Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy journal homepage: www.elsevier.com/locate/saa
Removal of methylene blue from aqueous solution by wood millet carbon optimization using response surface methodology Mehrorang Ghaedi ⇑, Syamak Nasiri Kokhdan Chemistry Department, Yasouj University, Yasouj 75918-74831, Iran
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Low-cost and non-toxic walnut
carbon (WC) was easily made. Walnut carbon was used to remove
methylene blue from aqueous solution. The dye removal was performed in a very short time (2 min). Walnut carbon was re-used after heating. Optimization was performed using response surface methodology.
a r t i c l e
i n f o
Article history: Received 19 April 2014 Received in revised form 8 July 2014 Accepted 18 July 2014 Available online xxxx Keywords: Adsorption Methylene blue (MB) Millet household carbon (MHC) Kinetic
a b s t r a c t The use of cheep, non-toxic, safe and easily available adsorbent are efficient and recommended material and alternative to the current expensive substance for pollutant removal from wastewater. The activated carbon prepared from wood waste of local tree (millet) extensively was applied for quantitative removal of methylene blue (MB), while simply. It was used to re-used after heating and washing with alkaline solution of ethanol. This new adsorbent was characterized by using BET surface area measurement, FT-IR, pH determination at zero point of charge (pHZPC) and Boehm titration method. Response surface methodology (RSM) by at least the number of experiments main and interaction of experimental conditions such as pH of solution, contact time, initial dye concentration and adsorbent dosage was optimized and set as pH 7, contact time 18 min, initial dye concentration 20 ppm and 0.2 g of adsorbent. It was found that variable such as pH and amount of adsorbent as solely or combination effects seriously affect the removal percentage. The fitting experimental data with conventional models reveal the applicability of isotherm models Langmuir model for their well presentation and description and Kinetic real rate of adsorption at most conditions efficiently can be represented pseudo-second order, and intra-particle diffusion. It novel material is good candidate for removal of huge amount of MB (20 ppm) in short time (18 min) by consumption of small amount (0.2 g). Ó 2014 Elsevier B.V. All rights reserved.
Introduction Main and significant environmental pollution source are highly color dyes suspended in organic solids or present in bulk solution ⇑ Corresponding author. Tel./fax: +98 741 2223048. E-mail address: [email protected] (M. Ghaedi).
[1]. Arrival and presence of these pollutants in water significantly damage aquatic life following generation of mutagenic and carcinogenic activity. Textile, paper and printing activities are some source of dye containing wastewater. Versatile (cheep and non toxic material are used for dyes removal present in various media [2–4]. It was found that complex (aromatic) structures dyes due to their high and remarkable stability toward degeneration (phot and
http://dx.doi.org/10.1016/j.saa.2014.07.048 1386-1425/Ó 2014 Elsevier B.V. All rights reserved.
Please cite this article in press as: M. Ghaedi, S.N. Kokhdan, Removal of methylene blue from aqueous solution by wood millet carbon optimization using response surface methodology, Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy (2014), http://dx.doi.org/10.1016/j.saa.2014.07.048
2
M. Ghaedi, S.N. Kokhdan / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy xxx (2014) xxx–xxx
Nomenclatures Ct Qe Ce V W k1 k2 H
a b C Kdif F Qm Ka
dye concentration (mg L1) at time t equilibrium adsorption capacity (mg g1) dye concentration (mg L1) at equilibrium volume of solution (L) weight of adsorbent (g) rate constant of pseudofirst order adsorption (min1) second-order rate constant of adsorption (mg g1 min1) second-order rate constants (mg g1 min1) initial adsorption rate (mg g1 min1) desorption constant (mg g1) intercept of intraparticle diffusion (related to the thickness of the boundary layer) rate constant of intraparticle diffusion (mg g1 min1/2) fraction of solute adsorbed at any time t (mg g1) maximum adsorption capacity reflected a complete monolayer (mg g1) in Langmuir isotherm model Langmuir constant or adsorption equilibrium constant (L mg1) that is related to the apparent energy of sorption
bio degradation), light, heat oxidizing agent and other approach can be considered as more and serious toxic agent for ecosystem. Dyes via direct destruction or inhibition of catalytic capabilities lead to depth in microorganism [5,6]. Among the well known dyes reduction pathways [7–27] more attention was devoted to adsorption that emerged from unique advantages such as highly porous and safe adsorbent. These material due to their unique properties are useful for dyes removal. Carbon simply can be obtained by burning and putting it into a sealed container with very cost effective and non-toxic manner without much energy consumption. All these characteristics of such carbon make it more suitable than commercial activated carbon for diverse application. Methylene Blue (MB) (Fig. 1) manly use for coloring paper, temporary hair colorant, dyeing cottons, wools, while reveal very harmful effects on living things [7–9]. Dyes contaminated ecosystem significantly can perturbed aquatic, plants and human lifes [10–27]. MB due to its chemical resistance to light and oxidizing agent hardly can be removal or eliminated by biological treatment and chemical precipitation [28]. The pestilent and dangerous intermediates/substances created by this dye after undergoing reduction and oxidation in water further increase the need for its removal from wastewater. Therefore, we were motivated to prepare carbon from millet that following subsequent activation is as an ideal MB removal agent. The full characterization of this AC by BET, FT-IR, pH determination at zero point of charge (pHZPC) and Boehm titration reveal and confirm its applicability to interact with large amount of material through various pathway and mechanism. The experimental runs and optimum conditions were achieved along at least experimental run by central composite design and response surface methodology (RSM). This methodology simply permits to estimate the main effect and interaction of variables through economic and repeatable pathway.
Fig. 1. Chemical structure of methylene blue green.
RL KF N T R B1 K Qm e E X2 qe,exp qe,calc R2
Dimensionless equilibrium parameter (separation factor) isotherm constant indicate the capacity parameter (mg g1) related to the intensity of the adsorption isotherm constant indicate the empirical parameter (g L1) related to the intensity of the adsorption absolute temperature in Kelvin universal gas constant (8.314 J K1 mol1) related to the heat of adsorption (B1 = RT/b) constant related to the adsorption energy at the D–R isotherm (mol2 kJ2) theoretical saturation capacity at the D–R isotherm Polanyi potential at the D–R isotherm mean free energy of adsorption chi-squared test statistic experimental data of the equilibrium capacity (mg g1) equilibrium capacity obtained by calculating from the isotherm model (mg g1) correlation coefficient
Subsequent objective of this study was investigation of kinetics and isotherms parameters correspond to MB adsorption onto AC. The whole results confirm the suitability and applicability of the Langmuir model for isothermal study elucidation while the real behavior of adsorption experimental data simply can be represented by combination of pseudo second order and interparticle diffusion model.
Experimental Materials and instrumentation The methylene blue with molecular formula C16H18N3SCl and IUPAC name 3,7-bis(Dimethylamino)-phenothiazin-5-ium chloride, (MW = 319.86 g/mol), with CAS Number of 61-73-4 (United States, Sigma–Aldrich) 100 mg/L as stock solution was supplied by dissolving 10 mg of MB in 100 mL double-distilled water. The effect of solution pH over 2–8 (adjusted via addition of dilute HCl and/or KOH solution) its adsorption removal was studied, while each pH maximum wave length was recorded. All the parameters and experimental data were studied using the software MINITAB (version 16.0).
Preparation of millet carbon The source material container was heated at 400 °C for 3 h and subsequent production of coal was then slowly cooled to room temperature and rinsed with distilled water. The carbon was grinded to mesh lower than 120 and subsequently was activated by various with mixture of concentrated HCl and HNO3 with 1:1 valuation. The produced AC was studied by SEM (see Fig. 2), BET and FT-IR. The pH corresponding to the AC point of zero charge (pHZPC) of was determined by the pH drift method [27,29,30]. Determination of oxygen containing functional groups was carried out as follow (Boehm titration method) [31]: 1.0 g of the AC mixed separately with 15 ml solution of NaHCO3 (0.1 M), Na2CO3 (0.05 M) and NaOH (0.1 M) for acidic groups and 0.1 M HCl for basic groups/sites respectively at room temperature for more than
Please cite this article in press as: M. Ghaedi, S.N. Kokhdan, Removal of methylene blue from aqueous solution by wood millet carbon optimization using response surface methodology, Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy (2014), http://dx.doi.org/10.1016/j.saa.2014.07.048
M. Ghaedi, S.N. Kokhdan / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy xxx (2014) xxx–xxx
3
Fig. 2. SEM of activated carbon prepared from wood millet carbon.
2 days. Subsequently, the aqueous solutions were back titrated with HCl (0.1 M) for acidic and NaOH (0.1 M) for basic groups. The number and type of acidic sites were calculated by considering that NaOH neutralizes carboxylic, lactonic and phenolic groups, Na2CO3 neutralizes carboxylic and lactonic groups and that NaHCO3 neutralizes only carboxylic groups. Carboxylic groups are calculated by direct titration with NaHCO3, while the difference between actual amount of Na2CO3 and NaHCO3 volume is used for estimation of lactones. The difference between the groups titrated with NaOH and Na2CO3 show the phenol content. Basic sites were determined by titration with HCl. Characterization of prepared adsorbent was performed and presented in Table 1 is usable for simultaneous combination with RSM optimization of variables. According to the analysis of response by ANOVA, the main effect and interaction of variable can be estimated.
Central composite design (CCD) and optimization of parameters Central composite design (CCD) as most accepted experimental designs methodology following. Application of CCD permit estimation of their and synergetic or antagonist effect factors including pH, amount of adsorbent, MB concentration and contact time on removal percentage. Thirty-one experiments based on well know procedure and formula were designed and run at different conditions (room temperature) [32–34]. The number of runs was obtained according the relation 2n + 2n + nc, where n is the number of factors (four factors), nc is the number of center points (seven replicates). The MB removal percentage (response) and its change with Coded and uncoded factor values are listed in Table 2. Grinded and small price of millet wood waste was totally rinsed with detergent. Subsequently, it was revealed that variable such as time
and temperature of heating and type concentration of concentrated acid are major factor. The optimal model predictor quadratic equation given as:
Y ¼ b0 þ
4 4 3 X 4 X X X b i xi þ bii x2i þ bij xi xj i¼1
i¼1
ð1Þ
i¼1 j¼iþ1
was used to optimize and predict the response (Y) against the parameters. Where xi, xj are the coded values of the factors and b0, bi, bii and bij are the constant, linear, quadratic and interaction coefficients respectively. Following the analysis of variance and according to, the coefficient of determination (R2), the probability p-value (95% confidence level) and fisher’s test, it is possible to take useful information about statistical significance and the characteristics of suitability of predicated mode 1 for predication of real behavior of adsorption system. Significant parameters were recognized and subsequently a predictive model as follow was developed.
Y ¼ 90:24 þ 30:48x2 17:63x3 29:39x22 10:41x23 þ 27:03x2 x3
ð2Þ
Positive coefficients mean that the corresponding terms affect the response positively and negative values affect it negatively. According to ANOVA result (Table 2) and well know role (P-values less than 0.05) confirm the significance of each parameter, while the P value of higher than 0.05 show high significance and more applicability of model for prediction and description of exact behavior of proposed system. The terms x2 ; x3 ; x22 ; x23 and x2x3 were found to be significant that suggest more contribution of parameters like the MB concentration, the amount of millet AC
Please cite this article in press as: M. Ghaedi, S.N. Kokhdan, Removal of methylene blue from aqueous solution by wood millet carbon optimization using response surface methodology, Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy (2014), http://dx.doi.org/10.1016/j.saa.2014.07.048
4
M. Ghaedi, S.N. Kokhdan / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy xxx (2014) xxx–xxx
conditions reveal the high repeat ability of method for prediction of real removal percentage with relative deviation less than 3%.
Table 1 Summary report of proposed sorbent. Summary report Surface area Single point surface area at p/p° = 0.206301245 BET surface area: Langmuir surface area: Pore volume Single point adsorption total pore volume of poresless than 1256.713 Å width at p/p° = 0.984352481: t-Plot micropore volume: BJH adsorption cumulative volume of pores between 17.000 Å and 3000.000 Å width: BJH desorption cumulative volume of pores between 17.000 Å and 3000.000 Å width: Pore size Adsorption average pore width (4V/A by BET): BJH adsorption average pore width (4V/A): BJH desorption average pore width (4V/A): pHZPC Acid soluble Water soluble Carboxylic(acidic functions) Phenol Lactones Basic sites
335.9 m2/g 30.8 m2/g 342.7 m2/g 0.041 cm3/g 0.026 cm3/g 0.037 cm3/g 0.026 cm3/g
35.9 Å 204.2 Å 205.7 Å 8.5 NO NO 0.805 mmol/g NO 0.418 mmol/g NO
affect the MB removal linear and in quadratic way and typical three dimensional response surface Fig. 3 confirm the presence of strong interaction among variable. The optimized values for pH, the amount of AC prepared from millet, MB concentration and contact time were found to be 6.5, 0.26 g, 7.1 mg/L and 30 min, respectively. At these conditions, the predicated MB removal percentage was more than 96% with desirability 1. Conduction of similar experiments at specified optimum
Results and discussion Characterization of carbon FT-IR Applications of powerful identification techniques like FT-IR (Fig. 4) make reveal the distinguished peaks that confirm presence of functional groups like OH, COOH, amine and carbonyl groups. Stretching vibration band around 1700 cm1 is assigned to carbonyl C@O group present in aldehyde, ester, ketone and acetyl derivatives. The strong band at 1500 cm1 may be due to C@C band. The peaks appearing between 423 cm1 and 634 cm1 are assigned to the metal–oxygen (MAO) stretching mode in structure of carbon. The broad band around 3000–3400 cm1 confirm OH group. The band around 2900 cm1 reveals the presence CH group in AC. Therefore, presence of these reactive sites and centers make candidate the present AC as good and suitable material for strong interaction with various material and suitable them for waste water treatment. BET analysis of activated carbon The N2/77K adsorption isotherms permit to study the porous structure of material and also to attain useful information about surface area its ability to adsorb a target compound [35]. Specific surface area (SSA), pore volume and accessible pore volume for each pores category a significant criterion make useful information about the ability of material for trapping molecules with distinct sizes. The interference by the surrounding phase is especially problematic for the Bruner–Emmet–Teller (BET) N2 adsorption/desorption isotherm method because the entire surface is by Table 1 and
Table 2 Coded and uncoded central composite design. Run
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Dye concentration (mg/L)
Time (min)
Uncoded
PH Coded
Uncoded
Coded
Uncoded
Coded
Adsorbent (g) Uncoded
Coded
Removal (%)
2 8 6 6 8 6 6 4 8 4 8 6 6 4 8 8 4 4 8 6 6 6 4 8 6 10 6 6 4 6 4
a +1 0 0 +1 0 0 1 +1 1 +1 0 0 1 +1 +1 1 1 +1 0 0 0 1 +1 0 +a 0 0 1 0 1
10 14 10 10 14 10 10 14 6 6 6 10 10 14 6 6 14 6 14 10 10 10 14 14 10 10 18 10 6 2 6
0 +1 0 0 +1 0 0 +1 1 1 1 0 0 +1 1 1 +1 1 +1 0 0 0 +1 +1 0 0 +a 0 1 a 1
20 11 20 20 29 20 20 29 29 29 11 20 20 11 11 29 11 11 11 20 2 38 29 29 20 20 20 20 29 20 11
0 1 0 0 +1 0 0 +1 +1 +1 1 0 0 1 1 +1 1 1 1 0 a +a +1 +1 0 0 0 0 +1 0 1
0.225 0.325 0.225 0.225 0.325 0.425 0.225 0.125 0.125 0.325 0.125 0.225 0.025 0.125 0.325 0.325 0.325 0.325 0.125 0.225 0.225 0.225 0.325 0.125 0.225 0.225 0.225 0.225 0.125 0.225 0.125
0 +1 0 0 +1 +a 0 1 1 +1 1 0 a 1 +1 +1 +1 +1 1 0 0 0 +1 1 0 0 0 0 1 0 1
85.41 91.54 88.61 90.82 94.55 99.12 93.45 58.62 91.32 99.98 87.21 92.65 25.34 55.16 99.18 99.87 94.36 98.74 53.21 97.45 82.17 92.18 95.96 51.36 86.45 94.99 59.78 88.13 85.46 94.36 84.36
Please cite this article in press as: M. Ghaedi, S.N. Kokhdan, Removal of methylene blue from aqueous solution by wood millet carbon optimization using response surface methodology, Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy (2014), http://dx.doi.org/10.1016/j.saa.2014.07.048
M. Ghaedi, S.N. Kokhdan / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy xxx (2014) xxx–xxx
5
Fig. 3. FT-IR spectrum of produced carbon.
Fig. 5. Pore size distribution of adsorbent prepared from millet carbon.
Fig. 5 show narrow micro-porosity structure of adsorbent surface area of AC around to be 32.58 m2/g with total pore volume of 0.029 cm3/g and average pore diameter less than 10 nm. The reasonable medium surface area and high volume assigned to low pore size show its applicability to adsorb more species.
The slope of the plot of ln qe vs e2 gives B (mol2/(kJ2)) and the intercept yields the adsorption capacity, (Qm (mg/g)). The mean free energy of adsorption (E) simple calculated by Eq. (3) [41]:
pHPZC Distribution of charge on AC has more contribution on its ability for removal and interaction of various compounds. The positive or negative charge of adsorbent significantly depend on variable such as pH and ionic strength one of most great important variables that significantly affect the behavior of adsorbent in dye removal for target transfer from bulk to the adsorbent surface known as diffusion into the micropores and mesopores. At pH above and below this value, the adsorbent get negative and positive charge, respectively, while the adsorption of anions is pHPZC accelerated at pH below the pHPZC [36,37].
Determination of oxygen containing functional groups The type and concentration of AC functional groups generally can be calculated according to Boehm titration technique [38]. In this general protocol, the NaHCO3 is used for determination of acidic carboxylic groups, while lactonic and carboxylic group is neutralized by sodium carbonate (Na2CO3). The acidic phenolic groups are determined following titration by sodium hydroxide. On the other hand titration of alkaline adsorbent media with HCl permit the calculation of basic group’s of AC such as pyrones [39,40]. The above mention titration has been carried out and content of each type functional group was estimated and presented in Table 1.
pffiffiffiffiffiffi E ¼ 1= 2B
ð3Þ
The calculated value of D–R parameters (Table 3) show that saturation adsorption capacity at different amount of adsorbents was in the range of 3.26–3.97, (good agreement with Langmuir value). The values of E based on Eq. (3) is between 1545.2 and 2915.8 mol1 strongly support physico-sorption nature of adsorption. Another important criterion for evaluating applicability of each model was carried out using the non-linear chi-square test statistic (v2) [42] as Eq. (10).
v2 ¼ Sumðqe;exp qe;cal =qe;cal Þ2
where qe,exp and qe,cal are experimental and calculated adsorption capacity value, respectively. The good agreement of predicted and experimental data tread to lower v2 value that support and confirm applicability of each model and vice versa. The obtained non-linear v2 value show that higher R2 value and smaller v2 value of Langmuir isotherm in comparison to other model show its superiority to other model for explanation of experimental equilibrium data. Similar value of other applied model confirms high efficiency of Langmuir isotherm to represent the experimental data at all conditions. The lower correlation coefficient (R2) of Freundlich model in comparison to Langmuir model, suggest confirm monolayer nature compare to multilayer adsorption. It is important to inaugurate the most appropriate correlations for the equilibrium data using conventional isotherm models like: Langmuir, Freundlich, Tempkin and Dubinin–Radushkevich. Applicability of the isotherm equations was tested by considering their correlation coefficient (R2) and error analysis.
Langmuir model :
C e =qe ¼ 1=ðK L qm Þ þ ð1=qm ÞC e
And Freundlich model :
Fig. 4. Central composite design (CCD) and optimization of parameters.
ð4Þ
qe ¼ K F C 1=n e
ð5Þ ð6Þ
KL is proportional to adsorption intensity and heat of adsorption (dm mg1) and qm is maximum amount of adsorbed analytes. The qm and KL value can be determined from the slope and intercept of respective line obtained by plotting log Ce/qe vs Log Ce. In Eq. (6), KF (constant related to the bonding energy) represents the quantity of dye adsorbed onto adsorbent. The slope 1/n, (between 0 and 1) is a measure of adsorption intensity or surface heterogeneity [43,44]. The value lower than 1 indicates a normal Freundlich isotherm while respective values higher than one indicate cooperative adsorption [45]. A plot of Ln qe vs Ln Ce enables the empirical constants KF and 1/n to be determined from the intercept and slope of the linear regression.
Please cite this article in press as: M. Ghaedi, S.N. Kokhdan, Removal of methylene blue from aqueous solution by wood millet carbon optimization using response surface methodology, Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy (2014), http://dx.doi.org/10.1016/j.saa.2014.07.048
6
M. Ghaedi, S.N. Kokhdan / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy xxx (2014) xxx–xxx
Table 3 Comparison of the coefficients isotherm parameters for MB adsorption MHC. Isotherm
Equation
Parameters
0.15 g
0.225 g
0.3 g
Langmuir
Ce/qe = 1/KaQm + Ce/Qm
Qm (mg g1) Ka (L mg1) RL X2 R2
4.937 0.736 0.042–0.186 0.554 0.997
4.315 1.105 0.033–0.149 0.987 0996
3.56 1.573 0.02–0.078 1.129 0.995
Freundlich
Ln qe = lnKF + (1/n)lnCe
1/n KF (L mg1) X2 R2
0.29 2.219 4.65 0.976
0.376 2.189 4.76 0.959
0.399 1.997 3.45 0.987
Tempkin
qe = Bl ln KT + BllnCe
Bl KT (L mg1) X2 R2
0.835 12.745 4.65 0.968
0.856 19.875 5.87 0.996
0.726 32.456 6.98 0.987
Dubinin and Radushkevich
Ln qe = ln Qs – Ke2
Qs (mg g1) K (108) E (kJ/mol) = 1/(2 K)1/2 X2 R2
4.54 19 1612.9 1.78 0.878
4.09 8 2706.7 0.987 0.965
3.07 7 2908.7 0.879 0.973
Result of Table 3 (Langmuir and Freundlich isotherm) show that experimental date well and fits adsorbent follow the Langmuir model (better fit as reflected by correlation coefficients (R2) of 0.997) with monolayer adsorption capacity [45–51]. It seems that the monolayer adsorption via different forces is the predominant mechanism of removal process. The adsorption of more layers on the first layer is significantly forbiben due to electrostatic repulsion between adsorbed layer and bulk molecules. Tempkin and Pyzhev [52] assumes the adsorption heat of in each a layer decreases linearly with surface coverage of adsorbent due to sorbate–adsorbate interactions. The linear form of the Tempkin isotherm equation is represented by the following equation [53]:
qe ¼ B1 ln K T þ B1 ln C e
ð7Þ
Values of B1 and KT were calculated from the plot of qe against Ln Ce (Table 3) that Tempkin isotherm with relatively high correlation coefficient (R2 > 0.9) is near the value of Langmuir. The Dubinin–Radushkevich (D–R) isotherm applied to estimate the porosity apparent free energy and the characteristic of adsorption [54,55] and its linear form can be shown in following equation:
Ln qe ¼ ln Q m K e2
ð8Þ
where e (Polanyi potential) can be calculated from Eq. (8):
e ¼ RT ln ð1 þ 1=C e Þ
ð9Þ
Adsorbent (g)
From the plot of ln qe vs e2, the K value was from its slope (mol2 (kJ2)1) and the intercept show the adsorption capacity (Qm; mg g1) simply can be calculated [56–61]. Kinetic study The correlation coefficient (R2) for the pseudo second-order kinetic at different initial MB concentrations was above 0.99 and the calculated qe values have good agreement with experimental values (Table 4). The equilibrium sorption capacity (qe) increase from 1.8 mg/g to 2.9 mg/g by raising the MB concentration from 6 mg/L to 18 mg/L at 0.25 g adsorbent value. The values of the rate constant were found to decrease from 1.069 g/mg min to 0.195 g/ mg min by raising the initial MB concentration from 6 mg/L to 18 mg/L at 0.25 g adsorbent. The high correlation coefficient (R2 values close or equal to 1) nearby value of theoretical and experimental adsorption capacity is criterion for judgment about suitability of kinetic models for explanation of experimental data. Comparison based on these two parameters show applicability of pseudo second order model for evaluating and fitting experimental data over entire adsorption stage [41–44]. The intraparticle diffusion model The linear relation of initial dye concentration with removal rate failed when pore diffusion is the predominant stage and limits the adsorption process [47–65]. Therefore, the possibility and
Table 4 Kinetic parameters of MB removal using 0.1–0.15 g of adsorbent over concentration in the range of 5–15 ppm at optima condition. Models
Parameters
5 ppm, 0.1 g
7 ppm, 0.15 g
10 ppm, 0.15 g
15 ppm, 0.15 g
First order kinetic model: Log(qeqt) = log(qe)–(K1/2.303)t
K1 qe (cal) R2
0.046 0.89 0.934
0.059 0.0.96 0.948
0.025 1.315 0.936
0.02116 1.894 0.921
Second order kinetic model: t/qt = 1/k2qe2 + (1/qe)t
K2 qe (cal) R2
0.298 2.89 0.998
0.597 2.18 0.999
0.276 3.76 0.998
0.214 4.54 0.997
Intraparticle diffusion qt = Kid t1/2 + C
Kdif C R2
0.089 1.95 0.879
0.099 1.67 0.907
0.143 1.86 0.976
0.199 2.34 0.945
Elovich qt = 1/b ln(ab) + 1/b ln (t)
b R2
5.37 0.966
7.48 0.990
4.99 0.990
3.07 0.967
Experimental date
qe (exp)
2.965
2.17
3.75
4.57
Please cite this article in press as: M. Ghaedi, S.N. Kokhdan, Removal of methylene blue from aqueous solution by wood millet carbon optimization using response surface methodology, Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy (2014), http://dx.doi.org/10.1016/j.saa.2014.07.048
M. Ghaedi, S.N. Kokhdan / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy xxx (2014) xxx–xxx
effective pH was 6 and the optimum adsorbent dose was found to be 0.25 g for 6 ppm of MB with removal percentage above 99%. Langmuir isotherm gave best fit to adsorption data than Freundlich isotherm using linear and non-linear methods that this fact may be attributed to high Langmuir surface area of adsorbent. The kinetic study of MB on AC was performed based on pseudo-first order, pseudo-second-order, Elovich and intraparticle diffusion equations. The data indicate that the adsorption kinetics follow the pseudo-second-order rate in addition to interparticle diffusion. The present study concludes that the WHC could be employed as low-cost adsorbents as alternatives to commercial activated carbon for the removal of color and dyes from water and wastewater. Further studies on quantitative characterization of this adsorbent and involved mechanisms, and feasibility of using this adsorbent for other dyes and for its possible industrial application are needed.
Table 5 Maximum mono layer sorption of several adsorbent. Adsorbent
Adsorption capacity (mg g1)
Refs.
ZnS-NP-AC Activated carbon (coconut shell fibbers) Activated carbon (olive stones) Cotton waste Date pits Fly ash Perlite Perlite Perlite (EP) Pyrophyllite Zeolite Activated carbon S. muticum seaweed Hydrilla verticillata Moss Moss Water hyacinth root Spirodela polyrrhiza duckweed Hexane-extracted spent bleaching earth Millet
90.9 19.59
[9] [66]
303 240 80.3 53.84 5.6–9.08 162.3 17.4–31.7 70.42 53.1 373.9 279.2 198.0 185,0 128.9 144.93 120.5
[67] [68] [69] [70] [71] [72] [73] [74] [75] [75] [76] [77] [78] [78] [79] [80]
3.745–4.739
This study
References
usability of this model for interpretation of experimental data was explored according to its well knows equation and conditions:
qt ¼ K dif t 0:5 þ C
7
ð10Þ
where C (mg g1) is the intercept and kdif is the intraparticle diffusion rate constant (in mg g1 min1/2). The values of qt were found to be linearly correlated with values of t1/2 and the rate constant kdif directly evaluated from the slope of the regression line (Table 4). The values of intercept C (Table 4) approximately give idea about the boundary layer thickness. The constant (C) was found to increase from 1.089 to 2.224 with increase in MB amount from 6 to 18 mg L1 at 0.15–0.3 g of AC is attributed to probable increase in boundary thickness and decrease in the chance of the external mass transfer. Both of these behavior support prominent increase in the amount of internal mass transfer. The high value of R2 shows suitability of this model to explain the experimental data. This may confirm that the rate-limiting step is the intraparticle diffusion process. The intraparticle diffusion rate constant, (kdif) value was in the range of 0.038–0.152 mg g1 min1/2 at 0.225 g adsorbent and has good positive correlation with initial dye concentration and this linear relationship show high contribution of intraparticle diffusion on the adsorption process. Generally, in kinetic studies passing the intraparticle diffusion plot through origin show that this mechanism solely limits the adsorption rate. This situation was not achieved in our studies that show another alternative model is coinciding in addition to intraparticle diffusion model to follow the adsorption data. Comparison with other adsorbents for MB The performance of adsorbent used in this work has been compared with other adsorbents. As is seen in Table 5 the proposed adsorbent is superior to literature in term of samples and maximum adsorption capacity. Conclusion This investigation show the applicability of MHC as good, green, low-cost and locally available adsorbent for the removal of MB from aqueous solutions in short time (
Contents lists available at ScienceDirect
Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy journal homepage: www.elsevier.com/locate/saa
Removal of methylene blue from aqueous solution by wood millet carbon optimization using response surface methodology Mehrorang Ghaedi ⇑, Syamak Nasiri Kokhdan Chemistry Department, Yasouj University, Yasouj 75918-74831, Iran
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Low-cost and non-toxic walnut
carbon (WC) was easily made. Walnut carbon was used to remove
methylene blue from aqueous solution. The dye removal was performed in a very short time (2 min). Walnut carbon was re-used after heating. Optimization was performed using response surface methodology.
a r t i c l e
i n f o
Article history: Received 19 April 2014 Received in revised form 8 July 2014 Accepted 18 July 2014 Available online xxxx Keywords: Adsorption Methylene blue (MB) Millet household carbon (MHC) Kinetic
a b s t r a c t The use of cheep, non-toxic, safe and easily available adsorbent are efficient and recommended material and alternative to the current expensive substance for pollutant removal from wastewater. The activated carbon prepared from wood waste of local tree (millet) extensively was applied for quantitative removal of methylene blue (MB), while simply. It was used to re-used after heating and washing with alkaline solution of ethanol. This new adsorbent was characterized by using BET surface area measurement, FT-IR, pH determination at zero point of charge (pHZPC) and Boehm titration method. Response surface methodology (RSM) by at least the number of experiments main and interaction of experimental conditions such as pH of solution, contact time, initial dye concentration and adsorbent dosage was optimized and set as pH 7, contact time 18 min, initial dye concentration 20 ppm and 0.2 g of adsorbent. It was found that variable such as pH and amount of adsorbent as solely or combination effects seriously affect the removal percentage. The fitting experimental data with conventional models reveal the applicability of isotherm models Langmuir model for their well presentation and description and Kinetic real rate of adsorption at most conditions efficiently can be represented pseudo-second order, and intra-particle diffusion. It novel material is good candidate for removal of huge amount of MB (20 ppm) in short time (18 min) by consumption of small amount (0.2 g). Ó 2014 Elsevier B.V. All rights reserved.
Introduction Main and significant environmental pollution source are highly color dyes suspended in organic solids or present in bulk solution ⇑ Corresponding author. Tel./fax: +98 741 2223048. E-mail address: [email protected] (M. Ghaedi).
[1]. Arrival and presence of these pollutants in water significantly damage aquatic life following generation of mutagenic and carcinogenic activity. Textile, paper and printing activities are some source of dye containing wastewater. Versatile (cheep and non toxic material are used for dyes removal present in various media [2–4]. It was found that complex (aromatic) structures dyes due to their high and remarkable stability toward degeneration (phot and
http://dx.doi.org/10.1016/j.saa.2014.07.048 1386-1425/Ó 2014 Elsevier B.V. All rights reserved.
Please cite this article in press as: M. Ghaedi, S.N. Kokhdan, Removal of methylene blue from aqueous solution by wood millet carbon optimization using response surface methodology, Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy (2014), http://dx.doi.org/10.1016/j.saa.2014.07.048
2
M. Ghaedi, S.N. Kokhdan / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy xxx (2014) xxx–xxx
Nomenclatures Ct Qe Ce V W k1 k2 H
a b C Kdif F Qm Ka
dye concentration (mg L1) at time t equilibrium adsorption capacity (mg g1) dye concentration (mg L1) at equilibrium volume of solution (L) weight of adsorbent (g) rate constant of pseudofirst order adsorption (min1) second-order rate constant of adsorption (mg g1 min1) second-order rate constants (mg g1 min1) initial adsorption rate (mg g1 min1) desorption constant (mg g1) intercept of intraparticle diffusion (related to the thickness of the boundary layer) rate constant of intraparticle diffusion (mg g1 min1/2) fraction of solute adsorbed at any time t (mg g1) maximum adsorption capacity reflected a complete monolayer (mg g1) in Langmuir isotherm model Langmuir constant or adsorption equilibrium constant (L mg1) that is related to the apparent energy of sorption
bio degradation), light, heat oxidizing agent and other approach can be considered as more and serious toxic agent for ecosystem. Dyes via direct destruction or inhibition of catalytic capabilities lead to depth in microorganism [5,6]. Among the well known dyes reduction pathways [7–27] more attention was devoted to adsorption that emerged from unique advantages such as highly porous and safe adsorbent. These material due to their unique properties are useful for dyes removal. Carbon simply can be obtained by burning and putting it into a sealed container with very cost effective and non-toxic manner without much energy consumption. All these characteristics of such carbon make it more suitable than commercial activated carbon for diverse application. Methylene Blue (MB) (Fig. 1) manly use for coloring paper, temporary hair colorant, dyeing cottons, wools, while reveal very harmful effects on living things [7–9]. Dyes contaminated ecosystem significantly can perturbed aquatic, plants and human lifes [10–27]. MB due to its chemical resistance to light and oxidizing agent hardly can be removal or eliminated by biological treatment and chemical precipitation [28]. The pestilent and dangerous intermediates/substances created by this dye after undergoing reduction and oxidation in water further increase the need for its removal from wastewater. Therefore, we were motivated to prepare carbon from millet that following subsequent activation is as an ideal MB removal agent. The full characterization of this AC by BET, FT-IR, pH determination at zero point of charge (pHZPC) and Boehm titration reveal and confirm its applicability to interact with large amount of material through various pathway and mechanism. The experimental runs and optimum conditions were achieved along at least experimental run by central composite design and response surface methodology (RSM). This methodology simply permits to estimate the main effect and interaction of variables through economic and repeatable pathway.
Fig. 1. Chemical structure of methylene blue green.
RL KF N T R B1 K Qm e E X2 qe,exp qe,calc R2
Dimensionless equilibrium parameter (separation factor) isotherm constant indicate the capacity parameter (mg g1) related to the intensity of the adsorption isotherm constant indicate the empirical parameter (g L1) related to the intensity of the adsorption absolute temperature in Kelvin universal gas constant (8.314 J K1 mol1) related to the heat of adsorption (B1 = RT/b) constant related to the adsorption energy at the D–R isotherm (mol2 kJ2) theoretical saturation capacity at the D–R isotherm Polanyi potential at the D–R isotherm mean free energy of adsorption chi-squared test statistic experimental data of the equilibrium capacity (mg g1) equilibrium capacity obtained by calculating from the isotherm model (mg g1) correlation coefficient
Subsequent objective of this study was investigation of kinetics and isotherms parameters correspond to MB adsorption onto AC. The whole results confirm the suitability and applicability of the Langmuir model for isothermal study elucidation while the real behavior of adsorption experimental data simply can be represented by combination of pseudo second order and interparticle diffusion model.
Experimental Materials and instrumentation The methylene blue with molecular formula C16H18N3SCl and IUPAC name 3,7-bis(Dimethylamino)-phenothiazin-5-ium chloride, (MW = 319.86 g/mol), with CAS Number of 61-73-4 (United States, Sigma–Aldrich) 100 mg/L as stock solution was supplied by dissolving 10 mg of MB in 100 mL double-distilled water. The effect of solution pH over 2–8 (adjusted via addition of dilute HCl and/or KOH solution) its adsorption removal was studied, while each pH maximum wave length was recorded. All the parameters and experimental data were studied using the software MINITAB (version 16.0).
Preparation of millet carbon The source material container was heated at 400 °C for 3 h and subsequent production of coal was then slowly cooled to room temperature and rinsed with distilled water. The carbon was grinded to mesh lower than 120 and subsequently was activated by various with mixture of concentrated HCl and HNO3 with 1:1 valuation. The produced AC was studied by SEM (see Fig. 2), BET and FT-IR. The pH corresponding to the AC point of zero charge (pHZPC) of was determined by the pH drift method [27,29,30]. Determination of oxygen containing functional groups was carried out as follow (Boehm titration method) [31]: 1.0 g of the AC mixed separately with 15 ml solution of NaHCO3 (0.1 M), Na2CO3 (0.05 M) and NaOH (0.1 M) for acidic groups and 0.1 M HCl for basic groups/sites respectively at room temperature for more than
Please cite this article in press as: M. Ghaedi, S.N. Kokhdan, Removal of methylene blue from aqueous solution by wood millet carbon optimization using response surface methodology, Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy (2014), http://dx.doi.org/10.1016/j.saa.2014.07.048
M. Ghaedi, S.N. Kokhdan / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy xxx (2014) xxx–xxx
3
Fig. 2. SEM of activated carbon prepared from wood millet carbon.
2 days. Subsequently, the aqueous solutions were back titrated with HCl (0.1 M) for acidic and NaOH (0.1 M) for basic groups. The number and type of acidic sites were calculated by considering that NaOH neutralizes carboxylic, lactonic and phenolic groups, Na2CO3 neutralizes carboxylic and lactonic groups and that NaHCO3 neutralizes only carboxylic groups. Carboxylic groups are calculated by direct titration with NaHCO3, while the difference between actual amount of Na2CO3 and NaHCO3 volume is used for estimation of lactones. The difference between the groups titrated with NaOH and Na2CO3 show the phenol content. Basic sites were determined by titration with HCl. Characterization of prepared adsorbent was performed and presented in Table 1 is usable for simultaneous combination with RSM optimization of variables. According to the analysis of response by ANOVA, the main effect and interaction of variable can be estimated.
Central composite design (CCD) and optimization of parameters Central composite design (CCD) as most accepted experimental designs methodology following. Application of CCD permit estimation of their and synergetic or antagonist effect factors including pH, amount of adsorbent, MB concentration and contact time on removal percentage. Thirty-one experiments based on well know procedure and formula were designed and run at different conditions (room temperature) [32–34]. The number of runs was obtained according the relation 2n + 2n + nc, where n is the number of factors (four factors), nc is the number of center points (seven replicates). The MB removal percentage (response) and its change with Coded and uncoded factor values are listed in Table 2. Grinded and small price of millet wood waste was totally rinsed with detergent. Subsequently, it was revealed that variable such as time
and temperature of heating and type concentration of concentrated acid are major factor. The optimal model predictor quadratic equation given as:
Y ¼ b0 þ
4 4 3 X 4 X X X b i xi þ bii x2i þ bij xi xj i¼1
i¼1
ð1Þ
i¼1 j¼iþ1
was used to optimize and predict the response (Y) against the parameters. Where xi, xj are the coded values of the factors and b0, bi, bii and bij are the constant, linear, quadratic and interaction coefficients respectively. Following the analysis of variance and according to, the coefficient of determination (R2), the probability p-value (95% confidence level) and fisher’s test, it is possible to take useful information about statistical significance and the characteristics of suitability of predicated mode 1 for predication of real behavior of adsorption system. Significant parameters were recognized and subsequently a predictive model as follow was developed.
Y ¼ 90:24 þ 30:48x2 17:63x3 29:39x22 10:41x23 þ 27:03x2 x3
ð2Þ
Positive coefficients mean that the corresponding terms affect the response positively and negative values affect it negatively. According to ANOVA result (Table 2) and well know role (P-values less than 0.05) confirm the significance of each parameter, while the P value of higher than 0.05 show high significance and more applicability of model for prediction and description of exact behavior of proposed system. The terms x2 ; x3 ; x22 ; x23 and x2x3 were found to be significant that suggest more contribution of parameters like the MB concentration, the amount of millet AC
Please cite this article in press as: M. Ghaedi, S.N. Kokhdan, Removal of methylene blue from aqueous solution by wood millet carbon optimization using response surface methodology, Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy (2014), http://dx.doi.org/10.1016/j.saa.2014.07.048
4
M. Ghaedi, S.N. Kokhdan / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy xxx (2014) xxx–xxx
conditions reveal the high repeat ability of method for prediction of real removal percentage with relative deviation less than 3%.
Table 1 Summary report of proposed sorbent. Summary report Surface area Single point surface area at p/p° = 0.206301245 BET surface area: Langmuir surface area: Pore volume Single point adsorption total pore volume of poresless than 1256.713 Å width at p/p° = 0.984352481: t-Plot micropore volume: BJH adsorption cumulative volume of pores between 17.000 Å and 3000.000 Å width: BJH desorption cumulative volume of pores between 17.000 Å and 3000.000 Å width: Pore size Adsorption average pore width (4V/A by BET): BJH adsorption average pore width (4V/A): BJH desorption average pore width (4V/A): pHZPC Acid soluble Water soluble Carboxylic(acidic functions) Phenol Lactones Basic sites
335.9 m2/g 30.8 m2/g 342.7 m2/g 0.041 cm3/g 0.026 cm3/g 0.037 cm3/g 0.026 cm3/g
35.9 Å 204.2 Å 205.7 Å 8.5 NO NO 0.805 mmol/g NO 0.418 mmol/g NO
affect the MB removal linear and in quadratic way and typical three dimensional response surface Fig. 3 confirm the presence of strong interaction among variable. The optimized values for pH, the amount of AC prepared from millet, MB concentration and contact time were found to be 6.5, 0.26 g, 7.1 mg/L and 30 min, respectively. At these conditions, the predicated MB removal percentage was more than 96% with desirability 1. Conduction of similar experiments at specified optimum
Results and discussion Characterization of carbon FT-IR Applications of powerful identification techniques like FT-IR (Fig. 4) make reveal the distinguished peaks that confirm presence of functional groups like OH, COOH, amine and carbonyl groups. Stretching vibration band around 1700 cm1 is assigned to carbonyl C@O group present in aldehyde, ester, ketone and acetyl derivatives. The strong band at 1500 cm1 may be due to C@C band. The peaks appearing between 423 cm1 and 634 cm1 are assigned to the metal–oxygen (MAO) stretching mode in structure of carbon. The broad band around 3000–3400 cm1 confirm OH group. The band around 2900 cm1 reveals the presence CH group in AC. Therefore, presence of these reactive sites and centers make candidate the present AC as good and suitable material for strong interaction with various material and suitable them for waste water treatment. BET analysis of activated carbon The N2/77K adsorption isotherms permit to study the porous structure of material and also to attain useful information about surface area its ability to adsorb a target compound [35]. Specific surface area (SSA), pore volume and accessible pore volume for each pores category a significant criterion make useful information about the ability of material for trapping molecules with distinct sizes. The interference by the surrounding phase is especially problematic for the Bruner–Emmet–Teller (BET) N2 adsorption/desorption isotherm method because the entire surface is by Table 1 and
Table 2 Coded and uncoded central composite design. Run
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Dye concentration (mg/L)
Time (min)
Uncoded
PH Coded
Uncoded
Coded
Uncoded
Coded
Adsorbent (g) Uncoded
Coded
Removal (%)
2 8 6 6 8 6 6 4 8 4 8 6 6 4 8 8 4 4 8 6 6 6 4 8 6 10 6 6 4 6 4
a +1 0 0 +1 0 0 1 +1 1 +1 0 0 1 +1 +1 1 1 +1 0 0 0 1 +1 0 +a 0 0 1 0 1
10 14 10 10 14 10 10 14 6 6 6 10 10 14 6 6 14 6 14 10 10 10 14 14 10 10 18 10 6 2 6
0 +1 0 0 +1 0 0 +1 1 1 1 0 0 +1 1 1 +1 1 +1 0 0 0 +1 +1 0 0 +a 0 1 a 1
20 11 20 20 29 20 20 29 29 29 11 20 20 11 11 29 11 11 11 20 2 38 29 29 20 20 20 20 29 20 11
0 1 0 0 +1 0 0 +1 +1 +1 1 0 0 1 1 +1 1 1 1 0 a +a +1 +1 0 0 0 0 +1 0 1
0.225 0.325 0.225 0.225 0.325 0.425 0.225 0.125 0.125 0.325 0.125 0.225 0.025 0.125 0.325 0.325 0.325 0.325 0.125 0.225 0.225 0.225 0.325 0.125 0.225 0.225 0.225 0.225 0.125 0.225 0.125
0 +1 0 0 +1 +a 0 1 1 +1 1 0 a 1 +1 +1 +1 +1 1 0 0 0 +1 1 0 0 0 0 1 0 1
85.41 91.54 88.61 90.82 94.55 99.12 93.45 58.62 91.32 99.98 87.21 92.65 25.34 55.16 99.18 99.87 94.36 98.74 53.21 97.45 82.17 92.18 95.96 51.36 86.45 94.99 59.78 88.13 85.46 94.36 84.36
Please cite this article in press as: M. Ghaedi, S.N. Kokhdan, Removal of methylene blue from aqueous solution by wood millet carbon optimization using response surface methodology, Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy (2014), http://dx.doi.org/10.1016/j.saa.2014.07.048
M. Ghaedi, S.N. Kokhdan / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy xxx (2014) xxx–xxx
5
Fig. 3. FT-IR spectrum of produced carbon.
Fig. 5. Pore size distribution of adsorbent prepared from millet carbon.
Fig. 5 show narrow micro-porosity structure of adsorbent surface area of AC around to be 32.58 m2/g with total pore volume of 0.029 cm3/g and average pore diameter less than 10 nm. The reasonable medium surface area and high volume assigned to low pore size show its applicability to adsorb more species.
The slope of the plot of ln qe vs e2 gives B (mol2/(kJ2)) and the intercept yields the adsorption capacity, (Qm (mg/g)). The mean free energy of adsorption (E) simple calculated by Eq. (3) [41]:
pHPZC Distribution of charge on AC has more contribution on its ability for removal and interaction of various compounds. The positive or negative charge of adsorbent significantly depend on variable such as pH and ionic strength one of most great important variables that significantly affect the behavior of adsorbent in dye removal for target transfer from bulk to the adsorbent surface known as diffusion into the micropores and mesopores. At pH above and below this value, the adsorbent get negative and positive charge, respectively, while the adsorption of anions is pHPZC accelerated at pH below the pHPZC [36,37].
Determination of oxygen containing functional groups The type and concentration of AC functional groups generally can be calculated according to Boehm titration technique [38]. In this general protocol, the NaHCO3 is used for determination of acidic carboxylic groups, while lactonic and carboxylic group is neutralized by sodium carbonate (Na2CO3). The acidic phenolic groups are determined following titration by sodium hydroxide. On the other hand titration of alkaline adsorbent media with HCl permit the calculation of basic group’s of AC such as pyrones [39,40]. The above mention titration has been carried out and content of each type functional group was estimated and presented in Table 1.
pffiffiffiffiffiffi E ¼ 1= 2B
ð3Þ
The calculated value of D–R parameters (Table 3) show that saturation adsorption capacity at different amount of adsorbents was in the range of 3.26–3.97, (good agreement with Langmuir value). The values of E based on Eq. (3) is between 1545.2 and 2915.8 mol1 strongly support physico-sorption nature of adsorption. Another important criterion for evaluating applicability of each model was carried out using the non-linear chi-square test statistic (v2) [42] as Eq. (10).
v2 ¼ Sumðqe;exp qe;cal =qe;cal Þ2
where qe,exp and qe,cal are experimental and calculated adsorption capacity value, respectively. The good agreement of predicted and experimental data tread to lower v2 value that support and confirm applicability of each model and vice versa. The obtained non-linear v2 value show that higher R2 value and smaller v2 value of Langmuir isotherm in comparison to other model show its superiority to other model for explanation of experimental equilibrium data. Similar value of other applied model confirms high efficiency of Langmuir isotherm to represent the experimental data at all conditions. The lower correlation coefficient (R2) of Freundlich model in comparison to Langmuir model, suggest confirm monolayer nature compare to multilayer adsorption. It is important to inaugurate the most appropriate correlations for the equilibrium data using conventional isotherm models like: Langmuir, Freundlich, Tempkin and Dubinin–Radushkevich. Applicability of the isotherm equations was tested by considering their correlation coefficient (R2) and error analysis.
Langmuir model :
C e =qe ¼ 1=ðK L qm Þ þ ð1=qm ÞC e
And Freundlich model :
Fig. 4. Central composite design (CCD) and optimization of parameters.
ð4Þ
qe ¼ K F C 1=n e
ð5Þ ð6Þ
KL is proportional to adsorption intensity and heat of adsorption (dm mg1) and qm is maximum amount of adsorbed analytes. The qm and KL value can be determined from the slope and intercept of respective line obtained by plotting log Ce/qe vs Log Ce. In Eq. (6), KF (constant related to the bonding energy) represents the quantity of dye adsorbed onto adsorbent. The slope 1/n, (between 0 and 1) is a measure of adsorption intensity or surface heterogeneity [43,44]. The value lower than 1 indicates a normal Freundlich isotherm while respective values higher than one indicate cooperative adsorption [45]. A plot of Ln qe vs Ln Ce enables the empirical constants KF and 1/n to be determined from the intercept and slope of the linear regression.
Please cite this article in press as: M. Ghaedi, S.N. Kokhdan, Removal of methylene blue from aqueous solution by wood millet carbon optimization using response surface methodology, Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy (2014), http://dx.doi.org/10.1016/j.saa.2014.07.048
6
M. Ghaedi, S.N. Kokhdan / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy xxx (2014) xxx–xxx
Table 3 Comparison of the coefficients isotherm parameters for MB adsorption MHC. Isotherm
Equation
Parameters
0.15 g
0.225 g
0.3 g
Langmuir
Ce/qe = 1/KaQm + Ce/Qm
Qm (mg g1) Ka (L mg1) RL X2 R2
4.937 0.736 0.042–0.186 0.554 0.997
4.315 1.105 0.033–0.149 0.987 0996
3.56 1.573 0.02–0.078 1.129 0.995
Freundlich
Ln qe = lnKF + (1/n)lnCe
1/n KF (L mg1) X2 R2
0.29 2.219 4.65 0.976
0.376 2.189 4.76 0.959
0.399 1.997 3.45 0.987
Tempkin
qe = Bl ln KT + BllnCe
Bl KT (L mg1) X2 R2
0.835 12.745 4.65 0.968
0.856 19.875 5.87 0.996
0.726 32.456 6.98 0.987
Dubinin and Radushkevich
Ln qe = ln Qs – Ke2
Qs (mg g1) K (108) E (kJ/mol) = 1/(2 K)1/2 X2 R2
4.54 19 1612.9 1.78 0.878
4.09 8 2706.7 0.987 0.965
3.07 7 2908.7 0.879 0.973
Result of Table 3 (Langmuir and Freundlich isotherm) show that experimental date well and fits adsorbent follow the Langmuir model (better fit as reflected by correlation coefficients (R2) of 0.997) with monolayer adsorption capacity [45–51]. It seems that the monolayer adsorption via different forces is the predominant mechanism of removal process. The adsorption of more layers on the first layer is significantly forbiben due to electrostatic repulsion between adsorbed layer and bulk molecules. Tempkin and Pyzhev [52] assumes the adsorption heat of in each a layer decreases linearly with surface coverage of adsorbent due to sorbate–adsorbate interactions. The linear form of the Tempkin isotherm equation is represented by the following equation [53]:
qe ¼ B1 ln K T þ B1 ln C e
ð7Þ
Values of B1 and KT were calculated from the plot of qe against Ln Ce (Table 3) that Tempkin isotherm with relatively high correlation coefficient (R2 > 0.9) is near the value of Langmuir. The Dubinin–Radushkevich (D–R) isotherm applied to estimate the porosity apparent free energy and the characteristic of adsorption [54,55] and its linear form can be shown in following equation:
Ln qe ¼ ln Q m K e2
ð8Þ
where e (Polanyi potential) can be calculated from Eq. (8):
e ¼ RT ln ð1 þ 1=C e Þ
ð9Þ
Adsorbent (g)
From the plot of ln qe vs e2, the K value was from its slope (mol2 (kJ2)1) and the intercept show the adsorption capacity (Qm; mg g1) simply can be calculated [56–61]. Kinetic study The correlation coefficient (R2) for the pseudo second-order kinetic at different initial MB concentrations was above 0.99 and the calculated qe values have good agreement with experimental values (Table 4). The equilibrium sorption capacity (qe) increase from 1.8 mg/g to 2.9 mg/g by raising the MB concentration from 6 mg/L to 18 mg/L at 0.25 g adsorbent value. The values of the rate constant were found to decrease from 1.069 g/mg min to 0.195 g/ mg min by raising the initial MB concentration from 6 mg/L to 18 mg/L at 0.25 g adsorbent. The high correlation coefficient (R2 values close or equal to 1) nearby value of theoretical and experimental adsorption capacity is criterion for judgment about suitability of kinetic models for explanation of experimental data. Comparison based on these two parameters show applicability of pseudo second order model for evaluating and fitting experimental data over entire adsorption stage [41–44]. The intraparticle diffusion model The linear relation of initial dye concentration with removal rate failed when pore diffusion is the predominant stage and limits the adsorption process [47–65]. Therefore, the possibility and
Table 4 Kinetic parameters of MB removal using 0.1–0.15 g of adsorbent over concentration in the range of 5–15 ppm at optima condition. Models
Parameters
5 ppm, 0.1 g
7 ppm, 0.15 g
10 ppm, 0.15 g
15 ppm, 0.15 g
First order kinetic model: Log(qeqt) = log(qe)–(K1/2.303)t
K1 qe (cal) R2
0.046 0.89 0.934
0.059 0.0.96 0.948
0.025 1.315 0.936
0.02116 1.894 0.921
Second order kinetic model: t/qt = 1/k2qe2 + (1/qe)t
K2 qe (cal) R2
0.298 2.89 0.998
0.597 2.18 0.999
0.276 3.76 0.998
0.214 4.54 0.997
Intraparticle diffusion qt = Kid t1/2 + C
Kdif C R2
0.089 1.95 0.879
0.099 1.67 0.907
0.143 1.86 0.976
0.199 2.34 0.945
Elovich qt = 1/b ln(ab) + 1/b ln (t)
b R2
5.37 0.966
7.48 0.990
4.99 0.990
3.07 0.967
Experimental date
qe (exp)
2.965
2.17
3.75
4.57
Please cite this article in press as: M. Ghaedi, S.N. Kokhdan, Removal of methylene blue from aqueous solution by wood millet carbon optimization using response surface methodology, Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy (2014), http://dx.doi.org/10.1016/j.saa.2014.07.048
M. Ghaedi, S.N. Kokhdan / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy xxx (2014) xxx–xxx
effective pH was 6 and the optimum adsorbent dose was found to be 0.25 g for 6 ppm of MB with removal percentage above 99%. Langmuir isotherm gave best fit to adsorption data than Freundlich isotherm using linear and non-linear methods that this fact may be attributed to high Langmuir surface area of adsorbent. The kinetic study of MB on AC was performed based on pseudo-first order, pseudo-second-order, Elovich and intraparticle diffusion equations. The data indicate that the adsorption kinetics follow the pseudo-second-order rate in addition to interparticle diffusion. The present study concludes that the WHC could be employed as low-cost adsorbents as alternatives to commercial activated carbon for the removal of color and dyes from water and wastewater. Further studies on quantitative characterization of this adsorbent and involved mechanisms, and feasibility of using this adsorbent for other dyes and for its possible industrial application are needed.
Table 5 Maximum mono layer sorption of several adsorbent. Adsorbent
Adsorption capacity (mg g1)
Refs.
ZnS-NP-AC Activated carbon (coconut shell fibbers) Activated carbon (olive stones) Cotton waste Date pits Fly ash Perlite Perlite Perlite (EP) Pyrophyllite Zeolite Activated carbon S. muticum seaweed Hydrilla verticillata Moss Moss Water hyacinth root Spirodela polyrrhiza duckweed Hexane-extracted spent bleaching earth Millet
90.9 19.59
[9] [66]
303 240 80.3 53.84 5.6–9.08 162.3 17.4–31.7 70.42 53.1 373.9 279.2 198.0 185,0 128.9 144.93 120.5
[67] [68] [69] [70] [71] [72] [73] [74] [75] [75] [76] [77] [78] [78] [79] [80]
3.745–4.739
This study
References
usability of this model for interpretation of experimental data was explored according to its well knows equation and conditions:
qt ¼ K dif t 0:5 þ C
7
ð10Þ
where C (mg g1) is the intercept and kdif is the intraparticle diffusion rate constant (in mg g1 min1/2). The values of qt were found to be linearly correlated with values of t1/2 and the rate constant kdif directly evaluated from the slope of the regression line (Table 4). The values of intercept C (Table 4) approximately give idea about the boundary layer thickness. The constant (C) was found to increase from 1.089 to 2.224 with increase in MB amount from 6 to 18 mg L1 at 0.15–0.3 g of AC is attributed to probable increase in boundary thickness and decrease in the chance of the external mass transfer. Both of these behavior support prominent increase in the amount of internal mass transfer. The high value of R2 shows suitability of this model to explain the experimental data. This may confirm that the rate-limiting step is the intraparticle diffusion process. The intraparticle diffusion rate constant, (kdif) value was in the range of 0.038–0.152 mg g1 min1/2 at 0.225 g adsorbent and has good positive correlation with initial dye concentration and this linear relationship show high contribution of intraparticle diffusion on the adsorption process. Generally, in kinetic studies passing the intraparticle diffusion plot through origin show that this mechanism solely limits the adsorption rate. This situation was not achieved in our studies that show another alternative model is coinciding in addition to intraparticle diffusion model to follow the adsorption data. Comparison with other adsorbents for MB The performance of adsorbent used in this work has been compared with other adsorbents. As is seen in Table 5 the proposed adsorbent is superior to literature in term of samples and maximum adsorption capacity. Conclusion This investigation show the applicability of MHC as good, green, low-cost and locally available adsorbent for the removal of MB from aqueous solutions in short time (
