Journal of Hazardous Materials 170 (2009) 417–424

Contents lists available at ScienceDirect

Journal of Hazardous Materials

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

Removal of methylene blue from aqueous solution by adsorption onto

pineapple leaf powder
Chih-Huang Weng a,∗ , Yao-Tung Lin b , Tai-Wei Tzeng b
a
Department of Civil and Ecological Engineering, I-Shou University, Da-Hsu Township, Kaohsiung 84008, Taiwan
b
Department of Soil and Environmental Sciences, National Chung Hsing University, TaiChung 40227, Taiwan

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

Article history: The ability of an unconventional bio-adsorbent, pineapple leaf powder (PLP) for the adsorption of methy-
Received 5 August 2008 lene blue (MB) from aqueous solution was studied. It was observed that intra-particle diffusion was
Received in revised form 20 April 2009 involved in the adsorption process and that the kinetic data fitted well with a pseudo-second-order
Accepted 21 April 2009
equation. Fitting parameters revealed that the rate of adsorption increased with decrease in dye con-
Available online 3 May 2009
centration and decrease in ionic strength while the mixing speed did not have a significant effect on
adsorption. The adsorption was favorable at higher pH and lower temperature, and the equilibrium data
Keywords:
were well fitted by the Langmuir isotherm. The maximum adsorption capacity varied from 4.68 × 10−4
Adsorption
Dye
to 9.28 × 10−4 mol/g when pH increases from 3.5 to 9.5. Thermodynamic parameters suggest that the
Methylene blue adsorption is a typical physical process, spontaneous, and exothermic in nature. The results revealed that
Pineapple leaf this agricultural waste has potential to be used as an economical adsorbent for the removal of methylene
blue from aqueous solution.
© 2009 Elsevier B.V. All rights reserved.

1. Introduction attention from a large number of researchers because: (1) it is

abundantly available; (2) most of the types of agriculture waste
Dyes are widely used as coloring agents in a variety of indus- are readily to be used and do not require a complex pretreatment
tries, such as textiles, cosmetics, pulp mills, leather, dye synthesis, step or activation process before applications; (3) regeneration of
printing, foods, and plastics. Because many organic dyestuffs are these adsorbents may not be necessary (unlike activated carbon,
harmful to human beings and hazardous to aquatic organisms, where regeneration is essential); and (4) less maintenance and
removal of dyestuffs from wastewater has received considerable supervision are required for the operation of the adsorption pro-
attention over the past decades. Among the numerous dye removal cess. Nevertheless, use of these cheap alternatives for wastewater
techniques, adsorption is the preferred method and gives the treatment remains limited due to both insufficient documentation
best results as it can be used to remove various types of color- in real wastewater systems and the necessity of post-usage disposal.
ing materials [1]. Activated carbon is the most widely employed Pineapple is largely cultivated in tropical countries, such as
adsorbent for the removal of color from textiles due to its effective- China, India, Thailand, Indonesia, and Taiwan. In Taiwan, the total
ness and high adsorption capacity; however, its use is still limited area of cultivated pineapple plants is 12,225 acres and 439,872 tons
because of high operating costs. The need for regeneration and dif- of pineapple are harvested annually, according to a 2006 report
ficulty of separation from the wastewater after use are also major from the Taiwan Council of Agriculture. Normally leaves on farm-
concerns associated with activated carbon. To lower the cost of land are used as a natural compost material or burned on site after
wastewater treatment, many researchers have focused on finding the harvest, creating a point source of air pollution. The major con-
non-conventional alternative adsorbents. Non-conventional adsor- stituents of pineapple leaf are cellulose (70–80%), lignin (5–12%),
bents or so-called low-cost adsorbents are generally referred to and hemicellulose [11]. Some studies have focused on improving
non-hazardous waste produced from industry, agriculture, and the use of this natural bio-adsorbent in short and fiber-reinforced
biosorbents. Extensive lists and reviews of these types of adsor- rubber composites [12,13] because the leaves of the pineapple plant
bent for contaminant removal have been made [1–10]. In particular, yield strong, fine silky fibers. Such uses of pineapple leaves are dou-
the use of agricultural waste in adsorption system has been drawn bly beneficial, providing a natural adsorbent and decreasing the
amount of agricultural waste.
Despite that many low-cost adsorbents have been studied for
∗ Corresponding author. Tel.: +886 7 6578957; fax: +886 7 6577461. decontamination purposes, studies of the pineapple leaf as an
E-mail address: [email protected] (C.-H. Weng). economical adsorbent for dye removal have not yet entered the

0304-3894/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.jhazmat.2009.04.080
418 C.-H. Weng et al. / Journal of Hazardous Materials 170 (2009) 417–424

scientific literature. Utilization of this waste for the treatment of was prepared. (2) Solution pH was adjusted to 6.5 using either with
wastewater is a win–win strategy because it not only converts the 0.1 M HNO3 or 0.1 M NaOH. (3) A concentration of 0.5 g/L PLP was
waste into a useful material but it also prevents on-site burning of added to the solution. (4) The resultant solution was agitated on a
the waste and saves on disposal costs. The aim of this work was magnetic stirrer at a constant speed and room temperature (27 ◦ C)
to evaluate the potential of pineapple leaf as a low-cost adsor- for 2 h. (5) A 5 mL liquor was removed at given time intervals and
bent in the removal of a basic dye, methylene blue (MB), from immediately filtered through the 0.45 ␮m membrane filter (supor-
aqueous solution. Methylene blue was chosen as a target contam- 450, Gelman Sci.) to collect the supernatant. (6) The residual MB
inant to characterize the adsorptive properties of pineapple leaf concentration in the supernatant was determined. Each run of the
as it is a common cationic dye used in the medical, textile and experiment was replicated and the average values were used in
printing industries. We investigated parameters that may affect the data analysis. Blank tests without adsorbent (PLP) were run in
the adsorption, including initial dye concentration, agitation speed, parallel to avoid possible adsorption on the reactor and the filter
ionic strength, pH, and temperature. Models to fit the adsorption apparatus.
equilibrium and kinetic data were also formulated. Thermody- Experiments to investigate the effect of mixing speed on adsorp-
namic parameters for the present system were determined based tion rate were carried out using the procedure described above
on isotherm data. Results of this study will be useful for using except that the mixing speed was maintained at 200, 400, and
this natural waste as an economic bio-adsorbent in the removal 500 rpm. Various initial MB concentrations (5 × 10−6 –4 × 10−5 M)
of methylene blue from wastewater. were prepared for gauging the effect of initial MB concentration
on adsorption rate. The effect of ionic strength on adsorption rate
was investigated using the procedures described above except that
2. Materials and methods
different ionic strengths (5 × 10−3 –7 × 10−2 M) were used.
The amount of MB adsorbed at contact time t (min), qt (mol/g),
2.1. Materials
was calculated using Eq. (1):
All reagents were of analytical grade chemicals. All solutions C0 − Ct
were prepared with double distilled water. The cationic dye used in qt = (1)
W
this study, MB or basic blue 9, was purchased from Riedel-de Haën
Co., Germany. MB has a molecular weight of 373.9 g/mol and its where C0 and Ct (mol/L) represent the concentrations of MB in the
chemical formula is C16 H18 ClN3 S. The cationic MB dye produces an solution at time t = 0 and t = t, respectively, and W (g/L) was the dose
intense molecular cation (C+ ) and reduced ions (CH+ ) in the aqueous of PLP.
solution. MB concentration in the sample solution was determined
using a spectrophotometer (HACH DR-2010, USA) at a wavelength 2.3. Equilibrium adsorption experiments
of 665 nm.
Mature pineapple leaves were collected from a tract of farm- Isotherm experiments were carried out to obtain the maximum
land near the campus of I-Shou University. The fresh leaves are adsorption capacity and thermodynamic parameters as affected
sword-shaped and measure about 75–90 cm in length and 5–7 cm by solution’s pH and temperature. The experimental procedures
in width. Upon collection, each leaf was first washed with tap water were as follows: (1) A series of 100 mL solution containing differ-
two or three times to remove dust and impurities. The leaves were ent MB concentrations with a constant ionic strength of 1 × 10−2 M
air dried and subsequently oven dried at 80 ◦ C for 2 days. Prior to (NaNO3 ). (2) Solution pH was adjusted to cover a range from 3.5
use, the dried leaves were powdered using a grinder. The pineapple to 9.5 using either HNO3 or NaOH. (3) A concentration of 0.3 g/L
leaf powder (PLP) was passed through an ASTM 200 mesh screen. of PLP was added to the solution. (4) The bottles were shaken on
Particles with diameter less than 75 ␮m were used for adsorp- a reciprocal shaker at 150 excursions/min and at room tempera-
tion studies. To characterize the PLP samples, a number of basic ture (27 ◦ C) for 80 min. This contact time was found to be adequate
properties were determined. The specific surface area and average for reaching equilibrium adsorption based on the results of a sep-
pore volume of each PLP sample were determined using a BET-N2 arate kinetic study. (5) At the end of shaking, the final pH of the
(Brunauer, Emmett, Teller) surface area analyzer (Beckman Coulter mixed liquor was recorded. (6) The liquor was filtered through the
SA3100). The pHzpc (pH at zero point of charge) of the PLP parti- 0.45 ␮m filter paper to collect the supernatant. (7) The residual MB
cles was determined with a zeta meter (Lazer Zee® model 500, Pen concentration in the supernatant was analyzed. The experimen-
Kem Inc.). The gravimetric density of the PLP particles was deter- tal procedure employed for studying the effect of temperature on
mined following the methods described in ASTM (American Society the dye adsorption isotherm was the same as that described above
for Testing and Materials) D854-92. The pH of each PLP sample except that the pH of the mixed solution was adjusted to 5.5 and the
was measured in water suspension at a 1:1 weight ratio of PLP temperature was controlled under isothermal conditions at 4, 14,
to distilled water using a pH meter. A scanning electron micro- 24, 34, 44, and 54 ◦ C by placing the mixtures in a water circulation
graph (SEM) (Philips 501 scanning electron micrograph) was used bath with temperature that remained constant within ±0.2 ◦ C. In
to characterize the morphology of the adsorbent. An FTIR (Fourier order to assure thermal equilibrium, the bottles were first put into
transform infrared spectroscopy) spectrophotometer Bruker Vec- the bath for about 30 min prior to the experiment. Control samples
tor 22 covering a wave number range of 400–4000 cm−1 was used without present MB in the mixed suspension were also used.
to identify the chemical components of the PLP.

2.4. Mathematical models

2.2. Kinetic adsorption experiments
The kinetic data was analyzed using pseudo-second-order (PSO)
Kinetic studies were carried out to establish the effect of time intra-particle diffusion (IPD) models. The PSO model is a well-
on the adsorption process and to quantify the adsorption rate. The known kinetic model proposed by Ho and McKay [14] based on
effects of mixing speed, initial MB concentration, and ionic strength the adsorption capacity, and is expressed as follows:
on the adsorption kinetics were investigated. The general experi-
mental procedures were as follows: (1) 1 L of solution containing dqt
= k(qe − qt )2 (2)
a fixed MB concentration with a constant ionic strength of NaNO3 dt
 C.-H. Weng et al. / Journal of Hazardous Materials 170 (2009) 417–424 419

where qt represents the amount of adsorbed MB on the PLP sample

(mol/g) at time t (min), qe is the amount adsorbed at equilibrium
(mol/g), and k is the PSO rate constant (g/(mol min)). After integrat-
ing the equation and applying boundary conditions qt = 0 at t = 0 and
qt = qt at t = t, Eq. (2) becomes
kq2e t
qt = (3)
1 + kqe t
In a rapidly stirred batch reactor, the IPD is likely a slow process
and is therefore rate-determining [15]. The IPD equation originally
proposed by Weber and Morris [16] was used to identify the diffu-
sion mechanism:
qe = kp t 1/2 + C (4)
where kp is the IPD rate constant (mol/(min1/2 g)) and C is the inter-
cept, which gives an indication of the thickness of the boundary
layer (i.e., the larger the intercept, the greater the boundary layer
effect [17]).
The Langmuir adsorption isotherm [18] was used to describe the Fig. 1. SEM image of PLP (1000×).
MB adsorption:
KL Ce Qm 3.2. Kinetic studies
qe = (5)
1 + KL Ce
where qe is the amount of MB adsorbed at equilibrium (mol/g), KL is 3.2.1. Effect of agitation speed
the Langmuir adsorption constant (L/mol), Ce is the equilibrium MB In the batch adsorption systems, agitation speed plays a sig-
concentration (mol/L), and Qm is the maximum adsorption capacity nificant role in affecting the external boundary film and the
(mol/g). distribution of the solute in the bulk solution [21]. The results
The graphical computer software, KaleidaGraphTM [19], was showed that the adsorption equilibrium of MB almost remained
used to perform a non-linear fit of the experimental data. In apply- unchanged as agitation speed was increased. A similar trend was
ing the above models to describe the kinetic data, the least-squares observed for varying agitation speeds at different times. These
correlation coefficient (r2 ) and normalized standard deviation (s, %) observations can be explained by the fact that the boundary layer
were calculated for each experiment. The value of s was calculated resistance was very small and the mobility of the system was high
using: under the experimental conditions. In other words, the diffusion
 of the MB ion from the solution to the surface of the PLP and into
2
[(qt,exp − qt,cal )/qt,exp ] the pores occurred quickly and easily. Since the uptake of MB was
s = 100 × (6)
n−1 not significantly influenced by the degree of agitation, an agitation
speed of 200 rpm was therefore used for all further experiments.
where qt,exp and qt,cal are, respectively, the measured and calculated
MB adsorbed at time t, respectively, and n is the number of data
points. 3.2.2. Effect of ionic strength
Fig. 2 shows the amount adsorbed as a function of contact time
at various initial concentrations. The amount adsorbed increased
3. Results and discussion

3.1. Adsorbent characteristics

The SEM micrograph of a typical PLP sample at 1000× magnifi-

cation depicted in Fig. 1 shows that the surface of PLP is porous. A
pore volume of 0.021 mL/g and a specific surface area of 5.24 m2 /g
were measured. The density of PLP is 1.53 g/cm3 , which shows that
the gravitation method can be used to separate the adsorbent. The
pH value of PLP is 6.42, indicating that this natural material can be
considered a neutral adsorbent. The result of zeta potential mea-
surements shows that the pHzpc of PLP sample is 2.3. Therefore,
at a solution pH greater than the pHzpc , the PLP possesses a neg-
atively charged surface, which is favorable for adsorbing cationic
MB. The main ingredients of PLP are cellulose, hemicellulose, and
lignin. The FTIR spectra of PLP show an intense and broad band
ranging from 3050 to 3750 cm−1 due to the hydrogen-bonded OH
vibration of the cellulosic structure of PLP [12]. A strong and sharp
band at 2910 cm−1 is attributed to the C–H stretching vibration from
CH2 group of cellulose and hemicellulose. The peak at 1735 cm−1 is
due to C–O stretching of carbonyl groups (>C O) in hemicellulose.
The peaks at 1435 and 1255 cm−1 indicate the existence of lignin
and hemicellulose structures, respectively [20]. Bands at 1400 and
1060 cm−1 reflect the C N and C O stretches [11]. From the FTIR
Fig. 2. Effects of ionic strength on the adsorption kinetics. Solid lines represent the
spectra, it is clear that some surface functional groups are present best fit of PSO equation. Conditions: initial MB 2 × 10−5 M, PLP 0.3 g/L, 27 ◦ C, final pH
on the PLP, but the specific surface functional groups are unclear. 6.5 ± 0.1.
420 C.-H. Weng et al. / Journal of Hazardous Materials 170 (2009) 417–424

5.00 × 10−5
1.82 × 104
0.999
0.3

1.9
2 × 10−5

7 × 10−2

5.16 × 10−5
2.18 × 104
0.999
0.3

1.7
2 × 10−5
−2
5 × 10

5.58 × 10−5
3.44 × 104
0.999
0.3

1.0
2 × 10−5
−2
1 × 10
Fig. 3. A linear relationship between rate constant and EDL thickness (1/).

5.84 × 10−5
with time and reached a plateau when dynamic equilibrium was

3.63 × 104
attained. The time required to reach the adsorption equilibrium

0.998
was 60 min for this concentration range. As shown, the adsorp-

0.3

1.7
2 × 10−5
−3
tion process was divided into three stages: (1) an initial stage with

5 × 10
adsorption occurring instantly, (2) subsequently slow adsorption,
and (3) a final stage with adsorption reaching equilibrium and

Pseudo-second-order adsorption rate constants and qe values obtained at different initial MB concentrations and ionic strengths.
remaining constant. The first stage can be attributed to the rapid
attachment of MB to the surface of the PLP by surface mass trans-

6.13 × 10−5
3.33 × 104
fer. At this stage, more than 80% of PLP adsorption was found in

0.999
all cases. The second stage was slower, possibly because many of

0.5

1.5
the available external sites was already occupied and because of
4 × 10−5
−2
1 × 10
the slow diffusion of MB molecules into the pore spaces of the
PLP. An asymptotic trend was found after approximately 20 min
regardless of the initial MB concentration applied to the adsorption
system. The amount adsorbed did not vary significantly at contact 3.18 × 10−5

times longer than the equilibrium time (60 min). The kinetics of the
7.22 × 104

adsorption process indicates that the adsorption of MB onto PLP can

0.999

be considered as a fast adsorption process because more than 95%

0.5

1.6
2 × 10−5
−2

of MB was adsorbed within 20 min. Such finding reveals the ben-

1 × 10

efits of using this low-cost adsorbent, or so-called eco-adsorbent,

for the treatment of wastewaters rich in dyes in general and MB in
particular.
The influence of ionic strength on the MB adsorption kinetic was
1.73 × 10−5
1.93 × 105

studied with a series of experiments at different NaNO3 concentra-

tions (5 × 10−3 –7 × 10−2 M). The results show that the amount of
0.999
0.5

1.2

MB adsorbed decreased in the presence of NaNO3 . The PSO model fit

1 × 10−5
−2
1 × 10

parameters (given in Table 1) show that an increase in ionic strength

Initial MB concentration (M)

led to a decrease in the adsorption rate constant. Ionic strength is

one of the key factors affecting the electrical double layer (EDL)
structure of a hydrated particulate. The thickness of EDL (m), 1/,
8.25 × 10−6
9.09 × 105

can be determined from the relationship:

 −0.5
0.999

1 2F 2 I × 1000
0.5

0.7

=
5 × 10−6

(7)
−2

 εεo RT
1 × 10

where F is the Faraday constant (96,500 C/mol), I is the ionic

strength (M), R is the molar gas constant (8.314 J/mol K), T is the
absolute temperature (K), ε is the dielectric constant of water
Ionic strength (M)

(78.5), and ε0 is the vacuum permittivity (8.854 × 1012 C/V m). Fig. 3
k2 (g/(mol min))

depicts the linear relationship between rate constant k2 and the

qe (mol/g)

thickness of the EDL. When ionic strength was increased from

PLP (g/L)
Table 1

5 × 10−3 to 7 × 10−2 M, the value of k decreased from 3.63 × 104

s (%)

to 1.82 × 104 (g/(mol min)) (Table 1). As indicated by Eq. (7), an

r2
 C.-H. Weng et al. / Journal of Hazardous Materials 170 (2009) 417–424 421

Table 2
Comparison of adsorption capacity of methylene blue on various adsorbents.

Adsorbents SSA (m2 /g) pH Temp. (◦ C) Qm (mol/g) K (L/mol) Refs.

Natural-based
Pineapple leaf powder 5.236 7.5 24 8.88 × 10−4 4.80 × 104 This study
Rice husk NA NA 20 8.36 × 10−4 2.00 × 106 [24]
Cotton waste NAa NAa 20 7.44 × 10−4 9.67 × 105 [24]
Peanut hull 72.35 5.0 20 1.82 × 10−4 5.87 × 104 [25]
Orange peel 22.1 7.2 30 5.51 × 10−5 6.72 × 106 [26]
Neem leaf powder NAa NAa 27 2.35 × 10−5 1.31 × 105 [22]

Carbon/polymer
Commercial powder ACb NAa 7.4 30 2.62 × 10−3 1.79 × 105 [16]
Rice husk ash 101.3 6.4 30 2.16 × 10−3 1.60 × 104 [27]
Rice husk carbon NA 7.4 30 9.18 × 10−4 1.34 × 105 [16]
H3 PO4 activated carbon 1184 NA Room 8.45 × 10−4 NA [28]
Coconut shell carbon NAa 7.4 30 7.43 × 10−4 3.40 × 104 [16]
Polymer, Jalshakti® NAa 5.5 NAa 4.61 × 10−4 1.96 × 105 [29]
Groundnut shell carbon NAa 7.4 30 4.41 × 10−4 4.79 × 104 [16]
Bamboo dust carbon NAa 7.4 30 3.82 × 10−4 4.49 × 104 [16]
Granular ACb (Nuchar C-190) NAa 5.5 NAa 2.79 × 10−4 1.20 × 104 [29]
Coconut shell fibers carbon 978 6.0 30 5.24 × 10−5 2.57 × 104 [30]
Coir pith carbon 167 6.9 20 1.57 × 10−5 3.48 × 105 [4]

Biosorbent
Sargassm muticum NAa 5.5 25 7.46 × 10−4 9.35 × 103 [31]
Diamionethane sporopollenin NAa NAa NAa 4.8 × 10−5 1.12 × 105 [32]
Sewage sludges NAa NAa 25 3.08 × 10−4 1.12 × 104 [33]
Ash/clay
Spent activated clay 171 6 28 4.93 × 10−4 6.47 × 104 [34]
Clay 30 NAa 20 1.56 × 10−4 NAa [35]
Kaolinite 27.49 NAa 27 3.75 × 10−5 NAa [36]
Sono-chemical treated fly ash 35.4 9.7 30 1.2 × 10−5 NAa [37]
Fly ash NAa 7.5 30 8.23 × 10−6 2.86 × 105 [38]
Petrified sediment 28.1 7.0 30 6.4 × 10−6 1.42 × 105 [7]
Sludge ash 3.7 9.8 24 5.00 × 10−6 4.29 × 105 [1]
a
NA: not available/not reported.
b
AC: activated carbon.

increase in ionic strength leads to a decrease in EDL thickness and process, involving adsorption on the external surface and diffu-
an increase in the amount of indifferent ions approaching the PLP sion into the internal pores of the PLP. The first intercepting line
surface. Thus, the results shown above can be attributed in part to represents IPD while the second indicates the solution was pro-
increased competition between MB and Na+ ions for surface sites gressively reaching equilibrium adsorption. The slope of the first
with increasing the ionic strength. linear portion gives the intra-particle rate constant, kp , and the
intercept of this portion is proportional to the thickness of boundary
3.2.3. Effect of initial MB concentration
Results show that increased initial MB concentration led to
increased MB adsorption. Apparently, the initial MB concentration
plays an important role in affecting the capacity of MB to absorb
onto PLP. The higher the MB concentration is, the stronger the driv-
ing force of the concentration gradient, and therefore the higher
the adsorption capacity. The corresponding PSO model fit param-
eters (given in Table 1) indicate that the data can be described by
this model with relatively low standard deviation and high correla-
tion coefficient. It was found that the adsorption rate was inversely
proportional to the initial MB concentration. When initial MB con-
centration increased from 5 × 10−6 to 4 × 10−5 M, the PSO rate
constant k2 decreased from 9.09 × 105 to 3.33 × 104 g/(mol min).
Higher initial MB concentration resulted in a lower diffusion effi-
ciency and more competition of adsorbing ions for a fixed activated
surface site; consequently, a lower k2 value was observed. Similar
results have also been reported by Crini et al. [21] for the adsorption
of dyes onto cyclodextrin-based peat.

3.2.4. Intra-particle diffusion

Normally, if a plot of the amount of dye adsorbed versus t1/2
shows two or more intercepting lines, the adsorption steps are inde-
pendent of one another [22]. The IPD plots and the fit parameters
of the effect of ionic strength are depicted in Fig. 4. These plots
clearly exhibit two distinguishable intercepting lines, which indi-
cate that the adsorption process was controlled by a multi-step Fig. 4. IPD plots and fit parameters at different ionic strengths.
422 C.-H. Weng et al. / Journal of Hazardous Materials 170 (2009) 417–424

layer. As seen in Fig. 4, values of kp increase slightly with increasing

ionic strength. This may imply that the diffusion rate is not signifi-
cantly affected by the changes in ionic strength. The boundary layer
parameter, C, decreased with increasing ionic strength, indicating
that the boundary layer effect becomes prominent at lower ionic
strength.
The driving force of diffusion is very important for the adsorp-
tion processes and is associated with the adsorbate concentration
in the bulk solution. Therefore, the fit parameters of the IPD model
for the effects of initial MB concentrations on adsorption were
also made. The resulting values of kp increase from 2.93 × 10−7 to
3.91 × 10−6 mol/(min1/2 g) when initial MB concentration increases
from 5 × 10−6 to 4 × 10−5 M. It appears that an increase in adsorbate
concentration results in an increase in the driving force, which leads
to an increase in the MB diffusion rate. The results show that the val-
ues of intercept C increase from 7.32 × 10−6 to 4.54 × 10−5 mol/g as
initial MB concentration increases from 5 × 10−6 to 4 × 10−5 M. This
indicates that a higher initial MB concentration results in a stronger
boundary layer effect in the adsorption process. The increase in
the value of C also provides a measure of the abundance of solute
adsorbed on the boundary layer. Similar boundary layer effects
were served for the adsorption of malachite green on cyclodextrin-
based adsorbent [21].

3.3. Equilibrium studies

3.3.1. Effect of pH
The effects of pH on the MB adsorption isotherms are shown
in Fig. 5a. As shown, the amount of MB adsorbed increases with
increasing pH. The maximum monolayer adsorption capacity (Qm )
and the corresponding Langmuir constants (KL ) are given in the
figure. The high value of r2 (>0.98) indicates that the experimental
data were well correlated to the Langmuir model output. As shown,
the values of KL remain constant while the values of Qm increase as
the pH increases. The adsorption of MB onto PLP seems to be greatly
affected much by the solution’s pH. Such pH-dependence may be
due in part to the fact that the surface charge of PLP is strongly
related to the solution pH.
Results from the FTIR analysis indicate that PLP consists of vari-
ous functional groups such as hydroxyl (–OH) and carbonyl (>C O)
that may be influenced by the pH. The pKa of MB is 0.04 [23]; hence,
Fig. 5. Effect of (a) pH and (b) temperature on the isotherm for MB adsorption onto
MB is completely ionized at pH greater than 0.04 and serves as a PLP. Solid lines are the best fit of Langmuir isotherm.
cationic species. At alkaline pH, a significant strong electrostatic
attraction exists between the positively charged cationic dye and
A comparison between MB adsorption capacities of various
the negatively charged PLP surface due to the ionization of PLP
adsorbents is given in Table 2. Our results (8.88 × 10−4 mol/g) are
functional groups. As shown previously, the pHzpc of PLP parti-
quite similar to similar investigations for coal (8.67 × 10−4 mol/g),
cles was 2.3. The PLP has a positively charged surface at pH less
H3 PO4 activated carbon (8.45 × 10−4 mol/g), and rice husk
than the pHzpc . On the contrary, when pH is greater than pHzpc ,
(8.36 × 10−4 mol/g). The MB adsorption capacity and the adsorp-
the negatively charged PLP surface is favorable for the adsorption
tion affinity constant of PLP are much higher than commercial
of cationic MB dye. Thus, less adsorption of basic MB dye occurs at
granular activated carbon (Nuchar C-190) and other natural-based
lower pH. Additionally, less adsorption of MB at acidic conditions
adsorbents such as rice husk, neem leaf powder, orange peel, and
can be attributed in part to the presence of excess H+ ions desta-
activated carbon derived from coconut shell fibers.
bilizing the basic dye and competing with cationic dye ions for the
adsorption sites. The effect of pH on the solid–liquid equilibrium
can be described by the following possible mechanisms: 3.3.2. Effect of temperature
The effect of temperature on the adsorption isotherm was inves-
tigated under isothermal conditions in the temperature range
(1) chemical interaction between the hydroxyl and carboxyl groups
of 4–54 ◦ C. The temperature dependence of the MB adsorption
of the active component of PLP and the reactive group (Cl− ) of
isotherm is shown in Fig. 5b. The experimental results indicate that
MB.
the magnitude of adsorption increases with decreasing tempera-
S–O− + MB+ = S–O–MB (8) ture, suggesting that the adsorption reaction was exothermic in
nature. The values of the adsorption constants obtained from the
where S represents the active surface sites of PLP and S–O− Langmuir model are in the figure.
represents the ionized surface functional groups of PLP; or To determine whether the process will occur spontaneously, a
(2) weak electrostatic interaction between the MB and the set of thermodynamic parameters for the MB–PLP adsorption sys-
electron-rich sites of the PLP surface. tem were calculated, including changes in the standard free energy
 C.-H. Weng et al. / Journal of Hazardous Materials 170 (2009) 417–424 423

Table 3
Thermodynamic parameters for the adsorption of MB onto PLP.

G◦ (kJ/mol) H◦ (kJ/mol) S◦ (J/(K mol))

◦ ◦ ◦ ◦ ◦ ◦
4 C 14 C 24 C 34 C 44 C 54 C

−24.76 −25.53 −26.12 −26.83 −27.51 −28.19 −5.93 76.4

(G◦ ), enthalpy (H◦ ), and entropy (S◦ ): References

G◦ = −RT ln KL (9) [1] S. Babel, T.A. Kurniawan, Low-cost adsorbents for heavy metals uptake
from contaminated water: a review, J. Hazard. Mater. B 97 (2003)
219–243.
S ◦ H ◦ 1 [2] G. Crini, Non-conventional low-cost adsorbents for dye removal: a review,
ln KL = − (10)
R R T Bioresour. Technol. 97 (2006) 1061–1085.
[3] C.H. Weng, Y.F. Pan, Adsorption characteristics of methylene blue from aqueous
The values of H◦ and S◦ were determined from the slopes and solution by sludge ash, Colloids Surf. A: Physicochem. Eng. Aspects 274 (2006)
154–162.
intercepts, respectively, of the van’ Hoff plot (ln KL versus 1/T). The
[4] D. Kavitha, C. Namasivayam, Experimental and kinetic studies on methylene
calculated thermodynamic parameters are given in Table 3. Because blue adsorption by coir pith carbon, Bioresour. Technol. 98 (2007) 14–21.
all G◦ values are negative, the adsorption of MB onto PLP is a [5] C.H. Weng, C.Z. Tsai, S.H. Chu, Y.C. Sharma, Adsorption characteristics of copper
spontaneous process, confirming the affinity of PLP for the dye. Gen- (II) onto spent activated clay, Sep. Purif. Technol. 54 (2007) 187–197.
[6] S.B. Bukallah, M.A. Rauf, S.S. Alali, Removal of methylene blue from aqueous
erally, a value of G◦ in between 0 and −20 kJ/mol is consistent with solution by adsorption on sand, Dyes Pigments 74 (2007) 85–87.
electrostatic interaction between adsorption sites and the adsorb- [7] A.Z. Aroguz, J. Gulen, R.H. Evers, Adsorption of methylene blue from aque-
ing ion (physical adsorption) while a more negative G◦ value ous solution on pyrolyzed petrified sediment, Bioresour. Technol. 99 (2008)
1503–1508.
ranging from −80 to −400 kJ/mol indicates that the adsorption [8] B.H. Hameed, Spent tea leaves: a new non-conventional and low-cost adsorbent
involves charge sharing or transferring from the adsorbent surface for removal of basic dye from aqueous solutions, J. Hazard. Mater. 161 (2009)
to the adsorbing ion to form a coordinate bond (chemisorption) 753–759.
[9] J. Febrianto, A.N. Kosasih, J. Sunarso, Y.H. Ju, N. Indraswati, S. Ismadji,
[39–41]. As shown, the magnitude of G◦ (−24.76 to −28.19 kJ/mol) Equilibrium and kinetic studies in adsorption of heavy metals using
for this adsorption system is a slightly above the physical adsorp- biosorbent: a summary of recent studies, J. Hazard. Mater. 162 (2009)
tion range but significantly below the chemisorption range. These 616–645.
[10] J. Wang, C.P. Huang, H.E. Allen, D.K. Cha, D.W. Kim, Adsorption character-
values suggest that the adsorption is a typical physical process
istics of dye onto sludge particulates, J. Colloid Interface Sci. 208 (1998)
enhanced by the chemical effect. The small negative value of H◦ 518–528.
(−5.93 kJ/mol) indicates that the adsorption is physical in nature [11] S.C. Saha, B.K. Das, P.K. Ray, C.M. Pandey, K. Goswami, Infrared spectra of raw and
chemically modified pineapple leaf fiber (annanus comosus), J. Appl. Polym. Sci.
involving weak attractive forces and is exothermic. This implies that
43 (1991) 1885–1890.
the adsorption process is energetically stable. The low value of H◦ [12] A.K. Monanty, P.C. Tripathy, M. Misra, S. Parija, S. Sahoo, Natural fiber pineap-
also implies loose bonding between the MB molecules and the PLP ple chemical modification of pineapple leaf fiber: graft copolymerization of
surface [41]. The positive value of S◦ (76.4 J/(K mol)) indicates an acrylonitrile onto defatted pineapple leaf fibers, J. Appl. Polym. Sci. 77 (2000)
3035–3043.
increase in the degree of freedom (or disorder) of the adsorbed [13] N. Lopattananon, K. Panawarangkul, K. Sahakaro, B. Ellis, Performance of
species. In general, the thermodynamic parameters indicate that pineapple leaf fiber-natural rubber composites: the effect of fiber surface treat-
the adsorption is spontaneous and exothermic. ments, J. Appl. Polym. Sci. 102 (2006) 1974–1984.
[14] Y.S. Ho, G. McKay, Pseudo-second order model for sorption processes, Process
Biochem. 34 (1999) 735–742.
[15] G. McKay, The adsorption of dyestuffs from aqueous solution using activated
4. Conclusions carbon: analytical solution for batch adsorption based on external mass transfer
and pore diffusion, Chem. Eng. J. 27 (1983) 187–196.
Present studies showed that the adsorption of MB onto [16] W.J. Weber Jr., J.C. Morris, Kinetic of adsorption on carbon from solution, J.
Sanitray Eng. Div. ASCE 89 (1962) 31–59.
PLP is favored at high pH, lower temperature, and low ionic [17] N. Kannan, M.M. Sundaram, Kinetics and mechanism of removal of methylene
strength. The adsorption kinetics followed a pseudo-second- blue by adsorption on various carbons—a comparative study, Dyes Pigments 51
order kinetic model and intra-particle diffusion was involved (2001) 25–40.
[18] I. Langmuir, The adsorption of gases on plan surfaces of glass, mica and plat-
in the adsorption process. Fitting of Langmuir isotherm data
inum, J. Am. Chem. Soc. 40 (1918) 1361–1403.
showed that the maximum adsorption capacity increases with [19] KaleidaGraphTM (Version 3.6), Synergy Software, Reading, PA, USA, 2003.
increasing pH and decreasing temperature. PLP has a maxi- [20] L.Y. Mwaikambo, M.J. Ansell, Chemical modification of hemp, sisal,
mum adsorption capacity of 8.88 × 10−4 mol/g at pH 7.5 and jute, and kapok fibers by alkalization, J. Appl. Polym. Sci. 84 (2002)
2222–2234.
24 ◦ C, which is much higher than many low-cost adsorbents. [21] G. Crini, H.N. Peindy, F. Gimber, C. Robert, Removal of C.I. basic green 4 (mala-
Thermodynamic parameters suggest that the adsorption is a typ- chite green) from aqueous solutions by adsorption using cyclodextrin-based
ical physical process, spontaneous, and exothermic in nature. adsorbent: kinetic and equilibrium studies, Sep. Purif. Technol. 53 (2007)
97–110.
This study revealed that the PLP can be used as an econom- [22] K.G. Bhattacharyya, A. Sharma, Kinetic and thermodynamics of methylene blue
ical natural-based adsorbent to remove MB dye from aqueous adsorption on neem (Azadirachta indica) leaf powder, Dyes Pigments 65 (2005)
solution. When the PLP used as adsorbent becomes saturated, 51–59.
[23] A. Albert, E.P. Sergeant, Ionization Constants of Acids and Bases. A Laboratory
it can be incinerated. Since the production of pineapple occurs Manual, Wiley, New York, NY, 1962, pp. 69–91.
year-round in most tropical countries, wasted pineapple leaves [24] G. McKay, J.F. Porter, G.R. Prasad, Removal of dye colours from aqueous solu-
can easily be acquired and used for future commercial applica- tions by adsorption on low-cost materials, Water Air Soil Pollut. 114 (1999)
423–438.
tions. [25] R. Gong, Y. Sun, J. Chen, H. Liu, C. Yang, Effect of chemical modifica-
tion on dye adsorption capacity of peanut hull, Dyes Pigments 67 (2005)
175–181.
Acknowledgements [26] G. Annadurai, R.L. Juang, D.J. Lee, Use of cellulose-based wastes for sorption of
dyes from aqueous solutions, J. Hazard. Mater. 92 (2002) 263–274.
[27] S. Chandrasekhar, P.N. Pramada, Rice husk ash as an adsorbent for methylene
This work was supported by the National Science Council
blue—effect of ashing temperature, Adsorption 12 (2006) 27–43.
of R.O.C. (Grant No. NSC 96-2221-E-214-013-MY3). The authors [28] A.A. Attia, B.S. Girgis, N.A. Fathy, Removal of methylene blue by carbons derived
express their thanks to the fellows of MANALAB, I-Shou University from peach stones by H3 PO4 activation: batch and column studies, Dyes Pig-
for supporting the SEM and EDS analyses. ments 65 (2008) 51–59.
424 C.-H. Weng et al. / Journal of Hazardous Materials 170 (2009) 417–424

[29] R. Dhodapkar, N.N. Rao, S.P. Pande, S.N. Kaul, Removal of basic dyes from aque- [35] A. Gurses, C. Dogar, M. Yalcin, M. Acikyildiz, R. Bayrak, S. Karaca, The adsorption
ous medium using a novel polymer: Jalshakti, Bioresour. Technol. 97 (2006) kinetics of the cationic dye, methylene blue, onto clay, J. Hazard. Mater. B131
877–885. (2006) 217–228.
[30] K.P. Singh, D. Mohan, S. Sinha, G.S. Tondon, D. Gosh, Color removal from wastew- [36] D. Ghosh, K.G. Bhattacharyya, Adsorption of methylene blue on kaolinite, Appl.
ater using low-cost activated carbon derived from agricultural waste material, Clay Sci. 20 (2002) 295–300.
Ind. Eng. Chem. Res. 42 (2003) 1965–1976. [37] S. Wang, Z.H. Zhu, Sonochemical treatment of fly ash for dye removal from
[31] E. Rubin, P. Rodriguez, R. Herrero, J. Cremades, I. Barbara, M.E.S. Vicente, Removal wastewater, J. Hazard. Mater. B 126 (2005) 91–95.
of methylene blue from aqueous solutions using as biosorbent Sargassum [38] V.V.B. Rao, S.R.M. Rao, Adsorption studies on treatment of tex-
muticum: an invasive macroalga in Eurpe, J. Chem. Bioresour. Technol. 80 (2005) tile dyeing industrial effluent by fly ash, Chem. Eng. J. 116 (2006)
291–298. 77–84.
[32] M. Kucukosmanoglu, O. Gezici, A. Ayar, The adsorption behaviors of methylene [39] M.J. Jaycock, G.D. Parfitt, Chemistry of Interfaces, Ellis Horwood Limited,
blue and methyl orange in a diaminoethane sporopollenin-mediated column 1981.
system, Sep. Purif. Technol. 52 (2006) 280–287. [40] M. Horsfall Jr., A.A. Abia, A.I. Spiff, Kinetic studies on the adsorption
[33] F. Rozada, L.F. Calvo, A.I. Garcia, J. Martin-Villacorta, M. Otero, Dye adsorption of Cd2+ , Cu2+ and Zn2+ ions from aqueous solutions by cassava (mani-
by sewage sludge-based activated carbons in batch and fixed-bed systems, hot sculenta cranz) tuber bark waste, Bioresour. Technol. 97 (2006)
Bioresour. Technol. 87 (2003) 221–230. 283–291.
[34] C.H. Weng, Y.F. Pan, Adsorption of a cationic dye (methylene blue) onto spent [41] D. Sigh, Studies of the adsorption thermodynamics of oxamyl on fly ash, Adsorpt.
activated clay, J. Hazard. Mater. 144 (2007) 355–362. Sci. Technol. 18 (2000) 741–748.