Introduction

Heavy metals have been detected in many industrial effluents such as fertilizers, petroleum refining, paper-making, mining, dyeing, and plastic products (Giwa et al. 2013). Due to the toxicity, accumulation in the environment, and non-degradability of heavy metals, they are considered as prior environmental pollutants and hazards to human as well (Deshmukh et al. 2017; Padilla-Ortega et al. 2014). Cadmium is a highly toxic metal element that causes health problems in kidney and nervous system (Zhu et al. 2007). Inhalation and injection are two possible pathways of human exposure with Cd, while 10–50% of this metal is adsorbed via breathing. The toxicity mechanism of cadmium occurs through oxidative stress and DNA changes, affecting Zn or Mg physiological function, inhibiting heme synthesis, and mitochondrial malfunction. Besides, protein and amino acid deficiency and apoptosis are of other clinical toxicology aspects of cadmium (Bernhoft 2013; Satarug et al. 2017) Therefore, it is important to eliminate this metal from effluents before discharging them into water bodies. A number of methods such as chemical decomposition, soluble condensation, ultrafiltration, electrolyte, and ion exchange have been applied to remove heavy metals from wastewater (Ruyter-Hooley et al. 2017; Tang et al. 2017; Zheng et al. 2008). Each method has its own economic limitations or low efficiency in very low levels of pollutants. Extensive studies up to now have been carried out on the application of natural and synthetic adsorbents due to their adsorption properties and ionic exchange characteristic (Argun 2008). In recent years, many efforts have been focused on synthesis or application of low-cost materials for heavy metals’ removal from industrial wastewater (Jorfi et al. 2017). Among minerals, zeolites have been widely used for eliminating the environmental contaminants. These materials have a three-dimensional aluminous silicate structure (Kocasoy and Şahin 2007; Shaheen et al. 2012). Selectivity is an advantage of zeolites which can be employed to remove harmful ions from effluents. However, one of the greatest problems of synthetic resins is their constraint in low-volume application. Also, the difficulty of storage and disposal of saturated resins, low or non-consistency against radiation, heat, and chemical oxidizing agents along with their high price could be raised as their other drawbacks (Kocasoy and Şahin 2007). Therefore, natural zeolite with their abundance and cost-effectiveness has become more predominant. Among the zeolites, three types including clinoptilolite, bentonite, and chabazite are more widely used than others. Clinoptilolite is one of the well-known types of zeolite with a microporous structure comprising from SiO4 and AlO4 banded to oxygen atoms. The high selective ion exchange capacity along with non-toxicity of this compound has led to its wide industrial application for eliminating toxic materials from water and wastewater (Murkani et al. 2015; Erdem et al. 2004). Also, the use of heterogeneous catalysts has been increased because of functioning in a wide range of pH and high stability. Fe3O4 is one of the most catalysts used to remove heavy metals, and its magnetic nanoparticles have high physical and chemical sustainability. In addition, when the nanoparticles are merged with polymeric materials due to increased specific surface area, adsorption capacity increases (Mallakpour and Behranvand 2016; Pasandideh et al. 2016). Recently, magnetic nanoparticles have been coated with organic and inorganic materials (Nalbandian et al. 2016). Another capability of magnetic iron nanoparticles on a porous substrate is the possibility of easy separation using a magnetic field so that no nanoparticle remains in the refined solution. Regarding the accessibility to abundant zeolite resources in Iran and the exploitation of these resources in areas such as Semnan, they could be used for decontamination of environmental pollutants in Middle East countries (Kocasoy and Şahin 2007; Inglezakis et al. 2002). Therefore, this study aimed to investigate the applicability of Fe3O4@Z nanocomposite in cadmium removal from aquatic environments.

This pilot scale study was performed in 2016 which was conducted from March to September. The project took place in Ahvaz Jundishapur University of Medical Sciences, Iran.

Materials and methods

Materials

Clinoptilolite zeolite (AlCaH6KNaO3Si) was purchased from Afrazand Mining Corporation. The chemicals including ferric chloride (FeCl3), ferrous chloride (FeCl2), and NaOH with 98% purity were purchased from Merck Company and were used without further pre-treatment.

Synthesis of Fe3O4/zeolite nanocomposite (Fe3O4@Z)

First, zeolite was washed for three times with deionized water and placed in oven at 100 °C for 24 h to remove the moisture. After crushing the particles, they were screened using standard ASTM sieve with a mesh size of 20–40. Then, 4 gr screened zeolite was added to 40 mL of distilled water and stirred for 10 min until a uniform suspension was obtained. In the next step, appropriate amount of FeCl3 and FeCl2 with a molar ratio of 2:1 was added to suspension and the mixture was stirred for 20 min under N2 gas. (Usually, to prevent nanoparticle agglomeration before coating on target materials, different methods are used such as surfactant and polymers (Alqadami et al. 2017). However, sonication was used in this study.) Then, 20 mL of NaOH solution (2.5 M) was added to the solution, while the process of stirring was extended to 30 min. The synthesized product was filtered followed by washing with distilled water and finally dried in an oven at 60 °C for 24 h (Jahangirian et al. 2013). Fe3O4@Z was generated according to Eq. 1.

$$ {\text{Z}}({\text{gr}}) + {\text{Fe}}^{2+} + 2{\text{Fe}}^{3+} + 8{\text{OH}}^{-} \mathop{\longrightarrow}^{{\text{N}}_{2}}{\text{Fe}}_{3} {\text{O}}_{4} @{\text{Z}} + 4{\text{H}}_{2} {\text{O}} $$
(1)

Characterization

To determine the crystalline structure of zeolite and nanocomposite, an X-ray diffraction (XRD) instrument (model: D6792 PHILIPS) was used. Also, the samples’ surface morphology and their chemical composition before and after sorption were determined using field emission scanning electron microscopy and energy-dispersive X-ray analysis (FESEM-EDS) (model: Hitachi S4160). The magnetization of nanocomposite was identified using a vibrating sample magnetometer (VSM model MDKB). To determine the different chemical bonds in nanocomposite before and after Cd(II) adsorption, Fourier transfer infrared (FTIR) spectra analysis was done in vibrational frequency ranging from 500 to 4000 cm−1 using a JASCO 460 plus FTIR spectrometer. The pore size distribution, specific surface area, and the pore volume of the nanocomposite were determined using the Brunauer–Emmett–Teller (BET) micrometrics (BELSORP MINI II, Japan).

Experimental procedure

This research was a fundamental-applied study conducted on a laboratory scale. All experiments were performed in a batch system in 200-mL laboratory jars. At the beginning, 500 mg/L of Cd stock solution was prepared by dissolving cadmium chloride (Cd (Cl)2·5H2O) in distilled water. The other solutions were daily prepared by diluting the stock solution. In order to mix the solutions, they were shaken at the speed equal to 150 rpm. At the end of equilibrium time, in order to separate the composite from solution, a magnet was used. The samples were centrifuged at 4000 rpm for 5 min and the pH of the supernatant liquid was lowered to 2 by adding nitric acid. All samples were stored in refrigerator before analysis. Finally, the residual Cd concentrations in the samples were measured by atomic absorption spectroscopy (AAS Analyst 700, America). Each experiment was performed twice, and the average values were reported as the final results. In this study, Cd adsorption was investigated by considering different parameters such as contact time (0–100 min), solution pH (2–9), adsorbent dosage (0.2–1 g/L), initial cadmium concentration (2–30 mg/L), and temperature (20–40 °C). 1 N solution of HCl and NaOH was used to adjust pH. Data analysis was performed using Excel software. The removal efficiency and adsorption capacity of Cd were calculated using Eqs. (2, 3), as follows:

$$ Re\,(\% ) = \frac{{(C_{0} - C_{t} )}}{{C_{0} }} \times 100 $$
(2)
$$ q_{\text{e}} = \frac{{(C_{0} - C_{t} )V}}{M} $$
(3)

where C0 and Ct are the initial and final concentrations of cadmium in solution (mg/L), respectively. qe is the amount of the adsorbed cadmium (mg/g), M is the dosage of adsorbent (g), and V is the solution volume (L).

Evaluation of the isoelectric point

After preparation of adsorbent, isoelectric point was determined. In this regard, a salt solution (0.01 M) was used as an electrolyte and 0.1 M HCl and NaOH solutions for pH adjustment. Then, 180 mL of 0.01 M salt solution was added to six laboratory jars (200 ml) and pH was adjusted in the range of 2–12. Afterward, 1 g/L adsorbent was added to all jars and they were put on a shaker at 150 rpm for 24 h. Then, solutions were placed on a magnet to precipitate the adsorbent. The final pH of each sample was measured using a digital pH meter. To determine pHzpc, the initial and final pH values were plotted on the graph and the cross-point of these two pH values was determined as pHzpc.

Optimization of adsorption parameters

In the present study, in order to determine equilibrium time 0.6 g/L adsorbent dose was added to initial Cd concentrations of 20 mg/L in pH = 4. In addition, to determine the optimized pH, the experiments were performed in pH range between 2 and 9, initial cadmium concentration of 20 mg/L, and an adsorbent dose of 0.6 g. By using the optimum pH, other parameters were optimized, afterward. In the next step, the adsorption isotherm and kinetics models were studied. Finally, the optimum temperature and thermodynamic parameters were determined by running adsorption process at different temperatures, while the incubator shaker was used to adjust the temperature.

Examination of isotherm and adsorption kinetics

In this study, three isotherm models including Freundlich, Langmuir, and Dubinin–Radushkevich (D–R) were used to describe the cadmium adsorption process on Fe3O4@Z nanocomposite. The equations and linear form of these isotherms are shown in Table 1. In order to set the isothermal models and evaluate the cadmium adsorption capacity, the experiments were performed in 200-mL laboratory jars by adding 1 g/L of adsorbent at pH = 6 and various cadmium concentrations. The jars were placed in a shaker incubator at 150 rpm, and the temperature was adjusted to 25 °C. In addition, first-order and second-order kinetic models were studied.

Table 1 Adsorption isotherm and kinetic models used in the study (Pourfadakari et al. 2017; Pourfadakari and Mahvi 2014)
Full size table

Thermodynamics studies

Thermodynamic parameters were used to evaluate the effect of temperature, physiochemical properties of the adsorption process, and also provide information on energy changes associated with adsorption. In this study, the variations of Gibbs free energy, enthalpy, and entropy were studied to predict the adsorption process (Baghapour et al. 2013).

$$ \Delta G^{ \circ } = - RT\ln (K_{\text{L}} ) $$
(4)
$$ \Delta G^{ \circ } = \Delta H^{ \circ } - T\Delta S^{ \circ } $$
(5)
$$ \ln (K_{\text{L}} ) = (\Delta S^{ \circ } /R) - (\Delta H^{ \circ } /RT) $$
(6)

where T is the temperature (K), R is the gas constant (8.314 J / mol K), and KL is the thermodynamic equilibrium constant (1/mol). The values of ΔH° and ΔS° can be calculated from the intercept and slope of van’t Hoff plots of ln KL versus 1/T, respectively (Rezaei Kalantry et al. 2016).

Modeling the reactor’s operational conditions

In order to better portray the performance of the reactor in Cd adsorption, the removal efficiency trends by changes in operational parameters were modeled. pH, adsorbent dose, and Cd concentration contributed as the independent variables and removal efficiency as the dependent variable in the modeling process. To this end, specific software, namely CurveExpert ver. 1.4, was employed. This software contains over 260 model functions and regression as well. The modeling results were reported as their model name, mathematical function, graphs, and fitness as R2.

Results and discussion

Characterization of zeolite and nanocomposite before and after adsorption process

Figuring out the adsorbent structure is one of the most important issues, which should be considered in adsorption studies. The chemical analysis of the crystalline structure of zeolite and Fe3O4@Z nanocomposite was determined by XRD diffraction, shown in Fig. 1a, b. Accordingly, in zeolite sample, the main peaks were observed at 2θ: 9.94, 11.27, 13.14, 17.44, 23.57, 27.32, 30.10, and 32.10 (JCPDS 01-079-1461). Also, in Fe3O4@Z nanocomposite sample, the main peaks were observed at 2θ:30.25, 43.42, 57.42, and 63.02. It is noticeable that sharp peak was identified in 2θ: 35.74 which corresponds to the standard card number (75-0033). The morphology and distribution particle size of zeolite and nanocomposite by FESEM-EDS analysis are shown in Fig. 1c–e. According to Fig. 1c, zeolite aggregated as large particles, which created a large external surface area with uniform distribution. On the other hand, the sample contained different elements such as oxygen, silica, aluminum, potassium, sodium, and calcium. In addition, in Fig. 1d, the uniform distribution of nanocomposite particle with small size in range 16.63–29.03 nm on zeolite before Cd(II) sorption was identified. The quantitative analysis indicated the presence of various elements such as oxygen, silica, potassium, calcium, and iron in the nanocomposite. The FESEM-EDS analysis in Fig. 1e shows the presence of elements such as oxygen, potassium, calcium, iron, and cadmium. The particle size in the loading nanocomposite with Cd(II) was between 22.53 and 72.35 nm. New peak for Cd indicates the uniform presence of Cd(II) ions distributed on the surface of nanocomposite after adsorption. This change is occurred due to ion exchange; in other words, after Cd(II) sorption, some elements replace with Cd ions. Magnetic characterization of the nanocomposite before and after Cd(II) adsorption is shown in Fig. 2a, b. As is clear, no difference was found in nanocomposite magnetic value and the saturation magnetization (MS) value for each two curve was to be 15.856 emu/g. Functional groups on nanocomposite surface before and after adsorption Cd(II) are shown in Fig. 2c. According to FTIR analysis, it was identified that all bands in the nanocomposite sample after adsorption are the same before of adsorption with low difference observed in the peaks. The large peaks at 3434 and 3429 cm−1 (before and after adsorption) are related to (–OH) hydroxyl groups. The maximum peak at 1063 and 1051 cm−1 (before and after adsorption) is associated with the stretching vibrations of C–O. The small peaks located at 2923.45, 2855.05, and 2924 cm−1 (before and after adsorption) related to (C–H) alkyl group and (C=O) carbonyl functional group was observed in the peaks 1632 and 1627 cm−1 (before and after adsorption). In addition, the adsorption band at the range 460–632 cm−1 corresponds to the existence of Fe–O groups. The functional groups decreased after adsorption probably referring to the fact that metal binding occurs through ion exchange (Poinern et al. 2016). The nanocomposite properties before and after sorption of Cd(II) ions were determined by BET analysis using N2 adsorption method which is reported in Table 2. It was identified that the average pore diameter, special surface area and pore volume of each of the two sample (before and after sorption) were 10.58 and 10.59 nm, 71.95 and 69.48 m2/g, and 16.53 and 15.98 cm3/g, respectively. The average pore size of adsorbent was found in range of 2–50 nm. Also Fig. 2d shows a type IV isotherm for as-prepared nanocomposite demonstrating the existence of mesoporous structure, based on the IUPAC category (Wang et al. 2009).

Fig. 1
figure 1

XRD patterns of a natural zeolite and b Fe3O4@Z nanocomposite, FESEM-EDX images of c natural zeolite, d Fe3O4@Z nanocomposite before and e after Cd(II) sorption

Full size image
Fig. 2
figure 2

VSM analysis a before sorption, b after sorption, c FTIR spectra, d adsorption/desorption isotherm of N2 on Fe3O4@Z nanocomposite

Full size image
Table 2 Physico-chemical characterization of Fe3O4@Z
Full size table

Effect of various parameters on the adsorption process

Effects of pH

Solution pH is a crucial parameter affecting the adsorption of contaminants onto the adsorbent surface. pHzpc is a critical feature of the adsorbent that should be carefully looked at. pHZPC is equal to the point at which the initial pH is equal to terminal pH which, in this study, was 5.7± 0.1. At this pHZPC, virtually no ion exchange occurs which means that the charge on the synthesized nanocomposite surface is zero. In pH values less than pHZPC, the amount of H+ ions on the adsorbent surface goes up and the negative charge (hydroxyl anions) decrease, which causes a competition between the H+ and Cd2+ ions for adsorption onto the adsorbent surfaces and consequently reduction in Cd adsorption. At pH values above this point, the adsorbent surface has a negative charge due to the existence of hydroxyl ions which leads to conversion of Cd2+ ions to Cd(OH)2 and reduction in the adsorption efficiency (Krishnan and Anirudhan 2003; Raji and Anirudha 1997). According to Fig. 3a, it was identified that in pH = 2 the adsorption efficiency reached to 17.6%. Furthermore, by increasing solution pH to 6, the positive electrical charge on the adsorbent surface was reduced which caused enhancing the efficiency of the cadmium adsorption capacity. Therefore, at pH = 6, contact time 60 min, adsorbent dose of 0.6 g/L, the removal efficiency for 20 mg/L of cadmium solution reached to 42.9%. Then, at pH values 7 and 9, the adsorption efficiency decreased to 23.4% and 14.36%, respectively. Therefore, pH = 6 was selected as optimal value. Results of Gupta et al. (2003) in removing cadmium and nickel from wastewater using bagasse ashes showed that the maximum adsorption of these materials was found at pH = 6 and 6.5, respectively (Gupta et al. 2003).

Fig. 3
figure 3

Effect of operational parameters on removal efficiency: a pH (nanocomposite dosage: 0.6 g/L, initial Cd concentration: 20 mg/L), b different dosages of nanocomposite (pH = 6 and initial Cd concentration: 20 mg /L), c initial Cd concentration (pH = 6, nanocomposite dosage: 1 g/L)

Full size image

Effect of nanocomposite dosage

The effect of various doses of nanocomposite (0.2–1 g/L) on Cd adsorptive removal was investigated. Based on Fig. 3b, in 20 mg/L of Cd solution the increase in adsorbent dosage from 0.2 to 1 g/L causes the cadmium removal level up from 24.4 to 68.9%. This is while the adsorption capacity decreased from 24.4 to 13.78 mg/g. The increase in cadmium removal by the adsorbent dosage would be a result of extended unsaturated active sites on the adsorbent surface that are occupied by cadmium ions (Tang et al. 2017). These results are consistent with the study of Pérez-Marín et al. (2007).

Effect of initial Cd concentration

The effect of initial Cd concentrations ranging from 2 to 30 mg /L at pH = 6 and the adsorbent dosage of 1 g/L on Cd adsorptive removal were investigated. As shown in Fig. 3c, with the increase in initial Cd concentration from 2 to 30 mg/L, the removal efficiency dropped from 94.4 to 58.2%, while the adsorption capacity increased from 1.88 to 17.46 mg/g. The reason is the saturation and consequently reduction in active sites on the adsorbent surface. In other words, at higher concentrations the availability of fewer adsorption sites and the amount of cadmium removal depend on its initial concentration. Therefore, a greater amount of adsorbent is required for sustaining proper removal at higher concentrations. The increase in adsorption capacity is due to an increase in the driving force of the concentration gradient by the increase in the initial concentration (Isaac and Sivakumar 2013; Takdastan et al. 2018). Decontamination of cadmium by different methods is shown in Table 3.

Table 3 Removal of Cd(II) by different adsorbents
Full size table

The effect of time, isotherms, and kinetic studies

The results showed that by increasing the contact time, Cd removal efficiency followed the same trend but with a steep slope during the first 30 min continuing by a steady increase in next 30 min until reaching to 60 min as the equilibrium time. This occurs due to the fact that in the primary stages, there is a great number of blank area on the adsorbent surface which, by passing the time, was saturated by cadmium ions. According to Fig. 4a, b, the obtained results stated that Freundlich equilibrium (R2 = 0.99) and pseudo-second-order kinetic models (R2 = 0.99) have been best fitted for describing the adsorption reaction. The number of isotherm constants, Kf, KL, and BT, and correlation coefficients of isotherm and kinetic equations is expressed in Table 4. In general, the values of n are in the range of 1–10 which illustrate that the adsorption condition was favorable. The obtained results of the Freundlich model indicate that constant n was greater than 1 that clearly shows the favorable removal conditions.

Fig. 4
figure 4

Modeling the adsorption isotherm of Freundlich and second-order kinetic at Cd adsorption a contact time 1 h and pH = 6 and temperature of 25 °C; b initial Cd concentration 20 mg/L at the different time, c effect of temperature and van’t Hoff regression plot for thermodynamic parameters and the adsorption of Cd onto nanocomposite in various temperatures and pH = 6

Full size image
Table 4 The study results of isotherms and their kinetics equations
Full size table

Effect of temperature and thermodynamic studies

The effect of temperature on cadmium removal was investigated in various temperatures (20 to 40 °C), 1 g/L dosage of adsorbent, and the initial Cd concentration of 20 mg/L. According to the obtained results, by moving the temperature from 20 to 40 °C, the Cd removal efficiency was enhanced from 51.7 to 78.8%. It indicates that the adsorption process is chemical. Temperature increase causes reduction in solution’s viscosity which motivates the mobility and increase in diffusion rate of cadmium ions. This condition leads to the bond between the adsorbent and adsorbate molecules which become stronger. This could be stemmed from the existence of available vacant surfaces during the early stages of the adsorption process (Javadian et al. 2015). According to Table 5, reduction in ΔG° indicates that the adsorption process has a better performance at higher temperatures and the positive value of ΔH° shows that the adsorption process is endothermic (Boparai et al. 2011). Figure 4c depicts the regressions of van’t Hoff plot for thermodynamic parameters.

Table 5 Thermodynamic parameters at different temperatures
Full size table

Desorption

To determine the reusability of the nanocomposite first, 20 mg/L of cadmium solution containing 1 g/L nanocomposite was shacked in optimum condition, and then, the solution was centrifuged and nanocomposite was separated from the solution. Finally, the amount of Cd(II) removal was determined by atomic absorption spectroscopy (AAS). To perform desorption from three desorbing agents HNO3, distilled water, and NaOH was used. At first, 20 ml of 1 M HNO3, distilled water, and NaOH solution under same condition were added to 1 g/L of adsorbent loaded with Cd(II) and the solution was stirred for 30 min. Then, the adsorbent was separated from the solution followed by drying in an oven at 110 °C for 60 min (Tesi 2014). The process was repeated for four times. The results obtained in Fig. 5 revealed that after four cycles, HNO3 was the best eluent in which nearly 38% Cd(II) ions were recovered. At acidic conditions (higher H+ ions concentration), the adsorbent surface tends to do the desorption process by replacing the adsorbed metal ions on the adsorbent surface. Decreased Cd(II) removal after four cycles was due to the saturated active sites on the nanocomposite surface along with an amount of nanocomposite lost during the sorption–desorption process (Ghoneim et al. 2014). Desorption ratio (DR %) was calculated through Eq. (7), as follows:

Fig. 5
figure 5

Reusability of Fe3O4@Z nanocomposite using desorbing agents at (pH = 6, initial Cd concentration = 20 mg/L)

Full size image
$$ {\text{DR}}\% = \frac{{{\text{Amount}}\,{\text{of}}\,{\text{desorbed}}\,{\text{metal}}\,{\text{ion}} \times 100}}{{{\text{Amount}}\,{\text{of}}\,{\text{adsorbed}}\,{\text{metal}}\,{\text{ion}}}} $$
(7)

Reactor’s performance modeling

In water and wastewater treatment processes, modeling each unit is beneficial to operators in a way that they could more effectively control their processes. In this study after studying the isotherms and thermodynamics of the proposed adsorption process, the reactor’s operational conditions were modeled. The CurveExpert outputs are presented in Fig. 6a–c. As can be seen, the removal efficiency had nonlinear relation with all the three operational parameters (pH, adsorbent dose, and Cd concentration) with a high level of fitness (over 0.99 for all the models) which was reciprocal quadratic for pH, a rational function for dose, and polynomial fit for concentration. It is noteworthy that all of the models were run at the optimum contact time (60 min).

Fig. 6
figure 6

Modeling the relation between the operational parameters and reactor’s removal efficiency in contact time = 60 min a pH, b nanocomposite dose, c Cd concentration

Full size image

Conclusion

In the current study, a new nanocomposite (Fe3O4@Z) was synthesized and applied for adsorptive removal of Cd under influence of various parameters such as solution pH, nanocomposite dosages, initial Cd concentrations, and contact time. The obtained results demonstrated that the maximum Cd uptake capacity was observed at pH = 6 and 60 min contact time. Furthermore, a direct relationship was observed between adsorbent dosage and adsorptive removal of Cd, while Cd removal efficiency showed a decreasing trend along with the enhancement of initial Cd concentrations. Among the applied equilibrium and kinetic models, Freundlich equilibrium and pseudo-second-order kinetic models represented the best performance regarding the fitness of the experimental data of Cd removal.