Journal of Environmental Chemical Engineering 2 (2014) 4655

Contents lists available at ScienceDirect

Journal of Environmental Chemical Engineering

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

Recovering urea from human urine by bio-sorption onto Microwave Activated

Carbonized Coconut Shells: Equilibrium, kinetics, optimization and eld studies
Mahesh Ganesa Pillai *, Prithvi Simha, Ashita Gugalia
Mass Transfer Laboratory, Chemical Engineering Division, School of Mechanical and Building Sciences (SMBS), VIT University, Vellore 632014, Tamil Nadu, India

A R T I C L E I N F O

A B S T R A C T

Article history:
Received 28 August 2013
Accepted 26 November 2013

Microwave Activated Carbonized Coconut Shell (MACCS) was used to recover urea from human urine.
Batch adsorption studies were conducted to evaluate the effect of initial adsorbate concentration (25%
100%), contact time, carbon loading (13 g) and shaking speed (150200 rpm) on the removal of urea at
30 8C. Microwave activation was performed at 180 W (microwave output power) for 10 min. The
sorption data were tted to Langmuir, Freundlich, Tempkin, FloryHuggins and DubininRadushkevich
isotherm models. Results showed that the maximum monolayer adsorption capacity of the MACCS
powder was 256.41 mg g1. The FloryHuggins model was found to best describe the urea uptake
process since it demonstrated the minimum deviations from the experimental data. The kinetic data was
tted to pseudo-rst-order, pseudo-second-order and intra-particle diffusion models, and was found to
follow closely the pseudo-rst order kinetic model. Based on the Central Composite Rotary Design, a ve
factor interaction model and a quadratic model were respectively developed to correlate the adsorption
variables to the adsorption capacity. Field studies were conducted to determine the percentage biomass
increase and relative agronomic effectiveness for soil treated with the urea adsorbed MACCS powder.
Microwave activated carbonized coconut shell was shown to be a promising adsorbent for recovery and
removal of urea from human urine solutions.
2013 Elsevier Ltd. All rights reserved.

Keywords:
Coconut shell
Human urine
Activated carbon
Response surface
Methodology
Urea

Introduction
The objective of any sanitation system is to protect and promote
human health by breaking the cycle of disease. Most of the existing
sanitary systems prevent exposure of humans to harmful
pathogens found in the excrement. These systems carry the waste,
remove pathogens and pollutants, and nally release the contents
back into the nature, often in large volumes of diluted wastes,
which leads to eutrophication. By shifting away from todays
paradigm which focuses on what must be removed from
wastewater, to a new paradigm focusing on what can be recovered
sanitation systems may begin to be described as Resource
Recovery Systems [1].
The key role of inorganic fertilizers (NPK) in the phenomenal
growth of food grain production is well established [2]. However,
majority of the nitrogen is made from natural gas which is subject
to price change and availability of methane, whilst, the global
potassium and phosphorous mines are set to run out in less than a
century [3]. Besides, the regular usage of inorganic fertilizers
contributes to a progressive increase in soil acidity. These factors

* Corresponding author. Tel.: +91 97902 99447; fax: +91 4162243092.

E-mail addresses: maheshgpillai@vit.ac.in, clickonmag@yahoo.co.in
(M.G. Pillai).
2213-3437/$ see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jece.2013.11.027

have made many farmers switch over to organic farming. However,

the progress of organic farming has been very slow due to rapid
decline in organic raw materials such as animal wastes, crop
residues and green manure [4]. The application of fresh human
urine as a source of plant nutrient is rapidly growing in agricultural
practices, and has already been successfully exploited in many
countries for cultivating a wide variety of crops [57]. The results
validated that the crops grown using human urine, recorded a
higher yield with greater nutritional value and taste similar to
crops grown in normal soil. In addition, they do not pose any
signicant hygienic threat or leave any distinct avor in food
products.
In particular, the use of human urine can help achieve a closed
loop fertility system that can re-circulate nutrients from human
beings back to agricultural elds. The initial phase of this system
involves separating urine from the waste stream, keeping the
excrement dry and speeding up the decomposition of pathogens,
while the latter phase involves recovery of valuable nutrients from
the source separated urine [8]. Though extensive research has
already been carried out on urine diversion systems, studies on the
recovery of nutrients (urea) are relatively few; particularly
separation of urea from urine has not been adequately investigated
[9]. Although conventional methods such as reverse osmosis,
chemical precipitation, electro-chemical process and ion exchange
can be used for the removal of urea from urine, strict operating

M.G. Pillai et al. / Journal of Environmental Chemical Engineering 2 (2014) 4655

Nomenclature
A
B
Bac
BDAP
Bt
Ce
C0
Es
kad
kid
k1
k2
Ka
Kf
KL
LoD
LoQ
n
qe
qm
qs
q, qt
R
RL
t
T
V
W

u
e
DG8

Tempkin isotherm equilibrium binding constant

(L g1)
Tempkin isotherm constant (J mol1)
biomass of plants in absolute control (g)
biomass of plants in di-ammonium phosphate (g)
biomass of plants in treatment (g)
equilibrium liquid-phase concentration of urea
(mg L1)
initial liquid-phase concentration of urea (mg L1)
mean sorption energy (kJ mol1)
DubininRadushkevich
isotherm
constant
(mol2 kJ1)
intra-particle
diffusion
rate
constant
(mg g1 min1/2)
rst order rate constant (min1)
second order rate constant (g mg1 min1)
equilibrium constant of adsorption
Freundlich isotherm constant related to adsorption
capacity (mg g1) (L g1)n
Langmuir isotherm constant (L mg1)
limit of detection (mg L1)
limit of quantitation (mg L1)
adsorption intensity
amount of urea adsorbed at equilibrium (mg g1)
maximum adsorption capacity for the solid phase
loading (mg g1)
theoretical maximum capacity (mg g1)
amount of urea adsorbed at time t (mg g1)
universal gas constant (8.314 J mol1 K1)
separation factor
adsorption time (min)
absolute temperature (K)
volume of the urine solution (L)
mass of dry adsorbent used (g)
degree of surface coverage
Polanyi potential
Gibbs free energy of sorption (kJ mol1)

conditions, high cost, long periods and bad impact from shock
loads make them undesirable to be used in practical industrial
applications [10]. Compared to these methods, adsorption has
drawn more attention by researchers due to its feasibility, high
safety and low cost [11].
Biomaterials have the potential to be used as low cost ecofriendly adsorbents because of the unused resources that are
widely available [12]. Coconut (Cocos nucifera) generates a huge
amount of solid waste, mostly in the form of ber and shell. It has
been reported that the processing of coconut shells into granular
activated carbon of sufcient density, hardness and porosity, can
potentially provide an inexpensive and renewable adsorbent
[13,14]. Hence, the feasibility of employing activated carbon
prepared from coconut shell toward the removal of urea from
human urine was analyzed. The study attempts to: (i) enhance the
surface area of coconut shell under microwave irradiation before
carbonization; (ii) investigate the effect of contact time, initial
concentration of adsorbate and adsorbent dosage on urea
adsorption capacity; (iii) analyze the adsorption equilibrium and
kinetics; (iv) characterize the Microwave Activated Carbonized
Coconut Shell (MACCS) powder; and (v) optimize the process

47

variables, using Response Surface Methodology (RSM) for the

particular batch adsorption system.
Materials and methods
Raw materials
The adsorbate used for the studies was urine obtained from ten
healthy young male volunteers (early twenties) with a wellbalanced diet. Fresh urine samples collected in air-tight containers
were immediately put on ice and stored at 20 8C for two days
until the start of batch tests. The urine samples were thawed just
before the investigation. Urine samples were characterized to
calculate its major constituents and results have been summarized
in Table 1. The coconut shells utilized for the preparation of
activated carbon were obtained locally. The shells of the fruit were
washed with distilled water to remove dirt from its surface and
dried at 105 8C for 24 h. Subsequently, the shells were crushed to a
size of 12 cm. Proximate analysis of the raw coconut shells used in
this study by ASTM-1762 standards revealed the following;
moisture content (5.64%), volatile matter (73.44%), xed carbon
content (20.29%) and ash content (0.63%). All chemicals used in the
study were purchased from Nice Chemicals Pvt. Ltd, Cochin, India
and were used without further purication.
Preparation and characterization of MACCS powder
The sieved coconut shells were exposed to microwave
irradiation at an output power of 180 W for 10 min (selected as
the heating period, based on preliminary runs) before carbonization. The samples were placed in porcelain boats and heated in a
furnace at a rate of 24 8C per min from room temperature to 500 8C,
and maintained at this temperature for 1 h. The carbon thus
obtained was cooled to room temperature, ground using a mortar,
sieved to 100 mesh size (0.149 mm) and stored in tightly closed
bottles for further analysis. The surface texture and the development of porosity for the precursor and the prepared activated
carbon were analyzed using Scanning Electron Microscopy (FESEM, SUPRA 55, Carl Zeiss) with a 20 kV electron source. The
surface organic structures of the samples were detected using
Fourier Transform Infra-Red spectroscope (IRAfnity-1, Shimadzu)
recorded at 4 cm1 of resolution and 60 scans min1 between 4000
and 400 cm1. The pore texture was characterized by N2
adsorption at 77 K using a BET surface area analyzer (Micromeritics ASAP 2020). Prior to adsorption, the carbon was degassed
at 300 8C in a vacuum condition for 2 h. The BET surface area was
calculated from the N2 adsorption isotherm using the Brunauer
EmmettTeller (BET) equation [15]. The samples were characterized for their total pore volume, pore size, and specic surface area
[16]. Iodine Number (mg iodine/g carbon) was determined by
using a 0.1 N standardized Iodine solution and by extrapolating to
0.02 N by an assumed isotherm slope using the standard method.

Table 1
Major constituents and their concentration in the urine samples.
S. No.

Constituents in urine

Concentration (mg L1)

1
2
3
4
5
6
7
8

Urea
Creatinine
Chlorides
Sodium
Potassium
Sulfates
Phophates
Ammonium

19,800
980
5945
3195
1480
810
685
512

Total solutes

33,407

48

M.G. Pillai et al. / Journal of Environmental Chemical Engineering 2 (2014) 4655

Apparent Density was determined using ASTM D2854 standard

method.

randomized in order to minimize the effects of the uncontrolled

factors. The level of variables considered for CCRD and the set of
experiments in terms of coded and actual values are given in Table 3.

Batch adsorption studies

25 ml of urine solutions with different initial concentrations of
urea (25%: 4950 mg L1, 50%: 9900 mg L1, 75%: 14,850 mg L1
and 100%: 19,800 mg L1) were prepared in Erlenmeyer asks
(250 ml). 13 g of MACCS powder was added to 25 ml of urine
solutions, and kept in an incubator shaker, where shaking speed
(125200 rpm) and time (0360 min) were varied continuously.
The asks were then removed from the shaker and sample was
withdrawn to be analyzed using a UVvisible spectrophotometer
(Shimadzu UV-1601, Japan) at 430 nm [17]. The amount of urea
adsorbed at equilibrium, qe (mg g1), was calculated using Eq. (1).


C 0  C e V
qe
(1)
W
where C0 and Ce (mg L1) are the initial and equilibrium liquidphase concentrations of urea, V is the volume of the solution (L)
and W is the mass of dry adsorbent used (g). Measurement of
absorbance for different urea solutions was performed with the
number of scans, xed at 10, to yield a Photometric accuracy of
0.002 Abs; Photometric repeatability of 0.001 Abs with a noise
level of 0.002 Abs. The LoD (Limit of Detection) and LoQ (Limit of
Quantitation) were calculated based on standard deviation of the
responses (s) and the slope (S) of three independent analytical curves
as dened by ICH. LoD and LoQ were calculated as 3.3(s/S) and 10(s/
S) and were found to be 0.1002 and 0.303 mg L1, respectively.
Design of experiments
Central Composite Rotary Design
The Response Surface Methodology (RSM) is a collection of
mathematical and statistical techniques used for modeling and
analysis of problems, in which a response of interest is inuenced
by several variables and its objective is to optimize this response
[18]. Central Composite Rotary Design (CCRD), a standard RSM
design was applied for seeking the optimum conditions for the
adsorption of urea from human urine on MACCS powder. CCRD was
employed due to its advantages of rotability and ability to analyze
the complex interaction between process parameters. It is well
suited for tting a quadratic surface, which usually works well for
process optimization. It consists of F factorial points, 2k axial
points and nc center points (six replicates), where k is the number
of independent variables. These points are used to estimate the
quadratic terms. Repetition of center points provides an estimate
of pure error. The axial points (a, 0, 0), (0, a, 0) and (0, 0, a)
dene the rotary design, where variance of the predicted response is
constant at all points equidistant from the center of the design. A ve
level four factor CCRD was performed using the software Design
Expert 8.0.7.1 to nd interactive effects of the four variables where
each variable was set at 5 levels: a, 1, 0, +1, +a, where a = 1.682
(Table 2) [19]. Since k = 4, Factorial points F = 24 = 16 points, Axial
points 2k = (2  4) = 8 and center points nc = 6, the total number of
experiments, n = 16 + 8 + 6 = 30. The experimental sequence was

Statistical method
A second-order model is useful in approximating a portion of
the true response surface with a parabolic curvature [20]. It
includes quadratic terms and all cross product terms along with
the usual rst order terms and is expressed as (Eq. (2)):
Y bo b1 x1 b2 x2 b3 x3 b4 x4 b11 x1 2 b22 x2 2
b33 x3 2 b44 x4 2 b12 x1 x2 b13 x1 x3 b14 x1 x4
b23 x2 x3 b24 x2 x4 b34 x3 x4

(2)

where Y is the predicted response, b0 is the intercept coefcient, bi

is the linear terms, bii is the squared terms, bij is the interaction
terms and Xi and Xj represent the coded independent variables.
Analysis of variance (ANOVA) was performed based on the
proposed model to nd out the interaction between the variables
and the response. The statistical signicance was checked by the Ftest and quality of the t for the regression model was expressed by
the coefcient of determination (R2). Model terms were selected or
rejected based on the probability value with 95% condence level.
Finally, the three dimensional response surfaces were drawn in
order to visualize the individual and the interactive effects of the
independent variables on adsorption of urea.
Field studies
The application of MACCS powder for plant (Vigna mungo and
Vigna radiate) growth was performed at Vellore Institute of
Technology, Vellore, India, which is located at 128550 12.7900 N and
79870 59.900 E at an altitude of 216 m above from sea level. The mean
annual maximum and minimum temperatures of the site are
33.5 8C and 23.6 8C with relative humilities of 47%65%, respectively and mean annual rainfall of 954 mm. The effect of MACCS
addition on plant growth was studied with four replications. The
urea adsorbed MACCS powder used for the pot trials was prepared
using the parameters obtained from optimization studies (RSM).
For each set, a pot of 0.2 m diameter and 0.2 m height were lled
with 2 kg of 0.002 m sieved dry soil. 20 seeds were sowed in ve
pots named as control (contained no MACCS), 0.1% (w/w) of urea
adsorbed MACCS, 0.2% (w/w) of urea adsorbed MACCS, 0.3% (w/w)
of urea adsorbed MACCS, and di-ammonium phosphate (DAP) at
60 kg P2O5/ha, respectively. All pots were watered regularly.
The relative agronomic efciencies of the treatments were then
estimated to indicate their agronomic effectiveness relative to
DAP. The percentage biomass increase for various treatments was
computed using absolute control as the standard while DAP was
considered as the standard for determining relative agronomic
efciencies (Eqs. (3) and (4)).


Bt  Bac
Percentage Biomass Increase
(3)
 100
Bac
Relative Agronomic Efficiency

Bt  Bac
BDAP  Bac

 100

(4)

Table 2
The variables and their levels for the Central Composite Rotary Design.
Independent variable

Symbol

Coded levels
1.682 (a)

1

+1

+1.682 (+a)

Initial adsorbate concentration (ml)

Adsorbent loading (g)
Rotations in shaker (rpm)
Adsorption time (min)

X1
X2
X3
X4

1.987
1.159
132.9
139.1

6.25
1.50
150
180

12.5
2.00
175
240

18.75
2.50
200
300

23.01
2.841
217.0
340.9

M.G. Pillai et al. / Journal of Environmental Chemical Engineering 2 (2014) 4655

49

Table 3
Independent variables and result for the adsorption of urea from human urine onto MACCS using Central Composite Rotary Design (CCRD).
Run

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
30

Coded variables

Response Y (mg g1)

Real variables

x1

x2

x3

x4

X1

X2

X3

X4

Predicted

Observed

0
1
1
0
1
0
1
1
0
1
0
0
1
1.682
1
0
1
1
0
1
1
0
1
0
1.682
1
0
1
1

0
1
1
0
1
0
1
1
0
1
0
1.682
1
0
1
0
1
1
1.682
1
1
0
1
0
0
1
0
1
1

0
1
1
0
1
0
1
1
0
1
0
0
1
0
1
1.682
1
1
0
1
1
1.682
1
0
0
1
0
1
1

0
1
1
0
1
1.682
1
1
0
1
0
0
1
0
1
0
1
1
0
1
1
0
1
1.682
0
1
0
1
1

12.5
18.75
18.75
12.5
6.25
12.5
6.25
18.75
12.5
18.75
12.5
12.5
6.25
23.01
6.25
12.5
6.25
18.75
12.5
6.25
18.75
12.5
18.75
12.5
1.987
6.25
12.5
18.75
6.25

2
1.5
2.5
2
1.5
2
2.5
2.5
2
1.5
2
2.84
1.5
2
2.5
2
2.5
1.5
1.16
2.5
2.5
2
1.5
2
2
1.5
2
2.5
1.5

175
150
150
175
200
175
200
200
175
150
175
175
200
175
200
132.95
150
200
175
150
200
217.05
200
175
175
150
175
150
150

240
180
300
240
180
139.08
180
180
240
300
240
240
300
240
300
240
300
300
240
180
300
240
180
340.92
240
180
240
180
300

176.1
220.07
172.92
174.15
87.538
154.25
72.963
200.52
175.34
222.43
172.11
157.41
89.657
236.26
72.495
130.27
62.84
237.5
237.15
60.22
202.31
181.31
235.25
172.72
25.62
75.985
173.1
175.22
76.847

177.04
219.99
174.42
177.04
92.72
152.4
75.884
199.24
177.04
224.06
177.04
161.38
97.583
233.42
79.259
130.94
60.864
246.38
217.28
58.026
202.36
164.73
241.77
158.66
12.56
80.481
177.04
171.84
84.807

where Bt refers to the biomass of plants in treatment (g), while Bac

and BDAP represents the biomass of plants in absolute control (g)
and in di-ammonium phosphate (g) respectively.
Results and discussion
Characterization
The microstructures of the raw coconut shell and the MACCS
surface were analyzed through SEM. The development of grain
structure was poor with no visible pores in the precursor (Fig. 1(a)).
The cellular structure of coconut shell was noticeable after
microwave activation and carbonization (Fig. 1(b)). This might
be due to the movement of moisture and volatile matter, leaving
voids that may be later transformed in the nal activated carbon
product. Microwave heating led to heat generation in the bulk
which is transported to the surface due to temperature gradient.
Since pressure is directly related to temperature, the moisture
transfer from bulk to surface takes place, catalyzing pore
formation. In addition, the thermal stress resulted in development
of cracks, crevices and slits in the matrix of the ultimate carbon
material. All these effects are clearly visible in Fig. 1(b) where
increased surface area resulted in sorption of urea molecules on the
MACCS surface, unlike Fig. 1(a). The N2 adsorption isotherms were
analyzed by BET method. The MACCS was found to have a welldeveloped porosity with BET specic area of 700 m2 g1. Due to the
loss of some macropores during microwave pretreatment, the
specic area was less than that of commercially available activated
carbon. However, microwave pretreatment caused a sharp rise in
temperature lead to the contraction of the carbon skeleton and the
development of a rich pore structure in the carbon (Table 4). In
addition a high iodine number indicated a high sorption capacity
for the MACCS.
The vibrational spectrum of a molecule is considered to be a
unique physical property and is characteristic of the molecule [21].
The obtained FTIR spectrum of raw coconut shells powder

(Fig. 2(a)) revealed the peaks at 3442.9, 1629.8, 1400.3 and

1066.6 cm1, which corresponds to the presence of dimeric OH
stretch, C5
5C stretch, tertiary alcohol bend and aromatic CH place
new bend in functional groups. It is clear that the adsorbent
displays a number of absorption peaks, reecting the complex
nature of the adsorbent. Fig. 2(b) shows the shifting or disappearance of some peaks (3132.40, 1593.20, 1396.46 and 1066.64 cm1)
with simultaneous addition of new peaks at lower wave numbers.
These changes indicate the involvement of the disappeared
functional groups present on MACCS powder in sorption process.
The surface chemistry of MACCS after adsorption illustrated
intensive peaks, at 1587.42, 1402.25, 1136, 1066.64 cm1 showing
removal of OH stretch while retaining the aromatic CH place
new bend (Fig. 2(c)).
Batch adsorption studies
Effect of initial adsorbate concentration on adsorption equilibrium
Initial concentration provides an important driving force to
overcome all the mass transfer resistance of urea between the urine
solution and the MACCS surface. The effect of initial adsorbate
concentration in the range of 25% (4950 mg g1 urea) to 100%
(19,800 mg g1 urea) on adsorption (investigated under the
specied conditions; carbon loading: 2 g, shaking speed: 130 rpm
and temperature: 30 8C) is shown in Fig. 3. It is depicted from the
gure that the amount of urea adsorbed increased from 60 to
225 mg g1, as the initial concentration increased from 25% to 100%.
The adsorption was rapid in the early stages but attained an
asymptotic range for larger adsorption time. Adsorption took place
until the surface functional sites were fully occupied, later; the urea
molecules diffused into the pores of the adsorbent for further
adsorption [22]. Urea removal efciency was greater than 90% for all
the four initial concentrations studied, suggesting that initial
concentration only affects urea uptake capacity (mg g1) and not
the percentage adsorption, as higher the amount of urea in the
adsorbate more is the capacity given a constant adsorbent loading.

50

M.G. Pillai et al. / Journal of Environmental Chemical Engineering 2 (2014) 4655

Fig. 2. FT-IR images of coconut shells. (a) Raw, (b) after microwave activation and
carbonization and (c) MACCS after urea adsorption.

300
25%
50%
75%

Fig. 1. SEM images of coconut shells. (a) Raw (20 kV, 25.43 K X); (b) MACCS surface
after urea adsorption (20 kV, 5.64 K X).

Effect of adsorbent loading on adsorption equilibrium

Adsorbent dosage is an essential factor in the adsorption of urea
molecules from aqueous solutions owing to its effect on the
amount of urea adsorbed per unit mass of the adsorbent [23]. The
experiments were carried out by varying adsorbent dosage at xed
adsorbate concentration (100% urine), speed of shaker (150 rpm)
and temperature (30 8C). The results of the dependence of urea
adsorption on the amount of MACCS powder are shown in Fig. 4. It
was found that with an increase in adsorbent loading from 1 g to
2 g, though the percentage adsorption of urea increased, there was
a signicant decrease in urea uptake capacity from the maximum
of 680210 mg g1. The higher percentage of adsorption in initial
stages may be due to the availability of more surface area while
decrease in uptake capacity may be attributed to signicant
increase in active sites per gram. Thus, the driving force for mass
transfer between the bulk liquid phase and the solid phase
decreases with the passage of time [24].

Adsorption capacity ( mg / g )

100%

225

150

75

0
0

100

200

300

400

Adsorption time (min)

Fig. 3. Effect of initial concentration on adsorption kinetics of urea on MACCS
powder (Carbon loading: 2 g, Shaking speed: 130 rpm and 30 8C).

Effect of shaker speed on adsorption equilibrium

Shaking speed can inuence the distribution of the solute in the
bulk solution and the formation of the external boundary lm [25].
It was observed from Fig. 5, that an increase in shaker speed from
150 to 175 rpm leads to an increase in adsorption capacity from 75
to 80 mg g1. This gradual increase in the adsorption capacity may
be due to decrease in the boundary layer resistance leading to

Table 4
Textural characteristics and pore structures of the carbon samples.
Parameter

Test method

CCS

MACCS

CAC

Apparent density (g cm3)

Iodine number (mg g1)
Pore volume (cm3 g1)
Surface area (m2 g1)
Pore diameter (nm)
Moisture content (%)

ASTM D2854
ASTM D4607-86
BET
BET
4  pore volume/surface area
ASTM 1762

0.451
825
0.1883
625.4
0.301
5.18

0.397
911
0.2018
700.3
0.288
4.27

0.352
1000

1050

M.G. Pillai et al. / Journal of Environmental Chemical Engineering 2 (2014) 4655

800

are based on a set of assumptions that are related to the

homogeneity or heterogeneity of the adsorbent, the type of
coverage, and possibility of interaction between adsorbate species.
In this study, the equilibrium sorption data obtained were
analyzed in terms of Langmuir, Freundlich, Tempkin, Flory
Huggins, and DubininRadushkevich isotherms [2730]. Table 5
enlists all isotherm model equations, parameters and their
respective values.
The normalized deviation and normalized standard deviation
between experimental and predicted values for each isotherm
model was calculated (Eqs. (5) and (6)).

1.0 g
1.5 g

Adsorption capacity ( mg / g )

2.0 g

600

400

Normalized Deviation

200

0
100

200

300

400

Adsorption time (min)

Fig. 4. Effect of carbon loading on adsorption kinetics of urea on MACCS powder
(Initial concentration: 100%, Shaking speed: 125 rpm and 30 8C).

increase in mobility, thus forcing the urea molecules toward the

MACCS surface [26]. However, it was also noticed that, the
adsorption time has a direct effect on uptake capacity given a
constant agitation speed. At low shaking speed, the adsorbent gets
accumulated in the sample, hiding the active sites present in the
lower layers and urea gets adsorbed only on the upper layer
adsorbent active sites. This implies that the rate of adsorption
should be sufcient to ensure that, all the surface binding sites are
readily available for uptake of urea. Moreover, an increase in
adsorption time leads to strong interaction between the adsorbate
and adsorbent, thus, lowering the concentration of vacant sites.
Finally the system reaches a point where all the vacant sites are
used up and no further adsorption takes place.
Batch equilibrium studies
Adsorption isotherms are mathematical models that describe
the distribution of adsorbate in the adsorbent and solution. They
90
150 rpm
175 rpm



100 Xqeexp  qepred


N
qeexp

(5)

Normalized Std Deviation 100

s
P
2
qeexp  qepred=qeexp

N
(6)
The Langmuir model ts the data, only at low concentration. A
method to assess the goodness of t of the experimental data to the
isotherm models is to determine the correlation coefcient R2.
Upon tting the sorption data it was observed that, four isotherms
had R2 value greater than 0.90 while, it was 0.79 for the Freundlich
model demonstrating its poor agreement with the experimental
data. R2 values were greater than 0.97 for both the Dubinin
Radushkevich and FloryHuggins model making it very difcult to
decide the best model to represent the experimental data based on
the correlation coefcient. A better criterion to test the goodness of
t is by evaluating the normalized deviation and normalized
standard deviation [31]. For MACCS powder, the minimum value of
normalized deviation and normalized standard deviation are
found for the Flory Huggins model (Table 5). It is noteworthy that
these values are very high for the Freundlich model further
conrming its unsuitability and conrming the aforementioned
result based on the correlation coefcients.
The value of qm (maximum adsorption capacity) was found to
be 256.41 mg g1 which is sufciently close to the measured value
of 227.75 mg g1. The Langmuir isotherm can be expressed in
terms of a dimensionless constant called separation factor (RL)
(Eq. (7)) which is an indicator of the isotherm being favorable [32]
if 0 < RL < 1. The separation factor was found to lie within the
favorable limit indicating that the chosen operating conditions
favor the urea sorption process.

200 rpm

75

Adsorption capacity ( mg / g )

51

RL
60

1
1 K LC0

(7)

The Freundlich isotherm constant n is also an indicator of the

favorability of adsorption. For benecial adsorption n should lie
between 0 and 1. A value of 3.931 for the present study indicated
easy separation of urea from the solution and favorable adsorption.
The apparent Gibbs free energy of sorption [33] DG8 which is a
fundamental criterion of spontaneity was evaluated (Eq. (8)). The
equilibrium constant (Ka) obtained from the FloryHuggins model
was used to compute the apparent Gibbs free energy change.

45

30

15

DG RTlnKa

0
0

100

200

300

400

Adsorption time (min)

Fig. 5. Effect of shaking speed on adsorption kinetics of urea on MACCS powder
(Carbon loading: 1.5 g, Initial concentration: 25% and 30 8C).

(8)

The negative value of DG8 in Table 5 conrms the feasibility of

the process and the spontaneous nature of the sorption. The mean
sorption energy Es (kJ mol1) (Energy required to remove a
molecule from its location in the sorption space to innity) was
computed by using the DubininRadushkevich model constant
(Eq. (9)) and found to be 7.071 kJ mol1, which indicated that, the

M.G. Pillai et al. / Journal of Environmental Chemical Engineering 2 (2014) 4655

52

Table 5
Isotherm model parameters, correlation coefcients, and comparison of deviations for various isotherm models.
Isotherm

Parameter

Parameter values
2

Langmuir
qm K L C e
qe 1K
Ce

Coefcient of determination (R )
Maximum adsorption capacity for the solid phase loading (qm)
Energy constant related to heat of adsorption (KL)
Separation factor (RL)
Normalized deviation
Normalized standard deviation

0.94
256.41 (mg g1)
0.017295 (L mg1)
0.002912
11.535
12.465

Freundlich
 
logqe logK f 1n log C e

Coefcient of determination (R2)

Adsorption intensity (n)
Isotherm constant related to adsorption capacity (Kf)
Normalized deviation
Normalized standard deviation

0.7983
3.931
38.69 (mg g1) (g1)3.931
21.715
23.051

Tempkin
qe = Bln(A) + Bln(Ce)

Coefcient of determination (R2)

Constant related to heat of sorption (B)
Equilibrium binding constant (A)
Normalized deviation
Normalized standard deviation

0.9187
34.175 (J mol1)
0.5989
16.453
20.665

FloryHuggins
log Cue logKa mlog1  u

Coefcient of determination (R2)

Equilibrium constant of adsorption (Ka)
Gibbs free energy of sorption (DG8)
Normalized deviation
Normalized standard deviation

0.9753
1.7546  105
27.856 (kJ mol1)
0.5394
0.6244

DubininRadushkevich
ln(qe) = ln(qs)  kade2

Coefcient of determination (R2)

Theoretical maximum capacity (qs)
DubininRadushkevich model constant (kad)
Mean sorption energy (Es)
Normalized deviation
Normalized standard deviation

0.9827
222.294 (mg g1)
0.01 (mol2 kJ1)
7.071 (kJ mol1)
6.101
6.558

process was physisorption [34].

!
1
Es p
2kad

(9)

Further, a positive value of the mean sorption energy indicated

that the sorption was endothermic with higher solution temperature favoring the adsorption.
Adsorption kinetics
To examine the controlling mechanism of sorption process, the
pseudo-rst order rate model, pseudo-second order rate model,
and intra-particle diffusion model were tested on the equilibrium
data. The pseudo-rst order equation proposed by Langergren and
Svenska [35] was used as in Eq. (10).
logqe  qt logqe  k1 t

q  t kid t 1=2 C

(12)

A plot of qt versus t1/2 was generated to calculate kid

(mg g1 min1/2), the intra-particle diffusion rate constant and C,
the intercept (Fig. 6). The linearity of the plot suggested that intraparticle diffusion was involved in the adsorption process but since
the regression lines did not pass through the origin, it indicated
some degree of boundary layer control and that intra-particle
diffusion was not the only rate limiting step. Table 6 enlists the
values of constants calculated for all the kinetic models tested.

(10)

where qe and qt are the amounts of urea adsorbed (mg g1) at

equilibrium and at time t (min), respectively and k1 is the rate
constant of adsorption (min1). The plot of log (qe  qt) versus time
indicated that the sorption data ts the rst order rate expression
for different urea concentrations.
The second-order kinetic model (Eq. (11)) can be expressed as
t
1
t

q k2 q2e qe

0.0603 min1 and are independent of initial concentration. Intraparticle diffusion model based on the theory proposed by Weber
and Morris was tested to identify the diffusion mechanism [36]. It
is an empirically found functional relationship, common to most
adsorption processes, where uptake varies almost proportionally
with t1/2 rather than the contact time t (Eq. (12)).

(11)

where k2 (g mg1 min) is the second order rate constant. The linear
plots of t/q versus time indicated that the data does not t the
model well. Though the values of qe calculated from equations
differ from the experimental values for both pseudo rst and
second order model, the magnitude of correlation coefcient for
pseudo rst order is comparatively higher which makes it an
adequate model to dene the given system. Moreover, values for k1
(0.0575 min1) are quite close to each other, for the entire
concentration range further conrming the suitability of the
0
pseudo-rst order kinetic model. k1 values lie between 0.0543 and

Optimization studies
Effect of process variables on the urea adsorption
RSM was employed to determine the complex interactive
effects of the operating variables on urea adsorption. To study the
interaction factors, experiments were performed varying physical
parameters, using experimental design. By applying multiple
regression analysis in Table 3 data, the experimental results of the
full factorial central composite design were tted to the quadratic
polynomial Eq. (2). The adjusted model obtained for urea
adsorption, as a function of the coded variables is shown in Eq. (13).
Y 177:04 65:65x1  16:62x2 10:04x3 1:86x4
 19:11x21 4:34x22  10:32x23  7:60x24  6:42x1 x2
2:39x1 x3  0:065x1 x4 1:4x2 x3  0:37x2 x4
0:13x3 x4

(13)

Adequacy check of the proposed model is a signicant

component of the analysis procedure. Good adequacy ensures

M.G. Pillai et al. / Journal of Environmental Chemical Engineering 2 (2014) 4655

53

100000
25 %
50 %
75 %

q.t ( mg.min / g )

75000

100 %

50000

25000

0
0

10

15

20

-25000

0.5

(min)

Fig. 7. Perturbation graph showing the effect of each of the independent variables
on urea adsorption while keeping other variables at their respective midpoint
levels.

0.5

Fig. 6. Intra-particle diffusion plot for adsorption of urea onto MACCS powder at
30 8C.

approximation to the real system [37]. The adequacy check was

performed by the test of signicance of individual and regression
model coefcients and test for lack-of-t. The signicant factors
can be ranked based on the F-value or P value. The larger the
magnitude of F-value and correspondingly the smaller the
Prob. > F value, the more signicant is the corresponding
coefcient. An F-value of 79.13 and corresponding P-value
<0.001 indicates that the model is statistically signicant.
Moreover, P-values (<0.05) shown in Table 5 suggested that the
signicant model terms are initial adsorbate concentration (X1),
adsorbent loading (X2), agitation speed (X3), the two level
interaction of initial adsorbate concentration and adsorbent
loading (X1X2) and the quadratic effect of initial adsorbate
concentration, adsorbent loading and agitation speed X12, X22
and X32 respectively. The equation in terms of signicant factors is

given as in Eq. (14).

Y 177:04 65:65x1  16:62x2 10:04x3  19:11x21
4:34x22  10:32x23  6:42x1 x2

(14)

The actual adsorption capacity is the measured value for a

particular run and the predicted value is evaluated from the model.
The accuracy and precision of the model was veried by the
coefcient of determination (R2 = 0.9866). The predicted R2 of
0.9329 was in reasonable agreement with the adjusted R2 of
0.9742, which advocates that the proposed model has adequate
approximation to the actual value.
Interaction effects of adsorption variables
The steep curvature in the perturbation plot between initial
concentration and adsorbent loading shows the response of
adsorption capacity was very sensitive (Fig. 7). The relatively at

Table 6
Comparison of adsorption rate constant, qe, and correlation coefcients for pseudo-rst order, pseudo-second order and intra-particle diffusion models.
Model

Model constants

Initial concentration of adsorbate

50%

75%

100%

Pseudo-rst order model

qe (mg g1)
k1 (min1)
R2

83.023
0.0543
0.9328

162.36
0.0587
0.9268

248.31
0.0603
0.9204

274.09
0.0568
0.9582

Pseudo-second order model

qe (mg g1)
k2 (g mg1 min1)
R2

84.033
3.77  104
0.8207

158.73
2.485  104
0.8676

227.27
1.987  104
0.8933

270.27
1.950  104
0.9205

Intra-particle diffusion

kid  103 (mg g1 min1/2)

R2

1.2279
0.9317

2.4526
0.9322

3.6351
0.9319

4.4915
0.9327

25%

Table 7
Biomass per plant, % biomass increase and relative agronomic efciency for direct application of urea adsorbed MACCS powder: study with Vigna mungo and Vigna radiate.
Ttm. no.

1
2
3
4
5

Treatment

Absolute control
Di-ammonium phosphate at 60 kg P2O5/ha
0.1% (w/w) of urea adsorbed MACCS
0.2% (w/w) of urea adsorbed MACCS
0.3% (w/w) of urea adsorbed MACCS

Biomass per plant (g)

% Biomass increase

Proof of concept

Residual effect

Proof of concept

Residual effect

Relative agronomic efciency

Proof of concept

Residual effect

2.27
4.45
3.18
3.95
4.16

2.79
5.09
3.84
4.76
4.95

96.03
40.01
74.01
83.25

82.43
37.63
70.61
77.42

0.41
0.77
0.86

0.45
0.85
0.93

54

M.G. Pillai et al. / Journal of Environmental Chemical Engineering 2 (2014) 4655

plots for urea adsorption represented different combinations of

any two test variables at a time, while other variables were
maintained at zero level. It can be observed that there is a strong
and signicant inuence of adsorbate concentration on the urea
uptake capacity. The trend is that, higher the adsorbate
concentration, higher is the urea uptake capacity. This indicates
that the MACCS powder adsorbs more of the urea at higher
adsorbate concentration. Fig. 8(ac) shows the dependency of
urea uptake capacity on MACCS. Though the adsorbate and
adsorbent concentration affected the uptake capacity, it was
found that the latter was slightly affected by adsorption
time and shaking speed. It was also observed from the initial
studies that the adsorption capacity of urea increases with time
up to 60 min, where the surface sites are almost occupied.
Further, increase in time has a signicant effect on urea
adsorption.
The study revealed that the optimal conditions required to
achieve the highest urea adsorption from human urine onto
MACCS powder were initial adsorbate concentration of 16.69 ml,
adsorbent loading of 1.5 g, shaking speed of 188.8 rpm and
adsorption time of 211.99 min. As such, the prediction capability
of the proposed model was veried by additional batch experiments conducted in the experimental scale of the CCRD. The
validation results clearly conrmed with 95% certainty that a ve
level four factor Central Composite Rotary Design is an effective
tool for mathematical modeling of urea adsorption onto MACCS
powder.
Field studies
To compare the plant growth, fresh biomass (excluding
the roots) was measured in each pot. Table 7 clearly indicates
that the addition of MACCS to the soil had a noteworthy effect
on the plant growth seen with an increase in biomass due to
increase in addition of MACCS to the soil. Also, the use of MACCS
resulted in agronomic efciency values (0.86 and 0.93 for 0.3%
(w/w) of MACCS) very close to 1, indicating it was almost as
effective as di-ammonium phosphate. With reference to all
treatments, the plants subjected to the treatment of absolute
control had the least values for all the growth parameters. In the
test for residual effects, increase in nitrogen content of soil
resulted in an increase in the biomass of the plant in all the
treatments.
Conclusion

Fig. 8. 3D response surfaces: (a) Interactive effects of initial adsorbent loading and
initial adsorbate concentration on adsorption capacity at adsorption time 240 min,
agitation speed 175 rpm and adsorption temperature 30 8C. (b) Interactive effects of
adsorption time and agitation speed on the adsorption capacity at initial adsorbate
concentration 12.5 mg L1 adsorbent loading 2.00 g and adsorption temperature
30 8C. (c) Interactive effects of initial adsorbent concentration and agitation speed
on adsorption capacity at adsorbent loading 2.00 g, adsorption time 240 min and
adsorption temperature 30 8C.

lines of adsorption time and shaking speed show insensitivity

of the response to change with these two parameters. The
interaction effects between all four adsorption variables were
determined from the plotted three dimensional curves. The 3D

The present research revealed that the microwave activated

carbon developed from coconut shell has the potential to adsorb
urea from human urine. The urea uptake capacity was found to
depend on the initial adsorbate concentration and MACCS loading
signicantly. Further, the SEM analysis conrmed that microwave
activation process induced well-developed porous structure in
the coconut shells. Although the adsorption process follows the
pseudo-rst order model, the linear regression and high correlation coefcient suggested that intra-particle diffusion was also
involved. Based on error analysis, FloryHuggins isotherm
showed the minimum deviation from experimental data. Moreover, a negative value of DG8 indicated the feasibility and
spontaneous nature of the process. The development of a
mathematical model for urea adsorption process optimization
using statistical design of experiments appears to be appropriate
for prediction of interaction effects between various adsorption
variables. The validation results clearly conrmed that the urea
recovery from source separated urine using MACCS powder could
play an important role in sustainable agricultural practices and
future sanitary systems.

M.G. Pillai et al. / Journal of Environmental Chemical Engineering 2 (2014) 4655

Acknowledgement
The authors acknowledge the nancial support provided by VIT
University, Vellore, India under the VIT-SMBS research grant
scheme (2040/VP-A-31082012) for conducting the initial experiments.
References
[1] J.S. Guest, S.J. Skerlos, J.L. Barnard, M.B. Beck, G.T. Daigger, H. Hilger, S.J. Jackson, K.
Karvazy, L. Kelly, L. Macpherson, J.R. Mihelcic, A. Pramanik, L. Raskin, M.C.M. Van
Loosdrecht, D. Yeh, N.G. Love, A new planning and design paradigm to achieve
sustainable resource recovery from wastewater, Environ. Sci. Technol. 43 (2009)
61266130.
[2] J.W. Erisman, M.A. Sutton, J. Galloway, Z. Klimont, W. Winiwarter, How a century
of ammonia synthesis has changed the world, Nat. Geosci. 1 (2008) 636639.
[3] D.A. Vaccari, Phosphorus a looming crisis, Sci. Am. 300 (2009) 5459.
[4] M. Tejada, J.L. Gonzalez, A.M. Garca-Martnez, J. Parrado, Effects of different green
manures on soil biological properties and maize yield, Bioresour. Technol. 99
(2008) 17581767.
[5] T. Karak, P. Bhattacharyya, Human urine as a source of alternative natural fertiliser
in agriculture: a ight of fancy or an achievable reality, Resour. Conserv. Recycl. 55
(2011) 400408.
[6] S.K. Pradhan, J.K. Holopainen, H. Heinonen-Tanski, Stored human urine supplemented with wood ash as fertilizer in tomato (Solanum lycopersicum) cultivation
and its impacts on fruit yield and quality, J. Agric. Food Chem. 57 (2009) 7612
7617.
[7] P. Tidaker, B. Mattsson, H. Jonsson, Environmental impact of wheat production
using human urine and mineral fertilizers a scenario study, J. Cleaner Prod. 15
(2007) 5262.
[8] R. Franceys, J. Pickford, R. Reed, A Guide to the Development of On-site Sanitation,
WHO, Geneva, 1992.
[9] S.A. Esrey, J. Gough, D. Rapaport, R. Sawyer, M. Simpson-Herbert, J. Vargas, U.
Winblad, Ecological Sanitation, rst ed., Sida, Stockholm, 1998, , ISBN: 91-58676-12-0.
[10] M. Ugurlu, M.H. Karaoglu, Adsorption of ammonium from an aqueous solution by
y ash and sepiolite: isotherm, kinetic and thermodynamic analysis, Microporous
Mesoporous Mater. 139 (2011) 173178.
[11] M. Li, X. Zhu, F. Zhu, G. Ren, G. Cao, L. Song, Application of modied zeolite for
ammonium removal from drinking water, Desalination 271 (2011) 295300.
[12] J.R. Deans, B.G. Dixon, Uptake of Pb2+ and Cu2+ by novel biopolymers, Water Res.
26 (1992) 469472.
[13] O.S. Amuda, A.O. Ibrahim, Industrial wastewater treatment using natural material
as adsorbent, Afr. J. Biotechnol. 5 (2006) 14831487.
[14] S. Babel, T.A. Kurniawan, Cr(VI) removal from synthetic wastewater using coconut
shell charcoal and commercial activated carbon modied with oxidizing agent
and/or chitosan, Chemosphere 54 (2004) 951967.
[15] S.J. Gregg, K.S.W. Sing, Adsorption, Surface Area and Porosity, Academic Press,
New York, 1982.

55

[16] K. Yang, J. Peng, C. Srinivasakannan, L. Zhang, H. Xia, X. Duan, Preparation of high

surface area activated carbon from coconut shells using microwave heating,
Bioresour. Technol. 101 (2010) 61636169.
[17] T.K. With, T.D. Petersen, B. Petersen, A simple spectrophotometric method for the
determination of urea in blood and urine, J. Clin. Pathol. 14 (1961) 202204.
[18] D.C. Montgomery, Design and Analysis of Experiments: Response Surface Method
and Designs, fourth ed., John Wiley and Sons, New York, 2005.
[19] J. Gan, W. Yuan, Operating condition optimization of corncob hydrothermal
conversion for bio-oil production, Appl. Energy 103 (2013) 350357.
[20] G.E.P. Box, W.G. Hunter, J.S. Hunter, Statistics for Experimenters, John Wiley, New
York, 1978.
[21] J. Coates, Interpretation of Infrared Spectra A Practical Approach, John Wiley &
Sons, Chichester, 2000.
[22] D. Mohan, K.P. Singh, Single and multi-component adsorption of cadmium and
zinc using activated carbon derived from bagasse an agricultural waste, Water
Res. 36 (2002) 23042318.
[23] P. Miretzky, C. Munoz, A. Carrillo-Chavez, Fluoride removal from aqueous solution
by Ca-pretreated macrophyte biomass, Environ. Chem. 5 (2008) 6872.
[24] S. Pradhan, S.S. Shukla, K.L. Dorris, Removal of nickel from aqueous solutions
using crab shells, J. Hazard. Mater. 125 (2005) 201204.
[25] M. Jansson-Charrier, E. Guibal, J. Roussy, B. Delanghe, P. Le Cloirec, Vanadium(IV) sorption by chitosan: kinetics and equilibrium, Water Res. 30 (1996) 465
475.
[26] D.N. Jadhav, A.K. Vanjara, Adsorption kinetics study: removal of dyestuff efuent
using sawdust, polymerized sawdust and sawdust carbon-II, Indian J. Chem.
Technol. 11 (2004) 4250.
[27] I. Langmuir, The adsorption of gases on plane surfaces of glass, mica and platinum,
J. Am. Chem. Soc. 40 (1918) 13611403.
[28] H.M.F. Freundlich, Over the adsorption in solution, J. Phys. Chem. 57 (1906) 385
470.
[29] M.I. Tempkin, V. Pyzhev, Kinetics of ammonia synthesis on promoted iron
catalyst, Acta Phys. Chim. USSR 12 (1940) 327356.
[30] M.M. Dubinin, L.V. Radushkevich, The equation of the characteristic curve of the
activated charcoal, Proc. Acad. Sci. USSR Phys. Chem. Sect. 55 (1947) 331337.
[31] R.K. Singh, S. Kumar, S. Kumar, A. Kumar, Development of parthenium based
activated carbon and its utilization for adsorptive removal of p-cresol from
aqueous solution, J. Hazard. Mater. 155 (2008) 523535.
[32] T.W. Webber, R.K. Chakkravorti, Pore and solid diffusion models for xed-bed
adsorbers, AIChE J. 20 (1974) 228238.
[33] K. Vijayaraghavan, T.V.N. Padmesh, K. Palanivelu, M. Velan, Biosorption of nickel
(II) ions onto Sargassum wightii: application of two-parameter and three parameter isotherm models, J. Hazard. Mater. 133 (2006) 304308.
[34] M.M. Dubinin, The potential theory of adsorption of gases and vapors for
adsorbents with energetically non-uniform surface, Chem. Rev. 60 (1960)
235266.
[35] S. Langergren, B.K. Svenska, Zur theorie der sogenannten adsorption geloster
stoffe, K. Svenska Vetenskapsakad. Handl. 24 (1898) 139.
[36] W.J. Weber, J.C. Morris, Kinetics of adsorption on carbon from solution, J. Sanit.
Eng. Div. 89 (1963) 3160.
[37] B.K. Korbahti, M.A. Rauf, Application of response surface analysis to the photolytic
degradation of basic red 2 dye, Chem. Eng. J. 138 (2008) 166171.