Journal of Cleaner Production 148 (2017) 958e968

Contents lists available at ScienceDirect

Journal of Cleaner Production

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

Equilibrium, kinetic, and thermodynamic studies of lead ion and zinc

ion adsorption from aqueous solution onto activated carbon prepared
from palm oil mill effluent
Ganiyu Abimbola Adebisi a, Zaira Zaman Chowdhury a, ** , Peter Adeniyi Alaba b, *
a
Nanotechnology & Catalysis Research centre (NANOCAT), University of Malaya, IPS Building, 50603 Kuala Lumpur, Malaysia
b
Department of Chemical Engineering, University of Malaya, 50603 Kuala Lumpur, Malaysia

a r t i c l e i n f o a b s t r a c t

Article history: An efficient activated carbon was prepared using palm oil mill effluent as a precursor. The adsorption
Received 19 November 2016 capacity of activated carbon for lead ion and zinc ion from aqueous media was investigated under
Received in revised form equilibrium conditions between 303.15 and 353.15 K. The activated carbon was analyzed using fourier
6 February 2017
transform infrared spectroscopy, field emission scanning electron microscope, energy dispersive X-Ray,
Accepted 6 February 2017
Available online 8 February 2017
and Nitrogen adsorption-desorption analysis. The adsorption capacity of activated carbon was studied by
varying adsorbent dosage, contact time, and temperature. The equilibrium time was attained after
50 min for both ions. The data analysis was performed with different isotherm and kinetic models.
Keywords:
Activated carbon
Pseudo-second-order kinetic rather than pseudo first-order model is best fitted for both lead ion and zinc
Adsorption ion removal from wastewater. Further, the rate-determining step for both metal ions is chemisorption
POME based on the suitability of the Elovich equation. Langmuir model provides the best fit for both removal of
Equilibrium lead ion and zinc ion. The thermodynamic parameters shows the feasibility of adsorption of both ions is
Isotherm endothermic.
Thermodynamic © 2017 Elsevier Ltd. All rights reserved.

1. Introduction cost-effective, simple in design, easy to operate and much more

environmentally benign (Naje et al., 2016). In view of this, sub-
Heavy metals are discharged into water bodies through various stantial attention has been given to development of cheap adsor-
industrial activities especially in the manufacturing and mining bents such as kaolin (Alaba et al., 2016a), bentonite, fly ash blast
industries. They are highly toxic and usually constitute health furnace slag and activated carbon for removal of the different heavy
hazards when consumed beyond the permissible and bearable metals at low concentration (Mishra and Patel, 2009).
amount. Heavy metal pollutants have been removed from waste- Palm oil is one of the most significant vegetable oils globally
waters using various technologies ranging from precipitation, with Malaysia being the largest producer and exporter (Sani et al.,
reverse osmosis, membrane filtration and ion exchange (Baccar 2015). During palm oil processing, several products and by-
et al., 2009). However, these technologies have several disadvan- products generated include palm oil, palm kernel, palm oil mill
tages such as requirement of expensive equipment and the need of effluent (POME), shell, empty fruit bunch (EFB) and fiber (Sani et al.,
chemicals, which may in turn pollute the water. Due to these lim- 2015). The palm oil mill industry is said to be the major source of
itations, there is vital need for a more environmentally benign and industrial pollution in Malaysia (Phang, 1990). They discharges
cost effective method. Several efforts were geared towards the use more than 15million ton of POME from more than 110 processing
of adsorption process. The use of adsorbents in heavy metal mill (Board, 2003) into the environment, thereby constituting a
removal from wastewaters has not only been found to be superior serious environmental pollution especially in the waterways
to other conventional methods, but it has equally been found to be (Phang and Kim-Chong, 1988). The discharge of POME, either un-
treated or poorly treated, into the environment can lead to serious
hazards ranging from pollution of groundwater, release of methane
* Corresponding author. into the atmosphere to release of unpleasant odor to the environ-
** Corresponding author. ment (Chavalparit, 2006). Several methods have been adopted to
E-mail addresses: adebisi@siswa.um.edu.my (G.A. Adebisi), dr.zaira.
treat POME before the discharge including sedimentation, filtration
chowdhury@um.edu.my (Z.Z. Chowdhury), adeniyipee@live.com (P.A. Alaba).

http://dx.doi.org/10.1016/j.jclepro.2017.02.047
0959-6526/© 2017 Elsevier Ltd. All rights reserved.
 G.A. Adebisi et al. / Journal of Cleaner Production 148 (2017) 958e968 959

and decolourization (Igwe and Onyegbado, 2007). Other techniques activated carbon were studied. For each of the variables, the sample
involve ion exchange, adsorption, reverse osmosis, selective crys- was fed into the urns of about 150 mm length and 40 mm diameter.
tallization, and coagulation processes. Interestingly, POME has been The urns was then transferred into the calcination furnace (Bam-
reported to be rich in protein, nitrogenous compounds, carbohy- stead/Thermolyne 48000) fitted with a temperature knob to
drate, lipids and minerals (Habib et al., 1997). POME equally con- regulate the temperature. The urns was equally connected to ni-
tains lignocellulosic wastes as well as mixture of oil and trogen gas cylinder while the sample pyrolysis was carried out
carbohydrates (Oswal et al., 2002). This makes this renewable, using a continuous nitrogen gas flow of 100 mL min-1 at varying
readily available and abundant agricultural waste a viable alterna- temperature between 500 and 800
C. After the regulated holding
tive precursor for activated carbon synthesis. Activated carbons time, the samples were cooled to room temperature under inert
have been in use as adsorbents for quite a while. Nonetheless, the atmosphere, and subsequently raked from the vessel. The sample
commercial activated carbons are relatively expensive; hence, the was then filtered with distilled water to eliminate residual chemical
need for utilization of lignocellulosic wastes for the synthesis of agents until the pH became neutral and then dried at 65
C for 12 h.
activated carbons (Nacu et al., 2015). A few of the agricultural
wastes that have been used as a precursor for synthesis of activated 2.2.1. Impregnation of the sample with H3PO4
carbon adsorbents include solid wastes (Alaba et al., 2015), coconut The sample was activated with H3PO4 and the influence of
shell (Amuda et al., 2007), kenaf core (Chowdhury et al., 2012a), H3PO4 concentration on the yield and porous characteristics of
bamboo culms (Ekebafe et al., 2012), oil palm empty fruit bunch POMEAC were studied. 20 g of the sample was impregnated with
(Foo and Hameed, 2011), coconut shell (Gratuito et al., 2008), rattan H3PO4 in varying impregnation ratio ranging from 30% to 50%.
sawdust (Hameed et al., 2007) and a host of others. However, the
quality (characteristics) and quantity (yield) of activated carbon 2.2.2. Carbonization of the impregnated samples
that can be produced from a particular raw material depend on two The samples impregnated with H3PO4 were carbonized and the
main factors, that is, the nature of the precursor itself and the influence of carbonization temperature on the porosity and yield of
successful manipulation of the process conditions as it will require activated carbon during activation process was observed. This was
an optimum conditions to obtain the required characteristics of the investigated by varying the carbonization temperature from 500 to
activated carbon (Gratuito et al., 2008). Considering the nature of 800
C for the chemically activated sample. The holding time was
the precursor, a precursor with high cellulose content will lead to also varied from 1 to 3 h.
production of a microporous activated carbon, while one with a
high lignin content will give activated carbon that will be highly 2.3. Physico-chemical characteristics of POMEAC
mesoporous. To obtain a balanced process conditions will require
strict monitoring of such variables as carbonization temperature, The functional groups and structure of the prepared POMEAC
holding time and chemical impregnation ratio. were obtained using the FTIR spectroscopy. The FTIR analysis of
This study is aimed to evaluate the efficiency of POME activated both the raw POME and POMEAC were obtained from 400 to
carbon (POMEAC) for removal of Pb (II) and Zn (II) ions from 4000 cm1 using FTIR- 2,000, Perkin Elmerc.
wastewater for the first time. The material was characterized using The surface morphology of both the raw POME and POMEAC
various microscopic and spectroscopic examinations and its per- were analyzed with field electron scanning electron microscope
formance was evaluated for Pb (II) and Zn (II) uptake from waste- (FESEM) (Zeiss SUPRA 35VP; Germany) coupled with energy
water. The experimental variables of influence on the sorption dispersive x-ray spectrometer for elemental analysis.
experiment like contact time, temperature and initial metal ions
concentration were investigated. The analysis of equilibrium
2.4. Sorption of Pb (II) and Zn (II) ions on the adsorbent
adsorption of the experimental data was performed by kinetic
models, and various isotherms. The thermodynamic variables (DH
,
The sorption of the adsorbates onto the adsorbents was studied
DS
and DG
) for sorption experiments were also determined. through batch adsorption studies. Variation of uptake level of Pb(II)
and Zn(II) ions with contact time was determined for each of the
2. Methodology
adsorbent. This was required in other to establish the equilibrium
time (time for the adsorbent to reach sorption equilibrium). To
2.1. Collection and treatment of palm oil mill effluent (POME)
accomplish this, 0.2 g each of the adsorbent were dispersed in
sample
50 cm3 of each of the initial concentrations of 150, 200, 250, 300
and 350 mg L1 solutions at a time range of 10e360 min in well
The POME sludge was obtained from a local palm oil mill plant
stoppered plastic containers. The minimum amount of adsorbent
in Osogbo, Osun state, Nigeria. The point of collection was situated
was used to examine the possibility of achieving a remarkable
at a site before the discharge of the POME into the treatment pond.
adsorption performance by varying other operating parameters.
After collection, the sample was thoroughly filtered with distilled
The mixture was then thoroughly agitated using a water bath
water to eliminate dirt such as mud and other impurities. Further,
shaker set at 150 rpm and at room temperature (30
C). Subse-
the sample was sun-dried in a wide-mouthed plastic container for
quently, the mixture was filtered and the residual concentration of
2 months after which it was preserved in a desiccator to avoid
the Pb(II) and Zn(II) ions (as the case may be) in the filtrate was
possible bio-degradation. Further, the sample was filtered with
determined by ICP (APHA 3125B, 20th edition, 1998). The quantity
distil-water, and dried in oven at 60
C for 24 h. The resulting
of Pb(II) or Zn(II) removed at equilibrium (qe) was estimated from
samples were pulverized to <1 mm mesh size with the aid of a jaw
the following equation.
crusher (Denver Product, Dagenham, U.K).
ðCo þ Ce ÞV
2.2. Preparation of activated carbon qe ¼ (1)
W

Towards synthesis of activated carbon, the effects of the where,qe (mg/g) ¼ quantity of adsorbate (Pb(II) or Zn(II) ions)
different process variables such as holding time, carbonization adsorbed at equilibrium, Co (mg/L) ¼ initial concentration of
temperature and H3PO4 concentration on the yield and porosity of adsorbate,Ce (mg/L) ¼ final concentration of adsorbate after
960 G.A. Adebisi et al. / Journal of Cleaner Production 148 (2017) 958e968

adsorption at equilibrium.V (L) ¼ Volume of solution.W (g) ¼ Mass H3PO4 as the activating agent using 25% acid concentration for 2 h.
of dry adsorbent. It is evident from the figure that the yield (%) decreased with in-
crease in carbonization temperature. This is as a result of such
3. Results and discussion losses in form of carbon burn-off and other volatiles at higher
temperatures (Mi et al., 2015). Increase in temperature from 500
C
3.1. Effects of process variables on the yield of activated carbon to 800
C leads to decrease in the yield of activated carbon from
43.76% to 25.59%, indicating that the optimum carbonization
The effects of the process variables on the quality (porous temperature is 500
C.
characteristics) and quantity (yield) on activated carbon prepared
from a particular precursor cannot be underestimated as the yield 3.1.2. Holding time
and development of porosity depends largely on such variables as In addition to other process variables, the holding time plays
carbonization temperature, holding time and impregnation ratio. important role in determining the yield (%) and porous character-
istics of activated carbon prepared from a precursor. In this case, by
keeping the carbonization temperature and impregnation ratio
3.1.1. Carbonization temperature
constant at 650
C and 37.50%, the holding time was varied be-
Fig. 1a below elucidates the influence of carbonization temper-
tween 1 he3 h.
ature on the preparation of activated carbon from POME using
Fig. 1b below shows the effect of holding time on the POME
activated carbon (PAC) preparation. It could be perceived from the
figure that the yield of POMEAC initially increases as the holding
45 time increases from 1 to 2 h, indicating that the optimum holding
time is 2 h. However, further increase in the holding time decreases
40 the yield. This is mainly due to the severity of carbon burn-off with
increase in holding time. Similar result was observed by (Mi et al.,
Yield, %

2015).
35

3.1.3. Acid concentration

30 The concentration of the impregnating agent (H3PO4) used for
the synthesis was varied from 25% to 50%. The yield of carbon from
25 a precursor depends largely on the amount of carbon released to
500 550 600 650 700 750 800 bind with oxygen and hydrogen atoms during transformation from
Temperature, °C lignocellulosic materials. The conversion from lignocellulosic
wastes into carbon usually involves the release of O and H atoms in
40 form of aldehydes, CO2, CO, CH4 and H2O (Tseng, 2007).
Fig. 1c below indicates the effect, which the impregnation ratio
as well as the impregnating agent has on the POME activated car-
bon (PAC) yield (%) at 650
C for 1 h. From the figure, it is vivid that
the yield of POMEAC decreased from 38.14 to 37.60% as the con-
centration of H3PO4 increased from 25% to 50%, indicating that the
Yield, %

35 optimum acid concentration is 25%. This is traceable to severe

leaching carbon atoms released in form of CO2, CO, CH4 and
aldehydes.

3.2. Characterization

30
3.2.1. Porosity analysis
1 1.5 2 2.5 3
Pore diameter, pore volume and Surface area of the POME and
Time, h
POMEAC obtained at optimum conditions were determined by
using a Micromeritics BET instrument using ultra-pure nitrogen
40 adsorption at 196
C. Average values of these properties were
given in Table 1 below.
Fig. 2 illustrates N2 adsorption-desorption isotherms
(at 196
C) of POME and POMEAC. Both isotherms are type IV
according to the IUPAC classification (Sing, 1985), which is char-
Yield, %

acterized by hysteresis loop. The high N2 adsorbed amount on

POMEAC indicates its high porosity. This is traceable to removal of
inorganic and water-soluble organic matters, which opens the pore
structure and enhances the porosity (Alaba et al., 2016b).

3.2.2. EDX analysis

35 The elemental composition of the samples were analyzed using
25 30 35 40 45 50
energy dispersive analysis system of X-ray (EDX) (EDX, Ltd., USA).
Concentration, %
EDX analysis was conducted to evaluate the adsorption of cadmium
on CCB beads. The spectrum of POMEAC shows the presence of C, O,
Fig. 1. Effect of (a) carbonization temperature, (b) holding time (c) acid concentration Al, Si, K, P, Fe and Ca in the structure (Fig. 3). Table 1 presents the
on the yield of POMEAC. elemental percentage composition of the samples. The effect of
 G.A. Adebisi et al. / Journal of Cleaner Production 148 (2017) 958e968 961

Table 1 conjugated with carbon-carbon double bounds (C]CeOeC) in

Textural properties and elemental composition of POME and POMEAC. olefinic or aromatic structures. This indicates that POMEAC is rich
Properties POME POMEAC in basic surface functional group, which are suitable for adsorption
SBET (m2/g) 1.8056 59.1989
of hydrophobic pollutants.
Smicro (m2/g) 0.3546 38.0977
Vmicro (cm3/g) 0.0001 0.0173 3.2.4. FESEM
Vtotal (cm3/g) 0.0060 0.0348
Dp (nm) 13.24 2.36
The FESEM analysis was performed using software controlled
EDX (Weight%) digital FESEM equipment. Fig. 5 presents the result of the micro-
CK 68.19 47.07 graph for the raw material and product. The FESEM of POME shows
OK 30.45 38.02 that it is amorphous. Further, the micrograph illustrates the
Al K 0.28 0.57
porosity of POMEAC with various sizes of pore openings on the
Si K 0.56 1.43
KK 0.25 0.65 surface, which increase the contact area, and facilitates mass
Ca K 0.25 0.34 transfer the pore during adsorption.
Fe K 0.02 0.03
PK 0 11.89
3.3. Batch adsorption study

From economic point of view, it is essential to study the influ-

H3PO4 activation on POME is evident by the presence of phos-
ence of adsorbent dosage to determine the minimum weight of
phorus (P) and reduction in the percentage composition of carbon
adsorbent required for effective metal ion removal. Fig. 6 (a) pre-
due to carbon burn-off.
sents the influence of adsorbent dosage on sorption of Pb(II) and
Zn(II). It is evident that the removal of both metal ion increases
3.2.3. FTIR with increase in the quantity of adsorbent used. Moreover, increase
The FTIR spectra of both the sample (POME) and the prepared in the adsorbent dosage also favors the equilibrium adsorption
POME activated carbon (POMEAC) are shown in Fig. 4 below. The capacity (mg g1) (Fig. 6 (b)). However, economic viability of the
POME spectra show different peaks corresponding to different process is of vital importance. We used 0.2 g to investigate the
functional groups. The spectra peaks at 3860.02 and 3433.06 cm1 feasibility of better adsorption performance by varying other pa-
correspond to eNH associated with amide in bonded form and rameters. This is in agreement with the submissions of the work of
eOH functional groups of phenol. The peaks at 2923.59 and (Lupea et al., 2012).
2854.71 cm1 designate the presence of eCeH stretching in methyl
and methylene groups asymmetric and symmetric stretching. 3.4. Adsorption kinetics
Other functional groups that are present, according to the spectra
peaks are the ether groups (R-OR-) at 1741.53 cm1, carbonyl groups To determine the adsorption equilibrium time for the two pol-
(C]O) at 1641.12 cm1, esters group (R-CeO-R-) at 1382.71 cm1, lutants, batch adsorption studies were carried out at varying times
eCeOe stretching at 1170.88 cm1 and group eC]O groups at ranging from 10 to 360 min, at 30
C and pH 5.5 as used by
1027.87 cm1 spectral peak. Chowdhury et al. (2012b,c). The choice of pH value is also based on
The activation of POME to POMEAC leads to decrease in in- the understanding from the study of Chigondo et al. (2013) on the
tensities of the above peaks except the peak at 1027.87 cm1, which adsorption of Pb(II) and Zn(II) ions onto POMEAC at T ¼ 30
C,
increases and shifts to 1037.52 cm1 due to in-plane deformation of t ¼ 210 min, adsorbate initial concentration ¼ 350 mg L1 and
aromatic eCH. Other bands above 600, which are ascribed to al- adsorbent dosage ¼ 0.2 shows that adsorption of the metal ions
kenes and aromatics out of plane bend with acyclic CeOeC groups rose steadily from 2 to 6 and then drops again steadily. At pH

25
Quantity Adsorbed (cm³/g STP)

20

15

10 POMEAC

POME
5

0
0 0.2 0.4 0.6 0.8 1
Relative Pressure (p/p°)
Fig. 2. Nitrogen adsorption-desorption isotherms of POME and POMEAC.
962 G.A. Adebisi et al. / Journal of Cleaner Production 148 (2017) 958e968

POME

POMEAC

Fig. 3. Energy dispersive X-ray (EDX) analysis of POME and POMEAC.

0.92
Transmittance

0.84

0.76

POMEAC
0.68
POME

0.6
4,000 3,500 3,000 2,500 2,000 1,500 1,000 500
Wavenumber, cm-1

Fig. 4. FTIR analysis of POME and POMEAC.

between 5 and 6, more negatively charged surface are available adsorption (Nguyen et al., 2013). Fig. 7 presents the results of batch
which no doubt contributed to improved metal ion adsorption. At adsorption carried out at these varying time. From the figures, it
lower pH values, an adsorbent surface is usually surrounded by was observed that the sorption level of POMEAC increased pro-
hydroxonium ions (Hþ) thereby blocking metal ions from normal gressively as the sorption time increases. The equilibrium time was
binding to adsorbents (Onundi et al., 2010). Equally, at higher pH attained after 50 min for both metal ions. This suggests that all the
values of 8 and above, the OHe present in the solution precipitates, available sorption sites have been occupied by the cation under
thereby blocking available sorption sites and resulting into reduced investigation. This is similar in trends to what had already been
 G.A. Adebisi et al. / Journal of Cleaner Production 148 (2017) 958e968 963

Fig. 5. FESEM images of POME and POMEAC.

80
(a)

60

Equilibrium Adsorption 40
qt (mg/g)

20 150mg/L
200mg/L
250mg/L
300mg/L
350mg/L
0

0 100 200 300 400

Time (min)

60
(b)
50
Equilibrium Adsorption

40
qt (mg/g)

30

20
150mg/L
200mg/L
10 250mg/L
300mg/L
350mg/L
Fig. 6. Effect of adsorbent dosage on (a) percentage metal ion removal, and (b) equi- 0
librium adsorption capacity.

0 100 200 300 400

observed by Lupea et al. (2012).
Time (min)
Quite a few kinetic models were used to determine the rate
controlling step and study the sorption mechanism of Pb (II) and Zn Fig. 7. Effect of initial concentration (a) Pb2þ and (b) Zn2þ ion and contact time (min):
(II) onto POMEAC. Intra-particle diffusion model was used to at temperature 30
C and pH 5.5.
compute the rate controlling step, while pseudo first-order, pseudo
second-order kinetic (Kumar et al., 2012), and Elovich models were
used to compute the rate constant. The prospect of the intra- adsorbent at time t, kid (mg g1 min1/2) is the intra-particle diffu-
particle diffusion was utilized as follows (Weber and Morris, 1964): sion rate constant, and C (mg g1) is the constant, which is pro-
portional to the boundary layer thickness; that is the value of C
qt ¼ kid t 1=2 þ C (2) increases with increase in the boundary layer thickness (Mckay
et al., 1980).
where qt (mg g1) is the amount of adsorbate adsorbed on the The diffusion controlled sorption process is presented by the
964 G.A. Adebisi et al. / Journal of Cleaner Production 148 (2017) 958e968

Table 2
Kinetic model parameters for the adsorption of Pb(II) onto POMEAC.

C0 (mg/L) qe, exp (mg/g) Pseudo-first-order Pseudo-second-order Elovich model Intraparticle diffusion

qe, cal (mg/g) K1 (g/mg h) R2 qe, cal (mg/g) K2 (g/mg h) R2 (1/b)ln(ab) (mg/g) 1/b (mg/g) R2 Kid C R2

150 34.72 0.83 0.0030 0.0006 46.51 0.0008 0.8736 34.70 0.07 0.0073 0.0126 34.53 0.4600
200 44.98 1.87 0.0161 0.8751 71.43 0.0002 0.7701 38.40 1.19 0.9025 0.2612 40.91 0.8737
250 52.30 1.86 0.0048 0.8874 67.57 0.0007 0.9324 47.34 1.05 0.8964 0.2427 49.46 0.9548
300 56.44 3.25 0.0094 0.8373 86.96 0.0003 0.8472 37.29 3.96 0.9443 0.8948 45.43 0.9646
350 62.19 5.85 0.0078 0.9878 116.28 0.0008 0.7625 19.99 14.99 0.9212 3.4284 10.43 0.9638

straight line graph of qt versus t1/2 gives a slope of kid and intercept
of C. Tables 2 and 3 gives the values of kid and C for Pb (II) and Zn (II). 1 k 1
¼ 1 þ (6)
If the plot passes through the origin, then intraparticle diffusion qt qe t qe
remains the only rate-controlling step, else it is not the only rate
determining factor due to the effect of boundary layer diffusion on where qe represents the quantity adsorbed at equilibrium, k1
adsorption. For the removal of both Pb(II) and Zn(II) using POMEAC, (min1) denotes the pseudo-first-order adsorption rate constant,
the linear curves deviated from the origin due to the variation in the and qt symbolizes the quantity absorbed at time t (min), both in
diffusion rate in the initial and final stages of the sorption. Since the mmol g1. The linear plot of ln 1/qt vs. 1/t gives the value of k1 and qe
plots did not pass through origin, then the intraparticle diffusion is (Table 1).
not the only factor that determines the rate for the two metal ions. The pseudo-second-order kinetic model is given by
Elovich model was used to describe sorption on the surface of an
t 1 1
adsorbent without product desorption. The model is considered ¼ þ t (7)
qt k2 q2e t qe
suitable for some chemisorption processes, which involves valence
forces via exchange or sharing of between the adsorbate and the
where k2 is the pseudo-second-order adsorption rate constant in g
adsorbent. The removal rate decreases with time because of the
mmol1 min1, qt and qe are the quantity of each heavy metal ion
coverage of the active sites (Aharoni and Tompkins, 1970). Elovich
adsorbed at time t (min) and equilibrium, both in mmol g1. The
model is given by
linear plot of ln t/qt vs. t gives the value of k1 and qt (Tables 2 and 3).
The validity of the model was checked using the values of R2).
dqt The values of qe experimental and qe calculated differs appre-
¼ a expð  bqt Þ (3)
dt ciably in pseudo-first-order kinetic model, while the values are
closer for pseudo-second-order kinetic model at almost all the
where a and b are the experimental constants and a is the initial
initial concentrations for both Pb (II) and Zn (II). Further, the R2
rate as dqt/dt approaches a and qt approaches 0.
values for pseudo-first-order model were less than that of the
Given that qt ¼ qt at t ¼ t and qt ¼ 0 at t ¼ 0, the integrated form
pseudo-second-order model for both ions signifying that the
of the above equation is
pseudo-second-order is better correlated than the pseudo-first-
    order model. The same kinetic behavior was reported for the
1 1
qt ¼ lnðt þ t0 Þ  ln t0 (4) adsorption of Pb (II) and Zn (II) onto various adsorbents, such as,
b b activated carbon (Mishra and Patel, 2009); . Maghemite nanotubes
(Roy and Bhattacharya, 2012), EDCMS (Ren et al., 2013), MNPs-NH2
where t0 ¼ 1/ab. If t » t0, the equation can be further simplified as (Tan et al., 2012), tetrasulfide-functionalized silica (Fan et al., 2012),
    magnetic Co0.6Fe2.4O4 (Duan et al., 2015), Fe3O4-TETA-CMCS
1 1 (Dabrowski et al., 2004).
qt ¼ lnðabÞ þ ln t (5)
b b
The intercept and slope of the linear plot of qt vs. ln t gives (1/b) 3.5. Adsorption isotherm modeling
ln(ab) and (1/b). Table 1 shows the parameter values as well as the
correlation coefficient (R2), which shows the uniformity of the Figs. 8 and 9 presents the adsorption isotherms of Pb(II) and
predicted values with the experimental data. The model shows Zn(II). Isotherms provide the equilibrium relations between the
good correlation for sorption of both Pb (II) and Zn (II) on POMEAC adsorbate concentration on the solid and the liquid phase. The
especially at initial concentration range of 200e350 mg L1 for Pb maximum adsorption capacity is obtainable from the isotherms
(II) and 250e300 mg L1 for Zn (II). This indicates that Chemi- (Ayanda et al., 2013). This provide insight on the adsorbent capacity
sorption may be the rate-determining step. or the quantity of the adsorbent needed to remove a unit mass of
Pseudo-first-order kinetic model is given by metal ion from aqueous medium under the process conditions.

Table 3
Kinetic model parameters for the adsorption of Zn(II) onto POMEAC.

C0 (mg/L) qe, exp (mg/g) Pseudo-first-order Pseudo-second-order Elovich model Intraparticle diffusion

qe, cal (mg/g) K1 (g/mg h) R2 qe, cal (mg/g) K2 (g/mg h) R2 (1/b)ln(ab) (mg/g) 1/b (mg/g) R2 Kid C R2

150 29.46 0.87 0.0147 0.8567 29.50 0.1223 1.0000 28.89 0.11 0.8642 0.0231 29.15 0.7348
200 39.19 1.80 0.0170 0.8652 39.68 0.0086 0.9998 30.65 1.64 0.8332 0.3214 34.51 0.6399
250 43.81 2.77 0.0064 0.9757 42.37 0.0024 0.9979 24.57 2.81 0.9430 0.6432 30.29 0.9855
300 49.63 2.44 0.0071 0.8414 68.49 0.0011 0.5951 36.37 2.58 0.9760 0.5730 41.78 0.9610
350 49.98 2.17 0.0030 0.8085 53.19 0.0036 0.9980 44.38 1.21 0.8579 0.3035 46.60 0.8207
 G.A. Adebisi et al. / Journal of Cleaner Production 148 (2017) 958e968 965

80
(a) 2

60 (b)
1.6
qe (mg/g)

40 1.2

Ce/ qe (L/g)
Experimental
0.8
Langmuir
20
Freundlich
0.4
0
0 20 40 60 80 100 0
Ce (mg/L) 0 50 100
Ce (mg/L)

1.9
(c)
-2.3
1.8
0.0028 0.003 0.0032 0.0034
-2.4
(d)
1.7 -2.5
Log qe

ln KL

-2.6
1.6
y = 754.41x - 4.9063
-2.7 R² = 0.7947

1.5 -2.8
1 1.2 1.4 1.6 1.8 2 2.2
Log Ce 1/T
Fig. 8. Adsorption isotherm data for Pb(II) onto POMEAC at 303 K: (a) theoretical and experimental data at pH 5.5, (b) a plot of Ce/qe vs. Ce, (c) linear form of Freundlich, (d) a plot of
lnKL vs. 1/T using Langmuir model.

Langmuir and Freundlich isotherms are the most commonly characteristics of Langmuir isotherm by a modeling a dimension-
used isotherms to describe the solideliquid adsorption system. The less constant known as the separation factor or equilibrium
Langmuir isotherms assumes that sorption occurs at a specific parameter, RL, which shows the nature of the adsorption process.
homogeneous site on the surface of POMEAC, and that not more
than one adsorbate molecule can occupy a site. Langmuir adsorp-
tion model is given by 1
RL ¼ (9)
1 þ KL C0

Ce 1 Ce where C0 (mg L1) is the initial adsorbate quantity and KL is the

¼ þ (8)
qe kL qm qm Langmuir sorption constant (L mg1) described above. Adsorption
is favorable when the value of RL is between 0 and 1. The RL
where qe is the quantity of solute adsorbed at equilibrium in mmol parameter is said to be a more consistent indicator of the adsorp-
g1, Ce is the equilibrium concentration of the adsorbate in aqueous tion nature. The value of RL classifies the isotherm to be either
solution (mmol L1), qm (mmol g1) is the maximum monolayer favorable (0 < RL < 1), linear (RL ¼ 1), unfavorable (RL > 1) or
adsorption capacity, KL is the Langmuir adsorption constant L irreversible (RL ¼ 0). The RL values for both Pb(II) and Zn(II)
mmol1, which indicates the energy of adsorption. The values of KL (Table 5) show that the adsorption process is favorable by POMEAC.
and qm are computed from the linear plot of Ce/qe vs. Ce. (Figs. 8b Moreover, the values of the R2, which determine the uniformity of
and 9b), and the values were displayed in Table 4. the predicted values with the experimental data shows that Lang-
The maximum monolayer sorption capacity (qm) increases with muir model is suitable for both Pb(II) and Zn(II).
increase in temperature from 303.15 to 353.15 K (Table 4). Pb (II) Freundlich isotherm is a special case of Langmuir model. It is an
exhibits 94.34 mg g1 at 353.15 K and 68.49 mg g1 at 353.15 K for empirical model that presents interactions between the adsorbed
Zn (II) indicating that this system is effective and results are com- molecules in a heterogeneous system. The model is applicable to
parable to the previous studies related to heavy metal ions removal. non-ideal adsorption on heterogeneous surfaces and multilayer
Weber and Chakravorti (1974) expressed the essential adsorption (Wang and Qin, 2005), and is given by
966 G.A. Adebisi et al. / Journal of Cleaner Production 148 (2017) 958e968

60
(a) 3.2
2.8 (b)
40 2.4
qe (mg/g)

Ce/ qe (L/g)
Experimental 1.6
20 Langmuir 1.2
Freundlich Langmuir
0.8
Experimental
0.4
0
0
25 75 125
25 75 125
Ce (mg/L) Ce (mg/L)

1.8
(c)

1.7 -2.6
0.0028 0.003 0.0032 0.0034
-2.8 (d)
1.6
Log qe

ln KL

-3
1.5
y = -941.8x - 0.0768
-3.2 R² = 0.9892
1.4
1.4 1.6 1.8 2 2.2 -3.4
Log Ce 1/T

Fig. 9. Adsorption isotherm data for Zn(II) onto POMEAC at 303 K: (a) theoretical and experimental data at pH 5.5, (b) a plot of Ce/qe vs. Ce, (c) linear form of Freundlich, (d) a plot of
lnKL vs. 1/T using Langmuir model.

Table 4 Table 5
Langmuir isotherm for sorption of Pb(II) and Zn(II) onto POMEAC at different Separation factor, RL values for the Pb(II) and Zn(II) adsorption.
temperatures.
C0 (mg/L) Lead Zinc
Temp Pb(II) Zn(II)
RL RL
qm KL R2 qm KL R2
150 0.0672 0.1397
K (mg/g) (L/mg) (mg/g) (L/mg) 200 0.0513 0.1086
250 0.0414 0.0888
303.15 69.4444 0.0925 0.9961 59.8802 0.0410 0.9811
300 0.0348 0.0751
333.15 90.0901 0.0642 0.9272 67.5676 0.0563 0.9728
350 0.0299 0.0651
353.15 94.3396 0.0669 0.9519 68.4932 0.0633 0.9686

Table 6
1 Freundlich isotherm for sorption of Pb(II) and Zn(II) onto POMEAC at different
ln qe ¼ ln kf þ ln Ce (10) temperatures.
n
Temp K Pb(II) Zn(II)
where qe is the equilibrium concentration of the solute adsorbed on 2
KF (mg/g) n R KF (mg/g) n R2
adsorbent (mmol g1), Ce is the solute equilibrium concentration of
the (mmol L1), Kf is the Freundlich constant, which indicates the 303.15 19.1136 3.6860 0.9294 10.9300 3.1017 0.7828
adsorption extent in mmol11/n L1/n g1, and n is the heterogeneity 333.15 15.4808 2.6062 0.9415 12.7764 2.9464 0.7615
353.15 16.0997 2.5536 0.9745 13.3886 2.9412 0.7037
factor, which indicates the removal effectiveness. The slope and
intercept of ln Ce vs. ln qe (Figs. 8c and 9c) gives the Values of n and
Kf for the metal ions adsorption on POMEAC (Table 6). Since the the relationship between the quantity of metal ions adsorbed by
values of R2 for both metal ions isotherms are lower than 0.9 for the POMEAC and its equilibrium concentration in the aqueous
Zn(II), the model cannot adequately give the detail description of medium. For Pb(II), the values of Kf decreases with increase in
 G.A. Adebisi et al. / Journal of Cleaner Production 148 (2017) 958e968 967

Table 7  The activated carbon was characterized by EDX, FTIR and FESEM
Thermodynamic parameters for sorption of Pb(II), and Zn(II). techniques.
Temp (K) Lead Zinc  The adsorption of Pb(II) and Zn(II) increase from 34.72 to
DG
DS
DH
DG
DS
DH
62.19 mg g1 and 29.46e49.98 mg g1 as the initial concen-
tration increases from 150 to 350 mg L1. Increase in the
303.15 5.9987 6.2726 0.0408 8.0480 -7.8301 0.0006
adsorbent dosage from 0.2 to 1 g resulted in gradual increase in
333.15 7.6037 7.9692
353.15 7.9400 8.1052 the equilibrium adsorption capacity (mg g1) of POMEAC from
17.23 to 22.46 fro Pb(II) and 15.08 to 20.45, and increases in the
percent removal efficiency.
 Based on R2 value and difference between the experimental and
temperature from 303.15 to 333 K indicating that adsorption pro-
the calculate values of qe, pseudo second-order kinetic rather
cess is exothermic in nature. However, further increase in tem-
than the pseudo-first-order kinetic model are best fitted for
perature to 353.15 K decreases the Kf value. For Zn(II), the values of
removal of P(II) and Zn(II) from wastewater. In addition, based
Kf increases with temperature increase from 303.15 to 353 K indi-
on the suitability of the Elovich equation for both Pb(II) and
cating that the nature of the adsorption process is endothermic. The
Zn(II) we can suggest that the rate determining step for both
values of n for both metal ions fall between 1 and 10 signifying
metal ions is chemisorption. The intraparticle diffusion model
favorable sorption.
for the both Pb(II) and Zn(II) removal with respect to the
The equilibrium observations for Pb(II) shows that both Lang-
intercept show greater boundary layer effect. The intraparticle
muir and Freundlich model remarkably describe the sorption pro-
diffusion is not the only rate controlling step.
cess. This is inline with the report of Mishra and Patel (2009).
 The equilibrium observations for Pb(II) shows that both Lang-
Meanwhile, the Langmuir isotherm model best describe Zn(II)
muir and Freundlich model remarkably describe the sorption
removal process, which corroborates the report of Roy and
process. Meanwhile, the Langmuir isotherm model best
Bhattacharya (2012).
describe Zn(II) removal process. The maximum monolayer Pb(II)
sorption capacity is 94.34 mg g1 at 353.15 K and 68.49 mg g1
3.6. Thermodynamic analysis
at 353.15 K for Zn(II) indicating that this system is effective and
results are comparable to the previous studies related to heavy
The thermodynamic parameters of the sorption process Gibbs
metal ions removal.
free energy change (DG
), the entropy change (DS
), and enthalpy
 The thermodynamic data show that the removal of Pb(II) and
change (DH
), were obtained from the Langmuir constant (KL)
Zn(II) onto POMEAC from wastewater is feasible and endo-
variation with temperature (T) using the following equations,
thermic in nature.


DG ¼ RT ln KL (11)
Based on the above findings, it is obvious that the synthesized




POMEAC is an effective biosorbent for removal of Pb(II) and Zn(II)
DG ¼ DH  T DS (12) from wastewater. POMEAC could be used as a biosorbent for other
divalent metal ions such as Cd2þ, Cu2þ and Ni2þ.



DS DH
ln KL ¼  (13)
R RT Acknowledgements

where R is the universal gas constant (8.314  103 kJ mol1 K1).

This study was carried out with the aid of a High impact
The enthalpy change due to adsorption of Pb2þ and Zn2þ ions by
research (HIR) grant (H-21001-F0032) under the University of
POMEAC over the temperature range investigated can be obtained
Malaya.
from the linear plots of ln KL against 1/T using the least squares
method (Figs. 8d and 9d). The values of DH
, DS
and DG
for
sorption of Pb(II) and Zn(II) by POMEAC at a temperatures range of References
303.15e353.15 K (Table 7), show that DG
is small but increases Aharoni, C., Tompkins, F., 1970. Kinetics of adsorption and desorption and the
with increase in temperature for Pb(II), while it show a little Elovich equation. Adv. Catal. Relat. Subj. 21, 1e49.
decrease with temperature increase from 303.15 to 333.15 K, but Alaba, P.A., Sani, Y.M., Daud, W.M.A.W., 2015. Kaolinite properties and advances for
further increase to 353.15 K increased the DG
. The positive values solid acid and basic catalyst synthesis. RSC Adv. 5 (122), 101127e101147.
Alaba, P.A., Sani, Y.M., Mohammed, I.Y., Abakr, Y.A., Daud, W.M.A.W., 2016a. Syn-
of DG
reveal that the process is less spontaneous and positive thesis and characterization of sulfated hierarchical nanoporous faujasite zeolite
values of DH
show the need energy input for the process from the for efficient transesterification of shea butter. J. Clean. Prod. 142, 1987e1993.
surrounding. Both the sorption of Pb(II) and Zn(II) on POMEAC http://dx.doi.org/10.1016/j.jclepro.2016.11.085.
Alaba, P.A., Sani, Y.M., Mohammed, I.Y., Daud, W., Ashri, W.M., 2016b. Insight into
undergo endothermic process. The positive value of DS
for Pb(II) catalyst deactivation mechanism and suppression techniques in thermocata-
removal suggested a more random interaction at the interface of lytic deoxygenation of bio-oil over zeolites. Rev. Chem. Eng. 32 (1), 71e91.
the solid/solution compared to Zn(II) removal with a negative value Amuda, O., Giwa, A., Bello, I., 2007. Removal of heavy metal from industrial
wastewater using modified activated coconut shell carbon. Biochem. Eng. J. 36
of DS
during the sorption of metal ions on POMEAC. Base on the R2 (2), 174e181.
value, Langmuir model is more suitable for determination of the Ayanda, O.S., Fatoki, O.S., Adekola, F.A., Ximba, B.J., 2013. Kinetics and equilibrium
thermodynamic parameters of Zn (R2 ¼ 0.989) compared to Pb models for the sorption of tributyltin to nZnO, activated carbon and nZnO/
activated carbon composite in artificial seawater. Mar. Pollut. Bull. 72 (1),
(R2 ¼ 0.794). 222e230.
Baccar, R., Bouzid, J., Feki, M., Montiel, A., 2009. Preparation of activated carbon
from Tunisian olive-waste cakes and its application for adsorption of heavy
4. Conclusion
metal ions. J. Hazard. Mater. 162 (2), 1522e1529.
Board, M.P.O., 2003. Oil Palm Fruit Grading Manual. Malaysian Palm Oil Board,
An activated carbon was produced from POME and utilized for Kajang, Malaysia.
sorption of Pb(II) and Zn(II) from wastewater over a wide range of Chavalparit, O., 2006. Clean Technology for the Crude Palm Oil Industry in Thailand.
Wageneningen University Wageneningen, The Netherlands.
adsorbent dosage and metal ion concentrations. The following in- Chigondo, F., Nyamunda, B., Sithole, S., Gwatidzo, L., 2013. Removal of lead (II) and
ferences were reached as the results of this work: copper (II) ions from aqueous solution by baobab (Adononsia digitata) fruit
968 G.A. Adebisi et al. / Journal of Cleaner Production 148 (2017) 958e968

shells biomass. IOSRJ. Appl. Chem. 5 (1), 43e50. using various adsorbentsdIII. Silica: rate processes. Water Res. 14 (1), 15e20.
Chowdhury, Z.Z., Zain, S.M., Khan, R.A., Islam, M.S., 2012a. Preparation and char- Mi, T., Chen, L., Xin, S.-z., Yu, X.-m., 2015. Activated carbon from the Chinese herbal
acterizations of activated carbon from kenaf fiber for equilibrium adsorption medicine waste by H3PO4 activation. J. Nanomater. 2015.
studies of copper from wastewater. Korean J. Chem. Eng. 29 (9), 1187e1195. Mishra, P., Patel, R., 2009. Removal of lead and zinc ions from water by low cost
Chowdhury, Z.Z., Zain, S.M., Khan, R.A., Rafique, R.F., Khalid, K., 2012b. Batch and adsorbents. J. Hazard. Mater. 168 (1), 319e325.
fixed bed adsorption studies of lead (II) cations from aqueous solutions onto Nacu, G., Bulgariu, D., Cristina Popescu, M., Harja, M., Toader Juravle, D., Bulgariu, L.,
granular activated carbon derived from Mangostana garcinia shell. Bio- 2015. Removal of Zn (II) ions from aqueous media on thermal activated
Resources 7 (3), 2895e2915. sawdust. Desalin. Water Treat. 1e12.
Chowdhury, Z.Z., Zain, S.M., Khan, R.A., Arami-Niya, A., Khalid, K., 2012c. Process Naje, S.A., Chelliapan, Shreeshivadasan, Zakaria, Zuriati, Ajeel, Mohammed A.,
variables optimization for preparation and characterization of novel adsorbent Alaba, P.A., 2016. A Review of Electrocoagulation Technology for the Treatment
from lignocellulosic waste. BioResources 7 (3), 3732e3754. of Textile Wastewater. http://dx.doi.org/10.1515/revce-2016-0019.
Dabrowski, A., Hubicki, Z., Podkoscielny, P., Robens, E., 2004. Selective removal of Nguyen, T., Ngo, H., Guo, W., Zhang, J., Liang, S., Yue, Q., Li, Q., Nguyen, T., 2013.
the heavy metal ions from waters and industrial wastewaters by ion-exchange Applicability of agricultural waste and by-products for adsorptive removal of
method. Chemosphere 56 (2), 91e106. heavy metals from wastewater. Bioresour. Technol. 148, 574e585.
Duan, S., Tang, R., Xue, Z., Zhang, X., Zhao, Y., Zhang, W., Zhang, J., Wang, B., Zeng, S., Onundi, Y.B., Mamun, A., Al Khatib, M., Ahmed, Y.M., 2010. Adsorption of copper,
Sun, D., 2015. Effective removal of Pb (II) using magnetic Co 0.6 Fe 2.4 O 4 nickel and lead ions from synthetic semiconductor industrial wastewater by
micro-particles as the adsorbent: synthesis and study on the kinetic and palm shell activated carbon. Int. J. Environ. Sci. Technol. 7 (4), 751e758.
thermodynamic behaviors for its adsorption. Colloids Surf. A: Physicochem. Oswal, N., Sarma, P., Zinjarde, S., Pant, A., 2002. Palm oil mill effluent treatment by a
Eng. Aspects 469, 211e223. tropical marine yeast. Bioresour. Technol. 85 (1), 35e37.
Ekebafe, L., Ekebafe, M., Erhuaga, G., Oboigba, F., 2012. Effect of reaction conditions Phang, S.-M., 1990. Algal production from agro-industrial and agricultural wastes in
on the uptake of selected heavy metals from aqueous media using composite Malaysia. Ambio 415e418.
from renewable materials. Am. J. Polym. Sci. 2 (4), 67e72. Phang, S.-M., Kim-Chong, O., 1988. Algal biomass production in digested palm oil
Fan, H.-T., Wu, J.-B., Fan, X.-L., Zhang, D.-S., Su, Z.-J., Yan, F., Sun, T., 2012. Removal of mill effluent. Biol. Wastes 25 (3), 177e191.
cadmium (II) and lead (II) from aqueous solution using sulfur-functionalized Ren, Y., Abbood, H.A., He, F., Peng, H., Huang, K., 2013. Magnetic EDTA-modified
silica prepared by hydrothermal-assisted grafting method. Chem. Eng. J. 198, chitosan/SiO 2/Fe 3 O 4 adsorbent: preparation, characterization, and applica-
355e363. tion in heavy metal adsorption. Chem. Eng. J. 226, 300e311.
Foo, K., Hameed, B., 2011. Preparation of oil palm (Elaeis) empty fruit bunch acti- Roy, A., Bhattacharya, J., 2012. Removal of Cu (II), Zn (II) and Pb (II) from water using
vated carbon by microwave-assisted KOH activation for the adsorption of microwave-assisted synthesized maghemite nanotubes. Chem. Eng. J. 211,
methylene blue. Desalination 275 (1), 302e305. 493e500.
Gratuito, M.K.B., Panyathanmaporn, T., Chumnanklang, R.-A., Sirinuntawittaya, N., Sani, Y.M., Raji-Yahya, A.O., Alaba, P.A., Aziz, A.R.A., Daud, W.M.A.W., 2015. Palm
Dutta, A., 2008. Production of activated carbon from coconut shell: optimization Frond and Spikelet as environmentally benign alternative solid acid catalysts
using response surface methodology. Bioresour. Technol. 99 (11), 4887e4895. for biodiesel production. BioResources 10 (2), 3393e3408.
Habib, M., Yusoff, F., Phang, S., Ang, K., Mohamed, S., 1997. Nutritional values of Sing, K.S., 1985. Reporting physisorption data for gas/solid systems with special
chironomid larvae grown in palm oil mill effluent and algal culture. Aquacul- reference to the determination of surface area and porosity (Recommendations
ture 158 (1), 95e105. 1984). Pure Appl. Chem. 57 (4), 603e619.
Hameed, B., Ahmad, A., Latiff, K., 2007. Adsorption of basic dye (methylene blue) Tan, Y., Chen, M., Hao, Y., 2012. High efficient removal of Pb (II) by amino-
onto activated carbon prepared from rattan sawdust. Dyes Pigments 75 (1), functionalized Fe 3 O 4 magnetic nano-particles. Chem. Eng. J. 191, 104e111.
143e149. Tseng, R.-L., 2007. Physical and chemical properties and adsorption type of activated
Igwe, J., Onyegbado, C., 2007. A review of palm oil mill effluent (POME) water carbon prepared from plum kernels by NaOH activation. J. Hazard. Mater. 147
treatment. Glob. J. Environ. Res. 1 (2), 54e62. (3), 1020e1027.
Kumar, N.S., Woo, H.-S., Min, K., 2012. Equilibrium and kinetic studies on bio- Wang, X.-s., Qin, Y., 2005. Equilibrium sorption isotherms for of Cu 2þ on rice bran.
sorption of 2, 4, 6-trichlorophenol from aqueous solutions by Acacia leucoce- Process Biochem. 40 (2), 677e680.
phala bark. Colloids Surf. B: Biointerfaces 94, 125e132. Weber, T.W., Chakravorti, R.K., 1974. Pore and solid diffusion models for fixed-bed
Lupea, M., Bulgariu, L., Macoveanu, M., 2012. Biosorption of Cd(II) from aqueous adsorbers. AIChE J. 20 (2), 228e238.
solutions on marine algae biomass. Environ. Eng. Manag. J. 11 (3), 607e615. Weber, W., Morris, J.C., 1964. Equilibria and capacities for adsorption on carbon.
Mckay, G., Otterburn, M., Sweeney, A., 1980. The removal of colour from effluent Sanit. Eng. Div. 90, 79e91.