Efficient removal of malachite green from wastewater

0 downloads 6 Views 5MB Size Report
Nov 3, 2018 - Malachite green (MG) one of these toxic and carcinogenic dyes, has ...... porous carbon nitride functionalized with melamine-based dendrimer amine, Int. J. ... [24] N.F. Robaina, S. Soriano, R.J. Cassella, Polyurethane foam ...

Applied Surface Science 469 (2019) 236–245

Contents lists available at ScienceDirect

Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc

Full Length Article

Efficient removal of malachite green from wastewater by using boron-doped mesoporous carbon nitride

T

Elham Boorboor Azimi, Alireza Badiei , Jahan B. Ghasemi ⁎

School of Chemistry, College of Science, University of Tehran, Tehran, Iran

ARTICLE INFO

ABSTRACT

Keywords: Carbon nitride Boron doping Malachite green adsorption Response surface methodology

In this study the adsorption performance of boron doped mesoporous carbon nitride (BMCN) was studied for the elimination of malachite green (MG) dye from wastewater. BMCN with various weight ratios of boron doping was prepared in two steps including using a hard template to synthesize MCN, and then addition of boric acid as the boron source to produce BMCN. 1 wt% boron doping displayed the highest removal efficiency and adsorption capacity. The new adsorbent was analyzed using XRD, N2 adsorption-desorption, TEM, SEM-EDX and its elemental mapping, and FT-IR. Four critical parameters optimized by response surface methodology (RSM) method were temperature, initial dye concentration, pH and sorbent weight for MG adsorption. The maximum removal (100%) predicted from RSM was reported for 18 mg of the sorbent at pH 5 and MG concentration of 20 mg L−1 at room temperature. The boron doped mesoporous carbon nitride showed a high maximum adsorption capacity about 310 mg g−1 and reached at the equilibrium within 30 min followed the pseudo-second-order kinetic model with 99.8% MG removal while the rate-limiting step is the intraparticle diffusion stage. The adsorption phenomena matched well to the heterojunction surface Freundlich, Koble-Corrigan, and Sips isotherm models. The BMCN1 showed excellent reusability and was applied for six cycles of regeneration effectively, while its removal efficiency remained at a high value.

1. Introduction In recent years, one of the most serious challenges that scientists are dealing with is environmental pollution. One of these challenges is omitting toxic and priority pollutants from effluents before releasing to the environment. Dyes hinder light penetration into streams, impede photosynthesis and also can chelate with metal ions, all resulting in toxicity of living organisms [1]. As a result, the treatment of effluent containing such dyes has become of great significance to researchers. Malachite green (MG) one of these toxic and carcinogenic dyes, has been found valuable in industries, to control fish disease and parasites, and for painting wool, nylon, leather, silk, jute, cotton, and paper. However, this useful dye enters aqueous media and causes environmental pollutions and also risks treated fish users, because of its damaging effects on the human health [2]. To eliminate dyes from wastewater, an extensive variety of remediation technologies have been used. Among them, adsorption process with low price, facile operation, and excellent removal efficiency is known as the most favorable and broadly used approach in the procedures of wastewater treatment [3]. Several adsorbents have been studied to diminish dye concentration from the polluted water [4–8]. The most important interactions in ⁎

adsorption process, are electrostatic interaction, hydrogen bonding, porosity and pore size, acid-base interactions, and π–π interactions [9]. The graphite-like carbon nitride (named as g-C3N4) as a nonmetal polymeric semiconductor with a narrow band gap of about 2.7 eV, which is applicable to visible light, contains great properties such as physical and chemical stability, high resistance to erosion, multiple surface functionalities, and so on. Hence, g-C3N4 has earned worldwide attention in many applications like hydrogen evolution [10,11], CO2 reduction [12,13], pollutants photodegradation [14,15], optical sensors [16,17], etc. In spite of numerous studies on g-C3N4 as a photocatalyst and catalyst in recent years, there are only a few papers about its adsorption features [18–22]. G-C3N4 with multiple surface features such as electronic properties, basic functions of its nitride groups, and hydrogen bonding motif can act as a perfect adsorbent. However, its low surface area (below 10 m2 g−1) limits its adsorption properties. Consequently, in this research, mesoporous carbon nitride (named as MCN) with great surface area, high pore volume and good mechanical stability was chosen as the adsorbent. To further improve its performance, heteroatoms doping has been a promising method which can enhance the density of surface charges by modifying the electronic configuration of MCN. As far as our group know, there is no research for nonmetal

Corresponding author. E-mail address: [email protected] (A. Badiei).

https://doi.org/10.1016/j.apsusc.2018.11.017 Received 11 April 2018; Received in revised form 17 October 2018; Accepted 2 November 2018 Available online 03 November 2018 0169-4332/ © 2018 Elsevier B.V. All rights reserved.

Applied Surface Science 469 (2019) 236–245

E. Boorboor Azimi et al.

Nomenclature C0 Ce V m Qm Ka KF

T R qt k1 k2 α β Kdif

initial concentration (mg L−1) equilibrium concentration (mg L−1) dye solution volume (L) adsorbent weight (g) maximum adsorption capacity (mg g−1) constant of adsorption equilibrium of Langmuir equation (L mg−1) Freundlich constant (L g−1), constant of sorption

absolute temperature (K) universal gas constant (8.314 J mol−1 K−1) adsorption capacity at time t (mg g−1) 1th order kinetic rate constant (min−1) 2nd order kinetic rate constant (g mg−1 min−1) initial adsorption rate (mg g−1 min−1) desorption coefficient (g mg−1) rate constant of intraparticle diffusion model (mg g−1 min−1/2)

carbonized in N2 flow at 600 °C with a rate of 3 °C min−1 for 5 h. After cooling it to room temperature, the hard template was removed by hydrofluoric acid (5 wt%), then filtered and washed with some distilled water and ethanol about 3 times and was put in the oven at 100 °C overnight [26]. To synthesis BMCN, 0.5 g of MCN was added to 40 mL solution of H3BO3 with a certain amount, and was put in an ultrasound bath for about 30 min. Then it was heated while stirring to evaporate water. Next, the remained solid was located in a crucible and heated with a rate of 3.3 °C min−1 under N2 atmosphere to reach 600 °C, kept at this temperature for 3 h, then chilled to room temperature. The product was named as BMCNx, so that x is the boric acid wt% to MCN.

doping of MCN to be performed for adsorption of malachite green. Among of the nonmetals, boron with electron lacking and Lewis acidic properties brings an acid-base contrast to carbon nitride, which may enhance the adsorption of several dyes, especially with basic features. Optimizing experimental processes has altered remarkably in recent years. Nowadays, researchers usually prefer using multivariate techniques to difficult univariate procedures. The main drawback of the onevariable-at-a-time technique is its disability to study the interactions between variables. Consequently, the complete effects of every factor on the response won’t be represented [23]. The multivariate methods have many advantages like the potential to attain information about interactions between variables, and the need to fewer essential experiments, which will lessen time and cost [24]. The most applicable multivariate method used in optimization is RSM, which is a mixture of statistical and mathematical techniques. In this work, the MG adsorption onto B doped MCN has been studied. To reach this goal, MCN was prepared, and modified with different B doping weight ratios to find the best adsorbent for MG and to investigate the influence of boron doped amount in the adsorption process. To optimize some important adsorption conditions, RSM was applied. Independently, adsorption kinetics and isotherms were studied using various models, and the possible mechanisms for MG adsorption onto BMCN1 were debated.

2.3. RSm Central composite design (called as CCD) is an appropriate RSM design applied in this research to recognize the effects of some independent variables for the MG adsorption onto the ideal adsorbent. The variables, chosen for this study, were the sorbent weight (A), pH (B), the concentration of MG dye (C) and temperature (D). The range used for these variables were recorded in Table 1. A 24 polynomial CCD was utilized for four variables each of them at five levels, and six repeats at the central point, which leads to 30 runs. 2.4. Adsorption studies

2. Experimental

A standard solution of MG dye with 1000 ppm concentration was prepared. Initial MG concentrations in aqueous solution were prepared from 20 to 500 mg L−1 by dilution of the stock solution. The MG adsorption by BMCN was investigated in batch style adsorption equilibrium tests. The adsorption experiments were executed in some beakers that contain 10 mL MG at various concentrations for 30 min with constant stirring speed of 500 rpm. A constant quantity of BMCN was poured in each beaker; the beakers were then closed with a lid to lessen evaporation to a minimum. UV–Vis spectrophotometer was utilized to determine the MG concentration at λmax = 617 nm, in the aqueous solution according to the standard curve method. Then the following equations were applied to calculate the amount of the adsorbed MG on BMCN per unit of adsorbent (milligram MG per gram of adsorbent) called qe, qt, and removal efficiency (R%) of MG:

2.1. Methods and materials All the chemicals were pure and used without any further treatment. X-ray diffraction patterns were achieved from 5 to 75°on a Bruker D5000 Advance X-ray diffractometer with Cu Kα radiation (λ = 1.5406 Å, 40 kV, and 40 mA). Small angle XRD patterns were recorded from 0.01 to 0.65° on a Hecus (model S3-MICROpix) X-ray diffractometer with Cu anode equipment (λ = 1.5406 Å, 30 kV, and 0.4 mA). Transmission electron microscopy (TEM) images were prepared by CM30 with an accelerating voltage of 300 kV, and scanning electron microscopy (SEM) micrographs were provided by MIRA3 TESKAN equipped with EDS and elemental mapping. Fourier transform infrared spectroscopy spectra of the samples were achieved by RAYLEIGH WQF- 510A apparatus in the wavenumbers of 4000–600 cm−1. Nitrogen adsorption-desorption isotherms were determined on a Micromeritics apparatus of model MicroActive for TriStar II Plus 2.03 at −196 °C. The specific surface areas were calculated by the BET method, and the pore size distributions were calculated by the BJH model.

q e=

(C 0-Ce )V m

(1)

Table 1 Summary of the design and levels of four independent variables.

2.2. Preparation of boron doped MCN (BMCN)

Factors

To prepare MCN, a hard template, SBA-15, was applied. SBA-15 was synthesized according to the previous researches [25]. Then 1 g of the calcined SBA-15 with 2.7 g of ethylenediamine (EDA) and 6 g of carbon tetrachloride (CTC) were mixed. This mixture was refluxed at 90 °C for about 6 h while stirring. At last, the product which was dark-brown was dried overnight at 60 °C, and then was powdered. After that, it was

Sorbent pH Concentration Temperature

237

Coded factors

A B C D

Units

mg – mg L−1 °C

Variable levels −2

−1

0

+1

+2

6 3 20 20

12 4 60 25

18 5 100 30

24 6 140 35

30 7 180 40

Applied Surface Science 469 (2019) 236–245

E. Boorboor Azimi et al.

qt =

(C0 Ct )V m

R% =

(C0 Ce) ·100 C0

1290, 1600, and around 3350 cm−1 corresponding to the stretching mode of CeN(eC)eC, aromatic rings and the stretching band of NeH or the adsorbed OH groups, respectively (Fig. 5) [28]. The band at 1600 cm−1 for MCN was transferred to 1580 cm−1 for BMCN1, implying that B doping changed the aromatic ring, which possibly means boron was doped in the place of carbon atoms correctly.

(2) (3)

Adsorption isotherms were applied using different MG concentrations from 20 to 500 mg L−1, a fixed amount of sorbent (18 mg), at a constant pH of 5, at room temperature during 45 min. Various isotherm models were experienced for this work, including Freundlich, Langmuir, Temkin, Koble–Corrigan, Sips and Dubinin–Radushkevich to study the experimental data, mechanism, and to recognize the best matched isotherm models. Kinetic experiments were executed at beakers containing 18 mg of the adsorbent in 10 mL of MG solution with a constant concentration (100 mg L−1), at pH 5 which was adjusted by using HCl (0.1 M) or NaOH (0.1 M).

3.3. Model fitting by RSM The obtained results from RSM displayed that the removal percentage is in the range of 63–99.8%. The maximum R% (99.8%) was found at the following experimental conditions: 18 mg BMCN1 (A), pH (B) of 5, MG concentration (C) of 20 mg L−1 and the room temperature (D). RSM resulted in a regression equation which connects the MG adsorption to the variables as following (regarding coded factors):

3. Results and discussion

(R\%)3 = 8.745 E+ 005 + 1.545 E+ 005 A+ 1.074 E+ 005 B 1104.67 D

3.1. Optimization of boron weight ratio

+ 1593.7CD

To find the best adsorbent for MG, MCN was modified with different weight ratios of boron, and adsorption tests were performed. The results, shown in Fig. 1 obviously stated that BMCN1 (1 wt% B doping) was the best adsorbent for MG with the highest removal efficiency (R%) and adsorption capacity (qe). First, with modifying MCN with boron, new acidic sites might be added to the MCN structure, which may attract amines as the basic functions in MG. However, further increasing the boron weight ratio (more than 1 wt%), decreased both R% and qe slowly that it might be because of the cationic nature of MG that repulses less electronic density, which is made by doping boron into the structure. Therefore, BMCN1 was selected as an ideal adsorbent for the following experiments.

88529.77 C

70661.34AB + 20037.99BC + 1170.10BD 58169.43A2

32227.4B2

10957.23C2

(4)

This equation used to predict the response (R%) for the desired levels of each variable. To have a more consistent and interpretable model a power transformation of three should be used. Table 3 shows the value of regression coefficients of the Eq. (4). A p-value < 0.05 and a high F-value would propose a more noteworthy effect on the response [29]. Table 3 displayed that all the coefficients were important, with very small p-values (p < 0.0005) unless D coefficient (temperature) and its interactions with the other variables that they were insignificant. The fitted data for R% for the chosen quadratic model was illustrated in Table 4. The determination coefficient (R2 = 0.9971) by ANOVA (analysis of variance) for the quadratic model designated that only 0.29% of the variations could not be clarified by the model. Another factor that confirms the reliability of the model is adjusted Rsquared that in this case was 0.9947, which stated this model was greatly significant [30]. The R2 was in perfect agreement with the adjusted R2 with a small difference. Evaluation of the errors resulted in a lack of fit value of 2.85 that was insignificant at the confidence level of 95%. Adequate precision, measuring the ratio of signal to noise that was much greater than 4, was desirable for this model. Moreover, the value of C.V. % was rather low, that was an advantageous [31]. Furthermore, the p-value of the model was < 0.0001 (Table 3), representing the model terms were significant. These factors values confirmed a very high accurateness and excellent, trustworthy experimental values, that’s why the selected model was reliable to predict.

3.2. Characterization of the adsorbent To investigate the structure of MCN and BMCN1, their XRD patterns were displayed in Fig. 2. In wide-angle X-ray diffraction patterns the wide-ranging diffraction peak around 25.8° that is associated with the interplanar stacking peaks, characteristic of aromatic compounds in graphitic form, is contributed to the plane of (0 0 2) [27]. This diffraction peak was retained with lower intensity for the XRD pattern of BMCN1, which illustrated that the structure of MCN was not altered after modification. In small-angle XRD patterns (Fig. 2 inset) three diffraction peaks were observed: one intensive (1 0 0) and two weak (1 1 0) and (2 0 0) diffraction peaks, representing that the 2D hexagonal porous structure of SBA-15 template was conserved. The textural features and porous structures of MCN and BMCN1 were examined by N2 adsorption-desorption analysis. In Fig. 3 the hysteresis loops of typical type IV isotherms in the mesoporous materials range were observed for both MCN and BMCN1 that again confirmed that boron doping did not change the porous structure in MCN. Fig. 3 inset showed the pore diameter distribution of the adsorbents. The textural properties of the adsorbents, shown in Table 2, illustrated that the structure of MCN was not changed considerably, possibly because the amount of B doping was low (only 1 wt%) and boron doping has likely accrued at the position of the carbon atoms. To determine the morphology of the surface and elemental distribution of BMCN1, one with the best adsorption performance, TEM and SEM images (Fig. 4a and b) were prepared. The TEM image showed a hexagonal structure of the applied template that was SBA-15, and its uniform mesochannels. The SEM-EDS analysis (Fig. 4c) proved well that B exists in the structure. The element mapping and its respective SEM image, shown in Fig. 4d proved that B atoms distributed well in the structure of BMCN1. FT-IR spectra of the adsorbents showed three characteristic peaks at

Fig. 1. Effect of boron weight ratio on MG adsorption onto MCN. 238

Applied Surface Science 469 (2019) 236–245

E. Boorboor Azimi et al.

MG adsorption by BMCN1. Consequently, the MG adsorption from 20 to 40 °C was neither endothermic nor exothermic. On the other hand, as it can be realized from these surface graphs, the adsorption percentage of MG increased by increasing pH value, possibly the reason is that in acidic media, amino and imine groups on BMCN1 got protonated; Accordingly, an electrostatic repulsion between MG molecules, which are cationic, and the positive protons on BMCN1 caused a reduction in the adsorption. However, the R% is still 80% at pH 3 displaying that electrostatic attraction was not the only parameter affecting the MG adsorption. Moreover, the adsorption percentage of MG increased by increasing the sorbent amount; the existence of more active sites at a higher amount of sorbent caused this enhancement. Additionally, the removal percentage enhanced with reducing the MG concentration, although from the other point of view, increasing the MG concentration increased the adsorption capacity (qe) of the dye, that has been proved by the adsorption isotherm studies. These phenomena may be justified in this way: an increase in the dye concentration accelerated the dye diffusion from the solution into the sorbent owing to a driving force created by the concentration slope; Thus a higher amount of the dye adsorbs on BMCN1, which means increasing the qe. On the other hand, the inverse effect of dye concentration on the R% can be explained by the fact that the active sites on the adsorbent got saturated at the more concentrated dye [2]. Eventually, 100% removal of MG can be predicted for the values of the variables as follows: 18 mg of BMCN1, the dye concentration of 20 mg L−1, pH 5, and the room temperature.

Fig. 2. Wide and small (inset) angle XRD patterns of MCN and BMCN1.

3.5. Adsorption isotherms study Adsorption is a collection of mass transfer procedures, and in general, is defined as the adhesion of the sample to a surface of a liquid or a solid (adsorbent). Relations between adsorbent and adsorbate at a known temperature at equilibrium are termed by isotherms of adsorption. Experimental data may correctly fit to several isotherm models to determine an appropriate model for the design process. The obtained parameters from various models provide important data about the mechanism, properties of the surface and the sorbent affinities [32]. In this work, the data obtained from experiments, were verified with some isotherm models to evalaute their applicability. The values of the coefficients, calculated by these isotherm models, were recorded in Table 5. The Langmuir’s isotherm model [33] was researched to estimate the Qm. It represents a comprehensive monolayer exposure on the sorbent’s surface. The linear form of Langmuir’s isotherm may be written as follows:

Fig. 3. N2 adsorption-desorption isotherm and pore diameter distributions (inset) of MCN and BMCN1. Table 2 Textural properties of MCN and BMCN1. Samples

BET surface area (m2 g−1)

BJH pore size (nm)

Total pore volume (cm3 g−1)

Ce 1 1 = + Ce qe K a Qm Qm

MCN BMCN1

439 431

5.1 4.9

0.50 0.46

therefore, with drawing a plot of Ce/qe against Ce (Fig. 7) the unknown coefficients will be found. The Freundlich’s isotherm [34] is an experimental equation that describes heterogeneous structures and is stated as follows:

3.4. Interactive effects of the parameters involved in MG adsorption

Ln(q e) = Ln(KF) +

RSM was applied to evaluate the effects of MG concentration, temperature, pH, and BMCN1 weight on the removal percentage. All the runs were performed at a constant contact time of 45 min to make sure that the adsorption reaches its equilibrium. The 3D surface plots explain the major effects of every variable on the response and on the other variables. These graphs were drawn by setting two variables at their middle levels and changing two other variables and predicting the response (Fig. 6). The results discovered that the interaction of the sorbent weight with pH, and with dye concertation (oval contour plots) were strong; whereas, the temperature in the selected range had an ignorable effect on other variables whether dependent or independent. These results have been already proved by their p-values. As a result, the selected range of temperature may not play a noteworthy role in the

1 Ln(Ce) nF

(5)

(6)

KF is associated with bonding energy and shows the amount of dye adsorbed on the adsorbent. The nF value designates the degree of nonlinearity, between dye concentration and adsorption so that: if nF = 1, the adsorption linearity is confirmed; if nF < 1, the adsorption process will be chemical and if nF greater than 1, this is devoted to a physical adsorption [35]. Concerning Table 5, the value of nF was greater than unity, confirming that the MG adsorption by BMCN1 might be a physical procedure. The above results and the R2 value approved that the equilibrium of the adsorption described by the Freundlich’s isotherm was better than Langmuir’s isotherm. As a result, the MG adsorption might carry out on the heterogeneous surface of BMCN1, while the adsorbed molecules 239

Applied Surface Science 469 (2019) 236–245

E. Boorboor Azimi et al.

Fig. 4. (a) TEM image, (b) SEM image (c) EDS analysis, and (d) Elemental mapping and its related SEM image of BMCN1. Table 4 Statistical parameters for R% for the selected quadratic model by ANOVA. Lack of Fit Adeq Precision C.V. %*

2.85 70.130 1.94

R-Squared Adj R-Squared Pred R-Squared

0.9971 0.9947 0.9870

* Coefficient of variation.

Table 3 ANOVA for response surface model of reduced quadratic. Sum of squares

Mean squares

F value

p value

A B C D AB AC AD BC BD CD A2 B2 C2

5.726E+011 2.766E+011 1.881E+011 2.929E+007 7.989E+010 6.281E+010 9.794E+006 6.424E+009 2.191E+007 4.064E+007 9.474E+010 2.908E+010 3.362E+009

5.726E+011 2.766E+011 1.881E+011 2.929E+007 7.989E+010 6.281E+010 9.794E+006 6.424E+009 2.191E+007 4.064E+007 9.474E+010 2.908E+010 3.362E+009

2934.67 1417.65 963.99 0.12 409.42 321.88 0.041 32.92 0.092 0.17 485.55 149.04 17.23

< 0.0001 < 0.0001 < 0.0001 0.7301 < 0.0001 < 0.0001 0.8417 < 0.0001 0.7653 0.6847 < 0.0001 < 0.0001 0.0005

aCne 1 + bCne

(7)

qe =

Qm K S Cens 1 + Ks Cens

(8)

all the parameters were evaluated by IBM SPSS Statistics version 24.0 software and listed in Table 5. The high values of correlation coefficients, 0.997 of the said models indicated the goodness of fit of the experimental data, achieved for the adsorption of dye on BMCN1, to these two models. Fig. 8 showed that these isotherms were the bestfitted models to the experimental data. For the Sips model in case nS = 1, the model readily converts to the Langmuir’s equation. Otherwise, when either Ce or KS approaches zero, the isotherm becomes a Freundlich’s isotherm. Subsequently, by KS about 0, Freundlich’s model was better matched with the adsorption system. Temkin’s isotherm equation [38] has a factor that considers adsorption interactions and was called heat of adsorption. Temkin’s equation is written as follows:

Fig. 5. FT-IR spectra of MCN and BMCN1.

Source

qe =

qe = BT LnAT + BT LnCe

(9)

where BT = RT/bT and bT is linked to the adsorption heat. The constants and coefficients of the linear isotherm were evaluated by drawing a plot of qe versus Ln(Ce) as presented in Fig. 7 and were stated in Table 5. Evaluation of the data expresses that Temkin’s isotherm was not applicable to the MG adsorption onto BMCN1 because of the relatively low correlation coefficient (Fig. 8). The last equation analyzed here was Dubinin–Radushkevich’s isotherm [39] which is stated as follows:

were interacting with each other. The three parameters models that they depend on the special combination of the Langmuir’s and Freundlich’s models, are KobleCorrigan and Sips [36,37] and presented as follows respectively:

Ln(qe ) = Ln(Qm)

K

2

(10)

where K is connected to the energy of adsorption and ε is the 240

Applied Surface Science 469 (2019) 236–245

E. Boorboor Azimi et al.

a

b

100

100

R% (%)

75

87.5 75

R% (%)

50

25

62.5 50

3

24

180

4

18

18

140

100

5

A: sorbent (mg)

12

6 6

c

30 24

30

B: pH

7

12

60

20

C: Conc. (ppm)

d

100 90

A: sorbent (mg)

6

100

R% (%)

R % (%)

87.5

80 70 60

75 62.5 50 7

30

40 36

24 32

D: Temp. (C)

140

12

24 20

A: sorbent (mg)

6

5

100

C: Conc. (ppm)

f

100

90

80

80

70 60

B: pH

4

60 20

3

100

90

R % (%)

R % (%)

6

18

28

e

180

70 60

20

7

20

24

6 5

B: pH

36 3

32

100

32 4

40 36

60

28

D: Temp. (C)

C: Conc. (ppm)

28

140

24 180

40

20

D: Temp. (C)

Fig. 6. 3D surfaces representing the effect of a) sorbent weight and pH, b) sorbent weight and MG concentration, c) sorbent weight and temperature, d) pH and MG concentration, e) pH and temperature, f) MG concentration and temperature.

Table 5 Comparison of the isotherm parameters for MG adsorption by BMCN1. Isotherm model

Parameter

Langmuir

Qm (mg g−1) Ka (L mg−1) R2 Qm (mg g−1) K × 10−4 E (kJ mol−1) R2 Qm (mg g−1) Ks ns R2

Dubinin–Radushkevich

Sips

309.6 2.06 × 10−3 0.944 119.73 8.39 24.41 0.842 205.32 0.001 1.275 0.997

241

Isotherm model

Parameter

Freundlich

nF KF (L g−1) R2 A B N R2 AT bT R2

Koble-Corrigan

Temkin

1.216 1.103 0.980 0.204 0.001 1.275 0.997 0.042 54.13 0.946

Applied Surface Science 469 (2019) 236–245

E. Boorboor Azimi et al.

Fig. 7. Plots of Langmuir, Freundlich, Dubinin-Radushkevich, and Temkin isotherm models for MG adsorption onto BMCN1.

Fig. 8. Plots of qe versus Ce achieved from experimental data and qe calculated from different isotherm models.

Fig. 9. The contact time effect on the adsorption capacity and removal percentage (inset) of MG adsorption, at 100 mg L−1 MG concentration, the temperature of 30 °C, pH 5 and 18 mg sorbent weight.

Polanyi’s potential calculated by:

= RTLn(1 +

1 ) Ce

Dubinin–Radushkevich’s equation denoted the poorest fit to the experiments.

(11)

The mean free energy (E) was calculated from the constant K as E = 1/√2K. All the parameters were written in Table 5, by drawing a plot of Ln(qe) versus ε2 (Fig. 7). The correlation coefficient value was much fewer than other isotherms R2 as was observed in Fig. 8, so

3.6. The effect of contact time on adsorption Fig. 9 expressed the influence of contact time on the removal efficiency and the adsorption capacity. The adsorption of the dye was 242

Applied Surface Science 469 (2019) 236–245

E. Boorboor Azimi et al.

Table 6 Comparison between maximum adsorption capacities for MG for different adsorbents and BMCN1. Entry *no.

Adsorbent

Qm (mg g−1)

Time (min)

References

1 2 3 4

Activated carbon Fe3O4/activated carbon Reduced graphene oxide Cellulose nanofibril aerogels Graphene oxide-Fe3O4 Chitin hydrogel Modified carbon nanotubes Rice bran Modified rice husk Powdered activated carbon Organoclay Lignin sulfonate-g-poly Nano-ZnO/pollen grain B doped MCN

200 36.36 476.2 212.7

20 40 75 40

[40] [41] [42] [43]

179.15 33.57 172

60 1800 150

[44] [45] [46]

147.47 17 222.22

50–60 20 15

[47] [2] [1]

40.48 107 145.9 309.6

180 150 180 30

[48] [49] [50] This work

5 6 7 8 9 10 11 12 13 14

Fig. 10. Intraparticle diffusion model for the MG adsorption on BMCN1.

qt =

Table 7 Comparison of kinetic model parameters. Pseudo-firstorder Elovich

qe (calc.) k1 R2 Β Α R2

62.64 0.185 0.976 0.126 154.3 0.876

Pseudo-second-order Intraparticle diffusion

qe (calc.) k2 × 10−3 R2 Kdif C R2

t 1 1 = + t qt qe k2 qe2

1

ln t

(14)

(15)

where C is associated with the width of the boundary layer. This model includes three steps: first, penetration of the adsorbate molecules from the aqueous solution to the solid surface, second, intraparticle diffusion, and third, ultimate equilibrium stage [55]. The plot of this equation showed these three linear regions, reflecting that more than one procedure affected the dye adsorption by the adsorbent (Fig. 10). The R2 value, evaluated from the second regression line (Table 7), was so close to 1, demonstrating the application of this model, which may approve that the rate-limiting step was the intraparticle diffusion stage.

A study about the adsorption kinetic is essential to evaluate the mechanism of reaction and to discover the information about optimum operating conditions for adsorption. Thus some kinetic models were employed for MG adsorption. The linear forms for pseudo-first-order [51] and pseudo-secondorder [52] kinetic models are described as follow respectively:

k1·t

)+

qt = K dif t 1/2 + C

3.7. Adsorption kinetic study

qt ) = Ln(qe )

ln(

The plot of this equation was shown in Fig. S3. Elovich’s kinetic model is attributed to chemical adsorption processes, and hence the R2 value of this equation was rather low (Table 4), this model was not applicable to MG adsorption by BMCN1. It meant that the adsorption system may be closer to a physical adsorption. To study the rate determining stage of the MG adsorption by BMCN1, the intraparticle diffusion kinetic model expressed as follows, was tested.

56.18 6.345 0.999 3.47 33.95 0.996

considerably quick in the first 10 min and after that the rate of the process decreased till complete saturation. This was likely owing to the higher amounts of active sites at primary times that reduced as the reaction pushed forward. The equilibrium time of the adsorption happened after 30 min that is a very short time for a pollutant adsorption process, and it would be a great advantage of the prepared adsorbent. Table 6 showed a comparison study between BMCN1 that was researched in this study and some other adsorbents of malachite green dye.

Ln(qe

1

(12)

(13)

The plots of these equations were displayed in Figs. S1 and S2. Table 7 approved that the value of qe calculated by pseudo-2nd-order reaction equation was closer to the qe found by experiments, that was 52.8 mg g−1. Thus the reaction kinetics was probably pseudo-2nd-order [35]. The R2 value of this kinetic model was about unity that confirmed this claim. According to a normal pseudo-2nd-order reaction, the adsorption process may depend on both MG and the adsorbent amount [53]. The Elovich’s kinetic model is another kinetic equation, assessing the MG adsorption, given as follows [54]:

Fig. 11. FTIR spectra of BMCN1 before and after MG adsorption. 243

Applied Surface Science 469 (2019) 236–245

E. Boorboor Azimi et al.

Acknowledgment The authors acknowledge the University of Tehran for assisting this research financially. Appendix A. Supplementary material Supplementary data to this article can be found online at https:// doi.org/10.1016/j.apsusc.2018.11.017. References [1] R. Malik, D.S. Ramteke, S.R. Wate, Adsorption of malachite green on groundnut shell waste based powdered activated carbon, Waste Manage. 27 (2007) 1129–1138. [2] S. Chowdhury, R. Mishra, P. Saha, P. Kushwaha, Adsorption thermodynamics, kinetics and isosteric heat of adsorption of malachite green onto chemically modified rice husk, Desalination 265 (2011) 159–168. [3] K.Y. Foo, B.H. Hameed, Insights into the modeling of adsorption isotherm systems, Chem. Eng. J. 156 (2010) 2–10. [4] M. Zarezadeh-Mehrizi, A. Badiei, Highly efficient removal of basic blue 41 with nanoporous silica, Water Resour. Ind. 5 (2014) 49–57. [5] M. Zarezadeh-Mehrizi, A. Badiei, A.R. Mehrabadi, Ionic liquid functionalized nanoporous silica for removal of anionic dye, J. Mol. Liq. 180 (2013) 95–100. [6] M. Zarezadeh-Mehrizi, A. Badiei, A. Shahbazi, Sulfonate-functionalized nanoporous silica spheres as adsorbent for methylene blue, Res. Chem. Intermed. 42 (2016) 3537–3551. [7] A. Banaei, S. Ebrahimi, H. Vojoudi, S. Karimi, A. Badiei, E. Pourbasheer, Adsorption equilibrium and thermodynamics of anionic reactive dyes from aqueous solutions by using a new modified silica gel with 2,2′-(pentane-1,5-diylbis(oxy))dibenzaldehyde, Chem. Eng. Res. Des. 123 (2017) 50–62. [8] A. Namane, A. Mekarzia, K. Benrachedi, N. Belhaneche-Bensemra, A. Hellal, Determination of the adsorption capacity of activated carbon made from coffee grounds by chemical activation with ZnCl2 and H3PO4, J. Hazard. Mater. 119 (2005) 189–194. [9] Z. Hasan, S.H. Jhung, Removal of hazardous organics from water using metal-organic frameworks (MOFs): plausible mechanisms for selective adsorptions, J. Hazard. Mater. 283 (2015) 329–339. [10] T. Bhowmik, M.K. Kundu, S. Barman, Palladium nanoparticle-graphitic carbon nitride porous synergistic catalyst for hydrogen evolution/oxidation reactions over a broad range of pH and correlation of its catalytic activity with measured hydrogen binding energy, ACS Catal. 6 (2016) 1929–1941. [11] Y. Cui, G. Zhang, Z. Lin, X. Wang, Condensed and low-defected graphitic carbon nitride with enhanced photocatalytic hydrogen evolution under visible light irradiation, Appl. Catal., B 181 (2016) 413–419. [12] D.O. Adekoya, M. Tahir, N.A.S. Amin, g-C3N4/(Cu/TiO2) nanocomposite for enhanced photoreduction of CO2 to CH3OH and HCOOH under UV/visible light, J. CO2 Util. 18 (2017). [13] A. Kumar, P. Kumar, R. Borkar, A. Bansiwal, N. Labhsetwar, S.L. Jain, Metal-organic hybrid: photoreduction of CO2 using graphitic carbon nitride supported heteroleptic iridium complex under visible light irradiation, Carbon 123 (2017) 371–379. [14] J. Su, L. Zhu, P. Geng, G. Chen, Self-assembly graphitic carbon nitride quantum dots anchored on TiO2 nanotube arrays: an efficient heterojunction for pollutants degradation under solar light, J. Hazard. Mater. 316 (2016) 159–168. [15] A. Akhundi, A. Habibi-Yangjeh, Graphitic carbon nitride nanosheets decorated with CuCr2O4 nanoparticles: novel photocatalysts with high performances in visible light degradation of water pollutants, J. Colloid Interface Sci. 504 (2017) 697–710. [16] R. Malik, V.K. Tomer, V. Chaudhary, M.S. Dahiya, A. Sharma, S.P. Nehra, S. Duhan, K. Kailasam, An excellent humidity sensor based on In-SnO2 loaded mesoporous graphitic carbon nitride, J. Mater. Chem. A 5 (2017) 14134–14143. [17] H. Tian, H. Fan, J. Ma, Z. Liu, L. Ma, S. Lei, J. Fang, C. Long, Pt-decorated zinc oxide nanorod arrays with graphitic carbon nitride nanosheets for highly efficient dualfunctional gas sensing, J. Hazard. Mater. 341 (2018) 102–111. [18] M. Anbia, M. Haqshenas, Adsorption studies of Pb(II) and Cu(II) ions on mesoporous carbon nitride functionalized with melamine-based dendrimer amine, Int. J. Environ. Sci. Technol. 12 (2015) 2649–2664. [19] T. Yan, H. Chen, X. Wang, F. Jiang, Adsorption of perfluorooctane sulfonate (PFOS) on mesoporous carbon nitride, RSC Adv. 3 (2013) 22480–22489. [20] H. Chen, T. Yan, F. Jiang, Adsorption of Cr(VI) from aqueous solution on mesoporous carbon nitride, J. Taiwan Inst. Chem. Eng. 45 (2014) 1842–1849. [21] M. Fronczak, M. Krajewska, K. Demby, M. Bystrzejewski, Extraordinary adsorption of methyl blue onto sodium-doped graphitic carbon nitride, J. Phys. Chem. C 121 (2017) 15756–15766. [22] J. Peng, W. Zhang, Y. Liu, Y. Jiang, L. Ni, J. Qiu, Superior adsorption performance of mesoporous carbon nitride for methylene blue and the effect of investigation of different modifications on adsorption capacity, Water, Air, Soil Pollut. 228 (2016) 9. [23] T. Lundstedt, E. Seifert, L. Abramo, B. Thelin, Å. Nyström, J. Pettersen, R. Bergman, Experimental design and optimization, Chemom. Intell. Lab. Syst. 42 (1998) 3–40. [24] N.F. Robaina, S. Soriano, R.J. Cassella, Polyurethane foam loaded with SDS for the adsorption of cationic dyes from aqueous medium: multivariate optimization of the loading process, J. Hazard. Mater. 167 (2009) 653–659.

Fig. 12. Reusability study of BMCN1.

3.8. Adsorption mechanism Adsorption process is related to the nature of the adsorbate and the surface properties of the adsorbent. BMCN with a mesoporous structure and high surface area is able to adsorb MG. Besides, boron added new acidic sites to the structure, which may attract more amine groups as the basic functions in MG as previously discussed. On the other hand, amine and hydroxyl groups on the end edge of the BMCN structure are ready to adsorb MG as a cationic nature dye. FTIR analysis of BMCN1 before and after adsorption of MG was provided in Fig. 11, which showed a shift from 3380 to 3280 cm−1 implying that amine and hydroxyl groups were involved with MG dye after adsorption. 3.9. Reusability The adsorbent reusability is one of the most incredible factors for being an applicable adsorbent from the viewpoint of economic and environment. To evaluate the reusability of the boron doped MCN for adsorption of malachite green, six adsorption-desorption cycles were executed. Desorption of the already adsorbed MG from BMCN1, was performed by washing the adsorbent with ethanol twice, and then another cycle of adsorption was executed. As was viewed in Fig. 12, no significant decrease was accrued even after six cycles. The perfect reusability of BMCN1, made it suitable to adsorb dyes from industrial wastewaters. 4. Conclusion In summary, mesoporous carbon nitride was synthesized via a hard template method with SBA-15, EDA and CTC, and modified by boron doping to remove an organic dye, malachite green, from polluted water. BMCN1 (1 wt% B) showed the highest removal percentage of MG with a pseudo-second-order kinetic model. The RSM results showed the order of the most significant variables varied as dye concentration, sorbent weight and pH for this model and temperature has no observed effect on the adsorption in the studied range. The MG adsorption onto BMCN1 was matched well to the Freundlich, Koble-Corrigan, and Sips isotherm models. The adsorption was a physical process, happened at the heterogeneous surface of BMCN1 and reached the equilibrium time very fast at just 30 min. The rate determining stage in the MG adsorption by BMCN1, was possibly the intraparticle diffusion stage. Moreover, BMCN1 can be facilely regenerated by washing with EtOH and reused repeatedly for MG adsorption without decreasing its efficiency. A very high adsorption capacity and removal efficiency, and excellent recyclability, makes BMCN1 a fascinating candidate for industrial applications as adsorbent. 244

Applied Surface Science 469 (2019) 236–245

E. Boorboor Azimi et al. [25] M. Hartmann, A. Vinu, Mechanical stability and porosity analysis of large-pore SBA15 mesoporous molecular sieves by mercury porosimetry and organics adsorption, Langmuir 18 (2002) 8010–8016. [26] X. Jin, V.V. Balasubramanian, S.T. Selvan, D.P. Sawant, M.A. Chari, G.Q. Lu, A. Vinu, Highly ordered mesoporous carbon nitride nanoparticles with high nitrogen content: a metal-free basic catalyst, Angew. Chem. Int. Ed. 48 (2009) 7884–7887. [27] A. Vinu, K. Ariga, T. Mori, T. Nakanishi, S. Hishita, D. Golberg, Y. Bando, Preparation and characterization of well-ordered hexagonal mesoporous carbon nitride, Adv. Mater. 17 (2005) 1648–1652. [28] E.J. Acosta, S.O. Gonzalez, E.E. Simanek, Synthesis, characterization, and application of melamine-based dendrimers supported on silica gel, J. Polym. Sci. Part A: Polym. Chem. 43 (2005) 168–177. [29] L. Quanhong, F. Caili, Application of response surface methodology for extraction optimization of germinant pumpkin seeds protein, Food Chem. 92 (2005) 701–706. [30] R. Mohammadi, M.A. Mohammadifar, A.M. Mortazavian, M. Rouhi, J.B. Ghasemi, Z. Delshadian, Extraction optimization of pepsin-soluble collagen from eggshell membrane by response surface methodology (RSM), Food Chem. 190 (2016) 186–193. [31] Y. Song, B. Du, T. Zhou, B. Han, F. Yu, R. Yang, X. Hu, Y. Ni, Q. Li, Optimization of extraction process by response surface methodology and preliminary structural analysis of polysaccharides from defatted peanut (Arachis hypogaea) cakes, Carbohydr. Res. 346 (2011) 305–310. [32] A. Khaled, A.E. Nemr, A. El-Sikaily, O. Abdelwahab, Removal of Direct N Blue-106 from artificial textile dye effluent using activated carbon from orange peel: adsorption isotherm and kinetic studies, J. Hazard. Mater. 165 (2009) 100–110. [33] I. Langmuir, The constitution, and fundamental properties of solids and liquids. Part I. Solids, J. Am. Chem. Soc. 38 (1916) 2221–2295. [34] H. Freundlich, Over the adsorption in solution, J. Phys. Chem. 57 (1906) 1100–1107. [35] G. Crini, H.N. Peindy, F. Gimbert, C. Robert, Removal of C.I. Basic Green 4 (Malachite Green) from aqueous solutions by adsorption using cyclodextrin-based adsorbent: kinetic and equilibrium studies, Sep. Purif. Technol. 53 (2007) 97–110. [36] R.A. Koble, T.E. Corrigan, Adsorption isotherms for pure hydrocarbons, Ind. Eng. Chem. 44 (1952) 383–387. [37] M. Belhachemi, F. Addoun, Comparative adsorption isotherms and modeling of methylene blue onto activated carbons, Appl. Water Sci. 1 (2011) 111–117. [38] M. Temkin, V. Pyzhev, Kinetics of ammonia synthesis on promoted iron catalysts, Acta Physicochim. URSS 12 (1940) 217–222. [39] C.I. Pearce, J.R. Lloyd, J.T. Guthrie, The removal of colour from textile wastewater using whole bacterial cells: a review, Dyes Pigm. 58 (2003) 179–196. [40] Y. Önal, C. Akmil-Başar, Ç. Sarıcı-Özdemir, Investigation kinetics mechanisms of adsorption malachite green onto activated carbon, J. Hazard. Mater. 146 (2007)

194–203. [41] D. Datta, Ö. Kerkez Kuyumcu, Ş.S. Bayazit, M. Abdel Salam, Adsorptive removal of malachite green and Rhodamine B dyes on Fe3O4/activated carbon composite, J. Dispersion Sci. Technol. 38 (2017) 1556–1562. [42] K. Gupta, O.P. Khatri, Reduced graphene oxide as an effective adsorbent for removal of malachite green dye: plausible adsorption pathways, J. Colloid Interface Sci. 501 (2017) 11–21. [43] F. Jiang, D.M. Dinh, Y.-L. Hsieh, Adsorption and desorption of cationic malachite green dye on cellulose nanofibril aerogels, Carbohydr. Polym. 173 (2017) 286–294. [44] M.S. Raghu, K. Yogesh Kumar, M.K. Prashanth, B.P. Prasanna, R. Vinuth, C.B. Pradeep Kumar, Adsorption and antimicrobial studies of chemically bonded magnetic graphene oxide-Fe3O4 nanocomposite for water purification, J. Water Process Eng. 17 (2017) 22–31. [45] H. Tang, W. Zhou, L. Zhang, Adsorption isotherms and kinetics studies of malachite green on chitin hydrogels, J. Hazard. Mater. 209 (2012) 218–225. [46] A. Awadallah-F, Adsorptive removal of malachite green chloride and reactive red198 from aqueous solutions by using multiwall carbon nanotubes-graft-poly (2acrylamido-2-methyl-1-propanesulfonic acid), J. Polym. Environ. 25 (2017) 258–276. [47] H.N. Bhatti, A. Jabeen, M. Iqbal, S. Noreen, Z. Naseem, Adsorptive behavior of rice bran-based composites for malachite green dye: Isotherm, kinetic and thermodynamic studies, J. Mol. Liq. 237 (2017) 322–333. [48] S. Arellano-Cárdenas, S. López-Cortez, M. Cornejo-Mazón, J.C. Mares-Gutiérrez, Study of malachite green adsorption by organically modified clay using a batch method, Appl. Surf. Sci. 280 (2013) 74–78. [49] Y. Tang, Y. Zeng, T. Hu, Q. Zhou, Y. Peng, Preparation of lignin sulfonate-based mesoporous materials for adsorbing malachite green from aqueous solution, J. Environ. Chem. Eng. 4 (2016) 2900–2910. [50] G. Tzvetkov, N. Kaneva, T. Spassov, Room-temperature fabrication of core-shell nano-ZnO/pollen grain biocomposite for adsorptive removal of organic dye from water, Appl. Surf. Sci. 400 (2017) 481–491. [51] H. Yuh-Shan, Citation review of Lagergren kinetic rate equation on adsorption reactions, Scientometrics 59 (2004) 171–177. [52] Y.S. Ho, G. McKay, Sorption of dye from aqueous solution by peat, Chem. Eng. J. 70 (1998) 115–124. [53] A. Ausavasukhi, C. Kampoosaen, O. Kengnok, Adsorption characteristics of Congo red on carbonized leonardite, J. Clean. Prod. 134 (2016) 506–514. [54] S.H. Chien, W.R. Clayton, Application of elovich equation to the kinetics of phosphate release and sorption in soils, Soil Sci. Soc. Am. J. 44 (1980) 265–268. [55] G. McKay, The adsorption of dyestuffs from aqueous solution using activated carbon: analytical solution for batch adsorption based on external mass transfer and pore diffusion, Chem. Eng. J. 27 (1983) 187–196.

245

Suggest Documents