- Email: [email protected]
Novel activated carbon from Manihot esculenta Crantz for removal of Methylene Blue
Accepted Manuscript Novel activated carbon from Manihot esculenta Crantz for removal of Methylene Blue Buscotin Horax Beakou, Kaoutar El Hassani, Mohammed Amine Houssaini, Mounir Belbahloul, Elhassan Oukani, Abdellah Anouar PII:
S2468-2039(17)30053-5
DOI:
10.1016/j.serj.2017.06.003
Reference:
SERJ 92
To appear in:
Sustainable Environment Research
Received Date: 22 February 2017 Revised Date:
10 April 2017
Accepted Date: 28 June 2017
Please cite this article as: Beakou BH, El Hassani K, Houssaini MA, Belbahloul M, Oukani E, Anouar A, Novel activated carbon from Manihot esculenta Crantz for removal of Methylene Blue, Sustainable Environment Research (2017), doi: 10.1016/j.serj.2017.06.003. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Received 23 February 2017 Received in revised form 10 April 2017 Accepted 28 June 2017
Blue
RI PT
Novel activated carbon from Manihot esculenta Crantz for removal of Methylene
SC
Buscotin Horax Beakou*, Kaoutar El Hassani, Mohammed Amine Houssaini,
M AN U
Mounir Belbahloul, Elhassan Oukani, Abdellah Anouar
Department of Applied Chemistry and Environment, University Hassan First, Settat 26000, Morocco
AC C
EP
TE D
Keywords: Adsorption; Manihot esculenta Crantz; Methylene Blue; Box Behnken
*Corresponding author Email: [email protected] 1
ACCEPTED MANUSCRIPT ABSTRACT The adsorptive removal of Methylene Blue by a novel bio char namely Cassava Rind Carbon was studied in a batch system. Moreover, several experiments were performed to optimize operating conditions of the adsorption process. Thermo
RI PT
gravimetric analysis investigates the thermal behaviour of the raw material. The surface functional groups were identified with Attenuated Total Reflectance Fourier Transform Infra-Red spectroscopy. The carbon surface was also characterised by BET surface area,
SC
scanning electron microscopy, X-ray diffraction and zero charge point of pH. Among the three kinetic models used to calculate the adsorption rate constants, the pseudo-
M AN U
second order was the most accurate. The experimental isotherms were analysed and Redlich-Peterson’s model provided the best fit. The maximum capacity according to Langmuir model was 565 mg g-1 at 25 °C. The thermodynamic study revealed that the adsorption is endothermic, physical and random. The temperature, the mass of
TE D
adsorbent and the concentration of dye were the three factors optimized in a Box Behnken Response Surface Methodology, with an optimal amount of adsorbed MB of
EP
771 mg g-1.
1. Introduction
AC C
Many adsorbents can be synthesized from waste materials of agricultural industry,
fruit waste and plant waste [1]. Agricultural residues as precursors for activated carbon production are the focus of many studies because they are sustainable adsorbents with high porosity and high reactivity [2]. Using agricultural waste peels as biomass for the production of low cost sorbents for water treatment applications removes an ecological burden for the society [3]. One of the applications of those lignocellulosic bio-sorbents is the removal of hazardous dyes from aqueous waste. Almost 700 kt of dyes are
2
ACCEPTED MANUSCRIPT produced worldwide each year. About half of this amount comes from textile industry. Among the 100,000 dyes available commercially, 70 wt% are azoic dyes [4]. Unfortunately, at least 10% of those synthetics dyes are released into the environment. They are one of the most hazardous species found in industrial effluents, so treatment is
RI PT
mandatory. They can reduce light penetration, precluding the photosynthesis of aqueous flora. In addition, some azoic dyes are able to cause allergic reactions and exhibit cytotoxicity [5]. They can show carcinogenic and mutagenic events [6].
SC
Different methods to remove dye such as adsorption on various sorbents (layered double hydroxides [7], lignite [8], natural adsorbent [9], activated carbon [10],
M AN U
nanomaterial [11,12]), chemical decomposition by oxidation, photo degradation and microbiological decolourisation [13]. All of them have advantages and drawbacks. Due to the high cost and disposal problems, many of these conventional methods for treating dye wastewater have not been widely applied in large scale in the textile and paper
TE D
industries. Low cost adsorbents may make adsorption the best solution for wastewater decolourisation [14]. However, the biomass needs to be activated to reach great performance. Despite its simplicity, physical activation generates higher amounts of
EP
volatile matter and requires more energy. On the contrary, chemical activation uses dehydrogenation reactions, which inhibit tar formation and reduce the volatile
AC C
substances generated [15]. Nevertheless, chemical activation requires an additional step and use of additional reagents. Among the various chemical reagents, phosphoric acid and zinc chloride have been used extensively for biomass precursors while potassium hydroxide is preferred for coal-based precursors. Moreover, phosphoric acid is preferred to zinc chloride due to its non-polluting nature [16]. The origin of cassava (Manihot esculenta Crantz) is well documented [17]. Taxonomically, it belongs to the Euphorbiaceae family and Manihot genus. Among the
3
ACCEPTED MANUSCRIPT crops, cassava has the lowest blue water footprint [18]. The biomass of our study, the thin outer layer, rough and brown on the outside of cassava root, is a non-valorised waste. This permits the reuse of the rest of the peel as better product for animal feed. There is no study using the so-called cassava rind as adsorbent. Therefore, in this work,
RI PT
the cassava rind was chemically activated with phosphoric acid (weight ratio 1:1) and carbonisation was done at 120 °C. Characterisation by Thermo Gravimetric Analysis (TGA),
Fourier
Transform
Infra-Red
through
Attenuated
Total
Reflectance
SC
(FTIR/ATR), Scanning Electron Microscopy (SEM), BET surface area and X-ray Diffraction (XRD) was performed. The determination of the point of zero charge of pH
M AN U
(pHZPC) and the pH influence were also carried out. Methylene Blue (MB), an azoic dye, was used as a model compound for evaluating the optimized adsorption capacity of synthesized bio char.
Adsorption mechanism of Cassava Rind Carbon (CRC) is highlighted through
TE D
adsorption isotherms, kinetics and thermodynamics. The efficiency of CRC for MB removal was optimized by altering three parameters namely temperature, mass of adsorbent and concentration of dye. The Box Behnken Surface Response Methodology
EP
allows the determination of the influence of those parameters.
AC C
2. Materials and methods 2.1 Carbon preparation
The cassava roots were purchased in a local farm in Djakotomey, Benin. The roots
were thoroughly washed with distilled water and peeled with knife. The rind was separated from the peel manually and dried in the sun until all the moisture had evaporated. Then a steel container was filled with the cassava rind and tightly closed. The product was ground and sieved to size between 63 and 90 µm. The activation
4
ACCEPTED MANUSCRIPT protocol [19] was used with some modifications as follows. A mixture of the raw material (10 g) and phosphoric acid (weight ratio 1:1) was heated at 120 °C for 14 h. Then the resulting char was ground and soaked in 200 mL of sodium hydroxide (1 M) for 24 h to remove residual acid and washed with distilled water until the pH was
RI PT
neutral (pH-meter pH-2005, J.P. Selecta, Spain). The reagents were of analytical grade (Sigma-Aldrich, USA). Finally, the carbon was dried at 60 °C for 6 h and sieved to size between 40 and 63 µm.
SC
2.2 Characterisation
The thermal behaviour of the raw material was studied using a LABSYS evo TGA
M AN U
1150 (Setaram Instrumentation, France). The TGA experiment was conducted under the following conditions: initial temperature 50 °C, final temperature 600 °C and heating rate of 10 °C min-1 with an Argon flow rate of 25 cm3 min-1.
A FTIR/ATR, Varian 800/gladiATR (Scimitar series, Australia/Pike Technologies,
TE D
USA), allows the determination of surface functional groups of four products (raw material, CRC, MB and CRC with MB loaded). The spectra were recorded in the range of 400–4,000 cm−1 with 50 scans collected at 4 cm−1 resolution.
EP
The BET equation was used to obtain the specific surface areas. Pore volume analysis was performed using the Barrett–Joyner–Halenda (BJH) method. Before each
AC C
measurement on QUADRASORB SI (Quantachrome, Germany), the samples were first degassed at 150 °C for 24 h. Then, the measurements were conducted through the adsorption of nitrogen at liquid nitrogen temperature onto the biopolymer surface. SEM analysis was performed on a Phenom Xl (Phenom-world, Netherland) with an accelerating voltage of 10 kV. CRC powder was spread on carbon tape adhered to a SEM stage. Before imaging, the samples were coated with a thin gold layer to improve the image quality.
5
ACCEPTED MANUSCRIPT XRD was also performed at room temperature under ambient air conditions, using Cu Kα1 radiation (λ = 1.5406 Å) at 10 mA, 30 kV and a 2θ angle ranging from 2 to 90° (PHASER D2 Diffractometer, Bruker, Billerica, Massachusetts, USA). 2.3 Adsorption measurements
RI PT
A 1,000 mg L-1 stock solution was prepared by dissolving MB dye (Riedel-de Haen, Germany) in distilled water. The different concentrations of MB were prepared from this stock solution. The solutions used for pH adjustment, NaOH 0.1 M and HCl 0.1 M,
SC
were from their corresponding analytical grade solutions (Sigma-Aldrich, USA). Batch adsorption experiments were done at 300 RPM. The supernatants were separated by
M AN U
centrifugation at 7,500 RPM for 2 min (Sigma Laborzentrifugen, Germany) and collected by using disposable syringes. The MB residual concentrations in the solutions were determined by monitoring the absorbance with a UV–Vis spectrophotometer DR 6000 with RFID Technology (Hach Lange, Germany) at 663 nm. The amount of MB
using Eq. (1). Q =
( )
TE D
adsorbed per unit mass of the adsorbent at equilibrium, Qe (mg g-1), was calculated
(1)
EP
where Ci and Ce are the initial and the equilibrium concentration of MB (mg L-1)
AC C
respectively, V is the volume of solution (L) and W is the mass of adsorbent (g). To determine the effect of pH, the adsorption of dye molecules was investigated
over a pH range of 3 to 11. The activated carbon (25 mg) was added into the dye solution (50 mL, 200 mg L-1) and mixing was carried out at room temperature (25 ± 1 °C) for 3 h before centrifugation. Moreover, in the same range of pH, the pHZPC was measured as follows: 50 mL of KNO3 solution (0.1 M) with 0.1 g of CRC for 48 h of agitation at room temperature (25 ± 1 °C) [20].
6
ACCEPTED MANUSCRIPT Adsorption kinetic experiment was carried out on 1000 mL dye solution (50 mg L-1, pH = 10) mixed with 0.2 g of activated carbon at room temperature (25 ± 1 °C) for 3 h. Several millilitres of solution were sampled using disposable syringes at various time intervals.
RI PT
Adsorption isotherms were determined by measuring the depletion of MB concentration after the adsorption reached equilibrium. For these experiments, CRC (25 mg) was added into dye solutions (50 mL) of different concentrations from 200 to 350
SC
mg L-1 with a pH value of 10. The mixing was carried out, in a double-jacketed beaker in which temperature was maintained with a thermostat (J.P. Selecta Frigiderm, Spain)
were 25, 30, 35, 40, 45 °C.
M AN U
for 3 h before centrifugation. The different temperature values used in the experiment
2.4 Box-Behnken Response Surface Methodology (BBRSM)
A Response Surface Methodology approach is a statistical method that uses
TE D
experimental data obtained from specified experimental design to model and optimize any process in which response of interest is influenced by several variables. It is intensively used to optimize dyes adsorption. Box-Behnken experimental design is a
EP
standard statistical tool mostly used for process optimization with minimum number of experiments [21]. The process variables were the adsorption temperature (X1), dose of
AC C
adsorbent (X2), and initial dye concentration (X3), while the response variable was the amount of MB adsorbed per unit mass of the adsorbent (Y). A minimum number of 15 experiments is recommended by the Box-Behnken design to optimize the process parameters. Among these 15 experiments, the last three runs were carried out under identical conditions. The upper limit, central point and lower limit of the variables coded “-1”, “0”, and “1” are (20, 45, 70 °C), (15, 20, 25 mg) and (200, 300, 400 mg L-1) respectively. Conditions and results of the experiment are listed in Table 1. The central
7
ACCEPTED MANUSCRIPT points, the upper and lower limits were fixed based on the potential observed on isotherm 45 °C results as well as on the preliminary experimental runs. The BBRSM polynomial equation used to model the amount of adsorbed MB (Y) from the water is
Y = β + ∑ β X ∑ β X X + ∑ β X
RI PT
represented in Eq. (2).
(2)
SC
where Y is the predicted response, β0 is a constant, βi is the linear coefficient, βij is the interaction coefficients, βii is the quadratic coefficients, and Xi and Xj are the coded
M AN U
values of the process variables. The results of experiments were analysed using statistical computing software Minitab 17 utilising the model equation and the analysis of variance (ANOVA).
3.1 Characterization
TE D
3. Results and discussion
The TGA curves of the raw material in (Fig. 1a) show two types of weight loss. The
EP
first one occurs at 100 °C and corresponds to the dehydration of the material. The
AC C
maximum depletion of mass (40%) is around 300 °C. The degradation of hemicellulose happens in the range 241–297 °C, while α-cellulose and lignin in the fibre degrades at 297–353 °C. The thermal decomposition of lignin occurs within 353–500 °C [22]. At 350 °C most of a lignin-cellulose mixture disappears [23] and after 400 °C total degradation is completed [24]. Infrared techniques are fast, accurate, and low-cost for analysis of functional groups in biomass [25]. FTIR/ATR allows attenuation of the incident radiation and provides infrared spectra without the water background absorbance. In Fig. 1b, the spectra after
8
ACCEPTED MANUSCRIPT activation have less adsorption bands than those of the raw material, indicating that various functional groups present in the raw material spectra disappear after activation. It can be noticed that C=O stretch of aldehyde of hemicellulose in the raw material at 1732 cm-1 disappears It is also observed in the raw material spectrum at 1156 cm−1 that
RI PT
there is C-O-C asymmetrical stretching of hemicellulose and cellulose. In the same spectrum at 1031 cm-1 C-O, C=C, and C-C-O stretching can be attributed to C-O-C asymmetrical stretching of hemicellulose, cellulose, and lignin [26]. In the CRC spectra,
SC
only cellulose and lignin bands were observed, indicating that hemicellulose content was lost during activation. At 980 cm-1, C-O valence vibration of cellulose can be
M AN U
observed [27]. In the raw material as well as CRC+MB (the activated carbon loaded with dye) spectra at 2917 and 2850 cm−1 there is C-H stretching of lignin [28]. At 1593 cm-1, it can be observed that there is a strong intensity band in both MB and CRC+MB spectra, corresponding to the skeletal stretching vibrations of the C=C and C=N bonds
TE D
in the heterocycle of MB [29]. More bands of the C-H bending vibrations of different types at 1174, 1222 and 1391 cm-1 are in the fingerprint region in both MB and CRC+MB spectra proving the adsorption of the dye.
EP
The SEM micrographs show the modification of the surface of the raw material (Fig. 1c). More porosity in CRC can be noticed (Fig. 1d). The specific surface area of
AC C
CRC given by BET model is 2.4 m2 g-1 and the volume of pore is 5.7 10-3 cm3 g-1 with an average pore diameter of 9.48 nm. A very low specific surface area was also found with other lignin materials [30] and sawdust [31]. The XRD pattern (Fig. 1e) depicts that CRC is amorphous in nature and the broad peaks indicate that CRC particles are small. 3.2 Effect of pH
9
ACCEPTED MANUSCRIPT In the adsorption process, pH is an important factor. Fig. 2a represents the influence of initial pH on the amount of MB adsorbed per unit mass of the adsorbent. The variation between final and initial pH was also plotted versus the initial pH to deduce graphically the pHZPC (6.98). When the solution is very acidic, the adsorption process is
RI PT
inhibited. At the pH values corresponding to that of initial MB solutions, the amount of MB adsorbed increases significantly. The basic pH zone is optimal and increases a little more of the amount of MB adsorbed. The pHZPC of an adsorbent is a fundamental
SC
characteristic to comprehend interfacial properties. The negative electrostatic forces may be favourable to the adsorption (change of surface charges with cationic dye MB
M AN U
and CRC). At pH values below pHZPC, MB uptake was low because of the presence of H+ ions that compete with cationic MB for similar adsorption sites. At pH > pHZPC the adsorption process is favourable and the amount of dye adsorbed is higher. 3.3 Adsorption kinetics
TE D
Three models were applied to investigate the adsorption kinetic processes of MB on the CRC. The Lagergren’s pseudo-first order model, the Ho’s pseudo-second order model [32] and Weber and Morris [33] intra particle model are used and expressed in
EP
Eqs. (3), (4) and (5) respectively. Q = Q (1 − e )
AC C
(3)
Q = !
(4)
Q = k # √t + c
(5)
where Qt (mg g-1) is the adsorption capacity at the time t (min). k1 (min-1), k2 (g mg-1 min-1) and k3 (mg g -1 min-0.5) are the dynamics constants and c (mg g -1) is a constant. Both pseudo-first and pseudo-second order models suit the experimental data as can be seen in Figs. 2b and 2c, respectively. The kinetics parameters for pseudo-first order are Qe = 242 mg g-1 and k1 = 0.25 min-1 and those for pseudo-second order are Qe = 251 10
ACCEPTED MANUSCRIPT mg g-1 and k2 = 0.0025 g mg-1 min-1. According to the values of R2 of 0.991 and 0.997 for pseudo-first and pseudo-second respectively, the pseudo-second order describes the kinetics of the reaction is slightly better. The R2 value (0.970) obtained for intra particle diffusion (Fig. 2d) indicated that the
RI PT
adsorption kinetics was controlled by more than one process. Although intra particle model does not suit entirely the experimental data, the linear portion of curve presented in Fig. 2d, from 5 to 25 min, allows the determination of the parameters (k3 = 20.5 mg
SC
g-1 min-0.5 and c = 141 mg g-1). Intra particle model illustrated the gradual slower
3.4 Adsorption isotherms
M AN U
adsorption when the adsorbate slowly diffused to the interior surface of the particles.
The equilibrium adsorption isotherm is of importance in the design of adsorption systems. Several isotherm equations are available and three important isotherm models were selected in this study. Freundlich, Langmuir and Redlich-Peterson models are
Q = K ) C Q = Q =
,-. /0
!/0 /1 3
!2
EP
+
TE D
given in Eqs. (6), (7) and (8), respectively.
(6) (7) (8)
AC C
where Qmax (mg g-1) is a single-layer maximum adsorption capacity. The constants are n, KF, KL, KR and α, while β index ranges between 0 and 1. The fitting of Freundlich, Langmuir and Redlich-Peterson models with experimental
data at 25 °C is shown at Fig. 2e. Table 2 lists the correlation parameters and fitting factors for all three adsorption isotherm models from 25 to 45 °C and Fig. 2f displays isotherm data for temperature in the range of 25 to 45 °C. The Redlich Peterson model is the best model for the simulation of the adsorption isotherm as evident from the
11
ACCEPTED MANUSCRIPT coefficient of determination (R2) values. The maximum capacity according to Langmuir model was 565 and 572 mg g-1 at 25 and 45 °C, respectively. The KF values of MB adsorption onto CRC increase from 372 to 412 mg g-1 (L mg-1) 1/n at 25 and 45 °C, respectively. Moreover, the difference between the values of R2 of Langmuir and
RI PT
Freundlich decreases with the increase in adsorption temperature. At 45 °C Freundlich model fits the isotherm data better than Langmuir model. Therefore, multilayer adsorption with a stage of intra particle diffusion on heterogeneous surface explains the
SC
increase in the amount of MB adsorbed with the increase in adsorption temperature observed as shown in Fig. 2f. The potential exhibited by the CRC for wastewater
M AN U
decolourization is more than that of many naturally available adsorbents (111, 218 and 313 mg g-1 for cucumber peel, palm kernel fibre and rice husk respectively) [34–36]. 3.5 Thermodynamic study
The thermodynamic parameters measured to establish the adsorption process
TE D
included changes in standard enthalpy (∆H ° ), standard entropy (∆S ° ), and Gibbs energy ( ∆G° ) from the transfer of unit mole solute from the solution to the solid–liquid interface.
EP
∆G° was calculated using the equation below: ∆G° = −RTlnK =
AC C
(9)
where R is the universal gas constant (8.314 J mol-1), T (K) is the absolute temperature of the solution, and Kd is the distribution coefficient, which is computed as: >? =
@A
(10)
BA
CD>? =
∆E ° F
−
∆G °
(11)
FH
The values of lnK = are plotted against 1/T with the values of ∆H ° and ∆S ° calculated from the slope and intercept of the plot. The negative ∆G° values obtained at all
12
ACCEPTED MANUSCRIPT temperatures (-4.86; -5.02; -5.47; -6.00 and -6.24 kJ mol-1 at 298.15; 303.15; 308.15; 313.15 and 318.15 K respectively) studied indicated that the adsorption process is feasible and spontaneous. The positive ∆H ° value (17.4 kJ mol-1) indicates that the adsorption interaction is endothermic. In addition, the ∆H ° value less than 20
RI PT
kJ mol suggests that the adsorption is physical adsorption and reversible. Moreover, the positive value of ∆S° (0.074 kJ mol-1 K-1) indicates that there is an increase in the
CRC [37]. 3.6 Box-Behnken Response Surface Methodology
SC
randomness in the system of solid/solution interface during MB adsorption process by
M AN U
The polynomial regression equation was constructed using the BBRSM to establish the correlation analysis between the process variables and the response variable. The MB amount (Y) was found to vary from 397 to 779 mg g-1. The final empirical model is given by Eq. (12).
TE D
L = 596.70 + 87.21U − 32.51U + 28.92U# − 4.21U + 0.35U − 76.25U# − 57.11U U + 68.02U U# + 78.65U U#
(12)
The effectiveness of a model equation in predicting the experimental responses can
EP
be assessed based on the R2. The R2 was estimated as 0.992 proving that the model is
AC C
valid. Moreover, the adjusted R2 (Radj2 = 0.978) suggests extensive and very acceptable correlation between the independent variables. In addition, the predict R2 (Rpred2 = 0.875) indicates that the model will predict new observations as accurately as it fits the existing data. The coefficients of the model equation as well as the model parameters are listed in Table 3. The lower the value of p or the higher the value of F, the more significant are the model parameters. Table 3 shows that the model quadratic parameters U and U are insignificant (p value more than 0.05, confidence level of 95%). A test on the validity of the model is compulsory as it is used to optimize the process. 13
ACCEPTED MANUSCRIPT The appropriateness of the model, in addition to R2, is based on the ANOVA. Table 4 shows the results of ANOVA for the model. ANOVA is a statistical technique that subdivides the total variation in a set of data into component parts associated with specific sources of variation for testing hypotheses on the parameters of the model [38].
RI PT
The F value of 70 with p value of 0.000 of the model indicates its validity. From the ANOVA results, it can be concluded that the model predictions using Eq. (12) is satisfactory and that the model can be utilised to identify the optimum process
SC
conditions.
Referring to Table 4, temperature, mass of adsorbent and initial concentration show
M AN U
significant effects on the adsorption of MB. Fig. 3 shows a 3D response surface plot of mass of adsorbent and initial concentration on the amount of adsorbed MB. When the initial concentration is low and the mass of the adsorbent is high, the quantity of MB uptake is at the lowest value. Therefore, increase in the initial concentration and
TE D
decrease in the mass of adsorbent permit the amount of MB adsorbed to reach its maximum. As inferred earlier in thermodynamic study, the increase in temperature increases the amount of MB adsorbed.
EP
The optimum process conditions estimated using the optimizer tool in Minitab 17 are at temperature of 70 °C, mass of adsorbent of 15 mg, and dye concentration of 311
AC C
mg L-1, with the resultant amount of adsorbed MB of 771 mg g-1. The maximum response value under the optimized conditions was 787 mg g-1 (2.2% error).
4. Conclusions Efficient MB adsorption in aqueous solution was carried out by using CRC prepared by phosphoric acid activation, and its characterisation was done by FTIR/ATR, BET, SEM, and XRD. The adsorption kinetics follows a pseudo-second order and its isotherm
14
ACCEPTED MANUSCRIPT follows Redlich-Peterson’s model. The maximum capacity according to Langmuir model was 565 mg g-1 at 25 °C. The thermodynamic study reveals that the adsorption is endothermic, physical and random. The optimum process conditions estimated by the BBRSM are at a temperature of 70 °C, mass of adsorbent of 15 mg, and dye
RI PT
concentration of 311 mg L-1. According to the high performance exhibited by CRC in this study, cassava rind can be used as a better and low-cost biomass for wastewater decolourisation.
Sharma P, Kaur H, Sharma M, Sahore V. A review on applicability of naturally
M AN U
[1]
SC
References
available adsorbents for the removal of hazardous dyes from aqueous waste. Environ Monit Assess 2011;183:151–95. [2]
Ioannidou O, Zabaniotou A. Agricultural residues as precursors for activated carbon production – a review. Renew Sust Energ Rev 2007;11:1966–2005. Bhatnagar A, Sillanpää M, Witek-Krowiak A. Agricultural waste peels as
TE D
[3]
versatile biomass for water purification – a review. Chem Eng J 2015;270:244–
[4]
EP
71.
Singh K, Arora S. Removal of synthetic textile dyes from wastewaters: a critical
AC C
review on present treatment technologies. Crit Rev Env Sci Technol 2011;41:807–78.
[5]
Przystaś W, Zabłocka-Godlewska E, Grabińska-Sota E. Biological removal of azo and triphenylmethane dyes and toxicity of process by-products. Water Air Soil Poll 2012;223:1581–92.
[6]
Chung KT, Cerniglia CE. Mutagenicity of azo dyes: structure-activity relationships. Mutat Res-Rev Genet 1992;277:201–20.
[7]
El Hassani K, Beakou BH, Kalnina D, Oukani E, Anouar A. Effect of 15
ACCEPTED MANUSCRIPT morphological properties of layered double hydroxides on adsorption of azo dye Methyl Orange: a comparative study. Appl Clay Sci 2017;140:124–31. [8]
Robles I, Bustos E, Lakatos J. Adsorption study of mercury on lignite in the presence of different anions. Sustain Environ Res 2016;26:136–41. Khamparia S, Jaspal D. Adsorptive removal of Direct Red 81 dye from aqueous
RI PT
[9]
solution onto Argemone mexicana. Sustain Environ Res 2016;26:117–23.
[10] Geçgel Ü, Üner O, Gökara G, Bayrak Y. Adsorption of cationic dyes on activated
SC
carbon obtained from waste Elaeagnus stone. Adsorpt Sci Technol 2016;34:512– 25.
M AN U
[11] Asfaram A, Ghaedi M, Hajati S, Goudarzi A. Synthesis of magnetic γ-Fe2O3-based nanomaterial for ultrasonic assisted dyes adsorption: modeling and optimization. Ultrason Sonochem 2016;32:418–31.
[12] Dil EA, Ghaedi M, Asfaram A. The performance of nanorods material as
TE D
adsorbent for removal of azo dyes and heavy metal ions: application of ultrasound wave, optimization and modeling. Ultrason Sonochem 2017;34:792– 802.
EP
[13] Forgacs E, Cserháti T, Oros G. Removal of synthetic dyes from wastewaters: a review. Environ Int 2004;30:953–71.
AC C
[14] Gupta VK. Application of low-cost adsorbents for dye removal – a review. J Environ Manage 2009;90:2313–42.
[15] Liu SL, Wang YN, Lu KT. Preparation and pore characterization of activated carbon from Ma bamboo (Dendrocalamus latiflorus) by H3PO4 chemical activation. J Porous Mat 2014;21:459–66. [16] Lim WC, Srinivasakannan C, Balasubramanian N. Activation of palm shells by phosphoric acid impregnation for high yielding activated carbon. J Anal Appl
16
ACCEPTED MANUSCRIPT Pyrol 2010;88:181–6. [17] Allem AC. The origin of Manihot esculenta Crantz (Euphorbiaceae). Genet Resour Crop Ev 1994;41:133–50.
bioenergy. P Nalt Acad Sci USA. 2009;106:10219–23.
RI PT
[18] Gerbens-Leenes W, Hoekstra AY, van der Meer TH. The water footprint of
[19] Rajeshwarisivaraj, Sivakumar S, Senthilkumar P, Subburam V. Carbon from cassava peel, an agricultural waste, as an adsorbent in the removal of dyes and
SC
metal ions from aqueous solution. Bioresource Technol 2001;80:233–5.
[20] Geçgel Ü, Kocabıyık B, Üner O. Adsorptive removal of Methylene Blue from
M AN U
aqueous solution by the activated carbon obtained from the fruit of Catalpa bignonioides. Water Air Soil Poll 2015;226: 1–14.
[21] Ferreira SLC, Bruns RE, Ferreira HS, Matos GD, David JM, Brandão GC, et al. Box-Behnken design: an alternative for the optimization of analytical methods.
TE D
Anal Chim Acta 2007;597:179–86.
[22] Sampathkumar D, Punyamurthy R, Bennehalli B, Venkateshappa SC. Physical characterization of natural lignocellulosic single areca fiber. Ciência Tecnologia
EP
dos Materiais. 2015;27:121–35.
[23] Singh K, Risse M, Das KC, Worley J. Determination of composition of cellulose
AC C
and lignin mixtures using thermogravimetric analysis. J Energy Resour-ASME 2009;131:219–26.
[24] Monteiro SN, Calado V, Rodriguez RJ, Margem FM. Thermogravimetric behavior of natural fibers reinforced polymer composites – an overview. Mat Sci Eng A-Struct 2012;557:17–28. [25] Xu F, Yu J, Tesso T, Dowell F, Wang D. Qualitative and quantitative analysis of lignocellulosic biomass using infrared techniques: a mini-review. Appl Energ
17
ACCEPTED MANUSCRIPT 2013;104:801–9. [26] Sills DL, Gossett JM. Using FTIR spectroscopy to model alkaline pretreatment and enzymatic saccharification of six lignocellulosic biomasses. Biotechnol Bioeng 2012;109:894–903.
RI PT
[27] Schwanninger M, Rodrigues JC, Pereira H, Hinterstoisser B. Effects of shorttime vibratory ball milling on the shape of FT-IR spectra of wood and cellulose. Vib Spectrosc 2004;36:23–40.
SC
[28] Kubo S, Kadla JF. Hydrogen bonding in lignin: a Fourier transform infrared model compound study. Biomacromolecules 2005;6:2815–21.
M AN U
[29] Li Z, Wang CJ, Jiang WT. Intercalation of Methylene Blue in a high-charge calcium montmorillonite – an indication of surface charge determination. Adsorpt Sci Technol 2010;28:297–312.
[30] Abdelaziz OY, Hulteberg CP. Physicochemical characterisation of technical
TE D
lignins for their potential valorisation. Waste Biomass Valori 2017;8:859–69. [31] Salazar-Rabago JJ, Leyva-Ramos R, Rivera-Utrilla J, Ocampo-Perez R, CerinoCordova FJ. Biosorption mechanism of Methylene Blue from aqueous solution
EP
onto White Pine (Pinus durangensis) sawdust: effect of operating conditions. Sustain Environ Rese 2017;27:32–40.
AC C
[32] Ho YS, McKay G. Pseudo-second order model for sorption processes. Process Biochem 1999;34:451–65.
[33] Weber WJ, Morris, JC. Kinetics of adsorption on carbon from solution. J Sanit Eng Div 1963;89:31-59. [34] Akkaya G, Güzel F. Application of some domestic wastes as new low-cost biosorbents for removal of Methylene Blue: kinetic and equilibrium studies. Chem Eng Commun 2014;201:557–78.
18
ACCEPTED MANUSCRIPT [35] Ofomaja AE. Sorption dynamics and isotherm studies of Methylene Blue uptake on to palm kernel fibre. Chem Eng J 2007;126:35–43. [36] McKay G, Porter JF, Prasad GR. The removal of dye colours from aqueous solutions by adsorption on low-cost materials. Water Air Soil Poll
RI PT
1999;114:423–38.
[37] Liu Y. Is the free energy change of adsorption correctly calculated? J Chem Eng Data 2009;54:1981–5.
SC
[38] Li HP, Zhao GQ, Niu ST, Luan YG. Technologic parameter optimization of gas quenching process using response surface method. Comp Mater Sci
AC C
EP
TE D
M AN U
2007;38:561–70.
19
ACCEPTED MANUSCRIPT Table 1 Experimental data of BBRSM
Table 2
M AN U
SC
RI PT
Run X1 (°C) X2 (mg) X3 (mg L-1) Y (mg g-1) 1 20 15 300 491 2 70 15 300 779 3 20 25 300 521 4 70 25 300 581 5 20 20 200 460 6 70 20 200 499 7 20 20 400 397 8 70 20 400 708 9 45 15 200 601 10 45 25 200 398 11 45 15 400 486 12 45 25 400 598 13 45 20 300 598 14 45 20 300 598 15 45 20 300 594
Isotherms parameters and fitting factors for the adsorption of MB on CRC Parameters
Temperature (°C) 25 30 35 40 45 -1 -1 1/n Freundlich KF (mg g (L mg ) ) 372 382 391 397 412 n 10.2 10.7 10.4 9.9 10.7 0.987 0.987 0.991 0.995 0.997 R2 -1 Langmuir Qmax (mg g ) 565 565 574 579 572 KL (L mg-1) 0.66 0.80 0.98 1.26 2.10 R2 0.999 0.999 0.999 0.998 0.994 Redlich KR(mg g-1) 324 417 589 1107 2639 α 0.5 0.7 1.0 2.2 5.6 Peterson β 1.02 1.01 0.99 0.95 0.94 R2 0.999 0.999 0.999 0.999 0.999
AC C
EP
TE D
Models
20
ACCEPTED MANUSCRIPT Table 3
SC
Term Coefficient F p Constant 596.7 66.1 0.000 X1 87.2 15.7 0.000 X2 -32.5 -5.8 0.002 X3 28.9 5.2 0.003 X12 - 4.2 -0.5 0.627 2 X2 0.4 0.0 0.967 X32 - 76.3 -9.7 0.000 X1X2 - 57.1 -7.3 0.001 X1X3 68.0 8.7 0.000 X2X3 78.7 10.1 0.000
RI PT
Estimated coefficients using BBRSM (R2 = 0.992, R2 adj = 0.978 and Rpred2 = 0.875)
M AN U
Table 4 Analysis of variance (ANOVA) for BBRSM model
Sum of Mean F Degree of Freedom (DF) Squares (SS) Squares (MS) 9 153,899 17,100 70 Model Linear 3 75,989 25,330 104 Square 3 21,614 7,205 29 Interaction 3 56,295 18,765 77 5 1224 245 Error Lack of fit 3 1209 403 56 Pure error 2 15 7 14 155,123 Total
AC C
EP
TE D
Source
21
p 0.000 0.000 0.001 0.000 0.018 -
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 1. Characterization. (a) TGA curves of cassava rind biomass; (b) FTIR/ATR of raw material, CRC, MB and MB + CRC; (c) SEM image of raw material; (d) SEM image of CRC; (e) X Ray Diagram of CRC.
22
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 2. Adsorption experiments. (a) Influence of initial pH and pH of point of zero charge of CRC; (b) Fit to the experimental data by pseudo-first-order kinetic model; (c) Fit to the experimental data by pseudo-second-order kinetic model; (d) Fit to the linear portion by intraparticle kinetic model; (e) Fit to the experimental data by three isotherm models; (f) Isotherm experimental data for temperature from 25 to 45 °C.
23
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
Fig. 3. 3D surface plot of dose of adsorbent and concentration of dye on MB uptake (Y).
24
S2468-2039(17)30053-5
DOI:
10.1016/j.serj.2017.06.003
Reference:
SERJ 92
To appear in:
Sustainable Environment Research
Received Date: 22 February 2017 Revised Date:
10 April 2017
Accepted Date: 28 June 2017
Please cite this article as: Beakou BH, El Hassani K, Houssaini MA, Belbahloul M, Oukani E, Anouar A, Novel activated carbon from Manihot esculenta Crantz for removal of Methylene Blue, Sustainable Environment Research (2017), doi: 10.1016/j.serj.2017.06.003. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Received 23 February 2017 Received in revised form 10 April 2017 Accepted 28 June 2017
Blue
RI PT
Novel activated carbon from Manihot esculenta Crantz for removal of Methylene
SC
Buscotin Horax Beakou*, Kaoutar El Hassani, Mohammed Amine Houssaini,
M AN U
Mounir Belbahloul, Elhassan Oukani, Abdellah Anouar
Department of Applied Chemistry and Environment, University Hassan First, Settat 26000, Morocco
AC C
EP
TE D
Keywords: Adsorption; Manihot esculenta Crantz; Methylene Blue; Box Behnken
*Corresponding author Email: [email protected] 1
ACCEPTED MANUSCRIPT ABSTRACT The adsorptive removal of Methylene Blue by a novel bio char namely Cassava Rind Carbon was studied in a batch system. Moreover, several experiments were performed to optimize operating conditions of the adsorption process. Thermo
RI PT
gravimetric analysis investigates the thermal behaviour of the raw material. The surface functional groups were identified with Attenuated Total Reflectance Fourier Transform Infra-Red spectroscopy. The carbon surface was also characterised by BET surface area,
SC
scanning electron microscopy, X-ray diffraction and zero charge point of pH. Among the three kinetic models used to calculate the adsorption rate constants, the pseudo-
M AN U
second order was the most accurate. The experimental isotherms were analysed and Redlich-Peterson’s model provided the best fit. The maximum capacity according to Langmuir model was 565 mg g-1 at 25 °C. The thermodynamic study revealed that the adsorption is endothermic, physical and random. The temperature, the mass of
TE D
adsorbent and the concentration of dye were the three factors optimized in a Box Behnken Response Surface Methodology, with an optimal amount of adsorbed MB of
EP
771 mg g-1.
1. Introduction
AC C
Many adsorbents can be synthesized from waste materials of agricultural industry,
fruit waste and plant waste [1]. Agricultural residues as precursors for activated carbon production are the focus of many studies because they are sustainable adsorbents with high porosity and high reactivity [2]. Using agricultural waste peels as biomass for the production of low cost sorbents for water treatment applications removes an ecological burden for the society [3]. One of the applications of those lignocellulosic bio-sorbents is the removal of hazardous dyes from aqueous waste. Almost 700 kt of dyes are
2
ACCEPTED MANUSCRIPT produced worldwide each year. About half of this amount comes from textile industry. Among the 100,000 dyes available commercially, 70 wt% are azoic dyes [4]. Unfortunately, at least 10% of those synthetics dyes are released into the environment. They are one of the most hazardous species found in industrial effluents, so treatment is
RI PT
mandatory. They can reduce light penetration, precluding the photosynthesis of aqueous flora. In addition, some azoic dyes are able to cause allergic reactions and exhibit cytotoxicity [5]. They can show carcinogenic and mutagenic events [6].
SC
Different methods to remove dye such as adsorption on various sorbents (layered double hydroxides [7], lignite [8], natural adsorbent [9], activated carbon [10],
M AN U
nanomaterial [11,12]), chemical decomposition by oxidation, photo degradation and microbiological decolourisation [13]. All of them have advantages and drawbacks. Due to the high cost and disposal problems, many of these conventional methods for treating dye wastewater have not been widely applied in large scale in the textile and paper
TE D
industries. Low cost adsorbents may make adsorption the best solution for wastewater decolourisation [14]. However, the biomass needs to be activated to reach great performance. Despite its simplicity, physical activation generates higher amounts of
EP
volatile matter and requires more energy. On the contrary, chemical activation uses dehydrogenation reactions, which inhibit tar formation and reduce the volatile
AC C
substances generated [15]. Nevertheless, chemical activation requires an additional step and use of additional reagents. Among the various chemical reagents, phosphoric acid and zinc chloride have been used extensively for biomass precursors while potassium hydroxide is preferred for coal-based precursors. Moreover, phosphoric acid is preferred to zinc chloride due to its non-polluting nature [16]. The origin of cassava (Manihot esculenta Crantz) is well documented [17]. Taxonomically, it belongs to the Euphorbiaceae family and Manihot genus. Among the
3
ACCEPTED MANUSCRIPT crops, cassava has the lowest blue water footprint [18]. The biomass of our study, the thin outer layer, rough and brown on the outside of cassava root, is a non-valorised waste. This permits the reuse of the rest of the peel as better product for animal feed. There is no study using the so-called cassava rind as adsorbent. Therefore, in this work,
RI PT
the cassava rind was chemically activated with phosphoric acid (weight ratio 1:1) and carbonisation was done at 120 °C. Characterisation by Thermo Gravimetric Analysis (TGA),
Fourier
Transform
Infra-Red
through
Attenuated
Total
Reflectance
SC
(FTIR/ATR), Scanning Electron Microscopy (SEM), BET surface area and X-ray Diffraction (XRD) was performed. The determination of the point of zero charge of pH
M AN U
(pHZPC) and the pH influence were also carried out. Methylene Blue (MB), an azoic dye, was used as a model compound for evaluating the optimized adsorption capacity of synthesized bio char.
Adsorption mechanism of Cassava Rind Carbon (CRC) is highlighted through
TE D
adsorption isotherms, kinetics and thermodynamics. The efficiency of CRC for MB removal was optimized by altering three parameters namely temperature, mass of adsorbent and concentration of dye. The Box Behnken Surface Response Methodology
EP
allows the determination of the influence of those parameters.
AC C
2. Materials and methods 2.1 Carbon preparation
The cassava roots were purchased in a local farm in Djakotomey, Benin. The roots
were thoroughly washed with distilled water and peeled with knife. The rind was separated from the peel manually and dried in the sun until all the moisture had evaporated. Then a steel container was filled with the cassava rind and tightly closed. The product was ground and sieved to size between 63 and 90 µm. The activation
4
ACCEPTED MANUSCRIPT protocol [19] was used with some modifications as follows. A mixture of the raw material (10 g) and phosphoric acid (weight ratio 1:1) was heated at 120 °C for 14 h. Then the resulting char was ground and soaked in 200 mL of sodium hydroxide (1 M) for 24 h to remove residual acid and washed with distilled water until the pH was
RI PT
neutral (pH-meter pH-2005, J.P. Selecta, Spain). The reagents were of analytical grade (Sigma-Aldrich, USA). Finally, the carbon was dried at 60 °C for 6 h and sieved to size between 40 and 63 µm.
SC
2.2 Characterisation
The thermal behaviour of the raw material was studied using a LABSYS evo TGA
M AN U
1150 (Setaram Instrumentation, France). The TGA experiment was conducted under the following conditions: initial temperature 50 °C, final temperature 600 °C and heating rate of 10 °C min-1 with an Argon flow rate of 25 cm3 min-1.
A FTIR/ATR, Varian 800/gladiATR (Scimitar series, Australia/Pike Technologies,
TE D
USA), allows the determination of surface functional groups of four products (raw material, CRC, MB and CRC with MB loaded). The spectra were recorded in the range of 400–4,000 cm−1 with 50 scans collected at 4 cm−1 resolution.
EP
The BET equation was used to obtain the specific surface areas. Pore volume analysis was performed using the Barrett–Joyner–Halenda (BJH) method. Before each
AC C
measurement on QUADRASORB SI (Quantachrome, Germany), the samples were first degassed at 150 °C for 24 h. Then, the measurements were conducted through the adsorption of nitrogen at liquid nitrogen temperature onto the biopolymer surface. SEM analysis was performed on a Phenom Xl (Phenom-world, Netherland) with an accelerating voltage of 10 kV. CRC powder was spread on carbon tape adhered to a SEM stage. Before imaging, the samples were coated with a thin gold layer to improve the image quality.
5
ACCEPTED MANUSCRIPT XRD was also performed at room temperature under ambient air conditions, using Cu Kα1 radiation (λ = 1.5406 Å) at 10 mA, 30 kV and a 2θ angle ranging from 2 to 90° (PHASER D2 Diffractometer, Bruker, Billerica, Massachusetts, USA). 2.3 Adsorption measurements
RI PT
A 1,000 mg L-1 stock solution was prepared by dissolving MB dye (Riedel-de Haen, Germany) in distilled water. The different concentrations of MB were prepared from this stock solution. The solutions used for pH adjustment, NaOH 0.1 M and HCl 0.1 M,
SC
were from their corresponding analytical grade solutions (Sigma-Aldrich, USA). Batch adsorption experiments were done at 300 RPM. The supernatants were separated by
M AN U
centrifugation at 7,500 RPM for 2 min (Sigma Laborzentrifugen, Germany) and collected by using disposable syringes. The MB residual concentrations in the solutions were determined by monitoring the absorbance with a UV–Vis spectrophotometer DR 6000 with RFID Technology (Hach Lange, Germany) at 663 nm. The amount of MB
using Eq. (1). Q =
( )
TE D
adsorbed per unit mass of the adsorbent at equilibrium, Qe (mg g-1), was calculated
(1)
EP
where Ci and Ce are the initial and the equilibrium concentration of MB (mg L-1)
AC C
respectively, V is the volume of solution (L) and W is the mass of adsorbent (g). To determine the effect of pH, the adsorption of dye molecules was investigated
over a pH range of 3 to 11. The activated carbon (25 mg) was added into the dye solution (50 mL, 200 mg L-1) and mixing was carried out at room temperature (25 ± 1 °C) for 3 h before centrifugation. Moreover, in the same range of pH, the pHZPC was measured as follows: 50 mL of KNO3 solution (0.1 M) with 0.1 g of CRC for 48 h of agitation at room temperature (25 ± 1 °C) [20].
6
ACCEPTED MANUSCRIPT Adsorption kinetic experiment was carried out on 1000 mL dye solution (50 mg L-1, pH = 10) mixed with 0.2 g of activated carbon at room temperature (25 ± 1 °C) for 3 h. Several millilitres of solution were sampled using disposable syringes at various time intervals.
RI PT
Adsorption isotherms were determined by measuring the depletion of MB concentration after the adsorption reached equilibrium. For these experiments, CRC (25 mg) was added into dye solutions (50 mL) of different concentrations from 200 to 350
SC
mg L-1 with a pH value of 10. The mixing was carried out, in a double-jacketed beaker in which temperature was maintained with a thermostat (J.P. Selecta Frigiderm, Spain)
were 25, 30, 35, 40, 45 °C.
M AN U
for 3 h before centrifugation. The different temperature values used in the experiment
2.4 Box-Behnken Response Surface Methodology (BBRSM)
A Response Surface Methodology approach is a statistical method that uses
TE D
experimental data obtained from specified experimental design to model and optimize any process in which response of interest is influenced by several variables. It is intensively used to optimize dyes adsorption. Box-Behnken experimental design is a
EP
standard statistical tool mostly used for process optimization with minimum number of experiments [21]. The process variables were the adsorption temperature (X1), dose of
AC C
adsorbent (X2), and initial dye concentration (X3), while the response variable was the amount of MB adsorbed per unit mass of the adsorbent (Y). A minimum number of 15 experiments is recommended by the Box-Behnken design to optimize the process parameters. Among these 15 experiments, the last three runs were carried out under identical conditions. The upper limit, central point and lower limit of the variables coded “-1”, “0”, and “1” are (20, 45, 70 °C), (15, 20, 25 mg) and (200, 300, 400 mg L-1) respectively. Conditions and results of the experiment are listed in Table 1. The central
7
ACCEPTED MANUSCRIPT points, the upper and lower limits were fixed based on the potential observed on isotherm 45 °C results as well as on the preliminary experimental runs. The BBRSM polynomial equation used to model the amount of adsorbed MB (Y) from the water is
Y = β + ∑ β X ∑ β X X + ∑ β X
RI PT
represented in Eq. (2).
(2)
SC
where Y is the predicted response, β0 is a constant, βi is the linear coefficient, βij is the interaction coefficients, βii is the quadratic coefficients, and Xi and Xj are the coded
M AN U
values of the process variables. The results of experiments were analysed using statistical computing software Minitab 17 utilising the model equation and the analysis of variance (ANOVA).
3.1 Characterization
TE D
3. Results and discussion
The TGA curves of the raw material in (Fig. 1a) show two types of weight loss. The
EP
first one occurs at 100 °C and corresponds to the dehydration of the material. The
AC C
maximum depletion of mass (40%) is around 300 °C. The degradation of hemicellulose happens in the range 241–297 °C, while α-cellulose and lignin in the fibre degrades at 297–353 °C. The thermal decomposition of lignin occurs within 353–500 °C [22]. At 350 °C most of a lignin-cellulose mixture disappears [23] and after 400 °C total degradation is completed [24]. Infrared techniques are fast, accurate, and low-cost for analysis of functional groups in biomass [25]. FTIR/ATR allows attenuation of the incident radiation and provides infrared spectra without the water background absorbance. In Fig. 1b, the spectra after
8
ACCEPTED MANUSCRIPT activation have less adsorption bands than those of the raw material, indicating that various functional groups present in the raw material spectra disappear after activation. It can be noticed that C=O stretch of aldehyde of hemicellulose in the raw material at 1732 cm-1 disappears It is also observed in the raw material spectrum at 1156 cm−1 that
RI PT
there is C-O-C asymmetrical stretching of hemicellulose and cellulose. In the same spectrum at 1031 cm-1 C-O, C=C, and C-C-O stretching can be attributed to C-O-C asymmetrical stretching of hemicellulose, cellulose, and lignin [26]. In the CRC spectra,
SC
only cellulose and lignin bands were observed, indicating that hemicellulose content was lost during activation. At 980 cm-1, C-O valence vibration of cellulose can be
M AN U
observed [27]. In the raw material as well as CRC+MB (the activated carbon loaded with dye) spectra at 2917 and 2850 cm−1 there is C-H stretching of lignin [28]. At 1593 cm-1, it can be observed that there is a strong intensity band in both MB and CRC+MB spectra, corresponding to the skeletal stretching vibrations of the C=C and C=N bonds
TE D
in the heterocycle of MB [29]. More bands of the C-H bending vibrations of different types at 1174, 1222 and 1391 cm-1 are in the fingerprint region in both MB and CRC+MB spectra proving the adsorption of the dye.
EP
The SEM micrographs show the modification of the surface of the raw material (Fig. 1c). More porosity in CRC can be noticed (Fig. 1d). The specific surface area of
AC C
CRC given by BET model is 2.4 m2 g-1 and the volume of pore is 5.7 10-3 cm3 g-1 with an average pore diameter of 9.48 nm. A very low specific surface area was also found with other lignin materials [30] and sawdust [31]. The XRD pattern (Fig. 1e) depicts that CRC is amorphous in nature and the broad peaks indicate that CRC particles are small. 3.2 Effect of pH
9
ACCEPTED MANUSCRIPT In the adsorption process, pH is an important factor. Fig. 2a represents the influence of initial pH on the amount of MB adsorbed per unit mass of the adsorbent. The variation between final and initial pH was also plotted versus the initial pH to deduce graphically the pHZPC (6.98). When the solution is very acidic, the adsorption process is
RI PT
inhibited. At the pH values corresponding to that of initial MB solutions, the amount of MB adsorbed increases significantly. The basic pH zone is optimal and increases a little more of the amount of MB adsorbed. The pHZPC of an adsorbent is a fundamental
SC
characteristic to comprehend interfacial properties. The negative electrostatic forces may be favourable to the adsorption (change of surface charges with cationic dye MB
M AN U
and CRC). At pH values below pHZPC, MB uptake was low because of the presence of H+ ions that compete with cationic MB for similar adsorption sites. At pH > pHZPC the adsorption process is favourable and the amount of dye adsorbed is higher. 3.3 Adsorption kinetics
TE D
Three models were applied to investigate the adsorption kinetic processes of MB on the CRC. The Lagergren’s pseudo-first order model, the Ho’s pseudo-second order model [32] and Weber and Morris [33] intra particle model are used and expressed in
EP
Eqs. (3), (4) and (5) respectively. Q = Q (1 − e )
AC C
(3)
Q = !
(4)
Q = k # √t + c
(5)
where Qt (mg g-1) is the adsorption capacity at the time t (min). k1 (min-1), k2 (g mg-1 min-1) and k3 (mg g -1 min-0.5) are the dynamics constants and c (mg g -1) is a constant. Both pseudo-first and pseudo-second order models suit the experimental data as can be seen in Figs. 2b and 2c, respectively. The kinetics parameters for pseudo-first order are Qe = 242 mg g-1 and k1 = 0.25 min-1 and those for pseudo-second order are Qe = 251 10
ACCEPTED MANUSCRIPT mg g-1 and k2 = 0.0025 g mg-1 min-1. According to the values of R2 of 0.991 and 0.997 for pseudo-first and pseudo-second respectively, the pseudo-second order describes the kinetics of the reaction is slightly better. The R2 value (0.970) obtained for intra particle diffusion (Fig. 2d) indicated that the
RI PT
adsorption kinetics was controlled by more than one process. Although intra particle model does not suit entirely the experimental data, the linear portion of curve presented in Fig. 2d, from 5 to 25 min, allows the determination of the parameters (k3 = 20.5 mg
SC
g-1 min-0.5 and c = 141 mg g-1). Intra particle model illustrated the gradual slower
3.4 Adsorption isotherms
M AN U
adsorption when the adsorbate slowly diffused to the interior surface of the particles.
The equilibrium adsorption isotherm is of importance in the design of adsorption systems. Several isotherm equations are available and three important isotherm models were selected in this study. Freundlich, Langmuir and Redlich-Peterson models are
Q = K ) C Q = Q =
,-. /0
!/0 /1 3
!2
EP
+
TE D
given in Eqs. (6), (7) and (8), respectively.
(6) (7) (8)
AC C
where Qmax (mg g-1) is a single-layer maximum adsorption capacity. The constants are n, KF, KL, KR and α, while β index ranges between 0 and 1. The fitting of Freundlich, Langmuir and Redlich-Peterson models with experimental
data at 25 °C is shown at Fig. 2e. Table 2 lists the correlation parameters and fitting factors for all three adsorption isotherm models from 25 to 45 °C and Fig. 2f displays isotherm data for temperature in the range of 25 to 45 °C. The Redlich Peterson model is the best model for the simulation of the adsorption isotherm as evident from the
11
ACCEPTED MANUSCRIPT coefficient of determination (R2) values. The maximum capacity according to Langmuir model was 565 and 572 mg g-1 at 25 and 45 °C, respectively. The KF values of MB adsorption onto CRC increase from 372 to 412 mg g-1 (L mg-1) 1/n at 25 and 45 °C, respectively. Moreover, the difference between the values of R2 of Langmuir and
RI PT
Freundlich decreases with the increase in adsorption temperature. At 45 °C Freundlich model fits the isotherm data better than Langmuir model. Therefore, multilayer adsorption with a stage of intra particle diffusion on heterogeneous surface explains the
SC
increase in the amount of MB adsorbed with the increase in adsorption temperature observed as shown in Fig. 2f. The potential exhibited by the CRC for wastewater
M AN U
decolourization is more than that of many naturally available adsorbents (111, 218 and 313 mg g-1 for cucumber peel, palm kernel fibre and rice husk respectively) [34–36]. 3.5 Thermodynamic study
The thermodynamic parameters measured to establish the adsorption process
TE D
included changes in standard enthalpy (∆H ° ), standard entropy (∆S ° ), and Gibbs energy ( ∆G° ) from the transfer of unit mole solute from the solution to the solid–liquid interface.
EP
∆G° was calculated using the equation below: ∆G° = −RTlnK =
AC C
(9)
where R is the universal gas constant (8.314 J mol-1), T (K) is the absolute temperature of the solution, and Kd is the distribution coefficient, which is computed as: >? =
@A
(10)
BA
CD>? =
∆E ° F
−
∆G °
(11)
FH
The values of lnK = are plotted against 1/T with the values of ∆H ° and ∆S ° calculated from the slope and intercept of the plot. The negative ∆G° values obtained at all
12
ACCEPTED MANUSCRIPT temperatures (-4.86; -5.02; -5.47; -6.00 and -6.24 kJ mol-1 at 298.15; 303.15; 308.15; 313.15 and 318.15 K respectively) studied indicated that the adsorption process is feasible and spontaneous. The positive ∆H ° value (17.4 kJ mol-1) indicates that the adsorption interaction is endothermic. In addition, the ∆H ° value less than 20
RI PT
kJ mol suggests that the adsorption is physical adsorption and reversible. Moreover, the positive value of ∆S° (0.074 kJ mol-1 K-1) indicates that there is an increase in the
CRC [37]. 3.6 Box-Behnken Response Surface Methodology
SC
randomness in the system of solid/solution interface during MB adsorption process by
M AN U
The polynomial regression equation was constructed using the BBRSM to establish the correlation analysis between the process variables and the response variable. The MB amount (Y) was found to vary from 397 to 779 mg g-1. The final empirical model is given by Eq. (12).
TE D
L = 596.70 + 87.21U − 32.51U + 28.92U# − 4.21U + 0.35U − 76.25U# − 57.11U U + 68.02U U# + 78.65U U#
(12)
The effectiveness of a model equation in predicting the experimental responses can
EP
be assessed based on the R2. The R2 was estimated as 0.992 proving that the model is
AC C
valid. Moreover, the adjusted R2 (Radj2 = 0.978) suggests extensive and very acceptable correlation between the independent variables. In addition, the predict R2 (Rpred2 = 0.875) indicates that the model will predict new observations as accurately as it fits the existing data. The coefficients of the model equation as well as the model parameters are listed in Table 3. The lower the value of p or the higher the value of F, the more significant are the model parameters. Table 3 shows that the model quadratic parameters U and U are insignificant (p value more than 0.05, confidence level of 95%). A test on the validity of the model is compulsory as it is used to optimize the process. 13
ACCEPTED MANUSCRIPT The appropriateness of the model, in addition to R2, is based on the ANOVA. Table 4 shows the results of ANOVA for the model. ANOVA is a statistical technique that subdivides the total variation in a set of data into component parts associated with specific sources of variation for testing hypotheses on the parameters of the model [38].
RI PT
The F value of 70 with p value of 0.000 of the model indicates its validity. From the ANOVA results, it can be concluded that the model predictions using Eq. (12) is satisfactory and that the model can be utilised to identify the optimum process
SC
conditions.
Referring to Table 4, temperature, mass of adsorbent and initial concentration show
M AN U
significant effects on the adsorption of MB. Fig. 3 shows a 3D response surface plot of mass of adsorbent and initial concentration on the amount of adsorbed MB. When the initial concentration is low and the mass of the adsorbent is high, the quantity of MB uptake is at the lowest value. Therefore, increase in the initial concentration and
TE D
decrease in the mass of adsorbent permit the amount of MB adsorbed to reach its maximum. As inferred earlier in thermodynamic study, the increase in temperature increases the amount of MB adsorbed.
EP
The optimum process conditions estimated using the optimizer tool in Minitab 17 are at temperature of 70 °C, mass of adsorbent of 15 mg, and dye concentration of 311
AC C
mg L-1, with the resultant amount of adsorbed MB of 771 mg g-1. The maximum response value under the optimized conditions was 787 mg g-1 (2.2% error).
4. Conclusions Efficient MB adsorption in aqueous solution was carried out by using CRC prepared by phosphoric acid activation, and its characterisation was done by FTIR/ATR, BET, SEM, and XRD. The adsorption kinetics follows a pseudo-second order and its isotherm
14
ACCEPTED MANUSCRIPT follows Redlich-Peterson’s model. The maximum capacity according to Langmuir model was 565 mg g-1 at 25 °C. The thermodynamic study reveals that the adsorption is endothermic, physical and random. The optimum process conditions estimated by the BBRSM are at a temperature of 70 °C, mass of adsorbent of 15 mg, and dye
RI PT
concentration of 311 mg L-1. According to the high performance exhibited by CRC in this study, cassava rind can be used as a better and low-cost biomass for wastewater decolourisation.
Sharma P, Kaur H, Sharma M, Sahore V. A review on applicability of naturally
M AN U
[1]
SC
References
available adsorbents for the removal of hazardous dyes from aqueous waste. Environ Monit Assess 2011;183:151–95. [2]
Ioannidou O, Zabaniotou A. Agricultural residues as precursors for activated carbon production – a review. Renew Sust Energ Rev 2007;11:1966–2005. Bhatnagar A, Sillanpää M, Witek-Krowiak A. Agricultural waste peels as
TE D
[3]
versatile biomass for water purification – a review. Chem Eng J 2015;270:244–
[4]
EP
71.
Singh K, Arora S. Removal of synthetic textile dyes from wastewaters: a critical
AC C
review on present treatment technologies. Crit Rev Env Sci Technol 2011;41:807–78.
[5]
Przystaś W, Zabłocka-Godlewska E, Grabińska-Sota E. Biological removal of azo and triphenylmethane dyes and toxicity of process by-products. Water Air Soil Poll 2012;223:1581–92.
[6]
Chung KT, Cerniglia CE. Mutagenicity of azo dyes: structure-activity relationships. Mutat Res-Rev Genet 1992;277:201–20.
[7]
El Hassani K, Beakou BH, Kalnina D, Oukani E, Anouar A. Effect of 15
ACCEPTED MANUSCRIPT morphological properties of layered double hydroxides on adsorption of azo dye Methyl Orange: a comparative study. Appl Clay Sci 2017;140:124–31. [8]
Robles I, Bustos E, Lakatos J. Adsorption study of mercury on lignite in the presence of different anions. Sustain Environ Res 2016;26:136–41. Khamparia S, Jaspal D. Adsorptive removal of Direct Red 81 dye from aqueous
RI PT
[9]
solution onto Argemone mexicana. Sustain Environ Res 2016;26:117–23.
[10] Geçgel Ü, Üner O, Gökara G, Bayrak Y. Adsorption of cationic dyes on activated
SC
carbon obtained from waste Elaeagnus stone. Adsorpt Sci Technol 2016;34:512– 25.
M AN U
[11] Asfaram A, Ghaedi M, Hajati S, Goudarzi A. Synthesis of magnetic γ-Fe2O3-based nanomaterial for ultrasonic assisted dyes adsorption: modeling and optimization. Ultrason Sonochem 2016;32:418–31.
[12] Dil EA, Ghaedi M, Asfaram A. The performance of nanorods material as
TE D
adsorbent for removal of azo dyes and heavy metal ions: application of ultrasound wave, optimization and modeling. Ultrason Sonochem 2017;34:792– 802.
EP
[13] Forgacs E, Cserháti T, Oros G. Removal of synthetic dyes from wastewaters: a review. Environ Int 2004;30:953–71.
AC C
[14] Gupta VK. Application of low-cost adsorbents for dye removal – a review. J Environ Manage 2009;90:2313–42.
[15] Liu SL, Wang YN, Lu KT. Preparation and pore characterization of activated carbon from Ma bamboo (Dendrocalamus latiflorus) by H3PO4 chemical activation. J Porous Mat 2014;21:459–66. [16] Lim WC, Srinivasakannan C, Balasubramanian N. Activation of palm shells by phosphoric acid impregnation for high yielding activated carbon. J Anal Appl
16
ACCEPTED MANUSCRIPT Pyrol 2010;88:181–6. [17] Allem AC. The origin of Manihot esculenta Crantz (Euphorbiaceae). Genet Resour Crop Ev 1994;41:133–50.
bioenergy. P Nalt Acad Sci USA. 2009;106:10219–23.
RI PT
[18] Gerbens-Leenes W, Hoekstra AY, van der Meer TH. The water footprint of
[19] Rajeshwarisivaraj, Sivakumar S, Senthilkumar P, Subburam V. Carbon from cassava peel, an agricultural waste, as an adsorbent in the removal of dyes and
SC
metal ions from aqueous solution. Bioresource Technol 2001;80:233–5.
[20] Geçgel Ü, Kocabıyık B, Üner O. Adsorptive removal of Methylene Blue from
M AN U
aqueous solution by the activated carbon obtained from the fruit of Catalpa bignonioides. Water Air Soil Poll 2015;226: 1–14.
[21] Ferreira SLC, Bruns RE, Ferreira HS, Matos GD, David JM, Brandão GC, et al. Box-Behnken design: an alternative for the optimization of analytical methods.
TE D
Anal Chim Acta 2007;597:179–86.
[22] Sampathkumar D, Punyamurthy R, Bennehalli B, Venkateshappa SC. Physical characterization of natural lignocellulosic single areca fiber. Ciência Tecnologia
EP
dos Materiais. 2015;27:121–35.
[23] Singh K, Risse M, Das KC, Worley J. Determination of composition of cellulose
AC C
and lignin mixtures using thermogravimetric analysis. J Energy Resour-ASME 2009;131:219–26.
[24] Monteiro SN, Calado V, Rodriguez RJ, Margem FM. Thermogravimetric behavior of natural fibers reinforced polymer composites – an overview. Mat Sci Eng A-Struct 2012;557:17–28. [25] Xu F, Yu J, Tesso T, Dowell F, Wang D. Qualitative and quantitative analysis of lignocellulosic biomass using infrared techniques: a mini-review. Appl Energ
17
ACCEPTED MANUSCRIPT 2013;104:801–9. [26] Sills DL, Gossett JM. Using FTIR spectroscopy to model alkaline pretreatment and enzymatic saccharification of six lignocellulosic biomasses. Biotechnol Bioeng 2012;109:894–903.
RI PT
[27] Schwanninger M, Rodrigues JC, Pereira H, Hinterstoisser B. Effects of shorttime vibratory ball milling on the shape of FT-IR spectra of wood and cellulose. Vib Spectrosc 2004;36:23–40.
SC
[28] Kubo S, Kadla JF. Hydrogen bonding in lignin: a Fourier transform infrared model compound study. Biomacromolecules 2005;6:2815–21.
M AN U
[29] Li Z, Wang CJ, Jiang WT. Intercalation of Methylene Blue in a high-charge calcium montmorillonite – an indication of surface charge determination. Adsorpt Sci Technol 2010;28:297–312.
[30] Abdelaziz OY, Hulteberg CP. Physicochemical characterisation of technical
TE D
lignins for their potential valorisation. Waste Biomass Valori 2017;8:859–69. [31] Salazar-Rabago JJ, Leyva-Ramos R, Rivera-Utrilla J, Ocampo-Perez R, CerinoCordova FJ. Biosorption mechanism of Methylene Blue from aqueous solution
EP
onto White Pine (Pinus durangensis) sawdust: effect of operating conditions. Sustain Environ Rese 2017;27:32–40.
AC C
[32] Ho YS, McKay G. Pseudo-second order model for sorption processes. Process Biochem 1999;34:451–65.
[33] Weber WJ, Morris, JC. Kinetics of adsorption on carbon from solution. J Sanit Eng Div 1963;89:31-59. [34] Akkaya G, Güzel F. Application of some domestic wastes as new low-cost biosorbents for removal of Methylene Blue: kinetic and equilibrium studies. Chem Eng Commun 2014;201:557–78.
18
ACCEPTED MANUSCRIPT [35] Ofomaja AE. Sorption dynamics and isotherm studies of Methylene Blue uptake on to palm kernel fibre. Chem Eng J 2007;126:35–43. [36] McKay G, Porter JF, Prasad GR. The removal of dye colours from aqueous solutions by adsorption on low-cost materials. Water Air Soil Poll
RI PT
1999;114:423–38.
[37] Liu Y. Is the free energy change of adsorption correctly calculated? J Chem Eng Data 2009;54:1981–5.
SC
[38] Li HP, Zhao GQ, Niu ST, Luan YG. Technologic parameter optimization of gas quenching process using response surface method. Comp Mater Sci
AC C
EP
TE D
M AN U
2007;38:561–70.
19
ACCEPTED MANUSCRIPT Table 1 Experimental data of BBRSM
Table 2
M AN U
SC
RI PT
Run X1 (°C) X2 (mg) X3 (mg L-1) Y (mg g-1) 1 20 15 300 491 2 70 15 300 779 3 20 25 300 521 4 70 25 300 581 5 20 20 200 460 6 70 20 200 499 7 20 20 400 397 8 70 20 400 708 9 45 15 200 601 10 45 25 200 398 11 45 15 400 486 12 45 25 400 598 13 45 20 300 598 14 45 20 300 598 15 45 20 300 594
Isotherms parameters and fitting factors for the adsorption of MB on CRC Parameters
Temperature (°C) 25 30 35 40 45 -1 -1 1/n Freundlich KF (mg g (L mg ) ) 372 382 391 397 412 n 10.2 10.7 10.4 9.9 10.7 0.987 0.987 0.991 0.995 0.997 R2 -1 Langmuir Qmax (mg g ) 565 565 574 579 572 KL (L mg-1) 0.66 0.80 0.98 1.26 2.10 R2 0.999 0.999 0.999 0.998 0.994 Redlich KR(mg g-1) 324 417 589 1107 2639 α 0.5 0.7 1.0 2.2 5.6 Peterson β 1.02 1.01 0.99 0.95 0.94 R2 0.999 0.999 0.999 0.999 0.999
AC C
EP
TE D
Models
20
ACCEPTED MANUSCRIPT Table 3
SC
Term Coefficient F p Constant 596.7 66.1 0.000 X1 87.2 15.7 0.000 X2 -32.5 -5.8 0.002 X3 28.9 5.2 0.003 X12 - 4.2 -0.5 0.627 2 X2 0.4 0.0 0.967 X32 - 76.3 -9.7 0.000 X1X2 - 57.1 -7.3 0.001 X1X3 68.0 8.7 0.000 X2X3 78.7 10.1 0.000
RI PT
Estimated coefficients using BBRSM (R2 = 0.992, R2 adj = 0.978 and Rpred2 = 0.875)
M AN U
Table 4 Analysis of variance (ANOVA) for BBRSM model
Sum of Mean F Degree of Freedom (DF) Squares (SS) Squares (MS) 9 153,899 17,100 70 Model Linear 3 75,989 25,330 104 Square 3 21,614 7,205 29 Interaction 3 56,295 18,765 77 5 1224 245 Error Lack of fit 3 1209 403 56 Pure error 2 15 7 14 155,123 Total
AC C
EP
TE D
Source
21
p 0.000 0.000 0.001 0.000 0.018 -
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 1. Characterization. (a) TGA curves of cassava rind biomass; (b) FTIR/ATR of raw material, CRC, MB and MB + CRC; (c) SEM image of raw material; (d) SEM image of CRC; (e) X Ray Diagram of CRC.
22
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 2. Adsorption experiments. (a) Influence of initial pH and pH of point of zero charge of CRC; (b) Fit to the experimental data by pseudo-first-order kinetic model; (c) Fit to the experimental data by pseudo-second-order kinetic model; (d) Fit to the linear portion by intraparticle kinetic model; (e) Fit to the experimental data by three isotherm models; (f) Isotherm experimental data for temperature from 25 to 45 °C.
23
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
Fig. 3. 3D surface plot of dose of adsorbent and concentration of dye on MB uptake (Y).
24