Preparation and characterization of an AC–Fe₃O₄–Au hybrid for the simultaneous removal of Cd²⁺, Pb²⁺, Cr³⁺ and Ni²⁺ ions from aqueous solution via complexation with 2-((2,4-dichloro-benzylidene)-amino)-benzenethiol: Taguchi optimization

Abstract

Activated carbon (AC) was magnetized with Fe₃O₄ nanoparticles (AC–Fe₃O₄-NPs) and loaded with Au nanoparticles (AC–Fe₃O₄–Au-NPs), and was fully characterized using different techniques such as XRD, XPS, VSM, TEM and SEM. 2-((2,4-Dichloro-benzylidene)-amino)-benzenethiol (DBABT), a complexing agent, was synthesized and characterized using ¹H-NMR, ES-MS and FT-IR analysis. Subsequently, AC–Fe₃O₄-NPs was modified with DBABT and applied for the ultrasound-assisted removal of Cd²⁺, Pb²⁺, Cr³⁺ and Ni²⁺ ions via complexation with DBABT. The influences of variables such as reaction time and adsorbent mass (equilibrium) were optimized simultaneously. The method under optimum conditions was set at pH 5, concentrations of 5, 15, 25 and 25 mg L⁻¹ for the Cd²⁺, Pb²⁺, Cr³⁺ and Ni²⁺ ions, respectively, 0.02 g for the adsorbent mass, 5 min sonication time and 6 mg L⁻¹ for the concentration of DBABT, producing removal percentages of 80.59, 93.85, 68.52 and 81.68 for Cd²⁺, Pb²⁺, Cr³⁺ and Ni²⁺ ions, respectively. Analysis of real experimental equilibrium data at various concentrations of analytes reveals the efficiency of the Langmuir model for good representation of experimental data, with maximum mono-layer adsorption capacities of 185.22, 135.14, 188.70 and 133.34 mg g⁻¹ for Cd²⁺, Pb²⁺, Cr³⁺ and Ni²⁺ ions respectively. The experimental data at various real times reveal that in most situations the systems reach equilibrium at contact times lower than 20 min, while the data fitted well to a combination of a pseudo second order kinetic model and intraparticle diffusion.

---

1. Introduction

Heavy metal ions present in higher content are toxic for the human body via damaging the nervous, reproductive and blood circulation systems.^1–3 Reactive centers such as COOH, amine groups and sulfur atoms in body cells are candidates for attacking metal ions (transition and/or heavy metal ions). Source for their arrival in the environment are sources like industrial, agricultural and domestic wastes.^4,5 In particular, Cd²⁺, Pb²⁺, Cr³⁺ and Ni²⁺ as more hazardous agents for humans namely as environmental contaminants which cannot be degraded or destroyed naturally.^6,7 Therefore, the detection and removal of these pollutants need preliminary novel and efficient methods designed to mitigate medical and environmental risks.⁸ Their efficient removal from various media to concentrations lower than their threshold limits is an urgent requirement. The threshold limits of the metal ions being studied here according to important criteria for evaluating drinking water quality such as the World Health Organization (WHO: 0.003 mg L⁻¹, 0.015 mg L⁻¹, 0.05 mg L⁻¹ and 0.03 mg L⁻¹ for Cd²⁺, Pb²⁺, Cr³⁺ and Ni²⁺ ions, respectively), US Environmental Protection Agency (EPA: 0.005 mg L⁻¹, 0.01 mg L⁻¹, 0.05 mg L⁻¹ and 0.06 mg L⁻¹ for Cd²⁺, Pb²⁺, Cr³⁺ and Ni²⁺ ions, respectively) and Iranian National Standards Organization (INSO: 0.005 mg L⁻¹, 0.05 mg L⁻¹, 0.05 mg L⁻¹ and 0.07 mg L⁻¹ for Cd²⁺, Pb²⁺, Cr³⁺ and Ni²⁺ ions, respectively) were identified. Various removal and clean up approaches like chemical precipitation, ion-exchange, membrane separation, reverse osmosis and adsorption^9,10 are good choices for this purpose that have attracted great attention in the purification and separation of metal ions in wastewater treatment.¹¹ Adsorption techniques due to their simplicity, rapidity and ecofriendly nature are highly recommended to achieve this.¹² Graphene and carbon based material, bio sorbents, silicates, zeolites, and chitosan have been extensively applied for metal ion sorption and/or preconcentration.^13,14 Activated carbon (AC) with (without) modification is a powerful sorbent used in clean up protocols and its ability for adsorption and removal of metal ions is enhanced following activation using a strong acid that is associated with formation of variations functional groups like NH₂, COOH, N, OH and C=O.^15,16 The type and number of reactive centers is strongly effected by the variation of the loading which helps the agent on the AC surface.¹⁷ In addition, some of the most influencial properties of each adsorbent in such procedures is the presence of various functional groups, high surface area and a porous structure of the adsorbent.¹⁸ Therefore, application of nanostructure based adsorbents leads to enhancement of the adsorption properties. The presence of metallic centers like soft Au-NP centers leads to greater binding of organic chelating agents on adsorbents.^19,20 One of the main problems is that separation based technique have problems associated with phase separation that can be simply overcome by usage of magnetic adsorbents based on Fe₃O₄.^21,22 In such systems circumrotation by simple exposure to a magnetic field achieves phase separation without tedious extra work. Composites of AC with other nanostructures as a sole adsorbent lack sufficient sites for efficient binding and accumulation of the metal ions being studied here. This drawback was simply overcome through modification of the adsorbent surface using a novel Schiff base chelating agent.^23,24 Magnetic nanoparticles have continued to draw considerable interest because of their great potential applications in magnetic fluids, catalysis, biotechnology/biomedicine, magnetic resonance imaging and magnetic recording devices in view of their paramagnetism, biocompatibility and safety.^25–27 Magnetic separation benefits from features like good recovery, high efficiency, low cost and easy phase separation allowing for convenient and economical magnetic separation that eliminates centrifuge and/or filtration steps.^28,29 Hybrid materials show the inherent properties of the adsorbents, in particular flexibility and strength, and also the high adsorption properties of the surface bonded nanoparticles.³⁰ In addition, gold nanoparticles due to soft–soft interaction with S–H groups can be modified to favor separation of metal ions from preconcentrated environmental samples.^31,32 Therefore, in this work, the hybrid AC–Fe₃O₄–Au-NPs was synthesized, characterized using SEM, XPS, VSM, TEM and XRD and used for simultaneous ultrasound assisted removal of Cd²⁺, Pb²⁺, Cr³⁺ and Ni²⁺ ions from aqueous media. Chelating agents loaded on AC–Fe₃O₄–Au-NPs through soft–soft interactions (π–Au metal), hydrogen bonding with AC functional groups or Fe₃O₄ centers and surface plasmons of Au-NP improved the removal efficiency. The effects of parameters were investigated with a Taguchi design and combined to search for and find optimum conditions that permit the maximum simultaneous removal process.

2. Experimental

2.1. Materials and apparatus

All chemicals were used as received without further purification from Merck, Dermasdat, Germany. Equipment was used according to the manufacturer’s recommendations corresponding to previous publications.^33–36 The concentrations of the metal (ions) being studied here were determined using a Varian model atomic absorption spectrophotometer.

2.2. Preparation of 2-((2,4-dichloro-benzylidene)-amino)-benzenethiol

A solution of 2-aminothiophenol (0.376 g, 3 mmol) in absolute methanol (10 mL) was added to a solution of 2,4-dichlorobenzaldehyde (0.525 g, 3 mmol) in absolute methanol (10 mL) and stirred for 4 h. The product was produced as a greenish white precipitate (chemical structure is shown in Scheme 1). The compound was filtered and washed with methanol twice to obtain it in 83% yield. IR (Fig. 1a) (KBr, cm⁻¹): 3405 (m), 3053 (w), 2922 (w), 2878 (w), 2142 (w), 1899 (w), 1682 (w), 1620 (s), 1591 (m), 1562 (s), 1488 (m), 1474 (w), 1458 (m), 1435 (m), 1404 (m), 1358 (m), 1315 (w), 1280 (m), 1250 (w), 1227 (w), 1187 (w), 1166 (m), 1089 (s), 1053 (w), 1034 (w), 1013 (m), 965 (m), 932 (w), 880 (m), 853 (s), 826 (m), 754 (s) 723 (m), 708 (w), 673 (m), 611 (m), 593 (m), 548 (m), 510 (w), 489 (m), 462 (m).

Scheme 1 Suggested mechanism for the interaction of M²⁺ and the DBABT ligand.

Fig. 1 FT-IR (a), ¹H-NMR (b) and ES-MS (c) spectra of the DBABT ligand.

¹H-NMR spectrum (Fig. 1b) (in DMSO-d6), δ (ppm): 8.74 (s, 1H), 8.15 (d, 1H, J = 7.6 Hz), 8.10 (d, 1H, J = 8.8 Hz), 7.02 (d, 1H, J = 8.4 Hz), 7.63 (d, 2H, J = 8.4 Hz), 7.57 (t, 1H, J = 7.6 Hz, J = 7.2 Hz), 7.48 (t, 1H, J = 7.6 Hz, J = 7.2 Hz). ES-MS (m/z) (Fig. 1c): 280 (M + 1), 252, 203, 195, 171, 122, 113, 97, 79 and 64.

The thiolic Schiff base ligand has two coordinator atoms (S (soft atom) and N (relatively hard atom)). These two donor atoms can easily bind to metal ions in 1 : 1 or 2 : 1 molar ratios as proposed in Scheme 1, the soft atom facilitates coordination to soft metal ions such as Cd²⁺ and Pd²⁺ while the hard donor atom of nitrogen intends to bind to the hard to moderate metal ions such as Ni²⁺ and Cr²⁺. Geometries of the formed complexes seemed to be tetrahedral or octahedral coordination compounds in which the coordinated ligands are the bidentate Schiff base and water molecules.

2.3. Preparation of AC–Fe₃O₄–Au-NPs

The preparation of a precursor solution for fabrication of the AC–Fe₃O₄–Au-NPs was as follows: 0.001 mol of FeCl₂ and 0.002 mol of FeCl₃ were mixed thoroughly with 200 mL of de-ionized water for 1 hour under sonication in a nitrogen atmosphere. In the next step, NH₃·H₂O was added to the above reaction mixture in an Erlenmeyer flask for 1 h at 50 °C in a nitrogen atmosphere. Subsequently, 1.0 g of AC was dispersed in 100 mL of absolute ethanol solution which was mixed thoroughly at room temperature for 18 h with 10 mL of the Fe₃O₄ nanoparticle suspension (22.0 mg mL⁻¹). Then, the reaction mixture was deposited using a magnet and washed several times with ethanol and distillated water and dried at room temperature. Finally, the Au nanoparticles were synthesized as follows: 100 mL of a 2.5 × 10⁻⁵ M HAuCl₄ aqueous solution, while stirring was heated to 90 °C. Then, 0.15 g of citric acid was added into the above solution and vigorously stirred for 40 min at 90 °C. Subsequently, the cloudy solution was added to 0.9 g of AC–Fe₃O₄-NPs and vigorously stirred for 20 h at room temperature. Finally, the reaction mixture was allowed to settle and rinsed several times with distillated water and dried in an oven at 50 °C.

2.4. Simultaneous adsorption experiments

Ultrasonic-assisted adsorption experiments were performed in batch mode to simultaneously increase the diffusion coefficients of ions and the mass transfer of aggregates to the surface and the vacant sites of AC–Fe₃O₄–Au-NPs as follows: according to experimental runs in the TD, different concentrations of metal ions and DBABT at pH 6 (optimum value) were added into 50 mL Erlenmeyer flasks containing specific amounts of AC–Fe₃O₄–Au-NPs which were dispersed thoroughly over 5 min in an ultrasonic bath at room temperature. Finally, AC–Fe₃O₄–Au-NPs was separated from the solution using an external magnetic field (1.3 T) in less than one minute and subsequently the total concentrations of metal ions in the effluent liquid phase were measured using a flame atomic absorption spectrometer. To evaluate the performance of the method, removal percentage and (R%_{metal ions}) were calculated.

2.5. Taguchi design

The Taguchi method can lower variations in a process through the design of experiments based on their simultaneous variation rather than changing one factor at a time. This design allows fast and accurate estimation of the individual main effects and their best combination. In this paper, Taguchi was applied to screen 8.0 important factors based on one 2-level factor and seven 3-level factor (1² × 7³) and the respective matrixes and responses are shown in Table 1. Data analysis was performed using statistical analysis of variance (ANOVA) and optimum conditions were found through the combination with a desirability function approach and validated using a confirmation test (n = 3 replicates at optimum). 3D response surfaces were plotted to visualize the variation of the response versus applied terms.

Table 1 Taguchi experimental design matrix^a

Run	Block	X1	X2	X3	X4	X5	X6	X7	X8	R%Cd(II)	R%Pb(II)	R%Cr(III)	R%Ni(II)
a X1, pH. X2, Cd(II) concentration (mg L^−1). X3, Pb(II) concentration (mg L^−1). X4, Cr(III) concentration (mg L^−1). X5, Ni(II) concentration (mg L^−1). X6, adsorbent mass (g). X7, sonication time (contact time) (min). X8, DBABT concentration (mg L^−1).
1	1	7	15	15	5	5	0.03	7	6	44.56 ± 1.3	74.89 ± 1.9	84.2 ± 0.8	38.73 ± 1.2
2	1	7	5	15	25	15	0.02	3	9	73.12 ± 1.0	84.00 ± 1.4	46.2 ± 1.3	45.61 ± 1.9
3	1	5	15	15	25	25	0.01	5	6	70.36 ± 1.8	43.29 ± 2.3	47.6 ± 1.7	18.15 ± 2.0
4	1	7	25	15	5	5	0.01	5	9	44.19 ± 1.1	46.86 ± 2.4	60.31 ± 2.1	88.19 ± 2.8
5	1	5	15	25	25	5	0.03	7	9	23.26 ± 2.1	78.36 ± 2.3	21.32 ± 1.9	83.56 ± 2.3
6	1	5	15	5	25	15	0.01	7	3	31.19 ± 1.3	35.96 ± 1.7	60.81 ± 2.0	25.30 ± 1.9
7	1	5	15	5	5	15	0.02	7	9	61.59 ± 1.2	34.86 ± 1.5	28.19 ± 1.6	54.39 ± 1.6
8	1	7	5	25	15	25	0.01	7	6	84.21 ± 2.0	11.76 ± 1.3	31.96 ± 1.5	35.79 ± 1.9
9	1	7	5	15	25	5	0.02	5	3	75.96 ± 2.0	83.71 ± 1.9	36.12 ± 2.5	98.60 ± 1.2
10	1	7	15	5	25	15	0.03	3	6	90.35 ± 1.3	37.43 ± 1.6	41.19 ± 2.2	21.89 ± 1.1
11	1	5	25	25	5	25	0.02	3	6	35.12 ± 1.1	50.86 ± 2.3	39.20 ± 0.9	32.96 ± 1.9
12	1	5	25	5	15	5	0.03	5	9	28.6 ± 1.2	24.82 ± 1.1	48.56 ± 2.1	41.60 ± 1.0
13	1	5	15	15	15	25	0.03	3	3	10.21 ± 1.6	86.97 ± 1.0	44.39 ± 1.9	17.06 ± 0.9
14	1	5	5	5	5	5	0.01	3	3	38.31 ± 1.5	77.65 ± 1.7	53.26 ± 1.4	41.10 ± 1.2
15	1	7	25	25	15	5	0.02	7	3	10.26 ± 1.3	16.03 ± 2.5	43.76 ± 1.1	13.39 ± 2.9
16	1	7	15	25	5	15	0.03	5	3	35.01 ± 1.4	50.74 ± 1.2	40.13 ± 1.8	29.30 ± 2.1
17	1	7	5	25	15	15	0.01	3	9	31.26 ± 1.8	55.43 ± 2.1	38.20 ± 1.3	37.16 ± 0.7
18	1	5	5	15	15	15	0.02	5	6	40.01 ± 1.9	68.96 ± 1.3	86.19 ± 2.2	27.19 ± 1.3

---

3. Results and discussion

3.1. Characterization of samples

FT-IR spectra of Fe₃O₄-NPs, AC, AC–Fe₃O₄-NPs and AC–Fe₃O₄–Au-NPs (Fig. 2) were recorded and it can be seen that the Fe₃O₄ spectrum features a broad peak at about 3300 cm⁻¹ which is attributed to the O–H stretching mode of water that was adsorbed from the air on the surface of Fe₃O₄, and the vibration peak of the Fe–O bond was observed at 590 cm⁻¹. The spectra of AC, AC–Fe₃O₄-NPs and AC–Fe₃O₄–Au-NPs exhibit a large broad band at 3000–3400 cm⁻¹ corresponding to O–H bond stretching of surface active carbon, the presence of a peak at around 1100 cm⁻¹ corresponds to the vibrational mode of H–O–H and the peak at around 1600 cm⁻¹ is due to H–O–H bending.³⁷ The peak at around 590 cm⁻¹ for AC–Fe₃O₄-NPs and AC–Fe₃O₄–Au-NPs indicates successful synthesis of the adsorbent in this study.

Fig. 2 FT-IR analysis of AC, Fe₃O₄-NPs, AC–Fe₃O₄-NPs and AC–Fe₃O₄–Au-NPs.

The XRD pattern for the Fe₃O₄ nanoparticles (Fig. 3a) consists of eight characteristic peaks at 2θ = 23.2, 27, 30.1, 35.4, 43.0, 44.5, 53.5 and 57.0, while that for AC–Fe₃O₄-NPs (Fig. 3b) exhibits a new broad peak at 25° (002) corresponding to the interlayer spacing of the AC. The peaks at 43° (100), 53° (004) and 78° (100) correspond to diffractions and reflections from the carbon atoms.³⁰ Modification of AC–Fe₃O₄-NPs with Au-NPs (Fig. 3c) causes the appearance of peaks at 38° (111) which belong to planes of the pure face-centered crystalline structure of Au. While peaks corresponding to Au at 44.4° (200), 64.6° (220) and 77.5° (311) overlap with the peaks of AC–Fe₃O₄-NPs that reflect successful modification and loading of Au-NPs onto AC–Fe₃O₄-NPs.

Fig. 3 XRD patterns of Fe₃O₄-NPs (a), AC–Fe₃O₄-NPs (b) and AC–Fe₃O₄–Au-NPs (c).

According to the saturation magnetization curve (Fig. 4a) it is obvious that there is no hysteresis in the magnetization sweep. Therefore, the magnetic coercivity of the nanostructures is low and suggests the presence of superparamagnetic Fe₃O₄.

Fig. 4 VSM (a) and XPS results (b) for the samples.

The XPS results in Fig. 4b show the main differences in the surface composition of the four different nanoparticles in the C 1s, O 1s, Fe 2p and Au 4f regions.

The SEM images of AC, Fe₃O₄, AC–Fe₃O₄ and AC–Fe₃O₄–Au (Fig. 5) show that AC (Fig. 5a) has a superficial area with a lightly ordered structure among the shape distribution. The morphology of Fe₃O₄ (Fig. 5b and c) shows a regular, ordered structure and a spherical shape that following combination with AC (Fig. 5d) leads to a change in the morphology of the AC which reflects successful modification. Finally, modification of the surface of AC–Fe₃O₄-NPs with Au-NPs (Fig .5e) leads to a significant change in morphology which is related to intermixing and immobilization of AU-NPs on AC–Fe₃O₄-NsP.

Fig. 5 SEM image of AC (a), TEM image of Fe₃O₄-NPs (b), SEM images of AC–Fe₃O₄-NPs (c) and AC–Fe₃O₄–Au-NPs (d).

3.2. Analysis of ANOVA under a Taguchi design

ANOVA was employed to test the statistical significant factors (see Table 2) based on their F and P values and comparison with respective threshold limits. Good model F values were obtained for the R% of the Cd²⁺, Pb²⁺, Cr³⁺ and Ni²⁺ ions (1.34 × 10⁵, 211.43, 3613.41 and 2610.81) while their P values were <0.0001, 0.0047, 0.0003 and 0.0004, respectively. All these data confirm the high accurately and efficiency of the constructed equation and model to correlate the R% values in this study to the 8.0 factors being investigated. However, the P values corresponding to terms like X₁, X₂, X₃, X₄, X₅, X_6, X₇ and X₈, for R% of all of the studied metal ions are less than 0.05 and justify their significant and serious contribution to the experimental response. Therefore, the final equations in terms of coded factors for R% of all the metal ions being studied are:

R%_Cd(II) = −45.98 + 8.35X₁ − 3.75X₂ + 17.19X₃ − 2.85X₄ − 1.77X₅ + 3.94X₆ + 0.42X₇ − 12.49X₈ (1)

R%_Pb(II) = 53.48 − 2.27X₁ + 19.69X₂ + 8.44X₃ + 2.5X₄ − 0.03X₅ − 8.32X₆ + 11.91X₇ + 5.03X₈ (2)

R%_Cr(III) = 47.31 − 0.41X₁ + 5.9X₂ − 7.43X₃ + 3.57X₄ + 6.62X₅ + 1.38X₆ − 3.57X₇ − 0.90X₈ (3)

R%_Ni(II) = 41.67 + 3.71X₁ + 12.73X₂ + 7.23X₃ + 5.78X₄ − 8.57X₅ − 0.72X₆ − 9.03X₇ − 4.21X₈ (4)

Table 2 ANOVA of Taguchi experimental results

Source	DF^a	Cd(II)				Pb(II)				Cr(III)				Ni(II)
		SS^b	MS^c	F-Value	p-Value	SS	MS	F-Value	p-Value	SS	MS	F-Value	p-Value	SS	MS	F-Value	p-Value
a Degree of freedom.b Sum of square.c Mean square.
Model	15	10 050.72	670.05	1.34 × 10^5	<0.0001	9903.43	660.23	211.43	0.0047	4922.67	328.18	3613.41	0.0003	10 456.96	697.13	2610.81	0.0004
X1	1	1254.5	1254.5	2.50 × 10^5	<0.0001	92.84	92.84	29.73	0.032	3.08	3.08	33.95	0.0282	252	252	943.77	0.0011
X2	2	875.29	437.64	87 237.91	<0.0001	4316.94	2158.47	691.22	0.0014	569.16	284.58	3133.35	0.0003	1501.53	750.76	2811.68	0.0004
X3	2	2781.07	1390.53	2.77 × 10^5	<0.0001	1523.73	761.87	243.98	0.0041	2419.47	1209.73	13 319.79	<0.0001	533.75	266.88	999.47	0.001
X4	2	2197.98	1098.99	2.19 × 10^5	<0.0001	869.37	434.69	139.2	0.0071	246.91	123.46	1359.31	0.0007	1519.15	759.57	2844.67	0.0004
X5	2	50.92	25.46	5074.64	0.0002	470.51	235.25	75.34	0.0131	708.62	354.31	3901.13	0.0003	2818.95	1409.48	5278.61	0.0002
X6	2	482.07	241.03	48 046.72	<0.0001	640.98	320.49	102.63	0.0096	17.13	8.56	94.29	0.0105	137.96	68.98	258.33	0.0039
X7	2	128.72	64.36	12 829.19	<0.0001	1646.09	823.05	263.57	0.0038	312.14	156.07	1718.4	0.0006	958.89	479.44	1795.56	0.0006
X8	2	2280.19	1140.09	2.27 × 10^5	<0.0001	342.96	171.48	54.91	0.0179	646.17	323.08	3557.32	0.0003	2734.73	1367.37	5120.91	0.0002
Residual	2	0.01	5.02 × 10^−03			6.25	3.12			0.18	0.091			0.53	0.27
Cor total	17	10 050.73				9909.67				4922.85				10 457.49

---

For the Cd²⁺, Pb²⁺, Cr³⁺ and Ni²⁺ ions, respectively, the values of the determination coefficient R² were 1.0, 0.9994, 1.0 and 0.9999, predicted R² were 0.9999, 0.9490, 0.9970 and 0.9959, and the adjusted R² were 1.0, 0.9946, 0.9997 and 0.9996. The Adeq. precision ratios of the models for the Cd²⁺, Pb²⁺, Cr³⁺ and Ni²⁺ ions, respectively, were 1200.376, 44.144, 228.310 and 175.235 indicating that all of the above values for predicted and adjusted R² and Adeq strongly prove the high applicability and efficiency of the models for the prediction of real sorption data and also suggest the sufficiency of the developed models being studied for predictions of the simultaneous removal of these metal ions.

3.3. Optimized conditions

After screening the effects of different factors, DF (from STATISTICAL 10.0 software) as one of the most effective statistical methods was applied to find the optimum conditions of the variables. The minimum, middle and maximum values of desirability were configured as 0.0, 0.5 and 1.0, respectively. A value closer to 1.0 strongly supports closeness of each point to the real optimum conditions. Optimum values were obtained for all of the investigated variables and the corresponding desirability and R% value for each metal ion are shown in Table 3.

Table 3 Optimum conditions

Variable	Optimum value
pH	6.0
Cd^2+ concentration (mg L^−1)	5.0
Pb^2+ concentration (mg L^−1)	15
Cr^3+ concentration (mg L^−1)	25
Ni^2+ concentration (mg L^−1)	25
Adsorbent mass (g)	0.02
Sonication time (min)	5.0
DBABT concentration (mg L^−1)	6.0

---

Desirability Value	0.85
R% of Cd^2+	80.6
R% of Pb^2+	93.9
R% of Cr^3+	68.6
R% of Ni^2+	82.0

---

3.4. 3D response surface plots

The 3D response surface plots were used to illustrate combined effects and identify the major interactions between variables and R% of the metal ions being studied. In this work, four typical 3D response surface plots are shown in Fig. 6.

Fig. 6 3D response surface plots.

The typical effect and interaction of sonication time and Cd²⁺ ion concentration on R%_Cd(II) shows that the maximum Cd²⁺ ion removal percentage is achieved in a short sonication time (contact time) that emerged from the unique role of ultrasound in accelerating mass transfer, good mixing and the contact of analytes and the adsorbent, while simultaneously enhancing the diffusion coefficient. In addition, the effect of the initial Cd²⁺ ion concentration on R%_Cd(II) clearly indicates that the high removal percentage at lower concentration strongly decreased at higher concentration which probably is due to saturation of the reactive center of the adsorbent or may be related to repulsive forces between bulk Cd²⁺ ions and those adsorbed on the adsorbent surface that was proven by the Langmuir isotherm.

Typically the effect and interaction of pH and the influence of the initial Pb²⁺ ion concentration on R%_Pb(II) (Fig. 6b) shows that the adsorbent and ligand has a constant contributory response without a significant correlation with pH. This fact may be related to the absence of ligand proton exchange or Pb²⁺ ion distribution at various pH and the absence of strong formation of Pb(OH)⁺ and/or Pb(OH)₂(s) over pH 5–7. The presence of amine and SH groups accelerate and facilate complex formation among the soft metal ions, while OH, COOH and NH groups and O atoms of the chelating agent and AC and/or Fe₃O₄ possibly bind and accumulate on the sorbent. At pH < 5.0 the removal percentage of the metal ions being studied is related to protonation of the chelating agent reactive sites such as the imine nitrogen or thiol groups. A decrease in the response at pH > 7.0 is attributed to the formation of metal hydroxide species such as soluble M(OH)⁺ and/or M(OH)_m^n± and insoluble precipitate of M(OH)₂. Therefore, the working pH was selected as 5–7. In addition, the effect of the initial Pb²⁺ ion concentration on R%_Pb(II) has a similar trend to the other metal ions.

Fig. 6d shows the effect of DBABT and Ni²⁺ concentration on R%_Ni(II) and is found to have a similar trend for all of the analytes. The investigation of the effect of initial DBABT concentration on R%_Pb(II) revealed that increasing DBABT can enhance the magnitude of complex formation among the metal ions being studied. It is known that the stoichiometry and concentration of the complex is strongly affected by ligand content. At lower DBABT content, probably due to insufficient reactive centers and also completion among AC functional groups for binding metal ions (in an irreversible pathway), leads to a reduction in the metal ion removal percentage. Increasing chelating agent content is associated with enhancement of complexation sites and leads to an enhancement in metal ion sorption/recovery. A higher content of chelating agent may possibly have changed the stoichiometry and also the polarity and lipophilicity of the complexes, with such deviation at high content being expected and may account for such a trend.

3.5. Isotherm study

The equilibrium adsorption isotherm was used to give useful information about the mechanism, properties and the tendency of the adsorbent for the metal ions being studied. The equilibrium data were fitted to isotherm equations such as Langmuir, Freundlich, Temkin, and Dubinin–Radushkevich (D–R). The constant parameters and correlation coefficients (R²) obtained from the plots of known equations for the above models are summarized in Table 4. Based on the linear form of the Langmuir isotherm model (according to Table 4), the values of K_a (the Langmuir adsorption constant (L mg⁻¹)) and Q_m (theoretical maximum mono-layer adsorption capacity (mg g⁻¹)) were obtained from the intercept and slope of the plot of C_e/q_e vs. C_e, respectively. The high correlation coefficient at the optimum amount of the adsorbent mass shows the applicability of the Langmuir model for interpretation of the experimental data over the whole concentration range. The parameters of the Freundlich isotherm model such as K_F (L mg⁻¹) and n (the capacity and intensity of the adsorption) were calculated from the intercept and slope of the linear plot of ln q_e versus ln C_e, respectively. The values of 1/n for the Freundlich isotherm were 0.73, 0.56, 0.71 and 0.56 for Cd²⁺, Pb²⁺, Cr³⁺ and Ni²⁺, respectively, and justify high metal ion adsorption onto AC–Fe₃O₄–Au-NPs, while lower R² values imply low suitability of this model for good prediction and representation of the experimental data.

Table 4 Calculated isotherm constant parameters and correlation coefficients

Isotherm	Parameters	Value
		Cd(II)	Pb(II)	Cr(III)	Ni(II)
		0.01 g	0.01 g	0.01 g	0.01 g
Langmuir	Qm (mg g^−1)	185.22	135.14	188.70	133.34
	Ka (L mg^−1)	0.1380	0.4999	0.2030	0.4687
	R^2	0.9964	0.9984	0.9981	0.9976
Freundlich	1/n	0.7377	0.5651	0.7134	0.5627
	KF (L mg^−1)	3.90	5.06	4.471	4.96
	R^2	0.9911	0.9745	0.9946	0.9865
Temkin	B1	34.53	29.59	36.01	28.94
	KT (L mg^−1)	1.78	5.00	2.57	4.80
	R^2	0.9851	0.9976	0.9794	0.9881
Dubinin–Radushkevich	Qs (mg g^−1)	77.39	85.21	83.76	81.39
	B × 10^−7	−4	−1	−2	−1
	E	1123.59	2272.72	1587.30	2272.72
	R^2	0.9172	0.9437	0.9156	0.9099

---

The values of the Temkin constants and the correlation coefficient are lower than the Langmuir values. Therefore, the Temkin isotherm presents a lower ability to fit the experimental equilibrium data than the Freundlich and Langmuir isotherms which have the best correlation and efficiency to predict the experimental data.

The D–R model was employed to estimate the porosity apparent free energy and the adsorption characteristics. The slope of the plot of ln q_e versus ε² gives K, and the intercept yields the Q_m value.

The value of the D–R correlation is lower than the other isotherm values, suggesting poor ability to present and describe the experimental equilibrium data.

3.6. Comparison with the literature

Comparison of the performance of AC–Fe₃O₄–Au-NPs in terms of adsorption capacity, pH value and adsorption time toward adsorbents for the removal of metal ions is presented in Table 5. As can be seen, the contact time for this method (5 min) is comparable and superior to all of the mentioned methods. This observation confirms the unique role of ultrasound for enhancement and rapid adsorption of the metal ions being studied. On the other hand, the adsorption capacity for this adsorbent (0.02 g) is preferable and superior to other adsorbents (Table 5). This observation confirms the unique role of metallic nanoparticles and also the high contribution of the DBABT ligand on the collectivity of metal ions on the adsorbent. It shows the high performance of the proposed method for the removal of the metal ions being studied here: Cd²⁺, Pb²⁺, Cr³⁺ and Ni²⁺ ions.^34–44 In other words, this table reveals that AC–Fe₃O₄–Au-NPs rather than AC and AC–Fe₃O₄-NPs showed better removal efficiency for the metal ions being studied here.

Table 5 Comparison of the proposed method with published methods

Sorbent	Ion	pH	Adsorption capacity (mg g^−1)	Detection method	Ref.
Innovative electronic	Cd(II)	4.0	236.4	ICP-AES	38
Polyaniline grafted chitosan	Cd(II)	6.0	12.87	—	39
α-Ketoglutaric acid-modified magnetic chitosan	Cd(II)	6.0	255.7	FAAS	40
BiOBr microspheres	Cd(II)	—	11.7	Spectrophotometry	41
Dead T. viride	Cd(II)	6.0	10.95	Spectrophotometry	42
Polyvinyl alcohol-chelating sponge	Cd(II)	5.5	17.83	FAAS	43
AC	Cd(II)	6.0	53.0	FAAS	This work
AC–Fe3O4-NPs	Cd(II)	6.0	110.11	FAAS	This work
AC–Fe3O4–Au-NPs	Cd(II)	6.0	185.22	FAAS	This work
Polyaniline grafted chitosan	Pb(II)	6.0	13.23	—	39
BiOBr microspheres	Pb(II)	—	6.5	Spectrophotometry	41
Agaricus bisporus	Pb(II)	5.5	86.4	FAAS	44
Mn3O4-coated activated carbon	Pb(II)	5.0	14.91	ICP-AES	45
Few-layered graphene oxide nanosheets	Pb(II)	6.5	1850	Spectrophotometry	46
DPDB immobilized onto mesoporous silica	Pb(II)	5.0	195.31	FAAS	47
AC	Pb(II)	6.0	17.8	FAAS	This work
AC–Fe3O4-NPs	Pb(II)	6.0	82.1	FAAS	This work
AC–Fe3O4–Au-NPs	Pb(II)	6.0	135.14	FAAS	This work
Mn3O4-coated activated carbon	Cr(III)	5.0	5.30	ICP-AES	45
BiOBr microspheres	Cr(III)	—	16.9	Spectrophotometry	41
Gelatin–montmorillonite	Cr(III)	5.5	52.91	FAAS	48
Garden grass	Cr(III)	4.0	19.4	FAAS	49
Polyacrylonitrile nanofibers	Cr(III)	2.0	137.6	ICP-OES
AC	Cr(III)	6.0	60.0	FAAS	This work
AC–Fe3O4-NPs	Cr(III)	6.0	123.5	FAAS	This work
AC–Fe3O4–Au-NPs	Cr(II)	6.0	188.70	FAAS	This work
Hemp-based material	Ni(II)	50.	160	FAAS	50
Polyvinyl alcohol-chelating sponge	Ni(II)	5.5	17.83	FAAS	43 and 51
Fe3O4 nanoparticles	Ni(II)	8.0	362.31	Spectrophotometry	52
Activated carbon from Choerospondias axillaris	Ni(II)	5.0	28.25	FAAS	53
Orange peel	Ni(II)	5.0	16.60	FAAS	54
Histidine modified chitosan	Ni(II)	5.0	55.6	FAAS	48
AC	Ni(II)	6.0	15.3	FAAS	This work
AC–Fe3O4-NPs	Ni(II)	6.0	79.7	FAAS	This work
AC–Fe3O4–Au-NPs	Ni(II)	6.0	133.34	FAAS	This work

---

3.7. Comparison with legal standard values and regeneration of the adsorbent

Comparison of the performance of AC–Fe₃O₄–Au-NPs for the removal of the metal ions studied here in terms of the average concentration after removal treatment with the WHO, USEPA and INSO legal concentrations is presented in Table 6. The results show that the average measured concentrations after removal treatment under optimal conditions are close to the standard levels and no significant difference is observed between them.

Table 6 Comparison of the average concentrations of the heavy metals ions in this study after removal under optimal conditions with their legal values

Metal ion	Legal concentration according to WHO (mg L^−1)	Legal concentration according to USÈPA (mg L^−1)	Legal concentration according to WHO (mg L^−1)	Average measured concentration (mg L^−1)
Cd^2+ ions	0.003	0.005	0.005	0.97
Pb^2+ ions	0.015	0.01	0.05	0.91
Cr^3+ ions	0.05	0.05	0.05	7.85
Ni^2+ ions	0.03	0.06	0.07	4.5

---

The AC–Fe₃O₄–Au-NPs was efficiently regenerated with 50 mL of 0.50 M HCl and subsequently neutralized by washing with double distillated water. Subsequently, it was successfully reused in the adsorption–desorption process 5.0 times with very little change in the adsorption efficiency, supporting its regeneration.

4. Conclusion

In this work, activated carbon (AC) was successfully magnetized with Fe₃O₄ nanoparticles, modified with Au nanoparticles and then fully characterized with different techniques such as XRD and SEM, TEM, XPS and VSM. 2-((2,4-Dichloro-benzylidene)-amino)-benzenethiol was successfully synthesized and characterized using different techniques such as ¹H-NMR, ES-MS and FT-IR analysis. Subsequently, AC–Fe₃O₄–Au-NPs was used for the ultrasound-assisted removal of Pb²⁺, Cr³⁺, Cd²⁺ and Ni²⁺ ions following complexation with DBABT. The Taguchi design was employed for selection of the statistically significant variables and in combination with DF was used to search for the optimum conditions and significant terms in the empirical equation, while the interactions and main effects of the investigated variables were estimated using ANOVA and 3D surface plots. Analysis of real experimental equilibrium data at various concentrations of the analytes reveal the efficiency of the Langmuir model for good representation of the experimental data with maximum mono-layer adsorption capacities of 185.22, 135.14, 188.70 and 133.34 mg g⁻¹ for the Cd²⁺, Pb²⁺, Cr³⁺ and Ni²⁺ ions, respectively. The adsorption capacities for AC and AC–Fe₃O₄-NPs were much lower than these values. On the other hand, the proposed method was revealed to have good potential for the efficient removal of these metal ions from aqueous solution compared to several other adsorbents and methods. In addition, comparison to the legal standard values showed that the average measured concentrations after removal treatment under the optimal conditions were close to the standard levels and no significant difference was observed between them.

Acknowledgements

The authors thank the Research Council of the Yasouj University and the Iran National Science Foundation for financially supporting this work.

References

1. M. Jamshidi, M. Ghaedi, K. Dashtian, S. Hajati and A. Bazrafshan, RSC Adv., 2015, 5, 59522–59532.
2. F. N. Azad, M. Ghaedi, K. Dashtian, M. Montazerozohori, S. Hajati and E. Alipanahpour, RSC Adv., 2015, 5, 61060–61069.
3. H. Mazaheri, M. Ghaedi, S. Hajati, K. Dashtian and M. Purkait, RSC Adv., 2015, 5, 83427–83435.
4. F. Zare, M. Ghaedi, A. Daneshfar, S. Agarwal, I. Tyagi, T. A. Saleh and V. K. Gupta, Chem. Eng. J., 2015, 273, 296–306.
5. M. Ghaedi, H. Mazaheri, S. Khodadoust, S. Hajati and M. Purkait, Spectrochim. Acta, Part A, 2015, 135, 479–490.
6. B. Barfi, M. Rajabi, M. M. Zadeh, M. Ghaedi, M. Salavati-Niasari and R. Sahraei, Microchim. Acta, 2015, 182, 1187–1196.
7. M. Ghaedi, M. Montazerozohori, M. N. Biyareh, K. Mortazavi and M. Soylak, Int. J. Environ. Anal. Chem., 2013, 93, 528–542.
8. S. Hajati, M. Ghaedi and S. Yaghoubi, J. Ind. Eng. Chem., 2015, 21, 760–767.
9. M. Hébrant, M. Rose-Hélène and A. Walcarius, Colloids Surf., A, 2013, 417, 65–72.
10. F. Ge, M.-M. Li, H. Ye and B.-X. Zhao, J. Hazard. Mater., 2012, 211, 366–372.
11. E. Da’na and A. Sayari, Desalination, 2012, 285, 62–67.
12. R. Sitko, E. Turek, B. Zawisza, E. Malicka, E. Talik, J. Heimann, A. Gagor, B. Feist and R. Wrzalik, Dalton Trans., 2013, 42, 5682–5689.
13. J. Aguado, J. M. Arsuaga, A. Arencibia, M. Lindo and V. Gascón, J. Hazard. Mater., 2009, 163, 213–221.
14. F. Fu and Q. Wang, J. Environ. Manage., 2011, 92, 407–418.
15. M. Ghaedi, M. reza Rahimi, A. Ghaedi, I. Tyagi, S. Agarwal and V. K. Gupta, J. Colloid Interface Sci., 2016, 461, 425–434.
16. Y. Li, Q. Du, T. Liu, X. Peng, J. Wang, J. Sun, Y. Wang, S. Wu, Z. Wang and Y. Xia, Chem. Eng. Res. Des., 2013, 91, 361–368.
17. K. A. Krishnan, K. Sreejalekshmi, V. Vimexen and V. V. Dev, Ecotoxicol. Environ. Saf., 2016, 124, 418–425.
18. G. Sethia and A. Sayari, Carbon, 2016, 99, 289–294.
19. M. Roosta, M. Ghaedi, N. Shokri, A. Daneshfar, R. Sahraei and A. Asghari, Spectrochim. Acta, Part A, 2014, 118, 55–65.
20. G. Karimipour, M. Ghaedi, R. Sahraei, A. Daneshfar and M. N. Biyareh, Biol. Trace Elem. Res., 2012, 145, 109–117.
21. Y. Wang, P. He, X. Zhao, W. Lei and F. Dong, J. Solid State Electrochem., 2014, 18, 665–672.
22. Q. Cai, F. Hu, S.-T. Lee, F. Liao, Y. Li and M. Shao, Appl. Phys. Lett., 2015, 106, 023107.
23. G. Zheng, L. Polavarapu, L. M. Liz-Marzán, I. Pastoriza-Santos and J. Pérez-Juste, Chem. Commun., 2015, 51, 4572–4575.
24. J. Liu, M. Ye, Z. Wei, F. Hu, J. Wang, G. Ge, Z. Hu, M. Shao and S.-T. Lee, Nanoscale, 2015, 7, 13427–13437.
25. F. X. Ma, H. Hu, H. B. Wu, C. Y. Xu, Z. Xu and L. Zhen, Adv. Mater., 2015, 27, 4097–4101.
26. L. Xu and J. Wang, Environ. Sci. Technol., 2012, 46, 10145–10153.
27. T. Wang, Z. Liu, M. Lu, B. Wen, Q. Ouyang, Y. Chen, C. Zhu, P. Gao, C. Li and M. Cao, J. Appl. Phys., 2013, 113, 024314.
28. S. Patra, E. Roy, R. Madhuri and P. K. Sharma, Environ. Sci. Technol., 2015, 49, 6117–6126.
29. C. Karadaş and D. Kara, Water, Air, Soil Pollut., 2014, 225, 1–10.
30. S. Kumar, B. Kumar, A. Baruah and V. Shanker, J. Phys. Chem. C, 2013, 117, 26135–26143.
31. F. N. Azad, M. Ghaedi, K. Dashtian, S. Hajati and V. Pezeshkpour, Ultrason. Sonochem., 2016, 31, 383–393.
32. Y. Cui, D. Hu, Y. Fang and J. Ma, Sci. China, Ser. B: Chem., 2001, 44, 404–410.
33. C. Namasivayam and D. Kavitha, Dyes Pigm., 2002, 54, 47–58.
34. V. Gupta, B. Gupta, A. Rastogi, S. Agarwal and A. Nayak, Water Res., 2011, 45, 4047–4055.
35. M. Gonçalves, M. C. Guerreiro, L. C. A. de Oliveira and C. S. de Castro, J. Environ. Manage., 2013, 127, 206–211.
36. M. Ghaedi, H. A. Larki, S. N. Kokhdan, F. Marahel, R. Sahraei, A. Daneshfar and M. Purkait, Environ. Prog. Sustainable Energy, 2013, 32, 535–542.
37. A. Goudarzi, A. D. Namghi and C.-S. Ha, RSC Adv., 2014, 4, 59764–59771.
38. M. Xu, P. Hadi, G. Chen and G. McKay, J. Hazard. Mater., 2014, 273, 118–123.
39. R. Karthik and S. Meenakshi, Chem. Eng. J., 2015, 263, 168–177.
40. G. Yang, L. Tang, X. Lei, G. Zeng, Y. Cai, X. Wei, Y. Zhou, S. Li, Y. Fang and Y. Zhang, Appl. Surf. Sci., 2014, 292, 710–716.
41. X. Wang, W. Liu, J. Tian, Z. Zhao, P. Hao, X. Kang, Y. Sang and H. Liu, J. Mater. Chem. A, 2014, 2, 2599–2608.
42. R. M. Hlihor, M. Diaconu, F. Leon, S. Curteanu, T. Tavares and M. Gavrilescu, New Biotechnol., 2015, 32, 358–368.
43. C. Cheng, J. Wang, X. Yang, A. Li and C. Philippe, J. Hazard. Mater., 2014, 264, 332–341.
44. Y. Long, D. Lei, J. Ni, Z. Ren, C. Chen and H. Xu, Bioresour. Technol., 2014, 152, 457–463.
45. M.-E. Lee, J. H. Park, J. W. Chung, C.-Y. Lee and S. Kang, J. Ind. Eng. Chem., 2015, 21, 470–475.
46. G. Zhao, X. Ren, X. Gao, X. Tan, J. Li, C. Chen, Y. Huang and X. Wang, Dalton Trans., 2011, 40, 10945–10952.
47. M. R. Awual and M. M. Hasan, Microporous Mesoporous Mater., 2014, 196, 261–269.
48. A. Eser, V. N. Tirtom, T. Aydemir, S. Becerik and A. Dinçer, Chem. Eng. J., 2012, 210, 590–596.
49. A. H. Sulaymon, A. A. Mohammed and T. J. Al-Musawi, Int. J. Chem. React. Eng., 2014, 12, 477–486.
50. G. Z. Kyzas, Z. Terzopoulou, V. Nikolaidis, E. Alexopoulou and D. N. Bikiaris, J. Mol. Liq., 2015, 209, 209–218.
51. G. Z. Kyzas, P. I. Siafaka, E. G. Pavlidou, K. J. Chrissafis and D. N. Bikiaris, Chem. Eng. J., 2015, 259, 438–448.
52. R. K. Gautam, P. K. Gautam, S. Banerjee, S. Soni, S. K. Singh and M. C. Chattopadhyaya, J. Mol. Liq., 2015, 204, 60–69.
53. R. M. Shrestha, M. Varga, I. Varga, A. P. Yadav, B. P. Pokharel and R. R. Pradhananga, J. Inst. Eng., 2014, 9, 166–174.
54. F. Gönen and D. S. Serin, Afr. J. Biotechnol., 2014, 11, 1250–1258.

---