Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

Water Air Soil Pollut (2013) 224:1682

DOI 10.1007/s11270-013-1682-2

Ultrafiltration Combined with Coagulation/Flocculation/


Sedimentation Using Moringa oleifera as Coagulant to Treat
Dairy Industry Wastewater
Dalila Maria Formentini-Schmitt & Álvaro Cesar Dias Alves &
Márcia Teresinha Veit & Rosângela Bergamasco &
Angélica Marquetotti Salcedo Vieira & Márcia Regina Fagundes-Klen

Received: 3 April 2013 / Accepted: 26 July 2013 / Published online: 13 August 2013
# Springer Science+Business Media Dordrecht 2013

Abstract Direct ultrafiltration and its combination with and 3 bars, using the cross-flow filtration principle in
pretreatment by coagulation/flocculation/sedimentation a membrane filtration unit. Process efficiency was eval-
using Moringa oleifera as coagulant to treat dairy in- uated in terms of chemical oxygen demand (COD),
dustry wastewater were investigated. A single-channel apparent color, and turbidity removal, along with ma-
tubular ceramic membrane with an average porosity of jor requirements such as average permeate flux, per-
0.1 μm was used at transmembrane pressures of 1, 2, centage of fouling, and contribution of different resis-
tances (resistances in series model) to the total resis-
tance of the membrane. The highest removals for the
D. M. Formentini-Schmitt : Á. C. D. Alves : M. T. Veit :
M. R. Fagundes-Klen
evaluated parameters occurred in the combined coag-
Department of Chemical Engineering, Postgraduate ulation/flocculation/sedimentation/ultrafiltration pro-
Program, NBQ, Western Paraná State University, cess. At a pressure of 2 bar, the removal of turbidity
Campus de Toledo, Rua da Faculdade, 645, Jd. Santa Maria, and apparent color was 99.9 % and that of COD was
85903-000 Toledo, Paraná, Brazil
98.5 %. For the combined process, the lowest percent-
D. M. Formentini-Schmitt age of fouling was 59.8 %, which occurred at 1 bar.
e-mail: dalila_formentini@hotmail.com
The fraction of resistance due to fouling, which may
Á. C. D. Alves indicate irreversible damage of the membrane, was
e-mail: alv_dias@yahoo.com.br
lower in the process of coagulation/flocculation/sedi-
M. T. Veit mentation using M. oleifera as coagulant followed by
e-mail: marcia_veit@yahoo.com.br
ultrafiltration than in the process that treated dairy
M. R. Fagundes-Klen wastewater with direct ultrafiltration for all pressures.
e-mail: fagundes.klen@gmail.com

R. Bergamasco Keywords Moringa oleifera . Coagulation/


Department of Chemical Engineering, flocculation . Ultrafiltration . Dairy
State University of Maringá, wastewater . Membrane fouling
Av. Colombo, 5790, 87020-900 Maringá, Paraná, Brazil
R. Bergamasco
e-mail: rosangela@deq.uem.br
1 Introduction
A. M. S. Vieira (*)
Department of Food Engineering,
State University of Maringá,
The requirements regarding control of environmental
Av. Colombo, 5790, 87020-900 Maringá, Paraná, Brazil pollution have grown exceptionally in many countries
e-mail: angelicamsalcedo@hotmail.com in recent years. Therefore, wastewater treatment is very
1682, Page 2 of 10 Water Air Soil Pollut (2013) 224:1682

important and deserves special attention in industry. oleifera practically does not change pH and conductiv-
Wastewater from dairy industry contains milk residues, ity of the treated water and produces smaller volume of
proteins, carbohydrates, fats, and residual cleaning sludge, whose disposal poses no problem (Bhatia et al.
agents (László et al. 2009). 2007).
In the dairy industry, process water produced in the If the coagulant is extracted from the seeds using
starting, equilibrating, stopping, and rinsing of the concentrated salt solutions (1 mol L−1), its efficiency in
processing units generate a large volume of effluents, terms of residual turbidity is about seven times higher
which contain diluted fractions of the products pro- than that of the coagulant extracted with distilled water
cessed by this type of industry, and contribute signifi- (Okuda et al. 2001).
cantly to the total wastewater production (Balannec The physicochemical step of coagulation/floccu-
et al. 2002). According to Tchamango et al. (2010), lation/sedimentation alone is not enough to remove
the volume may vary from 0.2 to 10 L of wastewater the amount of pollutants necessary to meet the stan-
per liter of processed milk. dards for releasing treated wastewater into receiving
Dairy industry wastewater is considered highly pol- water bodies, thus the study of a subsequent step is
luting not only due to the large volume, but also due to required.
its physicochemical characteristics, such as high chem- Ultrafiltration (UF) is a membrane separation pro-
ical and biochemical oxygen demand (COD and cess that has been widely studied for water and waste-
BOD), high concentration of nutrients and solids in water treatment. Its main feature is the retention of
suspension (Kushwaha et al. 2010), and a considerable suspended solids, colloids, and macromolecules, in
variation in pH from 2 to 12 (Sengil and Özacar 2006). addition to constituting a complete barrier for microbi-
All the aforementioned factors demonstrate the com- al species of human health concern (Ravazzini et al.
plexity required for the treatment of this wastewater and 2005), and therefore the quality of the water obtained is
the need for studies using modern and innovative treat- higher than that from conventional processes.
ment techniques, considering environmental sustain- The implementation of environmental legislation
ability, as the treatment techniques applied to this type with more and more restrictive emission standards
of wastewater are usually associated with treatment and the shortage of drinking water leading to the need
processes that combine traditional physical or physico- for water reuse are important enough reasons for using
chemical treatment with biological treatment, presenting an advanced technique like UF for wastewater treat-
disadvantages such as high energy demand and difficul- ment (Sánchez-Sánchez et al. 2010).
ty of operation (Kushwaha et al. 2010). This justifies the Purified water produced by membrane treatment
increasing search for systems that meet the standards could be reused in the dairy industry as heating/cooling
regarding the quality of treated wastewater. water, boiler make-up water, or for cleaning purposes
Another disadvantage of the conventional treatment (Vourch et al. 2008). Membrane filtration is an effec-
based on a physicochemical step using inorganic co- tive method of COD reduction, but the proteinaceous
agulants is the generation of large volumes of nonbio- material in this type of residue causes severe mem-
degradable sludge due to the presence of metals such as brane fouling, reducing dramatically the permeate flux
aluminum and iron. For this reason, alternatives have (László et al. 2009).
been researched for these inorganic coagulants, like the According to Bergamasco et al. (2011), factors
use of natural coagulants (Vieira et al. 2010). such as concentration polarization, cake formation,
The use of the extract of Moringa oleifera seeds, a solute adsorption, and plugging of the pores are also
plant from India, as natural coagulant for water treat- responsible for the decline in permeate flux by intro-
ment has been widely studied. This coagulant shows ducing additional resistance during the ultrafiltration
turbidity removal ranging from 80 to 99 % (Bhatia process. Resistance-in-series models that consider in-
et al. 2007); 90–99 % of bacteria can be removed trinsic membrane resistance, adsorption resistance,
(Muyibi and Evison 1995). concentration polarization resistance, and cake resis-
The coagulating effect of M. oleifera seeds is asso- tance have been applied in the description of such
ciated with the presence of densely charged cationic processes.
proteins with molecular weight from 6 to 16 kDa and There are few studies that use pretreatment for waste-
isoelectric pH of 10 (Ndabigengesere et al. 1995). M. water from the dairy industry prior to the membrane
Water Air Soil Pollut (2013) 224:1682 Page 3 of 10, 1682

separation process (László et al. 2009; Sarkar et al. 2.2 Coagulation/Flocculation/Sedimentation


2006) in order to mitigate the effects of fouling on the Experiments
process, so this study aims to assess the combined
process of coagulation/flocculation/sedimentation using The coagulant solution was prepared by grinding 5 g of
coagulant extracted from M. oleifera followed by ultra- M. oleifera seeds with 100 mL of 1 mol L−1 potassium
filtration in the treatment of wastewater from dairy chloride solution in a blender. The mixture was kept
industry and to compare this process with direct under magnetic stirring for 30 min for the extraction of
ultrafiltration. the active compounds from the seeds and then filtered
The efficiency of the processes was evaluated in under vacuum and used immediately in the coagula-
terms of apparent color, turbidity, and COD reduction, tion/flocculation tests. This procedure followed an
and percentages of fouling and resistances in series for adapted methodology (Beltrán-Heredia et al. 2009).
the combined and direct processes at different working The coagulant dose was taken directly from the
pressures in ultrafiltration. extract (filtered mixture). In the coagulation/floccul-
ation/sedimentation step, the following conditions were
applied: coagulant concentration of 1,500 mg L−1, rapid
mixing speed of 100 rpm for 2 min, slow mixing speed
2 Materials and Methods of 20 rpm for 10 min, and settling time of 60 min (Bhatia
et al. 2007; Bhuptawat et al. 2007). The temperature was
2.1 Wastewater Characterization 25 °C. This treated wastewater was named coagulation/
flocculation/sedimentation (CFM) and the volume used
The wastewater used in the experiments came from the in the experiments was 1.2 L.
washing of the pasteurizers and was collected at the
entrance of the treatment plant of a typical dairy indus-
try in the western region of Parana State, Brazil. A 2.3 Ultrafiltration Experiments
single sample was collected and then homogenized,
fractionated, and suitably stored under refrigeration The tests were carried out in an ultrafiltration module
for subsequent use. The sample was characterized by (Netzsch, Fig. 1) using cross-flow filtration with a
physicochemical parameters—apparent color and tubular single-channel ceramic membrane of average
COD (HACH DR/2010 spectrophotometer) and tur- porosity of 0.1 μm, surface area of 0.005 m2, and
bidity (HACH 2100 P turbidimeter)—immediately af- molecular weight cutoff of 4 kDa. Al2O3/ZrO2 ceramic
ter collection and after the coagulation/flocculation/ membranes (TAMI, France) were used in the filtration
sedimentation/ultrafiltration and direct ultrafiltration process. The filtration module was made of stainless
tests. steel. The system was equipped with manometers at the
All experiments to determine the physicochemical inlet and outlet to control the transmembrane pressure
parameters were performed in triplicate and followed and connected to a thermostatic bath for temperature
the methodology of the Standard Methods for the Ex- control of the solution contained in the feed tank. The
amination of Water and Wastewater (APHA 1994). output of permeate was collected by opening the valve
Table 1 presents the characteristics of a sample of the and the concentrate was returned to the feed tank by the
dairy industry wastewater. hose.
Temperature was maintained at 25 °C for the tests,
with transmembrane pressures of 1, 2, and 3 bars. The
Table 1 Physicochemical characteristics of dairy industry
tests were conducted using the raw dairy industry
wastewater wastewater (RDW) and that pretreated by coagulation/
flocculation/sedimentation using M. oleifera as coagu-
Parameter Value
lant (CFM). For the combined experiments, RDW
Apparent color (mgPt-CoL−1) 5,544 underwent coagulation/flocculation/sedimentation in
Turbidity (NTU) 897
jar test before being subjected to ultrafiltration. A peri-
COD (mgO2L−1) 3,190
staltic pump was used to transfer 10 L of CFM from the
jar test to the ultrafiltration module. Then, the
1682, Page 4 of 10 Water Air Soil Pollut (2013) 224:1682

suspension was allowed to percolate through the mem- membrane clogging and consequent reduction of per-
brane system as shown in Fig. 1. meate flux. The total resistance (Rt) is obtained by
The initial permeability of the membrane at different adding the different resistances, that is, Rm, concentra-
pressures (1, 2, and 3 bars) was determined, as well as tion polarization resistance (Rc) due to the deposition
the filtration profile that was divided into three stages: of a layer of solute on the membrane surface, and
initial flux (reverse osmosis water (ROW)), flux of the fouling resistance (Rf), which is caused by solute ad-
solution to be treated (RDW or CFM), and final flux sorption in the membrane pores and walls. The resis-
(ROW). Each step had an estimated time of 180 min. tances were calculated through Eqs. (4) to (7):
Lastly, the final permeability was determined.
The permeate flux (J) was calculated in predeter- Rt ¼ Rm þ Rc þ R f ð4Þ
mined time intervals. It depends on the mass of permeate
(m) collected during a certain period of time (Δt), the
density (ρ) of the solution to be treated at 25 °C, and the
membrane area (Am) according to Eq. (1): Rm ¼ ΔPT =ðη: J initial Þ ð5Þ

J L: h−1 :m−2 ¼ ½ðm=ρ25 C Þ=Δt: Am ð1Þ

ROW flux (Jinitial and Jfinal) was assessed at the


beginning of the experiment and after determining the R f ¼ ΔPT =ðη: J final Þ–Rm ð6Þ
flux with RDW or CFM. Time intervals after stabiliza-
tion of the permeate flux (from 70 to 180 min) were
used to determine membrane fouling according to
Eq. (2). 
Rc ¼ ΔPT =ðη: J Þ– Rm þ R f ð7Þ
% fouling ¼ ½1–ð J final =J initial Þ  100 ð2Þ
In order to better understand the permeate flux pro-
The resistance in series model used by Bergamasco
files, the relative flux was calculated by Eq. (8)
et al. (2011) was applied to evaluate the filtration
according to Reddy et al. (2005). The relative flux
characteristics. The intrinsic membrane resistance
allows rapid evaluation of how much the permeate flux
(Rm), also known as hydraulic resistance, was calculat-
of RDW and CFM has decreased in comparison with
ed from data obtained during ROW filtration. ROW
the initial ROW flux.
flux (Jinitial) is shown in Eq. (3), where ΔPT is the
transmembrane pressure (working pressure), and η is % relative flux ¼ ð J = J initial Þ  100 ð8Þ
the dynamic viscosity of the permeate.
J initial ¼ ΔPT =ðη: Rm Þ ð3Þ The percentage of membrane fouling, as well as its
resistances, related with the process efficiency in terms
Fouling and concentration polarization of the solute of apparent color, COD, and turbidity removal, were
at the surface of the membrane are responsible for used to determine the best working condition.

Fig. 1 Schematic diagram


of the micro/ultrafiltration
experimental unit. 1 Ther-
mostatic bath, 2 feed tank, 3
pump, 4 manometers, 5
membrane filtration module,
6 flowmeter (rotameter), 7
permeate, 8 concentrate
Water Air Soil Pollut (2013) 224:1682 Page 5 of 10, 1682

3 Results and Discussion cheaper separation process as it uses lower pressures


than reverse osmosis.
Table 2 shows removal efficiencies in terms of appar- In the process of direct UF of the wastewater with-
ent color, turbidity, and COD reduction for the raw out pretreatment (RDW-UF), apparent color and tur-
dairy industry wastewater directly subjected to ultrafil- bidity removal was above 99.3 % even with pressure
tration (RDW-UF) and for the wastewater pretreated by variation. COD reduction varied between 96.1 and
coagulation/flocculation/sedimentation prior to ultra- 98 %.
filtration (CFM-UF) at different transmembrane pres- Figures 2, 3, and 4 present filtration profiles
sures (1, 2, and 3 bars). The efficiency of the coagula- obtained during the experiments showing a reduction
tion/flocculation/sedimentation pretreatment using the in the permeate flux over time. These profiles showed a
coagulant extracted from M. oleifera seeds (CFM) is steep decline in permeate flux at the initial period of
also shown. ultrafiltration. After a certain time, which was charac-
According to Table 2, the coagulation/flocculation/ teristic of each test, permeate flux remained stable until
sedimentation pretreatment with M. oleifera showed the end of the test.
high efficiencies of turbidity and apparent color remov- Increasing the pressure enabled an increase in flux
al. However, it is apparent that this process was not for all tests (CFM-UF and RDW-UF). This result is in
efficient for COD removal as the COD value obtained accordance with Bergamasco et al. (2011), who have
after pretreatment was 1,933.1 mg L−1, which is too reported that an increase in flux is expected when
high for disposal into receiving water bodies, justifying pressure is increased, particularly at the beginning of
the need for a subsequent treatment such as ultrafiltra- ultrafiltration when there is no formation of a gel layer
tion to fit this parameter to the current disposal stan- on the membrane.
dards. The results also showed that separation by ul- ROW is only subject to the intrinsic resistance when
trafiltration provided removals exceeding 96 % for all passing through the membrane during the initial ultra-
evaluated parameters and at all conditions studied. It filtration process as it is pure water. RDW and CFM, on
should be highlighted the significant increase in COD the other hand, have particles that must be removed
removal, from only 39.4 % in the CFM process to more and that accumulate on the membrane surface or ob-
than 96 % in the ultrafiltration processes. struct membrane pores, so the flux of these substances
Similar results of COD removal were obtained by is much lower than that of ROW.
Sarkar et al. (2006) who used wastewater from dairy Table 3 shows the relative flux values obtained by
industry in their experiments, testing chitosan Eq. (8) after certain time intervals during the ultrafil-
(10 mg L−1) at pH 4 for the coagulation/flocculation tration tests. The relative fluxes of RDW and CFM at
step, followed by adsorption with powdered activated the beginning of the experiments ranged between 40
charcoal (1.5 g L−1), which has improved color and odor and 64 %. Stabilized fluxes occurred after 25 min, on
removal. The pretreated wastewater subsequently average, and remained between 14 and 26 %.
underwent a tertiary treatment of reverse osmosis, in The observation of the relative fluxes shown in Ta-
which a reduction of 98 % in COD was achieved. ble 3 aided in the evaluation of data presented in Figs. 2,
In the present study, pH adjustment was not neces- 3, and 4 as they clearly demonstrated the decrease in
sary for the coagulation/flocculation/sedimentation permeate flux of RDW and CFM in comparison with
pretreatment with M. oleifera. Therefore, fewer steps ROW flux. According to Reddy et al. (2005), this is
were used (CFM-UF or RDW-UF); ultrafiltration is a due to blocking of the larger pores of the membrane

Table 2 Removal efficiencies


for CFM, CFM-UF, and RDW- Parameter CFM % CFM-UF % RDW-UF %
UF at different pressures
1 bar 2 bar 3 bar 1 bar 2 bar 3 bar

Apparent color 93.3 100.0 99.9 99.8 99.9 99.3 99.7


Turbidity 97.6 99.9 99.9 99.8 99.9 99.6 99.8
COD 39.4 96.1 98.5 98.3 96.1 97.2 98.0
1682, Page 6 of 10 Water Air Soil Pollut (2013) 224:1682

Fig. 2 Filtration profiles for


CFM-UF and RDW-UF at
1 bar

caused by solute molecules present in RDW and CFM. RDW-UF and CFM-UF, which reached approximately
Higher relative flux was expected for CFM—as it was 9 % at 3 bar.
pretreated—than for RDW, but this was only observed Membrane fouling is an inevitable drawback of the
in the first minutes of the experiments. separation process. However, it can be minimized, and
The increase in pressure allowed an increase in the lower the fouling rate, the more advantageous the
permeate flux, but did not significantly alter the rela- process.
tive flux as RDW and CFM fluxes increased propor- The different pressures and solutes used in the tests
tionally to the ROW fluxes of the respective tests. The had a significant influence on the characteristics of
largest relative flux occurred in the experiment with membrane fouling. Figure 5 shows the percentage of
CFM at 1 bar. fouling for the RDW-UF and CFM-UF processes, with
The average permeate flux of RDW-UF (calculated different transmembrane pressures (1, 2, and 3 bars).
between 50 and 150 min) was higher than that of CFM- It is clear from Fig. 5 that the increase in membrane
UF for the three pressures studied as shown in Table 4. fouling was not proportional to the increase in pres-
The major difference between the average permeate sure. These results are in disagreement with
fluxes was observed at the pressure of 1 bar in which Bergamasco et al. (2011) who found in their studies
the RDW-UF flux was approximately 27 % higher than that the increase in pressure promoted an increase in
that of CFM-UF. The increase in pressure increased the permeate flux and intensified the effect of fouling.
average fluxes and decreased the difference between Turan (2004) also observed that the permeate flux

Fig. 3 Filtration profiles for


CFM-UF and RDW-UF at
2 bar
Water Air Soil Pollut (2013) 224:1682 Page 7 of 10, 1682

Fig. 4 Filtration profiles for


CFM-UF and RDW-UF at
3 bar

increased with the increase in transmembrane pressure. (2009), studying surface water treatment, observed that
However, this positive effect was reduced when the although the quality of treated water was improved, no
feed solution had higher COD and therefore, according gain was obtained in the performance of the mem-
to the author, higher concentration of pollutants that brane, as in this case there was an increase in mem-
cause fouling. brane fouling.
The percentages of fouling observed in Fig. 5 were In ultrafiltration, the mechanism of membrane ob-
lower for CFM-UF than for RDW-UF at all pressures. struction must be taken into account. This mechanism
COD removal (Table 2) was higher for CFM-UF than causes a drop in permeate flux and is the result of phe-
for RDW-UF at 2 and 3 bars; at 1 bar, it was the same nomena (Bergamasco et al. 2011) that are mainly re-
for both processes. As for the average permeate flux, lated to the characteristics of the membrane (hydropho-
the values shown in Table 4 were lower for CFM-UF bicity, charge, surface roughness, and porosity), oper-
than for RDW-UF at all pressures. ating conditions of the membrane system (constant
Nevertheless, the above observation represents an working pressure and flux), and quality of the feed
advantage for the combined CFM-UF process. In a water (physicochemical and microbiological character-
similar study by Katayon et al. (2007), using M. istics), as well as the hydrodynamics of the membrane
oleifera as a coagulant followed by microfiltration for system which is characterized by the permeate flux and
wastewater treatment, the combination of the methods the cut surface. Among the most relevant physico-
improved permeate flow and reduced membrane foul- chemical characteristics is the amount of organic mat-
ing, but no further removal of the evaluated parameters ter present in the feed water which is a primary con-
was obtained. On the other hand, Bergamasco et al. tributor to membrane fouling (Guo et al. 2010).
The mechanisms commonly attributed to organic
Table 3 Relative flux for RDW-UF and CFM-UF over time at matter removal by ultrafiltration are sieve retention—
different pressures particles are retained on the membrane surface and
Relative flux (%)
form a cake that grows in thickness as the filtration

RDW-UF CFM-UF
Table 4 Average permeate flux for RDW-UF and CFM-UF at
Time (min) 1 bar 2 bar 3 bar 1 bar 2 bar 3 bar different pressures

0.25 49 40 40 64 41 40 Process Average permeate flux (L h−1 m−2)


10 34 24 23 26 28 25
1 bar 2 bar 3 bar
25 27 20 24 28 19 19
50 26 18 23 23 16 19 RDW-UF 80.4 96.4 132.3
150 26 15 22 16 14 18 CFM-UF 59.0 85.4 120.1
1682, Page 8 of 10 Water Air Soil Pollut (2013) 224:1682

resistance observed in ultrafiltration using RDW and


CFM at 1, 2, and 3 bars. Comparing Rf at different
pressures in Table 5, it can be observed that it was
higher for RDW-UF than for CFM-UF. As for Rc, it
was higher for CFM-UF than for RDW-UF at all pres-
sures. As stated by Guo et al. (2009), concentration
polarization resistance is more easily removed than
fouling resistance, which can be irreversible. There-
fore, the combined CFM-UF process was found to be
appropriate to increase membrane lifetime.
As shown in Table 5, total resistance increased with
Fig. 5 Percentage of fouling for the tests with RDW-UF and increasing transmembrane pressure. This is due to the
CFM-UF at different pressures higher compression according to Bergamasco et al.
(2011). Table 6 shows the contributions of the different
progresses—and adsorption sequestration, which in- resistances to the total resistance of the membrane in
volves input and capture of particles into the membrane the RDW-UF and CFM-UF processes at 1, 2, and
matrix, allowing the removal of particles smaller than 3 bars. Rf had the greatest contribution to total resis-
the pores of the membrane (Guo et al. 2009). tance in the case of direct ultrafiltration of raw dairy
For Guo et al. (2009), the direct filtration of natural wastewater. In the combined process, Rc contributed
organic matter by ultrafiltration is through adsorption the most to the total resistance, except at 2 bar when Rf
sequestration, which causes severe membrane fouling had a slightly greater contribution.
that is not easy to remove and is often irreversible. If Tables 5 and 6 demonstrated that the use of M.
coagulation is combined with ultrafiltration, the fine oleifera for coagulation prior to ultrafiltration de-
particles are deposited on the membrane surface by creased Rf and increased Rc. Rc is easier to remove than
concentration polarization, which is considered as re- Rf and contributed with 60.1 % of Rt at 1 bar, 40.9 % at
versible fouling and is easy to clean by physical methods 2 bar, and 50.8 % at 3 bar in the CFM-UF process.
as the filtration progresses. Therefore it is important to Cross-flow filtration prevents excessive concentra-
note in Tables 5 and 6 the influence of different resis- tion of material on the membrane surface as most of the
tances on the ultrafiltration of RDW and CFM at differ- feed stream just passes along the membrane and is
ent pressures (1, 2, and 3 bars) and the relative contri- recirculated to the feed tank. The shear stress tangential
butions of each resistance to the total resistance in each to the membrane surface helps to reduce fouling phe-
process. nomena (Ravazzini et al. 2005). This can also be seen
Table 5 shows the results obtained through the re- in Tables 5 and 6, as the contribution of Rc to Rt for
sistances in series model for the different types of CFM was greater at 1 bar than at the other pressures.
Possibly due to different physicochemical characteris-
Table 5 Membrane resistances during ultrafiltration of RDW
and CFM at 1, 2, and 3 bars
tics of RDW and CFM solutes, the pressure of 1 bar
allowed a greater accumulation of material on the
Process Resistance×1012 (m−1) membrane for CFM than for RDW, and this effect
was minimized by increasing the pressure, that is, a
Rm Rf Rc Rt
higher pressure increased the flux and did not allow the
RDW-UF formation of such a thick cake layer on the membrane.
ΔP=1 bar 1.36 2.16 1.63 5.15 Lee et al. (2009) stated that permeability reduction
ΔP=2 bar 1.37 4.53 3.31 9.21 in ultrafiltration, when mainly attributed to cake for-
ΔP=3 bar 1.91 4.08 3.93 9.92 mation on the membrane surface, exhibits dual func-
CFM-UF tions while acting as a protective barrier for membrane
ΔP=1 bar 1.36 1.93 4.96 8.25 fouling in that it provides additional resistance to per-
ΔP=2 bar 1.37 4.26 3.90 9.53
meation. A similar behavior was observed in the pres-
ΔP=3 bar 1.91 3.30 5.38 10.58
ent study as the relative contribution of Rf to Rt at 1 bar
for CFM was the lowest in all experiments, and
Water Air Soil Pollut (2013) 224:1682 Page 9 of 10, 1682

Table 6 Relative contributions


of the different resistances to the ΔP (bar) RDW-UF CFM-UF
total membrane resistance during
ultrafiltration of RDW and CFM %Rm %Rf %Rc %Rm %Rf %Rc
at 1, 2, and 3 bars
1 26.4±0.3 41.9±0.5 31.7±0.6 16.5±0.1 23.4±0.3 60.1±1.6
2 14.9±0.2 49.2±0.0 35.9±0.1 14.4±0.2 44.7±0.4 40.9±0.8
3 19.2±0.2 41.2±0.4 39.6±0.6 18.0±0.4 31.2±0.3 50.8±1.0

consequently the permeate flux under these conditions Importantly, M. oleifera emerged as a natural coag-
was also the lowest one. ulant for this type of wastewater, as the sludge gener-
Stoller (2009) pointed out that the flocculation ated in the coagulation/flocculation/sedimentation pro-
process—due to reduction of solute concentration by cess is biodegradable, unlike the sludge generated in
sedimentation and particle size shifts—can significantly the processes that use inorganic coagulants. Coagula-
influence fouling. Therefore, the author suggests that the tion/flocculation using M. oleifera can be a pretreat-
flocculation process as pretreatment for membrane op- ment for a biological treatment or ultrafiltration. The
eration must be optimized to not produce aggregates choice should be based on the quality required for the
whose dimensions are similar to the pore size. Different treated water, which may be disposed into a receiving
coagulants perform different particle size shifts; this body of water or reused (irrigation, industrial water).
should be considered in any membrane filtration process For reuse purposes, regardless of the process
using coagulation/flocculation as a pretreatment. (RDW-UF or CFM-UF), ultrafiltration stands out as
The results obtained in the ultrafiltration processes an alternative to conventional treatments used for
using RDW and CFM were similar, although superior wastewater from the dairy industry, because the quality
performance was expected for the ultrafiltration pro- of the treated water is much higher than that obtained
cess combined with coagulation/flocculation using M. with conventional treatments.
oleifera when compared with direct ultrafiltration
according to some reports (Turan 2004; Sarkar et al.
2006; Katayon et al. 2007; László et al. 2009; Stoller 4 Conclusions
2009; Bergamasco et al. 2011).
As dairy industry wastewaters contain particles of Direct ultrafiltration and its combination with pretreat-
high molecular weight (milk proteins, fat, and whey ment by coagulation/flocculation/sedimentation using
protein) and minerals like calcium and phosphorus, M. oleifera as coagulant under the optimum conditions
milk-derived substances (which had been coagulated determined in the coagulation/flocculation tests achieved
and settled) may have been partially substituted for good results regarding COD removal, exceeding 96 % at
substances from the M. oleifera. The use of a M. all pressures (1, 2, and 3 bars).
oleifera coagulant dosage of 1,500 mg L−1 with a The highest removals for the evaluated parameters
highly concentrated saline solution (1 mol L−1) and a were obtained in the coagulation/flocculation/sedimenta-
membrane with molecular weight cutoff of 4 kDa tion process combined with ultrafiltration, especially at
could explain the similar performance of both process- higher pressures. At 2 bar, the removal of turbidity and
es (RDW-UF and CFM-UF), in which despite the apparent color was 99.9 % and that of COD was 98.5 %.
increase in pressure, there were no significant changes.
& For the combined CFM-UF process, the lowest
With the evaluation of data such as average perme-
percentage of fouling was obtained at 1 bar.
ate flux and total resistance, the direct ultrafiltration
& The fraction of resistance due to fouling (%Rf),
process with RDW was found to be more effective. In
which may cause irreversible damage of the mem-
contrast, the better removals obtained with CFM-UF,
brane, was lower in the CFM-UF process than in
the lower percentage of fouling, and the Rc values,
the RDW-UF process at all pressures.
which provided important information about the type
of fouling that can occur (reversible or irreversible), are The combined CFM-UF process showed many ad-
advantages of the CFM-UF process. vantages over conventional processes, as besides
1682, Page 10 of 10 Water Air Soil Pollut (2013) 224:1682

ensuring superior quality of treated water, it enables the Kushwaha, J. P., Srivastava, V. C., & Mall, I. D. (2010). Treat-
ment of dairy wastewater by commercial activated carbon
reuse of treated water for purposes other than potable
and bagasse fly ash: parametric, kinetic and equilibrium
water within the dairy industry, reducing the consump- modelling, disposal studies. Bioresource Technology, 101,
tion of drinking water, which is a major environmental 3474–3483.
problem of this type of industry. László, Z., Kertész, S., Beszédes, S., Hovorka-Horváth, Z.,
Szabó, G., & Hodúr, C. (2009). Effect of preozonation on
the filterability of model dairy waste water in nanofiltration.
Acknowledgments The authors thank CNPq–National Coun- Desalination, 240, 170–177.
cil for Scientific and Technological Development and Araucaria Lee, B.-B., Choo, K.-H., Chang, D., & Choi, S.-J. (2009).
Foundation for Support to Scientific and Technological Devel- Optimizing the coagulant dose to control membrane fouling
opment of Paraná for the financial support. in combined coagulation/ultrafiltration systems for textile
wastewater reclamation. Chemical Engineering Journal,
155, 101–107.
Muyibi, S. A., & Evison, L. M. (1995). Moringa oleifera seeds
References for softening hard water. Water Research, 29, 1099–1105.
Ndabigengesere, A., Narasiah, K. S., & Talbot, B. G. (1995).
Active agents and mechanism of coagulation of the turbid
APHA. (1994). Standard methods for the examination of water waters using Moringa oleifera. Water Research, 29, 703–710.
and wastewater. New York: American Public Health Okuda, T., Baes, A. U., Nishijima, W., & Okada, M. (2001).
Association. Isolation and characterization of coagulant extracted from
Balannec, B., Gesan-Guiziou, G., Chaufer, B., Rabiller-Baudry, Moringa oleifera seed by salt solution. Water Research, 35,
M., & Daufin, G. (2002). Treatment of dairy process waters 405–410.
by membrane operations for water reuse and milk constit- Ravazzini, A. M., van Nieuwenhuijzen, A. F., & van-der Graff, J.
uents concentration. Desalination, 147, 89–94. H. M. J. (2005). Direct ultrafiltration of municipal waste-
Beltrán-Heredia, J., & Sánchez-Martín, J. (2009). Removal of water: comparison between filtration of raw sewage and
sodium lauryl sulphate by coagulation/flocculation with primary clarifier effluent. Desalination, 178, 51–62.
Moringa oleifera seed extract. Journal of Hazardous Ma- Reddy, A. V. R., Trivedi, J. J., Devmurari, C. V., Mohan, D. J.,
terials, 164, 713–719. Singh, P., Rao, A. P., et al. (2005). Fouling resistant mem-
Bergamasco, R., Bouchard, C., Silva, F. V., Reis, M. H. M., & branes in desalination and water recovery. Desalination,
Fagundes-Klen, M. R. (2009). An application of chitosan as 183, 301–306.
a coagulant/flocculant in a microfiltration process of natural Sánchez-Sánchez, A., Garrido, J. M., & Méndez, R. (2010). A
water. Desalination, 245, 205–213. comparative study of tertiary membrane filtration of indus-
Bergamasco, R., Konradt-Moraes, L. C., Vieira, M. F., trial wastewater treated in a granular and a flocculent sludge
Fagundes-Klen, M. R., & Vieira, A. M. S. (2011). Perfor- SBR. Desalination, 250, 810–814.
mance of a coagulation-ultrafiltration hybrid process for Sarkar, B., Chakrabarti, P. P., Vijaykumar, A., & Kale, V. (2006).
water supply treatment. Chemical Engineering Journal, Wastewater treatment in dairy industries—possibility of
166, 483–489. reuse. Desalination, 195, 141–152.
Bhatia, S., Othman, Z., & Ahmad, A. B. (2007). Pretreatment of Sengil, I. A., & Özacar, M. (2006). Treatment of dairy wastewa-
palm oil mill effluent (POME) using Moringa oleifera ters by electrocoagulation using mild steel electrodes. Jour-
seeds as natural coagulant. Journal of Hazardous Mate- nal of Hazardous Materials, 137, 1197–1205.
rials, 145, 120–126. Stoller, M. (2009). On the effect of flocculation as pretreatment
Bhuptawat, H., Folkard, G. K., & Chaudhari, S. (2007). Innova- process and particle size distribution for membrane fouling
tive physico-chemical treatment of wastewater incorporat- reduction. Desalination, 240, 209–217.
ing Moringa oleifera seed coagulant. Journal of Hazardous Tchamango, S., Nanseu-Njiki, C. P., Ngameni, E., Hadjiev, D., &
Materials, 142(1–2), 477–482. Darchen, A. (2010). Treatment of dairy effluents by
Guo, X., Shao, H., Hu, W., Gao, W., & Chen, X. (2010). Tannin electrocoagulation using aluminium electrodes. Science of
and polyacrylic acid polarity and structure influence on the the Total Environment, 408, 947–952.
performance of polyvinylchloride ultrafiltration membrane. Turan, M. (2004). Influence of filtration conditions on the per-
Desalination, 250, 740–744. formance of nanofiltration and reverse osmosis membranes
Guo, X., Zhang, Z., Fang, L., & Su, L. (2009). Study on ultra- in dairy wastewater treatment. Desalination, 170, 83–90.
filtration for surface water by a polyvinylchloride hollow Vieira, A. M. S., Vieira, M. F., Silva, G. F., Araújo, A. A.,
fiber membrane. Desalination, 238, 183–191. Fagundes-Klen, M. R., Veit, M. T., et al. (2010). Use of
Katayon, S., Noor, M. J. M. M., Tat, W. K., Halim, G. A., Moringa oleifera seed as a natural adsorbent for wastewater
Thamer, A. M., & Badronisa, Y. (2007). Effect of natural treatment. Water, Air, and Soil Pollution, 206, 273–281.
coagulant application on microfiltration performance in Vourch, M., Balannec, B., Chaufer, B., & Dorange, G. (2008).
treatment of secondary oxidation pond effluent. Desalina- Treatment of dairy industry wastewater by reverse osmosis
tion, 204, 204–212. for water reuse. Desalination, 219, 190–202.

You might also like