Ireche Vol 5 N 4

International Review of
Chemical Engineering
Rapid Communications
(IRECHE)
Contents
Performance Evaluation of Ducts with Non-Circular Shapes: 265

Laminar Fully Developed Flow Subjected to T Boundary Condition
by Valentin M. Petkov, Ventsislav D. Zimparov
Performance Evaluation of Photo-Fenton and Fenton Processses for Dairy Effluent Treatment 280
by Carla C. A. Loures, Hélcio J. Izário. Filho, Gisella R. Lamas Samanamud, André L. Souza,
Rodrigo F. S. Salazar, André L. C. Peixoto, Oswaldo L. C. Guimarães
Influence of Magnetic Field on the Efficiency of the Coagulation Process 289
to Remove Turbidity from Water
by Waleed M. Sh. Alabdraba, Mohamed B. A. Albayati, Ahmed Y. Radeef, Mustafa M. Rejab
Evaluation of a System for Dairy Wastewater Treatment Composed 294
by Photochemical (TiO2 fixed / UVsolar) and Biological (Aeration Pond) Reactors
by Rodrigo F. S. Salazar, Marcos F. Oliveira, Marco A. K. Alcântara, Hélcio J. Izário Filho
A Research Note on Oil Extraction from Jatropha Curcas Seeds 301
by Akhihiero E. Thelma
Analysis of the Water Quality in the Water Distribution System, Based on Logistic Regression 305
by Camelia Gavrila, Alina Elisabeta Sandu, Elena Burchiu
Biosurfactant Production from Industrial Residues Using Microorganisms Isolated from Oil Wells 310
by Ana Katerine de C. Lima Lobato, Andréa F. Almeida, Márcio S. Bezerra, Laerte M. Barros Júnior,
Luiz Carlos L. Santos, Gorete R. Macedo
BSCF 5582 and LSCF 6428 Mixed Conducting Membranes for Syngas Production 317
by Luiz Carlos L. Santos, Ana Katerine de C. Lima Lobato, Caetano Moraes
Copyright © 2013 Praise Worthy Prize S.r.l. - All rights reserved

International Review of Chemical Engineering (I.RE.CH.E.), Vol. 5, N. 4
ISSN 2035-1755 July 2013
Performance Evaluation of Ducts with Non-Circular Shapes:

Laminar Fully Developed Flow Subjected to T Boundary Condition
Valentin M. Petkov, Ventsislav D. Zimparov
Abstract – Extended performance evaluation criteria (ExPEC) have been used to assess the
performance characteristics of single-phase fully developed laminar flow through bundle of ducts
with non-circular shape - rectangular, isosceles triangular, elliptical, trapezoidal and hexagonal.
The bundle of circular tubes has been used as a reference heat transfer unit. Constant wall
temperature has been selected as the thermal boundary condition. The performance
characteristics of the heat transfer unit with non-circular ducts have been evaluated and
compared to those of the reference unit for different objectives and constraints imposed. As a
common constraint, the hydraulic diameter of the non-circular duct has been specified.
The results from this study clearly show that the rectangular, trapezoidal or hexagonal duct
configuration can compete, in some cases, with the circular tube configuration. The choice of the
tube shape and geometrical details depends on the constraints imposed and the objectives
pursued. The results obtained from the present study revealed that they differed from those
obtained with a common constraint of specified cross sectional area of the non-circular duct.
Copyright © 2013 Praise Worthy Prize S.r.l. - All rights reserved.
Keywords: Performance Evaluation Criteria, Single-Phase Laminar Flow, Non-Circular Ducts,

Entropy Generation
Nomenclature  Dynamic viscosity [ Pas ]

 Fluid density [ kg m-3 ]
A Heat transfer surface area [ m 2 ]
Af Cross-sectional flow area [ m 2 ] Dimensionless groups
cp Specific heat capacity [ J kg -1K -1 ] A* Dimensionless heat transfer surface, A / Ac
D Reference circular tube diameter [m] Dimensionless cross-sectional area,
A*f A f / A f ,c
Dh Hydraulic diameter [m]
k Thermal conductivity [ W m-1 K -1 ] D* Dimensionless tube diameter, Dh / Dc
Lt Tube length [m] L* Dimensionless tube length, L / Lc
m t Mass flow rate in tube [ kg s ] -1
f Fanning friction factor
Nt Number of tubes Nu Nusselt number
P Pumping power [W] Nu* Nusselt number ratio, Nu / Nuc
p Wetted perimeter [m] NS Augmentation entropy generation number
p Pressure drop [Pa] N* Ratio of number of tubes, Nt / N t ,c
Q Heat transfer rate [W] NTU Heat transfer units, 4St L / D
Qt Tube heat transfer rate [W] Pr Prandtl number
S gen Entropy generation rate [ W K -1 ]
P* Dimensionless pumping power, P / Pc
T Temperature [K] Q* Dimensionless heat transfer rate, Q / Q c

V Volume of tubes [ m3 ] Re Reynolds number
Re* Reynolds number ratio, Re/ Rec
W Mass flow rate in heat exchanger [ kg s-1 ]
x Axial distance along the tube [m] f Re * Ratio f Re  /  f Re c
St Stanton number
Greek symbols V* Volume ratio, V / Vc
 Temperature difference, Tw  T W* Dimensionless mass flow rate, W / Wc
Manuscript received and revised June 2013, accepted July 2013 Copyright © 2013 Praise Worthy Prize S.r.l. - All rights reserved
265
i* Dimensionless inlet temperature difference, ducted flows, particular in the laminar regime.
i / i,c In most of the heat exchangers in service, especially in
Dimensionless outlet temperature difference, shell and tube type, circular duct is used. The primary
o* objective of the heat exchanger designer is to use duct
o / o,c
geometries that yield:
 Shape factor, p / Dh (i) a high value of heat transfer area to volume ratio,
* Ratio of shape factors,  /  c (ii) a high value of heat transfer coefficient,
(iii) a corresponding value of friction factor.
* Ratio of heat exchanger effectiveness,  /  c
Offering maximum compactness, i.e., highest surface
 Dimensionless temperature difference, area to volume ratio, however, is not enough for selection
T / T of duct geometry. A designer should also take into
o Irreversibility distribution ratio consideration the overall thermal hydraulic behavior of
Tm* Dimensionless mean temperature difference, the flow through the ducts.
Tm / Tm,c The performance of conventional heat exchangers, for
single-phase flows in particular, can be substantially
improved by many augmentation techniques, resulting in
Subscripts
the design of high-performance thermal design systems.
c Circular tube
Heat transfer augmentation devices are commonly
f Fluid
employed to improve the performance of an existing heat
i Value at x  0
exchanger or to reduce the size and cost of a proposed
m Mean heat exchanger. On the basis of the first-law analysis
o Value at x  L Webb [11] and Webb and Bergles [12] have proposed
w Wall performance evaluation criteria (PEC) that define the
performance benefits of an exchanger having augmented
I. Introduction surfaces, relative to standard exchanger with smooth
surfaces subject to various objectives and design
The importance of the non-circular ducts to the constraints.
practical design of a variety of compact heat exchangers On the other hand, it is well established that the
cannot be overstated. Because of size and volume minimization of the entropy generation in any process
constraints in applications for aerospace, nuclear, leads to the conservation of useful energy. A
biomedical engineering and electronics, chemical and thermodynamic basis to evaluate the merit of
process industries, it may be required to use non-circular augmentation techniques by second-law analysis has
flow-passage geometries. There has been considerable been proposed by Bejan [13], [14] who developed the
work on laminar forced-convective heat transfer in non- entropy generation minimization (EGM) method.
circular ducts reported in the literature. The method has been extended by Zimparov [15], [16]
Shah and London [1] and Shah and Bhatti [2] give including the effect of fluid temperature variation along
extended reviews of a large number of these studies. In the length of a tubular heat transfer unit, and new
the more recent literature, several different flow cross- information has been added assessing two objectives
section geometries for newer compact heat exchanger simultaneously. The EGM method combined with the
applications have been studied. They include double-sine first law analysis provides the most powerful tool for the
[3], circular segment [4], semi-circular [5] and several analysis of the thermal performance of any augmentation
other unusual duct shapes. Duct geometries as single- technique.
and double-trapezoidal (hexagonal) represent flow In many instances, the designer is faced with existing
channels of a variety of compact heat exchangers. equipment, where the space occupied by the cooling
The double-trapezoidal duct shape is encountered in passage is at a premium and the heat and mass flow rates
lamella type compact heat exchanger, which find are limited by the size of an existing or retrofit pump or
extensive usage in chemical industry [6]-[8]. Plate heat fan.
exchangers are also used in a wide range of applications In these situations, where a coolant passage must be
including food and chemical processing, refrigeration, designed so that the volume of the passage is restricted to
and waste-heat recovery [9]. The single-trapezoidal some value and the heat rate and mass flow rate of the
channel is employed in plate-fin heat exchangers [8], and coolant are dictated by the available equipment, the
micro-channel electronic cooling modules [10]. Due to designer may ask the question: Is there an optimal cross-
smaller system dimensions, the hydraulic diameter of sectional area and optimal length for the coolant passage
flow channels in such heat exchangers are small and the that minimizes entropy generation and allows for the best
length-to-diameter ratio, L / Dh is relatively large. performance? A number of studies have been focused on
Due to these length scales and the viscous nature of the calculation and minimization of entropy generation in
the fluids being handled, the flow is usually laminar with fully-developed laminar flows in non-circular ducts [17]-
fully developed conditions. It is therefore important to [25].
investigate the performance characteristics of different The objective of all these studies is the minimization
Copyright © 2013 Praise Worthy Prize S.r.l. - All rights reserved International Review of Chemical Engineering, Vol. 5, N. 4
266
of entropy generation of the flow in single duct by has been used as a reference heat transfer unit. Using the
optimization of the regime and geometrical parameters. first and second laws simultaneously, the performance
Circular ducts are generally used in tubular heat characteristics of units with non-circular ducts have been
exchangers and most of the augmentation techniques evaluated for different objectives and constraints
have been applied to such tubes. As to the ducts with imposed and compared to those of the reference unit with
non-circular cross section the improvement of the circular tubes.
performance of one geometry with respect to another
depends on the duct geometry, inlet-to-wall temperature y
d    c / 2b 
ratio and the operating Reynolds number for a given fluid
and a certain duct length. h y
Most of the analyses related to non-circular cross c
sections (namely, square, equilateral triangle,  x
rectangular, sinusoidal, etc.) and with single-phase
b
laminar flow have been conducted on the basis of the
(a)
second law analysis.
In a recent paper, Chakraborty and Ray [26] have y
evaluated the performance of a square duct with rounded d   c / b 
corners, for single-phase laminar flow using the PEC
h y
identified by Webb [11] and Webb and Bergles [12]
taken as objective functions for each case, and second c
law analysis in an attempt to find out an optimal  x
operating point, i.e., a particular radius of curvature for b
the corners, which is advantageous from both first and
second law analyses. In this paper, for the first time
performance characteristics of a non-circular duct have
(b)
been compared to those of the circular duct using PEC.
This study is further evidence of the statement that the Figs. 1. Coordinate system and geometrical details of:
first and second law analyses should be used (a) trapezoidal duct and (b) hexagonal duct
simultaneously to assess the thermal performance of any
heat transfer unit. For any duct with non-circular shape,
the size is determined by either the hydraulic diameter II. Equations Based on the Entropy
Dh , or the cross-sectional area A f , since these Production Theorem
parameters are related through the shape factor Consider the energy and entropy balance of the general
  4 A f / Dh2 . In this regard, two different common internal flow configuration, where fluid flows through a
duct with a cross sectional area A f , a perimeter p , and
constraints can be imposed – specified cross-sectional
area A*f  1 , or specified hydraulic diameter of the ducts, hydraulic diameter Dh  4 A f / p . The shape of the cross
D*  1 . Performance evaluation of laminar fully- section is arbitrary but constant over the entire length of
the duct. The flow is single-phase, fully developed,
developed flow in ducts with non-circular shapes incompressible, and Newtonian. Following the model
subjected to H1 boundary condition and common developed by Zimparov [15], for fully-developed laminar
constraint A*f  1 have been recently presented in [27], flow in a tube bundle, the rate of entropy generated in the
[28]. The results of performance evaluation criteria for heat unit can be expressed as:
fully-developed laminar flow through bundle of
Q o 8W  f Re   Lt
2
rectangular, isosceles triangular and elliptical ducts, S gen   (1)
subjected to constant wall temperature boundary Nt TiTo N t2  2 pDh3 Tw
condition and common constraint A*f  1 can be found
in [29]. The rationale of the present study is to evaluate where:
the thermal performance of laminar fully-developed flow
in a bundle with different non-circular ducts as Q  Nt Qt , W  m t N t , A  pLt N t , o  i exp   NTU 
rectangular, isosceles triangular, elliptical, trapezoidal
and double trapezoidal (hexagonal). Figs. 1 present the The first and second term on the right-hand side of Eq.
geometrical details of trapezoidal and double trapezoidal (1) represent the entropy generation due to heat transfer
(hexagonal) duct [30]. The boundary condition is across the finite temperature difference and due to fluid
constant wall temperature with a common constraint, friction, respectively. Following Bejan [13], [14], the
D*  1 . In this case, the cross- sectional area of the duct augmentation entropy generation number N S can be
presented as:
is a consequence, A*f  * . The bundle of circular tubes
267
Sgen N S ,T  o N S ,P III. Performance Evaluation Criteria

NS   
S gen,circle 1  o The performance evaluation criteria, as suggested by
(2)
1 Webb [11] and Weeb and Bergles [12] and extended by

1  o
 N S ,T  o N S ,P  Zimparov [15] have been considered in this study.
These criteria are based on the use of first and second
law analyses in the pursuit of two objectives
where [15]: simultaneously.
In this study, the geometrical and regime parameters
Q* o To,c Q** of the reference channel (smooth circular tube) are
N S ,T   (2a)
N* o,c To N* To* selected to fulfill the requirement of
4 Lc /  Dc Re Pr   1 , corresponding to the fully-
T Q  T  developed laminar flow in tube.
To*   i,c  * 1  i,c   (2b) The performance characteristics of the bundles with

 To,c W*  To,c   non-circular tubes, such as rectangular, isosceles
triangular, elliptical, trapezoidal and hexagonal shapes
 are compared to the reference heat unit with circular
 St L*  
o*  exp  NTU c 1   (2c) tubes. The values of the shape factor  , friction factor
  Stc D*   f and Nusselt number Nu of non-circular ducts are
taken and calculated from Shah and London [1] and
W*2 L*  f Re * Tw,c Sadasivam et al. [30]. While obtaining the augmentation
N S ,P   P* (2d)
N*2 p* D*3 Tw entropy generation number, the irreversibility
distribution ratio for the circular configuration, o ,
8m c  Lc Ti,cTo,c varies in the range 103  o  103 ( 0.001  Be  0.999 ).
o  f Re c (2e)
 2 Dc4 c p o,c To  Ti c Tw
III.1. Fixed Geometry Criteria (FG)
The numerical value of the irreversibility distribution
These criteria involve a replacement of circular tubes
ratio o  Sgen,P / Sgen,T ,
 c
describes the by tubes with non-circular shape of equal length and
thermodynamic mode in which the circular tube passage cross sectional area. The FG-1 cases seek increased heat
is meant to operate. This mode can be expressed also by duty for constant exchanger flow rate and heat transfer
area. The FG-2 criteria have the same objective as FG-1,
the value of Bejan number, Be  S / S or gen,T 
gen c but requires that the non-circular tube design to operate
Be  1 / 1  o  . The PEC as suggested by Webb [11] at the same pumping power as the reference circular tube
design.
and Webb and Bergles [12] characterize nearly all the
PEC. Some of them will be considered below, namely,
FG-1b, FG-2b, FN-1, and FN-2. The other cases as FG- III.1.1. Case FG-1b
1a, FG-2a, VG-1 and VG-2a have already been The objective functions of the case FG-1b are lower
presented in [31]. The equations are developed for ducts
with different cross sectional shape and the equations (in inlet temperature difference i*  1 , decreased entropy
dimensionless form) for inside fully developed laminar generation number N S  1 , and simultaneous effect of
flow are: both of them as general performance criterion N S i*  1 .
A*  p* L* N*  * D* N* L* (3) The constraints imposed are: Q*  1 , W*  1 , A*  1 ,
L*  1 , D*  1 . The consequences of these constraints
W*2 L*  f Re * are the same as in case Fig. 1(a).
P*  (4) The Eqs. (2)-(5) yield N*  *1 , P*  * f* , To*  1 ,
N*2 * D*4
o*  exp 3.657 1  Nu*   , and:
Q*  W* * Ti* (5)
i*  *1 (8)
W*  Re* p* N*  Re* * D* N* (6)
where *  1.0265 1  exp   Nu   , N S ,T  *o* ,
where:
*  p* / D* N S ,P  P*  *  f Re * . The augmentation entropy
(7)
generation number becomes:
268
i FG-1b
1 
NS  *o*  o *  f Re *  (9)
1  o  1.0
4
The calculated values of i* by Eq. (8) for θ  30 °

45 °
rectangular, isosceles triangular and elliptical ducts are 60 °
75 °
shown in Fig. 2(a). As seen, the bundle with isosceles
1
triangular or elliptical ducts, Fig. 2(a), cannot reach the
first objective i*  1 , whereas the bundle with
rectangular ducts can achieve some benefit of order
i*  0.975 , if *  1.5 . Consequently, the goal i*  1
.96
could be achieved only for rectangular ducts with higher 1 5 9 
values of the aspect ratio, *  1.5 . Figs. 2(b) and 2(c)
represent the variation of i* for bundle with trapezoidal
*
Fig. 2(b). The variation of i with * for trapezoidal ducts
and hexagonal ducts.
As seen, the performance characteristics of bundle  i FG-1b
with trapezoidal ducts, Fig. 2(b), and hexagonal ducts,
1.01
Fig. 2(c), are similar. For smaller values of * ( *  5
for trapezoidal ducts and *  2.5 for hexagonal ducts),
1 θ  30 °
45 °
i* depends on * and  and the greater is  the 60 °
75 °
smaller is i* (the best performance with   75o ). .99
When * increases, i* achieves almost constant

.98
value of i*  0.975 . If the bundle is equipped with
trapezoidal ducts the smallest value of i* is achieved for .97
1 5 9 
*  5 , whereas for the bundle with hexagonal ducts it is
reached for *  3 . In this respect, the bundle with *
Fig. 2(c). The variation of i with * for hexagonal ducts
elliptical ducts behaves the same way as reference unit
with circular tubes whereas the bundle with isosceles Figs. 3 present the variation of the general criterion
triangular ducts behaves remarkably worse, i*  1.0 , N S i* with * and o as a parameter for the bundle
and the value of i* increases with the increase of * . with rectangular ducts, Fig. 3(a), isosceles triangular
The use of bundle with rectangular, trapezoidal or ducts, Fig. 3(b), and elliptical ducts, Fig. 3(c). The
hexagonal ducts can improve the performance of the unit conclusions that can be derived from these figures are as
in the order i*  0.975 . follows:
(i) The design of a heat unit as a bundle of ducts with
isosceles triangular shape, Fig. 3(b), is completely
i* FG-1b
inefficient since N S i*  1 for all values of * and
o and this performance becomes worse with the
Isosceles
Rectangular increase of * and decrease of o .
Elliptical
1.2 (ii) The design of a heat unit with elliptical ducts, Fig.
3(c), is also inefficient since N S i*  1 for the values
of *  1 , and this inefficiency increases with the
increase of * and o .
1
(iii) The performance of a heat unit with rectangular
ducts, Fig. 3(a), and *  1.55 does not depend on
the value of o : all curves intersect in one point.
.8
1 3 5  The benefit, N S i*  1 can be only achieved if
*  2 and o  101 ( 0.9  Be  1 ). For o  1
*
Fig. 2(a). The variation of i with * for rectangular, isosceles ( 0  Be  0.5 ) the circular tube configuration
triangular, and elliptical ducts performs better. The benefit increases with the
269
increase of * and decrease of o . The heat unit N S i* with * and o as a parameter, for the bundle
with *  3 (aspect ratio a / b  8 , Ref. [1]) with trapezoidal ducts and different values of  ,
realizes the greatest efficiency.   30o , Fig. 4(a),   45o , Fig. 4(b),   60o , Fig.
4(c), and   75o , Fig. 4(d).
N Si* o  103 FG-1b
101
100
101 N Si FG-1b
103
3
3
θ  30 °
2 o  103
101
1 100
101
103
1
0
1 2 3 
* 0
Fig. 3(a). The variation of the general criterion N S i 1 10 
with * and o for rectangular ducts
Fig. 4(a). The variation of the general criterion N Si* with *
102
FG-1b and o for trapezoidal ducts (   30o )
N Si* o  103
101
100 N Si FG-1b
101
103
3
101
2 θ  45 °
o  103
101
100
101
103
1
100
1 2 3 4 5 
Fig. 3(b). The variation of the general criterion N S i

* 0
1 10 
with * and o for isosceles triangular ducts
Fig. 4(b). The variation of the general criterion N Si* with *
and o for trapezoidal ducts (   45o )
N Si* FG-1b
N Si FG-1b
o  103
10 1
3 100
101 3
103
2
2
θ  60 °
o  103
10 1
100
1 101
1 103
0
1 1.5 2 2.5  0
1 10 
*
Fig. 3(c). The variation of the general criterion N S  i
with * and o for elliptical ducts Fig. 4(c). The variation of the general criterion N Si*
with * and o for trapezoidal ducts (   60o )
Figs. 4 present the variation of the general criterion
270
N Si FG-1b   30o , N S i*  0.335 ; for   45o ,

N S i*  0.269 ; for   60o , N S i*  0.246 ; for
3
  75o , N S i*  0.26 , i.e., the benefit is greater
compared to that for a bundle with trapezoidal ducts,
2 θ  75 °
and almost one and the same despite the value of  .
o  103
101
100
101 N Si FG-1b
1 103
0 2
1 10 
Fig. 4(d). The variation of the general criterion N Si*

θ  30 °
with * and o for trapezoidal ducts (   75o ) o  103
1 101
100
101
The conclusions that can be made from these images 103
are as follows:
(i) All units with trapezoidal ducts are inefficient when
0
o  101 despite the values of * and  . Benefits 1 10 
can be only obtained for small values of o

*
Fig. 5(a). The variation of the general criterion N S i
( o  101 ), and the smaller the value of o the
o
with * and o for hexagonal ducts (   30 )
greater is the benefit.
(ii) When the benefit is available ( o  101 ), the curve
N Si FG-1b
of general criterion N S i* has a minimum and an
optimal value *,opt exists (the maximum
thermodynamic efficiency). With the increase of  , 2
the value of *,opt decreases (the point moves to the
left), and the benefit increases.
θ  75 °
For instance ( o  103 ), for   30o , *  11 , and 1
o  103
101
100
N S i*  0.67 ; for   45o , *  8.3 , and 101
103
N S i*  0.49 ; for o

  60 , *  6.7 , and
N S i*  0.40 ; for o
  75 , *  5.7 , and 0
1 10 
N S i* 0.35 .
(iii) All curves of the general criterion intersect in one Fig. 5(b). The variation of the general criterion N Si* with *
point. With the increase of  , the place of this point and o for hexagonal ducts (   75o )
moves to the left to the smaller values of * .
Figs. 5 present the variation of the general criterion
III.1.2. Case FG-2b
N S i* with * and o as a parameter at different values
of  for the bundle with hexagonal ducts. The objective functions of the case FG-2b are lower
If the results are compared with those obtained for the inlet temperature difference i*  1 , decreased entropy
bundle with trapezoidal ducts, one can see, that the generation number N S  1 , and simultaneous effect of
images are similar and the conclusions look like the
same. the both of them N S i*  1 .
However, there are two important differences: The constraints imposed are: Q*  1 , P*  1 , A*  1 ,
(i) The value of *,opt almost does not depend on the L*  1 , D*  1 . The consequences of these constraints
value of  ; for example: for o  103 , *,opt  5 . are, N*  1 , W*  1 , A*f ,tot  1 , and V*  1 .
(ii) The benefit is greater: the general criterion The Eqs. (2)-(5) yield N*  *1 ,
N S i* ( *,opt , o  103 ) has the next values: for 1 / 2
W*  Re*   *  f Re *  , and:
271
1/ 2
 *  f Re *  i FG-2b
i*  (10)
* 4
where:
3

  
*  1.0265 1  exp   Nu  *  f Re * 
1/ 2
  2 θ  30 °
45 °
60 °
1/ 2 75 °
To*  0.981  0.019  *  f Re *  1
 
o*  exp 3.657 1  Nu*  *  f Re * 
1/ 2
 0
1 6 11 
N S ,T  *o* To* 1 , N S ,P  P*  1 Fig. 6(b). The variation of i* with * for trapezoidal ducts
The augmentation entropy generation number N S

The variation of the general criterion N S i* with *
becomes:
and o as a parameter for bundles with different non-
1 circular ducts are presented in Figs. 7.
NS 
1  o

*o* To* 1  o  (11)
N Si FG-2b
The calculated values of i*

by Eq. (10), are shown in 2
Figs. 6. As seen, the use of bundles with isosceles
triangular, elliptical or rectangular ducts, Fig. 6(a),
cannot reach the first objective i*  1.0 , namely, for
bundle with rectangular ducts i*  1.063  1.566 , for 1
isosceles triangular ducts i*  1.158  1.763 , and for o  103
101
elliptical ducts i*  1.0  1.921 . The bundles with 100
101
trapezoidal, Fig. 6(b), or hexagonal ducts (similar image 103
like Fig. 6(b)) perform even worse. For all bundles 0

1 2 3 
considered i*  1.0 and the greater is * the greater is
i* . The reason for that is the constraint P*  1 that Fig. 7(a). The variation of the general criterion N S i
*
requires a substantial decrease of the mass flow rate with * and o for rectangular ducts
W*  1 and number of tubes, N*  1 .
N Si FG-2b
o  103
i FG-2b 101
100
101
103
2 101
Isosceles
1 Rectangular
Elliptical
100
1 2 3 4 5 
1 2 3 4 5  *
Fig. 7(b). The variation of the general criterion N S i with * and o
for isosceles triangular ducts
*
Fig. 6(a). The variation of i with * for rectangular,
isosceles triangular, and elliptical ducts
272
N Si FG-2b entropy generation number N S caused by the substantial

o  103
101 decrease of the mass flow rate W*  1 . Consequently, in
100
101 the case FG-2a, no benefits should be expected if the
2 103
standard circular pipes in the bundle are replaced with
the non-circular ducts in the search of better
thermodynamic performance.
1
III.2. Fixed Number of Tubes Criteria (FN)
These criteria maintain constant number of tubes. The
objective of FN-1 is reduced surface area, via reduced
0
1 2 3 4 5  tubing length, for constant pumping power. Reduced
flow rate will probably be required to satisfy the constant
Fig. 7(c). The variation of the general criterion N S i
*
z with * pumping power criterion. The objective of FN-2 is
reduced pumping power.
and o for elliptical ducts
N Si FG-2b III.2.1. Case FN-1
2
The objective functions of the case FN-1 are reduced
heat transfer area A*  1 ( L*  1 ), decreased entropy
generation number N S  1 , and simultaneous effect of
θ  60 °
o  10 3 the both of them N S A*  1 . The constraints imposed are:
10 1
1
100
101 N*  1 , Q*  1 , P*  1 , i*  1 , and D*  1 . The
103
consequences of the constraints are A*f ,tot  1 , W*  1 ,
Re*  1 . The Eqs. (2)-(5) yield:
0
1 10  W*  * Re*  *1 (12)
*
Fig. 7(d). The variation of the general criterion N S i    2 2 
*  1.0265 1  exp   Nu * *  (13)
with * and o for trapezoidal ducts (   60 )
o
   f Re * 
 
N Si FG-2b
* *2
L*  (14)
2  f Re *
2
θ  75 ° A*  * L* 
  
* *
(15)
o  103
1
101 f Re *
100
101
103
Eq. (13) is to be solved by an iterative procedure for
any * :
0
1 10  To*  0.981  0.019*
*
Fig. 7(e). The variation of the general criterion N S i    2 3  
o*  exp 3.657 1  Nu* * *  
with * and o for hexagonal ducts (   75 )
o
   f Re *  
It seems that in some cases, o  1 , the use of bundles N S ,T  o* To* 1 , N S ,P  P*  1

with rectangular, elliptical, trapezoidal or hexagonal
ducts could be beneficial since N S i*  1 . In these cases, The augmentation entropy generation number N S
however, the small values of N S i*  1 are due to the becomes:
small values of the entropy generation and augmentation
273
1  o*  As seen, it is not possible to decrease the heat transfer

NS   *  o  (16) area A*  1 ( L*  1 ) of the unit with the non-circular
1  o  To  channels used in this study. Figs. 9 present the variation
of N S A* with * and o as a parameter for all bundles
The calculated values of A* by Eq. (15) are shown in
with non-circular ducts. As seen, for o  1 ( Be  0.5 ),
Figs. 8.
N S A*  1 whereas for o  1 ( Be  0.5 ), N S A*  1 and
A
FN-1 the values substantially decrease with the decrease of o .
The second very interesting fact is that the curves of
8
N S A* show the presence of minimum which move to
the left (lower values of * ) with the decrease of o .
6
This trend of behavior is due to the small values of the
entropy generation and augmentation entropy generation
4 number N S caused by the substantial decrease of the
Isosceles
mass flow rate W*  1 and domination of entropy
2 Rectangular
Elliptical
generation due to heat transfer at final temperature
difference.
0
1 2 3 
N S A o  103 FN-1
Fig. 8(a). The variation of heat transfer surface area 101
100
with * for rectangular, isosceles triangular and elliptical ducts 101 101
103
A FN-1
102 10-1
θ  30 °
101 45 °
60 °
75 ° 10-3
1 2 
100 Fig. 9(a). The variation of the general criterion N S A*

1 10 
Fig. 8(b). The variation of heat transfer surface area
with * for trapezoidal ducts N S A FN-1
A FN-1 101
o  103
101
102 100
101
10-1 103
101
θ  30 °
45 °
60 °
75 ° 10-3
1 2 3 4 5 
100
1 10 
Fig. 9(b). The variation of the general criterion N S A*
Fig. 8(c). The variation of heat transfer surface area
with * and o for isosceles triangular ducts
with * for hexagonal ducts
274
N S A FN-1 pumping power P*  1 , decreased entropy generation

o  103
101
number N S  1 , and simultaneous effect of the both of
100
101
them N S P*  1 .
103
101
The constraints imposed are: Q*  1 , A*  1 , i*  1 ,
N*  1 , D*  1 . The consequences of the constraints are
W*  1 , L*  1 , A*f ,tot  1 .
10-1
The Eqs. (2)-(5) yield L*  *1 , A*f  * ,
1
Re*   * *  , W*  *1 :
10-3
1 2 3  *  1.0265 1  exp   Nu*   (17)
Fig. 9(c). The variation of the general criterion N S A*  f Re *

P*  (18)
with * and o for elliptical ducts
 * * 2
N S A FN-1
o*  exp 3.657 1  Nu* *   , To*  0.981  0.019*
102
 f Re *
101 N S ,T  o* To* 1 , N S ,P  P*  .
 * * 2
100
Eq. (17) is to be solved by iterative process for any
10 -1
θ  30 ° 3
Nu . The augmentation entropy generation number N S
o  10
101 becomes:
100
10-2 101
103
1  *  f Re * 
NS   o*  o  (19)
10-3
0 10 20 30 
1  o  To  * * 2 
Fig. 9(d). The variation of the general criterion N S A* The calculated values of P* by Eq. (18) are presented
with * and o for trapezoidal ducts in Figs. 10, for rectangular, elliptical and isosceles
triangular ducts, Fig. 10(a), and trapezoidal ducts, Fig.
N S A FN-1 10(b).
102
P FN-2
1
10
.8
100
Isosceles
Rectangular
Elliptical
10-1
θ  75 ° 3
o  10 .4
10-2 101
100
101
103
10-3
0 5 10 15 
0
Fig. 9(e). The variation of the general criterion N S A* 1 2 3 4 5 
with * and o for hexagonal ducts
Fig. 10(a). The variation of the pumping power P*
with * rectangular, isosceles triangular and elliptical ducts
III.2.2. Case FN-2
The objective functions of the case FN-2 are lower
275
100 transfer performance of the unit. The difference between

P FN-2
the use of trapezoidal or hexagonal ducts is that if the
bundle uses trapezoidal ducts its performance depends on
 and o in a different way compared to that of the
bundle with hexagonal ducts.
θ  30 °
45 ° N S P FN-2
10-1 60 °
75 °
o  103
101
100
101
103
10-1
10-2
1 5 9 
Fig. 10(b). The variation of the pumping power P*

with * for trapezoidal ducts
10-2
1 2 3 
The variation of P* with * for hexagonal ducts is
the same as that for trapezoidal ducts. The calculated Fig. 11(a). The variation of the general criterion N S P*
values of P* by Eq. (18), Figs. 10, revealed that P*
varied as follows:
For rectangular ducts: N S P FN-2
P*  0.474  0.036 o  10 3
100 10 1
100
For isosceles triangle: 101
103
P*  0.236  0.019 10-1
For elliptical channels:

10-2
P*  1.0  0.005
For trapezoidal and hexagonal ducts:

10-3
1 2 3 
P*  0.889  0.00135
As seen, bundle with any of the non-circular ducts Fig. 11(b). The variation of the general criterion N S P*
studied can diminish the pumping power required and the with * and o for elliptical ducts
higher is * the greater is benefit. The variations of
N S P* with * and o as a parameter are presented in N S P FN-2
Figs. 11. As seen, for bundle with rectangular ducts, Fig.

11(a), N S P*  1 for *  1 , despite the value of o , and 100
substantially decreases with increasing of * and o .

The same tendency is observed for bundle with
10-1
elliptical ducts, Fig. 11(b).
In this case, however, the benefit stronger depends on
the value of o and it increases with the increase of o . o  10 3
10-2 10 1
For bundle with isosceles triangular ducts, Fig. 11(c), the 100
101
benefit can be only obtained if o  1 and it increases 103
with the increase of * and o . Figs. 12 and 13 present 10-3

1 2 3 4 5 
the variations of N S P* with * and o for bundle with
trapezoidal and hexagonal ducts. The use of these non- Fig. 11(c). The variation of the general criterion N S P*
circular ducts is also beneficial to improve the heat with * and o for isosceles triangular ducts
276
N S P FN-2 N S P FN-2
100 100
10-1 θ  30 °
10-1 θ  75 °
o  10 3 o  103
10 1 101
100 100
101 101
10-2 103 103
10-2
10-3
10-3
10-4
1 5 9  10-4
1 5 9 
Fig. 12(a). The variation of the general criterion N S P*
Fig. 13(b). The variation of the general criterion N S P*
o
with * and o for bundle with trapezoidal ducts (   30 ) o
with * and o for bundle with trapezoidal ducts (   75 )
N S P FN-2
IV. Discussion and Conclusions
10 0
Extended performance evaluation criteria (ExPEC)
have been used to assess the performance characteristics
of single-phase fully developed laminar flow through
10-1 θ  75 °
bundle of ducts with non-circular shape such as
o  103
101 rectangular, isosceles triangular, elliptical, trapezoidal
100
101 and hexagonal under different objectives and constraints
10-2 103
imposed.
These performance characteristics have been
10-3 compared to those of the heat unit with circular tubes.
The constant wall temperature has been selected as
thermal boundary condition. The results can be
10-4 summarized as follows:
1 5 9 
(i) For the case FG-1b, the bundle with isosceles
Fig. 12(b). The variation of the general criterion N S P* triangular or elliptical ducts cannot reach the first
o objective i*  1 , whereas the bundle with rectangular
with * and o for bundle with trapezoidal ducts (   75 )
ducts can achieve some benefit of order i*  0.975 , if
*  1.5 . The use of bundle with trapezoidal or
N S P FN-2
hexagonal ducts can also improve the performance of the
100
unit in the same order, i*  0.975 .
The design of a heat unit as a bundle of ducts with
isosceles triangular shape or elliptical ducts is completely
10-1 θ  30 °
o  103 inefficient since N S i*  1 for all values of * and o
101
100 and this performance becomes worse with the increase of
101
10-2 103 * and decrease of o .
For heat unit with rectangular ducts the benefit,
10-3 N S i*  1 can be only achieved if *  2 and o  101 .
For o  1 , the circular tube configuration performs
10 -4 better. The benefit increases with the increase of * and
1 5 9 
decrease of o . The heat unit with *  3 (aspect ratio
Fig. 13(a). The variation of the general criterion N S P* a / b  8 , Ref. [1]) realizes the greatest efficiency.
o The bundle with trapezoidal ducts are inefficient when
with * and o for bundle with hexagonal ducts (   30 )
o  101 despite the values of * and  .
277
Benefits can only be obtained for small values of o ducts, in: S. Kakaç, R.K. Shah, W. Aung (Eds.), Handbook of
Single-Phase Convective Heat Transfer, Wiley, New York, 1987
1
( o  10 ), and the smaller the value of o the greater is (Chapter 3).
[3] R.M. Manglik, J. Ding, Laminar flow heat transfer to viscous
the benefit. The performance characteristics of bundle power-law fluids in double-sine ducts, Int. J. Heat Mass Transfer
with hexagonal ducts are similar to those of bundle with 40 (1997) 1379-1390.
trapezoidal ducts but the benefit is greater and almost one [4] R.M. Manglik, A.E. Bergles, Fully developed laminar heat
and the same despite the value of  . transfer in circular-segment ducts with uniform wall temperature,
Numerical Heat Transfer Part A 26 (1994) 499-519.
(ii) For the case Fig. 2(b), bundles with isosceles [5] T.M. Harms, M.A.Jog, R.M. Manglik, Effect of temperature
triangular, elliptical or rectangular ducts cannot reach the dependant viscosity variations and boundary conditions on fully
first objective i*  1.0 . Bundles with trapezoidal or developed laminar forced convection in a semicircular duct,
Journal of Heat Transfer 120 (1998) 600-605.
hexagonal ducts perform even worse. The reason for that [6] G. Walker, Industrial Heat Exchangers, Hemisphere, New York,
is the constraint P*  1 that requires a substantial 1990.
[7] E.A.D. Saunders, Heat Exchangers: Selection, Design and
decrease of the mass flow rate W*  1 and number of Construction, Longman, Harlow, UK, 1998.
[8] R.K. Shah, D.P. Sekulić, Fundamentals of Heat Exchanger
tubes, N*  1 . Design, John Wiley & Sons, Inc., Hoboken, New Jersey, 2003.
When the general criterion N S i* is used, it seems in [9] R.M. Manglik, Plate heat exchangers for process industry
applications: enhanced thermal- hydraulic characteristics of
some cases, o  1 , that bundles with rectangular, chevron plates, in: R.M. Manglik, A.D. Klaus (Eds.), Process,
Enhanced and Multiphase Heat Transfer, Begell House, New
elliptical, trapezoidal or hexagonal ducts could be York, 1996, pp-267-276.
beneficial since N S i*  1 . In these cases, however, the [10] S.M. Flockhart, R.S. Dhariwal, Experimental and numerical
investigation into the flow characteristics of channels etched in
small values of N S i*
 1 are due to the small values of silicon, Journal of Fluids Engineering 120 (1998) 291-295.
[11] R.L. Webb, Performance evaluation criteria for use of enhanced
the entropy generation and augmentation entropy heat transfer surfaces in heat exchanger design, Int. J. Heat Mass
generation number N S caused by the substantial Transfer 24 (1981) 715-726.
decrease of the mass flow rate W*  1 . Consequently, in [12] R.L. Webb, A.E. Bergles, Performance evaluation criteria for
selection of heat transfer surface geometries used in low Reynolds
the case Fig. 2(a), no benefits should be expected if the number heat exchangers, in Kakac, S., Shah, R.K., Bergles, A.E.
standard circular pipes in the bundle are replaced with (Eds) Low Reynolds Number Flow Heat Exchangers, Hemisphere
Publ. Corp., Washington, D.C., 1983, pp. 735-742.
the non-circular ducts in the search of better [13] A. Bejan, Entropy Generation through Heat and Fluid Flow,
thermodynamic performance. John Wiley & Sons, New York, 1982.
(iii) For the case FN-1, it is not possible to decrease [14] A. Bejan, Entropy Generation Minimization, CRC Press, Boca
the heat transfer area A*  1 ( L*  1 ) of the unit with the Raton, 1996.
[15] V.D. Zimparov, Extended performance evaluation criteria for
non-circular ducts used in this study. The use of the enhanced heat transfer surfaces: heat transfer through ducts with
general criterion N S A*  1 shows that for o  1 , constant wall temperature, Int. J. Heat Mass Transfer 43 (2000)
3137-3155.
N S A*  1 , whereas for o  1 , N S A*  1 and the values [16] V.D. Zimparov, Extended performance evaluation criteria for
substantially decrease with the decrease of o . This trend enhanced heat transfer surfaces: heat transfer through ducts with
constant heat flux, Int. J. Heat Mass Transfer 44 (2001), 169-180.
of behavior is due to the small values of the entropy [17] D.H Richardson, D.P. Seculic, A. Campo, Low Reynolds number
generation and augmentation entropy generation number flow inside straight micro channels with irregular cross-sections,
Heat Mass Transfer 36 (2000) 187-193.
N S caused by the substantial decrease of the mass flow [18] S.Y. Wu, Y.R. Li, Y. Chen, L. Xiao, Energy transfer
rate W*  1 and domination of entropy generation due to characteristics of forced convective heat transfer through a duct
with constant wall temperature, Energy 32 (2007) 2385-2395.
heat transfer at final temperature difference. [19] H.F. Oztop, Effective parameters on second law analysis for
(iv) For the case FN-2, bundles with any of the non- semicircular ducts in laminar flow and constant wall heat flux,
circular ducts studied can diminish the pumping power Int. Comm. Heat Mass Transfer 32 (2005) 266- 274.
[20] H. F. Oztop, I. Dagtekin, A.Z. Sahin, Second law analysis of fully
required and the higher is * the greater is benefit. The developed laminar flow for rectangular ducts with semicircular
variations of N S P* with * and o as a parameter also ends, Int. Comm. Heat Mass Transfer 36 (2009) 725-730.
[21] D.P. Sekulic, A. Campo, J.K. Morales, Irreversibility phenomena
show that the use of these non-circular ducts is associated with heat transfer and fluid friction in laminar flows
beneficial. through singly connected ducts, Int. J. Heat Mass Transfer 40
As expected, the analysis of the thermal performance (1997) 905-914.
[22] E.B. Ratts, A.G. Raut, Entropy generation minimization of fully
of heat units with non-circular ducts revealed that the developed internal flow with constant heat flux, ASME J. Heat
selection of the optimal shape of the unit ducts strongly Transfer 126 (2004) 656-659.
depends on the geometric and thermal-hydraulic [23] T. A. Jankowski, Minimizing entropy generation in internal flows
constraints, and the objectives pursued. by adjusting the shape of the cross section, Int. J. Heat Mass
Transfer 52 (2009) 3439-3445.
[24] S. Jarungthammachole, Entropy generation analysis for fully
developed laminar convection in hexagonal duct subjected to
References constant heat flux, Energy 35 (2010) 5374 - 5379.
[1] R.K. Shah, A.L. London, Laminar Flow Forced Convection in [25] V.D. Zimparov, V.M. Petkov, Second law analysis of fully
Ducts, 1978, Academic Press. developed internal flows with different shape of the cross-
[2] R.K. Shah, M.S. Bhatti, Laminar convective heat transfer in
278
section, (2011) International Review of Chemical Engineering Authors' information

(IRECHE), 3 (2), pp. 256-264.
[26] S. Chakraborty, S. Ray, Performance optimization of laminar Technical University of Gabrovo,
fully developed flow through square ducts with rounded corners, 4 Hadji Dimitar, 5300 Gabrovo, Bulgaria.
Int. J. Therm. Sciences 50 (2011) 2522-2535.
[27] V.M. Petkov, V.D. Zimparov, A.E. Bergles, Performance Valentin Petkov is an Assistant Professor of
Evaluation of Laminar Fully Developed Flow through Ducts With thermodynamics and heat transfer at Gabrovo
Non-Circular Shapes Subjected to H1 Boundary Condition. Part Technical University, Bulgaria. He received his
1, (2013) International Review of Chemical Engineering BSc and PhD degrees from Gabrovo Technical
(IRECHE), 5 (1), pp. 1-15. University. His main research interests are
[28] V.M. Petkov, V.D. Zimparov, Performance Evaluation of enhanced heat transfer, compact heat
Laminar Fully Developed Flow through Ducts With Non-Circular exchangers, and thermodynamic optimization.
Shapes Subjected to H1 Boundary Condition. Part 2, (2013) E-mail: petkov_v@tugab.bg
International Review of Chemical Engineering (IRECHE), 5 (2),
pp. 74-87. Ventsislav Zimparov is a Professor of
[29] V.M. Petkov, V.D. Zimparov, A.E. Bergles, Performance thermodynamics and heat and mass transfer in
evaluation of ducts with non- circular shapes and laminar fully Gabrovo Technical University, Bulgaria. He
developed flow, (2012) International Review of Chemical received his PhD in 1990 from Technical
Engineering (IRECHE), 4 (4), pp. 379-391. University of Sofia, Bulgaria and DSc degree in
[30] R.S. Sadasivan, Raj M. Manglik, M.A. Jog, Fully developed 2004. Head of the Division of Fluid Power and
forced convection through trapezoidal and hexagonal ducts, Int. J. Heat Transfer Engineering. His main research
Heat Mass Transfer 42 (1999) 4321-4331. interest are enhanced heat transfer, compact heat
[31] V.M. Petkov, V.D. Zimparov, A.E. Bergles, Performance exchangers, heat exchanger networks, constructal theory and
evaluation of ducts with non- circular shapes: Laminar fully thermodynamic optimization. He has published more than 80 articles in
developed flow and constant wall temperature, Int. J. Thermal well-recognized journals and proceedings.
Sciences (2013) (submitted). E-mail: ventsi.zimparov@gmail.com
279
ISSN 2035-1755 July 2013
Performance Evaluation of Photo-Fenton and Fenton Processses

for Dairy Effluent Treatment
Carla C. A. Loures, Hélcio J. Izário. Filho, Gisella R. Lamas Samanamud, André L. Souza,
Rodrigo F. S. Salazar, André L. C. Peixoto, Oswaldo L. C. Guimarães
Abstract – This study aimed to evaluate the efficiency of Photo-Fenton and Fenton processes in
reducing organic matter of dairy effluent. An Orthogonal Array L9 Taguchi was used to determine
optimal conditions of acidity media, temperature, Fenton concentration and UV radiation
intensity. Reaction time was set up to 60 min. Optimized parameters were: pH 3.0, temperature,
Fenton reagent concentration and UV radiation at the highest level. The Dissolved Organic
Carbon percentage reduction (DOC) was 91 %. An effective degradation study was carried out, in
which, the reduction percentage was found to be less than the most efficient DOC removal. A
cost/benefit evaluation of the AOP process employed on the in natura dairy effluent treatment
showed that the reagent consumption the main cost of the process. The highest efficiency
experiment for the dairy effluent of this study had operational costs lower than US$ 0.50 to Photo-
Fenton and Fenton processes, respectively. Copyright © 2013 Praise Worthy Prize S.r.l. - All
rights reserved.
Keywords: Dairy Effluent, Photo-Fenton, Taguchi Method, Economic Viability, Efficient Organic
Degradation, Cost/Benefit Ratio
I. Introduction Among biological methods, methanization is by far

the most interesting since it transforms the organic matter
The dairy industries are associated with generation of in milk to methane, a compound well known for its
large amounts of wastewater. Water, in these industries, unquestionable combustion properties. Moreover, high
is used in all steps, such as: cleaning, sanitization, removal rates of the COD are obtained [11] even though
heating, cooling and washing floors [1]-[2]. A key this process does not eliminate phosphorus and nitrogen
feature of this sewage is bio organic matter and non- compounds contained in the dairy effluents [12].
biodegradable one due to the presence of organic Generally, biological effluent treatment processes
compounds such as: fatty acids, esters, alcohols, (aerobic and anaerobic processes) offer the lowest cost
aldehydes, ketones, amines and other nitrogen per unit volume of wastewater treated or per kg of COD
compounds [3]-[4]. removed. The unit cost is normally lower than $5 per m3
Wastewaters from agro-industries are characterized by for biological processes and higher than $10 per m3 for
high chemical oxygen demand (COD) due to their high AOP. Consequently, biological processes are the
level of organic contents [5]. The dairy industry is preferred choice of the industry for wastewater treatment
particularly concerned: as a matter of fact it generates a [13]. However, effluents containing non-biodegradable
huge amount of wastewaters: approximately 0.2 L to 10 compounds pose a challenge to biological treatment
L of waste per liter of processed milk [6]. In most cases, schemes.
these effluents are not treated and are simply thrown into Biological processes do not normally remove non-
rivers where they contribute to eutrophication by biodegradable chemicals and in some cases, high
phosphorus and nitrogen compounds [7]-[9]. concentration of such chemicals may inhibit the
Treating dairy effluents is thus of crucial importance biological processes resulting in poor performance.
not only for the environment, but also for the purpose of Non-biodegradable compounds can be recalcitrant
recycling water for use in industrial processes [10]. and/or toxic to microorganisms. The presence of such
Nowadays, many physicochemical and biological compounds requires non-biological processes for
methods are used to treat dairy effluents, with the effective elimination and AOP have such a capability.
particular aim of reducing the volume of the produced Lately, new techniques have been proposed in order to
sludge. The physicochemical processes suffer the solve some biological system limitations and improve the
disadvantage that reagent costs are high and the soluble organic matter degradation and other compounds on
COD removal is low [5]. Moreover, chemical treatments different effluents [14], [15]. Advanced Oxidation
could induce a secondary pollution due to the fact that Processes (AOPs) are very efficient technologies for
chemical additives may contaminate the treated water.
280
Carla C. A. Loures et al.
some organic compounds that are especially difficult to Finally, signal-to-noise (SN) in the Taguchi method
degrade. can be used to optimize the process and to reduce the
During the past decade, studies using AOPs have been process variability [31].
widely researched for different types of wastewater, The aim of this study consisted in evaluating the
including membranes enhanced by nanotechnologies efficiency of bench-scale reactors from dairy effluent
[16]-[20]. AOPs refer specifically to processes that can treatment by AOP – Photo-Fenton process. An economic
produce hydroxyl radicals (•OH). Due to high standard evaluation of the Photo-Fenton process was also
reduction potential (1), this radical can oxidize a variety performed to balance the relation between energy and
of organic compounds to CO2, H2O and inorganic ions reagent consumption taking into account the chemical
from hetero-atoms [21]: process only.
•OH + e- + H+ → H2O Eo = 2.730 V (1)

II. Experimental
2+
Photo-Fenton’s reaction (UV/Fe + H2O2), has II.1. Sampling and Preservation
demonstrated to be one of the most efficient process for
effluent treatment due to high oxidation power. When Fe The effluent studied comes from a dairy factory in
(III) complexes are irradiated, a central-orbit electron is Vale do Paraiba region, city of Guaratinguetá, São Paulo.
promoted and a ligand bonds to a central metal, called This sample was collected from the production line
Ligand-to-Metal Charge Transfer (LMCT), which that feeds the treatment plant before the acid
implies in the reduction of Fe (III), Fe (II) and ligand pretreatment. 400 L of raw effluent were collected and
oxidation (2), forming hydroxyl radical [22]: stocked in a cold chamber at 4 ºC.
Due to complex characteristics some methodologies
Fe(OH)2+ + hv → Fe2+ + •OH (2) were adapted to improve the accuracy and precision of
analytical results [31]. To determine COD in dairy
Fe2+ generated by irradiation in the presence of effluents, some adjustments were made to 5220 D Closed
hydrogen peroxide reacts promoting the sequence of Reflux colorimetric method of [32].
Fenton’s reaction. In the catalytic reaction, a cycle of To measure Biochemical Oxygen Demand (BOD),
regenerated Fe2+ should be established. The use of Winkler’s modified method was used. For COD and
Fe2+/Fe3+ in the presence of hydrogen peroxide under BOD determinations, preliminary tests with diluted
irradiation is called Photo-Fenton. deionized water were used due to specific concentrations
Generally, the optimum molar ratio H2O2/Fe2+ on each methodology. The dissolved organic carbon
recommended for a Fenton treatment is from 10 to 40 (DOC) determination was performed in a TOC-VCPH
[23], [24]. The hydrogen peroxide dosage is important to Total Organic Carbon Analyzer (high sensitivity) from
obtain the best degradation efficiency while the Shimadzu based on high temperature catalytic oxidation.
concentration of Ferrous ions is important to the reaction The statistic and graphic analysis were done using
kinetics [24]. However, the excess of any of these software Minitab R15. Ferrous ions mass and H2O2
compounds may decrease the degradation efficiency, estimation used in the Photo-Fenton treatment
since both H2O2 and Fe2+ can capture hydroxyls radicals To determine the preliminary mass proportion
[25]. The H2O2 / Fe2+ stoichiometric ratio depend on between H2O2 and Fe2+ of Fenton reagent, the H2O2
physical-chemicals characteristics of the wastewater. weight (30 % w/w) necessary to the radical formation
In the same way, design of experiments has been •OH as shown in (3) [16]:
widely used to optimize process parameters and to
enhance the quality of products by applying engineering C + 2 H2O2 → CO2 + 2 H2O (3)
and statistics concepts [26]. In 1987, Taguchi [27]
developed and published a statistical tool for designing- From the theoretical mass of H2O2 a ratio of
of-experiments (DOE) to meet the above requirements. H2O2/Fe2+ was estimated for each level of Fenton
The Taguchi technique provides an efficient and reagent. Seeking for a better mass proportion, some
systematic method to optimize designs for performance, preliminary experiments were conducted under the
quality and cost [28]. It has been successfully employed following conditions: pH = 3, 0; temperature = 25 ºC;
in designing reliable, high quality products at relatively reaction time = 2 h; UV irradiation = 28 W, and aliquots
low costs [29]. The advantage of employing Taguchi taken every 30 min.
technique has been summarized by [30]. Primarily, the It was verified that degradation efficiency of the
method requires limited number of experiments to pollutant linearly grows with an increase in the H2O2
conduct the experimental design. Another important dosage [33]-[34].
point is that many different variables can be examined However, in some cases, the increase in concentration
simultaneously. This means that predominant parameters of peroxide does not cause a significant increase in the
can be investigated deeply whereas, secondary degradation efficiency [29]-[30] and might even promote
parameters can be overlooked. Therefore, time, energy the reduction in the process efficiency [30], since it acts
and resources can be saved. as scavenger.
281
II.2. Degradation of Dairy Effluent proposed for this stage were: pH, temperature,
by Photo-Fenton Process concentrations of H2O2 and Fe2+ in Fenton’s reagent and
UV radiation source. Table I shows selected levels
Photochemical treatment was performed in a semi
variables per liter of effluent treated dairy.
batch reactor. Fenton’s reagent was added according to
The levels selection were based on [23]. Most studies
the following stock concentrations: 0.82 mol L-1 of
performed with Photo-Fenton showed pH of 3.0–4.0,
FeSO47H2O, H2O2 at 30 % w/w. After thermal
which is considered outstanding preventing ferric salts
conditioning of the effluent, ferrous ions, and H2O2
hydrolysis. For this exploratory experimental design, a
solutions were added at the system, simultaneously.
higher hydrogen ion concentration ratio (pH 3.0 to 5.0)
These additions were performed by dosing pump
was used to evaluate the effluent behavior under
during 50 min in 1h of reaction. Sulfuric acid and NaOH
oxidative conditions in different values of acidity.
(5.0 mol L-1, both) solutions were used to maintain the
Temperature values were selected considering two
medium acidity during all reaction time. This control was
important aspects: average local temperature (~30 ºC)
performed by a potentiometer with borosilicate glass
and, peroxide solution degradation temperature.
electrode. The electrode was kept in the reaction bath.
The levels for Photo-Fenton’s reagents were based on
The operational stages with Photo-Fenton/Fenton
preliminary studies. Low-pressure Hg lamps of 15 and 28
processes were as follows:
W were used for ultraviolet radiation, evaluating the
1°- 3.0 L of dairy effluent kept at room temperature,
radiation of Photo-Fenton process in the dairy effluent
homogenized and placed in a glass container;
characterized by high concentrations of dissolved solids
2º- Thermostatic bath and centrifugal pump were turned
and high turbidity. The variable response was given in
on;
percentage of DOC reduction, according to the Eq. (4):
3º- Temperature was adjusted according to experimental
design; % reduction = 1- DOCt/DOC0
4º- pH was regulated according to experimental records;
5º- Lamp was started to emit Ultraviolet (UV) radiation; where: DOC0 = Initial DOC;
6º- In parallel, ferrous and peroxide solution were DOCt = Treated DOC.
continuously added during 50 min of a total of 1 h
reaction, where the highest rate of addition was 2.1
ml / min and the lowest 1.81 ml / min; II.4. Study of the Effective Degradation
7º- After adjusting all parameters and setting the time of the Effluent After AOP
control for 1 hour treatment, pH was kept constant; When an AOP with a Fenton reagent is applied,
8º- Aliquots (20 mL) were taken every 10 minutes; another operation must be performed afterwards,
9º- The pH of both rates were adjusted from 8.0 to 9.0 ensuring the removal of iron ions and conditioning. This
for ferrous ions precipitation and filtered in consists of altering the pH from 8.0 to 9.0 of the treated
quantitative filter paper; solution (NaOH 5 mol L-1). Basically, this precipitation
10°-Each sample of dairy effluent was submitted to process consisted in the following steps:
determine concentrations of COD, H2O2 residual 1º- Alkaline precipitation of iron in an exact volume of
concentrations and DOC. Dilutions owing to pH treated effluent by AOP, filtration and drying of the
adjustment (AOP and precipitation) were taken into precipitate in an oven at 150 ºC for 3 h;
account to the concentration calculation in all 2º- Maceration of dry residue in a porcelain capsule until
analytical determinations. a thin and homogeneous mass was obtained, using
around 50 mg;
II.3. Experimental Design for Dairy Effluent 3º- Heat digestion of the mass with 2 mL of aqua regia
Degradation by AOP – Taguchi’s L9 Orthogonal solution and 10 mL of deionized water (ferrous ion
Array Method oxidation);
4º- Concentrated sulfuric acid and excess of K2Cr2O7 1
A factorial statistic planning was performed to eq L-1 solution added to the dissolved residual product
optimize the parameters represented by Taguchi’s L9 reacting for 30 min under heat conditions by its own
orthogonal array, in duplicate, enabling a more simple exothermic property;
and standardized method of fractional factorial 5º- Titration of the excess of dichromate (after cooling)
experiments [36]. Independent variables and factors by a Ferrous II solution, previously standardized.
TABLE I
CONTROL FACTORS AND LEVELS OF AN EXPLORATORY STUDY FOR DAIRY EFFLUENT TREATMENT
WITH THE PHOTO-FENTON PROCESS BY USING TAGUCHI’S L9 ORTHOGONAL ARRAY METHOD
Factor Symbology Level 1 Level 2 Level 3
Temperature (ºC) A 20 30 35
pH B 3.0 4.0 5.0
Fenton’s reagent C 0.255 mol H2O2 L-1 + 0.294 mol H2O2 L-1 + 0.343 mol H2O2 L-1+
+ 0.0108mol Fe2+ L-1 + 0.0143 mol Fe2+ L-1 + 0.0215 mol Fe2+ L-1
UV D Without 15 W 28 W
282
The dosage result according to the chemical confirming the increase of dairy effluent biodegradation
equivalence is expressed in mg C Kg-1, considering that through this treatment.
the dichromate had reacted only with the organic load Color and turbidity results were satisfactory, making
present in the weighted mass. The organic load obtained the visual characteristics of the treated effluent as clear
in the precipitate is subtracted from the organic load and colorless.
determined by DOC analysis from the filtrated solution They presented a removal of 95.5 % and 98.8 %,
being its result the effective organic load degraded. respectively. With the reduction of organic load and
consequently of the concentration of total solids (total
removal), the effluent did not show any characteristic
III. Results and Discussions smell of the dairy effluent. Also, there has been an
III.1. Characterization of in Natura Effluent oxidation of the nitrogenous compounds, with possibly a
NOx formation, since the concentration of nitrates and
Physical-chemical analysis of dairy effluent samples nitrites were not changed.
were conducted to COD, BOD5, DOC, BOD5/COD, pH, In all experiments by AOP, after 60 min time of
turbidity and color determinations according to [32]. As reaction, the oxidized effluent showed no residual
shown in Table II, in which the results of organic and concentration of peroxide, which shows a correct
inorganic parameters used in dairy effluent analysis after determination of H2O2 levels in the experiment. Oxidized
being submitted to photochemical treatment are effluent, in all AOP experiments and after reactional time
displayed to the following experimental conditions: pH of 60 min, did not show residual peroxide which suggests
3,0, temperature at 35 ºC, Fenton reagent in the a correct determination of H2O2 levels in the design of
proportions 0.343 mol H2O2 L-1 + 0.0215 mol Fe2+ L-1. experiment.
In general, the physicochemical results were
expressive, with significant reductions in percentages.
Regarding the organic load, there has been a DOC III.2. Delineating Experiments
degradation of 90.86%, showing a significant efficiency The delineating experiment was conducted to evaluate
in the oxidation process of the dairy effluent, even at the photocatalytic reaction of the dairy effluent relative
high concentrations of interfering ions such as chloride to the concentration of Fenton reagent (mH2O2 mFe2+),
[37], [38]. shown below. Therefore, a concentration of H2O2 (0.343
According to the Federal Law and the State of São mol L-1 ) was fixed owing to its best result for DOC and
Paulo Law, there is not a specific concentration value of Chemical Oxygen Demand (COD) percentage
COD; however, it is recommended a BOD value of < 60 degradation considering the absence of peroxide residual
mg L-1 or minimum efficient reduction of 80 % for after 2 h of reaction. For a better assessment, the results
treatment process. In general, homogeneous of DOC and COD were compared after the treatments as
photocatalytic treatments of DOC and COD parameters can be seen on Table III.
were efficient. The experimental conditions of these delineating
Another important parameter to be analyzed is the experiments were: pH 3.0, temperature 25 °C, Fenton
relation BOD5/COD. According to [36], this relation is reagent in accordance as can be seen in Table III and
different for several residues, which can be altered, ultraviolet radiation of 28 w.
specially, by biological treatment. The relation shows It can be observed in Table III that the photocatalytic
effective oxidation degradation in a determined organic oxidation reaction during the first 60 min presented the
load. highest range of reduction of the two response factors
Biodegradability was evaluated [39], which refers to a being 78 % and 50 % for COD and DOC, respectively
relation of BOD5/COD > 0.4 characteristically for (the maximum of reduction after two hours of reaction
biodegradable effluent. Thus, Table II shows that the was 86 % of COD and 53 % of DOC).
result obtained by AOP (Photo-Fenton) reached 0.69
TABLE II
RESULTS OF PHYSICAL-CHEMICAL ANALYSIS OF IN NATURA EFFLUENT AND AFTER PHOTO-FENTON TREATMENT
Effluent Releasing
Parameters Literature Values References
in natura Treatment Patterns[36] (mg L-1)
Aspect Turbid Limpid - - Absence
Color (mg Pt-Co L-1) 432 28.68 - - Absence
ST (mg L-1) 5680 Absence 545-5720 [18] 8
pH 6.0 – 6.4 8.0-8.50 5.25-8.0 [18], [35] -
Odor Irritant Absence - - 60
COD (mg O2 L-1) 9000–10000 930-939 797-8000 [18], [35] -
BOD5 (mg O2 mg L-1) 2300 – 2500 643.0-652.0 1292-60000 [18] -
BOD5/COD 0.25 0.69 - - -
DOC (mg /L-1) 1513-1800 137.0-143.0 2500-5000 [18] -
Chlorides (mg L-1) 1301.8 27 - - -
-1
N-NH3 (mg L ) 158.0 0.03 0.25-657 [18] -
N oganic (mg L-1) 180.0 0.05 16.5 – 1048 [18], [35]
283
TABLE III
RESULTS OF COD AND DOC PERCENTAGE REDUCTION OF EXPLORATORY STUDY IN THE PHOTOCATALYTIC REACTION
OF DAIRY EFFLUENT IN RELATION TO THE CONCENTRATION (MOL H2O2 L-1 + MOL FE2+ L-1)
Fenton reagent % of reduction of COD x time (min) % of reduction of DOC x time (min)
molH2O2 L-1 + mol Fe2+ L-1 Relation Molar 30 60 90 120 30 60 90 120
0,343+0,030 11.40 70 % 78 % 84 % 85 % 28 % 41 % 44 % 45 %
0,343+0,024 14.30 72 % 78 % 85 % 86 % 35 % 50 % 52 % 53 %
0,343+0,018 19.00 73 % 76 % 82 % 84 % 30 % 36 % 37 % 39 %
0,343+0,012 28.60 63 % 69 % 75 % 78 % 27 % 31 % 36 % 37 %
Therefore, it was decided to work with an L9 matrix According to Table I, it can also be noticed that the
with time reaction of 1 h and using as concentrations as temperature (level 1 = 20 ºC, level 2 = 30 ºC and level 3
shown in Table I for the respective levels 1, 2 and 3 of = 35 ºC) and the measurement of UV radiation source in
Fenton reagent of L9 array. W (level 1 = absence, level 2 =15 and level 3 = 28) are
significant.
Aiming to obtaining the variable value response (DOC
III.3. DOC Results for Taguchi’s L9 Method
percentage reduction), the following configuration
Fig. 1 shows the study of the process variability by the should be considered: high level for Fenton’s reagent and
graph of the average of the responses in relation to low level for pH. Table IV shows Analysis of Variance
signal-to-noise (S/N). The results shown in Fig. 1 were (ANOVA) involved factors in dairy effluent treatment
calculated from the situation "bigger is better", since it with Photo-Fenton process, according to Taguchi’s L9
seeks the highest percentage reduction of COT for each orthogonal array experiment.
experimental condition of the matrix L9. According to the analysis of variance on Table IV, all
factors show significant effect (F > 2) in the percentage
Main Effects Plot for DOC Variation
of DOC reduction, pH being the most significant, with F
Data Means equal to 79.824 and p-value equal to 2 x10-6. Less
T pH
80
significant are the concentration of Fenton (F= 13.911),
UV radiation (F=7.156) and the temperature (F = 4.499).
70
Thus, according to F tests values, the effect of pH is
60 approximately 6 times more significant to Fenton’s
variable, 11 times more significant to UV and 18 times
Mean
1 2 3 1 2 3
80
Fent on UV
more significant to temperature, based on DOC reduction
in ANOVA.
70
It was also observed the high confidence value given
60 by the test of p-value with the same sequence of
1 2 3 1 2 3
significance on test F, being the higher percentage
superior to 99.99 % for pH and 95.60 % for Fenton’s
Fig. 1. Percentage variation results of DOC reduction in factors used for
reagent and temperature.
dairy effluent treatment according to Taguchi’s L9 method According to Table V the experimental conditions
showing more DOC percentage reduction (about 90.86
Taguchi’s treatment analysis based on the graphic of %) in the dairy effluent was experiment 7, with Fenton’s
effects (Fig. 1) indicates that Fenton’s reagent factors reaction equal to 0.343 mol H2O2 L-1 + 0.0215 mol Fe2+
(level 1= 0,255 mol H2O2 L-1 0.0108 mol Fe2+ L-1level 2 L-1, ultraviolet radiation at 28 W, pH = 3.0 and
= 0.294 mol H2O2 L-1 + 0.0143 mol Fe2+ L-1 and level 3 temperature at 35 °C.
=( 0.343 mol H2O2 L-1 + 0.0215 mol Fe2+ L-1) and pH It should be mentioned that better results treated with
(level 1 = 3.0, level 2 = 4.0 and level 3 = 5.0) were Photo-Fenton process, as shown on Table V, were
relevant factors in the process. Notice that the effluent obtained due to the pH between 3.0 and 4.0. It can also
degradation process has a better performance in pH with be observed that tests 5 and 7 presented high percentage
more acidity (3.0). of DOC reduction for the same concentration value of
Fenton’s reagent and the lamp of same power.
TABLE IV
VARIANCE ANALYSIS OBTAINED FROM AVERAGE DOC REDUCTION VALUES ACCORDING
TO L9 ORTHOGONAL ARRAY FOR DAIRY EFFLUENT TREATMENT BY PHOTO-FENTON PROCESS
Degrees of
Factors Sum of Squares Mean sum of Squares F P
Freedom
1- Temperature 112.034 2 56.0168 4.49977 0.044199
2 – pH 1987.436 2 993.7180 79.8242 0.000002
3 – Fenton 346.342 2 173.1708 13.9106 0.001765
4 – UV 178.171 2 89.0857 7.1562 0.013803
Residual 112.040 9 12.4488
284
TABLE V
MEAN PERCENTAGE OF DOC REDUCTION FOR SPECIFIC FACTORS AND EXPERIMENTS ACCORDING TO TAGUCHI’S L9 ORTHOGONAL ARRAY
AFTER DAIRY EFFLUENT TREATMENT BY FENTON AND PHOTO-FENTON’S PROCESSES
Exp Temp (°C) pH Fenton (mg L-1) UV (W) meanDOC (%)
1 1 1 1 1 67.40
2 1 2 2 2 69.98
3 1 3 3 3 56.75
4 2 1 2 1 82.97
5 2 2 3 3 72.74
6 2 3 1 2 54.96
7 3 1 3 3 90.86
8 3 2 1 2 65.79
9 3 3 2 1 52.58
[40] pointed pH out as a decisive factor for an TABLE VI

DOC MEAN PERCENTAGES AND EFFECTIVE DEGRADATION
efficient reaction. The result of this study is practically
Mean of DOC Mean of effective degradation Difference in mean
based on the value of pH adopted, independently from (%) DOC (%) DOC
the type of effluent to be treated. Degradation speed 67.40 53.9 13.5
tends to be higher than pH, which is around 3.0, and 69.98 55.3 14.7
tends to diminish with the increase of pH due to 56.75 42.6 14.2
82.97 63.9 19.1
formation of ferrous species catalytically decomposing 72.74 55.3 17.5
hydrogen peroxide in oxygen and water, preventing the 54.96 47,0 7.96
formation of hydroxyl radicals. 90.86 69.0 22.0
The efficiency of Photo-Fenton process is related to 65.79 50.3 15.5
52.58 41.6 11.0
the optimum quantities of ferrous and hydrogen peroxide
to be used in the oxidative process as well as pH in the
reaction media. Thus, it can be noticed that the precipitation in
Through this concept, we noted that experiments 1, 4 alkaline pH had influenced the result of the employed
and 7, being performed with pH of 3.0 and concentration process (Photo-Fenton), owing to the quantity of organic
of Fenton’s reagent (0.255 mol H2O2 L-1 + 0.0108 mol matter added to the precipitate.
Fe2+ L-1) and (0.294 mol H2O2 L-1 + 0.0143 mol Fe2+ L-1) It can be observed as in experiment 7 that the Photo-
and (0.343 mol H2O2 L-1 + 0.0215 mol Fe2+ L-1) had a Fenton AOP real effective degradation, in percentage
DOC percentage removal of 65, 66, and 91%, terms, was 69 % that is, a difference of around 22 % of
respectively. DOC percentage degradation, a significant result.
These results show that the more quantity of species To the following processes to the AOP, specifically to
(concentration of oxidant agent and catalyst) used in the Fenton process, where only the filtrated is used, the
Fenton’s reagent, the more the reaction efficiency is. total DOC reduction will be considered.
Experiments 1 and 4 were performed without UV
radiation, which could also indicate less efficiency of the III.5. AOP Degradation Rate
process because the presence of UV radiation maximize
hydroxyl radical’s reaction. To express DOC behavior reduction among Taguchi
After the statistic analysis of the values obtained in experimental replication planning, a first order fit
Taguchi’s L9 array, a linear regression model for the exponential decay was used (5), obtaining correlations
percentage of DOC reduction was adjusted with coded higher than 0.97 representing a good adjustment among
variable levels given in (4): its mean points. It can be pointed out that this
mathematical adjustment has the function of evaluating
DOC variation = 70.4 + 2.52 T – 12.8 pH+ the behavior of the organic load reduction to dairy
(4) effluent oxidation studied in relation to time:
+ 5.37 [Fenton] + 3.85 UV
Y  Y0  A1e
 x / t1 
(5)
III.4. Effluent Effective Degradation After AOP
Due to physical-chemical characteristics, a The exponential outlines of dairy effluent organic load
complementary study was performed with the effluent degradation were shown in Fig. 2.
treated owing to quantify the total organic matter present In this block the highest DOC percentage degradation
(co-precipitation). After analytical characterization of the values were found. It is also significant the influence of
precipitate, the effective reduction of DOC of each pH in it. Good correlations were also found in the DOC
experiment of the Taguchi L9 array is shown in Table VI. mean adjustment results (R2 > 0.999), except for
During the precipitation of ferrous ions, co- experiment 4 (R2 < 0.9), in which, the equation does not
precipitated species (complexed anions or cations) were express a good mathematical adjustment. Lower
dragged by the hydroxide or adsorbed on the surface of deviation values between DOC mean results were also
Fe(OH)3 and/or Fe(OH)2 by electrostatic interaction. observed, being 2.4 %, 2.5 % and 3.5 % for experiments
285
4, 5 and 7, respectively, as showed in Fig. 2. Taguchi L9 orthogonal array was carried out considering
the chemical process only (Photo-Fenton).
DOC reduction of optimized time for 3 L of effluent
was considered in the final calculation of energy and
reagent consume proportionally to the relation
cost/benefit (less is better), in each experiment to
Taguchi L9 array as shown in Table VII.
Analyzing the results in Table VII, it can be observed
that the best result is the one that shows a better
cost/benefit ratio. On experiment 7, DOC percentage
degradation was of 90.86 %, however, its cost/benefit
value of 4.32 is the fourth smaller among the others.
Another relevant factor is the cost of inputs (reagents)
bigger in relation to the energy for all experiments.
Fig. 2. Experimental outlines of oxidation of dairy effluent by An individual analysis among Photo-Fenton and
Photo-Fenton process of experiments 4, 5 and 7 of Taguchi design in Fenton processes and its cost/benefit can be performed in
relation to time and DOC means
relation to the final destination of the treated effluent in
two specific evaluations.
It can be observed that there is no stationary phase in First of all, if the destination is the direct disposal to
the reaction, that is, after a 60 min period of dairy
rivers, the investment cannot be any different of
effluent degradation, the reaction could still be carried on experiment 7. Even though it is lower than experiment 7,
and the reduction of DOC be superior of 83 %. experiment 4 (Fenton Process) is still consistent to
Experiment 5 had the lowest organic load reduction
national laws (reduction higher than 80 %). The
(72.4 %) of this block. It can be observed that the cost/benefit value being less than experiment 7.
reaction is carried out in about 40 min (70 % of DOC Specifically, if the effluent treated by Photo-Fenton is
reduction). In experiment 7, the outline of DOC
later treated biologically (AOP hybrid – activated
degradation is steady by the end of the reaction (around
sludge), the temperature parameter must be evaluated
50 min), exactly at the end of H2O2 addition that can thoroughly for it is the one which makes the process
indicate a higher DOC percentage reduction (> 91 %) in more expensive.
case there is more peroxide in the reaction system.
Treatments with lower temperature must be the most
The pH values of this block in the lowest level (pH significant factor in the process. Experiments 1, 2 and 3,
3.0) presented the highest values of organic load for example, used the oxidation reaction at 20 oC. Among
degradation that confirms the Photo-Fenton process
them, experiment 2 is the one that shows better
mechanisms even in an environment with high matrix cost/benefit (4.03), showing the second best value in all 9
complexity. For a process of 1 h that is only experiments of Taguchi planning. Being all the variables
photoirradiated, using 3 L of dairy effluent, the
at intermediate levels and a DOC percentage reduction of
percentages of DOC reduction were: 13.5 % with UV
69.98 %.
radiation of 15 W and 18.2 % with 28 W. These Assessing concomitantly the energy and reagents
experimental results show the importance in combining a consume, experiment 6 may be considered the best one.
catalyzed oxidation power to the photolysis, especially
In this experiment, even using intermediate
when the organic load is high and of difficult temperature (30 oC), the cost/benefit value was 4.87, that
degradation. was compensated by using the lower Fenton reagent
(level 1) allied to pH 5.0 and UV lamp of 15 W.
III.6. Economic Evaluation of Photo-Fenton Process This experiment reached a DOC percentage reduction
of 54.96 %.
The economic evaluation (energy consume and
reagents) of dairy effluent treatment according to
TABLE VII
VALUES OF ENERGY AND REAGENT CONSUME OF 3 L DAIRY EFFLUENT TREATED BY PHOTO-FENTON PROCESS
OF EACH OPTIMIZED EXPERIMENT OF TAGUCHI L9 ARRAY
Energy Consume[41] Reagent Consume Total Value[42] Mean DOC Relation
Experiment
US$ / 3L US$/3L US$/3L Reduction % US$/%red.(x1000)
1 0.132 0.188 0.314 67.40 4.66
2 0.078 0.204 0.282 69.98 4.03
3 0.079 0.226 0.305 56.75 5.37
4 0.105 0.224 0.329 82.97 3.96
5 0.080 0.226 0.306 72.74 4.2
6 0.08 0.188 0.268 54.96 4.87
7 0.167 0.226 0.393 90.86 4.32
8 0.134 0.188 0.322 65.79 4.89
9 0.139 0.240 0.339 52.58 6.44
286
IV. Conclusion treatment in dairy industries—possibility of reuse, Desalination

195 (2006), 141–152.
DOC, COD and BOD5 percentage reduction were [2] K. Baskaran, L. M. Palmowski, B. M. Watson, Wastewater reuse
and treatment options for the dairy industry, Water Sci. Technol.
verified after the photochemical treatment, which varied Water Supply 3 (2003) 85–91.
from 91; 91 and 74 %, respectively. This demonstrates [3] Y. Laor, A. Jacek, Cal. L. Kozlel, Ravid, J. Air &Waste Manage.
that a big part of organic load present in the dairy Assoc., 58 (2008) 1187-1197.
effluent was degraded by the photochemical process, [4] C.H. Comninellis, Teorías y aplicaciones para el control de la
contaminación. (2010) p.18, Querétaro, Qro.
which means a significant contribution of AOP process [5] B. Demirel, O. Yenigun, T. T. Onay, Anaerobic treatment of dairy
for the degradation of dairy effluent. wastewaters. Process Biochem, 40 (2005) 2583–2595.
The analysis of ANOVA from the Taguchi’s L9 [6] B. Balannec, M.Vourch, M. Rabiller-Baudry, B. Chaufer,
experimental design, made it possible to identify the Comparative study of different nanofiltration and reverse osmosis
membranes for dairy effluent treatment by dead-end filtration, Sep
most significant factors in DOC percentage reduction. In Purif Technol, 42 (2005) 195–200.
this case, all factors were significant, being pH the one of [7] J. M. Chimenos, A. I. Fernandez, A. Hernandez, L. Haurie, F.
higher influence. An effective DOC degradation showed Espiell, C. Ayora, Optimization of phosphate removal in
a coprecipitation of organic species to the formed anodizing aluminum wastewater, Water Res, 40 (2006) 137–143.
[8] S. Irdemez, N. Demircioglu, Y. Sevki, Z. Bingul, The effects of
precipitate, in the treated effluent by Photo-Fenton and current density and phosphate concentration on phosphate
respectively alkalinisation, may vary 22.0 to 7.96 %. removal from wastewater by electrocoagulation using aluminum
In relation to the evaluation of cost/benefit, two and iron plate electrodes Sep Purif Technol, 52 (2006) 218–223.
aspects were considered. First, if the treated effluent by [9] A. K. Golder, A. N. Samanta, S. Ray, Removal of phosphate from
aqueous solutions using calcined metal hydroxide sludge
AOP were released directly in superficial water (river), wastewater generated from electrocoagulation. Sep Purif Technol,
the variables and respective levels of the process should 52 (2006) 102–109.
be: 30 oC, pH in 3.0, 0.294 mol H2O2 L-1 + 0.0143 mol [10] A. Hamdani, M. Mountadar, O. Assohei, Comparative study of
Fe2+ L-1 and without UV irradiation, obtaining a DOC the efficacy of three coagulants in treating dairy factory waste
water Int J Dairy Technol, 58 (2005) 83–88.
percentage reduction of 82.97 %. In case the effluent [11] A. Haridas, S. Suresh, K. R. Chitra, V. B. Manilal, The buoyant
treated by Photo-Fenton is treated afterwards by a filter bioreactor: a high rate anaerobic reactor for complex
biological treatment, the best experiment in relation to wastewater-process dynamic with dairy effluent, Water Res, 39
cost/benefit must be processed in the following variables (2005) 993–1004.
[12] J. L. Rico, H. Garcίa, C. Rico, I. Tejero, Characterisation of solid
and respective levels: 30 oC, pH in 5.0, 0.255 mol H2O2 and liquid fractions of dairy manure with regard to their
L-1 + 0.0108 mol Fe2+ L-1and UV irradiation of 15 W, component distribution and methane production Bioresour
being the DOC reduction of 54.96 %. Technol, 98 (2007) 971–979.
In light of the significant results obtained via AOP, [13] C.B. Chidambara Raj, H. Li Quen, Advanced oxidation processes
for wastewater treatment: Optimization of UV/H2O2 process
based on homogeneous photocatalysis (photo-Fenton and through a statistical technique, Chemical Engineering Science, 60
Fenton process), it can be said that this method has a (19) (2005) 5305-5311.
significant potential of application for the treatment of [14] A. Rey, M. Faraldos, J. A. Casas, J. A. Zazo, A. Bahamonde, J.
dairy effluent, which combined with an activated sludge Rodryguez, J. Appl. Catal. B: Environ. (2008).
[15] G. Lamas Samanamud, C. C. A. Loures, A.L. Souza, et al.,
type biological process can degrade its organic load Heterogeneous Photocatalytic Degradation of Dairy Wastewater
within the process efficiency values established by Using Immobilized ZnO, ISRN Chemical Engineering, (2012)
Article 18 (CETESB), enabling its disposal. Article ID 275371, 8 pages doi:10.5402/2012/275371
Of the two processes and considering the best cost [16] J. Pignatello, J. Environ.Sci.Technol. 26 (1992) 944.
[17] S. Parsons, Advanced oxidation processos for water treatment. 2nd
benefit, the Fenton process is more advantageous. ed. (United Kingdon 2005).
The researchers of this review have worked with some [18] J. R. Banu, S. Anadan, S. Kaliappan, I. T. Yeom, Treatment of
environmentally important themes. Especially with dairy wastewater using anerobic and solar photocatalytic methods.
oxidative processes applied to industrial effluents and Solar Energy, 82 (9) (2008) 812-819.
[19] H. Kusic, N. Koprivanac, L. Srsan, J. Photochem. Photobio. A:
wastes, considering homogeneous and heterogeneous Chem., 181 (2007) 195.
catalysis, in batch and semi-continuous processes. They [20] M. T. M. Pendergast and E. M. V. Hoek A review of water
work with the statistical evaluation of experimental treatment membrane nanotechnologies. Energy Environ. Sci., 4
planning, and have also applied a mathematical model to (2011) 1946.
[21] C. H. Langford, J. H. Carey, Can. J. Chem., 43 (1975) 2430.
determine the importance of each parameter to the model [22] G. Ruppert, R. Bauer, Chemosphere., 28 (8) (1994) 1447.
and its costs. [23] A. Goi, M. Trapido, Chemosphere., 46 (2002) 913.
[24] E. Chamarro, A. Marco, S. Esplugas, Water Res., 35 (4)
(2001)1047.
Acknowledgements [25] W. Tang, S. Tassos, Water Res.31 (5) (1997), 1117.
[26] T. Y. Wang, C. Y. Huang, European J. Operational Res. (2007)
To the laboratory of Environment coordinated by Prof. 1052.
[27] G. Taguchi, System of Experimental Design, vols. 1 and
PhD Messias Borges Silva. To CAPES for the financial 2Unipub-Kraus/ASI (New York/Dearborn, MI 1988).
support. [28] G. Taguchi, Introduction to Quality Engineering: Designing
Quality into Products and Processes Unipub-Kraus/ASI (New
York/Dearborn, MI 1987).
References [29] A.M.F.M. Guedes, L. M.P. Madeira, R.A.R. Boaventura, C.A.V
Costa, Water Res. 37 (2003) 3061.
[1] B. Sarkar, P. P. Chakrabarti, A.Vijaykumar, V. Kale, Wastewater [30] M. W. Weiser, K. B. Fong, Am. Ceram. Soc. Bull., 73 (1994) 83
287
[31] L. S. Lima, H.J. Izário Filho, F.J.M. Chaves, Rev. Analytica. 25 Rodrigo Fernando dos Santos Salazar. PhD in
(2006) 52. Science with emphasis in Analytical Chemistry
[32] APHA–AWWA. Standard Methods for the Examination of Water at Federal University of São Carlos (Brazil,
and Wasterwater. 21st ed. (New York: American Public Health 2013). He has experience in Chemical
Association 2005) Engineering with emphasis on wastewater
[33] B. G. Kwon, D. S. Lee, N. Kang, J. Yoon, Water Res. 33 (9) treatment, and Chemistry, with an emphasis on
(1999) 2118. Analytical Chemistry. Prof. Salazar is assistant
[34] Y. W. Kang, Water Res.34 (10) (2000) 2786. professor at Franciscan University (UNIFRA)
[35] P. M. Ndegwa, L. Wang, V. K. Vaddella, Proc. Biochem. 42 (9) and in the State University of Rio Grande do Sul (UERGS) in the area
(2007) 1272. of Sciences and Technology.
[36] CETESB, Variáveis de Qualidade das Águas (2010). Available E-mail: r.f.s.salazar@gmail.com
in: http://www.cetesb.sp.gov.br/Agua /rios/variaveis.asp#dbo
[37] F. J. Baumann, Dichromate reflux chemical oxygen demand. André Luís de C. Peixoto was born in São José
Proposed method for chloride correction in highly saline wastes. dos Campos, Brazil, in 1981. He received his
Anal. Chem., 46 (9) (1974) 1336–1338. degree as a chemical engineer in 2006 and
[38] M. D. Porter, B. Vaidya, S. W. Watson, S. J. Coldiron, Reduction obtained his MSc. degree in 2008 from the
of chloride ion interference in chemical oxygen demand (COD) Engineering School of Lorena (University of
determinations using bismuth-based adsorbents. Analytica São Paulo), Brazil. On September 2008, he
Chimica Acta (1997). became a member of the Center of Studies on
[39] D. E. Kritikos N.P. Xekoukoulotakis E. Psillakis , D. Chemical Engineering as a PhD candidate and
Mantzavinos, Photocatalytic degradation of Reactive Black 5 in workes at Federal Institute of Education, Science and Technology
aqueous solution: Effect of operating conditions and coupling (IFSP, Brazil). Mr. Peixoto is member of the American Institute of
with ultrasound irradiation. Water Research, 41(10) (2007) 2236– Chemical Engineering and the Brazilian Society of Chemical
2246. Engineering.
[40] Y. Zhang, J. Choi, C. Huang, J. Hazardous Mat. B125 (2007) E-mail: alcpeixoto@yahoo.com.br
166.
[41] Portal Business Brasil (2012) Energia Elétrica a terceira mais cara Oswaldo Luiz Cobra Guimarães was born in
do Brasil. Bananal, Brazil, in 1962. He received his degree
https://sites.google.com/site/portalbusinessbrasil/home/energiabra as a chemical engineer in 1986, MSc. degree in
sileira 2002 from University of Santa Catarina, Brazil
[42] UOL (2012) Cotação do dólar. Available in: and Ph.D. in 2005 from the University of
http://economia.uol.com.br/cotacoes. Taubaté. Dr. Guimaraes teaches disciplines
related to mathematics and computer
programming, also working on research related
Authors’ information to advanced oxidation processes, mathematical modeling and neural
networks at the School of Engineering of Lorena, University of São
Carla Loures was born in Juiz de Fora, Brazil, Paulo.
Master of Science in Chemical Engineering E-mail: cobraguimaraes@uol.com.br
(University of São Paulo) in environmental areas
(wastewater treatment and characterization of
solid and liquid effluents)/2011. Doctor of
Science in Production Engineering area by the
State University Paulista Júlio de Mesquita
Filho. Mrs Loures operates on the following
subjects: Mathematical Methods and Statistical Engineering, with
emphasis in Advanced Oxidation Processes.
Tel.: (55)12-3159-5076
E-mail: carlaloures@dequi.eel.usp.br
Hélcio Izário Filho. Doctor of Science in Chemistry from by the State

University of Campinas. He is currently a professor in the School of
Engineering of Lorena Universidade de São Paulo. Escola de
Engenharia de Lorena, University of São Paulo. Dr. Izario Filho has
experience in chemistry with emphasis in analytical instrumentation,
specifically absorption spectrometry and atomic emission
spectrophotometry, acting on the following themes: characterization of
the inorganic environment (analytical characterization of solid and
liquid waste and Advanced Oxidation Processes).
E-mail: helcio@dequi.eel.usp.br
Gisella Lamas Samanamud. Masters in Chemical Engineering by the

University of São Paulo/2011. She is currently a researcher / student at
the University of Texas at San Antonio (UTSA), TX, USA. MSc Lamas
Samanamud currently works in the Laboratory of Environmental
Engineering on the development of wastewater treatment, Advanced
Oxidation Processes, photochemical processes and molecular biology.
E-mail: fgn366@my.utsa.edu
André Luiz de Souza. MSc in Chemical Engineering from the

University of São Paulo, 2011. He is lately working as professor and
High School Coordinator.
E-mail: andrelsfonte@hotmail.com
288
ISSN 2035-1755 July 2013
Influence of Magnetic Field on the Efficiency

of the Coagulation Process to Remove Turbidity from Water
Waleed M. Sh. Alabdraba, Mohamed B. A. Albayati,

Ahmed Y. Radeef, Mustafa M. Rejab
Abstract – The effect of magnetic field on the coagulation process and the possibility of reduce
the coagulant dose with different cases of shedding the magnetic field on the two samples of water
was studied, and the results showed that the magnetic field affects clearly and strongly on
increasing the efficiency of removing the impurities from water (turbidity reduced from 662 NTU
to 0.36 NTU and from 45.25 NTU to 0 NTU), and on reduction of the chemical coagulant (reduced
optimal dose about 10 mg/L) by subjected water to magnetic field with optimal intensity in the
periods of rapid and slow mixing and sedimentation at the third bottom of the tank. Copyright ©
2013 Praise Worthy Prize S.r.l. - All rights reserved.
Keywords: Magnetic Field, Turbidity, Coagulation Process
I. Introduction A magnet is a material or object that produces a

magnetic field. A magnetic field is generated when
Water is a basic necessity of life, not only for people electric charge carriers such as electrons move through
but for every type of plant and animal as well. Much of space or within an electrical conductor. This magnetic
our fresh water is also used outdoors for watering lawns, field is invisible but is responsible for the most notable
flower beds, and vegetable gardens, as well as washing property of a magnet [5]. Liquid water is affected by
cars and filling swimming pools, therefore the water magnetic fields [6], [7] and such fields can assist its
must be treated to make it safe for human consumption, purification [8]. Water is diamagnetic and may be
aesthetically acceptable, and suitable for use by affected by very high magnetic fields (10 Tesla (T),
industries and other uses [1]. compare Earth's magnetic field 50 μT) [9]. Other studies
Many processes used for treatment of surface water show an increase in cluster size in liquid water is caused
and ground water consists of primary sedimentation, by a magnetic field [7]. Salt mobility is enhanced in
coagulation, filtration, disinfection, conditioning, strong magnetic fields (1-10 T) causing some disruption
softening, fluoridation, removal of tastes and odors, to the hydrogen bonding [10].
corrosion control, algae control, and aeration [2]. However this only causes a net reduction in hydrogen
Chemical coagulation is an important process in water bonding at high salt concentrations (for example 5 M
treatment that helps produce clear, finished water which NaCl), whereas at lower concentrations (1 M NaCl) the
is aesthetically acceptable to the consumer. Flocculation increase in water hydrogen bonding in the presence of
is gentle stirring or agitation to encourage the particles such high magnetic fields more than compensates for this
thus formed aggregates with masses large enough to effect [10]. Lo and others studied high turbidity raw
settle or be filtered from solution. In these processes water treated by magnetic aggregation method with
colloidal particles (20-2000Å)[1 Å = 10-4 μm = 10-10 m] applied magnetic field and the operating parameters were
and very fine solid suspensions initially present in water appropriate dosage of magnetite particles, water quality
are combined into larger agglomerates that can be parameters and the strength of applied magnetic field, the
separated by primary sedimentation, flocculation, modified jar tests were used in this study by adding pre-
filtration, or other separation methods [3]. chemosynthesis (Fe3O4) nanoparticles into the high
The colloids commonly found in water are stable turbidity water. Magnetite particles (Fe3O4) were
because of the very large surface area relative to their synthesized by chemical precipitation method in this
mass and the electrical charge that they carry. The charge study.
of colloids can be positive or negative. However, most The results indicate that it has better efficiency of
colloidal particles in water have a negative charge. magnetic aggregation after magnetic particles being
Because of their large surface area and electrical charge, further magnetized and the method of magnetic
colloidal particles settle very slowly. Aluminum or iron aggregation and separation can reduce the turbidity of
salts are utilized to neutralize these surface charges and raw water of different initial turbidity and from different
to cause the colloids to coalesce and become large areas at different pH values [4].
enough so that they will readily settle [4].
289
Waleed M. Sh. Alabdraba, Mohamed B. A. Albayati, Ahmed Y. Radeef, Mustafa M. Rejab
While Banejad and Abdosalehi studied the effect of The hardness were measured in the laboratory of
magnetic field on water hardness reducing. In this study, Water and Environment, in department of
they have been chosen amounts of water influent 4 L/h Environmental Engineering In Tikrit University, by
and 30 L/h and magnetic field intensities of zero titration method.
Tesla,0.05 (Tesla), 0.075 (Tesla), and 0.1 (Tesla), with
doing examination by 3 times and analyze the results,
they have shown that changing magnetic field intensity, III. Experimental Procedure
amounts of water influent, and also together, influence on Water samples were collected over a period for more
water hardness, and the result shown that it will increase than a month as the characteristics of the river were
the efficiency of magnetic treatment, when magnetic variable depending on rainfall periods and change in the
field intensity is changing and\or amounts of water amount of rain.
influent with significant effects at level of 99% on Water samples, in the laboratory, in all studied cases
reducing of water Hardness [11]. Monica Chin and Fan were put in the six circular jars (1L) of the jar test device
studied the effects of applied magnetic field strength that in each one there are a paddle with (6cm) length and
during settling of aggregates and total organic carbon (2cm) width, and in all experiments the jar test was used
(TOC) concentration, on the removal of turbidity that with 250 rpm with gradient velocity (G= 62/s) as a Rapid
exceeds 10,000 NTU resulted from typhoon and that mixing for 1 minute, to produce a good mixed solution,
caused to shut down the water treatment plants. After and then 30 rpm as a slow mixing for 30 minutes with
addition of proper dosage of magnetite nanoparticles gradient velocity (G= 15.8/s) and (G·T=2.84×104) as a
(Fe3O4) (1 g/L), turbid particles settled down more period of fluctuation, then the jars were left for 20
quickly and the turbidity was significantly lowered from minutes as a period of sedimentation [14], [15].
about 10000 to less than 300 NTU, and in the presence of These samples were subjected to the magnetic fields
external magnetic field during sedimentation, the with different intensities [measured in Tesla unit (T)]
turbidity could be reduced to about 30 NTU and all that (1.38 T, 2.76 T, 4.14 T, 5.52 T, 6.9 T, 8.28 T) by using
is under 1000 NTU which is acceptable for potable water number of magnetic bars, which tied around the beakers
treatment plant to work [12]. (jars). The number of experiments was worked, in all
Alkhazan and Saddiq studied the two cases of experiments Alum (Aluminum Sulfate) was used as a
magnetic field applied to the water (static and shaking) coagulant, (which prepared by dissolved 1 gm of dry
and found The results of pH value increase with alum in 1 liter of distilled water), the physical and
increasing magnetic intensity in static and shaking chemical analyzes (turbidity, pH, EC,TDS, temperature
treatment. decrease in EC values with an increase of both and Hardness) have been made for all samples before the
magnetic intensity and time, and decreases in (P, Ca, Cu, jar test and after it for all experiments.
Na and Cl ) concentration were (82, 27, 30, 6 and 18%), The research was carried out in number of stages on a
respectively in static state, while they were ( 82, 37, 54, 2 laboratory scale as below:
and 12%), respectively in shaking state [13]. The First Stage included the examination of the jar
The main objectives of this study was: without exposure to a magnetic field and then choose the
1- Study the effect of the magnetic field on the optimal dose of alum.
coagulation process. The Second Stage included choosing the optimal
2- Selection the perfect case with appropriate depth for magnetic field with the optimal dose for the same
shed the magnetic field. sample(same water sample properties).
3- Study the possibility of reducing the optimal dose of The Third Stage includes the use of static magnetic
coagulant by using a magnetic field. intensity with a fixed dose and study the different
situations of exposure to magnetic field at different
heights (at the bottom third and middle third and upper
II. Materials and Methods
third) and tying magnets once vertically and once
- Sampling site: horizontally (to choose optimal tying situation of
Water samples were collected from The Tigris River magnetic bars) and also included exposure to the
near the location of suction pipes that transport the raw magnetic field from the bottom and the case of tying
water from the river to The Tikrit University Plant For magnets on the Paddle surface.
Water Supply, where the physical and chemical The Fourth Stage included the use of optimal dose
characteristics of the water affected by pollution that with optimal situation of tying the magnetic bars with
happened to river by the rain and the disposal of the optimal magnetic intensity with change the period of
wastewater. exposure to the magnetic field to show where the
- Physical and Chemical Analysis of Water Samples: magnetic field strongly affects exactly (in rapid mixing
The physical properties of the samples such as or in slow mixing or in sedimentation period or in two of
turbidity, pH, temperature, EC and TDS, were measured them or in all steps).
by using turbidity meter HANNA model 93703 and pH, These stages were held twice, once on a high sample
temperature, EC and TDS, measured by using turbidity and again on a low sample turbidity, but in low
multi_parameter PCS model 35. turbidity samples we chose the lowest dose to show the
290
effects of magnetic field obviously, and to show the

possibility of reduce the optimal dose by using optimal
magnetic field.
The Fifth Stage included using the optimal magnetic
intensity for specific water quality (previous samples)
with other new water sample (different water quality)
without choose a new magnetic intensity for this new
sample, to study the possibility of use the optimal
magnetic intensity.
IV. Result and Discussion

Table I shows the characteristics of the two samples of Fig. 2. The relation between magnetic intensity and turbidity for sample
water used in this study. (1) with alum dose 10(mg/L) and sample (2) with alum dose 30 (mg/L)
TABLE I
THE CHEMICAL AND PHYSICAL CHARACTERISTICS
OF THE TWO WATER SAMPLE
Properties Sample (1) Sample (2)
TU (NTU) 45.25 662
pH 7.99 7.40
EC (µS/cm) 456 495
SALT (ppm) 324 236
TDS (ppm) 215 351
◦
T ( C) 14.3 16.7
Hardness (mg/L) 240 240
As it shows that characteristics of the sample (1) had

changed after rain, as shown in the characteristics of the
sample (2), the effect of magnetic field was studied by Fig. 3. The relation between the turbidity and the diffrent sites
and cases of shed the bar
using these two samples. The first step is to choose the
optimum dose of coagulant by jar test for the two As can be seen the magnetic intensity must be shed in
samples. The optimal dose for a sample (1) was (20 mg / the bottom third of the jar and with the horizontal
L) and (30 mg / L) for the sample (2) as showing in the position of the bars, and the worst case is shedding the
Fig. 1. Note: the optimal dose was reduced in the first magnetic field down the jar or on the Paddle surface,
sample to show the effect of the magnetic field clearly (because of the effect of the magnetic field on the
and to examine the possibility of using a lower dose than movement of the paddle) as shown in Fig. 3. After the
the optimal dose. In the second stage a several magnetic choice of the optimum dose of coagulant and appropriate
intensity were used (1.38, 2.76, 4.14, 5.52, 6.9 and 8.28) magnetic field intensity and appropriate depth to shed the
Tesla. The tests show that the optimal intensity for magnetic field, appropriate period to begin the process to
sample (1) is 4.14 T, and 5.52 T, as shown in Fig. 2. shed the magnetic field are choosing, and had been
In the third stage by using the optimum dose of shown that the magnetic field affected largely in the
coagulant in the samples with shed optimum magnetic sedimentation (same result as compared with [11]) and
intensity (fixed dose with constant magnetic intensity) to lesser affected in the slow mixing and rapid mixing, but
choose an appropriate depth to shed magnetic intensity it is better to be a shed the magnetic field during the three
and appropriate way to tying magnetic bars. processes (slow and rapid mixing and sedimentation) to
achieve high efficiency, as shown in the Figs. 4 and 5.
Fig. 4. The relation between turbidity and the different periods

Fig. 1. The relation between alum dose and turbidity of exposure to the optimal magnetic field for sample (1)
for samples (1) and (2)
291
and rapid mixing, but it is better to be shed the

magnetic field during the three processes (slow and
rapid mixing and sedimentation) to achieve high
efficiency.
 The magnetic field must be shed in the bottom third
of the jar and with the horizontal position of the bars.
 The worst case of magnetic field shedding when the
magnetic field shed from down the jar or from the
Paddle surface.
 We can reduce the optimal dose of coagulant by using
Fig. 5. The relation between turbidity and the different periods an optimum magnetic field and thus reducing the
of exposure to the optimal magnetic field for sample (2) health impact of coagulant on the health of
consumers.
The effect of the magnetization can be observed
clearly by comparing the efficiency of removal on the
same sample with and without the shedding of the References
magnetic field, where shown that with the shedding of
[1] Diersing, Nancy, Water Quality: Frequently Asked Questions,
the magnetic field with optimal intensity increase the PDA. NOAA (2009).
efficiency of coagulation process, as shown in the Fig. 6 [2] De Zuane, J., Handbook Of Drinking Water Quality ( Jone Wiley
for the first sample and Fig. 7 for the second sample. & Sons Inc. 2nd'' edition 1997 ).
[3] Lanubusch, E. J., Water Quality and Treatment (McGraw-Hill, 3rd
edition. 1971).
[4] Lo, Shang-Lien, Yung-Li Wang and Ching-Yao Hu, High
Turbidity Reduction During The Storm Period By Applied
Magnetic Field, J. Environ. Eng. Manage. 17 (2007) 365-370.
[5] Griffiths, David J, Introduction to Electrodynamics ( Prentice
Hall 3rd edition 1999).
[6] Pang, X. F. and Deng, B., Investigation Of Changes In Properties
Of Water Under The Action Of A Magnetic Field, Sci China Ser
G-Phys Mech Aston 51(2008) 1621-1632.
[7] Cai R. Yang H, He J, Zhu W., The Effects Of Magnetic Fields On
Water Molecular Hydrogen Bonds, Journal of Molecular
Structure 938 (2009) 15-19.
[8] Ambashta, R. D. and Sillanpää, M., Water Purification Using
Magnetic Assistance, J Hazard Mater. 180 (2010) 38-49.
[9] Zaslavsky, B. Y, Aqueous Two-Phase Partitioning, Journal of
Dispersion Science and Technology 16 (1995) 393-394.
Fig. 6. Comparison for sample (1) after jar test between [10] Chang, K.-T. and Weng C.-I., An Investigation Into The
using coagulant only and coagulant with magnetic field Structure Of Aqueous Nacl Electrolyte Solutions Under Magnetic
Fields, Computational Materials Science 43 (2008) 1048–1055.
[11] Banejad, H. and Abdosalehi, E., The Effect Of Magnetic Field
On Water Hardness Reducing, IWTC 13 (2009) 117-128.
[12] Monica Chin, Ching-Ju and Fan Zhen-Guo,, Magnetic Seeding
Aggregation Of High Turbid Source Water, J. Environ. Eng.
Manage., 20 (2010), 145-150.
[13] Alkhazan, Molouk and Saddiq, Amna, The Effect Of Magnetic
Field On The Physical, Chemical And Microbiological Properties
Of The Lake Water In Saudi Arabia, Journal of Evolutionary
Biology Reasearch 2 (2010) 7-14.
[14] American Water Works Association, Operational Control of
Coagulation and Filtration Processes (Manual of Water Supply
Practices – M37 Edition 1992).
[15] Ammirtharajah, A. and Trusler, S.L., Destabilization of Particles
by Turbulent Rapid Mixing, J. Environ. Eng. 112 (1986) 1085-
1108.
Fig. 7. Comparison for sample (2) after jar test between using coagulant Authors’ information
only and coagulant with magnetic field
Waleed M. Sh. Alabdraba born in Iraq-Mosul
Jan-6-1976. PhD Environmental Engineering,
V. Conclusion Mosul University, College of Engineering,
Mosul, Nenava, Iraq 2005. He research in the
 The efficiency of coagulation process greatly increase field of Water and Wastewater Treatment,
Biofuel Cell, Solid Waste Management and
when the magnetic field with optimal intensity is Industrial Wastewater. Asst. Prof. Alabdraba is
used. a faculty member at Environmental Engineering
 The magnetic field affected greatly in the Department, Tikrit University, member of Iraqi Engineers Union.
sedimentation and lesser effect in the slow mixing E-mail: walabdraba@tu.edu.iq
292
Mohamed B. A. Albayati born in Iraq-Kirkuk

Jan-6-1984. M.Sc. Environmental Engineering,
Tikrit University, College of Engineering, Tikrit,
Salah-Aldeen, Iraq 2010. He research in the field
of Water and Wastewater Treatment, Industrial
Wastewater. Asst. Lect. Albayati is a faculty
member at Environmental Engineering
Department, Tikrit University, member of Iraqi
Engineers Union.
E-mail: mohbureng@tu.edu.iq
293
ISSN 2035-1755 July 2013
Evaluation of a System for Dairy Wastewater Treatment

Composed by Photochemical (TiO2 fixed / UVsolar)
and Biological (Aeration Pond) Reactors
Rodrigo F. S. Salazar, Marcos F. Oliveira, Marco A. K. Alcântara, Hélcio J. Izário Filho
Abstract – The efficiency of a system composed with photochemical reactor based on Advanced
Oxidation Process (AOP) followed by treatment in bioreactor was investigated as an alternative
for treatment of dairy effluent in order to reduce the organic load. The influence of organic load in
crude dairy effluent and pre-treated dairy wastewater were also investigated upon efficiency of
biodegradation system based on Aerated Pond System (APS). The AOP was conducted using
titanium dioxide (TiO2) as photocatalyst and solar irradiation. The catalytic bed was prepared by
using TiO2, coating evenly with polyurethane, and fixing on a flat metal plate. Effluent was
percolated and fixed at 23º angle of the North on a catalytic plate where it was irradiated for a
three-hour period. The biodegradation tests were conducted during 72 hours. Chemical Oxygen
Demand and Total Organic Carbon were employed to evaluate the total organic carbon reduction,
respectively. Biodegradation of crude dairy effluent with an approximate COD value of 3,800 mg
O2 L-1 presented around 26.01 ± 5.23% of carbon reduction. The AOP – APS system, which
showed higher efficiency of organic load degradation, presented 93.70 ± 0.10% of total carbon
reduction, characterized by a COD of entry into 3782.5 ± 37.6 mg O2 L-1 and a COD for the
treated effluent output at 236.8 ± 0.1 mg O2 L-1. This data demonstrated the necessity of dairy
wastewater pretreatment and the efficiency of pretreatment system based on TiO2/UVsolar.
Copyright © 2013 Praise Worthy Prize S.r.l. - All rights reserved.
Keywords: AOP, Photocatalysis, TiO2, Dairy Wastewater
I. Introduction 3) treatment systems based on enzymatic hydrolysis

using lipase to increase the BOD/COD ratio and,
Among food industries, the dairy sector companies thereby, increase the degradation efficiency of the
promote higher environmental impact due to the organic pre-treated effluent in a subsequent biological
load of their waste, generated liquid waste volume and treatment [5]-[11]. From the presented systems, the
high water consumption [1]. biological degradation remains as one of the most
Brazil is one of the five largest milk producers with a promising options for the removal of organic material
significant increase of 37% volume of produced milk, from dairy effluents.
from 23.5 billion liters to 32.1 billion liters of milk in However, these systems have the limitation of a
2011 [2]. On the first semester of 2012, 5.7 billion liters narrow range to operate with great degradation efficiency
of milk have been purchased and, from this total, about due to factors such: restricted pHoptimum range, sudden
99.7% have been processed into milk powder and dairy variations of the organic load, as well as the physical and
products [3]. chemical characteristics of the effluent, which can lead to
Concurrently with the growth of this industry, the an overloading of the treatment system and sludge
generation of dairy effluents and derivatives as well as its swelling [12].
management, recycling and treatment of these wastes In any case, the biological treatment systems,
have become an increasingly concern in Brazil and in the especially those done by activated sludge, have a great
world [4]-[5]-[6]-[7]. Several studies and literature refers potential for being combined with other biodegradation,
that liquid wastes, arising from milk processing, propose physical-chemical processes among others.
treating their waste by using: Among the new developed purification technologies
1) ecological system treatment: consisting of aerobic and processes in order to treat complex effluents,
reactors, anaerobic, clarifiers and wetlands [8]-[9], Advanced Oxidation Processes used with efficient results
2) physical treatment systems, such as the use of in the remediation of chemical species, mainly
ultrafiltration membranes and reverse osmosis that, in recalcitrant chemical species [13] must be highlighted.
most cases, include the reuse of the water used in the The AOPs were defined by Glaze (1987) as physical
process [1]-[10], and chemical processes that involve the generation of
294
transient species with a high oxidizing power, pointing acid from Carlo ErbaReagenti P.A.-A.C.S. and Vetec
out the hydroxyl radical (•OH) among others. brands were used respectively.
This radical has a high oxidizing power (EPHHO•/HO‾ ~ Solid reagents used for the preparation of solutions
+2.8V, 25 ºC), and can be generated by photochemical and standards were of high analytical grade (PA). Water:
means, including sunlight or by other forms of energy, resistivity of 18.2 m cm obtained by a Millipore
mineralizing organic pollutants in non-toxic forms as system, Simplicity model.
CO2 e H2O [13]. Some AOPs, as heterogeneous
photocatalysis, radiolysis and other advanced techniques,
allow transformations on toxic contaminants that are less II.1. Advanced Oxidative Process (TiO2/UVsolar)
susceptible to oxidation, such as metal ions and halogen As a pretreatment for the degradation of the dairy
compounds [13]-[14]. Most of AOPs studies reveal that effluent, a system based on heterogeneous photocatalysis
the heterogeneous photocatalysis using TiO2 and ZnO with fixed titanium dioxide and solar radiation was used.
may have an important role with regard to emerging An open reactor in a continuous system, bench scale,
technologies for water treatment, due to large number of was used in the photocatalytic process for the solar UV
investigations on the subject compared to other AOPs radiation absorption. TiO2 used had a treatment with
studied [7]-[13]-[15]-[16]-[17]-[18]. silicon oxide and was manufactured by DuPont from
As a result of the details above, this study aims to Brazil S/A. A 5.0 L volume and 13 L.min-1 of leachate
develop a hybrid system consisting of a heterogeneous flow were constantly kept.
photocatalytic pretreatment (TiO2 / UVsolar), which This system, according to Fig. 1, was basically
increases the biodegradation ration of the dairy effluent composed of 28 × 20.5 × 26 cm glass tank, a BOMAX
for further treatment with aeration pond system, seeking centrifugal pump, Model NH-30PX-T - 13 L min-1 and a
to reduce the hydraulic retention time of the biological 24 cm wide and 75 cm long metal plate coated with paint
treatment and a better disposal quality of this effluent. containing TiO2. A Fractional Factorial Design of
Experiments (DOE - 24-1) was employed to optimize the
pretreatment via Advanced Oxidation Process. On this
II. Material and Methods study the parameters evaluated were: reaction time (2
For collection, preservation and storage of the dairy and 3 hour of reaction time), phase of TiO2 (Anatase and
effluent, samplings were necessary in 20.0 L pumps of Rutile), medium acidity (pH 5 or 7), and the organic load
the effluent treatment station homogenization tank. These of dairy wastewater (crude effluent or 1:1 v/v with
ones were stocked in a cold room at 4°C in the water). The response variable was the percentage of the
Engineering School of Lorena. chemical oxygen demand (%CODreduced). This way, the
Due to the complex characteristics of the dairy kinetic efficiency of the effluent degradation was
effluent samples, adaptations of the methods for the verified. For statistical analysis, software MINITAB®
determination of chemical oxygen demand were Release 14.12.0 - Statistical Software, 2004 Oringin 6.0
necessary. For COD determination in the dairy effluent, and Excel (Windows 2007) were used. The paint applied
adjustments were required [19]-[20] on method 5220 D. on the metal plates was produced with polyurethane resin
Closed Reflux, Colorimetric Method of APHA Standard and high concentration of 0.025 g TiO2 cm-2 catalyst to
Methods [21]. Measurements of total organic carbon enable greater efficiency of the oxidative process and
were performed in a Shimadzu TOC-VCPH total organic consequent degradation of the effluent. After application,
carbon analyzer based on catalytic oxidation at elevated metal plates were placed in a stove at 90°C for 30
temperatures and CO2 determination by infrared minutes. An aluminum support that permitted the
spectroscopy. A calibration curve was prepared from a exchange of stainless steel plate with one of the two
potassium biphthalate standard, covering the range from types of Anatase or Rutile TiO2 was directed towards the
0 to 500 mg L-1. For the inorganic carbon, the curve was North with a direction angle of 23°, as shown in Fig. 1.
prepared with a standard mixture of Na2CO3 and The substrate to be treated was pumped up to the top
NaHCO3 in the range between 0 and 500 mg L-1. of the plate, uniformly percolated while being photo-
Samples were prepared from an aliquot of 1.0 mL of irradiated. Monitoring, correction and adjustment of the
effluent, previously filtered through a membrane (0.45 pH parameter during the experiments were performed by
µm) and diluted to 25.0 ml with distilled water. a pH meter adapted to the reservoir, and temperature by a
After homogenization, the sample was injected into a digital infrared thermometer INCOTERM® model
chamber at high temperature of 680 °C, containing MULT TEMP. Previous to photocatalytic tests,
platinum adsorbed in alumina to determine the total evaporation effects in an open system reactor was
carbon. Another aliquot of the sample was injected into evaluated by placing a system in parallel, using a non-
the device in another reaction chamber with addition of coating metal plate. Follow-ups were carried out on
hydrochloric acid for the determination of inorganic evaporated percentage from the solution and also 'blank
carbon. For both steps, CO2 was determined by a non- tests' of catalyst bed without TiO2 to assess possible
dispersive infrared analyzer and TOC determined by the degradation effects by photolysis. It was found that no
difference between TC and IC. Reagents and solutions as photolysis occurred in the dairy effluent during the
65% m/m of nitric acid and 95 - 97% m/m of sulfuric photocatalytic treatment [20].
295
phase).
TABLE I
ANALYSIS OF VARIANCE RESULTS OBTAINED DUE TO FACTORIAL
DESIGN (24-1) TO EVALUATE AOP SYSTEM
Sources Gl1 SQ2 MQ3 F4 p5
Main effects 4 776.6 194.15 3.49 0.062
Interactions 3 2465.1 821.71 14.79 0.001
Residual error 8 444.5 55.56
Total 15 3686.2
1
degrees of flexibility; 2quadratic sum; 3quadratic average; 4 F Factor;
5
Fig. 1. Schematic of a solar reactor with TiO2 fixed on a plate [17] degree of confidence
Results show that the more the reaction time

II.2. Biodegradation Process (Aeration Pond System) increases, the better the fixed parameters efficiency for
Biological treatment process was conducted in the organic material degradation is. The only first level
Erlenmeyers with a 0.5 L nominal capacity, containing a variable that showed influences on photocatalytic
Qair= 0.75 L min-1 air diffuser. Dairy effluent diluted efficiency was the TiO2 phase used in this study.
with distilled water (1:1 v/v) was employed to optimize According to Carp, Rodriguez and co-workers [16]-
the biodegradation system. The second used design [22], the anatase crystalline form presents a higher
corresponded to a 23 Full Factorial having the following photocatalytic activity due to the larger band gap (3.23
as input variables: eV, 384 nm) when compared to the rutile form (3.02 eV,
1) The sludge concentration (32.0 or 64.0 mg biomass 411nm). This smaller band gap of rutile form frequently
L-1); favors the electron/avoidance recombination. This factor,
2) Sludge conditioning (with or without lactose); combined with TiO2 effects of "superhydrophilicity"
3) Nutrient solution (addition or not) to obtain the best when fixed in a catalytic bed and doped with silicon
parameters for the biological system. oxide [16]-[23], have been considered the main causes
The response variable was also the percentage of the for higher photocatalytic activity in the anatase[16]. As
reduced oxygen chemical demand (% CODreduced). described before, it was observed that the best
The nutrient solution was prepared by adding 1.0 ml degradation results were obtained in photocatalyzed
of the solutions of ferric chloride (0.25 mg L-1), calcium processes with TiO2 in anatase form compared to those
chloride (36.42 mg L-1) and magnesium sulfate (22.5 mg performed with TiO2 in the rutile form.The hydrogen
L-1) in 1.0 L of distilled water. For the sludge potential (pH) is considered an important variable in the
conditioning activated with lactose, periodic additions of evaluation of photocatalytic activity in aqueous medium,
20 mL of lactose solution (1.0 mol L-1) were done with a influenced in the adsorption and dissociation of the
total of 4 additions within 20 conditioning days. substrate [16]-[22]. Banu and co-workers [24] reported
the effects of pH on the photocatalytic activity with TiO2
powder (DuPont) for the degradation of the synthetic
III. Results and Discussion dairy effluent, previously treated in upflow anaerobic
sludge blanket reactor (UASB). An increase of the
III.1. Advanced Oxidative Process (TiO2/UVsolar)
degradation rate was observed as pH increased, showing
A fractional factorial design to evaluate the a maximum of degradation at pH 5. Based on this
investigated variables of AOP process was applied. The experimental design, a statistical model that predicted the
response factor was the percentage of CODreduced. photocatalytic process behavior was proposed [19]. The
For each experimental condition of the fractional following Eq. (1) represents the obtained modeling, with
factorial design, 2.0 L of the pre-treated effluent were % CODreduced as a response factor:
collected and subjected to biological treatment in the
previously optimized system. Results concerning the % = 20.3838 + 3.30875 ∙ +
optimization of biodegradation system will be discussed −4.37875 ∙ − 3.71625 ∙ + 2.4625 ∙ + (1)
at the following session. For a better interpretation of −6.55375 ∙ − 9.24125 ∙ + 5.07125 ∙
effects and interactions of these assessed parameters for
the photocatalytic system optimization, the analysis of where factor A refers to reaction time. Factor B
variance (ANOVA) process was performed (Table I). represents TiO2 phase. Factor C refers to medium acidity
From the analysis of variance and the proposed model, (pH) and factor D is the organic load dilution rate in the
it is observed that interaction effects were analytically medium. The proposed model showed deviation (S),
more significant on the response signal (99.9% correlation coefficient (R2) and adjusted correlation
confidence), in relation to the main effects (93.8% coefficient (R2adjusted) equal to 7.45, 87.94% and 77.39%,
confidence). These results were confirmed by Pareto respectively. By testing the proposed model, Fig. 2 can
analysis with 95% confidence, which shows significant be elaborated. This contains the relation between the
influence of three interactions: reaction time/pH, reaction proposal of the model and the analytical data, observing
time/TiO2; reaction time/effluent and a main effect (TiO2 a good statistical model adjustment.
296
60
Prediction by statistic modelling
Experimental data
It is observed that the conditioning promoted a
removal efficiency of about 60 % CODreduced, while
50 processes conducted without previous conditioning of the
40
sludge presented an organic load removal efficiency of
%CODreduced
approximately 40%. Similar behaviors were observed in

30
the second interaction order involving "prior
20
conditioning” factor. To determine the degree of
confidence of factors studied in the aeration pond system
10
optimization a study of analysis of variance has been
0 carried out (Table III). With full factorial and replicas, it
0 1 2 3 4 5 6 7 8 9
4-1
was estimated the deviation for deeming the effects,
Fractional Design of Experiment (2 ) - AOP (TiO 2\UV solar)
beyond t = 2.306. The main effects refer to factors of 1st
Fig. 2. Evaluation of the statistical model proposed for predicting the order investigated through the full factorial design (23)
%CODreduced in the photo-reactor compared with data obtained during for activated sludge system evaluation. With a 99.9%
photodegradation tests confidence (p ≤ 0.001), it was found that the main effects
were those which had more influence on the reduction of
According to results obtained, analysis of variance and the organic load by APS. According to the obtained
prediction model, the best parameters for the system results a statistical model of factorial experiment for the
degradation by AOP presented a three-hour reaction process conducted via APS was proposed. The equation
time; TiO2 anatase phase; performance of the process in (2) represents the model obtained with % CODreduced as a
acidic medium (pH 5) and organic load of effluent in response factor. The proposed model (S), (R2) and
natura. We found a %CODreduced by 54.77% ± 1.66 and a (R2adjusted) equal to 5.716, 92.13% and 85.25%,
reduction percentage of the total organic carbon by 38.37 respectively:
± 0.37%.
% = 48.8250 − 2.4375 ∙ +
III.2. Parameters Evaluation of Biological +0.6875 ∙ + 13.0875 ∙ + 1.0 ∙ + (2)
Treatment System +3.05 ∙ + 1.675 ∙ − 0.7375 ∙
A full factorial design (23) to determine the best Factor A refers to the nutrients addition. Factor B is
parameters of aeration pond system (APS) was applied. the amount of sludge used at the beginning of the
For a better evaluation of the investigated parameters, biodegradation while C factor refers to the prior
results were reported in terms of (%) CODreduced (Table conditioning of the sludge in a medium containing
II). Analysis of standard deviations for each duplicate lactose. By testing the proposed model it was possible to
shows a good reproducibility of the experimental data. It elaborate Fig. 3, containing the relation between what the
is also noted that microbial flora, previously conditioned model proposed and the analytical data. We could
in lactose, favored a greater efficiency for the observe a good adjustment of the statistical model.
degradation of the tested effluent. There is a major According to the data obtained from the full factorial
influence on previous conditioning of digesting culture in design (23) the best APS conditions used for the
lactose, compared to other effects and interactions. treatment of the pretreated effluent by AOP are found in
Table IV.
TABLE II
RESULTS IN % CODREDUCED OF THE FULL FACTORIAL DESIGN (23) TO DETERMINE THE APS FACTORS
Sludge initial Early Conditioning % CODreduced in 72 h
Experiment Nutrients
concentration (mg L-1) in lactose (1st) batch (2nd) batch Average Deviation
1 Without 32.0 Without 46.4 41.5 43.95 3.46
2 With 32.0 Without 34.9 24.1 29.50 7.64
3 Without 64.0 Without 46.1 30.9 38.50 10.75
4 With 64.0 Without 32.8 29.2 31.00 2.55
5 Without 32.0 With 63.3 55.1 59.20 5.80
6 With 32.0 With 63.1 56.7 59.90 4.53
7 Without 64.0 With 66.0 60.8 63.40 3.68
8 With 64.0 With 66.0 64.3 65.15 1.20
TABLE III
ANALYSIS OF VARIANCE OF RESULTS OBTAINED ACCORDING TO FACTORIAL DESIGN (23) FOR AERATION POND SYSTEM EVALUATION
Sources Gl1 SQ2 MQ3 F4 p5
Main effects 3 2843.15 947.716 29.01 0.000
Interaction (2nd order) 3 209.73 69.910 2.14 0.173
Interaction (3rd order) 1 8.70 8.702 0.27 0.620
Residual error 8 261.39 32.674
Total 15
1
degrees of flexibility; 2quadratic sum; 3quadratic average; 4F Factor; 5degree of confidence
297
TABLE IV After photo-treatment, the effluent samples obtained

PARAMETERS USED FOR CONDUCTING THE BIODEGRADATION
under different conditions of the fractional factorial
PROCESS BY AERATION POND SYSTEM
Variables and adopted
design (24-1) were subjected to the biodegradation system
Parameters by optimized APS. Results were reported in terms of
conditions
Volume of effluent 0,5 L %COD reduced and can be found in Table VI.
Reaction Time 72 hours A greater efficiency of the organic load
-1)
Sludge initial Concentration (mg L 32.0 (~ 25.0 mL)
Medium for digesting flora In lactose
biodegradation for samples coming from the
Nutrients Without addition pretreatment related to experimental conditions 1, 2, 6, 7
Aeration 0.75 L min-1 of AOP-based treatment was observed. Under these
conditions, the used catalyst bed consisted of TiO2 in
Prediction by statistic modelling
Experimental Data (APS) anatase phase.
80
TABLE V
70
RESULTS OF COD, TOC AND %COD REDUCED FOR THE DAIRY
60 EFFLUENT TREATMENT IN THE HYBRID SYSTEM WITHOUT CATALYST
BED OF TIO2 (N = 2).
%CODreduced
50
Factors Initial organic load After AOP After APS
40
COD (mg L-1) 2186.4 ± 81.8 2161.4 ± 70.3 1613.4 ± 53.5
TOC (mg L-1) 612.9 ± 3,0 606.8 ± 2.0 466.3 ± 2.7
30 % CODreduced - 1.14 ± 1.27 23.91 ± 0.81
20
TABLE VI
10 PERCENTAGE OF CODreduced IN PRE-TREATED DAIRY EFFLUENT
ACCORDING TO THE CONDITIONS OF FRACTIONAL FACTORIAL DESIGN
0 (24-1) AND SUBSEQUENT BIOLOGICL TREATMENT
0 1 2 3 4 5 6 7 8 9
IN OPTIMIZED APS (n = 2)
3
Desig of Experiment (2 ) - biodegradation optimization % COD(reduced)
Experimental After AOP After APS
Fig. 3. Evaluation of the statistical model proposed for predicting Conditions Average Deviation Average Deviation
%CODreduced in the APS compared to the obtained data during the 1 14.45 2.90 86.19 0.50
biodegradation tests 2 12.30 2.40 78.20 1.38
3 27.70 18.81 55.32 1.50
4 18.50 2.97 30.92 1.51
III.3. Evaluation of the Hybrid System (Advanced 5 7.02 3.27 30.31 2.18
Oxidation Process – Aeration Pond System) 6 17.50 7.07 83.94 2.33
7 54.77 1.66 89.41 1.30
After evaluation and adjustment of the aeration pond 8 10.80 1.13 29.47 1.69
system, the samples coming from the photocatalytic
treatment were unfrozen, and 0.5 L was used for each Using only the best photocatalysis experimental
AOP experimental condition for the biodegradation in condition (experimental condition 7 - Reaction time: 3
the biological optimized system. After 72 hours of hours; phase of TiO2: anatase; pH:5; effluent organic
treatment by APS, we observed the COD reduction by load: in natura) a %CODreduced by 54.77 ± 1.66 and a
25.19 ± 4.91% with respect to the initial load in the total organic carbon percentage reduction by 38.37 ±
biological system. Obtained results of this hybrid system 0.37% were observed. Only the degradation obtained by
treatment are presented in terms of COD, TOC and photocatalysis was higher than the treatment obtained by
%CODreduced (Table V). Seventy and two hours after APS to treat the same kind of effluent (effluent in
initiating the processing operation, the conjugated system natura). Below it is found an analytical sequence used
was analyzed and it was observed an overall efficiency of for the dairy effluent biodegradation coming from the
26.01 ± 5.23% in the reduction of the organic load. pre-treatment with AOP for the experimental condition 7
Similar behaviors were observed by Ndegwa, Salazar (Table VII).
and co-workers during the assessment of different
TABLE VII
parameters that influence the biodegradability of dairy CINETIC DEGRADATION OBSERVED OF PRETREATED DAIRY EFFLUENT
effluents under conditions with and without oxygen BY OPTIMIZED AOP DURING BIOTREATMENT UNDER THE BEST
supply and depending on the initial organic load for CONDITION OF BIOLOGICAL PROCESS
digestion [25]-[26]. Reaction time (h) pH T (ºC) COD (mg L-1)
The authors observed that the higher the concentration 0 5.01 ± 0.02 25.8 ± 0.9 1723.4 ± 34.6
24 6.87 ± 0.01 26.3 ± 0.1 1710.4 ± 45.6
of lactose or oil and grease, the longer the adaptation and 48 7.74± 0.03 28.4 ± 0.7 620.9 ± 16.3
digestion time of organic material present in the dairy 72 8.12 ± 0.02 25.9 ± 0.2 236.1 ± 2.2
effluent is, and in anaerobic conditions, the digestion
time could be more than 72 h. These results highlight the Results obtained from the treatment of this hybrid
need to perform a pre-treatment of the dairy effluent, system in the best conditions are presented in terms of
allowing a higher biodegradation if it is submitted to a COD, TOC and %CODreduced (Table VIII). After
biological treatment, and in a shorter hydraulic retention photochemical treatment, COT/COD ratio increased
time. from 0.26 to 0.35.
298
TABLE VIII aeration pond biodegradation system, present a great and

RESULTS OF COD, COT AND %CODreduced FOR THE DAIRY EFFLUENT
potentially hybrid system for treating effluents and
TREATMENT IN THE HYBRID SYSTEM WITH HIGHER DEGRADATION
EFFICIENCY OF THE ORGANIC LOAD (n = 2) wastewater in the dairy sector.
Factors Initial organic load After AOP After APS
COD (mg L-1) 3782.5 ± 37.6 1710.4 ± 0.3 236.8 ± 0.1
COT (mg L-1) 983.6 ± 2.6 606.8 ± 2.0 28.3 ± 0.3 Acknowledgements
% CODreduced - 54.8 ± 2.3 93.7 ± 0.1
The authors thank to CAPES (Coordination for the
Due to this observation, it was found that the Improvement of the Higher Education Person) for the
pretreatment system with AOP (TiO2 UVsolar) increased financial support, Prof. MSc. Gerônimo V. Tagliaferro
the dairy effluent biodegradability ratio. This fact may from Yakult and Antônio Rodriguez de Lima, Eng. for
have contributed to increasing the degradation efficiency providing the effluent used in this project and for the
of the biological system employed. It was also observed technical support.
that the digester flora needed a 24-hour period to adapt
itself to the environment. After adjustment, significant
References
reductions of the organic load were shown. When
evaluating the hybrid system (Table VIII) to reduce [1] M. Vourch, B. Balannec, B. Chaufer, G. Dorange. Treatment of
effluent organic load, it was possible to obtain dairy industry wastewater by reverse osmosis for water reuse.
Desalination,219(2008) 190 – 202.
%CODreduced equal to 93.7% ± 0.1% and % of reduced [2] M. C. Pinho, K. B. Siqueira, E. S. Mercês, Produção de leite teve
total organic carbon of 97.12 ± 0.04% after 72 hours of aumento de 4,5% em 2011. Panorama do Leite – Embrapa Gado
reaction. de Leite (2012) 5 – 6.
[3] Instituto Brasileiro de Geografia e Estatística – IBGE. In:
INDICADORES IBGE: Estatística da Produção
Agropecuária/Junho de 2012.(2012) 1-5.
IV. Conclusion [4] B. Demirel, O.Yenigun, T. T. Onay. Anaerobic treatment of dairy
wastewater: a review. ProcessBiochemistry, (8) 40 (2005) 2583 –
The use of an open photochemical reactor showed 2595.
effluent evaporation during the oxidative process [5] M. C. M. R. Leal, D. M. G. Freire, M. C. Cammarota, G. L.
mischaracterizing the degradation kinetics. By evaluating Sant’AnnaJr,Effect of enzymatic hidrolysis on anaerobic
the influence of the parameters investigated in AOP, with treatment of dairy wastewater. ProcessBiochemistry (5) 41 (2006)
1173 – 1178.
the use of a plate without TiO2 presence (blank test), it [6] R. D. Villa, M. R. A. Silva, R. F. P. Nogueira, Potencial de
was found that the addition of ~ 180 mL distilled water aplicação do processo foto-fenton/solar como pré-tratamento de
to correct the evaporation rate did not alter the reduction efluente da indústria de laticínios. Química Nova (8)30 (2007)
of COD values. 1799 – 1803.
[7] P. Abreu, E. L. Pereira, C. M. M. Campos, F. L. Naves,
Also, the blank test proved a low reaction yield of Photocatalytic Oxidation Process (UV/H2O2/ZnO) in the treatment
organic compounds photolysis in the dairy effluent, and sterilization of dairy wastewater. ActaScientiarum, (1) 35
which did not interfere on the results of AOP statistical (2013) 75-81.
design. The fractional factorial design (DOE 24-1) [8] S. Lansing, J. F. Martin. Use of an ecological treatment system
(ETS) for removal of nutrients from dairy wastewater. Ecological
permitted to evaluate and to determine the best analytical Engineering (3) 28 (2006) 235 – 245.
conditions of Advanced Oxidation Processes. The [9] J. Morgan, J. F. Martin, Performance of an ecological treatment
experimental condition 7 of factorial design consisting of system at three strengths of dairy wastewater loading. Ecological
three-hour reaction, TiO2anatase, pH 5 and effluent in Engineering, (3-4) 33 (2008) 195 – 209.
[10] T. H. Bae, S. S. Han, T. M.Tak, Membrane sequencing batch
natura showed the best result of organic material reactor system for the treatment of dairy industry wastewater.
degradation in the dairy effluent, resulting in a ProcessBiochemistry, (2) 39 (2003) 221 – 231.
%CODreduced of 54.77% ± 1.66% and a COT % of 38.37 [11] A. A. Mendes, E. B. Pereira, H. F. Castro, Biodegradação de
± 0.37%. Through a full factorial design (DOE 23), the Águas Residuárias do Laticínios Provenientemente Tratadas por
Lipases.Brazilian Journal of Food Technology (2) 9 (2006) 143-
best analytical conditions of the biological system for 149.
pretreated effluent treatment with AOP were the sludge [12] W. Janczukowicz, M. Zielinski, M. Debowski, Biodegradability
initial concentration equal to 32,0 mg L-1 and prior evaluation of dairy effluents originated in selected sections of
conditioning of the digesting culture in lactose for a 72- dairy production, Bioresource Technology (10)99 (2008) 4199 –
4205.
hour-reaction time. This resulted in a % CODreduced by [13] X. Domènech, W. F.Jardim,M. I. Litter, Procesos avanzados de
59.20 ± 5.80 %, representing a COD reduction of 3342.2 oxidación para la eliminación de contaminantes.In: Eliminación
± 184.2 mg L-1 for 1359,4 ± 32.2 mg L-1. The ‘prior de Contaminantes por Fotocatálisis Heterogênea. La Plata: Rede
conditioning’ factor was the most significant for the CYTED, 2001.Cap. 1
[14] S. H. Pan, K. V. Lo, P. H. Liao,H. Schreier, Microwave
increase of organic material degradation present in the pretreatment for enhencement of phosphorus release from dairy
samples. The use of optimized AOP-ASS treatment manure.Journal of Environmental Science and Health, Part B,41
systems showed a meaningful reduction of COD (2006) 451 – 458.
percentage equal to 93.7 ± 0.1% and TOC % by 97.12 ± [15] A. Fujishima, T. N. Rao, D. A. Tryk. Titanium dioxide
photocatalysis. Journal of Photochemistry and Photobiology C:
0.04%. These results combined to the treatment with Photochemistry Reviews. (1) 1 (2000) 1 – 21.
Advanced Oxidation Process and based on [16] O. Carp, C. L. Huisman, A. Reller, Photoinduced reactivity of
heterogeneous photocatalysis (TiO2/UVsolar), with titanium dioxide,Progress in Solid State Chemistry, (1) 32 (2004)
33 – 177.
299
[17] J. S. Carrocci, R. Y. Mori, O. L. C. Guimarães, R. F. S. Salazar, Marcos Fernandes de Oliveira, São Paulo, SP,
M. F. Oliveira, A. L. C. Peixoto, H. J. Izário Filho, Application of Brazil (São Paulo - Brazil, 30/03/1966). Ph.D.
heterogeneous catalysis with TiO2 photo irradiated by sunlight degree in Materials Engineering at Polytechnic
and latter activated sludge system for the reduction of School of the University of São Paulo-USP,
vinasseoranic load. Engineering4 (2012) 746-760. Brazil (2006). Post Graduated Industrial
[18] M. P. D. Costa, J. V. S.Pancotto, M. A. K. de Alcântara, A. S. Administration (Carlos Vanzolini Foundation –
Cavalcanti, O. L. C. Guimarães, H. J. Izário Filho. Combination USP).Graduated in Industrial Chemist in 1990
of sunlight irradiated oxidative processes for landfill leachate: at Guarulhos University, São Paulo,
heterogeneous catalysis (TiO2) versus homogeneous catalysis Brazil.Current Specialist Chemist atAxalta Coating Systems (ex-
(H2O2) (Portuguese), Ambiente&Água-An Interdisciplinary DuPont Performance Coatings).Since 1988 works in OEM coatings
Journal of Applied Science(1) 8 (2013) 290-306. R&D for car manufacturers (Toyota, Honda, Mitsubishi, Volvo,
[19] A. L. C. Peixoto, R. A. Brito, R. F. S. Salazar, O. L. C. General Motors, Volkswagen and Ford).Includes waterborne
Guimarães, H. J.Izário Filho, Predição da demanda química de technology and pigment dispersions formulation experience. On 2002
oxigênio em chorume maduro contendo reagente de fenton, por and 2003, received the DuPont Safety Health Environment Prize for
meio de modelo matemático empírico gerado com planejamento projects to reduce the volatile organic content in OEM coatings.On
fatorial completo,Química Nova (7) 31(2008) 1641 - 1647. 2004, 2006 and 2010 he received the 9th, 10th e 12th Science and
[20] R. F. S. Salazar, H. J. IzárioFilho, Aplicação de processo Technology Paint Award from Brazilian Paint Manufacturers
oxidativo avançado baseado em fotocatálise heterogênea Association (ABRAFATI) and Petrobras, for works with organosilanes
(TiO2/UVsolar) para o pré-tratamento de afluente to reduce the corrosion rate in automotive carbon steel, carbon black
lácteo, Augmdomus1 (2009) 27-44. pigments studies to avoid corrosion in steel tanks, and magnesium
[21] APHA-AWWA. Standart methods for examination of water and oxide as ecological corrosion inhibitor, respectively. One of the authors
wastewater. 20.ed. Washington DC: American Public Health of the book Tintas, Ciência e Tecnologia, 2009 edition (Coatings -
Association – American Water works association, 1998. Science and Technology) with chapters about carbon black and
[22] J. Rodrìguez, R. J. Candal, J. Solís,W. Estrada, M. A. Blesa, El pigments dispersions technology. Researcher about silanes as corrosion
fotocatalizador: síntesis, propriedades y limitaciones.Solar Safe inhibitors (University of São Paulo – USP).Professor of Materials
Water (2006) 135 – 152. Science for Coatings Technology Post Graduation Course - Oswaldo
[23] D. M. Tobaldi, A. Tucci, G. Camera-Roda, G. Baldi, L Cruz College – São Paulo, Brazil.
Esposito,Photocatalytic activity for exposed building E-mail: marcosfernandes1@yahoo.com.br
materials.Journal of the European Ceramic Society28, (2008)
2645 – 2652. Marco Aurélio Kondracki de Alcântara.
[24] J. R. Banu, S. Anadan, S. Kaliappan, I. T. Yeom, Treatment of Postdoctoral at the Center for Research and
dairy wastewater using anerobic and solar photocatalytic methods. Development of Soils and Environmental
Solar Energy (9) 82 (2008) 812 – 819. Resources of the InstitutoAgronômico de
[25] P. M. Ndegwa, L. Wang, V. K. Vaddella, Potential strategies for Campinas. He is currently a Professor in the
process control and monitoring of stabilization of dairy School of Engineering of Lorena, University of
wastewaters in batch aerobic treatment systems. Process São Paulo. Areas of expertise: Soils, with
Biochemistry (9) 42 (2007) 1272 – 1278. emphasis on soil pollution, environmental
[26] R. F. S. Salazar, J. S. Carrocci, H. J. Izário Filho, Employment of impacts, focus turned to analytical chemistry. Address: Universidade de
factorial design to evaluate the organic loading and aeration of São Paulo. Escola de Engenharia de Lorena. Departmento de Ciências
biological systems in the degradation of dairy Básicas e Ambientais. Estrada Municipal do Campinho, s/n. Lorena –
wastewater, Ambiente&Água-An Interdisciplinary Journal of SP. ZIP 12.602-810.
Applied Science(3) 6 (2011) 98-109, 2011. E-mail: marko@usp.br
Hélcio José Izário Filho. Lorena – SP, Brazil.

Authors’ information Doctor of Science in Chemistry from by the
State University of Campinas. He is currently a
Rodrigo Fernando dos Santos Salazar. Santa Professor in the School of Engineering of
Maria – RS, Brazil (Rio de Janeiro - Brazil, Lorena - University de São Paulo. He has
07/11/1982). Bachelor in Biochemical experience in chemistry with emphasis in
Engineering at School of Engineering of Lorena analytical instrumentation, specifically
USP (Brazil, 2006), MSc in Chemical absorption spectrometry and atomic emission
Engineering at University of São Paulo (Brazil, spectrophotometry, acting on the following themes: characterization of
2009). PhD in Science with emphasis in the inorganic environment (analytical characterization of solid and
Analytical Chemistry at Federal University of liquid waste and Advanced Oxidation Processes). Escola de Engenharia
São Carlos (Brazil, 2013). He has experience in Chemical Engineering de Lorena, University of São Paulo. Departmento de Engenharia
with emphasis on wastewater treatment, and Chemistry, with an Química. Estrada Municipal do Campinho, s/n. Lorena – SP. ZIP
emphasis on Analytical Chemistry.He has publications on the following 12.602-810.
topics: atomic absorption spectrometry, trace elements, ICP-MS, E-mail: Hélcio@dequi.eel.usp.br
physicochemical characterization, oxidative processes advanced
(POAs) and wastewater treatment.Prof. Salazar is Assistant Professor at
Franciscan University (UNIFRA, Conjunto II, Centro
UniversitárioFranciscano, Silva Jardim St, 1323 – ZIP97010-491 –
Santa Maria – RS – Brazil) and in the State University of Rio Grande
do Sul (UERGS, Andrade Neves St, 336 – ZIP 98025-810 – Cruz Alta
– RS).
E-mails: r.f.s.salazar@gmail.com
salazar@unifra.br
300
ISSN 2035-1755
2035 1755 July 2013
A Research Note on Oil Extraction ffrom

rom Jatropha Curcas Seeds
Akhihiero E. Thelma
Abstract - The non-edible

non edible seed oil of Jatropha curcas is gaining so much popularity in its usage
as alternative fuel and in biodiesel production. Jatropha seed oil extraction is an inevitable unit
operation process prior to its usage in biodiesel production, soap making, candle making and
other uses. Oil of Jatropha seeds have been extracted by mechanical method with a hydraulic
press and also by solvent extraction using nn-hexane
hexane as a solvent. The percentage yield of oil
obtained by hydraulic press is 34.5% while that with solvent extraction is 44.313%. The pphysical
hysical
and chemical properties of the oil are also highlighted. Copyright © 2013 Praise Worthy Prize
S.r.l. - All rights reserved.
Keywords:
Keywords Jatropha,
Jatropha, Biodiesel, Extraction, N-Hexane
N Hexane
I. Introduction In the case of mechanical extraction using a hydraulic

press or screw press the cruscrushed
hed seeds are stock into
Jatropha is a genus of over 170 plants from the sacks and heated to a temperature of over 50 OC to enable
Euphorbiaceae or spurge family, native to the central the oil flow out easily from the seed mass for maximum
America but commonly found and utilized across most of oil yield. The use of organic solvents such as n hexane,
the tropical and subtropical regions of the world [1]. isohexane, methanol or ethanol to dissolve out oil from
Among the different species of Jatropha, JatrophaJatropha seeds popularly known as solvent extraction is another
curcas has a wide range of uses and promises various way of extracting vegetable oils from plants. Solvent
significant ben
benefits
efits to human and industry [1]
[1]--[3].
[3]. Due to extraction involves the use of a soxhlet extractor.
the fact that Jatropha curcas
curcas belongs to the family In soxhlet extraction, the volatile solvent for example
euphorbiacease
euphorbiacease,, it is closely related to some other n-hexane
n hexane is vapourize into the top of the cha chamber
mber
important cultivated plants like rubber tree, castor etc. it containing the crushed material to be extracted, where it
has been reported that Jatropha ccurcas urcas will help in is condensed and falls down on the bag containing the
meeting the challenges of global biofuel demand (37 material, in the process dissolving and carrying the oil.
billion gallons) by 2016 [4]. The mixture of solvent and oil is leached out of the
Presently, the Nigerian government is showing great cake in the round bottomed flask placed on a heated
interest in biofuel plants especially Jatropha ccurcas
urcas with mantle which provides heating for the solvent to
the aim of reducing over dependence on fossil fuel and continuously vapourise until extraction is completed.
also reduce the environmental hazards caused by fossil Solvent extraction generally consist of a sequence of
fuel utilization, together with
with ultimate need to encourage four operations. Physical removal of oil from seed, husk
micro,
icro, small or medium enterprises in biofuel production etc in the extractor, desolventising
desolventising–– toasting of the dede--
for sustainable development. The seed oil of Jatropha oiled meal, distillation to remove the solvent of the
curcas can be used directly or blended with petrol diesel extracted oil and recovery of the sol
solvent
vent for re
re--use
use in the
to fuel diesel engines [5], or may be converted
conver ed to reactor.
biodiesel by transesterification [6 [6].
]. The process of Other methods of extracting oils from plants includes
biodiesel production from vegetable oil involves two steam distillation implored
implored for extract
extracting
ing essential oil [7].
[7].
phases. The first phase is the extraction of crude oil from Vegetable oils extracted from plants contains 90 90-98%
98%
the plant material and the second is the transesterification triglycerides [8],
[8], [9]. They sometimes contains nn-alkanes
alkanes
of the crude oil into biodiesel [3]. and polynuclear aromatic hydrocarbons [10],[11].
[10] [11].
The aim of oil extraction ffromrom plants is to separate Triglycerides are esters of three fatty acid and one
cellular or fluid lipids from other constituents such as glycerol [12]. When vegetable oil are reacted with
proteins, polysaccharides, small molecules like amino methanol in the presence of a suitable catalyst, methyl
acids, sugars and so on. esters (biodiesel) are produce with gglycerol
lycerol molecule as a
Commercially, vegetable oils are cold pressed; byproduct [13],
[13], [14]. Vegetable oils extracted with
expeller pressed, extracted by hexane and/or ot other
her mechanical means using hydraulic or screw press may be
solvents and sometimes extracted using liquid carbon preferred for biodiesel production than those extracted
dioxide as a solvent. Before extraction the oil seeds are with solvent due to lower cost of extraction and
husked and cleaned and then crushed into powdering avoidance of the risk of solvent contamination. This
form.
Manuscript received and revised June 2013, accepted July 2013

Manuscript Copyright © 2013 Praise Worthy Prize S.r.l. - All rights reserved
301
Akhihiero E. Thelma
paper reports the percentage yield of oil extracted from into the cup of the tester to the mark and it was followed
Jatropha curcas seeds using the hydraulic press and by the replacement of the cup and the cup cover with the
solvent extraction using n-hexane as solvent. left hand pointing towards the left front corner of the test
The properties of the oil extracted by mechanical unit. Stirrer driver drive was fitted into the tester properly
mean together with the fatty acids composition are and it was followed by the connection of the resistance
highlighted. The average molecular weight of the oil was thermometer probe. Flame and the pilot light were
also calculated from its saponification and acid values. carried out by lighting and the draught screen was closed.
The tester was put on and the heater temperature was
regulated and switch for the stirrer was on
II. Materials Used for Experiments simultaneously with the tester for homogeneity. The
Materials used in this research work were soxhlet flash occurs when a large flame was observed on the cup
extractor, heating mantle, round bottomed flask, and the temperature at which this occurs was recorded as
hydraulic press, weighing balance, storage bottles, n- the flash point for the oil sample.
hexane, gas chromatography and other chemicals of
analytical grades such as standard potassium
IV. Results and Discussion
hydroxide(0.1M), Benzene, Neutral alcohol, alcoholic
potassium hydioxide (0.5M), hydrochloric acid (0.5M), The results obtained from extraction processes are
Sodium Carbonate Solution (0.5M), Conical flasks, shown below: For hydraulic press:
Beakers, Deionised water, Pensky-Martin closed cup Total weight of crushed seeds = 673.9 g;
tester. Weight of Jatropha curcas oil = 232.6 g;
Volume of Jatroph curcasa oil = 260 ml;
Percentage yield=
III. Experimental Methods
232.6g 100
Mature Jatropha seeds harvested from a farm in = × = 34.5%
southern Nigeria were dried, dehulled and grinded. 673.9g 1
The ground seeds were weighed, packed into sacks
For hot solvent extraction:
and subjected to pressure under the hydraulic press to
Weight of crushed Jatropha seed = 66.2379g;
extract the oil. The hydraulic press used was locally
Weight of beaker = 98.1800 g;
fabricated from local materials such as wood and
wrought iron from unserviceable vehicle. The oil Weight of beaker + Jatropha oil = 169.6170g;
extracted was stored in a bottle. For the solvent Weight of Jatropha seed oil = 73.6652 g;
extraction, dehulled Jatropha curcas seeds were weighed Weight of seed cake = 92.5727 g;
and grinded. The ground seeds were packed into a small Percentage oil in hexane extract =
sack used to replace the thimble of the soxhlet extractor. 73.6652 100
The soxhlet carrying the thimble is mounted on the = × = 44.313%
round bottomed flask sitted on the heating mantle. 166.2379 1
The extraction began at a temperature of 550C, For masteration:
pressure of 1 atmosphere and for 6 hours when the liquid Weight of crushed seeds = 100 g;
falling from the thimble into the flask was colourless Weight of beaker = 47.21 g;
showing that extraction was completed. At the end of Weight of beaker + Jatropha oil = 58.27 g;
extraction, the solution in the round bottomed flask (oil + Weight of Jatropha oil = 11.06 g;
hexane) was distilled using a rotary evaporator to recover Percentage oil yield =
the solvent. In the absence of a rotary evaporator, the
solution can be evaporated under solar heat at 11.06 100
atmospheric pressure. The solvent vaporizes leaving the = × = 11.06%
100 1
oil behind in the container. This process is called hot
extraction. Cold extraction with hexane (masteration TABLE I
FATTY ACID PROFILE OF JATROPHA CURCAS SEED OIL
method) was also carried out. Hundred grams of the Jatropha curcas seed oil
seeds were soaked in 400ml of n hexane for 48 hours (2 Fatty Acids % composition
days). The solution was filtered and evaporated under the Myristic acid C14:0 0.260
sun to recover the oil. The gas chromatography was used Palmitic acid C16:0 14.398
Palmitoleic acid C16:1 0.704
to analyse the oil for its fatty acid content. The chemical
Margaric acid C17:0 0.332
properties of the oil including its saponification value Stearic acid C18:0 6.65
and acid value were determined from standard analytical Oleic acid C18:1 43.39
methods [15]. The flash point of the oil sample was Linoleic acid C18:2 32.984
Linolenic acid C18:3 0.351
carried out using the pensky-martens closed cup tester
Arachidic acid C20:0 0.218
(Kehler-model k-16270). The oil sample was dried in the Behenic acid C22:0 0.117
laboratory on the bench to remove the traces of the water Erucic acid C22:1 0.506
moisture in the sample. The dried sample was poured Lignocenic C24:01 0.087
302
Akhihiero E. Thelma
TABLE II of mechanical extracted oil = 3.079 mgKOH/g oil as

FUEL PROPERTIES OF JATROPHA CURCAS SEED OIL
shown in Table II.
Fresh Seed oil
Specific gravity 0.919
Hence the average molecular weight of the mechanical
Density 0.919 g/dm3 extracted oil is found to be:
Acid value 3.079
% free fatty acid 1.40% 56.1 × 1000 × 3
Saponification value 182.33mg/g Average molecular weight = =
Kinematic viscosity at 40oC 36.02 cSt 182.33 − 3.079
Moisture 1% = 938.9 g/mol
Flash point 254oC
Pour point 15oC The determination of the average molecular weight of
Due point 0.oC
Temp. of gaseous fuel 23oC
the oil is very important as it helps to know the volume
pH 5.48 or weight of alcohol required for a specific molar ratio of
alcohol to oil in biodiesel reaction. It is therefore so
necessary that the average molecular weight of the oil be
The percentage oil yield obtained from the hydraulic
determined first before biodiesel production process.
press is in agreement with the 28 – 40% oil yield
In case of biodiesel production from Jatropha curcas
reported in literature[16][3] Also according Shweta,
seed oil, the average molecular weight of the oil can be
Vinod, and others, the oil content in Jatropha curcas
easily determined from its saponification and acid values
seeds is 40 – 60% by weight of seed kernel [17]-[19].
rather than many schools, college higher institutions in
The 44.313% yield obtained by hot solvent extraction
Nigeria cannot afford a gas chromatography to determine
is in agreement with Shweta’s report. Others also
the fatty acids content of the oil and hence its average
reported that Jatropha curcas seed oil content is 30 –
molecular weight.
50% by weight of seed [20], [5]. The percentage yield of
Since Jatropha curcas seed oil differ in composition
both the hydraulic press extract and solvent extract agree
from location to location, [10][1] there is the need to
with their report. However the percentage yield of both
always determined the average molecular weight of
the hydraulic press extract and solvent extract falls below
every oil extracted before use in biodiesel production.
the yield obtain by some other authors whose reports
state that the seed could contain up to 50 – 60% oil [21],
[6]. This is so because some seeds could contain more oil V. Conclusion
than others due to geographical location, soil content,
origin, prevailing environmental conditions etc. [10], [1]. The seed oil of our local Jatropha curcas in southern
The percentage yield obtained by masteration method or Nigeria was extracted by mechanical method and
cold solvent extraction is so low and discouraging. chemical method. The oil yield of the mechanical extract
Therefore the use of cold solvent extraction or is 34.5% while the oil yield of the solvent extract is
masteration in extracting oil from Jatropha curcas seeds 44.313%. The average molecular weight of the hydraulic
should be discouraged. The three predominant fatty acids extract as determined from its saponification and acid
in the oil are Oleic, Linoleic and Palmitic acids, with values is 938.9g/mol.
Oleic acid being the most abundant. The most abundant fatty acid in the oil as determined
This is also in agreement with the reports of Shweta by gas chromatographic analysis is Oleic. The high
and others [17] [22]. saponification value of the oil shows its usefulness in
Although the oil yield from the mechanical extraction detergent and soap making. It is recommended that prior
is lower than that from hot solvent extraction, the to biodiesel production from Jatropha curcas oil, the
mechanical extracted oil is better recommended for average molecular weight of the oil can be determined
biodiesel making because it is purer and free from the easily from its saponification and acid values, in the
risk of solvent contamination. The high saponification absence of a gas chromatography which is lacking in
value of the oil is an indication of its usefulness in soaps many colleges and higher institutions in Nigeria.
and detergents making [23]. If the saponification value
and acid value of the oil are known, the average
molecular weight of the oil could be determined using
References
the equation below. [1] Sepidar, S., Zurina Z.A., Robiah Y. and Azhari M. Extraction of
Average molecular weight of: oil from Jatropha seeds, optimization and kinetics. American
Journal of Applied Sciences 6(7): (2009) 1390 – 1395.
[2] Shambhu, V.B., Bhattacharya .T.K., Nayak L.K. and Das S.
56.1 × 1000 × 3 studies on Characterization of Raw Jatropha Oil and its Biodiesel
ℎ oil = [23] with Relevance of Diesel. International Journal of Emerging
−
Technology and Advanced Engineering Vol. 3, Issue 4 (2013) 48
– 54.
where = saponification value [3] Uriarte, A.F. Biofuel from Plants oils. Published by the ASEAN
(mgKOH/g oil); Foundation, Jakarta, Indonesia. (2010) ISBN No. 978-979-19684-
=Acid value (mgKOH/g oil); 1-6.
of mechanical extracted oil = 182.33 mgKOH/g oil; [4] Adebayo, G.B., Ameen O. M and Abass L.T. Physico-Chemical
Properties of biodiesel produced from Jatropha Curcas oil and
303
Akhihiero E. Thelma
fossil diesel. Journal of Microbiology and Biotechnology Authors’ information

Research. 1(1): (2011) 12-16.
[5] Pramanik, K. Properties and use of Jatropha Curcas oil and diesel Department of Chemical Engineering, Delta State University,
fuel blends in compression ignition engine: Renewable Energy, Abraka (Oleh Campus), Nigeria.
28(2). (2003) 239 – 248. E-mail: ejirogheneakhihiero@yahoo.co.uk
[6] Kazi, M.R., Mohammad M., Md. Roknuzzaman and Asadullah Al
Galib Biodiesel from Jatropha oil as an Alternative fuel for diesel Akhihiero Ejiroghene Thelma, Chemical
engine. International Journal of Mechanical and Mechatronics. Engineer, M.Eng, in Chemical Engineering by
Vol. 10, No. 03. (2007) 1-6. University of Benin, Benin City, Nigeria A .Ph.
[7] Akhihiero E.T. , Production of Perfume from Leaves of Ocimum D Student in Chemical Engineering, University
Gratissimum Proceedings of GASAT Africa Conference (2000) of Benin, and lecturer I in Delta State
307-308. University, Nigeria. Main research interests:
[8] Demirbas, A. Global biofuel strategies. Energy Edu Sci Techno. Renewable energy.
17: (2006) 33 – 63.
[9] Pankaj, T., Rayeev K. and Sanjeev G. Transesrification,
Modelling and simulation of Batch Kinetics of Non-edible
vegetable oil for Biodiesel production. Department of Chemical
engineering, ITT Kampur, 208016, India, (2007) 1 – 20.
[10] Akhihiero, E.T., Aluyor E.O., Audu T.O.K., composition of N-
alkanes and Polynuclear aromatic hydiocarbon in Jatropha seed
oil from different location. Accepted for Publication in Journal of
Advanced Material Research, (2012).
[11] Meher, L.C., Sagar, D.V, Naik S.N Technical aspects of biodiesel
production by transesterification – a Review – Renewable and
sustainable energy Reviews, 10: (2006) 248 – 268.
[12] Balat, M. Fuel characteristics and the used of biodiesel as a
transport fuel. Energy sour 28: (2006) 855 – 64.
[13] Gerpen, J. V. Biodiesel Processing a production. Fuel Processing
Technology. 86: (2005) 1097 – 1107.
[14] Das L.M. Potential of Biofuels. Rountable on Sustainable biofuel.
Indian Institute of Technology, Delhi (2008).
[15] AOAC. Official methods of Analysis. Association of Official
Analytical Chemists. 15 Edition. Washington, D.C. USA, 1990.
[16] Alexander, J., Benford, D., A. Cock Burn, Jean – Pierre C., E.
Dogliotti, A. Di Donmenico, M.L. Fernandez-Cruz, Peter Furst, J.
Funk-Gremmels et al, Ricin (from Ricinus Communis) as
undersirable substances in animal feed. Scientific Opinion of the
panel on contaminants in the food chain. The European Food
safety Authority Journal, 726 (2008) 1-38.
[17] Shweta S., Sharma S. and Gupta M.N Biodiesel Preparation by
Lipase-Catalyzed Transesterification of Jatropha Oil. Energy and
Fuels 18 (2004) 154-159.
[18] Vinod K., Babu J., Pramod W. R., Abiu M. and Firdaus J. Cold
active lipase catalyzed production of biodiesel from Jatropha oil
in a solvent free system .J. Chem, Pharm, Res. 3(2): (2011) 226 –
233. ISSN No: 0975 – 7384.
[19] Meka P.K., Tripathi and V, and Singh R.P J. Oleo Sci, 56: (2007)
9-12.
[20] Kandpal, J. B. and Madan M. Jatropha curcas: a renewable
source of energy for meeting future energy needs Review. Energy
6(2): (1995) 159-160.
[21] Abigor, D.R., Okpefa E., Okiy D.A and Uadia P.O. Hydrolysis of
palm oil, palm kernel oil and coconut oil by lipase from Jatropha
curcas L. Proceedings, Oleochemical Conference, PORIM,
International Palm Oil Congress, (1999) 252 – 265.
[22] Siddharth J. and Sharma M.P. Kinetics of acid base catalysed
transesterification of Jatropha Curcas oil. Bioresource
Technology 10: (2010) 7701-7706.
[23] Hawash S., El Diwani G., Abdel Kader E. Optimization of
Biodiesel Production from Jatropha oil by heterogeneous Base
catalysed transesterification. International Journal of Engineering
Science and Technology. Vol 3, No. 6: (2011) 5242 – 5251.
304
ISSN 2035-1755 July 2013
Analysis of the Water Quality in the Water Distribution System,

Based on Logistic Regression
Camelia Gavrila, Alina Elisabeta Sandu, Elena Burchiu
Abstract – In this paper we describe the evaluation of various parameters in establishing their
prognostic value of the water quality in the water distribution system. Our application was
validated using real data from the actual status of distribution networks in Romania. The study
was performed on percentages of the equipment deficits, age equipment and incorrect sample and
the statistical analysis was based on logistic regression. We concluded by using mathematical and
statistical methods, that an accurate age equipment is one of the most important and essential
information in the analysis report. Copyright © 2013 Praise Worthy Prize S.r.l. - All rights
reserved.
Keywords: Logistic Regression, Age Equipment, Water Quality
I. Introduction The logistic regression, however, deals with the case

where the basic random variable Y of interest is a
The pursuit of safe drinking water has long been and dichotomous variable taking the value 1 with probability
remains a major concern of public-health officials and  and the value 0 with probability 1    .
water-treatment operators [1]-[6]. This requires a
thorough understanding of the factors that cause the Such a random variable is called a point-binomial or
deterioration in drinking water, [4]. Bernouilli variable, and it has the simple discrete
For controlling the ever increasing complexity of probability distribution:
these fields, a proper understanding of their processes, as 1 y
well as the ability to reason about them is two important Pr Y  y    y 1    , y  0,1
factors. Moreover, these processes include the effects of
individual as well as random variation. Essentially they Suppose that for the ith individual of a sample
are uncertain; the uncertainties involved render an overall  i  1, 2,...,n  , Yi is a Bernoulli variable with:
understanding hard to achieve and reasoning a daunting
task. 1 yi
Pr Yi  yi    iyi 1   i  , yi  0,1
Models capturing these processes and methods using
these models are thus called for to support decision
making in real-life practice. We need statistics because The logistic regression analysis assumes that the
we want to draw more valid conclusions from limited relationship between  i and the covariate value xi is
amounts of data and significant differences are often described by the logistic function:
masked by experimental imprecision. On the other hand,
the human mind excels at finding patterns and 1
i  , i  1, 2 ,...,n
relationships and tends to generalize too much. 1  exp     0  1 xi  
The basic logistic function is given by:

II. Logistic Regression
Logistic regression allows one to predict a discrete 1
f z 
outcome, such as group membership, from a set of 1  e z
variables that may be continuous, discrete, dichotomous,
or a mix of any of these. Generally, the dependent or where, as in this simple regression model:
response variable is dichotomous, such as
presence/absence or success/failure. Thus, in instances zi   0  1 xi
where the independent variables are a categorical, or a
or, in the multiple regression model:
mix of continuous and categorical, logistic regression is
preferred. The usual regression analysis goal is to k
describe the mean of a dependent variable Y as a zi   0   1 x ji
function of a set of predictor variables. j 1
305
C. Gavrila, A. E. Sandu, E. Burchiu
representing an index of combined risk factors. i k

Under the simple logistic regression model, the ln   0    j x ji
1 i j 1
likelihood function is given by:
y This leads to the likelihood function:

n n exp   0  1 xi   i
L  Pr Yi  yi    , yi  0,1
1  exp   0  1 xi  yi
  i 1  j x ji 
i 1 i 1  exp   k
n 0
L  , yi  0,1
from which we can obtain maximum likelihood estimates
of the parameters  0 and 1 . As mentioned previously,
i 1 1  exp   0  
k
j 1
 j x ji 
the logistic model has been used both extensively and
successfully to describe the probability of developing Also similar to the univariate case, exp  i 
Y  1 some disease over a specified time period as a represents one of the following:
function of a risk factor X . 1. The odds ratio associated with an exposure if X i is
For the case of the logistic regression model, the binary;
logistic function for the probability  i can also be 2. The odds ratio due to a 1-unit increase if X i is
expressed as a linear model in the log scale (of the odds): continuous.
After ̂i and its standard error have been obtained, a
i
ln   0  1 xi 95% confidence interval for the odds ratio above is given
1 i by:
  
exp  ˆ i  1.96 SE ˆ i 

II.1. Multiple Regression Analysis
The effect of some factor on a dependent or response These results are necessary in the effort to identify
variable may be influenced by the presence of other important risk factors for the binary outcome. Of course,
factors through effect modifications (i.e., interactions). before such analyses are done, the problem and the data
Therefore, to provide a more comprehensive analysis, have to be examined carefully. If some of the variables
it is very desirable to consider a large number of factors are highly correlated, one or fewer of the correlated
and sort out which ones are most closely related to the factors are likely to be as good predictors as all of them;
dependent variable. In this section we discuss a information from similar studies also has to be
multivariate method for risk determination. incorporated so as to drop some of these correlated
This method, which is multiple logistic regression explanatory variables.
analysis, involves a linear combination of the The use of products such as X1 X 2 and higher power
explanatory or independent variables; the variables must
be quantitative with particular numerical values for each terms such as X12 may be necessary and can improve the
patient. A covariate or independent variable, such as a goodness of fit. It is important to note that we are
patient characteristic, may be dichotomous, polytomous, assuming a (log) linear regression model, in which, for
or continuous (categorical factors will be represented by example, the odds ratio due to a 1-unit increase in the
dummy variables). value of a continuous X i is independent of x .
Examples of dichotomous covariates are gender and Therefore, if this linearity seems to be violated, the
presence/absence of certain co morbidity. Polytomous incorporation of powers of X i should be seriously
covariates include race and different grades of
considered. The use of products will help in the
symptoms; these can be covered by the use of dummy
investigation of possible effect modifications. Finally,
variables. Continuous covariates include patient age and
there is the messy problem of missing data; most
blood pressure. In many cases, data transformations (e.g.,
packaged programs would delete a subject if one or more
taking the logarithm) may be desirable to satisfy the
covariate values are missing.
linearity assumption.
Suppose that we want to consider k covariates
simultaneously; the simple logistic model of previous III. Results and Discussion
section can easily be generalized and expressed as:
The purpose of this study is to develop a mathematical
1 model that gives a prediction of water quality in
i  , i  1, 2 ,...,n distribution systems, namely, whether or not it complies
  k  with the standards required by law.
1  exp     0 
 
  j x ji  
 The study is based on real and actual data regarding
j 1  the poor structure and the age older than 30 years of
distribution networks in Romania, as well as water
or, equivalently: quality [5].
306
Using the data mentioned above were taken into

account the following parameters: ED - equipment
deficits [%], AE – age equipment [%] and IS - Incorrect
Sample [%]. The data were processed using the SPSS [6]
(Statistical Package for the Social Sciences). This
application was originally developed in 1965 at Stanford
University in California. Since then, it has become the
most widely used data analysis application from the field,
coming to dominate the field. SPSS is updated
continuously, both in terms of the range of statistical
techniques presented and in that of the detail of the
results obtained.
To determine the mathematical tools we can apply, first
we have studied, the repartition of this parameters (if the
variables are or not distributed normal), Figs. 1-3.
Because the variables are not distributed normal, we
propose the utilization of logistic regression.
The logistic regression model the relation between an Fig. 3. Distribution of the variable AE
set of independent variables and an dependent variable
(in this case, the quality of water - % incorrect sample). The logistic regression can be binomial or
multinomial. Binomial logistic regression can be used if
exists just two categories of variables which have to be
predicted. Multinomial logistic regression is used when
there are three or more categories.
The predictors (i.e., independent variables) can be
score variable, nominal variable (or categories) or may
be a combination of these two categories. The best
predictors of the category of one case have B significant
coefficients (i.e., regression coefficients).
The value of B of the logistic regression is applied, in
fact, at the natural logarithm of a value named the odds
ratio, which is the ratio of the frequencies of two
alternative results.
This logarithm is known under the name logit, there
the term logistic regression. The odds ratio it is actually
the probability of belonging of one category and not the
other. Most researchers rarely need to calculate these
values, so that the value logit have more theoretically
importance then practical.
Fig. 1. Distribution of the variable IS SPSS program generates classification tables, these
indicating predict belonging to a certain category, based
on predictor variables. They are a good indicator of the
quality of prediction, because in the table it is precisely
indicated the number of correct classifications. The
logistic regression analysis includes a big number of
„goodness-of-fit” statistics, based on chi-square
coefficient.
This has more functions, but the most important it is
represented by the fact that indicates an improved
correspondence between predicted category
appurtenance and the real appurtenance in a category.
An useful predictor should improve the
correspondence degree between predicted appurtenance
and real categories, which different cases belong. In this
study, were considered the independent variables ED and
AE and the dependent variable IS. In the table below are
present the correlations between variables.
It appears that the correlation (0,385) between AE and
Fig. 2. Distribution of the variable ED IS it is greater than correlation (0,064) between ED and
307
IS, which means that the age of materials older than 30 The values obtained indicate o pretty good prediction.
years influences more the quality of water than poor The performance of the proposed method is illustrated
structure of the distribution system. using three water distribution networks in Ialomita
Using the coefficients from Table II was obtained county, Table V.
following logistic regression equation:
TABLE V
VALIDATION OF THE MODEL USING DATA FROM IALOMITA COUNTY
ln  ODDS _ quality  
AE>30 P P
Locality ED [%]
 5,605  0 ,0208   ED   0 ,105   AE  year[%] computed measured
Slobozia 60,2 64,23 53 53,48
Slobozia
0 0 100 100
ODDS_quality represents the ratio between the quality Noua
of conforming water and the quality of nonconforming Bora 0 0 100 100
water and it is given by the following expression: p - The probability of water quality to comply with the standards
required by law
exp  5, 605  0 ,0208   ED   0 ,105   AE  

IV. Conclusion
TABLE I
CORRELATIONS In order to evaluate the water quality in the water
ED AE IS distribution system, we have applied the Logistic
ED Pearson Correlation 1 ,416* ,064 Regression method, from which we can conclude that the
Sig. (2-tailed) ,048 ,773 model which contains the variables ED, AE, IS explains
N 23 23 23 its evolution.
AE Pearson Correlation ,416* 1 ,385
Sig. (2-tailed) ,048 ,070 The correlations between variables are presented in
N 23 23 23 Table II and it can be observed that Age Equipment has a
IS Pearson Correlation ,064 ,385 1 importance in water quality evaluation.
Sig. (2-tailed) ,773 ,070 The model developed is unique and can be applied to
N 23 23 23
distribution networks in Romania and has been tested
* Correlation is significant at the 0.05 level (2-tailed).
with good results in three areas located in Ialomita
TABLE II County.
THE VARIABLES FROM LOGISTIC REGRESSION EQUATION Using calculation algorithm described above to
B S.E. Wald df Sig. construct water quality prediction model for distribution
Intercept 5.605 .144 .958 1 .032 networks in Romania may also be used in other countries
ED 0.0208 .177 .464 1 .049 where we have the necessary data pertaining to countries
AE -0.105 .121 1.264 1 .026
concerned.
Mathematical models which have the capacity to
Thus, if in water distribution network there are known accurately describe the correlations between equipment
percentages of poor material, respectively percentages on deficits, age equipment and incorrect sample are
the age of the materials of the network (older than 30 excellent instruments for analysis and interpretation of
years) can find the probability, p, that the water does experimental data as well.
comply with its quality standards (Table III).
This model was almost the same exact, regarding
conform water quality (correct in proportion of 90,3%), References
like in non-conformable water quality case (87,4%).
[1] J. L. Rodgers and W. A. Nicewander, Thirteen ways to look at the
TABLE III correlation coefficient. The American Statistician, 42(1) (1988),
EXAMPLES 59–66.
[2] Dowdy, S. and Wearden, S., Statistics for Research, (Wiley.
ED [%] AE>30 [%] ODDS_quality P [%]
1983).
70,43 71,4 0,652427 39 [3] Francis, DP; Coats AJ, Gibson D., How high can a correlation
36,43 13,9 134,72 99 coefficient be?, Int J Cardiol 69 (1999) 185–199.
30 2 411,16 100 [4] T. Le Chap, Introductory Biostatistics, (New York: Wiley 2003).
1,4 0 279,81 100 [5] National Report, (ARA, 2010).
[6] Jaba E., Grama A., Statistical analysis with SPSS on Windows,
(Bucharest, Ed. Polirom, 2004).
The Cox and Snell coefficient pseudo R 2 was 0,975
(Table IV) indicating a good correlation between the
model and the actual data. Authors’ information
1
TABLE IV Hydraulics, Thermal Installations and Atmosphere Protection
PSEUDO R-SQUARE Department of Technical University of Civil Engineering, Bucharest,
Method Coefficient Romania.
Cox and Snell ,975 2
Mathematics and Informatics Department of University of Civil
Nagelkerke ,979
Engineering, Bucharest, Romania.
McFadden ,674
308
Camelia Gavrila Romania, 3 October 1960,

Bachelor of Mathematics (1983) at University of
Bucharest, Faculty of Mathematics, Romania,
Ph.D. in the field of civil engineering (2003) at
Technical University of Civil Engineering
Bucharest, Romania.
She has ISI publications in the JOAM, OAM–
RC, SPIE journals, Chemistry Magazine,
Proceeding Conference IEEE, SPIE, WSEAS.
She has research agreements and projects in the field of: Identification
Methods for Nonlinear System; Numerical Methods for solving PDE’s:
finite difference method, discrete volume method, finite element
method; Modelling Water Quality; Environment Protection; Optics and
lasers; Abel modelling; Modelling and simulation with COMSOL
Multiphysics. Assoc. Prof. Dr. GAVRILA is a membership to
professional bodies: The Romanian Association of Water (ARA), The
International Water Association (IWA), IEEE, OSA.
Alina Elisabeta Sandu Romania, 1 September

1977, Bachelor of Mathematics (2000) at
University of Bucharest, Faculty of
Mathematics, Romania, Ph.D. student in the
field of civil engineering at Technical University
of Civil Engineering Bucharest, Romania.
Elena Burchiu Romania, 14 January 1961,

Bachelor of engineering installations (1985) at
Technical University of Civil Engineering
Bucharest, Romania, Ph.D. in the field of civil
engineering (2012) at Technical University of
Civil Engineering Bucharest, Romania. She has
publications in the Mathematical Modelling in
Civil Engineering. She has research agreements
and projects in the field of civil engineering.
309
ISSN 2035-1755 July 2013
Biosurfactant Production from Industrial Residues

Using Microorganisms Isolated from Oil Wells
Ana Katerine de C. Lima Lobato1, Andréa F. Almeida2, Márcio S. Bezerra3,

Laerte M. Barros Júnior4, Luiz Carlos L. Santos5, Gorete R. Macedo6
Abstract – Biosurfactants can be used in several industries, such as the petrochemical industry.
It is therefore of great interest to improve biosurfactant production processes to make
biosurfactants economically competitive compared to synthetic surfactants. To achieve economic
competitiveness, it is important to select microorganisms that can produce biosurfactants, to
establish an appropriate and economical media and optimized cultivation conditions for the
biosurfactant synthesis. In this work, the production of biosurfactant by microorganisms isolated
from oil wells was studied in a medium composed of sugar cane molasses and petroleum as
carbon sources. The kinetics of microbial growth were evaluated by monitoring the cell
concentration, substrate consumption and surface tension variation, which could indicate
biosurfactant production. A reduced cultivation time of 8 hours with a significant surface tension
reduction (30.77%) was achieved with molasses as a low-cost substrate in the presence of
petroleum. The obtained surface tension reduction was similar to that obtained by the previous
literature. However, to the best of our knowledge, no report has yet been published with a reduced
cultivation time of 8 hours and a significant surface tension reduction using molasses as a low-
cost substrate in the presence of petroleum. Copyright © 2013 Praise Worthy Prize S.r.l. - All
rights reserved.
Keywords: Biosurfactant, Microorganism, Molasses, Petroleum, Surface Tension
I. Introduction Compared to synthetic surfactants, biosurfactants

perform better under extreme temperatures and pH and at
Biosurfactants are amphoteric substances produced high salinity [12], [13]. In addition, biosurfactants have
from different sources of carbon by microorganisms such low toxicity and are biodegradable [3], [14], [15], [16].
as bacteria, yeasts and filamentous fungi. Little is known In spite of these advantages, the economical
about what leads microorganisms to produce these production of biosurfactants has several barriers that
metabolites, but one reason may be the need to solubilize must be overcome, such as low yield, high cost for large-
insoluble substrates in media [1]. scale production, the need for sterilization, and process
These molecules act preferentially on the interface control problems including foam formation, product
between fluid phases with different polarities, and for the recovery and purification problems and the difficulty in
case of an oil reservoir, would link the oil/water or conducting a chemical analysis of the products, given the
water/oil interfaces and reduce the surface and interface complex nature of the produced metabolite [17].
tensions between these phases [2], [3], [4]. According to Kosaric [6], the greatest use of
This action strongly affects the rheology of the biosurfactants in the oil industry is for MEOR (Microbial
medium and improves the oil recovery factor [5]. The Enhanced Oil Recovery).
main classes of biosurfactants are glycolipids, This tertiary oil recovery technique consists of using
lipopeptides, lipoproteins, phospholipids, fatty acids, and microorganisms or some of their metabolic products to
polymeric and particulate surfactants [6]. Some recover residual oil. Polymers and surfactants produced
researchers prefer to group biosurfactants by their by these microorganisms could reduce the oil-rock
molecular weights [7], [8]. surface tension; this reduction mitigates capillary forces
For example, low-molecular-weight biosurfactants are and improves oil flow through the rock pores [11], [14],
generally glycolipids, and rhamnolipids are the main [18].
example of this class [9]. For an economical biosurfactant production process, it
This type of biosurfactant is produced by different is important to identify the microorganisms that produce
species of Pseudomonas [10]. biosurfactant and to optimize the cultivation medium and
Biosurfactants compete with synthetic surfactants in the fermentation process itself [16]. It is estimated that
three main aspects: cost, functionality and production raw materials account for 10 to 30% of the overall
volume (productivity) [11]. production cost of biosurfactants [19], [20].
310
Ana Katerine de C. Lima Lobato et al.
Several renewable substrates, especially those from sub-master culture and incubating the cells at 38ºC for 48
residue-generating industries, have been used for hours. After incubation, the sub-master culture was
microorganism cultivation and biosurfactant production stored at 5ºC.
on an experimental scale in an attempt to make the
produced metabolite more competitive with its synthetic
counterpart [4], [16], [21], [22], [23], [24]. The main II.4. Cell Propagation
drawback of using alternative substrates as a culture Erlenmeyer flasks with capacities of 500 mL and
medium is finding a residue with high nutrient value that containing 150 mL of culture medium were inoculated
allows cell growth and product accumulation [19]. with cells from the sub-master culture. The flasks were
Olive Oil Mill Effluent (OOME), animal fat, frying placed in a shaker (New Brunswick) at 38ºC for 24
oil, soapstock, molasses, milk serum and other starch- hours, under agitation at 250 rpm. This culture was used
rich residues [3], [25] have been used as substrates for as an inoculum for the biosurfactant production. A period
biosurfactant production. Manioc liquid residue, soy of 24 hours was established (in previous work) because
seed, beet, sweet potato and potato sugars have also been at this time, the cells are in the exponential phase.
used. Other residues include wheat bran, rice and wheat
stalks; soy, corn and rice husks; sugar cane and manioc
bagasse; coffee processing residues such as coffee pulp; II.5. Cultivation Procedures
fruit processing residues such as apple and grape pulp; Erlenmeyer flasks with aeration rates of 0.4
pineapple and carrot processing wastes, banana wastes; (Vmedium/Vflask) were inoculated with 7% (v/v) of
and oil mill wastes, such as coconut, soy, peanut and inoculum. After inoculation, the flasks were placed in the
canola [19], [26]. shaker and were agitated at 250 rpm and 38ºC for 96
These metabolites produced from inexpensive hours. Samples were taken every 4 hours for the first 12
renewable substrates, obtained through high-yield lead to hours and every 12 hours afterwards to determine
economically viable processes and make biosurfactants a biomass concentration, substrate concentration and
promising product. surface tension variation.
In this work, Pseudomonas aeruginosa sp. isolated
from oil wells was studied for biosurfactant production.
It was cultivated in Erlenmeyer flasks under agitation to II.6. Growth Kinetics
identify, through kinetic variables, a better culture To study the growth kinetics of the microorganism,
medium with a substrate of sugar cane molasses and several experiments were conducted to accompany
petroleum. culture growth in different carbon sources, such as
sucrose, molasses and petroleum, and the conditions of
the inoculum and the medium were varied (see Table I).
II. Materials and Methods
TABLE I
II.1. Microorganism DESCRIPTION OF THE EXPERIMENTS PERFORMED
The Pseudomonas aeruginosa sp. strain isolated from Experiment Inoculum medium Culture medium
PETROBRAS oil wells in the AÇU100 zone, Mossoró, 1 Medium A, 1%S Medium A, 1%S
2 Medium A, 1%S Medium A, 1%S, 1%P
Brazil, was selected as a promising microorganism. The 3 Medium A, 1%M, 1%YE Medium A, 1%M
strain belongs to UN-RNCE (Rio Grande do Norte/Ceará 4 Medium A, 1%M, 1%YE Medium A, 1%M, 2%P
Business Unit) and is maintained by the Department of *S – Sucrose; M – Molasses; P – Petroleum; YE – Yeast Extract
Antibiotics of the Federal University of Pernambuco.
II.7. Cell Concentration
II.2. Cultivation Media
The cell concentration was obtained with the dry
Cultivation media was prepared from yeast extract, weight method [27]. Samples (2 ml) taken from the
agar, sucrose, molasses, petroleum (from AÇU100 zone shaker were centrifuged (16.000 g for 15 min) and were
– PETROBRAS - Mossoró/Brazil) and medium A, which washed with distilled water. After a second washing, the
was prepared from 100 mM phosphate buffer (pH 7) with samples were placed in a furnace at 105°C for 24 hours
1 g/l of (NH4)2SO4, 0.25 g/l of MgSO4.7H2O and 1 ml/l and weighed until a constant weight (dry weight) was
of a stock salts solution containing (g/100 ml) 0.01 of obtained. Therefore, the cell concentration (g/l) was
EDTA, 0.30 of MnSO4.H2O, 0.01 of FeSO4.7H2O, 0.01 obtained by dividing the dry weight by the sample
of CaCl2, 0.01 of CoCl2.6H2O and 0.01 of ZnSO4.7H2O. volume.
II.3. Culture Maintenance II.8. Substrate Concentration

The culture was maintained in solid medium with The DNS method is based on the reduction of 3,5
PCA and 25% (w/v) of agar. Cell renewal was performed dinitrosalicylic acid to 3-amino-5-nitrosalicylic acid with
by transferring the cells from the master culture to the the simultaneous oxidation of the aldehyde group of the
311
sugar to a carboxylic group; a red color develops and is bioremediation of oil spills. The carbon source, medium
measured spectrophotometrically at 600 nm [28]. The composition, cultivation time and knowledge of the
original procedure determines glucose and fructose. kinetics are essential variables in the overall assessment
Because our sample contains sucrose, a preliminary of the process [32], [33].
hydrolysis is needed. The hydrolysis was performed in a The kinetic study of the cell growth (experiments 1 to
50 mL volumetric balloon by the addition of 1 mL of the 4) is shown in Fig. 1.
sample, 0.5 mL of HCl and 6 mL of water. The final Fig. 1 shows that the exponential growth phase occurs
solution was heated at 70°C for 10 minutes and then in the first 24 hours of cultivation, with the exception of
cooled to ambient temperature and neutralized with 4N experiment 2. In experiment 2, no yeast extract was used
NaOH solution. The flask was filled to the mark with in the medium during preparation of the inoculum, and
distilled water. oil was added to the cultivation medium. In this case, the
existence of a lag phase was observed; this phase
indicated the reaction of the microorganism to the
II.9. Surface Tension medium, which is now more complex.
The surface tension was measured in a Du Noüy However, this was not observed in the experiment 4,
tensiometer through the ring method [29]. The analyses in which the medium contained petroleum and the
were conducted at 30ºC on the supernatant obtained after inoculum preparation medium contained yeast extract.
centrifugation of the raw sample. The inclusion of this source of essential amino acids
The surface tension reduction provides an indirect and proteins, which are responsible for cell growth, was
measurement of biosurfactant production. strategic for the elimination a lag phase in a more
complex medium. For most of the conditions studied,
after 36 hours of cultivation, the cell concentration
II.10. Emulsification Index begins to decline phase.
The emulsification activity was determined by This decline is attributed to cell death as a function of
measuring the emulsification index (E24) at 25°C [24]. the substrate limitation or to biosurfactant action on the
In general, 4 ml of crude oil was poured into a test cell wall. The adherence of biosurfactants to cell surfaces
tube containing 6 ml of biosurfactant solution or caused deterioration in the integrity of the cell membrane
commercial surfactant. After being vigorously vortexed and also cause breakdown in the nutrition cycle [34],
for 2 min, the test tube was kept still for 24 h and the [35].
heights of the emulsion, oil and aqueous zones were Experiments 1 to 4 show that in most of the cases, the
measured. The emulsion index (E24) was then calculated cells developed adequately under the conditions studied.
from the ratio of the height of the emulsion zone to the The substrate was exhausted, and the biomass
total of the heights of the oil, emulsion, and aqueous concentration reached a plateau after 48 hours of
zones. cultivation (Fig. 2).
The reduced surface tension coincides with the
exponential growth phase between 0 and 12 hours (Fig.
II.11. Kinetic Parameters 3).
The maximum specific growth rate was estimated 18
with the software Lissage [30] based on linearization of
the exponential phase of cell growth. The factor for the 16
conversion of substrate to cell was calculated by dividing
14
the change in biomass by the substrate concentration.
Biomass (10 g/L)
12
-3
III. Results and Discussion 10
The evaluation of biosurfactant production processes 8

using low-cost substrate and real environmental
conditions is important from an economic and 6
application perspective [31]. To this, the study explored a
4
set of variations in medium composition, from the
presence of simple substrates, such as pure sucrose, to a 2
potentially more economical carbon source easily 0 4 8 12 24 36 48 60 72 84 96 108
assimilated by the cells (sugar cane molasses) and a Time (h)
combination of those substrates with crude oil. Experiment 1 Experiment 2
The use of crude oil in the medium is expected Experiment 3 Experiment 4
because the working strain has been isolated from oil
Fig. 1. Biomass production curves for experiments 1 to 4 at 250 rpm
wells and also because of the possibility for using the and 38°C. Experiment 1 (Medium A, 1% S); Experiment 2 (Medium A,
produced biosurfactant in oil industry activities, such as 1% S, 1% P); Experiment 3 (Medium A, 1% M); Experiment 4
MEOR (Microbial Enhanced Oil Recovery) and (Medium A, 1% M, 2% P)
312
14
rate occurred in experiment 4 (Xmax = 0.0885). In
12 addition, experiment 4 presented a surface tension
reduction percentage (STR) of 30.77% for the medium
10
containing molasses and petroleum. This result indicated
8 that the microorganism was well adapted to the proposed
medium.
Substrate (g/L)
6
Experiments 1 and 3, which substitutes sucrose for
4,8
molasses, show reductions in the surface tension of
4,0
12.9% and 44.44%, respectively.
3,2
2,4
This result suggests a substantial increase in
1,6
biosurfactant and also indicates the potential for
0,8
molasses as substrate in this process. Ghurye et al. [36]
0,0 produced biosurfactant from a mixed culture with a
carbon source of molasses, and they obtained a surface
0 4 8 12 24 36 48 60 72 84 96 108
Time (h) tension reduction of approximately 30% after 48 hours of
Experiment 1 Experiment 2
cultivation.
Experiment 3 Experiment 4 Considering molasses as substrate, Makkar and
Cameotra [23] showed a surface tension reduction in the
Fig. 2. Substrate consumption curves for experiments 1 to 4 at 250 rpm medium of 50% after 24 hours. Kashkouli et al. [37]
and 38°C. Experiment 1 (Medium A, 1% S); Experiment 2
applied an artificial neural network to model the
(Medium A, 1% S, 1% P); Experiment 3 (Medium A, 1% M);
Experiment 4 (Medium A, 1% M, 2% P) fermentation parameters for biosurfactant production by
Bacillus subtilis ATCC 6633 from sugar cane molasses.
70 In this bioconversion process, the highest reduction of
surface tension (45%) was noticeable at 48 hours.
Experiment 3 achieved a tension reduction (44.44% in 24
Surface tension (dyna/cm)
60
hours) comparable to the result obtained by these
authors.
50 Several studies have shown significant reductions in
surface tension when they adopted low-cost substrates
40
for biosurfactant production; Fox and Bala [22] managed
to reduce surface tension by 60% in 72 hours by
cultivating Bacillus subtilis with potato.
30 By cultivating Pseudomonas aeruginosa, Santa Anna
et al. [38] observed that the use of babassu oil resulted in
0 4 8 12 24 36 48 60 72 84 96 108 a surface tension reduction of 31% in 168 hours whereas
Time (h) with glycerol, a reduction of 48.2% was observed in the
same time period.
Nitschke and Pastore [39] evaluated a cassava flour
Fig. 3. Surface tension variation curves for experiments 1 to 4 at 250 processing effluent as a substrate for surfactant
rpm and 38°C. Experiment 1 (Medium A, 1% S); Experiment 2 production by two Bacillus subtilis strains; a surface
(Medium A, 1% S, 1% P); Experiment 3 (Medium A, 1% M);
Experiment 4 (Medium A, 1% M, 2% P)
tension reduction in the medium of 44.9% was obtained
in 24 hours with B. Subtilis ATCC 21332. Jayachandran
This relation suggests that the biosurfactant synthesis and George [40] analyzed rhamnolipid biosurfactants
is associated with the growth because parallel relations produced by submerged fermentation with orange fruit
could be observed between biomass production, substrate peelings as the sole carbon source, and they found a
consumption, and the decrease in the surface tension of surface tension reduction of approximately 45% after 168
the culture broth. hours of fermentation.
Experiment 1 shows only a small reduction in surface Rocha et al. [41] evaluated cashew apple juice for
tension (Fig. 3) when only sucrose is used as a carbon production of biosurfactants by Bacillus subtilis
source. This result is most likely because of the easy LAMI008 and achieved a surface tension reduction of
assimilation of sucrose, which resulted in the absence of 21.37% in 24 hours.
biosurfactant synthesis.
TABLE II
Some cultivation parameters, such as the maximum CULTIVATION PARAMETERS FOR EXPERIMENTS 1 TO 7
specific growth rate (Xmax), the substrate-to-cell Xmax STR STmin
Experiment YX/S
conversion factor (YX/S), the surface tension reduction (h-1) (%) (dyn.cm-1)/t(h)
percentage (STR %) and the relation between the 1 0.0223 0.0009 12.90 27/36
2 0.0429 0.0017 47.69 34/48
minimum surface tension and time (STmin/t), were 3 0.0467 0.0028 44.44 30/24
calculated, and they are shown in Table II. 4 0.0885 0.0013 30.77 36/8
As shown in Table II, the highest specific cell growth
313
With respect to the addition of petroleum to the and petroleum in the medium (Medium A, 1% M, 2% P)
medium, experiment 2 (1% sucrose and 1% petroleum) was 100%. The ability to form emulsions with oil
showed the greatest reduction in surface tension suggests that the biosurfactant produced by fermentation
(47.69%) followed by experiment 4 (30.77%), where the with Pseudomonas aeruginosa sp. is a good candidate for
medium contained 1% molasses and 2% petroleum. a cleaning and emulsifying agent and that the
Some studies were reported in the literature on surface biosurfactant has potential applications in microbial
tension reduction when more complex substrates are used enhanced oil recovery (MEOR), environment
for the growth of microorganism producers of preservation [12], medicine and the cosmetics industry.
biosurfactants; Bento et al. [42] isolated biosurfactant- Although some investigators have managed to obtain
producing microorganisms from soil contaminated with similar surface tension reductions, the cultivation time
diesel oil and evaluated the capacity for reduction of the required was far greater than that observed in the present
surface tension of the medium when cultivated in a study; this result indicates the potential of the
medium containing diesel oil as the only carbon source. Pseudomonas aeruginosa sp. strain and the conditions
Among the microorganisms, Acinetobacter junii adopted here.
showed the highest surface tension reduction (37.33%)
after 168 hours of cultivation. Samadi et al. [43]
cultivated Brevibacterium sp. S-34 in different carbon IV. Conclusion
sources (glucose, glycerol, molasses, gasoline, canola oil The Pseudomonas aeruginosa sp. strain proved to be a
and residual oil) and found that the greatest reduction in promising producer of biosurfactants from sugar cane
the surface tension of the medium was obtained when molasses. The highest STR of the medium coincides with
glycerol was the only carbon source (56.52%) after 72 the exponential growth phase; thus, biosurfactant
hours of cultivation. synthesis is associated with cell growth. The optimum
The highest substrate-to-cell conversion factor was cultivation conditions for biosurfactant synthesis were
obtained in experiment 3 (YX/S = 0.0028). In this case, found to use inoculum with 1% molasses and 1% yeast
only molasses was used in the medium; this result extract in medium A, and a cultivation medium with the
indicates that the absence of petroleum favors the same composition and 2% petroleum; under these
conversion of sugar into cells. conditions, a STR of 30.77% was reached after only 8
By analyzing the ratio of the surface tension reduction hours of cultivation.
to the cultivation time (STmin/t), it was found that for Compared to the current literature, the surface tension
experiment 2, with sucrose and petroleum in the medium, reduction was similar to that obtained by other
a STR of 47.69% was obtained, and the lowest point of researchers that used an alternative substrate for
the surface tension was at 48 hours of cultivation. biosurfactant production. However, minimization of the
The experiment that contains only molasses as a cultivation time is another important process parameter
carbon source (experiment 3), had a STR of 44.44%, and for reducing the production cost of biosurfactant. To the
the lowest surface tension was recorded at 24 hours of best of our knowledge, no report has yet been published
cultivation. However, in experiment 4, with molasses and with a reduced cultivation time of 8 hours and a surface
petroleum in the medium, a STR of 30.77% was reached tension reduction of approximately 30% with a low-cost
at only 8 hours of fermentation. Considering petroleum substrate in the presence of petroleum.
as part of the substrate, Silva et al. [44] studied the
biosurfactant production potential of the Pseudomonas
fluorescens strain and found a surface tension reduction Acknowledgements
of 57.14% after 60 hours of cultivation.
Notably, although the highest surface tension The authors thank FINEP, CTPETRO and
reduction occurred in experiment 2 (47.69%), which uses PETROBRAS for financial support.
sucrose and petroleum as the carbon source, this
reduction only occurred after 48 hours of cultivation. By References
using only molasses as the carbon source (experiment 3),
the surface tension reduction was practically the same as [1] F. Sineriz, R. K. Hommel, H. P. Kleber, Production of
biosurfactants. In: Encyclopedia of Life Support Systems (Eolls
in experiment 2 (44.44%) but in half the time (24 hours) Publishers, 2001).
of cultivation. [2] G. Bognolo, Biosurfactants as emulsifying agents for
Experiment 4, which uses molasses and petroleum in hydrocarbons, Colloids Surf. A. 152 (1999) 41-52.
the medium, also showed a significant surface tension [3] J. D. Desai, I. M. Banat, Microbial production of surfactants and
their commercial potential, Microbiol. Mol. Biol. Rev. 61 (1997)
reduction (30.77%) after only 8 hours of cultivation. The 47-64.
latter experiment is of greatest interest to this study [4] R. S. Makkar, S. S. Cameotra, An update on the use of
because a satisfactory surface tension reduction was unconventional substrates for biosurfactant production and their
attained in a lower time and because molasses, which is a new applications, Appl. Microbiol. Biotechnol. 50 (1998) 520-
529.
non-conventional substrate, and petroleum were used in [5] M. Benincasa, A. Abalos, I. Oliveira, A. Manresa, Chemical
the medium. The emulsification activity of the structure and biological activities of the biosurfactant produced
biosurfactant produced in the experiment 4 with molasses by Pseudomonas aeruginosa LBI from soapstock, Antonie Van
Leeuwenhoek. 85 (1) (2004) 1-8.
314
[6] N. Kosaric, Biosurfactants in Industry, Pure Appl. Chem. 64 (11) Microb. Cell Fact. 4 (2005) 1-16.
(1992) 1731-1737. [32] W. Schmidell, U. A. Lima, E. Aquarone, W. Borzani, Engenharia
[7] H. Bach, Y. Berdichevsky, D. Gutnick, An exocellular protein Bioquímica - Biotecnologia Industrial (Edgard Blücher, 2001).
from the oildegrading microbe Acinetobacter venetianus RAG-1 [33] M. H. Pinto, R. G. Martins, J. A. V. Costa, Avaliação cinética da
enhances the emulsifying activity of the polymeric bioemulsifier produção de biossurfactantes bacterianos, Quim. Nova 32 (2009)
Emulsan, Appl. Environ. Microbiol. 69 (5) (2003) 2608-2615. 2104-2108.
[8] E. Z. Ron, E. Rosenberg, Biosurfactants and oil bioremediation, [34] S. T. Hingley, A. T. Hastie, F. Kueppers, Effect of ciliostatic
Curr. Opin. Biotechnol. 13 (2002) 249-252. factors from Pseudomonas aeruginosa on rabbit respiratory cilia,
[9] N. Christofi, I. B. Ivshina, Microbial surfactants and their use in Infect. Immun. 51 (1986) 254-258.
field studies of soil remediation, J. Appl. Microbiol. 93 (2002) [35] E. Z. Gomaa, Antimicrobial activity of a biosurfactant produced
915-929. by Bacillus licheniformis strain M104 grown on whey, African
[10] G. Soberón-Chavéz, F. Lépine, E. Déziel, Production of Afr J Microbiol Res 6(20) (2012) 4396-4403.
rhamnolipids by Pseudomonas aeruginosa, Appl Microbiol [36] G. L. Ghurye, C. Vipulanandan, R. C. Willson, A practical
Biotechnol 68 (2005) 718–725. approach to biosurfactant production using nonaseptic
[11] I. M. Banat, R. S. Makkar, S. S. Cameotra, Potencial commercial fermentation of mixed cultures, Biotechnol. Bioeng. 44 (1994)
application of microbial surfactants, Appl. Microbiol. Biotechnol. 661-666.
53 (2000) 495-508. [37] Y. R. Kashkouli, A. Mogharei, S. Mousavian, F. Vahabzadeh,
[12] F. A. Kronemberger, C. P. Borges. D. M. G. Freire, Fed-batch Performance of Artificial Neural Network for predicting
biosurfactant production in a bioreactor, I.RE.CH.E 2(4) (2010) fermentation characteristics in biosurfactant production by
513-518. Bacillus subtilis ATCC 6633 using sugar cane molasses, Int J
[13] R. S. Makkar, S. S. Cameotra, I. M. Banat, Advances in Food Eng 7(6) (2011) 1-19.
utilization of renewable substrates for biosurfactant production, [38] L. M. Santa Anna, G. V. Sebastian, E. P. Menezes, T. L. M.
AMB Express 1(5) (2011) 1-19. Alves, E. P. Menezes, D. M. G. Freire, Production of
[14] M. Nitschke, G. M. Pastore, Biossurfactantes: propriedades e biosurfactants from Pseudomonas aeruginosa PA1 isolated in oil
aplicações, Quim. Nova 25 (5) (2002) 772-776. environments, Braz. J. Chem. Eng 19 (2002) 159-166.
[15] S. S. Cameotra, R. S. Makkar, J. Kaur, S. K. Mehta, Adv. Exp. [39] M. Nitschke, G. M. Pastore, Biossurfactantes: propriedades e
Med. Biol., 672 (2010) 261-280. aplicações, Appl. Biochem. Biotechnol. 112 (2004) 163-172.
[16] D. F. Coelho, S. A. Barbosa. E. Silveira, R. R. Souza, E. B. [40] S. George, K. Jayachandran, Analysis of rhamnolipid
Tambourgi, Biosurfactant production from unconventional biosurfactants produced through submerged fermentation using
resources: a short overview, (2012) International Rewiew of orange fruit peelings as sole carbon source, Appl. Biochem.
Chemical Engineering (IRECHE) 4(2), pp. 175-183. Biotechnol. 158 (2009) 694–705.
[17] M. G. Healy, C. M. Devine, R. Murphy, Microbial production of [41] M. V. P. Rocha, R. V. Barreto, V. M. M. Melo, L. R. B.
biosurfactants, Resour. Conservat. Recycl. 18 (1996) 41-57. Gonçalves, Evaluation of cashew apple juice for surfactin
[18] A. Perfumo, I. Rancich, I.M. Banat, Possibilities and challenges production by Bacillus subtilis LAMI008, Appl. Biochem.
for biosurfactants use in petroleum industry, Adv. Exp. Med. Biol., Biotechnol. 155 (2009) 366–378.
672 (2010) 135-145. [42] F. M. Bento, F. A. O. Camargo, B. C. Okeke, W. T.
[19] R. S. Makkar, S. S. Cameotra, An update on the use of Frankenberger Jr, Diversity of biosurfactant producing
unconventional substrates for biosurfactant production and their microorganisms isolated from soils contaminated with diesel oil,
new applications, Appl. Microbiol. Biotechnol. 58 (2002) 428- Microbiol. Res. 160 (2005) 249-255.
434. [43] N. Samadi, N. Abadian, A. Akhavan, M. R. Fazeli, A. Tahzibi, H.
[20] S. N. Al-Bahry, Y. M. Al-Wahaibi, A. E. Elshafie, A. S. Al- Jamalifar, Biosurfactant production by the strain isolated from
Bemani, S. J. Joshi, H. S. Al-Makhmari, H.S. Al-Sulaimani, contaminated soil, J Biol Sci. 7 (2007) 1266-1269.
Biosurfactant production by Bacillus subtilis B20 using date [44] T. A. L. Silva, H. W. C. Araújo, E. B. Tambourgi, C. A. A. Silva,
molasses and its possible application in enhanced oil recovery, G. M. C. Takaki, Potencial biotecnológico de uma nova linhagem
Int. Biodeterior. Biodegrad. 81 (2013) 141–146. de Pseudomonas fluorescens na produção de biossurfactante
[21] P. S. Babu, A. N. Vaidya, A. S. Bal, Kinetics of biosurfactant utilizando petróleo como substrato, Exacta 7 (2009) 31-37.
production by Pseudomonas aeruginosa strain BS2 from
industrial wastes, Biotechnol. Lett. 18 (3) (1996) 263-268.
[22] S. Fox, G. A. Bala, Production of surfactant from bacillus ATCC Authors’ information
21332 using potato substrates, Bioresour. Technol. 75 (2000)
1
235-240. Federal University of Bahia – Postgraduate Program of Chemical
[23] R. S. Makkar, S. S. Cameotra, Utilization of molasses for Engineering - R. Prof. Aristides Novis, 2, 2º andar, Federação, CEP
biosurfactant production by two bacillus strains at thermophilic 40210-630, Salvador/BA, Brazil. Email: katycarvalho@hotmail.com
conditions, J Am Oil Chem Soc. 74 (1997) 887-889.
2
[24] R. M. Patel, A. J. Desai, Biosurfactant production by Federal University of Paraíba – Department of Biotechnology -
Pseudomonas aeruginosa GS3 from molasses, Lett. Appl. Campus I, Cidade Universitária, Castelo Branco, CEP 58.051-900 João
Microbiol. 25 (1997) 91-94. Pessoa/PB, Brazil.
[25] S. Maneerat, Production of biosurfactants using substrates from
3
renewable-resources, Songklanakarin. J. Sci. Technol. 27 (3) Federal Institute of Education, Science and Technology of Rio Grande
(2005) 675-683. do Norte - R. Manoel Lopes Filho, 773, Valfredo Galvão, CEP: 59380-
[26] S. Mukherjee, P. Das, R. Sem, Towards commercial production 000, Currais Novos/RN, Brazil.
of microbial surfactants, Trends Biotechnol. 24 (11) (2006) 509-
4
515. Petrobras – E&P Business Unit for Rio Grande do Norte & Ceará - R.
[27] W. Borzani, W. Schmidell, U. A. Lima, E. Aquarone, Euzébio Rocha,1000, Cidade da Esperança, CEP 59164-100, Natal/RN,
Biotecnologia Industrial – Fundamentos (Edgard Blucher, 2001). Brazil.
[28] G. L. Miller, Use of dinitrosalicylic acid reagent for determination
5
of reducing sugar, Anal. Chem. 31 (1959) 426–428. Federal University of Bahia - Department of Materials Science and
[29] P. L. Du Noüy, An interfacial tensiometer for universal use, J. Technology - R. Prof. Aristides Novis, 2, 3º andar, Federação, CEP
Gen. Physiol. 7 (1925) 625-633. 40210-630, Salvador/BA, Brazil.
[30] D. A. Simões, G. Arroyo, J. L. Uribelarrea, Logiciel de Lissage
6
(1994). Software. Federal University of Rio Grande do Norte – Department of Chemical
[31] B. Hahn-Hägerdal, K. Karhumaa, C.U. Larsson, M. Gorwa- Engineering - Campus Universitário, CEP 59.072-970 Natal/RN,
Grauslund, J. Görgens, W. H. Van Zyl, Role of cultivation media Brazil.
in the development of yeast strains for large scale industrial use,
315
Ana Katerine de Carvalho Lima Lobato Luiz Carlos L. Santos (Rio de Janeiro, 1976)
(Natal, 1973) has a bachelor’s degree in has a bachelor’s degree in Chemical
Chemical Engineering from the Federal Engineering from the Federal University of Rio
University of Rio Grande do Norte, Natal, Brazil Grande do Norte, Natal, Brazil (1999), a
(1999), a master's in Chemical Engineering from master's in Chemical Engineering from the
the Federal University of Rio Grande do Norte, Federal University of Rio Grande do Norte,
Natal, Brazil (2003) and a PhD in Chemical Natal, Brazil (2002) and a PhD in Chemical
Engineering from the Federal University of Rio Engineering from the University of Manchester,
Grande do Norte, Natal, Brazil (2010) with a sandwich period at the United Kingdom (2007). He has experience in Chemical Engineering
University of Manchester, United Kingdom. Currently, she is a with an emphasis in Oil and Gas. His research interests cover the
researcher and an assistant professor in the Post-Graduate Program in following subjects: petroleum, biofuels, surfactants, microemulsions,
Chemical Engineering at the Federal University of Bahia. adsorption, catalysis and ceramic membranes. Dr. Santos is a member
She has experience in Chemical Engineering with an emphasis in of the Society of Petroleum Engineering.
Biochemical Engineering. Her research interests cover the following
subjects: petroleum, biofuels, biosurfactant production, antibiotic Gorete Ribeiro de Macedo has a bachelor’s
production and metabolic flux analysis. degree in Chemical Engineering from the
Prof. Lobato is a member of the technical chamber of biological Federal University of Rio Grande do Norte
products at the general management of medicines national agency of (1978), a master's in Chemical Engineering from
sanitary surveillance. the University of São Paulo (1982) and a Ph.D.
in Chemical Engineering from the University of
Andrea Farias de Almeida has degree in São Paulo (1998). She is currently professor at
Chemical Engineering from the Federal the Federal University of Rio Grande do Norte.
University of Rio Grande do Norte (2003), a She has experience in Chemical Engineering with an emphasis in
master’s in Chemical Engineering from the Biochemical Processes in the following subjects: enzyme production,
Federal University of Rio Grande do Norte bioethanol, biosurfactants, obtaining oligosaccharides by fermentation,
(2005) and a Ph.D. in Chemical Engineering cultivation of recombinant microorganisms, biological treatment of
from the Federal University of Rio Grande do liquid and solid waste. Prof. Macedo was the coordinator of the
Norte (2010). She is currently a researcher and Graduate Program in Chemical Engineering Federal University of Rio
assistant professor in the Department of Biotechnology at the Grande do Norte (2001-2005) and of the Ph.D in Biotechnology of the
Biotechnology Center of the Federal University of Paraíba. Northeast Biotechnology Network (2008-2012).
She has experience in Chemical Engineering with an emphasis on
biochemical processes in the following subjects: fermentative
processes, production of animal cells, in vitro production of
baculoviruses, AgMNPV and SfMNPV, and biopesticides.
Márcio Silva Bezerra (São Paulo, 1979) has a

bachelor’s degree in Chemical Engineering from
the Federal University of Rio Grande do Norte
(2004), a master’s in Chemical Engineering
from the Federal University of Rio Grande do
Norte (2006) and a Ph.D. in Chemical
Engineering from the Federal University of Rio
Grande do Norte (2012). He is currently a
professor at the Federal Institute of Education, Science and Technology
of Rio Grande do Norte (IFRN). He has experience working on the
following topics: biosurfactants, agroindustrial residues and
fermentation in submerged cultures.
Laerte Medeiros Barros Júnior has a

bachelor’s degree in Chemical Engineering from
the Federal University of Rio Grande do Norte
(1999), a master’s in Chemical Engineering
from the Federal University of Rio Grande do
Norte (2001) and a Ph.D. in Chemical
Grande do Norte (2008). He is currently a
process engineer at PETROBRAS.
He has experience in Chemical Engineering with an emphasis in
Environmental Engineering in the following topics: utilities,
wastewater treatment and the manufacturing process of urea.
316
ISSN 2035-1755 July 2013
BSCF 5582 and LSCF 6428 Mixed Conducting Membranes

for Syngas Production
Luiz Carlos L. Santos1, Ana Katerine de C. Lima Lobato2, Caetano Moraes3
Abstract – The oxygen permeation performance of three mixed ionic-electronic conducting

membranes was investigated. Ba0.5Sr0.5Co0.8Fe0.2O3- (BSCF 5582) membrane was chosen due to
its reported high oxygen permeation flux, electronic conductivity and good chemical stability.
They were fabricated using powders synthesised using both sol-gel and solid-state methods. A
commercial La0.6Sr0.4Co0.2Fe0.8O3- (LSCF 6428) membrane, fabricated with a powder synthesised
using combustion spray pyrolysis, was also used for comparative studies. X-ray diffraction,
scanning electron microscopy and energy dispersive analysis X-ray emission were used to
characterise the powders and membranes. Oxygen permeation fluxes through the membranes were
measured using pure helium on one side of the reactor and air on the other side in the temperature
range 650 ºC to 850 ºC. The oxygen permeation fluxes are compared to target fluxes available in
the literature. Furthermore, difficulties arising from imperfect sealing of the BSCF 5582 and
LSCF 6428 membranes in the reactor are discussed. Copyright © 2013 Praise Worthy Prize S.r.l.
- All rights reserved.
Keywords: Perovskite, BSCF 5582 and LSCF 6428, Mixed Conducting Membranes, Oxygen
Permeation, Syngas Production
I. Introduction An additional advantage of mixed-conducting

membranes is that they may eliminate the necessity of
Synthesis gas or syngas (H2 + CO) can be produced by catalytically active electrodes on the membrane surface.
different methods such as steam reforming, CO2 Mixed ionic-electronic conductors (MIEC) often
reforming and partial oxidation of methane (POM). The possess sufficient catalytic activity to promote the
first two methods present high costs associated with their oxygen dissociation process [4] and catalyse methane
highly endothermic nature [1]. Alternative carbon-based oxidation.
feedstock such as hydrocarbons, coal, petroleum coke Several factors can affect the performance of the
and biomass can be used to produce syngas although membranes such as temperature, partial pressures,
natural gas provides the lowest cost routes at present [2]. composition of the membrane and membrane thickness
In the case of partial oxidation, membrane technology [5], [6]. Zeng et al. [7] reported that the oxygen release
can be used to provide pure oxygen from air while rate during the POM process could strongly affect the
simultaneously excluding nitrogen from the reaction product distribution. They observed that if oxygen is
mixture eliminating the necessity of a separate oxygen quickly released from the solid, the ratio of O2/CH4 near
production plant and eliminating NOX pollutants. When the membrane is high and the total oxidation of methane
air and natural gas (CH4) are passed over either side of occurs preferentially. If the rate of oxygen release is low
the membrane the established oxygen chemical potential then partial oxidation reactions are more likely.
difference results in oxygen ions being transported However, if the rate of oxygen release is very low, the
through the membrane to the fuel side where they react amount of oxygen is not enough to carry out the partial
with methane to form syngas. oxidation reaction and methane cracking takes place.
The perovskite mixed metal oxides have attracted For example, Kendall et al. [8] studied the effect of
much attention for oxygen separation for the POM route dilution of the methane feed entering a solid oxide fuel-
to syngas [3]. Of general formula AIIBIVO3 or AIIIBIIIO3 cell (SOFC) at the anode side. They reported that the
the cubic lattice can be modified with group II cations or reaction changed from partial oxidation to complete
lanthanum at the corner A site while the body centred B oxidation when the methane was diluted.
site is occupied by smaller first row transition metal Other studies have focused on the importance of the
cations. membrane microstructure. A study by Zhang et al. [9] on
The former modification creates oxygen ion vacancies membranes formed from the SrCo0.8Fe0.2O3- system
while the latter produces electronic conductivity. In this looked into the relationship between the grain size and
way, both oxygen ionic conductivity and electronic oxygen permeation flux. They reported that the oxygen
conductivity are simultaneously present. permeation flux increased with decreasing average grain
317
L. C. L. Santos, A. K. de C. L. Lobato, C. Moraes
size of the membrane i.e., increasing grain boundary housing [20]. A good seal should be chemically inert (no
volume. Geffroy et al. [10] have reported that the oxygen reactions with the membrane), it should present a thermal
semi-permeation performances of dense expansion coefficient compatible with the membrane and
La0.6Sr0.4Fe0.6Ga0.4O3- membranes may be likely reactor and for laboratory studies it needs to be easily
improved by a porous coating on oxygen poor face such removable without damaging the reactor [21]. Workers
as La0.6Sr0.4Co0.8Fe0.2O3- perovskite. Another important in the field have used such methods as gold or silver
concern is membrane stability. Some perovskite-type rings that are deformed into position by compression
materials show good oxygen permeability at elevated prior to experimental studies, spring fitting arrangements
temperatures [11]. However, under reducing conditions that push the membrane tightly into position and ceramic
typical of the conditions associated with the methane- or glass composite cements that are applied at room
side, phase changes of the crystal lattice or segregation of temperature with subsequent curing to complete the seal.
metal oxides at the membrane surface can occur. Qi et al. [21] observed that a ceramic-glass composite
Since the early work of Teraoka et al. [12], [13] on the formed using a mixture of 40-50 % ceramic membrane
La1-xSrxCo1-yFeyO3- perovskite system much work has material powder, 20-50 % Pyrex glass and 5-20 % of an
been carried out on this family, with particular attention additive such as sodium aluminate or boron oxide can
given to LSCF 6428. (The formulae of complex result in a very good seal when working with dense
perovskites are often abbreviated using the first letter of ceramic membranes. It is the aim of this contribution to
the element symbol of each of the metal cation firstly compare the oxygen permeation performance of
components. This is followed by a list of numbers which the two types of membranes, LSCF 6428 and BSCF
is the first significant figure of the stoichiometry of the 5582, and compare their oxygen permeation performance
corresponding metal cation. This system is not official with targets already published in the literature [22]. In
and variations do occur in the literature). Studies of addition any problems with leaks under high temperature
membrane performance have focused on important operation were to be quantified.
quantifiable characteristics such as the surface exchange
coefficient (k) of gaseous oxygen with surface oxide
anion species and the bulk chemical diffusion coefficient II. Materials and Methods
( D ) of O2- through the membrane [14]. II.1. Membrane Preparation
Balachandran et al. [15] have used the
The BSCF 5582 perovskite material was chosen
La0.2Sr0.8Co0.2Fe0.8O3- (LSCF 2828) membrane material because of its excellent ionic and electronic properties
for POM to syngas at 850 ºC. They found that after a few [5], [11], [23], [24], [25].
minutes of operation the membrane showed chemical Two different methods of powder preparation were
instability under the reaction conditions employed, used in this work, the traditional solid-state method and
breaking into several pieces. The LSCF 6428 system has the sol-gel (citrate) method. The procedures employed
also been investigated by Jin et al. [16]. Membranes of were based on the work of Tan et al. [5]. The LSCF 6428
this material proved to be unstable for POM. After 3-7 used in this work is a commercial powder from Praxair
hours of exposure to methane at 850 ºC the membrane Specialty Ceramics formed by the combustion spray
broke into pieces. Pei et al. [17] studied the Sr(Co, pyrolysis (CSP) method and was used without further
Fe)Ox-type oxygen membrane reactors for generating treatment.
syngas. Using the SrCo0.8Fe0.2O3- (SCF) membrane they
found that two types of failure occurred. The first
involved a structural change causing lattice mismatch II.1.1. Solid-State Method
which was followed by chemical decomposition into This method consists of mechanical mixing of the
metals, metal carbonates and metal oxides. A different precursors of the membrane followed by calcining.
stoichiometry of SCF membrane (SrFeCo0.5Ox a non- Barium oxide (BaO, 97 %), strontium oxide (SrO, 99.9
perovskite metal oxide) was used by Ma et al. [18]. The %), cobalt II oxide (CoO, 99.99 %) and iron III oxide
researchers found that this membrane demonstrated high (Fe2O3, 99.98 %) were used as precursors for the
ionic and electronic conductivities and good structural preparation of the BSCF 5582 membrane by the solid-
stability in oxidizing and reducing atmospheres. To state method. Stoichiometric amounts of oxides were
improve the stability Shao et al. [19] partially substituted mixed and milled for 24 hours in a mortar grinder
Sr in SCF with Ba. They found high oxygen permeation (Fritsch Company). About 30 ml of ethanol were added
fluxes, 1.4 ml/cm2.min at an operating temperature of to the oxides to improve mixing. The powders were
950 ºC, could be attained1. calcined at 950 ºC for 5 hours with a heating and cooling
An essential requirement for the use of membrane rate of 2 ºC/min.
ceramic reactors is to find a good seal for the system i.e.,
between the membrane and any membrane or reactor
II.1.2. Citrate Method
1
Where the authors report fluxes in volumetric terms we report the
conditions of temperature and pressure to which this flux applies only In contrast to the solid-state method, sol-gel
if the authors have reported such conditions. techniques use aqueous solutions in order to obtain
318
homogeneous fine powders. ceramic membrane reactor, mass flow controllers to

The precursors are hydrolysed while simultaneously a accurately control the gas flow and an online gas
condensation or polymerization reaction occurs [26]. chromatograph (Varian CP 3800 with a molecular sieve
In this work, citric acid was used as the chelating column, 5A, and TCD) for outlet gas analysis. The gas
agent in a process denoted “citrate method”. Barium chromatograph (GC) was calibrated for oxygen and
nitrate (Ba(NO3)2, 99+ %) strontium nitrate (Sr(NO3)2, nitrogen using pure air, air in helium (50:50 ratio) and a
99+ %) cobalt II nitrate hexahydrate (Co(NO3)2.6H2O, standard gas of 4 % oxygen and 4% nitrogen in helium.
98+ %) and iron III nitrate nonahydrate (Fe(NO3)3.9H2O, The relative error within the measurements was
99 %) were used as precursors in this method. estimated to be around 5% of the measured mole
Applying magnetic stirring, the four nitrates were fraction.
introduced into distilled water with citric acid at a ratio The reactor is made of two chambers formed by dense
of 5:1 (citric acid:metal ions). As citric acid does not inert alumina tubes with the test membrane placed
possess a strong chelating ability [27], ethylene glycol between them. Air was supplied to one side of the
was used to help the polymerization process [28] at a membrane using an alumina tube of internal diameter
ratio of ethylene glycol:metal ions = 1:1. NH3.H2O was 12.7 mm.
used to control the pH (pH = 9). The final solution was On the other side high purity helium (the “sweep gas”)
heated at 70 ºC for several hours to promote the was supplied. The membrane was cemented to a dense
polymerization and to remove excess solvent. The alumina tube of diameter 25.4 mm using an alumina disc
solution became viscous without any visible with a central hole over which the membrane was fixed
precipitation. The purple gel was pyrolised at 350 ºC for into position.
2 hours resulting in a black powder. Finally the powders This tube contains an internal dense alumina tube
were calcined at 950 ºC for 5 hours with the heating and which has an internal diameter of 12.7 mm and was used
cooling rate of 2 ºC/minute. to deliver the helium sweep gas to the membrane.
After preparation of the BSCF 5582 powders, the Stainless steel Swagelok connections and tubes with
materials were uniaxially pressed at 425 MPa for ca 5 graphite ferrules were used to connect the feed gases to
minutes to obtain disc membranes of 20 mm diameter. the reactor and were positioned well outside the furnace
The green membranes were sintered at 1150 ºC for 5 in order to minimise the possibility of leaks developing
hours with a heating and cooling rate of 1 ºC/min. at the operating temperature. The testing rig is illustrated
The resulting discs were of 18 mm diameter and 1 mm in Fig. 1.
thick. LSCF 6428 membranes were formed by pressing The sealing agent employed in this work was a binary
at 30 MPa for ca 5 minutes to give discs of 20 mm filler and binder cement supplied by Fortafix (UK). The
diameter. liquid binder comprises of a solution of water-soluble
The membrane was then sintered at 1200 ºC for 5 sodium and potassium silicate glasses of around 40 % by
hours with a heating and cooling rate of 1 ºC/min. The weight.
diameter of these disc membranes after sintering was 16 The powder filler consists of a mixture of alumina,
mm with a thickness of 1 mm. For both materials the silica and sodium fluorosilicate. A 1:1 ratio of binder to
shrinkage was attributed to densification resulting from filler was used. Upon application the cement was allowed
the sintering process. to set for 24 hours at room temperature before
commencing permeation studies.
II.2. Membrane Characterization
Three different characterization techniques were used
in this work. Firstly, X-ray diffraction (XRD) analysis
using a PW 3710 X-ray diffractometer (Philips company)
was used to characterize the crystal structure of the
powders. A Quanta 200 (FEI company) electron
microscope was used for energy dispersive X-ray
emission analysis (EDAX) to confirm the constituent
metal cation stoichiometry of the powders and
membranes.
The same instrument was used for scanning electron
microscopy (SEM) studies in order to ascertain
morphology and particle size of the powders and to
investigate the surface microstructure of the disc
membranes.
II.3. Reactor Setup

The oxygen permeation testing rig comprises of a Fig. 1. Illustration of the testing rig used for oxygen permeation
using disc membranes
319
III. Results and Discussion III.1.1. SEM Analysis

III.1. Characterization of the Powders Figs. 3 show the SEMs of the morphologies of the
and Disc Membranes XRD Analysis BSCF 5582 powders synthesised using the solid-state
and citrate methods and also the commercial LSCF 6428
Figs. 2 show the XRD results for the BSCF 5582 powder.
powders prepared by the solid-state and citrate methods
and for the commercial LSCF 6428 powder.
As can be seen, the BSCF 5582 powders synthesised
using the two different methods show the same pattern
indicating that the perovskite crystalline structure is the
same for the two different methods of preparation.
These results were compared to XRD powder patterns
(a) (b) (c)
for BSCF 5582 previously published in the literature [5]
and it was confirmed that the peaks were at the same 2- Figs. 3. SEM of the BSCF 5582 powders synthesized by (a) the solid-
theta values. state method, (b) citrate method and (c) commercial LSCF 6428
Shao et al. [19] also studied the BSCF 5582 perovskite
and they obtained the same XRD pattern with the As can be seen the BSCF 5582 powder formed from
structure corresponding to cubic perovskite with the the solid-state method presents visible grains. However,
lattice parameter a = 3.9795 Å. The commercial LSCF the grain distribution is not uniform and large plates were
6428 also demonstrated a single-phase perovskite observed. For BSCF 5582 powders synthesised by the
structure. The indexing of the peaks is given for all the citrate method a different structure can be seen. It shows
powders. a solid material in which there are no visible grains. In
comparison, the commercial LSCF 6428 powder shown
reveals small particles of less than 1m in size and the
(110)
presence of agglomerates of around 2m in size.

The SEMs showing the surface microstructure of the
two BSCF 5582 and the LSCF 6428 membranes are
(211)
(200)
presented in Figs. 4.
(221,300)
(220)
(111)
(310)
(100)
(210)
(311)
20 30 40 50 60 70 80
2
(a)
(110)
(a) (b) (c)
Figs. 4. SEM of the BSCF 5582 membranes formed from powders of

(211)
(a) solid-state method (b) citrate method and (c) commercial LSCF
6428
(111)
(220)
(200)
(221,300)
Clear grain boundaries are observed for all

(310)
(311)
(100)
(210)
membranes. The grain boundaries of the BSCF 5582

membrane formed from the powder made by the solid-
20 30 40 50 60 70 80
2 state method present occluded particles. In the case of the
(b) BSCF 5582 membrane synthesised by the powder
formed from the citrate method it is possible to see a
number of isolated surface features ca 4 m in diameter.
(110)
The LSCF 6428 membrane shows grain particle sizes

in the range 0.5 to 2 m with a surface absent of such
surface cavities. The micrograph image appears ill-
(200)
defined because of the limit of the low-vacuum imaging

(211)
(221,300)
(111)
conditions employed.
(100)
(310)
(220)
(210)
20 30 40 50 60 70 80 III.1.2. EDAX Analysis

2
(c) Initially this technique was used to determine the
stoichiometry of the powders after calcination. It is a
Figs. 2. XRD powder patterns for: (a) BSCF 5582 synthesised by solid-
state method (b) BSCF 5582 synthesised by citrate method and
bulk technique in which the electron beam of variable
(c) commercial LSCF 6428 spot size is used to analyse a volume of the sample
extending to beneath the surface.
320
In this study the smallest spot size used to analyse the where:
finer details of the samples was of the order of 1 m2 Jo2: solid-state oxygen permeation flux through the
with a penetration depth of about 1 m and so a volume membrane (ml(STP)/cm2.min).
of the mixed metal oxide of 1m3 was analysed. The FSG: flow rate of the outlet sweep gas (ml(STP)/min).
obtained experimental stoichiometry for BSCF 5582 S: active area of the membrane (cm2).
material was found to be close to the nominal
Ba:Sr:Co:Fe ratio of 0.50:0.50:0.80:0.20. The solid-state III.3. Oxygen Flux Measurements
method gave a ratio of 0.54:0.50:0.79:0.22 while the
citrate method gave 0.54:0.51:0.79:0.21. EDAX analysis The oxygen flux measurements were performed at 650
was also applied to the commercial LSCF 6428 material ºC, 750 ºC and 850 ºC with the helium sweep gas and air
giving a La:Sr:Co:Fe ratio of 0.64:0.41:0.20:0.75 close to as the inlet gases at either side of the membrane. The
the nominal ratio of 0.60:0.40:0.20:0.80. flow rate of helium and air was kept constant at 30
As previously mentioned the grain boundaries of the ml(STP)/min and 15 ml(STP)/min respectively for all
BSCF 5582 membrane formed from the powder made by experiments. At ambient temperature no significant leak
the solid-state method presented occluded particles. was observed. The membrane was adequately gas-tight
EDAX showed that the grain boundaries and occlusions and only a trace amount, <0.1 % nitrogen, was detected
possessed the same metal ion stoichiometry as that for in the outlet of the sweep gas. The system was then
the grains. heated at a low ramp rate (1 ºC/min) until the desired
temperature was reached. However, on increasing the
temperature, leakage of nitrogen into the helium side was
III.2. Analysis of Leak Contribution observed.
In this work, an alumina based cement was used to The leak rate increased with temperature for all three
minimize the difference of thermal expansion between membrane materials tested and it was more pronounced
the dense alumina reactor tubes and the membrane itself. for the BSCF 5582 membranes. This was most likely due
Leak development was monitored in the experimental to differences in the thermal expansion of the
operating temperature range of 650ºC - 850ºC. membranes, alumina tube and the cement causing
This involved simply analysing the outlet gases for mechanical failure and indicating the unsuitability of the
any trace of nitrogen as air is supplied to one side of the binary cement for the LSCF 6428 and BSCF 5582
membrane. membranes used in this work.
During the experiments, the concentration of nitrogen The leakage of oxygen into the helium sweep side
in the outlet sweep gas was measured by the GC and would also lower the oxygen chemical potential
attributed to imperfect sealing of the membrane at high difference across the membrane, which would result in a
temperatures. Therefore the mole fraction of oxygen in lowering of the oxygen flux. Figs. 5 show the total
the outlet sweep gas measured at the GC due to the leak oxygen and the oxygen due to leakage as the percentage
contribution can be calculated using Eq. (1), where 0.79 detected in the outlet sweep gas for all three membranes.
and 0.21 refer to the mole fraction of oxygen and The extracted oxygen permeation fluxes for the
nitrogen, respectively, in air: membranes tested in this work, given in
ml(STP)/cm2.min, are presented in Figs. 6.
 0.21  The BSCF membrane formed from the powder made
OL     NL (1) by the solid-state method gave an oxygen flux of 0.26
 0.79  ml(STP)/cm2.min at 850ºC and is much lower than the
value reported in the literature of 1.4 ml(STP)/cm2.min at
where: 850ºC [5]. The micrograph of this membrane, as
OL: mole fraction of oxygen due to the air leak. previously stated, presents occluded material within the
NL: mole fraction of nitrogen due to the air leak. grain boundaries. The occlusions could be a hindrance to
Therefore, the mole fraction of oxygen as a result of oxygen permeation. The fact that the metal ion
solid-state oxygen permeation through the membrane can stoichiometry at the grain boundaries is the same as that
be calculated (Eq. (2)): for the grains does not mean that the same phase is
present in the two locations. Indeed the occlusions may,
OMP  OS  OL (2) for instance, comprise of unreacted metal oxides. In such
a case we might expect oxygen conductivity through the
where: grain boundaries to be reduced. The BSCF 5582
OMP: mole fraction of oxygen permeation through the membrane synthesised by the citrate method gave a
membrane. permeation flux of 0.40 ml(STP)/cm2.min at 850 ºC. This
OS: mole fraction of oxygen in the sweep gas. value is less than that previously reported in the
Consequently: literature, 1.3 ml(STP)/cm2.min at 850 ºC [5], but is
significantly greater than that observed in this work for
FSG  OMP the BSCF 5582 membrane formed from the solid-state
J O2  (3) powder under the same operating conditions.
S
321
(a) (a)
(b)
(b)
(c)
Figs. 5. Graphs showing: (a) BSCF 5582 membrane formed from the (c)
powder produced by the solid-state method; (b) BSCF 5582 membrane
formed from the powder produced by the citrate method; (c) LSCF Figs. 6. Extracted oxygen permeations fluxes (a) BSCF 5582
6428 membrane. (■) refers to total detected oxygen, () refers to membrane formed from the powder produced by the solid-state
oxygen leak contribution method, (b) BSCF 5582 membrane formed from the powder produced
by the citrate method and (c) the LSCF 6428 membrane
It is worth noting that no occluded material was
observed in the grain boundaries by SEM for the
membrane formed by the powder from the citrate IV. Conclusion
method. At 850 ºC the LSCF 6428 membrane gave an Three MIEC perovskite-based membranes, BSCF
oxygen permeation flux of 0.09 ml(STP)/cm2.min. 5582 formed by powders produced by the solid-state and
Similar results have previously been reported in the sol-gel methods, and a commercial LSCF 6428 formed
literature in the range 0.04 ml/cm2.min to 0.13 from powder produced by combustion spray pyrolysis,
ml(STP)/cm2.min though not under exactly the same were investigated for oxygen permeation. Of the three
conditions as employed in this study; these other membranes tested, BSCF 5582 synthesised by the citrate
experimental conditions having involved the use of method gave the highest oxygen flux. However, the
tubular membranes, working at 1000 oC and using maximum oxygen flux obtained was less than that
membranes of different thicknesses (0.96 to 1.5 mm) previously reported in the literature [5]. The BSCF 5582
[29], [30], [31]. Successful membrane materials must membrane formed from the solid-state powder gave an
possess both long-term stability and provide a constant oxygen flux lower than the BSCF 5582 membrane
oxygen flux suitable for the reaction process. The oxygen prepared by the citrate method. The flux for LSCF 6428
flux of 5-10 ml(STP)/cm2.min [22] has been deemed to was comparable to previously reported values [29]. The
be required for commercial-scale syngas production. The oxygen permeation of the three membranes investigated
performance of the BSCF 5582 and LSCF 6428 fell short of the desired 5-10 ml(STP)/cm2.min. In this
membranes used in this study fell short of this value with sense, hollow fibre membranes might be considered as a
0.40 ml(STP)/cm2.min being the highest flux obtained promising membrane geometry as it can provide a much
using the BSCF 5582 membrane synthesised by the higher surface area per unit volume [32]. Tan et al. [33]
citrate method.
322
have reported that the oxygen permeation process lanthanum cobaltite perovskite-type membrane reactors for partial
oxidation of methane to syngas, J Membrane Sci 166 (2000) 13-
through LSCF 6428 hollow fibre membranes can be
22.
significantly improved after some surface modifications [17] S. Pei, M. S. Kleefisch, T. P. Kobylinski, J. Faber, C. A. Udovich,
helping to achieve anticipated oxygen permeation fluxes V. Zhang-McCoy, B. Dabrowski, U. Balachandran, R. L.
of commercial interest. Mieville, R. B. Poeppel, Failure mechanisms of ceramic
membrane reactors in partial oxidation of methane to synthesis
gas, Catal Lett 30 (1994) 201-212.
[18] B. Ma, U. Balachandran, Oxygen nonstoichiometry in mixed-
Acknowledgements conducting SrFeCo0.5Ox, Solid State Ionics 100 (1997) 53-62.
[19] Z. Shao, W. Yang, Y. Cong, H. Dong, J. Tong, G. Xiong,
Luiz Carlos L. Santos would like to thank Investigation of the permeation behavior and stability of a
CAPES/Brazil for financial support. The authors would Ba0.5Sr0.5Co0.8Fe0.2O3−δ oxygen membrane, J Membrane Sci 172
like to thank Ian Metcalfe, Ronald Hughes, Richard (2000) 177-188.
Plaisted, Patrick Hill, Rebwar Salar and Alan Thursfield [20] A. Vivet, P. M. Geffroy, V. Coudert, J. Fouletier, N. Richet, T.
Chartier, Influence of glass and gold sealants materials on oxygen
and for their practical assistance. permeation performances in La0.8Sr0.2Fe0.7Ga0.3O3−δ perovskite
membranes, J Membrane Sci (2011) 366, 132-138.
[21] X. Qi, F. T. Akin, Y. S. Lin, Ceramic–glass composite high
References temperature seals for dense ionic-conducting ceramic membranes,
J Membrane Sci 193(2001) 185-193.
[1] M. Anderson, Y. S. Lin, Carbon dioxide separation and dry [22] R. Bredesen, J. Sogge, A technical and economic assessment of
reforming of methane for synthesis of syngas by a dual-phase membrane reactors for hydrogen and syngas production, SINTEF
membrane reactor, AIChE Journal 59 (6) (2013) 2207-2218. report S96017, 1996 in: H. J. M. Bouwmeester, Catal Today 82
[2] D. J.Wilhelm, D. R. Simbeck, A. D. Karp, R. L. Dickenson, (2003) 141-150.
Syngas production for gas-to-liquids applications: technologies, [23] Z. Shao, G. Xiong, H. Dong, W. Yang, L. Lin, Synthesis, oxygen
issues and outlook, Fuel Process Technol 71 (2001) 139-148. permeation study and membrane performance of a
[3] A. Thursfield, I. S. Metcalfe, The use of dense mixed ionic and Ba0.5Sr0.5Co0.8Fe0.2O3−δ oxygen-permeable dense ceramic reactor
electronic conducting membranes for chemical production, J for partial oxidation of methane to syngas, Sep Purif Technol 25
Mater Chem 14 (2004) 2475-2485. (2001) 97-116.
[4] H. Wang, Y. Cong, W. Yang, Investigation on the partial [24] Z. Shao, H. Dong, G. Xiong, Y. Cong, W. Yang, Performance of
oxidation of methane to syngas in a tubular Ba0.5Sr0.5Co0.8Fe0.2O3−δ a mixed-conducting ceramic membrane reactor with high oxygen
membrane reactor, Catal Today 82 (2003) 157-166. permeability for methane conversion, J Membrane Sci 183 (2001)
[5] L. Tan, X. Gu, L. Yang, W. Jin, L. Zhang, N. Xu, Influence of 181-192.
powder synthesis methods on microstructure and oxygen [25] H. Dong, Z. Shao, G. Xiong, J. Tong, S. Sheng, W. Yang,
permeation performance of Ba0.5Sr0.5Co0.8Fe0.2O3−δ perovskite- Investigation on POM reaction in a new perovskite membrane
type membranes, J Membrane Sci 212 (2003) 157-165. reactor, Catal Today 67 (2001) 3-13.
[6] L. Roses, G. Manzolini, S. Campanari, E. De Wit, M. Walter, [26] A. J. Burggraaf, K. Keizer, in: R. R., Bhave, Inorganic
Techno-economic Assessment of Membrane Reactor membranes: synthesis, characteristics and applications (Van
Technologies for Pure Hydrogen Production for Fuel Cell Vehicle Nostrand Reinhold, 1991).
Fleets, Energy&Fuel, Article ASAP [27] S. Liu, X. Tan, K. Li, R. Hughes, Synthesis of strontium cerates-
dx.doi.org/10.1021/ef301960e (2013). based perovskite ceramics via water-soluble complex precursor
[7] Y. Zeng, S. Tamhankar, N. Ramprasad, F. Fitch, D. Acharya, R. routes, Ceram Int 28 (2002) 327-335.
Wolf, A novel cyclic process for synthesis gas production, Chem [28] M. Gaudon, C. Laberty-Robert, F. Ansart, P. Stevens, A. Rousset,
Eng Sci 58 (2003) 577-582. Preparation and characterization of La1–xSrxMnO3+δ (0 x 0.6)
[8] K. Kendall, C. M. Finnerty, G. Saunders, J. T Chung, Effects of powder by sol–gel processing, Solid State Sci 4 (2002) 125-133.
dilution on methane entering an SOFC anode, J Power Sources [29] J. A. Lane, S. J. Benson, D. Waller, J. A. Kilner, Oxygen
106 (2002) 323-327. transport in La0.6Sr0.4Co0.2Fe0.8O3−δ, Solid State Ionics 121 (1999)
[9] K. Zhang, Y. L. Yang, D. Ponnusamy, A. J. Jacobson, K. Salama, 201-208.
Effect of microstructure on oxygen permeation in [30] S. Li, H. Qi, N. Xu, J. Shi, Tubular dense perovskite type
SrCo0.8Fe0.2O3−δ, J Mater Sci 34 (1999) 1367-1372. membranes. Preparation, sealing, and oxygen permeation
[10] P. M. Geffroy, J. M. Bassat , A. Vivet, S. Fourcade, T. Chartier, properties, Ind Eng Chem Res 38 (1999) 5028-5033.
P. N. Del Gallo Richet, Oxygen semi-permeation, oxygen [31] S. Diethelm, J. Van Herle, Oxygen transport through dense
diffusion and surface exchange coefficient of La0.6Sr0.4Fe0.8Co0.2O3-δ perovskite-type permeation membranes, J
La(1−x)SrxFe(1−y)GayO3−δ perovskite membranes, J Membrane Sci Eur Cera. Soc 24 (2004) 1319-1323.
354 (2010) 6-13. [32] L. C. L. Santos, C. Moraes, R. Hughes, Characterization of
[11] H. Wang, Y. Cong, W. Yang, Oxygen permeation study in a hollow fibre membranes for oxygen permeation and parcial
tubular Ba0.5Sr0.5Co0.8Fe0.2O3-δ oxygen permeable membrane, J oxidation reactions, BJPG 5(1) (2011) 045-054.
Membrane Sci 210 (2002) 259-271. [33] X. Tan, Z. Wang, H. Liu, S. Liu, Enhancement of oxygen
[12] Y. Teraoka, H. M. Zhang, S. Furukawa, N. Yamazoe, Oxygen permeation through La0.6Sr0.4Co0.2Fe0.8O3−δ hollow fibre
permeation through perovskite-type oxides, Chem Lett 14 (1985) membranes by surface modifications, J Membrane Sci 324 (2008)
1743-1746. 128–135.
[13] Y. Teraoka, H. M. Zhang, K. Okamoto, N. Yamazoe, Mixed
ionic-electronic conductivity of La1-xSrxCo1-y FeyO3−δ, Mat Res
Bull 23 (1988) 51-58.
[14] R. J. Chater, S. Carter, J. A. Kilner, B. C. H. Steele, Development Authors’ information
of a novel SIMS technique for oxygen self-diffusion and surface 1
Federal University of Bahia - Department of Materials Science and
exchange coefficient measurements in oxides of high diffusivity,
Technology - R. Prof. Aristides Novis, 2, 3º andar, Federação, CEP
Solid State Ionics 53-56 (1992) 859-867.
40210-630, Salvador/BA, Brazil.
[15] U. Balachandran, J. T. Dusek, R. L. Mieville, R. B. Poeppel, M.
S. Kleefisch, S. Pei, T. P. Kobylinski, C. A. Udovich, A. C. Bose, 2
Federal University of Bahia – Postgraduate Program of Chemical
Dense ceramic membranes for partial oxidation of methane to
Engineering - R. Prof. Aristides Novis, 2, 2º andar, Federação, CEP
syngas, Appl Catal, A 133 (1995) 19-29.
40210-630, Salvador/BA, Brazil.
[16] W. Jin, S. Li, P. Huang, N. Xu, J. Shi, Y. S. Lin, Tubular
323
3
Federal University of Rio de Janeiro - Department of Chemical
Engineering – Av. Horário Macedo, 2030, Ilha do Fundão, CEP 21941-
909, Rio de Janeiro/RJ, Brazil.
Luiz Carlos L. Santos (Rio de Janeiro, 1976)

has a bachelor’s degree in Chemical
Grande do Norte, Natal, Brazil (1999), a
master's in Chemical Engineering from the
Federal University of Rio Grande do Norte,
Natal, Brazil (2002) and a PhD in Chemical
Engineering from the University of Manchester,
United Kingdom (2007). He has experience in Chemical Engineering
with an emphasis in Oil and Gas. His research interests cover the
following subjects: petroleum, biofuels, surfactants, microemulsions,
adsorption, catalysis and ceramic membranes.
Dr. Santos is a member of the Society of Petroleum Engineering.
E-mail: lclsantos@ufba.br
Ana Katerine de Carvalho Lima Lobato

(Natal, 1973) has a bachelor’s degree in
Chemical Engineering from the Federal
University of Rio Grande do Norte, Natal, Brazil
(1999), a master's in Chemical Engineering from
the Federal University of Rio Grande do Norte,
Natal, Brazil (2003) and a PhD in Chemical
Grande do Norte, Natal, Brazil (2010) with a sandwich period at the
University of Manchester, United Kingdom. Currently, she is a
researcher and an assistant professor in the Post-Graduate Program in
Chemical Engineering at the Federal University of Bahia.
She has experience in Chemical Engineering with an emphasis in
Biochemical Engineering. Her research interests cover the following
subjects: petroleum, biofuels, biosurfactant production, antibiotic
production and metabolic flux analysis.
Prof. Lobato is a member of the technical chamber of biological
products at the general management of medicines national agency of
sanitary surveillance.
Caetano Moraes (Rio de Janeiro, 1953)

graduated in Chemistry from the University of
Brasilia (1974), Master's degree in Chemical
de Janeiro (1978) and Ph.D. in Chemical
Engineering from the University of Salford -
United Kingdom (1983). He is currently
Associate Professor of the Federal University of
Rio de Janeiro. He has experience in Chemical Engineering, with
emphasis in Industrial Processes of Chemical Engineering, mainly in
the following areas: combustion reactions, catalysts based on
palladium/alumina, alumina-based catalysts supported nickel and
production of acetone.
324
2035-1763(201307)5:4;1-4
Copyright © 2013 Praise Worthy Prize S.r.l. - All rights reserved

Ireche Vol 5 N 4

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Ireche Vol 5 N 4

Hochgeladen von

Copyright:

Verfügbare Formate

International Review of

Performance Evaluation of Ducts with Non-Circular Shapes: 265

Copyright © 2013 Praise Worthy Prize S.r.l. - All rights reserved

Performance Evaluation of Ducts with Non-Circular Shapes:

Valentin M. Petkov, Ventsislav D. Zimparov

Keywords: Performance Evaluation Criteria, Single-Phase Laminar Flow, Non-Circular Ducts,

Nomenclature  Dynamic viscosity [ Pas ]

T Temperature [K] Q* Dimensionless heat transfer rate, Q / Q c

Sgen N S ,T  o N S ,P III. Performance Evaluation Criteria

The calculated values of i* by Eq. (8) for θ  30 °

When * increases, i* achieves almost constant

Fig. 3(b). The variation of the general criterion N S i

N Si FG-1b   30o , N S i*  0.335 ; for   45o ,

Fig. 4(d). The variation of the general criterion N Si*

can be only obtained for small values of o

N S i*  0.49 ; for o

The augmentation entropy generation number N S

The calculated values of i*

like Fig. 6(b)) perform even worse. For all bundles 0

N Si FG-2b entropy generation number N S caused by the substantial

N Si FG-2b III.2.1. Case FN-1

It seems that in some cases, o  1 , the use of bundles N S ,T  o* To* 1 , N S ,P  P*  1

1  o*  As seen, it is not possible to decrease the heat transfer

100 Fig. 9(a). The variation of the general criterion N S A*

N S A FN-1 pumping power P*  1 , decreased entropy generation

Fig. 9(c). The variation of the general criterion N S A*  f Re *

100 transfer performance of the unit. The difference between

Fig. 10(b). The variation of the pumping power P*

P*  0.236  0.019 10-1

For elliptical channels:

For trapezoidal and hexagonal ducts:

Figs. 11. As seen, for bundle with rectangular ducts, Fig.

substantially decreases with increasing of * and o .

with the increase of * and o . Figs. 12 and 13 present 10-3

section, (2011) International Review of Chemical Engineering Authors' information

Performance Evaluation of Photo-Fenton and Fenton Processses

I. Introduction Among biological methods, methanization is by far

•OH + e- + H+ → H2O Eo = 2.730 V (1)

[40] pointed pH out as a decisive factor for an TABLE VI

IV. Conclusion treatment in dairy industries—possibility of reuse, Desalination

Hélcio Izário Filho. Doctor of Science in Chemistry from by the State

Gisella Lamas Samanamud. Masters in Chemical Engineering by the

André Luiz de Souza. MSc in Chemical Engineering from the

Influence of Magnetic Field on the Efficiency

Waleed M. Sh. Alabdraba, Mohamed B. A. Albayati,

Keywords: Magnetic Field, Turbidity, Coagulation Process

I. Introduction A magnet is a material or object that produces a

effects of magnetic field obviously, and to show the

IV. Result and Discussion

As it shows that characteristics of the sample (1) had

Fig. 4. The relation between turbidity and the different periods

and rapid mixing, but it is better to be shed the

Mohamed B. A. Albayati born in Iraq-Kirkuk

Evaluation of a System for Dairy Wastewater Treatment

Rodrigo F. S. Salazar, Marcos F. Oliveira, Marco A. K. Alcântara, Hélcio J. Izário Filho

Keywords: AOP, Photocatalysis, TiO2, Dairy Wastewater

I. Introduction 3) treatment systems based on enzymatic hydrolysis

Results show that the more the reaction time

approximately 40%. Similar behaviors were observed in

TABLE IV After photo-treatment, the effluent samples obtained

TABLE VIII aeration pond biodegradation system, present a great and

Hélcio José Izário Filho. Lorena – SP, Brazil.

A Research Note on Oil Extraction ffrom