12 14 16
I • Experiment atrazine • Experiment dimethoate --SK model atrazine&dimethoate I
Figure 37: Permeate flux plotted against pressure using the experimental data and
predicted results from Spiegler-Kedem model.
Concentration Polarization Profile
Figure 38 provides the concentration polarization profiles for atrazine and dimethoate at
different operating pressure. The profiles were gauged based on the ratio of membrane wall
concentration to bulk concentration (Cm/Cb) (Ahmad et at., 2007). Based on the membrane wall
concentration calculated from the Spiegler-Kedem model, the concentration polarization profile
was depicted to increase with the increasing pressure. Both solutes demonstrated similar trends on
concentration polarization where the solute concentrations increased from the initial bulk
concentration to the maximum concentration (wall concentration) at the maximum pressure
applied. Although previous results showed that the rejection increased with the increasing
pressure, these profiles show that the effect of concentration polarization would be magnified
with the increasing pressure. The same trend was also observed by Bhattacharjee and Datta
(2003). Thus, due consideration should be given when choosing the suitable applied pressure for
nanofiltration system.
18
16
14
12
'" 100-E0 8
6
4
2
00 2 4 <if. 6 8 10 12 14 16
Pressure x 105 (Pa)
I- .+ - .Atrazine • Dimethoate I
Figure 38: Concentration polarization profile plotted against pressure.
;,
Reference:
Agbekodo, K M., Legube, B. and Dard, S. (1996). Atrazine and simazine removal mechanisms
by nanofiltration: influence of natural organic matter concentration. Water Research,
30(11), 2535-2542.
Ahmad, A.L., Chong, M.F. and Bhatia, S. (2005). Mathematical modelling and simulation of
multiple solutes system for nanofiltration process. Journal ofMembrane Science, 253 (1
2), pp 103-115.
Ahmad, A.L., Chong, M.F. and Bhatia, S. (2007). Mathematical modeling of multiple solutes
system for reverse osmosis process in palm oil mill effluent (POME) treatment. Chemical
Engineering Journal [In Press].
Ahn, KH., Song, KG., Cha, H.Y. & Yeom, LT. (1999). Removal of ions in nickel
electroplating rinse water using low-pressure nanofiltration. Desalination, 122, p.77-84.
AI-Bastaki, N. and Abbas, A. (2001). Use of fluid 1nstabilities to enhance membrane performance:.a review. Desalination, 136, 255-262. .,
Bandini, S. and Mazzoni, C. (2005). Modelling the amphoteric behaviour of polyamide
nanofiltration membranes. Desalination, 184, 327-336.
Baker, R.W. (2000). Membrane technology and applications. New York: McGraw-Hill. 162-173.
Bellona, C., Drewes, J. E., Xu, P. aIJft Amy, G. (2004). Factors affecting the rejection of organic
solutes during NF/RO treatment-a literature review. Water Research, 38, 2795-2809.
Bhattacharjee, C. and Datta, S. (2003). Analysis ofpolarized layer resistance during ultrafiltration
of PEG-6000: an approach based on filtration theory. Separation and Purification
Technology, 33, 115-126. \
Bhattacharya, A. and Ghosh, P. (2004). Nanofiltration and reverse osmosis membranes: theory
and application in separation of electrolytes. Reviews in Chemical Engineering, 20(1-2),
111-173.
Braeken, L., Ramaekers, R., Zhang, Y., Maes, G., Van der Bruggen, B. and Vandecasteele, C.
(2005). Influence of hydrophobicity on retention in nanofiltration of aqueous solutions
containing organic compounds. Journal ofMembrane Science, 252, 195-203.
Causserand, C., Aimar, P., Cravedi, J.P. and Singlande, E. (2005). Dichloroaniline retention by
nanofiltration membranes. Water Research, 39,1594-1600.
Chang, S.,Waite, T.D., Schafer, AI., Fane, AG., 2002. Adsorption of trace steroid estrogens to
hydrophobic hollow fibre membranes. Desalination, 146,381-386.
Chooi, L.K. (2005). Personal communication. 27 May.
Freger, V., Arnot, T.C. and Howell, J.A (2000). Separation of concentrated organic/inorganic
salt mixtures by nanofiltration. Journal ofMembrane Science, 178, 185-193
Freger, V., Bottino, A, Capannelli, G., Perry, M., 6itis, V. and Belfer, S. (2005).
Characterization of novel acid-stable·Ny membranes before and after exposure to acid
using ATR-FTIR, TEM and AFM. Journal ofMembrane Science, 256,134-142.
Hilal, N., AI-Zoubi, H., Darwish, N.A and Mohammed, A.W. (2004). A comprehensive review
of nanofiltration membranes: Treatment, pretreatment, modelling, and atomic force
microscopy. Desalination, 170,281-308.
I
Hilal, N., AI-Zoubi, H., Darwish N.A and Mohammed, AW., (2005). Characterization of-I..
nanofilteration membranes membranes using atomic force miscoscopy, Desalination, 177,
187-199.
Idris, A, Kormin, F. and Noordin, M.Y. (2006). Application of response surface methodology in
describing the performance of thin fllm composite membrane. Separation and
Purification Technology, 49, 271-280.
Lefebvre, X., Palmeri, J., Sandeaux, I., Sandeaux, R., David, P., Maleyre, B., Guizard, C.,
Amblard, P., Diaz, J.-F. and Lamaze, B. (2003). Nanofiltration modeling: a comparative
study of the salt filtration performance of a charged ceramic membrane and an organic
nanofilter using the computer simulation program NANOFLVX. Separation and
Purification Technology, 32, 117-126.
Kim, T.-V., Amy, G. and Drewes, lE. (2005). Rejection of trace organic compounds by high
pressure membranes. Water Science and Technology, 51, 335-344.
Kim, T.-V. (2006). Personal communication. 17 October.
Kimura, K, Amy, G., Drewes, l and Watanabe, Y. (2003). Adsorption of hydrophobic
compounds onto NF/RO membranes: an artifact leading to overestimation of rejection.
Journal ofMembrane Science, 221, 89-101.
Kiso, Y., Nishimura, Y., Kitao, T. and Nishimura, K (2000). Rejection properties of non
phenylic pesticides with nanofiltration membranes. Journal ofMembrane Science, 171,
229-237.
Kiso, Y., Sugiura, Y., Kitao, T. and Nishiml,lra, K (2001). Effects of hydrophobicity and
molecular size on rejection of aromatic pesticides with nanofiltration membranes.
Journal ofMembrane Science, 192,1-10.
Kosutic, K, Furac, L., Sipos, L. and Kunst, B. (2005). Removal of arsenic and pesticides from
drinking water by nanofiltration membranes. Separation and Purification Technology,
42,137-144.
<If.
Krieg, H.M., Modise, S.l, Keizer, K and Neomagus, H.W,J.P. (2004). Salt rejection in
nanofiltration for single and binary salt mixtures in view of sulphate removal.
Desalination, 171,205-215.
Malaysian CropLife and Public Health Associatidn (MCPA) (2005) [Online]. [Accessed on 2
June 2005]. Available from World Wide Web: http://www.mcpa.org.my/index.php
Noordin, M.Y., Venkatesh, V.C., Sharif, S., Elting, S. and Abdullah, A. (2004). Application of
response surface methodology in describing the performance of coated carbide tools
when turning AISI 1045 steel. Journal ofMaterials Processing Technology, 145, 46-58.
Nystrom, M., Kaipia, L. and Luque, S. (1995). Fouling and retention ofnanofiltration membranes.
Journal ofMembrane Science, 98, 249-262.
Ooi, B.S. (2005). Development of polymeric composite nanofiltration membrane: synthesis,
characterization and its evaluation on copper sulfate rejection. Ph.D thesis, Universiti
Sains Malaysia.
Plakas, KV., Karabelas, A.J., Wintgens, T. and Melin T. (2006). A study of selected herbicides
retention by nanofiltration membranes-The role of organic fouling, Journal of
Membrane Science, 284, 291-300.
Puasa, S.W. (2006). Micellar-enhanced ultrailftration for removal of reactive dyes from an
aqueous solution. M.Sc. thesis, Universiti Sains Malaysia.
Santos, J.L.C., Beukelaar, P., Vankelecom, I.F.l, Velizarov, S. and Crespo, J.G. (2006). Effect of
solute geometry and orientation on the rejection of uncharged compounds by
nanofiltration. Separation and Purification Technology, 50,122-131.
Schaep, l, Van der Bruggen, B., Vandecasteele, C. and Wilms, D. (1998). Influence of ion size
and charge in nanofiltration, Separation and Purification Technology, 14, 155-162.
Schwarzenbach, R.P., Gschwend, ;eM. and Imboden, D.M. (1993). Environmental Organic
Chemistry. New York: John Wiley & Sons, 194-200.
Spiegler, KS. and Kedem, O. (1966). Thermodynamics of hyperfiltration (reverse osmosis):
Criteria for efficient membranes. Desalination, 1, 311-326.I,
Syamal, M., De, S. and Bhattachrya, P.K (1997) Phenol solubilization by cetyl pyridinium
chloride micelles in micellar enhanced ultrafiltration. Journal ofMembrane Science. 137,
pp.99-107.
Thanuttamavong, M., Yamamoto, K, Oh, l I., Choo, K H. and Choi, S. J. (2002). Rejection
characteristics of organic and inorganic pollutants by ultra low-pressure nanofiltration of
surface water for drinking water treatment. Desalination, 160,257-264.
Van der Bruggen, B., Schaep, J., Maes, W., Wilms, D. and Vandecasteele, C. (1998).
Nanofiltration as treatment method for the removal of pesticides from ground waters.
Desalination, 117,139-147.
Van der Bruggen, B., Everaert, K, W., Wilms, D. and Vandecasteele, C. (2001) Application of
nanofiltration for removal of pesticides, nitrate and hardness from ground water: rejection
properties and economic evaluation. Journal ofMembrane Science, 193, 239-248.
Van Gauwbergen, D. and Baeyens, l (1998). Modelling reverse osmosis by irreversible
thermodynamics. Separation and Purification Technology, 13, 117-128.
Williams, M.E., Hestekin, lA., Smothers, C.N. and Bhattacharyya, D. (1999). Separation of
organic pollutants by reverse osmosis ana nanofiltration membranes: mathematical.models and experimental verification, Ind" Eng. Chem. Res., 38 (10), 3683 -3695.
Wong, S.S. (2006). Treatment ofpulp and paper mill wastewater by coagulation-jlocculation
method. M.Sc. thesis, Universiti Sains Malaysia.
Zhang, Y, Bruggen, B. V., Chen, G.X., Braeken, L. and Vandecasteele, C. (2004). Removal of
pesticides by nanofiltratio~: effect of the water matrix. Separation and Purification
Technology, 38,163-172.
,,
APPENDIX 3
Journal Publications
•
,..
ELSEVIER
Available online atwww.$ciencedirect.com..~",-.;- ScienceDirect
Journal of Hazardous Materials xxx (2007) xxx-xxx
Journal ofHazardousMaterials
www.elsevier.com/locate/jhazmat
Abstract
Dimethoate and atrazine retention from aqueoussolution by nanofiltration membranes
A.L. Ahmad *, L.S. Tan, S.R. Abd. ShukorSchool a/Chemical Engineering, Engineering Campus, Universiti Sains Malaysia,
14300 Nibong Tebal, Seberang Prai Selatan, Pulau Pinang, Malaysia
Received 14 December 2006; received in revised form 15 May 2007; accepted 16 May 2007
In order to produce sufficient food supply for the ever-increasing human population, pesticides usage is indispensable in the agriculture sectorto control crop losses. However, the effect of pesticides on the environment is very complex as undesirable transfers occur continually amongdifferent environmental sections. This eventually leads to contamination ofdrinking water source especially for rivers located near active agriculturepractices. This paper studied the application ofnanofiltration membrane in the removal of dimethoate and atrazine in aqueous solution. Dimethoatewas selected as the subject of study since it is being listed as one of the pesticides in guidelines for drinking water by World Health Organization.Nevertheless, data on effectiveness of dimethoate rejection using membranes has not been found so far. Meanwhile, atrazine is classified as one ofthe most commonly used pesticides in Malaysia. Separation was done usin&, a small batch-type membrane separation cell with integrated magneticstirrer while concentration of dimethoate and atrazine in aqueous solution was analyzed using high performance liquid chromatography (HPLC).Four nanofiltration membranes NF90, NF200, NF270 and DK were Tested for their respective performance to separate dimethoate and atrazine.Of all four membranes, NF90 showed the best performance in retention of dimethoate and atrazine in water.©2007 Elsevier B.V. All rights reserved.
rfeywords: Nanofiltration; Membrane technology; Pesticides; Dimethoate; Atrazine
1. Introduction
Malaysia is an active player in agriculture practice, plantng oil palm, paddy, fruit, vegetables and many other products'or local consumption and some for export purpl>ses. PesticidesIre also part and parcel of agriculture sector as' a mean of pest:ontrol for sustainability of the industry. Annuall~ sales figIre of approximately RM 300 million is recorded by Malaysian;ropLife and Public Health Association [1]. The huge amount oflesticides used is emerging as contaminants in water. This is noturprising because pesticides sprayed on crops can be drifted by~ind into nearby water source while pesticides applied directly) the soil can be washed off by rain into nearby surface waterodies or percolate through the soil to lower soil layers androundwater [2].
• Corresponding author. Tel.: +6045937788; fax: +6045941013.E-mail addresses:chlatif@eng.usm.my (A.L. Ahmad),
n.liansee@gmail.com(L.S. Tan),Isyamrizal@eng.usm.my (S.R.Abd. Shukor).
104-3894/$ - see front matter © 2007 Elsevier B.V. All rights reserved.li:1O.l0l6/j.jhazmat.2007.05.047
In this study, the pollutants selected were dimethoate andatrazine. Dimethoate is a type of organophosphorus insecticidethat has been identified as one of the chemicals from agriculture activities for which guideline value has been established byWorld Health Organization in the guidelines for drinking water[3]. In fact, its presence in water is not a surprise since it is highlysoluble in water and adsorbs very weakly to soil particles, thus,subjecting it to considerable leaching [2]. Although this woul~
normally cause minute concentration of pesticides presence inwater, its chronic effect to the livings has been of more concern.Doull [4] reported that dimethoate could cause oncogenicity,mutagenicity, fetotoxicity and reproductive effects. Meanwhile,the other pollutant studied is atrazine as it is among the mostc,ommonly used pesticides in Malaysia especially for its usagea~ herbicide in plantations. Although atrazine is considered to bea low toxic herbicide, extensive amount of its usage has rankedit among the most common pesticides found in surface waterand groundwater [5]. This situation has warranted urgent globalattention to abate theirpresence in drinking water. Recent reportshave revealed that high doses of atrazine induce abnormalitiesand deformities in non-target organisms. Furthermore, the syn-
2 A.L. Ahmad et at. / Journal ofHazardous Materials xxx (2007) xxx-xxx
,.
Nomenclature
A membrane areaCf concentration of feedCp concentration of permeatebt time differenceb V cumulative volume differenceKow octanollwater partition coefficientLv membrane permeabilityR percentage of pesticide rejectionVw permeate flux
ergy effect of dimethoate-atrazine is more lethal than the effectof the individual pesticide since the toxicity of dimethoate wasenhanced significantly when they are in binary combination [6].
Traditionally, removal of pesticides for the production ofdrinking water was done by activated carbon filtration. It waseffective, but expensive and required frequent regeneration [7].Over the past few years, nanofiltration membranes have beenstudied as potentially useful means of pesticide removal considering the fact that the molecular weights (MWt) of mostpesticides are more than 200 Da [2,5].
Nanofiltration has been successfully applied in drinkingwater treatment plant in Mery-sur-Oise, France [8], Leiduin [9].and Heemskerk [10] in Holland as well as Saffron Walden inEngland [11]. However, there is still a long list of pesticidesin guidelines for drinking water by World Health Organization[3] but lack of data for their effective separation using membrane, including dimethoate. Therefore, there are still room forthe investigation ofthe feasibility ofusing membrane technologyto remove dimethoate from water, with addition to observationfor binary mixture of dimethoate-atrazine.
Thus, the objective of this study is to examine the performance of nanofiltration membranes to retaindimethoate andatrazine in aqueous solution. Four nanofiltrati&n membraneswere subjected to stirred dead-end filtration. The effect of feedconcentration and operating pressure on the permeafe flux andfeed-based rejection of dimethoate and atrazine were investigated.
Table 1Properties of dimethoate and atrazine [2]
2. Materials and methods
2.1. Pesticides
Dimethoate with 99.8% purity and atrazine with 97.4% puritywere purchased from Riedel-de Haen (Germany). The molecularstructures of both pesticides are presented in Table 1.
2.2. Membranes
Three types of nanofiltration membranes provided byDowlFilmtec (USA) and another type of nanofiltrationmembrane purchased from Osmonics (USA) were used inthis experiment. The thin film polyamide membranes fromDowlFilmtech used were NF90, NF200 and NF270, while thethin film polyamide membrane from Osmonics used was DK.Polyamide membranes were used in this study because they wereable to achieve good pesticides retention [12,13]. Table 2 provides the specification of the membranes used as given by themanufacturers.
2.3. Membrane stirred cell
A 300-mL stirred cell (Sterlitech), model Sterlitech™HP4750, USA, was used to conduct the dead-end filtrationexperiments. The membrane diameter was chosen to be 0.049 mwith effective membrane area ofl,46 x 10-3 m2. The maximumoperating pressure for this cell was 69 x 10+5 Pa.
2.4. Experimental setup and procedure
Dead-end filtration experiments were carried out with thestirred cell (Sterlitech™ HP4750). The pesticide solution inthe cell was stirred by a Teflon-coated magnetic bar. The cellwas pressurized using compressed high purity nitrogen gas. Thepressure in the permeate side was approximately atmosphericunder all condition. The transmembrane pressures used duringexperiments were 6 and 12 x 10+5 Pa. The concentration of pesticide was set to be at 2 and 20 mglL. This concentration washigher than the usual concentration found in the case of run-offdue to consideration of the membrane in case of accidental spillofpesticides in water source. The stirring speed was set constantat 1000 rpm.
AtrazinePesticide
Chemical structure
Molecular weight (Da)Solubility in waterLog octanol/water partition coefficient, Kow
" [16].
Dimethoate
s \II _/'/~
/0-P1-:S- II CH.3CH 0
3 0/
CH:J229.2825 gIL at 21°C0.70
TN, N--l~~~N~
H3C H N H CH3
215.6920 mglL at 20°C2.61"
AL. Ahmad et al. / Journal ofHazardous Materials xxx (2007) xxx-xxx 3
Table2Specification of membrane used
Membrane NF90 NF200 NF270 DK
ManufacturerMaterialContact angle (0)"pure water permeabilityb (m3{(m2 sPa))Maximum operating pressure (Pa)Maximum operating temperature (OC)
pH range
DowlFilmtecPolyamide
1.90 X 10-11
41 x 10+5
453-10
DowlFilmtecPolyamide26±21.17 x 10-11
41 x 10+5
453-10
DowlFilmtecPolyamide28±23.20 x 10-11
41 x 10+5
453-10
OsmonicsPolyamide
7.84 X 10-12
40 X 10+5
383-10
" [5].b Our measurements.
Fig. 1. Schematic diagram of experimental setup.
where ~V is the cumulative volume difference (m3), ~t is thetime difference (s) andA is the membrane area (m2), respectively.
All experiments were conducted at room temperature(25 ± 2 QC). A schematic diagram of the expertnental setup isshown in Fig. I. .
(2)
2.5. Analytical method
Concentration of dimethoate and atrazine in feed and permeate was analysed using high performance liquid chromatography(HPLC) by Perkin Elmer (USA). The HPLC column usedwas Zorbax SB-CN (5fJo, 4.6mm Ld. xl50mm long, AgilentTechnologies). The mobile phase was a mixture of 35% acetonitrile and 65% deionized water while the flow rate was set at1.0 mL/min. The UV detector was operated at a wavelength of200 nm. The peak for dimethoate was detected at around 3.5 minwhile the peak for atrazine was detected at around 5.3 min. Percentage of rejection was obtained with the following equation:
where R is the percentage of pesticide rejection, Cp is the concentration of permeate (mglL) and Cf is the concentration offeed (mglL)
3.1. Retention ofdimethoate and atrazine
3. Results and discussion
The retention performance of dimethoate and atrazine byNF90, NF200, NF270 and DK at different pressure and concentration is presented in Figs. 2 and 3, respectively. From thesefigures, it is obvious that the retention of both dimethoate andatrazine tend to be better when the pressure was increased from6 to 12 x 10+5 Pa. It could be seen that NF90 produced the bestretention performance for the operating pressure and feed concentration tested, at approximately 85% for dimethoate and morethan 95% retention for atrazine. The performance of DK wasthe second highest of all four membranes tested while NF200showed slightly lower retention than DK when both were oper~ted at the same pressure and feed concentration. NF270 showedthe lowest rejection performance out of the four membranestested, especially for dimethoate retention. Higher retention wasobserved at higher pressure due to the increased water flux. Theconcentration of permeate became diluted with the increasedwater flux as the solute molecule was rejected by molecularsieving effect.
Meanwhile, the concentration effect was less significant onrejection of dimethoate and atrazine as compared to the effectof pressure as there was only slight increment of rejection per-
R = (1 - Cp) x 100%
" Cf
(1)
Pel'lllc.tc
MombrlUloStirred Coil
SmilllessSteolFecd.Tollk
~@ :Pressuroregu!a1or
~ :V.t"e
The cell contained a nanofiltration membrane with an effective area of 1.46 x 10-3 m2. The membrane was immersed for24 h in deionized water before being used in any experimentalwork. Membrane permeability was determined by initially filtering it using deionized water at 12 x 10+5 Pa for approximately8 h for compaction to avoid compression effect in the later stageof experiment. Then, stabilized water flux at different operating pressures was obtained and membrane permeability values(Lp) could be determined from the slope of flux against pressuregraph.
For separation process, the same compaction process wascarried out before the test cell was emptied and 1.8 L of feedsolution was filled into the test cell and solution reservoir. Thecell was then pressurized at the operating pressure indicated bya pressure regulator. Permeate from the bottom of the cell wascollected and its weight was measured with an electronic balance·of ±0.01 g accuracy. The cumulative weight were converted tocumulative volume and the permeate flux could be obtained.Permeate flux, Vw (m3/m2 s), was obtained using Eq. (1):
~VVw=--
~t· A
A.L. Ahmad et al. / Journal ofHazardous Materials xxx (2007) xxx-xxx
100 100(b)
80 80c c0
~ 60 0 60(I) ~'iii' (I)
0:: 'iii'40
?fl-40 0::
?fl-
20 20
0 06 12 6 12
Pressure x 10+5 (Pa) Pressure x 10+5 (Pa)
100 100(c) (d)
80 80
c c0 0 60:;:l 60 :;:lu u(I) (I)'iii' 'l0:: 40 40?fl- ?fl-
20 20
0 06 12 6 12
Pressure x 10+5 (Pa) Pressure x 10.5 (Pa)
1m2 mg/L lEI 20 mg/L IFig. 2. Rejection of dimethoate by Nf'iJQ (a), NF200 (b), NF270 (c), and DK (d).
100 100(b)
80 80
c c0 60 0 60~
:;:lu
(I) (I)
'iii'40
'iii'0:: 0:: 40?fl- ::!!0
20 20
0 06 12 6 12
Pressure x 10·f (pa) Pressure x 10+5 (Pa)
100 100(c) (d)
80 80
c c0 60 0 60:;:l :;:lu u(I) (I)
'l..,..
40(I)'0::' 40
::!! ?fl-o
20 20
0 06 12 6 12
Pressure x 10+5 (Pa) Pressure x 10+5 (Pa)
1m2 mg/L 1SlI20 mg/L IFig. 3. Rejection of atrazine by NF90 (a), NF200 (b), NF270 (c), and DK (d).
A.L. Ahmad et al. / Journal ofHazardous Materials xxx (2007) xxx-xxx 5
formance although the concentration was increased 10 times ascompared to two times increment of pressure. This finding is inagreement with work done by Causserand et a1. [13] and Zhanget a1. [14]. This shows that in practical terms, the membraneshave almost the same efficiency level for dimethoate rejectioneven though the feed concentration varies as much as 10 timesfrom time to time. However, atrazine retention performance ofNF200 and NF270 of around 80% was obtained by Plakas eta1. [5] when the concentration of atrazine was between 0.150and 0.300 mg/L. This suggests that while effect of concentration did not pose much impact if compared to effect of pressure,it was still a valid gradient for transport of solute through membrane. NF90 was found to be a more robust membrane in viewof atrazine retention since its retention was almost equal even atsuch high concentration of 2-20 mg/L.
Overall, all four membranes tested showed better retentionfor atrazine than dimethoate although dimethoate has slightlyhigher molecular weight than atrazine. Several rcports [15-18]suggested that although molecular sieving effect must not beneglected, hydrophobicity of the solutes played a very importantrole in determining the retention performance by membrane.The higher the value of log Kow , the better the rejection wouldbe. This behaviour was shown in this study since atrazine has
higher hydrophobicity than dimethoate. Moreover, dimethoatehas aliphatic molecular structure compared to the heterocyclicaromatic structure of atrazine. Kiso et aI. [16] reported that nonphenylic pesticides were rejected at a relatively lower degreethan phenylic pesticides.
3.2. Permeate flux performance
Figs. 4 and 5 show the flux performance of the membranes fordimethoate and atrazine retention, respectively. Based on thesefigures, it was obvious that the increase in pressure had significant effect on permeate flux for both dimethoate and atrazineretention tests. All membranes tested experienced approximately double increment of permeate flux when the operatingpressure was doubled from 6 to 12 x 10+5 Pa. This shows thatpermeate flux increment corresponded linearly to the pressureapplied to the solution. Meanwhile, concentration of feed hadvery little effect on the permeate flux as compared to operating pressure. It showed no significant trend although it causedslightly lower permeate flux when it was increased for certainmembrane especially at P= 12 x 10+5 Pa run. Thus, effect ofconcentration can be excluded from consideration when it comesto flux performance.
4.00 4.00(a) (b)
3.50 3.50
ur 3.00 ur 3.00N N
E E 2.50;0- 2.50 ;0-
.§. .§."I 2.00 "I 2.000 0.... ....
1.50>< 1.50 ><>< ><:::I :::I 1.00u: 1.00 u:
0.50 0.50
0.00 0.00
6 12 6 12
Pressur~ x 10+5 (Pa) Pressure x 10+5 (Pa)
4.00 4.00(d)
3.50 3.50
ur 3.00 ur 3.00N N
E E;0- 2.50
;0- 2.50.§. .§."I 2.00 "I 2.000 0.... ....>< 1.50 >< 1.50>< ~ )(:::I ':::I
u: 1.00 u: 1.00
0.50 0.50
0.00 0.006 12 6 12
Pressure x 10+5 (Pa) Pressure x 10+5 (Pa)
1m2 mg/L 12120 mg/L IFig. 4. Flux performance on dimethoate by NF90 (a), NF200 (b), NF270 (c), and DK (d).
A.L. Ahmad et al. / Journal ofHazardous Materials xxx (2007) xxx-xxx
4.00 4.00(a) (b)
3.50 3.50
U! 3.00 U! 3.00N N
E E... 2.50 ... 2.50.§. .§.'7 2.00 '7 2.00Cl Cl.... ....>C 1.50 >C 1.50>C >C:::l :::l
u::: 1.00 u::: 1.00
0.50 0.50
0.00 0.006 12 6 12
Pressure x 10+5 (Pa) Pressure x 10+5 (Pa)
4.00 4.00(d)
3.50 3.50
U! 3.00 U! 3.00N N
E E... 2.50 ... 2.50.§. .§.'7 2.00 '7 2.00Cl Cl.... ....>C 1.50 >C 1.50>C >C:::l :::l
u::: 1.00 u::: 1.00
0.50 0.50
0.00 0.006 12 6 12
Pressure x 10+5 (Pa) Pressure x 10+5 (Pa)
11m2 mg/L liS 20 mg/L IFig. 5. Flux perfonnance on atrazine by NF90 (a), NF200 (b), NF270 (c), and DK (d).
NF270 produced the highest permeate flux for all conditionstested. This was especially obvious at operating pressure of12 x 10+5 Pa. NF90 showed the second highestpermeate flux outof the four membranes with approximately 40% IO'fer comparedto permeate flux by NF270. Meanwhile, NF200 snowed considerably low flux rate compared to NF270 while DK,.producedthe lowest permeate flux performance as it has approximately300% lower flux compared to NF270. Based on the publisheddata, NF270 had average pore size of 0.71 nm, NF90 had average pore size 0.55 nm while NF200 had average pore sizeof 0.38 nm [19,20]. Hence, the results obtained in this studycorresponded to the average pore size reported in the literature.
However, this also showed that while 0.55 nm average poresize for NF90 was sufficient to retain dimethoate and atrazinewith high percentage of rejection, solute-membrane interactionfactor was also important [15,18,21] as DK and NF200 couldnot sustain as much rejection as NF90 although DK had similaraverage pore size with NF90 [22] while NF200 had smalleraverage pore size. The interaction between membrane materialfor DK and pesticides tested was believed to contribute to thecrossing of solutes through the membrane because it had lowerpercentage of retention of pesticides compared to NF90. This
validated the claim by the manufacturer that NF90 is suitablefor pesticides and herbicides removal [23].
3.3. Retention performance ofNF90 foratrazine-dimethoate
Since NF90 showed good rejection for both atrazine anddimethoate individually, rejection of atrazine-dimethoate wastested at pressure of 6 x 10+5 Pa and stirring rate of 1000 rpmto examine if the membrane would have the same goodperformance when the two pesticides co-exist. The ratio ofatrazine:dimethoate was set at 20:80,50:50 and 80:20 for a totalof 10 mg/L pesticides. Fig. 6 shows that NF90 still maintained itsgolild performance of retention for both atrazine and dimethoatein the presence of binary mixture of pesticides, although therewas slightly lower retention observed compared to the singlesolute condition. This observation was in line with observation by Plakas et al. [5] which suggested that simultaneousfiltration of more than one pesticide resulted in a kind of competitive adsorption on the membrane surface and, thus, created agreater passage to the permeate side. These results were also inagreement with the report by Kiso et al. [16] which found thatherbicides displaying higher rejection in single solute solutions
Fig. 6. Rejection performance of atrazine-dimethoate for NF90.
References
7
[2] M.A. Kamrin, Pesticide Profiles: Toxicity, Environmental Impact and Fate,CRC Press, Boca Raton, 1997.
[3] World Health Organization. Guidelines for drinking-water qualit:'2004, third ed. http://www.who.int/water_sanitation..health/dwq/gdwq3/enlindex.html. Accessed on 16 June 2005.
[4] J. Doull, Pesticide Carcinogenicity, in: N.N. Ragsdale, R.E. Menzer (Eds.),Carcinogenicity and Pesticides: Principles, Issues and Relationship, American Chemical Society, Washington DC, 1989, pp. 1-5.
[5] K.V. Plakas, A.J. Karabelas, T. Wintgens, T. Melin, A study of selected herbicides retention by nanofiltration membranes-the role of organic fouling,J. Membr. Sci. 284 (2006) 291-300.
[6] T.D. Anderson, KY. Zhu, Synergistic and antagonistic effects of atrazineon the toxicity of organophosphorodithioate and organophosphorothioate insecticides to Chironomustentans (Diptera: Chironomidae), Pestic.Biochem. Physiol. 80 (2004) 54-64.
[7] B. van der Bruggen, J. Schaep, W. Maes, D. Wilms, C. Vandecasteele,Nanofiltration as treatment method for the removal of pesticides fromground waters, Desalination 117 (1998) 139-147.
[8] B. Cyna, G. Chagneaub, G. Bablon, N. Tanghe, Two years ofnanofiltrationat the Mery-sur-Oise plant, France, Desalination 147 (2002) 69-75.
[9] P.A.C. Bonne, E.F. Beerendonk, J.P. van der Hoek, J.A.M.H. Hofman,Retention of herbicides and pesticides in relation to aging RO membranes,Desalination 132 (2000) 189-193.
[10] J.A.M.H. Hofman, E.F. Beerendonk, H.C. Folmer, J.C. Kruithof, Removalof pesticides and other micropollutants with cellulose acetate, polyamideand ultra-low pressure reverse osmosis membranes, Desalination 113(1997) 209-214.
[11] E. Wittmann, P. Cote, C. Medici, J. Leech, A.G. Turner, Treatment of ahard borehole water containing low levels of pesticide by nanofiltration,Desalination 119 (1998) 347-352.
[1'2] S.S. Chen, S.T. James, L.A. Mulford, C.D. Norris, Influences of molecular weight, molecular size, flux, and recovery for aromatic pesticideremoval by nanofiltration membranes, Desalination 160 (2004) 103111.
[13] C. Causserand, P. Aimar, J.P. Cravedi, E. Singlande, Dichloroanilineretention by nanofiltration membranes, Water Res. 39 (2005) 15941600.
[14] Y. Zhang, B. van der Bruggen, G'x. Chen, L. Braeken, C. Vandecasteele,Removal of pesticides by nanofiltration: effect of the water matrix, Sep.Purlf. Techno!. 38 (2004) 163-172.
[15] C. Bellona, J.E. Drewes, P. Xu, G. Amy, Factors affecting the rejection oforganic solutes during NF/RO treatment - a literature review, Water Res.38 (2004) 2795-2809.
[16] Y. Kiso, Y. Nishimura, T. Kitao, K Nishimura, Rejection properties ofnon-phenylic pesticides with nanofiltration membranes, J. Membr. Sci. 171(2000) 229-237.
[17] Y. Kiso, Y. Sugiura, T. Kitao, K. Nishimura, Effects of hydrophobicityand molecular size on rejection of aromatic pesticides with nanofiltrationmembranes, J. Membr. Sci. 192 (2001) 1-10.
[18] K Kosutic, L. Furac, L. Sipos, B. Kunst, Removal of arsenic and pesticidesfrom drinking water by nanofiltration membranes, Sep. Purif. Techno!. 42(2005) 137-144.
[19] N. Hilal, H. AI-Zoubi, N.A. Darwish, A.W. Mohammed, Characterisationof nanofiltration membranes using atomic force microscopy, Desalination177 (2005) 187-199.
[2P] X. Lefebvre, J. Palmeri, J. Sandeaux, R. Sandeaux, P. David, B. Maleyre,, C. Guizard, P. Amblard, J.-F. Diaz, B. Lamaze, Nanofiltration modeling: a
comparative study of the salt filtration performance of a charged ceramicmembrane and an organic nanofilter using the computer simulation programNANOFLVX, Sep. Purif. Techno!. 32 (2003) 117-126.
[21] T.-V. Kim, G. Amy, J.E. Drewes, Rejection of trace organic compounds byhigh-pressure membranes, Water Sci. Techno!. 51 (2005) 335-344.
[22] J.L.C. Santos, P. Beukelaar, I.F.J. Vankelecom, S. Velizarov, J.G. Crespo,Effect of solute geometry and orientation on the rejection of unchargedcompounds by nanofiltration, Sep. Purif. Techno!. 50 (2006) 122-131.
[23] The Dow Chemical Company, nanofiltration products, Filmtec1M membranes, http://www.dow.com/liquidseps/prod/app..nano.htm. Accessed on20 November 2006.
10080
-A.L. Ahmad et al. / Journal ofHazardous Materials xxx (2007) xxx-xxx
40 60% Atrazlne
20
100
80
i 60
~ 40
~ 20
00
Authors would like to thank Universiti Sains Malaysia forfunding this research with short-term grant (Account 6035167).Appreciation also goes to DowlFilmtec for providing the membranes.
4. Conclusion
Acknowledgements
[I] Malaysian CropLife and Public Health Association. http://www.mcpa.org.my/index.php. Accessed on 2 June 2006.
The performance of nanofiltration membrane to retaindimethoate and atrazine in aqueous solution was examinedin this study. Four nanofiltration membranes, NF90, NF200,NF270 and DK, which have molecular weight cut-off of around200 were subjected to stirred dead-end filtration and the effect offeed concentration and operating pressure on the permeate fluxand feed-based rejection of dimethoate was investigated. It was· •found that increasing the transmembrane pressure posed positiveeffect on dimethoate and atrazine rejection and permeate flux.However, effect of feed concentration had little significance onthe performance of the membranes tested.
NF90 showed the best retention performance while NF270showed the highest permeate flux out of the four membranestested. However, good retention quality should be the primaryproperty in choosing the appropriate nanofiltration membranefor application in pesticides treatment from water. Therefore,despite its high permeate flux, NF270 is not suitable especiallyfor dimethoate retention as it showed the poorest Jetention quaIity. NF90 is deemed the more suitable nanofiltration membranefor dimethoate and atrazine retention from aqueo~ solutionsince, it showed the highest retention of dimethoate and atrazinecoupled with considerably good permeate flux. Furthermore,although there was slight reduction of retention performance forNF90 in binary atrazine-dimethoate solution, it stilI managed tomaintain its robust retention performance.
may permeate more in mixed solute systems. However, resultsobtained in this study showed that while there was slight reduction of pesticides retention, the performance of NF90 was stilIcommendable even though it was in mixed solute system.
ELSEVIER
Available online at www.sciencedirect.com...-""'@;' ScienceDirect
Journal of Hazardous Materials xxx (2007) xxx-xxx
Journal ofHazardousMaterials
www.elsevier.com/locate/jhazmat
ysia, 14300 Nibong Tebal,
pted 23 October 2007
atmosphere and water. The presence of pesticides in water hasbeen reported by previous researchers [5-9]. Low-level residue.;of pesticides in water generally may not present acute toxicityproblems, but chronic effects will likely be ofconcern [10]. Thisis because pesticides could have chronic effects such as cancer[11-13], reproductive effects, fetal damage, delayed neurologicmanifestations and possible immunologic disorders [12].
In view ofthis scenario, many studies on separation of pesticides using nanofiltration membranes have been done in recentyears. Size exclusion by a nanofiltration membrane is recognized to be the main retention mechanism for pesticides. Otherparameters such as hydrophobicity, dipole moment, polarity andcharge of a molecule have also been found to influence the rejection performance [14-18]. On the other hand, according to Chenet al. [19], rejection of pesticides was dependent on operationalflux and recovery as well. For a particular pesticide in the tw...oPerational fluxes and recoveries, the highest percent rejectionoccurred at high flux and low recovery, and the lowest percentrejection occurred at low flux and high recovery. Meanwhile, astudy done by Zhang et al. [20] found that pore narrowing byion adsorption and water matrix influenced rejections.
So far, not much attention has been given to the changesin nanofiltration performance during nanofiltration of pesti-
plext envi
air mayan effec
away fromay locations
to til; soil mayce water or per
groundwater [2].tended to urban
that the movementy a function of water
sorption capacities of
esticides is, it will evenreat to human's health via
37788; fax: +6045941013..my (A.L. Ahmad),amrizal@eng.usm.my
The role of pH in nanofiltration of atrazine anddimethoate from aqueous solution
A.L. Ahmad*, L.S. Tan, S.R. ASchool ofChemical Engineering, Engineering Campus, Universiti
Seberang Prai Selatan, Pulau Pinan
Received 9 July 2007; received in revised form 23 Octo
Keywords: Nanofiltration; Membrane teclmology; Pesticides; Dime
The effect of pesticides on the environment ias undesirable transfers occur continually amoronmental sections. Pesticides that are spraeventually end up in soils or water. The attive medium which can move airbornetheir application sites and redeposit the[1]. On the other hand, pesticides appliebe washed ofI by rain into nearby bodiGolate through the soil to lower soil IPesticides uses and transfers haveized catchments [3]. However, itofpesticide in and through the soisolubility of the pesticides anthe soil type [4].
No matter where the app .tually end up becoming a p
Abstract
This study examined the performance of nanofiltration membranes to ine and dimethoate in aqueous solution under different pHconditions. Four nanofiltration membranes, NF90, NF200, NF270 and D ted to be examined. The operating pressure, feed pesticide andstirring rate were kept constant at 6 x 105 Pa, 10mglL and 1000 rpm. It t increasing the solution's pH increased atrazine and dimethoaterejection but reduced the permeate flux perfOlmance for NF200, NF K. However, NF90 showed somewhat consistent performance inboth rejection and permeate flux regardless of the solution's pH. N :1i!;tained above 90% of atrazine rejection and approximately 80% ofdimethoate rejection regardless of the changes in solution's pH. Thus NF90 is deemed the more suitable nanofiltration membrane for atrazine anddimethoate retention from aqueous solution compared to NF200, d DK.© 2007 Published by Elsevier B.V.
1. Introduction
-• COlTesponding author. TelE-mail addresses: chla .
lan.liansee@gmail.com(L.S.(S,R.Abd. Shukor).
Q1
I 0304-3894/$ _ see front matter © 2007 Published by Elsevier B.V.I doi:l0.l0l6/j.jhazmat.2007.l0.073
z A.L. Ahmad et al. / Journal ofHazardous Materials xxx (2007) xxx-xxx
215.6920mgIL @ 20 Ge1.7b
2.61c
5. Experimental set-up and procedure
2.2. pH adjustment
Atrazine
2.3. Membranes
The chemicals used to adjust the pH of the pesticide solutions for filtration experiments were hydrochloric acid, HC137%(w/w) and sodium hydroxide, NaOH (l M). These chemicalswere obtained from Merck.
Three types of nanofiltration membranes provided byDowlFilmtec (IJSA) and one type of nanofiltration membranepurchased frQ1h Water Technologies (USA) with molecular weight cttf¥· (MWCO) of around 200 Da were used inthis expe . e thin film polyamide membranes fromDowlFil were NF90, NF200 and NF270 while thethin fil e membrane from GE Water Technologiesused Table 2 provides the specification of the mem-bran given by the manufacturers.
hnL stirred cell (Sterlitech), model Sterlitech™50, USA, was used to conduct the dead-end filtration·ments. The effective membrane area is 1.46 x 10-3 m2.
The maximum operating pressure for this cell was 69 x 105 Pa.
Dead-end filtration experiments were carried out with thestirred cell (Sterlitech™ HP4750). The pesticide solution in thecell was stirred by a Teflon-coated magnetic bar. The cell waspressurized using compressed high purity nitrogen gas. Thepressure in the permeate side was approximately atmospheric underall condition. The pesticides solution, prepared using deionizedwater, was adjusted to different initial pH by adding 1M NaOHor 37% (w/w) HCI. The pH measurement was conducted usingpH meter (Mettler Toledo Delta 320 pH Meter). The operating
H
S ~~o-~-s I'"'-CH3
~ I 'H3C 0 0
/H3C
229.2825 gIL @ 21 °e2.0'0.70
Nomenclature
A membrane areaCf concentration of feedCp concentration of permeateKow octanollwater partition coefficientLp membrane permeabilitypKa acid disassociation constantR percentage of pesticide rejection6.t time differenceVw permeate fluxb.V cumulative volume difference
Pesticide
Dimethoate with 99.8% purity and atrazine wiwere purchased from Riedel-de Haen (Germanstructures of both pesticides are presented in
2. Materials and methods
cides in aqueous solution when there are changes in its pH.However, this factor must not be neglected as the role of pHis also important in determining the stability of membrane[21,22]. Therefore, the objective of this study is to investigatethe performance of nanofiltration membranes to retain atrazineand dimethoate in aqueous solution under different pH conditions. The effect of initial solution's pH for pesticide rejectionand permeate flux were obtained and examined. This study isa continuation from a previous study which focused on theeffect feed concentration and operating pressure on the permate flux and rejection of dimethoate and atrazine from aquesolution [23].
2.1. Pesticides
Table IProperties of dimethoate and atrazine [2]
Chemical structureMolecular weight (Da)SolUbility in waterAcid disassociation constant, pLogKow-a [30].
b [31].c [32].-
A.L. Ahmad et al. / Journal ofHazardous Materials xxx (2007) xxx-xxx 3
Table 2.Specification of membrane used
Membrane
ManufacturerMaterialContact angle" CO)Surface charge (pH 7)pure water permeabilitye (m3{(m2 sPa»Maximum operating pressure (Pa)Maximum operating temperature (DC)
pH range
NF90 NF200 NF270 DK
DowlFilmtec DowlFilmtec DowlFilmtec OsmonicsPolyamide Polyamide Polyamide Polyamide
26±2 28±2Negativeb Negativeb NegativeC Negatived1.90 x 10-11 1.17 x 10-11 3.20 X 10-11 7.84 X 10-12
41 X 105 41 X 105 41 X 105 40 x lOS45 45 45 383-10 3-10 3-10 3-10
Concentration of atr dimethoate in feed and perme-ate was analysed usin ormance liquid chromatography(HPLC) by Perkin-Elmer SA). The HPLC column usedwas Zorbax SB-CN (5/-L, 4.6 mm i.d. x 150 mm long, AgilentTechnologies). The mobile phase was a mixture of 35% acetonitrile and 65% deionized water while the flow rate was set at
II
(2)
AnalyticalBalance
Pornlcate
MembraneStirred Cell
Magnetic Stirrer
StainlessSteelFeed Tank
Fig. 1. Diagram of experimental set-up.
~® :Pressureregullltor
~ :VllIve
where R is the percentage of pesticide rejection, Cp is the concentration of permeate (mg/L) and Cf is the concentration offeed (mg/L)
1.0 mLimin. The UV detector was operated at a wavelength of200 nm. The peak for dimethoate was detected at around 3.5 minwhile the peak for atrazine was detected at around 5.3 min. Percentage of rejection was obtained with the following equation:
R = (1 - ~:) x 100%
3. Results and discussion;,
The effect of initial solution's pH on the atrazine anddimethoate rejection at fixed operating pressure, pesticide concentration and stirring rate are presented in Figs. 2 and 3. Fromthe figures, it can be seen that the rejection performance foratrazine and dimethoate by NF200, NF270 and DK increasedas the pH was increased while the rejection trend for NF90 wasalmost constant regardless of the pH condition. The percent-
3.1. Rejection ofatrazine and dimethoate
(1)
ce (m3), ~t is the(m2), respectively.for four times and
les were used as thee conducted at room
diagram of the experi-
" [331.b [341.C [351.d [361e Our measurements.
~Vv --
w - ~tA
where ~V is the cumulative volumetime difference (s) andA is the mem
Samples were collected at evthe average values obtained fromresults in this work. All expe 'etemperature (25 ± 2 QC). A smental set up is shown in Fi
2.6. Analytical method
pressure, feed pesticide and stirring rate were kept constant at6 x 105 pa, lOmg/L and 1000rpm while the initial solution'spH were varied at 4, 7 and 9.
The cell contained a nanofiltration membrane with an effective area of 1.46 x 10-3 m2. The membrane was immersed for24 h in deionized water before being used in any experimentalwork. Membrane permeability was determined by initially filtering it using deionized water at 12 x 105 Pa for approximatel8h for compaction to avoid compression effect in the later stageof experiment. Then, stabilized water flux at different oping pressures was obtained and membrane permeability(Lp) could be determined from the slope of flux against pgraph.
For separation process, the same compaction prcarried out before the test cell was emptied and 1.solution was filled into the test cell and solution rcell was then pressurized at the operating pressura pressure regulator. Permeate from the bottomcollected and its weight was measured with an eof ±0.01 g accuracy. The cumulative weightcumulative volume and the permeate fluxPermeate flux, Vw (m3/m2 s), was obtain
two membranes are slightly different although with the samepolyamide thin-film composite. NF270 has a very thin semiaromatic piperazine-based polyamide active layer while NF90consists of a fully aromatic polyamide active layer [24]. Thisslight difference of membrane structures can be one of reasonsthat NF90 showed superior rejection characteristics comparedto other nanofiltration membranes tested at the experimentalconditions.
Puasa [25] reported that polyamide thin-film composite membranes have charge characteristics that influence the separationcapabilities, wbich can be altered by the solution's pH and itwas reported~n e isoelectric point of polyamide membraneis generally o~t n 4 and 5. The occurrence of an isoelectricpoint me er pH than the isoelectric point, the mem-brane is charged and vice-versa. Hence, in the caseof polym ranes, surface membrane charge is typicallynegati pH values, it increases as the pH decreases andswit itive values at low pH's [26].
Ho contrary to the usual phenomenon which occursfo ecies whereby at isoelectric point, the flux is usu-
ighest while the rejection is at the lowest [27], theserved for the uncharged pesticides molecules is some
t different. In the case of uncharged molecules, instead ofinfluenced by the changes in membrane surface charge,
it is elieved that it was the changes of the membrane struc-es and/or formation of high molar mass complexes which
ificantly affected the performance of solute rejection andrmeate flux [28]. Nevertheless, the possibility of formation
of high molar mass complexes at high pH is sidelined in thisresearch since the rejection of atrazine and dimethoate onlyincreased at high pH for NF200, NF270 and DK while NF90showed a slight decrease of rejection at high pH.
Hence, it is deduced that the trend of atrazine and dimethoaterejection obtained for NF200, NF270 and DK in this experimentwas due to the changes of the membrane structures caused bythe solution's pH. The results obtained were in agreement withobservation done by Freger et al. [29] whereby the rejectionof lactate decreased with the decrease of pH. In another workby Freger et al. [22], it was concluded that at low pH, acidichydrolysis disrupted the chemical bonds in the membrane polymer matrix. This condition reduced the degree of crosslinking(i.e., rigidity) of the polymer matrix which eventually caused thedecrease of rejection. At the same time, acidic hydrolysis alsocaused the increase of the hydrophilic sites at the membrane[22]. The increase of hydrophilic sites would cause the increaseof permeate flux. On the other hand, the increase of atrazine anddimethoate rejection at high pH observed for NF200, NF270 and
\ DK could be caused by the hydration swelling of the membrane, skin layer [29]. This could result in shrinking of membrane pore
size, and thus, reduced the permeation of solute through thepores of the membrane. Meanwhile, it is believed that NF90was rather chemical-resistant as it showed somewhat consistent performance regardless of the solution's pH. There wasonly a drop of about 3% of rejection perfOlmance for NF90compared to the obvious increase or reduction of rejection performance shown by the rest of the nanofiltration membranestested.
ionor clearer
er.Jtjectedefmolec
iscussed inhydrophoof atrazine
10
- .... ·NF
-·.··N-w- NF270__ OK
10
- ..... NF90··.··NF200-w- NF270__ OK
9
9
to the alteration of initial
8
AL. Ahmad et al. / Journal ofHazardous Materials xxx (2007) xxx-xxx
7
pH
pH
6
85
5
...,.!.',.'! .....
.................... ,.. ,.........
4
4
i- ._.. _._- -' _.-. - .... -. -'- -.. - 'i
.-----_..-._._'-.'-'-'-'--§
100
90
80
70c:0g 60
'ii'a: 50'IJl
40
30
20
103
4
100
90
80
c: 700
! 60'aYa: 50.,.
40
30
20
103
Dimethoate
pH4 pH9
NF90 2.71 3.43 4.33NF200 24.55 22.35 24.11NF270 25.12 41.59 26.04DK 8.67 7.65 17.73
Membrane
Table 3Percentage of changes in rejection perfsolution's pH
Fig. 3. Effect of initial solution's pH on rejection of dimet
Fig. 2. Effect of initial solution's pH on rejection of atrazine.
age of changes in rejection performance due to tof initial solution's pH is also presented is Tascrutiny. It is noted that atrazine was consistenthan dimethoate although dimethoate has sligular weight than atrazine. This behaviour hour previous work whereby it was due to tbicity (log Kow) and heterocyclic aroma .[23].
Meanwhile, Nghiemetal. [24]repoNF90 is smaller than that of NF270
A.L. Ahmad et al. / Journal ofHazardous Materials xxx (2007) xxx-xxx 5
2.50E'05
2.00E·05
•.. ' 1.50E-05.§"ei 1.00E·05
ii:
P .... _ .... _ .... _ ... _ ........ - ..
.. -....-..- ..... NF90··.··NF200...... - NF270
... ·_·_·_·_·_·_·.. ·_· ,·_·-·~OK..,. , .
2;50E·05 i-..-..............-.. ..~ .... .... .._.._..-.
•· ... ··-·-·_·_·-.. ·_·... ·_·-ll.................................... - .... ·NF90··.··NF200... '" -NF270~DK
5.00E.06Ie II 5.00E·OS )(
3.2. Permeate flux peiformance
Fig. 4. Effect of initial solution's pH on permeate flux during rejection ofatrazine.
10986 7pH
5
itial solution's pH on pemleate flux of pure water.
4O;OOE+OO+--......---._-..-_-...._-.-_-,,..---.
3
Authors would like to thank Universiti Sains Malaysia forfunding this research with ShortTerm Grant (Account 6035 167).Appreciation also goes to DowlFilmtec for providing the membranes.
solution's pH as it showed almost constant fluxregardless of the pH condition.
Acknowledgements
t IS study, the feasibility of nanofiltration membranes for¥ction ofpesticides in aqueous solution was evaluated with
perspective of understanding the performance of nanofiltramembranes in different pH conditions. Four nanofiltration
m mbranes, NF90, NF200, NF270 and DK, with molecularwe!ght cut-off of around 200 were subjected to stirred dead-endfiltration and the effect of initial pH's solution on the permeate flux and feed-based rejection of atrazine and dimethoatewas investigated. It was found that increasing the solution's pHincreased atrazine and dimethoate rejection but at the same time,reduced the permeate flux performance for NF200, NF270 andDK. However, effect of solution's pH had rather small significance on the performance of NF90.
From the results, it can be concluded that the NF90 had thehighest rejection of all the membranes tested. It managed tomaintain above 90% of atrazine rejection and approximately80% of dimethoate rejection regardless of the changes in solution's pH. Besides, it was rather chemical-resistant as it showedsomewhat consistent performance in both rejection and permeate flux regardless of the solution's pH. This finding strengthensthe conclusion from our study [23] that NF90 is deemed the moresuitable nanofiltration membrane for atrazine and dimethoateretention from aqueous solution compared to NF200, NF270andDK.
10
- NF90
•• NF200
-"'- NF270~OK
solumpared
ceo Thus,s not that
e, but it wases thr~gh thewhat resistant
987pH
s4
Ie
.... _._._. - .-'''' -t- ..... _._ ....
•..................•............•
..._..-..-..-.. ....._..- .....2:50E005
2:001:;·05
5.00E·08
cr'Iii i.50e·os
gS Moe·osii:
The effect of initial solution's pH on the permeate flux duringrejection of atrazine and dimethoate at fixed operating pressure,feed pesticide concentration and stirring rate are presented inFigs. 4 and 5, respectively. As the acid hydrolysis at low pH orswelling of membrane skin layer at high pH, as explained in thprevious section, is believed to be responsible for the increaseor decrease in pesticide r~jection for NF200, NF270 andit is expected that the permeate flux would be as much affby solution's pH as the pesticide rejection performancethe acid hydrolysis and hydration swelling. The effectsolution's pH on permeate flux ofpure water is showSimilar trend ofpermeate flux was observed with thpesticide at different pH which further supportedthat the variation of trend observed was due tothe membrane structures.
However, it seemed that except for NF270,tion's pH seemed not to be as much on permeato the degree ofchanges seen in the rejectionit is believed that the difference in permobvious because the changes at the polysufficient to efficiently retain or allowmembrane. Again, NF90 showed that'
References
[1] M.S. Majewski, Sources, movement, and fate of airborne pesticides, in:H. Frehse (Ed.), Pesticide Chemistry: Advances in International Research,Development, and Legislation, VCH Verlagsgesellschaft mbH, German)'1991.
986 7pH
54O;OOE+OO +--..,--.....,.--.---...---,..-.....---,
3
Fig. 5. Effect of initial solution's pH on permeate flux during rejection ofdimethoate.
A.L. Ahmad et al. / Journal ofHazardous Materials xxx (2007) xxx-xxx
I,
[19] S.S. Chen, S.T. James, L.A Mulford, C.D. Norris, Influences of molecularweight, molecular size, flux, and recovery for aromatic pesticide removalby nanofiltration membranes, Desalination 160 (2004) 103-111.
[20] Y. Zhang, B. Van del' Bruggen, G.X. Chen, L. Braeken, C. Vandecasteele,Removal of pesticides by nanofiltration: effect of the water matrix, Sep.Purif. Techno\. 38 (2004) 163-172.
[21] S. Bandini, J. Drei, D. Vezzani, The role of pH and concentration on theion rejection in polyamide nanofiltration membranes, J. Membr. Sci. 264(2005) 65-74.
[22] V. Freger, A Bottino, G. Capannelli, M. Perry, V. Gitis, S. Belfer, Characterization of novel acid-stable NF membranes before and after exposure toacid using ATR-FrIR, TEM and AFM, J. Membr. Sci. 256 (2005) 134-142.
[23] AL. Ahmad, .S. Tan, S.R. Abd. Shukor, Dimethoate and Atrazine Reten-tion from Solution by Nanofi)tration Membranes, J. Hazard.Mater., in Q:
[24] L.D. Ngh' hafer, M. Elimelech, Nanofilration ofhorrnone mim-icking c contaminants, Sep. Sci. Techno\. 40 (2005) 2633-2649.
[25] S.w. lar-enhanced ultrailftration for removal ofreactive dyesfrom lution, M.Sc. thesis, Universiti Sains Malaysia, Penang,20
[26] S. Mazzoni, Modelling the amphoteric behaviour ofpolyamidemembranes, Desalination 184 (2005) 327-336.
evelopment of polymeric composite nanofiltration membrane:characterization and its evaluation on copper sulfate rejection.
thesis, Universiti Sains Malaysia, Penang, 2005.strom, L. Kaipia, S. Luque, Fouling and retention of nanofiltratioll
embranes, J. Membr. Sci. 98 (1995) 249-262.. Freger, T.C. Arnot, J.A. Howell, Separation of concentrated,~ganic/inorganic salt mixtures by nanofiltration, J. Membr. Sci. 178 (2000)185-193.
] HC Agrochemicals Co., Ltd. http://www.hcchemica\.com/techlinsecticideslDimethoate.htm. Accessed on 10 September 2007.
1] B. Yaron, R Calvet, R Prost, Soil Pollution: Processes and Dynamics,Springer-Verlag, Heidelberg, 1996.
[32] Y. Kiso, Y. Nishimura, T. Kitao, K, Nishimura, Rejection properties ofnon-phenylic pesticides with nanofiltration membranes, J. Membr. Sci. 171(2000) 229-237.
[33] K,V. Plakas, AJ. Karabelas, T. Wintgens, T. Melin, A study ofselected herbicides retention by nanofiltration membranes-the role oforganic fouling,J. Membr. Sci. 284 (2006) 291-300.
[34] C. Bellona, J.E. Drewes, The role of membrane surface charge and solutephysico-chemical properties in the rejection of organic acids by NF membranes, J. Membr. Sci. 249 (2005) 227-234.
[35] L.D. Nghiem, A.1. Schafer, M. Elimelech, Phamlaceutical retention mechanisms by nanofiltration membranes, Environ. Sci. Techno\. 39 (2005)7698-7705.
[36] AI.C. Morao, AM.B. Alves, M.D. Afonso, Concentration of c1avulanicacid broths: influence of the membrane smface charge density on NFoperation, J. Membr. Sci. 281 (2006) 417-428.
[2] M.A. Kamrin, Pesticide Profiles: Toxicity, Environmental Impact, and Fate,CRC Press, Boca Raton, 1997.
[3] H. Blanchoud, F. Farrugia, J.M. Mouchel, Pesticides uses and transfers inurbanized catchments, Chemosphere 55 (2004) 905-913.
[4] E.P. Lichtenstein, Persistence and fate ofpesticides in soil, water and crops:significance to humans, in: A.S. Tahori (Ed.), Pesticides Chemistry. Volume VI: Fate of Pesticides in Environment, Gordon and Breach ScientificPublishers, Great Britian, 1972.
[5] R Gotz, H. Bauer, P. Friesel, K, Roth, Organic trace compounds in thewater of the River Elbe near Hamburg. Part n, Chemosphere 36 (9) (1998)2103-2118.
[6J S. EI-Kabbany, M.M. Rashed, M.A. Zayed, Monitoring of the pesticidelevels in some water supplies and agricultural land, in El-Haram, Giza(ARE.), J. Hazard. Mater. A72 (2000) 11-21.
[7] Z. Zhang, H. Hong, X. Wang, J. Lin, W. Chen, L. Xu, Determination andload of organophosphorus and organochlorine pesticides at water fromJiulong River Estuary, China, Mar. Poll. Bull. 45 (2002) 397-402.
[8] I.K. Konstantinou, D.G. Hela, T.A Albanis, The status of pesticide pollution in surface waters (rivers and lakes) of Greece. Part I. Review onoccurrence and levels, Environ. PoUut. 141 (2006) 555-570.
[9J M.T. Moore Jr., RE. Lizotte, S.S. Knight Jr., S. Smith, C.M. Cooper,Assessment of pesticide contamination in three Mississippi Delta oxbowlakes using Hyalella azteca, Chemosphere 67 (11) (2007) 21842191.
[IOJ RF. Carsel, C.N. Smith, Impact of pesticides on ground water contamination, in: G.J. Marco, RM. Hollingworth, W. Durham (Eds.), Silent SpringRevisited, American Chemical Society, Washington DC, 1987.
[II] J.E. Davies, R Doon, Human health effects of pesticides, in: G.J. Marco,R.M. Hollingworth, W. Durham (Eds.), Silent Spring Revisited, AmericanChemical Society, Washington D. C, 1987.
[12] J. DouU, Pesticide carcinogenicity, in: N.N. Ragsdale, RE. Menzer (Carcinogenicity and Pesticides: Principles, Issues and Relationship, Aican Chemical Society, Washington DC, 1989.
[13] M. Younes, H. Galal-Gorchev, Pesticides in drinking water-a casFood Chern. Toxico\. 38 (2000) S87-S90.
[l4J B. Van der Bruggen, J. Schaep, D. Wilms, C. Vandecasteele, Inmolecular size, polarity and charge on the retention of organiby nanofiltration, J. Membr. Sci. 156 (1999) 29-41.
[15] Y. Kiso, Y. Sugiura, T. Kitao, K, Nishimura, Effects ofand molecular size on rejection of aromatic pesticides wimembranes, J. Membr. Sci. 192 (2001) 1-10.
[16] L. Braeken, R Ramaekers, Y. Zhang, G. Maes, B. VaVandecasteele, Influence of hydrophobicity on retenof aqueous solutions containing organic compou(2005) 195-203.
[17] C. Causserand, P. Aimar, J.P. Cravedi, E. Singlantion by nanofiltration membranes, Water Res. 3
[18] K. Kosutic, L. Furac, L. Sipos, B. Kunst, Remofrom drinking water by nanofiltration membr(2005) 137-144.
Elsevier Editorial System(tm) for Chemical Engineering Journal
Manuscript Draft
Manuscript Number:
Title: Modeling of the retention of atrazine and dimethoate with nanofiltration
Article Type: Original Article
Section/Category: Environmental Chemical Engineering
Keywords: Nanofiltration; dimethoate; atrazine; organic molecules; Spiegler-Kedem.
Corresponding Author: Dr. Abdul Latif Ahmad,
Corresponding Author's Institution:
First Author: Abdul Latif Ahmad
Order of Authors: Abdul Latif Ahmad; Lian See Tan; Syamsul Rizal Abd. Shukor
Abstract: The present study aims to investigate the viability of using Spiegler-Kedem model to predict the
retention of atrazine and dimethoate with nanofiltration using NF90 in stirred cell condition. Spiegler-Kedem
model is the thermodynamics of irreversible processes in which no particular mechanism of transport and
structure of membrane is specified. The ~piegler-Kedem transport equations were used to derive the
reflection coefficient and solute permeabilitYilf the system. The model was successfully applied on the
modeling of the organic molecules tested. It was found that Spiegler-Kedem model provided a good
estimation of experimental value. The coefficient of determination (R2) obtained for the fitted data was
0.9871 and 0.9692 for atrazine and dimethoate, respectively.
Suggested Reviewers: Nidal Hilal PhD "
Head of Advance Water Treatment Research Group (AWTG), School of Chemical, Environmental and
Mining Engineering, The University of Nottingham
nidal.hilal@nottingham.ac.uk
He is actively involved in research in membrane and modeling field
Abdul Wahab Mohamad PhD
Head of Department, Department Of Chemical and Process Engineering, Universiti Kebangsaan Malaysia
wahabm@eng.ukm.my
He is an active researcher in membrane and modeling field
Mohamed Khayet PhD
Lecturer, Department of Applied Physics I, Faculty of Physics, Universiti Complutense of Madrid
khayetm@fis.ucm.es
He is an active researcher in membrane and modeling field
Opposed Reviewers:
,,
Date: 31 st October 2007
Prof. L. R. Weatherley,Chair & Professor,Department of Chemical & Petroleum Engineering,The University of Kansas,Learned Hall,1530 W. 15th St., Lawrence,KS 66045, USA.
Dear Prof,
MS entitled; "Modeling of the retention of atrazine and dimethoatewith nanofiltration"
I am submitting a Manuscript for possible publication in Chemical Engineering Journal.The submission requirements are listed in the table below:
No Requirements Information1 Manuscript Modeling of the retention ofatrazine and dimethoate
Title with nanofiltration2 Corresponding/ Prof. Dr. Abdul LatifAhlllad
submitter's School of Chemical Engineering, Engineering CampusInformation Universiti Sains Malaysia, Seri Ampangan, 14300 Nibong Tebal
Penang, MalaysiaEmail : chlatif@eng.usm.myTelephone: +60(4) 5941012Fax : +60(4) 5941013
3 Co-authors Lian See Tan, Syamsul Rizal Abd Shukor4 Keywords Nanofiltration; dimethoate; atrazine; organic molecules;
Spiegler-Kedem
Kindly acknowledge me the rec~ipt of the manuscript. If you have any enquiries, pleasedo not hesitate to contact at"" the above address or through my email atchlatif@eng.usm.my. Your cooperation regarding this matter is very much appreciated.
Thank You.
Yours sincerely,,,
PROF. ABDUL LATIF AHMAD
Suggested Reviewer
1. Euro Ing. Dr. Nidal HilalLeader in Chemical Engineering,Head of Advance Water Treatment Research Group (AWTO),School of Chemical, Environmental and Mining Engineering,The University ofNottingham,University Park, Nottingham NO? 2RD,United Kingdom.Tel: 44 (0)1159514168Fax: 44 (0)115951 4115E-mail: nida1.hilal@nottingham.ac.uk
2. Prof. Ir. Dr. Abdul Wahab MohamadHead of DepartmentDepartment Of Chemical and Process Engineering,Universiti Kebangsaan Malaysia,43600 UKM, Bangi, Selangor, Malaysia.Tel: 60 (3) 89296410Fax: 60 (3) 89252546Email: wahabm@eng.ukm.my
3. Prof. Dr. Mohamed Khayet,Department of Applied Physics I,Faculty of Physics, Universiti Complutense ofMadrid,Avda. Complutense sin, 28040 Madrid, SpainTel: +34 394 4511Fax: +34 394 5248E-mail: khayetm@fis.ucm.es
,,
Modeling of the retention of atrazine and dimethoate
with nanofiltration
A.L. Ahmad*, L.S. Tan, S.R. Abd. Shukor
School of Chemical Engineering, Engineering Campus, Universiti Sains Malaysia
14300 Nibong Tebal, Seberang Prai Selatan, Pulau Pinang, Malaysia.
Email address:chlatif@eng.usm.my (corresponding author*)
Phone: +60(4)5937788
Fax: +60(4)5941013
ABSTRACT,.
The present study aims to investigate..~e viability ofusing Spiegler-Kedem model
to predict the retention of atrazine and dimethoate with nanofiltration using NF90 in
stirred cell condition. Spiegler-Kedem model is the thermodynamics of irreversible
processes in which no particular mechanism of transport and structure of membrane is
specified. The Spiegler-Kedem transport equations were used to derive the reflection
coefficient and solute permeabifIty of the system. The model was successfully applied on<1/..
the modeling of the organic molecules tested. It was found that Spiegler-Kedem model
provided a good estimation of experimental value. The coefficient of determination (R2)
obtained for the fitted data was 0.9871 and 0.9692 for atrazine and dimethoate,
respectively. ,,
Keywords: Nanofiltration; dimethoate; atrazine; organic molecules; Spiegler-Kedem.
1
1. Introduction
Public attention on the potential long-term consequences ofpesticides on human
health and environment has started since 1962 when Carson [1] highlighted the matter in
her book 'Silent Spring'. Ballantyne and Marrs [2] stated that the word 'pesticides' is
used to cover substances that control organisms (insects, fungi, plants, slugs, snails,
weeds, micro-organism, nematodes, etc) which destroy plant life and interfere with food
chain, and which act as vectors to disease organism to man and animals. Pesticide
pollution in water may arise from runoff and leaching [3, 4].
The implementation on the control of water quality is important because different
type ofpesticides have different decaying period. Unlike heavy metals and other
pollutants, pesticides are lethal to the envirom)J.~nt even at micro level of concentrations
[5]. Nanofiltration is a promising membrane technique with a growing number of
applications for the treatment of drinking water and wastewater [6]. Nanofiltration
membranes differ from reverse osmosis membranes mainly because they are designed to
selectively remove compounds such as multivalent ions or organic contaminants while
allowing other compounds to pass~[7]. Furthermore, the energy requirements are much.,.
lower for nanofiltration than with reverse osmosis because the transmembrane pressures
applied in nanofiltration are significantly lower than those in reverse osmosis [8].
Some nanofiltration models take into account the mechanism of transport while
other models are independent of the mechanism transport. The solution-diffusion model,
solution-diffusion imperfection and extended Nernst-Planck model belong to the former
2
category while the Spiegler-Kedem (SK) model represents the latter [9]. Spiegler-Kedem
model is the thermodynamics of irreversible processes which indicates that the flow of
each component in a solution is related to the flows of other components. In this model,
the membrane is treated as a 'black box' in which no particular mechanism oftransport
and structure of membrane is specified [10]. It was relatively slow process near
equilibrium where the mechanisms of transport and the structure of the membrane are
ignored [11]. The nature of the membrane such as charge and compactness also does not
affect the transport activities through it [9]. The Spiegler-Kedem model has been
extensively used in predicting data for the transport of charged and unchanged solute
through the membrane in nanofiltration system [6, 10, 12-14]. However, modeling on the
retention of organic molecules has received less attention so far [6].
Atrazine was selected as subject of study because this herbicide is commonly used
in the plantations around the world as well as in Malaysia [15]. Extensive amount of its
usage has ranked it among the most common pesticides found in surface water and
groundwater [16]. On the other hand, dimethoate is also widely used in Malaysia and it is
being regulated in guidelines for drl~king water by World Health Organization.
Nevertheless, data on effectiveness ofdfmethoate retention using membranes has not
been found so far [17]. Previous studies [18, 19] found that NF90 showed superior
rejection characteristics for atrazine and dimethoate compared to other nanofiltration
membranes tested. Therefore, the objective of the pre~ent study is to investigate the,
viability ofusing Spiegler-Kedem model to predict the retention of atrazine and
dimethoate with NF90 in stirred cell condition. The measurable objectives are:
3
(1) To estimate the parameters ofthe model from the experimental data obtained from the
nanofiltration system.
(2) To validate the proposed model by comparing the simulated results with the
experimental results.
2. Theory
The Spiegler-Kedem model states that the fluxes of solute and solvent are directly
related to the chemical potential differences between the two sides of the membrane. The
chemical potential gradient is caused by either concentration or pressure gradient. The
solvent transport is due to the pressure gradient across the membrane and the solute
transport is due to the concentration gradient and/or convective coupling of the volume
flow [9].
2.1 Transport Equations
The transport equation expressed by Spiegler-Kedem model is as follows [20]:
For solvent,
-(dP d7T)J =-P --CF-v wdx sdx
For solute,
J =_p_ dCs+(l_CF)cJs s dx s s v
(1)
(2)
Diffusion is represented by the first term in Equation (2); the second term represents the
contribution of convection to the transport ofuncharged molecules [8]. In an ideally
semipermeable membrane, CF = 1. In an entirely unselective membrane in which a
4
concentration gradient does not cause volumetric flow at all, (j = O. Thus, (j is a measure
of the degree of semipermeability of the membrane reflecting its ability to pass solvent in
preference to solute and solvent and (j = 0 indicates complete coupling [9].
2.2 Model Development
The transport phenomena of nanofiltration membranes in the pressure-driven
process can be described by the irreversible thermodynamics. In general, the transport
equations for the components through a nanofiltration membrane consist of two
component which is diffusion component and convection component. This is reflected by
the transport equation of Spiegler-Kedem. For a system involving a single solute in
aqueous solution, the solute retention can be described by three transport coefficients:
1. Specific hydraulic permeability, Pw'
11. Local solute permeability, p.
iii. Reflection coefficient, (js
Permeability is the flux ofa component (solvent or solute) through the membrane
per unit driving force (the effective tfansmembrane pressure).The reflection coefficient is
a measure of the portion ofthe membrane through which the solute cannot be transferred
[12]. The assumptions made for this work are:
i. The Spiegler-Kedem model is assumed to adequately predict the transport ofI,
solutes and solvent regardless the type of solutes and its charges, solvent and
membrane.
11. The pressure and concentration gradient are the driving forces.
5
111. Solutes present in the system are semi-permeable to the membrane.
IV. In the concentration polarization layer thickness, each solute has its
independent value of the diffusion and mass transfer coefficients.
v. Pw ' P. and (j s are assumed to be constants across the uncharged membranes
so that the equation for the integration of Equation (1) and (2) of the
membrane can be simplified.
The simplified version of model transport equation can be written as [21]:
For solvent,
For solute,
Osmotic pressure, 1r, can be estimated using tHe Vant-Hoffs equation [10]:
Equation (3) and (4) can be sini~lifiedas
(3)
(4)
(5)
(6)
(7)
(8)
The imperfection of the membrane is chamcterized by the reflection coefficient,
(j • This reflection coefficient can express the degree of solute-membrane interaction
whose values are in the range of O.:s (j .:sl. An osmotic difference (111r) across an
imperfectly semi permeable membranes is compensated by an applied pressure ( 11P) so
6
that the solvent flux is zero (Jv =0) and M is smaller than I1n. The ratio between these
two is defined as (]" .
(9)
Where,
(]" = 1, for in ideally semi permeable membrane (l00% rejection)
(]" = 0, no rejection
Reflection coefficient, (]", is characteristic of the convective transport of the
solute. An (]" of 100 % indicates that the convective solute transport is totally hindered or
that no transport by convection takes place at all. This is the case for ideal RO
membranes where the membranes have a dense structure and no pores are available for,.
convective transport. The retention may how~v~rbe lower than 100 % because solute
transport may take place by solution-diffusion. As it has been shown that nanofiltration
membranes have pores, a reflection coefficient below 100 % will be found if the solutes
are small enough to enter the membrane pores.
I
The Spiegler-Kedem model assumes the membrane to be uncharged. In neutral~.
membranes, solute permeability, ~ and the reflection coefficient, (]" have constant
values characterizing a given solute-membrane system. At low pressure, both terms
contribute to the transport of solute through the membrane. However, at higher pressure,,,
the relative importance of convection in the transport will be higher. In the hypothetical
case ofan infinite pressure, diffusion is negligible compared to the infinite convection
7
flux. Since diffusion of solutes will result in an increase of transport relative to the water
transport, the relative transport solutes is at the lowest at infinite pressure. The
permeation for solute is defined as [10]:
CR =1-~
s Cms
The true rejection in term of reflection coefficient, 0" and solute driving force, F:
R = O"s(I-FJs I-O"sFs
Where Fs is defined as,
(10)
(11)
(12)
The observed retention coefficient, Ros is defined by the solute concentration in feed, eft.and the permeate Cps'
(13)
As the Spiegler-Kedem bodel relates the membrane surface concentration to the
.".permeate concentration, it needs to be combined with concentration polarization if the
permeate concentration is to be related to the bulk feed concentration. This results in the
Combined Film Theory-Spiegler-Kedem or CFSK models. This phenomenon is
expressed in the Film Theory Model [10]. Mass transfer coefficient, ks ' is an important
parameter for concentration polarization where this parameter is dependent on several
factor like feed flow rate, temperature and cell geometry.
8
Relationship between membrane surface concentration to the permeate concentration in
concentration polarization is expressed in Equation (14). The concentration polarization
usually exists in nanofiltration process because of the formation of a boundary layer
separating the membranes surface from the bulk solution [10].
Cms -Cps (J ]__----'-_ = exp _v
Cbs - Cps ks
which ks is defined as,
k = Dsw
s 8
where Dsw is the diffusion coefficient of solute in water and 8 is the concentration
(14)
(15)
polarization layer thickness. By using the rejection fractions instead of concentrations,
the Film Theory Model can be expressing as:
Ros Rs (Jv]-=--=--exp --I- Ros 1-Rs ks
By substituting the equation (11) into equation (16):
Ros (J"s(l-Fs) (Jv]-=--= exp --I- Ros 1- (J"s ks
Substitute equation (12) to (17) give~,
(16)
(17)
\,On the other hand, the following equation is applied for estimation of diffusion
coefficient in water, D w [22]:
9
(18)
D = 2.7xlO-4w mO.71
Meanwhile, for estimation for mass transfer coefficient, ks [23]:
Sherwood number, Sh:
Reynold number, Re:
Schmidt number, Sc:
3. Experimental
3.1 Materials
Dimethoate with 99.8% purity and atrazine with 97.4% purity were purchased
from Riedel-de Haen (Germany). TIle molecular structures and properties of both.,.
pesticides are presented in Table 1. The nanofiltration membrane used in this study is
(19)
(20)
(21)
(22)
(23)
NF90 (Dow/Filmtec). Table 2 provides the specification of the membrane used as given
by the manufacturers.
I,
10
3.2 Membrane Stirred Cell
A 300 mL stirred cell (Sterlitech), model Sterlitech™ HP4750, USA, was used to
conduct the dead-end filtration experiments. The effective membrane area is 1.46 x 10.3
m2• The maximum operating pressure for this cell was 69xl05 Pa.
3.3 Experimental Set-up and Procedure
Dead-end filtration experiments were carried out with the stirred cell (Sterlitech™
HP4750). The pesticide solution in the cell was stirred by a Teflon-coated magnetic bar.
The cell was pressurized using compressed high purity nitrogen gas. The pressure in the
permeate side was approximately atmospheric under all condition. The feed pesticide and
stirring rate were kept constant at 10 mg/L and 1000 rpm while the operating pressure
were varied from 5xl05 to 15xl05 Pa. The membran'e was immersed for 24 hours in
deionized water before being used in any experimental work. Membrane permeability
was determined by initially filtering it using deionized water at 16xl05 Pa for
approximately 8 hours for compaction to avoid compression effect in the later stage of
experiment. Then, stabilized water flux at different operating pressures was obtained and
membrane permeability values (Lp)~could be determined from the slope of flux against
pressure graph.
For separation process, the same compaction process was carried out before the
test cell was emptied and 1.8 litres of feed solution ",as filled into the test cell and,
solution reservoir. The cell was then pressurized at the operating pressure indicated by a
pressure regulator. Permeate from the bottom of the cell was collected and its weight was
11
measured with an electronic balance of ± 0.01 g accuracy. The cumulative weight were
converted to cumulative volume and the permeate flux could be obtained. Permeate flux,
vw (m3/m2.s), was obtained using equation (24):
LlVv =-
W Llt.A(24)
where Ll V is the cumulative volume difference (m3), M is the time difference (s) and A is
the membrane area (m2), respectively.
Samples were collected at every 20 minutes for four times and the average values
obtained from the samples were used as the results in this work. All experiments were
conducted at room temperature (25 ± 2 QC). A schematic diagram of the experimental set
up is shown in Fig. 1.
3.4 Analytical Method
Concentration ofatrazine and dimethoate in feed and permeate was analysed
using High Performance Liquid Chromatography (HPLC) by Perkin Elmer (USA). The
HPLC column used was Zorbax S~-CN (51!, 4.6mm i.d.x 150 mm long, Agilent
Technologies). The mobile phase wata mixture of35% acetonitrile and 65% deionized
water while the flow rate was set at 1.0mllmin. The UV detector was operated at a
wavelength of200nm. The peak for dimethoate was detected at around 3.5 minute while
the peak for atrazine was detected at around 5.3 mi~te. The value of retention was
obtained with the following equation:
(25)
12
where R is the pesticide retention, Cp is the concentration of permeate (mg/L) and Cf is
the concentration of feed (mgIL)
4. Results and Discussions
4.1 Parameter Estimation
The estimation ofparameters for the membrane transport model is an important
aspect of this study. The results obtained from the experimental test of the membrane
system were employed for parameter estimation for the model. The Spiegler-Kedem
model was characterized by the hydraulic permeability of the membrane, Pw ' reflection
coefficient, (J' , solute permeability, p. and mass transfer coefficient, ks ' Result of the
parameter estimation for NF90 is shown in Table 3.,.
The retention ofpesticide against permeate flux curve is presented in Fig. 2. It can
be seen from the comparison between the experimental data and predicted data that the
Spiegler-Kedem model provided good regression based on the model applied. Thus, the
parameters estimated can be accepted. In fact, the coefficient of determination (R2)
I
obtained for the fitted data was 0.9871 and 0.9692 for atrazine and dimethoate,.,.respectively.
The reflection coefficient, (J' , was in good agreement with the results obtained in\
the experimental work as it showed that NF90 had the value of an almost ideal membrane.
This is because the value close to 1 meant that it had high ability to pass solvent in
preference to solute [20], resulting in high retention of solute by NF90. Meanwhile,
13
atrazine had slightly higher mass transfer coefficient than dimethoate due to its slightly
lower molecular weight [22]. However, atrazine had obviously lower solute permeability,
p., compared to dimethoate. This lower solute permeability value possessed by atrazine
explains its higher retention compared to dimethoate. This behaviour was due to the
higher hydrophobicity (log Kow) and heterocyclic aromatic structure of atrazine [18]
4.2 Comparison between Experimental and Modeling Data
As confirmed by the irreversible thermodynamics model, the retention of solute
increased with the increasing permeate flux (i.e. increasing applied pressure). The solute
retention and permeate flux are plotted against applied pressure in Fig. 3 and Fig. 4,
respectively. It is observed that while the solute retention against pressure curve for
predicted value by Spiegler-Kedem model fitte&well with the experimental data, the
model was unable to match the slope of the experimental flux against pressure curve as
good as it did in the case for retention. However, the trend of experimental data was still
in agreement with the predicted data by the model and did not deviate far from each other.
I4.3 Concentration Polarization Profile
"if.
Fig. 5 provides the concentration polarization profiles for atrazine and dimethoate
at different operating pressure. The profiles were gauged based on the ratio of membrane
wall concentration to bulk concentration (Cm/Cb) [24]. Based on the membrane wall
concentration calculated from the Spiegler-Kedem model, the concentration polarization
profile was depicted to increase with the increasing pressure. Both solutes demonstrated
similar trends on concentration polarization where the solute concentrations increased
14
from the initial bulk concentration to the maximum concentration (wall concentration) at
the maximum pressure applied. Although previous results showed that the retention
increased with the increasing pressure, these profiles show that the effect of concentration
polarization would be magnified with the increasing pressure. The same trend was also
observed by [25]. Thus, due consideration should be given when choosing the suitable
applied pressure for nanofiltration system.
5. Conclusions
Comparisons were made between the Spiegler-Kedem model with the
experimental data obtained for nanofiltration ofatrazine and dimethoate. The model was
successfully applied on the modeling of the organic molecules tested. It was found that
Spiegler-Kedem model provided a good estimation of the experimental value. The
coefficient of determination (R2) obtained for the fitted data was 0.9871 and 0.9692 for
atrazine and dimethoate, respectively. Although the model was unable to match the slope
of the experimental flux against pressure curve as good as it did for the case for retention,
the trend of experimental data was still in agreement with the predicted data by the model
and did not deviate far from eac~ other. It was also found from concentration polarization
profiles that the effect of concentration polarization would be magnified with the
increasing pressure.
,,
15
Acknowledgement
Authors would like to thank Universiti Sains Malaysia for funding this research
with Short Term Grant (Account 6035167). Appreciation also goes to Dow/Filmtec for
providing the membrane.
Nomenclature
J v solvent fluxes (mls)
Pw specific hydraulic permeability (m4/N.s)
dx vertical distance from the membrane surface (m)
J s solute flux (mollm2.s)
~ 2s local solute permeability (m Is)
dCs solute concentration different in solution (mollm3)
Cs solute concentration in solution (mollm3)
Pw hydrodinamic permeability (m3/N.s)
M hydrostatic pressure driverl difference (N/m2)
J s solute flux (mollm2.s)
~ solute permeability (mollN.s)
Cs average solute concentration in solution (mollm3)
\,
a an osmotic constant (m3Palg),
Rg ideal gas constant (8.314 m3pa/mo1.K),
16
T
m
Sh
operating temperature (Kelvin),
solute molar mass (g/mol)
true rejection
solute concentration at the permeate side (mg/L)
solute concentration at the wall of feed side (mg/L)
solute bulk concentration (mg/L)
mass transfer coefficient (m/s)
Sherwood number (dimensionless)
diffusion coefficient in water (m2/s)
rsc radius of stirred cell (m)
Kow octanol/water partition coefficient
Lp membrane permeability
Vw permeate flux
~v cumulative volume difference
M time difference
A membrane area of·
R pesticide retention
Cp concentration ofpermeate
Cf concentration of feed
Greek letters
P density of solution (kg/m3)
OJ stirring rate (S·l)
17
\,
f.l viscosity of solution (kg/m.s)
!1;r osmotic pressure difference (N/m2)
a s reflection coefficient (dimensionless)
References
[1] R. Carson, Silent Spring, Houghton Mifflin, Boston, 1962.
[2] B. Ballantyne and T. C. Marrs, Pesticides: An overview of fundamentals. In: B.
Ballantyne and T. C. Marrs (Eds.), Pesticide toxicology and international regulation,
John Wiley and Sons, England, 2004.
[3] 1. L. Schnoor, Chemical Fate and Transport in the Environment. In: 1. L. Schnoor
(Ed.), Fate of Pesticides and Chemicals in the Environment, John Wiley and Sons,,.
United States ofAmerica, 1992.
[4] H. Beitz, H. Schmidt and F. Herzel, Occurrence, toxicological and ecotoxicological
significance ofpesticides in groundwater and surface water. In: H. Bomer (Ed.),
Pesticides in Groundwater and Surface Water, Springer-Verlag, Germany, 1994.
[5] S. Yedla and A.K. Dikshit, Abatement of Pesticide Pollution: Removal ofOrganoI
Chlorine Pesticide from Water Environment, Narosa Publishing House, New Delhi,.,.2005.
[6] B. Van der Bruggen and C. Vandecasteele, Modelling of the retention ofuncharged
molecules with nanofiltration, Water Res., 36 (2002) 1360-1368.
\
[7] K. Kosutic and B. Kunst, Removal of organics from aqueous solutions by commercial
RO and NF membranes of characterized porosities, Desalination, 142 (2002) 47-56.
18
[8] B. Van der Bruggen, J. Schaep, D. Wilms and C. Vandecasteele, A comparison of
models to describe the maximal retention oforganic molecules in nanofiltration, Sep.
Sci. and Tech., 35 (2000) 169-182.
[9] A Bhattacharya and P. Ghosh, Nanofiltration and reverse osmosis membranes: theory
and application in separation of electrolytes, Rev. Chern. Eng., 20 (2004) 111-173.
[10] AL. Ahmad, M.F. Chong and S. Bhatia, Mathematical modelling and simulation of
multiple solutes system for nanofiltration process, J. Membr. Sci., 253 (2005) 103
115.
[11] D. Van Gauwbergen and J. Baeyens, Modelling reverse osmosis by irreversible
thermodynamics, Sep. Puri. Tech., 13 (1998) 117-128.
[12] S. Wadley, C.J. Brouckaert, L.AD. Baddock and C.A Buckley, Modelling
ofnanofiltration applied to the recovery of s..alt from waste brine at a sugar
decolourisation plant, J. Membr. Sci., 102 (1995) 163-175.
[13] H. AI-Zoubi, N. Hilal, N.A Darwish and AW. Mohammad, Rejection and
modelling of sulphate and potassium salts by nanofiltration membranes: neural
network and Spiegler-Kedem model, Desalination, 206 (2007) 42-60.
[14] V.K. Gupta, S.T. Hwang, W.B. Krantz and AR. Greenberg, Characterization of
nanofiltration and reverse oSn\osis membrane performance for aqueous salt solutions
using irreversible thermodynamics, Desalination, 208 (2007) 1-18.
[15] L.K. Chooi, Personal communication, 27 May 2005.
[16] K.V. Plakas, AJ. Karabelas, T. Wintgens, l;. Melin, A study of selected herbicides
retention by nanofiltration membranes-The role of organic fouling, J. Membr. Sci.
284 (2006) 291-300.
19
[17] World Health Organization. Guidelines for drinking-water quality 2004, third ed.
http://www.who.int/water sanitation hea1th/dwq/gdwq3/en/index.html. Accessed on
16 June 2005.
[18] AL. Ahmad, L.S. Tan and S.R. Abd. Shukor, Dimethoate and Atrazine Retention
from Aqueous Solution by Nanofiltration Membranes, J. Hazard. Mater. (2007) [In
Press].
[19] AL. Ahmad, L.S. Tan and S.R. Abd. Shukor, The role ofpH in nanofiltration of
atrazine and dimethoate from aqueous solution, J. Hazard. Mater. (Accepted).
[20] K.S. Spiegler and O. Kedem, Thermodynamics of hyperfiltration (reverse osmosis):
Criteria for efficient membranes, Desalination, 1 (1966) 311-326.
[21] Z.V.P. Murthy and S.K. Gupta, Estimation ofmass transfer coefficient using a
combined nonlinear membrane transport and film theory model, Desalination, 109
(1997) 39-49.
[22] R.P. Schwarzenbach, P.M. Gschwend and D.M. Imboden, Environmental Organic
Chemistry, John Wiley & Sons, New York, 1993.
[23] AG Fane, Module design and operation. In: AI. Schafer, AG. Fane and T.D. Waite
(Eds.), Nanofiltration-printiples and application, Elsevier, Oxford, 2005.
[24] AL. Ahmad, M.F. Chong antS. Bhatia, Mathematical modeling of multiple solutes
system for reverse osmosis process in palm oil mill effluent (POME) treatment.
Chern. Eng. J., 132 (2007) 183-193.
[25] C. Bhattacharjee and S. Datta, Analysis ofpplarized layer resistance during
ultrafiltration ofPEG-6000: an approach based on filtration theory, Sep. Puri. Tech.,
33 (2003) 115-126.
20
[26] M.A. Kamrin, Pesticide Profiles: Toxicity, Environmental Impact, and Fate, CRC
Press, Boca Raton, 1997.
[27] Y. Kiso, Y. Nishimura, T. Kitao, K. Nishimura, Rejection properties ofnon-phenylic
pesticides with nanofiltration membranes, J. Membr. Sci. 171 (2000) 229-237.
"if.
I,
21
List of figures
Fig. I: Diagram of experimental set-up.
Fig. 2: Solute retention against permeate flux curve from experimental data and the
predicted results from Spiegler-Kedem model.
Fig. 3: Solute retention plotted against pressure using the experimental data and predicted
results from Spiegler-Kedem model.
Fig. 4: Permeate flux plotted against pressure using the experimental data and
predicted results from Spiegler-Kedem model.
Fig. 5: Concentration polarization profile plotted against pressure.
I,
Figure 1
Permeate
GasNz
Legend:@ :Pressure regulator
t>i<l :Valve
StainlessSteelFeed Tank Membrane
Stirred Cell
IMagnetic Stirrer I
\,
AnalyticalBalance
Figure 2
3.00E-05 3.50E-05
.--.--.----+.~~~~.-~-.
./
/
5.00E-06 1.00E-05 1.50E-05 2.00E-05 2.50E-05
Flux (m3/m2.s)
1.00
0.90
0.80
0.70
a 0.60~.e 0.50
~ 0.40
0.30
0.20
0.10
0.00 _---,-------r----,.-----,------,----,.--------,
O.OOE+OO
• Experiment atrazine
--_.~~- SK model atrazine
• Experiment dimethoate
--SK model dimethoate
,,
Figure 3
1614121086
+. - .. - . -+- .. - .. -+ +
2 4
,
:" _-----1.-------111----,,,,
,,,,,,,,,,,,,,,,,
,
1.00
0.90
0.80
0.70
5 0.60;c 0.50
~... 0040
0.30
0.20
0.10
0.00 -----,----.,.------,----,-----,----,-----,-------,
oPressure x105 (Pa)
+ Experiment atrazine•••••. SK model atrazine
• Experiment dimethoate
--- SK model dimethoate
'If.
I,
Figure 4
4.00E-05
3.50E-05
3.00E-05
.......II! 2.50E-05....
",:€2.00E-05E.....
)(~ 1.50E-05ii:
1.00E-05
5.00E-06
O.OOE+OO
0 2 4 6 8 10 12 14 16
Pressure x1 05 (Pa)
I • Experiment atrazine • Experiment dimethoate -- SK model atrazine&dimethoate I
;,
Figure 5
18
16
14
12
... 10()
E() 8
6
4
2
00 2 4 6 8 10 12 14 16
Pressure x105 (Pa)
I· . + ..Atrazine ---Dimethoate 1
I,
List of tables
Table 1: Properties of dimethoate and atrazine [26]
Table 2: Specification of membrane used
Table 3: Parameters estimated based on the experimental results
I,
Table 1
Pesticide Dimethoate Atrazine
Chemical structure
Molecular weight (Da)
Solubility in water
LogKow
229.28
25 giL @21°C
0.70
;,
215.69
20 mglL @ 20°C
Table 2
Membrane
Manufacturer
Material
Maximum operating pressure, Pa
Maximum operating temperature, °C
pH range
aour measurements.
NF90
Dow/Filmtec
Polyamide
1.90xl0-11
45
3-10
,,
Table 3
Parameter Value
Atrazine Dimethoate
Hydraulic permeability, Pw 2.3611E-ll 2.3611E-ll
Reflection coefficient, (J' 0.9835 0.9560
Solute permeability, P. 3.4317E-07 2.4142E-06
Mass transfer coefficient, ks 1.2894E-05 1.2524E-05
"if.
I,
APPENDIX 4
Award
Tan Datuk Dr Augustine S. H. OngPresident
Malaysian Invention and Design Society
heldfrom
18th- 20th lYIay 2007,
~Y~~~~Y~~~Y~T~T~~Y~
• 0 f~~. ~,~',:~, ~ ;.... ~~.'." c~.· ~ /.' _ Invention - Innovation Q ~r;~ '_"'_-':..__.._~"6'j .L 1 t::-. /\ _ Industrial Design - Technology ~~ ... "",.~ ..'. ~ i'~ MINISTRY OF SCIENCE, . MAL A Y 5 I A 0 ~, TECHNOLOGY & INNOVATION • ~ t, MIN0 S "
~ ~~ Ce~tlficate of~wa~d ~~ ~~ This is to certify that .~~ ~~ PROF ABDUL LATIF AHMAD, TAN LIAN SEE, ~~ DR SYAMSUL RIZAL,ABD SHUKOR ~~ ~~ has been awarded the ~~ ~~ ITEX SILVER MEDAL ~I~ • ~~. for the 'invention .~~ ~"~ A NOVEL NANOFILTRATION TREATMENT SYSTEM ~~, FOR PESTICIDES CONTAMINATED WATER: L~~ ~
l ENHANCEMENT OF WATER TREATMENT PLANT ~
at the ~
I ~18th International Invention, Innovation & Technology Exhibition ~
..,- ITEX 2007 .~
Kuala Lumpllr, Malaysia ~
~
~~~f~~)l
~~,
~~~
~~~~~~~~~.A.~#
~ROFESORABDUL LATIF AHMAD
304.P.TKIMIA.6035167rtJMLAH GERAN ;- 39,86620
PrIOPROJEK ~-
F:~:.:~:S - ,~~~r~~~K
~J~I:5I:.N~1 ,
UNIT KUMPULAN WANG AUANAHUNIVERSm SAINS MAlAYSIA
KAMPUS KEJURUTERAANSERl AMPANGAN
PENVATA KUMPULAN WANG
TEMPOH BERAf(HIR 311121 2007
~.I;,o!)tJLA1"8) !\!ANOJ:Tl,.TRATION PROCESS FOR PESTICIt>ES 'lltEATMENT
-D'T'e"'M~
Tempoh ProjeIc15f10~05-1~/'~tib\t4 ~a
UNIT EITD",J...,.,ih s..... M....Vfj~~
PENAJA;- JANGKA PENOEK
Pemelanjaan~m..qg.g~
31/1212006(bJ
\
Tanggungan Perbelanjaen Jumlah Jumleh Bald Peruntukansemasa Semasa Perbelanjaan Perbetanjsan Semasa
2007 ~-~~ :::;7 T~$!#L~::-'" 2007f-.",<
(e) (d) (c+d) (b+c+d) (a-(b+c+d)
--0.00 6,500.00 6,500.00 6,500.00 2,436.00
0.00 2,249.4{I 2,249.40 2,249.40 1,15fl.60
0.00 0.00 0.00 4.80 195.2fl
0.00 0.00 0.00 0.00 400.00
731.00 11,071.51 11,808.51 21,059.84 (4,92.9.84)
0.00,· 0.00 0.00 0.00 1,200.00
0.00 • 2,441.60 2,441.60 5,664.50 (1,664.50)
0..00 0.00 0.00 1,145.70 3,854.30
737.00 22.262.51 22,999.51 36,624.24 3.241.76
737.00 22.262.51 22,999.51 36,624.24 3,241.16
a.on0.00
4.80
0.00
9,251.33
0.00
3,222.90
1,145.10
13.624.73 -----
13,624.13
39,866.00
39,866.00
8,936.00
4,000.00
200.00 ~
400.00
16.130.00
1,200.00
4,000.00
5,000.00
(8)
~!"Jf'lhl"l/In
Jumlah Besar
Vat
~Jti_GAJI KAKlTANGAN AWAM
~J PERBEtANJAAN PERJA.tANAN DAN SARAHI
~~: PERHUBUNGAN DAN UTIUTI
~~_SEWMN
~BEKALAN DA.N AtAT PAKAI HABlS
~PENYELENGGARMN & PEMBAIKAN KECLL
~~~ PERKHIDMATAN IKTISAS & HosprrAUTL
~;HARTA-HARTAMODAlLAIN
Pagel UfJ· ~.g~e:-tT/).
Senarai edaran
Profe~or Madya Hamidi Abdul Aziz @ Abdul RahmanProfesor Madya Mohamad Razip SelamatProfesor Madya Ismail AbustanProfesor Madya Dr Ir. Hj. Mohd Nordin AdlanDr. Mohd Suffian YusoffProfesor Madya Sr. Mohd Sanusi S. AhamadProfesor Madya Fauziah AhmadCik Nor Habsah Md. Sabiani
Pusat Pengajian Kejuruteraan Awam
Profesor Abdul Latif AhmadProfesor Madya Ridzuan ZakariaDr. Mashitah Mat DonDr. Ahmad Zuhairi AbdullahDr. Lee Keat TeongDr. Mohd Azmier Ahmad
Pusat Pengajian Kejuruteraan Kimia
Profesor Madya Zainal Alimuddin Zainal AlauddinPusat Pengajian Kejuruteraan Mekanik
Profesor Hanafi IsmailProfesor Madya Luay Bakir HussainDr. Norlia BaharunDr. Zulkifli Mohamad Ariff
Pusat Pengajian Kejuruteraan Bahan dan Sumber Mineral
Profesor Mohd Omar Abd. KadirProfesor Wan Rosli Wan DaudProfesor Madya Mahamad~Hakimi IbrahimDr. Norli Ismail
Pusat Pengajian Tekn1)logi Industri
Profesor Mohd Asri Mohd NawiProfesor Lim Poh EngProfesor Madya Ahmad Md. NoorProfesor Madya Seng Chye EngDr. Amat Ngilmi Ahmad Sujari
Pusat Pengajian Sains Kimia
I,
Profesor Madya Mohd Nawawi Mohd NordinPusat Pengajian Sains Fizik
Profesor Madya Nik Norulaini Nik Ab. RahmanProfesor Madya Misni SurifDr. Issham Ismail
Pusat Pengajian Pendidikan Jarak Jauh
top related