Open Access

Response surface modeling of lead (׀׀) removal by graphene oxide-Fe3O4 nanocomposite using central composite design

  • Mohammad Khazaei1,
  • Simin Nasseri1, 2Email author,
  • Mohammad Reza Ganjali3, 4,
  • Mehdi Khoobi5,
  • Ramin Nabizadeh1,
  • Amir Hossein Mahvi1, 2, 6,
  • Shahrokh Nazmara1 and
  • Elham Gholibegloo7
Journal of Environmental Health Science and Engineering201614:2

https://doi.org/10.1186/s40201-016-0243-1

Received: 27 September 2015

Accepted: 12 January 2016

Published: 22 January 2016

Abstract

Background

Magnetic graphene oxide (Fe3O4@SiO2-GO) nanocomposite was fabricated through a facile process and its application as an excellent adsorbent for lead (II) removal was also demonstrated by applying response surface methodology (RSM).

Methods

Fe3O4@SiO2-GO nanocomposite was synthesized and characterized properly. The effects of four independent variables, initial pH of solution (3.5–8.5), nanocomposite dosage (1–60 mg L−1), contact time (2–30 min), and initial lead (II) ion concentration (0.5–5 mg L−1) on the lead (II) removal efficiency were investigated and the process was optimized using RSM. Using central composite design (CCD), 44 experiments were carried out and the process response was modeled using a quadratic equation as function of the variables.

Results

The optimum values of the variables were found to be 6.9, 30.5 mg L−1, 16 min, and 2.49 mg L−1 for pH, adsorbent dosage, contact time, and lead (II) initial concentration, respectively. The amount of adsorbed lead (II) after 16 min was recorded as high as 505.81 mg g−1 for 90 mg L−1 initial lead (II) ion concentration. The Sips isotherm was found to provide a good fit with the adsorption data (KS = 256 L mg−1, nS = 0.57, qm = 598.4 mg g−1, and R2 = 0.984). The mean free energy Eads was 9.901 kJ/mol which confirmed the chemisorption mechanism. The kinetic study determined an appropriate compliance of experimental data with the double exponential kinetic model (R2 = 0.982).

Conclusions

Quadratic and reduced models were examined to correlate the variables with the removal efficiency of Fe3O4@SiO2-GO. According to the analysis of variance, the most influential factors were identified as pH and contact time. At the optimum condition, the adsorption yield was achieved up to nearly 100 %.

Keywords

Graphene oxideAdsorptionLead (II)OptimizationCentral composite design

Background

Effluents containing Lead and other toxic metals (׀׀) are increasingly discharged into the water supplies due to the expansion of industries [1]. The maximum levels lower than 15 ppb for lead (׀׀) in drinking waters has been mandated by many environmental agencies and national standard organizations [24]. The strict limitations on discharging effluents contained lead (׀׀) to the natural water bodies are attributed to the lead (׀׀) potential health effects on children and adults [3].

Many processes such as precipitation, membrane filtration, adsorption, and ion exchange have been applied to remove lead (׀׀) and other toxic metals from the industrial effluents [5]. Only a few methods such as using functionalized adsorbents and membrane technologies can be adopted to capture low concentrations around 1 mg L−1, which is commonly occurred in drinking water sources [6]. Although, adsorption processes are useful in removal low concentrations of metal ions from aqueous solutions, but there are two main limitations regarding to the use of them; 1. low adsorption capacity [7, 8], and 2. difficult separation of adsorbent from treated water after the end of adsorption process [911].

Graphene oxide is an emerging carbon-based nonmaterial that has revealed the promising adsorptive properties [12, 13]. Graphene oxide (GO) creates a highly stable aqueous dispersion which prepares an excellent situation for effective contacts with target contaminants without needing to vigorous mechanical mixing [14]. The GO flakes have high specific surface area ranging from 600 to 3500 m2 g−1 [15, 16]. The dispersibility property of GO is attributed to the plenty of hydrophilic functional groups on the GO flakes [15]. The GO flake surface contains different functional groups including epoxide and hydroxide, whereas, the edge of flakes are mainly contained a hedge of carboxylic groups [14].

Using magnetic agents like Fe3O4 has been considered as a way to separate the GO nanosheets from aqueous solution when the adsorption process is finished [17, 18]. Some methods employed for the adding of Fe3O4 on the GO surface are generally led to form reduced GO (rGO) [19, 20]. Because of the elimination of functional groups during the reduction process, rGO represents weak dispersity [14]. Hence, preserving the GO dispersibility in the aqueous solution as well as adding the magnetic property for separation purposes is under consideration.

Few literatures were reported applying non reduced Fe3O4/GO for the adsorption purposes [17, 21, 22]. Among them, some synthesis approaches have relied on the formation of covalent bonds between the GO sheets and Fe3O4 nanoparticles [17, 22] which has more stability than those methods based on physical attraction [21].

This research aimed to fabricate the covalent bond Fe3O4@SiO2-GO nanocomposite as a highly dispersible and easy separatable adsorbent for the elimination of lead (׀׀) from aqueous solution. Other purpose of the study was determining the optimal operational condition using response surface methodology (RSM) to achieve satisfactory lead (׀׀) removal. The conventional optimization method, which altered one variable at a time by keeping the other variables constant, is a time consuming and costly approach that can not consider the interactive effects between variables. RSM technique is an empirical statistical approach used to evaluate the relationship between a set of controlled experimental variables and observed results. It can be applied to optimize and identify the performance of adsorption process. Minimum experimental runs are achievable by using RSM. Applying RSM reduces the experiment runs and the reagents consumption. It also facilitates the execution of experiments necessary for the construction of the response surface.

Methods

Materials

Graphite powder (particle size ˂ 20 μm), tetraethyl orthosilicate (TEOS), (3-aminopropyl) triethoxysilane (APTES), n- hydroxysuccinimide (NHS) and 1- ethyl-3- (3-dimethyl aminopropyl) carbodiimide (EDC.HCl) were purchased from Sigma- Aldrich, Ltd. Co. All other chemicals such as sodium nitrate (NaNO3), potassium permanganate (KMnO4), sulfuric acid (H2SO4), hydrochloric acid (HCl), hydrogen peroxide aqueous solution (H2O2), iron chloride hexahydrate (FeCl3, 6 H2O), and iron chloride tetrahydrate (FeCl2, 4 H2O) were of reagent grade and used without further purification.

Preparation of graphene oxide (GO)

Graphene oxide was synthesized from the graphite powder by the modified Hummers et al. method [23]. Briefly, 2.0 g of graphite powder and 2.0 g of NaNO3 were mixed with 92 mL of H2SO4 (98 %) in a flask and stirred in an ice bath vigorously for 0.5 h, and then 12.0 g of KMnO4 was added to the above solution slowly. After stirring for 0.5 h, the ice bath was removed and the solution was stirred in a water bath at 35 °C for 6 h. After that, 160 mL of the DI water was added slowly to the flask. Then, the obtained mixture was stirred at 900Cfor 2 h. Afterward 400 mL of DI water was added and followed by addition of 12 mL of H2O2 (30 %), Upon which the color of mixture turned to bright yellow. The obtained suspension was washed with 1:10 HCl solution (150 mL) and DI water several times to remove metal ions [24]. The resultant dispersion was sonicated at 130 KHz for 2 h and centrifuged to obtain exfoliated graphene oxide [25].

Preparation of Fe3O4@SiO2-NH2

The Fe3O4 magnetic nanoparticles were synthesized using a coprecipitation method [26]. For the synthesis of Fe3O4@SiO2- NH2, 1.0 g of the obtained Fe3O4MNPs was dispersed in a mixture of 40 mL ethanol and 10 mL of DI water using an ultrasonic water bath. After that, o.5 mL TEOS and 2 mL NH3.H2O (25 %) were added, and the mixture was stirred at 50 °C for 6 h. The solid product was collected by an external magnetic field, washed with ethanol and dried under vacuum. In the next step, 1 g of the obtained Fe3O4@SiO2 was dispersed in 25 mL dried toluene and treated with addition of 1 mL APTES [27]. The mixture was refluxed for 24 h under nitrogen atmosphere. The product was washed with ethanol and then dried to obtain Fe3O4@SiO2-NH2 [28].

Preparation of Fe3O4@SiO2-GO

The condensation reaction between amine groups of Fe3O4@SiO2-NH2 and carboxyl groups of GO was performed [28]. Typically, 0.2 g GO was dispersed in 50 mL DI water containing of 0.1 g NHS and 0.2 g EDC.HCl by ultrasonication for 2 h. Subsequently, 0.5 g Fe3O4@SiO2- NH2was added to the above mixture and stirred for 12 h at room temperature. The solid product was collected and washed with DI water and ethanol by magnetic separation and then dried under vacuum [17, 29]. Figure 1 shows a schematic view of the synthesis process.
Fig. 1

Schematic of the chemical path to the synthesis of Fe3O4@SiO2-GO nanocomposite. (TEOS: tetraethylorthosilicate, APTES: 3 aminopropyltriethoxysilane)

To confirm the stability of nanocomposite, concentration of Iron after the adsorption process was measured. As shown in Tables 2, and 3, the leaching of Iron into the aqueous solution after contact times was negligible.

Characterization

The SEM images were taken with Hitachi- S4160 scanning microscope (Tokyo, Japan) to survey the morphological pattern and surface structural aspects of GO and Fe3O4@SiO2-GO nanocomposite. A Nanoscope V multimode atomic force microscope (Veeco Instruments, USA) were used to perform AFM measurements. The AFM images were taken from samples which prepared by deposition a dispersed GO/methanol solution (70 mg mL−1) onto a mica surface and allowing them to dry in air [25]. The images were taken under ambient condition by adjusting the instrument on the tapping mode.

Batch adsorption experiments

Using a thermostatic shaker, batch experiments were conducted in 100 mL Erlenmeyer flasks to study the removal of lead (׀׀) on the Fe3O4@SiO2-GO nanocomposite. The different volumes contained known quantities of as-dispersed Fe3O4@SiO2-GO nanocomposite were added to 20 mL solution having predominated concentrations of Pb2+. All solutions underwent constant mixing at the 300 rpm for different contact times determined by the experimental design. After ending the adsorption process, the nanocomposite was eliminated from the aqueous solution by a magnet. The equation (1) was applied to determine the removal efficiency of lead (׀׀):
$$ \mathrm{R}\left(\%\right)=\frac{\left({C}_0-{C}_e\right)}{C_0}\times 100 $$
(1)

Where, R (%) is the removal efficiency, C 0 and C t are the concentrations (as mg L−1) of lead (׀׀) at 0 and t minutes after the contact time, respectively.

The equilibrium adsorption capacity was also obtained as equation (2):
$$ {\mathrm{q}}_e=\frac{\left({C}_0-{C}_e\right)}{x_{ads}}\times 1000 $$
(2)

Where, q e is the equilibrium capacity (mg g−1), x ads is the nanocomposite concentration in aqueous solution (mg L−1), and 1000 is converting factor (mg g−1).

Lead (׀׀) measurements in the aqueous solution were performed by using a Spectro Arcos ICP-optical emission spectrometer (SPECTRO Analytical Instruments, Kleve, Germany) based on radial plasma observation. The Spectro Arcos has a Paschen–Runge mount which equipped with 32 linear CCD detectors. The CCD detectors supply the ability of simultaneous monitoring of line intensities at wavelengths between 130 and 770 nm.

Isotherm and kinetic constants were obtained using the Solver “add-in” with Microsoft Excel spreadsheet program [30] according to the nonlinear forms of the equations.

Experimental design

Central composite design (CCD) was used to investigate the lead (׀׀) removal. The RSM was employed to evaluate the combined effects of pH (X 1 ), GO-Fe3O4 dose (X 2 ), contact time (X 3 ), and initial lead (׀׀) concentration (X 4 ) on the adsorption process. The experimental conditions of independent variables, which were derived from CCD, are summarized in Table 1. The lead (׀׀) removal efficiency (Y) were served as output responses. Applying two blocks, one cube block and one star block, total 44 experiments were carried out, consisting of 20 center points, 24 = 16 design point, and 2 × 4 = 8 axial points.
Table 1

The original and coded levels of independent variables

Original factors

Coded levels

−1

0

+1

pH-X1

3.5

4.75

6

7.25

8.5

Fe3O4@SiO2-GO -X2 (mg/L)

1

15.75

30.5

46.25

60

Time-X3 (min)

2

9

16

23

30

Initial Pb2+ Concentration-X4 (ppm)

0.5

1.625

2.75

3.87

5

The details of 44 experiments are presented in Table 2. The chosen independent factors applied in the study were coded based on Eq. (3):
Table 2

Observed and predicted values for the quadratic model (T = 298 K)a

Run no.

Observed values (%)

Predicted values (%)

Residual

Run no.

Observed values

Predicted values

Residual

1

53.433

57.917

−4.484

23

98.579

100.000

−1.421

2

98.904

67.699

31.205

24

90.133

83.418

6.715

3

95.230

96.303

−1.074

25

80.601

85.100

−4.500

4

95.600

23.773

71.826

26

98.995

100.000

−1.005

5

95.317

94.748

0.570

27

99.263

95.067

4.196

6

95.375

94.747

0.628

28

94.576

94.739

−0.163

7

95.626

94.757

0.869

29

94.923

94.742

0.182

8

94.884

94.749

0.135

30

89.334

88.112

1.222

9

98.796

94.709

4.087

31

98.009

100.000

−1.991

10

92.002

94.739

−2.737

32

41.095

49.543

−8.448

11

87.971

87.427

0.544

33

93.478

94.836

−1.359

12

56.850

61.083

−4.234

34

95.341

94.742

0.599

13

98.432

66.637

31.796

35

94.931

94.798

0.133

14

99.493

91.859

7.634

36

96.083

99.816

−3.733

15

94.866

94.770

0.096

37

99.695

93.757

5.938

16

99.589

54.557

45.031

38

86.848

85.821

1.026

17

95.032

94.750

0.281

39

95.115

94.757

0.359

18

95.019

94.759

0.260

40

94.402

94.747

−0.345

19

92.438

84.665

7.773

41

94.495

98.033

−3.537

20

92.624

84.440

8.184

42

95.071

94.947

0.124

21

93.237

83.804

9.433

43

95.290

94.968

0.322

22

95.020

94.797

0.223

44

99.006

100.000

−0.994

aThe average concentration of Iron (measured by ICP-OES) after ending contact times for 44 adsorption runs: 4.32 ± 1.01 ppb

$$ {x}_i=\frac{\left({X}_i-{X}_0\right)}{\varDelta X} $$
(3)
Where, x i is a dimensionless coded value of the ith independent variable, X 0 is the center point value of X i and ΔX is the step change value. A quadratic (second order) model as shown in Eq. (4) was applied to approximate the interaction between the response (Y) and four independent variables:
$$ Y={b}_0+{\displaystyle \sum_{i=1}^k}{b}_i{X}_i+{\displaystyle \sum_{i=1}^k}{b}_{ii}{X}_i^2+{\displaystyle \sum_{i=1}^{k-1}}{\displaystyle \sum_{j=2}^k}{b}_{ij}{X}_i{X}_j+c $$
(4)

Where, Y represents the dependent variable (lead (׀׀) removal efficiency), b0 is a constant value, bi, bii, and bij refer to the regression coefficient for linear, second order, and interactive effects, respectively, Xi, and Xj are the independent variables, c denotes the error of prediction.

The above mentioned CCD analysis plus to the statistical analysis, such as ANOVA, F-test, and t-test were obtained using R software (version 3.0.3: 2014-03-06).

Results and discussion

Characterization of GO-Fe3O4 nanocomposite

As shown in Fig. 2, a UV- visible spectrum obtained for the GO aqueous dispersion (orange line) displays a plasmon peak at 231 nm which is related to the π → π* transitions due to the aromatic C − C bonds. Also, a hump can be detected around 300 nm approving the n → π* transitions of C = O bonds [31, 32]. Gradually adding the lead (׀׀) aqueous ions into the GO dispersion resulted in producing a growing humpy pattern around 300 nm which can be attributed to the affinity between lead (׀׀) and C = O bonds relating to the carboxylic groups in the GO structure [33].
Fig. 2

UV-visible spectra for GO dispersion in water before (Orange line) and after adding the different concentrations of Lead (II). Lead (II) initial concentration = 1 μM, pH = 6.8, T = 298 K

Fabricated Fe3O4@SiO2-GO was characterized with different techniques. Figure 3 shows the field emission SEM images of GO and Fe3O4@SiO2-GO. Figure 3a illustrated the fluffy nature of GO layers which turned to agglomerated morphology after the formation of covalent bonds with Fe3O4@SiO2-NH2 and producing Fe3O4@SiO2-GO (Fig. 3b). The magnetic properties of the prepared nanocomposite were characterized using vibration sample magnetization (VSM). Figure 4 shows the magnetization curve patterns of Fe3O4, Fe3O4@SiO2-NH2, and Fe3O4@SiO2-GO. As inferred from Fig. 4, the maximum saturation magnetizations of Fe3O4, Fe3O4@SiO2-NH2, and Fe3O4@SiO2-GO were 60.2, 43.8, and 22.3 emu g−1, respectively.
Fig. 3

Typical FE-SEM of GO (a), and Fe3O4@SiO2-GO (b)

Fig. 4

Magnetic hysteresis loops of Fe3O4 (a), Fe3O4@SiO2-NH2 (b), and Fe3O4@SiO2-GO (c). Photographic image of 1.4 mg mL−3 Fe3O4@SiO2-GO dispersions in water before (left image) and after (right image) exposing to the external magnetic field for 15 s (d)

The VSM curves (Fig. 4) shows that, the magnetic power of Fe3O4 nanoparticles were dropped into the one third of original state which was due to both the SiO2-NH2 coverage and the GO covalent bonds. But, the remaining 22.3 emu g −1 of saturation magnetization can still be considered as a powerful magnetic field to separate the nanocomposite from the aqueous solution, as shown in the Fig. 4d. Also, the coercivity and remanence were not observed after removing the magnetic field. From Fig. 4d, the yellow brown color of the GO dispersion revealed that the oxygenation of the graphene nanosheets has been effectively occurred during the synthesis [16, 34]. After 3 months from the GO preparation, there is no any visible sign of sedimentation which shows long-term dispersibility of GO in water.

The AFM image of GO nanosheets is illustrated in Fig. 5a. Also, Fig. 5b shows the distribution of GO obtaining 210 nanosheets found in a certain area of the mica surface which confirms preparing a well distributed dispersion. As shown in Fig. 5b, the average thickness measured was 2.74 nm revealed producing few layered (1 and 2 layers) GO [35].
Fig. 5

Tapered mode AFM topography scan. Exfoliated graphene oxide deposited on a freshly cleaved mica surface (a). Histogram of platelet thicknesses from images of 210 platelets (the mean thickness is 2.74 nm) (b). Height profile through the green line (Line 1) presented in (a). Cross-section A-A through the sheet shown in (a) exhibiting a height minimum of 0.748 nm (c)

As shown in Fig. 5c, the thickness of a random GO sheet measured using the height profile (Line 1) in the AFM image, is about 0.75 nm. This sub-nanometer thickness confirms producing the GO monolayer [14].

Figure 6 shows the FT-IR spectra obtained for GO, Fe3O4, and Fe3O4@SiO2-GO materials. Figure 6a depicts the characteristic features illustrated in the FT-IR spectrum for GO which are contained the adsorption bands attributed to the C-O stretching at 1055 cm−1, the C–OH stretching at 1226 cm−1, and the C-O carbonyl stretching at 1733 cm1 [3638]. Furthermore, the O-H hydroxide stretching vibrations appear at 3419 cm−1. Also, the adsorbed water molecules stretching at 1621 cm−1, although it may feature due to the skeletal vibrations of un-oxidized graphitic remnants [39, 40].
Fig. 6

FT-IR spectra of GO (a), Fe3O4 (b), and Fe3O4@SiO2-GO (c)

As can be seen from Fig. 6b, the spectrum of the Fe3O4 shows the Fe-O stretching vibration at 591 cm−1, and an intense OH band around 3400 cm−1. The OH band is attributed to the stretching vibrations of Fe-OH groups attached on the Fe3O4 surface and also can be assigned to the remaining water that was not eliminated from the surface of the Fe3O4 nanoparticles [41].

As depicted in Fig. 6c, the peak at 3401 cm−1 of Fig. 5c attributed to the –NH2 vibration. Comparing Fig. 6c with Fig. 6a, the peak at 1733 cm−1 was almost disappeared, and a new broad peak was emerged at 1641 cm-1 corresponding to C = O characteristic stretching band of the amide group. The stretching band of the amide C–N peak appears at 1230 cm−1 [42]. Meanwhile, as shown in Fig. 6c, the peaks at 802 and 1110 cm−1 were obviously observed due to the Si–O vibrations. From these findings, it is recommended that APTES functionalized Fe3O4 was covalently bonded to GO through the amide linkage [22].

Response surface methodology model analysis

Predicted values for lead (׀׀) removal efficiencies (%) applying quadratic model and reduced quadratic model were represented in Tables 2 and 3, respectively. The statistical significance of models was depicted in Tables 4 and 5 which represent the analysis of variance (ANOVA) for the quadratic model and reduced quadratic model, respectively.
Table 3

Observed and predicted values for the reduced cubic model (T = 298 K)a

Run no.

Observed values (%)

Predicted values (%)

Residual

Run no.

Observed values

Predicted values

Residual

1

53.433

59.426

−5.994

23

98.579

100.000

−1.421

2

98.904

68.393

30.511

24

90.133

83.437

6.696

3

95.230

96.805

−1.575

25

80.601

86.087

−5.486

4

95.600

25.705

69.894

26

98.995

100.000

−1.005

5

95.317

94.994

0.323

27

99.263

95.744

3.519

6

95.375

94.994

0.381

28

94.576

94.994

−0.418

7

95.626

94.994

0.632

29

94.923

94.994

−0.070

8

94.884

94.994

−0.110

30

89.334

89.160

0.175

9

98.796

93.827

4.969

31

98.009

94.994

3.015

10

92.002

94.994

−2.992

32

41.095

50.054

−8.959

11

87.971

86.087

1.884

33

93.478

94.994

−1.516

12

56.850

59.426

−2.577

34

95.341

94.994

0.347

13

98.432

66.406

32.026

35

94.931

94.994

−0.063

14

99.493

93.430

6.063

36

96.083

100.000

−3.917

15

94.866

94.994

−0.128

37

99.695

93.715

5.980

16

99.589

56.162

43.426

38

86.848

88.655

−1.807

17

95.032

94.994

0.038

39

95.115

94.994

0.122

18

95.019

94.994

0.025

40

94.402

94.994

−0.591

19

92.438

84.254

8.184

41

94.495

100.000

−5.505

20

92.624

84.254

8.370

42

95.071

94.994

0.077

21

93.237

83.437

9.800

43

95.290

94.994

0.296

22

95.020

94.994

0.026

44

99.006

94.994

4.012

aThe average concentration of Iron (measured by ICP-OES) after ending contact times for 44 adsorption runs: 2.44 ± 0.12 ppb

Table 4

Analysis of variance (ANOVA) for the quadratic model

Model formula in rsm (X1,X2,X3,X4)

DF

Sum of squares

Mean square

F-value

Probability (P)

First-order response

4

3307.4

826.85

31.7283

<0.0001

Two-way interactions

6

999.2

166.53

6.3903

0.0002

Pure quadratic response

4

1173.2

293.30

11.2546

<0.0001

Residuals

29

755.7

26.06

-

-

Lack of fit

28

750.5

26.80

5.1449

 

Multiple R2 = 0.879; Adjusted R2 = 0.820; Lack of fit: 0.337

Table 5

Analysis of variance (ANOVA) for the reduced quadratic model

Model term

DF

Sum of squares

Mean square

F-value

Probability (P)

pH

1

2859.39

2859.39

120.547

<0.0001

Adsorbent

1

204.08

204.08

8.6037

0.0058

Time

1

241.06

241.06

10.162

<0.0003

pH2

1

1165.24

1165.24

49.124

<0.0001

pH × Adsorbent

1

200.98

200.98

8.4728

<0.01

pH × Time

1

219.01

219.01

9.2330

<0.005

Adsorbent × Time

1

491.94

491.94

20.739

<0.0001

Residuals

36

853.93

23.72

-

-

Multiple R2 = 0.863; Adjusted R2 = 0.836; Lack of fit: 0.355

The reduced quadratic model was applied by omitting the variables assigned to the P-values more than 0.05 in the quadratic model [43].

The values of the determination coefficient (multiple R2) shown in Tables 4 and 5 indicated that 87.9 and 86.3 % of the variability in the response could be explained by the quadratic and reduced quadratic models, respectively.

If there have been various terms in the model and also the sample size has not been very large, the adjusted correlation coefficient (adjusted R2) may represent values considerably smaller than the multiple correlation coefficients (Multiple R2) [44]. In this experiment, the adjusted correlation coefficient value (adjusted R2 = 0.836) are also noticeable to support the high significance of the models and approves a satisfactory adjustment for the reduced quadratic model to the experimental data [4547].

The values of the determination coefficient (multiple R2) shown in Tables 4 and 5 indicated that 87.9 and 86.3 % of the variability in the response could be explained by the quadratic and reduced quadratic models, respectively.

As revealed from Tables 4 and 5, the “lack of fit (LOF)” values were 0.337 and 0.355 for the quadratic and reduced quadratic models, respectively. The insignificant values of LOF (>0.05) and the significant P-values for both models prove that applying models is eligible to interpret the lead (׀׀) removal process and also, the reduced model is the better choice because of the higher adjusted R2 (0.836) and the higher value obtained for LOF (0.355) [43, 46].

Table 6 represents the regression analysis obtained from the reduced quadratic model. As shown, the significant of each coefficient was determined by P-value. Also, Values of “Prob > |t|” less than 0.05 in Table 6 indicate that all the model terms are significant. The bigger amounts of the t-values beside the smaller amounts of P-values show the more significant of the corresponding coefficients [46]. Based on the t- and P –value results, pH and time can be considered as the substantial effective factors on the lead (׀׀) adsorption. The effect of each model term, also, can be observed from the coefficient estimate values represented in Table 6.
Table 6

Regression analysis for the reduced quadratic model

Model term

Coefficient estimate

Std. error

t value

Pr (>|t|)

P-value

Intercept

−205.76106

27.09

−7.59

5.5 × 10−9

<0.001

pH

65.73077

7.051

9.32

3.9 × 10−11

<0.001

Adsorbent dosage

2.21028

0.444

4.98

1.6 × 10−5

<0.001

Time

4.62763

0.920

5.03

1.3 × 10−5

<0.001

pH^2

−3.69755

0.527

−7.01

3.20 × 10−3

<0.001

pH × Adsorbent

−0.19222

0.066

−2.90

6.1 × 10−9

<0.01

pH × Time

−0.42283

0.139

−3.04

4.4 × 10−3

<0.01

Adsorbent × Time

−0.05370

0.012

−4.50

5.8 × 10−5

<0.001

Contour plots depicted in Fig. 7 are the graphical illustrations of the regression analysis (Table 6) which represent the simultaneous effects of adsorbent-pH (a), time-pH (b), and time-adsorbent (c) on lead (׀׀) removal efficiency as the response factor. As noted above, the interaction effects of pH (X1) and Fe3O4@SiO2-GO dose (X2) on the lead (׀׀) removal is shown in Fig. 7a. The contact time and lead (׀׀) initial concentration were fixed constant at 16 min and 2.49 mg/L, respectively. As shown, the lead (׀׀) removal increased with increasing the Fe3O4@SiO2-GO dosage. The maximum lead (׀׀) removal was obtained in the range of pH from 6.5 to 8.5. In that range, the lead (׀׀) removal was independent from the adsorbent dosage. In the pH values between 3–6.5, almost a direct relationship between pH and lead (׀׀) removal was observed.
Fig. 7

Contour plots for the effect of factors on the lead (II) removal. Adsorbent dose (mg L−1) and adsorption pH (a), contact time (min) and adsorption pH (b), adsorption time (min) and adsorbent dose (mg L−1) (c)

The effects of pH (X1) and contact time (X3) on lead (׀׀) removal are shown in Fig. 7b. The adsorbent dose and lead (׀׀) initial concentration were fixed constant at 30.5 mg/L and 2.49 mg/L, respectively. It is inferred from Fig. 7b that around the neutral pH, the lead (׀׀) removal process was almost completed during the contact time up to 10 min. But for pH values less than 4, after the contact time more than 30 min, the removal efficiency near 70 % was achieved.

The effects of Fe3O4@SiO2-GO dosage (X2) and contact time (X3) on lead (׀׀) removal can be observed in Fig. 7c. The pH and lead (׀׀) initial concentration were fixed constant at 6 and 2.49 mg/L, respectively. Figure 7b shows that, for the adsorbent doses more than 40 mg/L, regardless the contact time, the lead (׀׀) removal efficiency revealed the levels permanently beyond 95 %.

Lead (׀׀) removal showed to be very sensitive to changes in the pH both in low and high adsorbent dose. The removal capacity of Fe3O4@SiO2-GO nanocomposite was rapidly increased when the pH increased from 3.5 to 8.5; as it was also reported by Madadrang [15]. pH influences both the surface charges of the functional groups on the surface of graphene oxide and also the species of lead ion in the aqueous solution [48].

Increasing the protonation of functional groups on graphene oxide surface would be happen at acidic conditions and electropositivity of Fe3O4@SiO2-GO surface would retard the adsorption rate, and finally, the removal efficiency of lead (׀׀) can be reduced [18]. As inferred from Fig. 7a-b, when pH is less than 5, the lead (׀׀) removal efficiency was weak. However, the adsorption of lead (׀׀) was enhanced with the increasing pH from 5 to 8.5. Normally, the adsorption capacities of metal ions for most carbon based nanomaterials would increase with increases in the pH value. In this case, lead (׀׀) can be adsorbed onto the graphene oxide surface by reacting lead (׀׀) with − COOH and − OH groups [15, 49].

As shown in the contour plot exhibited in Fig. 7c, regardless the adsorbent dose, more than 75 % of lead (׀׀) adsorption was achieved during the contact time less than 10 min and the adsorption process was completely done after passing twenty minutes. These findings revealed that the fast adsorption rate of the lead (׀׀) can be attributed to the high affinity of lead (׀׀) ions to the hydroxide (−OH), epoxide (−O−) and carboxylic (−COOH) groups on the GO nanosheets [15, 50, 51].

The optimum values of pH, adsorbent dose, and contact time determined by applying the reduced quadratic model were 6.9, 30.49 mg/L, and 16.01 min, respectively. The optimum concentration of initial lead (׀׀) was not obtained from the reduced quadratic model because it was omitted from the model. But the results of quadratic model suggested 2.49 mg L−1 as the optimum value. Further studies such as isotherm and kinetic experiments were investigated according to the abovementioned optimum values obtained from the model.

Adsorption isotherms

In order to investigate the adsorption equilibrium of lead (׀׀), isotherm models consisting of Langmuir (Eq. 5), Freundlich (Eq. 6), and Sips (Eq. 7) were applied. The original forms (nonlinear) of models can be expressed as follows:
$$ {q}_e=\frac{q_m{K}_L{C}_e}{1+{K}_L{C}_e} $$
(5)
$$ {q}_e={K}_f{C}_e^{1/{n}_F} $$
(6)
$$ {q}_e=\frac{q_m{\left(\ {K}_S{C}_e\right)}^{n_S}}{1 + {\left({K}_S{C}_e\right)}^{n_S}} $$
(7)

Where, qe is the amount of lead (׀׀) adsorbed on the absorbent at equilibrium (mg g−1), Ce describes the equilibrium lead (׀׀) concentration (mg L−1), KL is the Langmuir adsorption constant (L mg−1) and qm denotes the maximum adsorption capacity attributeing to the complete monolayer coverage of the adsorbent (mg g−1). KF is the Freundlich constant related to the maximum sorption capacity (mg g−1), also, nF is the Frendlich constant related to the heterogeneity factor. KS (L g−1) is the affinity constant and nS denotes the surface heterogeneity. If nS value is equal to the unity, the Sips isotherm is turned to the Langmuir isotherm, and consequently, the homogeneous adsorption can be modeled. Also, any deviation of nS value from the unity (more than or less than unity) predicts the heterogeneous surface [52, 53].

The adsorption mechanism can be expressed by applying Dubinin–Radushkevich isotherm (Eq.8) which is based on the potential theory and assumes a Gaussian energy distribution.
$$ {q}_e={q}_m \exp \left(-{B}_{DR}{\epsilon}^2\right) $$
(8)
Where, ε is the Polanyi potential given by:
$$ \epsilon = RTln\left(1+\frac{1}{C_e}\right) $$
(9)
The B DR constant (mol2/J2) is related to the mean free energy E ads of the adsorption per molecule when it is transferred to the surface from infinity of the bulk phase.
$$ {E}_{ads}=\frac{1}{\sqrt{2{B}_{DR}}} $$
(10)
Plotting the isotherm models versus the experimental results are depicted in Fig. 8a. The parameters of isotherm models can be observed in Fig. 8b. These parameters were determined according to the nonlinear regression by plotting qe versus Ce assisted by Solver Add-Ins MS Excel [30].
Fig. 8

Fitting results of different isotherms for lead (II) adsorption by Fe3O4@APTES -GO (a), and parameters obtained for the isotherm models (b). q m,exp: 505.8, Adsorbent Dose: 30.5 mg L−1, pH: 6, C0,Lead(II) : 0.5–90 mg L−1, time : 16 min

As shown in Fig. 8b, the Sips isotherm model represents the higher correlation coefficient (R2 = 0.984) comparing with the Langmuir (R2 = 0.964) and Freundlich (R2 = 0.952) models.

The Sips model includes three parameters and has the capability to apply for both the homogeneous and heterogeneous systems [54]. The Sips model (Eq. 5) integrates parameters from both the Langmuir and the Freundlich isotherm. The heterogeneous surface of adsorbent can be considered if the deviation of n S value from the unity be occurred [53, 55]. However, the Sips isotherm moves toward a constant level at high concentrations whereas a pattern of Freundlich model can be observed at low concentrations [55].

According to the experimental data, the maximum adsorption uptake (qm) was 505.8 mg g−1 which indicates the adsorption capacity higher than those reported by studies applying magnetic GO as lead (׀׀) adsorbent [56] and is comparable with the studies using pristine GO [15, 49, 57]. The maximum adsorption uptake (qm) obtained from Sips isotherm model was found to be 598.4 mg g−1 which was more than the values achieved both from the Langmuir model and the experimental data (qm.Langmuir = 497.8 mg g−1, qm,exp = 505.8 mg g−1). This indicates that the Sips model overestimates the qm value which can be due to the heterogeneity characteristic considered in the Sips model. As shown in Fig. 8b, the deviation of nS value from the unity (nS = 0.57) as well as the nF value more than unity (nF = 4.28) can be assigned to the crosslinking effects beside the amount of functionalities such as -COOH and -OH on the adsorbent surface (see FTIR-spectra in Fig. 6). The isotherm curves were L-shaped, which shows the high affinity of surface groups towards lead (׀׀) ions both at low and high concentrations [55]. As revealed from Fig. 8b, the mean free energy Eads was 9.901 Kj/mol which seems to be the evidence of predomination the chemisorption mechanism [58].

Adsorption kinetics

The experimental data were fitted with the original (nonlinear) forms of Lagergren-first-order (Eq. 11), pseudo-second-order (Eq. 12), and Double-exponential kinetic (Eq. 13) equations.
$$ {q}_t={q}_e\left(1- exp\kern0.5em \Big(-{k}_1t\right)\Big) $$
(11)
$$ {q}_t=\frac{K_2{q}_e^2t}{1+{q}_e{k}_2t} $$
(12)
$$ {q}_t={q}_e-\frac{D_1}{x_{ads}} \exp \left(-{k}_{D_1}t\right)-\frac{D_2}{x_{ads}} \exp \left(-{k}_{D_2}t\right) $$
(13)

Where, qt and qe are the sorption capacity (mg g−1) at time t and at the equilibrium time, respectively. k1 and k2 are the pseudo-first-order and pseudo-second-order rate constants, respectively. D 1 and D 2 (mg L−1) are the rapid and slow steps, K D1 and K D2 (min−1) are constants controlling the mechanism of slow and rapid phases, respectively. x ads is the adsorbent dosage (g L−1).

Figure 9a presents the nonlinear curves attributed to the kinetic models by applying equations 11, 12, and 13. Also, the kinetic parameters of lead (׀׀) removal were illustrated in Fig. 9b.
Fig. 9

Fitting results of different kinetic models for lead (II) adsorption by Fe3O4@APTES -GO (a), and parameters obtained from the kinetic models (b). Adsorbent Dose: 30.5 mg L−1, pH: 6, C0,Lead(II) : 2.36 mg L−1

According to the regression coefficient values of kinetic models, it was found that the double exponential kinetic model (R2 = 0.982) obtains a better description to predict the kinetic data of lead (׀׀) than both pseudo-first-order and pseudo-second-order models.

The values of constant parameters of double-exponential kinetic model revealed that the both external diffusion and internal diffusion have substantial effects on the lead (׀׀) sorption using Fe3O4@SiO2-GO nanocomposite [50].

Conclusions

Magnetic Fe3O4@SiO2-GO nanocomposite was synthesized and applied to elimination the lead (׀׀) from aqueous solution. Due to the high loading capacity of GO for metal ions, Fe3O4@SiO2-GO revealed excellent performance in treatment the lead (׀׀) contaminated waters. The removal process was found to be quick and facile, and the lead (׀׀) adsorption process was almost completed up to 10 min contact time.

Main advantages of Fe3O4@SiO2-GO nanocomposite include quick separation performed by using an external magnetic field and the noticeable lead (׀׀) removal capacity (506 mg g−1).

A central composite design (CCD) was applied to investigate the effects of four adsorption variables, namely pH, adsorbent dose, contact time, and initial lead ion concentration on the removal efficiency of lead (׀׀).

Both the quadratic and reduced quadratic models were applied to correlate the variables to the response values. Results from the analysis of response surfaces indicated that pH, time, and the adsorbent dose were found to have significant effects on the removal efficiency of lead (׀׀). The optimization of process was performed and the experimental values were found to be agreed satisfactorily with the predicted values.

The adsorption isotherms and kinetics were also investigated. Equilibrium adsorption data had best fit by the Sips isotherm model and chemisorption mechanism was predominated. Kinetic studies indicated that the double-exponential kinetic model is the preferred model to explain the equilibrium adsorption over the time.

Declarations

Acknowledgements

This research was part of a PhD dissertation of the first author and has been financially supported by a grant (NO, 28232-27-01-94) from Tehran University of Medical Sciences, Tehran, Iran. The authors would like to express their thanks to the Department of Environmental Health Engineering, School of Public Health, Tehran University of Medical Sciences for their collaboration.

Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Authors’ Affiliations

(1)
Department of Environmental Health Engineering, School of Public Health, Tehran University of Medical Sciences
(2)
Center for Water Quality Research, Institute for Environmental Research, Tehran University of Medical Sciences
(3)
Center of Excellence in Electrochemistry, Faculty of Chemistry, University of Tehran
(4)
Biosensor Research Center, Endocrinology & Metabolism Molecular-Cellular Sciences Institute, Tehran University of Medical Sciences
(5)
Department of Medicinal Chemistry, Faculty of Pharmacy and Pharmaceutical Sciences Research Center, Tehran University of Medical Sciences
(6)
Center for Solid Waste Research, Institute for Environmental Research, Tehran University of Medical Sciences
(7)
Department of Chemistry, Faculty of Science, University of Zanjan

References

  1. Yang X, Xu J, Tang X, Liu H, Tian D. A novel electrochemical DNAzyme sensor for the amplified detection of Pb2+ ions. Chem Commun. 2010;46(18):3107–9.View ArticleGoogle Scholar
  2. ISIRI. Drinking water physical and chemical specifications. Iran: Institute of Standards and Industrial Research; 2014.Google Scholar
  3. EPA. Drinking water contaminants. National Primary Drinking Water Regulations 2015 [cited 2015 9/13/2015]; Available from: http://www.epa.gov/safewater/contaminants/index.html, 1/30/2010.
  4. Ju X-J, Zhang S-B, Zhou M-Y, Xie R, Yang L, Chu L-Y. Novel heavy-metal adsorption material: ion-recognition P(NIPAM-co-BCAm) hydrogels for removal of lead(II) ions. J Hazard Mater. 2009;167(1–3):114–8.View ArticleGoogle Scholar
  5. Li Y-H, Di Z, Ding J, Wu D, Luan Z, Zhu Y. Adsorption thermodynamic, kinetic and desorption studies of Pb2+ on carbon nanotubes. Water Res. 2005;39(4):605–9.View ArticleGoogle Scholar
  6. Turan M, Mart U, Yüksel B, Çelik MS. Lead removal in fixed-bed columns by zeolite and sepiolite. Chemosphere. 2005;60(10):1487–92.View ArticleGoogle Scholar
  7. Mahvi A, Gholami F, Nazmara S. Cadmium biosorption from wastewater by Ulmus leaves and their ash. Eur J Sci Res. 2008;23(2):197–203.Google Scholar
  8. Maleki A, Mahvi AH, Zazouli MA, Izanloo H, Barati AH. Aqueous cadmium removal by adsorption on barley hull and barley hull ash. Asian J Chem. 2011;23(3):1373–6.Google Scholar
  9. Sekar M, Sakthi V, Rengaraj S. Kinetics and equilibrium adsorption study of lead(II) onto activated carbon prepared from coconut shell. J Colloid Interface Sci. 2004;279(2):307–13.View ArticleGoogle Scholar
  10. Sprynskyy M, Buszewski B, Terzyk AP, Namieśnik J. Study of the selection mechanism of heavy metal (Pb2+, Cu2+, Ni2+, and Cd2+) adsorption on clinoptilolite. J Colloid Interface Sci. 2006;304(1):21–8.View ArticleGoogle Scholar
  11. Ho YS, McKay G. The kinetics of sorption of divalent metal ions onto sphagnum moss peat. Water Res. 2000;34(3):735–42.View ArticleGoogle Scholar
  12. Rezaee R, Nasseri S, Mahvi A, Nabizadeh R, Mousavi S, Rashidi A, et al. Fabrication and characterization of a polysulfone-graphene oxide nanocomposite membrane for arsenate rejection from water. J Environ Health Sci Eng. 2015;13(1):1–11.View ArticleGoogle Scholar
  13. Norouzi P, Larijani B, Ganjali M. Ochratoxin A sensor based on nanocomposite hybrid film of ionic liquid-graphene nano-sheets using coulometric FFT cyclic voltammetry. Int J Electrochem Sci. 2012;7:7313–24.Google Scholar
  14. Dreyer DR, Park S, Bielawski CW, Ruoff RS. The chemistry of graphene oxide. Chem Soc Rev. 2010;39(1):228–40.View ArticleGoogle Scholar
  15. Madadrang CJ, Kim HY, Gao G, Wang N, Zhu J, Feng H, et al. Adsorption behavior of EDTA-graphene oxide for Pb (II) removal. ACS Appl Mater Interfaces. 2012;4(3):1186–93.View ArticleGoogle Scholar
  16. Stankovich S, Dikin DA, Dommett GH, Kohlhaas KM, Zimney EJ, Stach EA, et al. Graphene-based composite materials. Nature. 2006;442(7100):282–6.View ArticleGoogle Scholar
  17. Chen F, Yan F, Chen Q, Wang Y, Han L, Chen Z, et al. Fabrication of Fe3O4@SiO2@TiO2 nanoparticles supported by graphene oxide sheets for the repeated adsorption and photocatalytic degradation of rhodamine B under UV irradiation. Dalton Trans. 2014;43(36):13537–44.View ArticleGoogle Scholar
  18. Cui L, Wang Y, Gao L, Hu L, Yan L, Wei Q, et al. EDTA functionalized magnetic graphene oxide for removal of Pb(II), Hg(II) and Cu(II) in water treatment: Adsorption mechanism and separation property. Chem Eng J. 2015;281:1–10.View ArticleGoogle Scholar
  19. Alvand M, Shemirani F. Preconcentration of trace cadmium ion using magnetic graphene nanoparticles as an efficient adsorbent. Microchim Acta. 2014;181(1–2):181–8.View ArticleGoogle Scholar
  20. Teo PS, Lim HN, Huang NM, Chia CH, Harrison I. Room temperature in situ chemical synthesis of Fe3O4/graphene. Ceram Int. 2012;38(8):6411–6.View ArticleGoogle Scholar
  21. Liu Q, Shi J, Cheng M, Li G, Cao D, Jiang G. Preparation of graphene-encapsulated magnetic microspheres for protein/peptide enrichment and MALDI-TOF MS analysis. Chem Commun. 2012;48(13):1874–6.View ArticleGoogle Scholar
  22. Li Y, Wang X-Y, Jiang X-P, Ye J-J, Zhang Y-W, Zhang X-Y. Fabrication of graphene oxide decorated with Fe3O4@SiO2 for immobilization of cellulase. J Nanoparticle Res. 2015;17(1):1–12.View ArticleGoogle Scholar
  23. Hummers Jr WS, Offeman RE. Preparation of graphitic oxide. J Am Chem Soc. 1958;80(6):1339–9.Google Scholar
  24. Marcano DC, Kosynkin DV, Berlin JM, Sinitskii A, Sun Z, Slesarev A, et al. Improved synthesis of graphene oxide. ACS Nano. 2010;4(8):4806–14.View ArticleGoogle Scholar
  25. Cote LJ, Kim F, Huang J. Langmuir − Blodgett assembly of graphite oxide single layers. J Am Chem Soc. 2008;131(3):1043–9.View ArticleGoogle Scholar
  26. Hamadi H, Gholami M, Khoobi M. Polyethyleneimine-modified super paramagnetic Fe3O4 nanoparticles: an efficient, reusable and water tolerance nanocatalyst. 2011.Google Scholar
  27. Tartaj P, Serna CJ. Synthesis of monodisperse superparamagnetic Fe/Silica nanospherical composites. J Am Chem Soc. 2003;125(51):15754–5.View ArticleGoogle Scholar
  28. He F, Fan J, Ma D, Zhang L, Leung C, Chan HL. The attachment of Fe3O4 nanoparticles to graphene oxide by covalent bonding. Carbon. 2010;48(11):3139–44.View ArticleGoogle Scholar
  29. Wang W, Ma R, Wu Q, Wang C, Wang Z. Magnetic microsphere-confined graphene for the extraction of polycyclic aromatic hydrocarbons from environmental water samples coupled with high performance liquid chromatography–fluorescence analysis. J Chromatogr A. 2013;1293:20–7.View ArticleGoogle Scholar
  30. Hossain MA, Ngo HH, Guo W. Introductory of Microsoft Excel SOLVER function-spreadsheet method for isotherm and kinetics modelling of metals biosorption in water and wastewater. J Water Sustain. 2013;3(4):223–37.Google Scholar
  31. Paredes JI, Villar-Rodil S, Solís-Fernández P, Martínez-Alonso A, Tascón JMD. Atomic force and scanning tunneling microscopy imaging of graphene nanosheets derived from graphite oxide. Langmuir. 2009;25(10):5957–68.View ArticleGoogle Scholar
  32. Maiti R, Manna S, Midya A, Ray SK. Broadband photoresponse and rectification of novel graphene oxide/n-Si heterojunctions. Opt Express. 2013;21(22):26034–43.View ArticleGoogle Scholar
  33. Yang D, Velamakanni A, Bozoklu G, Park S, Stoller M, Piner RD, et al. Chemical analysis of graphene oxide films after heat and chemical treatments by X-ray photoelectron and Micro-Raman spectroscopy. Carbon. 2009;47(1):145–52.View ArticleGoogle Scholar
  34. Li D, Muller MB, Gilje S, Kaner RB, Wallace GG. Processable aqueous dispersions of graphene nanosheets. Nat Nano. 2008;3(2):101–5.View ArticleGoogle Scholar
  35. Zhao G, Li J, Ren X, Chen C, Wang X. Few-layered graphene oxide nanosheets as superior sorbents for heavy metal ion pollution management. Environ Sci Technol. 2011;45(24):10454–62.View ArticleGoogle Scholar
  36. Titelman GI, Gelman V, Bron S, Khalfin RL, Cohen Y, Bianco-Peled H. Characteristics and microstructure of aqueous colloidal dispersions of graphite oxide. Carbon. 2005;43(3):641–9.View ArticleGoogle Scholar
  37. Stankovich S, Piner RD, Nguyen ST, Ruoff RS. Synthesis and exfoliation of isocyanate-treated graphene oxide nanoplatelets. Carbon. 2006;44(15):3342–7.View ArticleGoogle Scholar
  38. Abkenar SD, Khoobi M, Tarasi R, Hosseini M, Shafiee A, Ganjali MR. Fast removal of methylene blue from aqueous solution using magnetic-modified Fe3 O4 nanoparticles. J Environ Eng. 2014;141(1):04014049.View ArticleGoogle Scholar
  39. Cataldo F. Structural analogies and differences between graphite oxide and C60 and C70 polymeric oxides (Fullerene Ozopolymers). Fullerenes Nanotubes Carbon Nanostruct. 2003;11(1):1–13.View ArticleGoogle Scholar
  40. Sun X, Liu Z, Welsher K, Robinson JT, Goodwin A, Zaric S, et al. Nano-graphene oxide for cellular imaging and drug delivery. Nano Res. 2008;1(3):203–12.View ArticleGoogle Scholar
  41. Yang X, Wang Y, Huang X, Ma Y, Huang Y, Yang R, et al. Multi-functionalized graphene oxide based anticancer drug-carrier with dual-targeting function and pH-sensitivity. J Mater Chem. 2011;21(10):3448–54.View ArticleGoogle Scholar
  42. Xu Y, Liu Z, Zhang X, Wang Y, Tian J, Huang Y, et al. A graphene hybrid material covalently functionalized with porphyrin: synthesis and optical limiting property. Adv Mater. 2009;21(12):1275–9.View ArticleGoogle Scholar
  43. Lenth RV. Response-Surface Methods in R, using rsm. J Stat Soft. 2009;32(7):1–17.View ArticleGoogle Scholar
  44. Myers RH, Montgomery DC, Anderson-Cook CM. Response surface methodology: process and product optimization using designed experiments. vol. 705:2009. Hoboken, New Jersey: John Wiley & Sons.Google Scholar
  45. Zhu H, Fu Y, Jiang R, Yao J, Xiao L, Zeng G. Optimization of copper(II) adsorption onto novel magnetic calcium alginate/maghemite hydrogel beads using response surface methodology. Ind Eng Chem Res. 2014;53(10):4059–66.View ArticleGoogle Scholar
  46. Turkyilmaz H, Kartal T, Yildiz SY. Optimization of lead adsorption of mordenite by response surface methodology: characterization and modification. J Environ Health Sci Eng. 2014;12(1):5.View ArticleGoogle Scholar
  47. Amini M, Younesi H, Bahramifar N. Statistical modeling and optimization of the cadmium biosorption process in an aqueous solution using Aspergillus niger. Colloids Surf A Physicochem Eng Asp. 2009;337(1):67–73.View ArticleGoogle Scholar
  48. Luo S, Xu X, Zhou G, Liu C, Tang Y, Liu Y. Amino siloxane oligomer-linked graphene oxide as an efficient adsorbent for removal of Pb(II) from wastewater. J Hazard Mater. 2014;274:145–55.View ArticleGoogle Scholar
  49. Zhao G, Ren X, Gao X, Tan X, Li J, Chen C, et al. Removal of Pb (II) ions from aqueous solutions on few-layered graphene oxide nanosheets. Dalton Trans. 2011;40(41):10945–52.View ArticleGoogle Scholar
  50. Hadi Najafabadi H, Irani M, Roshanfekr Rad L, Heydari Haratameh A, Haririan I. Removal of Cu2+, Pb2+ and Cr6+ from aqueous solutions using a chitosan/graphene oxide composite nanofibrous adsorbent. RSC Adv. 2015;5(21):16532–9.View ArticleGoogle Scholar
  51. Wu W, Yang Y, Zhou H, Ye T, Huang Z, Liu R, et al. Highly efficient removal of Cu (II) from aqueous solution by using graphene oxide. Water Air Soil Pollut. 2013;224(1):1–8.View ArticleGoogle Scholar
  52. Repo E, Warchol JK, Kurniawan TA, Sillanpää MET. Adsorption of Co(II) and Ni(II) by EDTA- and/or DTPA-modified chitosan: kinetic and equilibrium modeling. Chem Eng J. 2010;161(1–2):73–82.View ArticleGoogle Scholar
  53. Guo L, Li G, Liu J, Yin P, Li Q. Adsorption of aniline on cross-linked starch sulfate from aqueous solution. Ind Eng Chem Res. 2009;48(23):10657–63.View ArticleGoogle Scholar
  54. Repo E, Warchoł JK, Bhatnagar A, Sillanpää M. Heavy metals adsorption by novel EDTA-modified chitosan–silica hybrid materials. J Colloid Interface Sci. 2011;358(1):261–7.View ArticleGoogle Scholar
  55. Repo E. EDTA-and DTPA-functionalized silica gel and chitosan adsorbents for the removal of heavy metals from aqueous solutions. 2011. Lappeenranta University of Technology Laboratory of Green Chemistry.Google Scholar
  56. Ma Y, La P, Lei W, Lu C, Du X. Adsorption of Hg(II) from aqueous solution using amino-functionalized graphite nanosheets decorated with Fe3O4 nanoparticles. Desalin Water Treat. 2015;57:1–9. doi:10.1080/19443994.2014.998292.Google Scholar
  57. Yang Y, Xie Y, Pang L, Li M, Song X, Wen J, et al. Preparation of reduced graphene oxide/poly (acrylamide) nanocomposite and its adsorption of Pb (II) and methylene blue. Langmuir. 2013;29(34):10727–36.View ArticleGoogle Scholar
  58. Kakavandi B, Kalantary RR, Jafari AJ, Nasseri S, Ameri A, Esrafili A, et al. Pb (II) adsorption onto a magnetic composite of activated carbon and superparamagnetic Fe3O4 nanoparticles: experimental and modeling study. Clean Soil Air Water. 2015;43:1157–66. doi:10.1002/clen.201400568.View ArticleGoogle Scholar

Copyright

© Khazaei et al. 2016

Advertisement