Wastewater treatment has now become one of the most serious environmental challenges. Heavy metal pollutants are mostly exposed to the environment with the activities of various industries such as textile, mining, battery manufacturing, metal plating industries, etc . Compared with other wastewaters, heavy metal-polluted waters are very dangerous since they directly expose human health to severe risks. The effects of lead on kidney, nervous system, reproductive system, liver, and brain ; mercury on the central nervous system, lung, and kidney ; cadmium on lung, kidney, liver, and reproductive organs ; and nickel on skin, lung, mucous membranes, and nervous system  are some of the associated health problems. Besides; unlike other contaminants, heavy metals cannot be processed by microorganisms and may accumulate in living organisms.
Faced with the mentioned problems, various physical and chemical methods; such as chemical precipitation, solvent extraction, ion exchange, coagulation and sedimentation flotation, membrane filtration, adsorption, and photoreduction have been applied for removing soluble metals from various aqueous solutions [6-12]. However, many of these techniques have some disadvantages. Low efficiency and production of a large amount of sludge in precipitation methods, requirements of expensive materials and high energy consumption in membrane methods are some of these limitations. In contrast, along with all these physicochemical methods; adsorption is recognized as an effective, low-cost, and easy to operate method . Moreover; in some physical adsorption processes, the adsorbent can be regenerated and recycled by suitable desorption process. Development of cost-effective adsorbents with high efficiency has attracted a lot of interest in recent decades. The most current adsorbents for heavy metal removal can be categorized into activated carbon [14, 15], zeolites [16, 17], low-cost adsorbents (such as agricultural wastes and industrial byproducts) , and bioadsorbents .
Activated carbon (AC) adsorbents and zeolites have been widely used in adsorption processes; nevertheless, they are relatively expensive and sometimes have complicated synthesis methods. On the other hand, although bioadsorbents and low-cost adsorbents have partly reduced the cost of operation, but the removal capacity of these materials is quite low.
In recent years, nano-adsorbents have attracted a considerable attention because of their enhanced adsorption rate, capacity, and selectivity for heavy metals. Nanosized metal oxides, especially magnetic nanoparticles and their nanocomposites with organic compounds [20-22], and carbon nanotubes [2, 23] are the most studied nano-adsorbents. However, these adsorbents are usually had expensive synthesizing processes.
Calcium carbonate nanoparticles have shown a good capacity for the removal of heavy metals [24, 25]. Recently, a novel low-cost method for generation of calcium carbonate nanoparticles has been developed based on colloidal gas aphron (CGA) system . In this experimental work, the prepared calcium carbonate nanoparticles generated in a CGA system were characterized and their applicability in lead and iron ions removal was investigated. Their adsorption kinetics and capacities in aqueous solutions for the undesirable ions were reported.
MATERIALS AND METHODS
All the chemicals used in the study were of analytical grade and used as received without further purification. Calcium chloride dihydrate (CaCl2.2H2O), sodium carbonate (Na2CO3), lead(II) nitrate (Pb(NO3)2), and iron(II) sulfate heptahydrate (FeSO4.7H2O) were purchased from Merck. Hydrochloric acid (HCl) was obtained from Nirou Chlor Co. (Isfahan, Iran). Absolute ethanol (>99.5%) was obtained from Zanjan Co. (Zanjan, Iran). Sodium dodecyl sulfate (SDS, 99% purity) was purchased from FSA (England). The water in this work was Milli-Q ultrapure water.
Preparation and characterization of CaCO3 nanoparticles
Synthesis of CaCO3 nanoparticles was carried out using a colloidal gas aphron (CGA) generator similar to that described by the others [27-29]. Briefly, 212 mg (1mmol) of Na2CO3 and 200 mg of SDS were dissolved in 450 mL water as the mother solution for CGA preparation. Then 50 mL of a solution containing 147 mg (1mmol) of CaCl2.2H2O was steadily injected to the prepared CGAs for 15 minutes. The precipitated particles from the totally collapsed foam were centrifuged (9000 rpm, 15 min) and rinsed several times with the copious amount of cold ethanol and distilled water for removing any residual surfactant. The detailed characterization results of the obtained nanoparticles were reported in our previous work .
Heavy metal ions adsorption experiments
Stock solutions of heavy metal ions (1000 mg/L) were prepared by dissolving the precise amount of Pb(NO3)2 and FeSO4.7H2O in distilled water. These solutions were diluted to 10–200 mg/L for adsorption experiments. The initial pH of solutions was adjusted using diluted HNO3 when needed. All the adsorption tests were performed batch-wise using a thermostatic shaker with controlled temperature and shaking speed (120 rpm). Briefly, 0.025-0.075 g of CaCO3 was added to 50 ml of metal solution (10-200 mg/L) and shaken for 2 hours. The final concentrations of heavy metal ions in the samples were measured using atomic adsorption spectroscopy (VGP-210, Buck Scientific Inc). The adsorption capacity (qe, mg/g) and removal efficiency (Re (%)) of heavy metal ions were calculated using the following equations:
Where C0 and Ce are the initial and equilibrium concentrations of solute (mg/L) in the solution, V is the solution volume (mL) and w is the mass of adsorbent (CaCO3) (mg).
Design of experiments
The effects of temperature, heavy metal concentration, and adsorbent dosage on the removal capacity of CaCO3 nanoparticles were evaluated by the response surface methodology (RSM) using DESIGN EXPERT 7.0.0 software (Stat-Ease Inc, Minneapolis, MN, USA). Optimum conditions were determined using a 3-level, 3-factor Box-Behnken Design (BBD). Table 1 represents the coded and uncoded ranges of selected parameters.
The adsorption capacity of calcium carbonate (Y) was expressed as a function of selected independent variables using a second order polynomial equation (Eq. 3).
Where Y is the predicted response, Xi and Xj (i, j=1…k) are independent parameters, β0 is the intercept constant, βi, βii and βij (i, j=1…k) are linear, square, and interactive coefficients respectively and ε is a random error.
RESULTS AND DISCUSSION
Characterization of CaCO3
Physical properties of the adsorbent were reported elsewhere . The produced particles had monosized vaterite polymorph with the main particle size of 100 nm and a surface area and average pore diameter of 76 m2 g-1 and ~6.1 nm (SEM, XRD, and BET analysis).
Effects of pH and contact time
The removal percentage of CaCO3 (1g/L) for Pb2+ and Fe2+ ions (200ppm) was evaluated in a pH range of 2.5-6 and not more than 6 for avoiding their precipitation (Fig. 1). It can be observed that the adsorption increases with an increase in pH, and reaches a plateau value at the pH range of 5.5-6. Lower pH decreases the removal efficiency of heavy metals due to an increase in solubility of each ion in the solution. Moreover, lower pH values partially dissolve the calcium carbonate adsorbent particles. However, in pH values more than the precipitation limits of heavy metal ions in aqueous solutions, precipitation interferes with the adsorption process (pH>6). According to Fig. 1, the pH of 5.5 was selected for all the following experiments.
The kinetics of adsorption is shown in Fig. 2. The removal percentage of both metals on the adsorbent increases with contact time. It can be seen that at all the selected concentrations, more than 80% of the process was completed during the first hour and it achieves to equilibrium in the second hour. Therefore, a contact time of 2 hours was chosen for all the following experiments.
In order to investigate the mechanism of reactions, the kinetic parameters were evaluated using pseudo-first-order (PFO)  and pseudo-second-order (PSO)  kinetic models. The obtained data from Fig. 3 were utilized for this purpose. The linear forms of PFO and PSO equations can be expressed as:
Where qe and qt are the sorption capacity (mg/g) at equilibrium and time t (min), and k1 (min-1) and k2 (g/mg.min) are the rate constants of the PFO and PSO equations respectively.
The adsorption kinetic constants calculated based on the fitting results of PFO and PSO equations (Fig. 3) are summarized in Table 2. The higher R-squared values of the PSO model (R2>0.99) compared with the PFO model show that the experimental data were better described by this model.
Adsorption isotherms and effect of temperature
Fig. 4 demonstrates the adsorption isotherm of Pb2+ and Fe2+ adsorbed onto CaCO3 nanoparticles over a wide range of initial concentration (10–4000 mg/L) and various temperatures. It can be seen that the equilibrium adsorption capacity of adsorbent increases at higher metal concentration. The maximum equilibrium uptake for Pb2+ ion (1210±30 mg/g) was noticeably more than Fe2+ ion (845±8 mg/g). However, no significant changes in the isotherms were observed with increasing the temperature.
Langmuir and Freundlich adsorption isotherms were used to analyze the mechanism of adsorbate−adsorbent interactions (Eqs. 6 and 7). The linear forms of Langmuir and Freundlich equations can be expressed as follow:
Where Ceq is the equilibrium concentration of adsorbate ions in the solution (mg/L), qmax is the maximum adsorption uptake of metal ions at equilibrium (mg/g), and KLang (L/mg) and KFreu (mg/g) are the Langmuir and Freundlich constants. The calculated Langmuir and Freundlich parameters are presented in Table 3. The R-squared values show that the experimental data are better fitted to Langmuir isotherm model (R2>0.99). The values of Freundlich constants, KFreu and n indicate the favorable adsorption of metal ions by CaCO3 nanoparticles over the entire studied concentrations (n>1) .
The extraordinary amounts calculated for the maximum adsorption uptake obtained from the Langmuir isotherm were in good agreement with similar studies [24, 25]. Ma et al. and Cai et al. mentioned that the CaCO3 adsorption mechanism is not just a simple monolayer adsorption and transformation of precipitation between CaCO3 and heavy metal carbonates could be the explanation for the large adsorption capacity of CaCO3. However, natural types of calcium carbonate adsorbents employed in some studies did not show such great adsorption uptakes [33-36]. For instance; the maximum adsorption uptake reported for lead, varies from 0.0167 mg/g to 3242.48 mg/g using limestone  and hierarchical CaCO3–maltose meso/macroporous hybrid materials , respectively. We found that the size and crystal structure of calcium carbonate are important to improve the transformation of precipitation between CaCO3 and heavy metal carbonates. Calcite is the most thermodynamically stable polymorph of calcium carbonate in nature and has a lower solubility in comparison with other polymorphs of calcium carbonate like vaterite and amorphous calcium carbonate (ACC) . Moreover, nanosized particles may also show higher solubility in aqueous solutions rather than large ones . Therefore, nanosized and more soluble polymorphs of CaCO3 used by Ma et al. and Cai et al. may have facilitated the adsorption mechanism proposed by them [24, 25]. The large adsorption capacity of vaterite nanoparticles observed in this study could be the result of a similar transformation of precipitation between CaCO3 and heavy metal carbonates.
Analysis of Experimental Data
The effects of temperature, adsorbent dosage, and heavy metal initial concentration were evaluated by using an experimental procedure based on Box-Behnken design with five center point replicates, which resulted in 17 experimental runs (Table 4).
According to the RSM results, quadratic model equations were employed to express the association between the response and independent variables (Eqs 8 and 9). The insignificant variables are excluded in these equations.
Interactions among the independent variables and obtained responses were analyzed using the analysis of variance (ANOVA) (Table 5). Results of the Fisher’s F-test with a low probability (P-value<0.0001) confirm that both models are highly significant. The F-values of different independent variables indicate that the only significant model terms are X2, X3, X2X3, and X22 (p-value<0.05). The fit quality of the models was further evaluated by the R2 value. The R2 values for Pb2+ and Fe2+ removal models were 0.9682 and 0.9723 respectively. This indicates that there is a good agreement between the experimental and predicted values. The adequacy of regression models with the experimental data is also graphically shown in Fig. 5.
Three-dimensional (3D) response surface and the corresponding contour plots were utilized to assess the interaction between two significant independent variables and their effects on the response (Fig. 6). Fig. 6 (a and b) shows that the removal uptake was increased with increasing heavy metal concentration and decreasing CaCO3 dosage. According to Fig. 6 (c-f), the temperature was found to be insignificant on the response for both Pb2+ and Fe2+ removal uptakes. The optimum values for maximizing the removal capacity of CaCO3 were also obtained using Design- Expert 7.0.0 software. The optimized parameters suggested by the software for temperature, CaCO3 dosage, and metal concentration were 65 °C, 0.5 g/L and 200 mg/L for both Pb2+ and Fe2+ ions which were in good agreement with the experimental data.
Removal of Pb and Fe ions from aqueous solutions by calcium carbonate adsorbent was examined in batch-wise experiments. It was found that the maximum observed adsorption capacities of particles were 1210±30 mg/g for Pb(II) and 845±8 mg/g for Fe(II), at 25°C and pH=5.5. Analyzing the experimental data by kinetic and thermodynamic models revealed that the high removal capacity of the nano adsorbent is not just because of a simple monolayer adsorption process but it could be the result of a transformation/ precipitation mechanism between the CaCO3 and heavy metal carbonates. The most important parameters were found to be adsorbent dosage and heavy metal concentration respectively. The effect of temperature was found to be negligible within the experimental range.
The authors wish to acknowledge the financial support from the research deputy at the Isfahan University of Technology in Isfahan, Iran.
CONFLICT OF INTEREST
The authors declare that there are no conflicts of interest regarding the publication of this manuscript.