Adsorption of Calcium(II), Cadmium(II) and Copper(II) Ions from Water by Prepared Silicon-Aluminum-Zirconium Oxide Using Sol-Gel Process

Ternary oxide silica-alumina-zirconia (SiO2/Al2O3/ZrO2) obtained by the sol-gel process was applied as novel adsorbent of ions in aqueous solution. Batch isotherms were performed using standard Cu, Cd and Ca solutions at variable pH and different concentrations. Adsorption capacity was best at weakly acidic pH. The maximum adsorption capacities were 2.28 mg g for Ca, 9.89 mg g for Cd, and 14.88 mg g for Cu. The adsorption data fitted well to the single and dual-site Langmuir-Freundlich isotherm models. However, since the adsorption is very low in the sites with high energy, the single-site Langmuir-Freundlich provided a good fit as well. Whereas, the site responsible for the most ion adsorption was assigned to the Zr−OH group. A real sample of produced water was also tested. The new adsorbent showed a wide ability to retain many metal ions (alkali-earth, transition, and p-block metals) and even some anions were also caught by the adsorbent. The ternary oxide silica-alumina-zirconia was found to be a promising alternative adsorbent for metal ions in aqueous media.


Introduction
The presence of potentially toxic metal ions in wastewater, sediments, and soils has an anthropic origin as their main source and is nowadays an important environmental issue due to their toxicity to living beings. 1 Among the various treatment technologies, adsorption is a fast and wide-ranging method for removing metal ions with high efficiency and low cost. Various materials, such as zeolites, 2,3 nanomaterials 4-9 and polymeric materials, 10,11 have been used as adsorbents for the removal of toxic and nontoxic metal ions from wastewater. 12,13 Calcium is the most abundant metal in the human body, corresponding to 1 to 2 wt.% of its mass. This element is essential to the formation and maintenance of the bone matrix and stabilization of the membranes of excitable cells such as muscles and nerves. It also takes part in the blood clotting process, the activity of several enzymes and the growth and development of bones and teeth. The concentration of calcium in tissues is variable and in cells, it is found mainly in the endoplasmic reticulum and mitochondria. 14 Cadmium is a non-essential and highly toxic metal. The main anthropic sources are electroplating, mining, pigment manufacturing, alkaline batteries, and metallurgical activities. 15 Prolonged exposure to this element and its salts can cause kidney failure, anemia, cardiovascular disease, growth impairment, and loss of taste and smell. For these reasons, environmental regulatory agencies define severe limitations on the maximum concentration of cadmium in natural water bodies, as well as the maximum concentration allowed for the discharge of wastewater. 16 Copper is considered an essential trace element for human health. It is found in all fluids and various tissues, being of vital importance for many metabolic processes. 17 However, an excess of this element is toxic due to its affinity with the sulfhydryl (−SH) groups of several proteins and enzymes, being associated with diseases such as Wilson's disease, epilepsy, melanoma, and rheumatoid arthritis, as well as loss of taste. 17,18 Exploration and production of oil and natural gas, in onshore or offshore platforms produce wastes and effluents, among which produced water stands out, which is extracted together with oil and the gas. This residue consists of the formation water naturally present in the geological origin of the oil reservoir, and injected water into the reservoir to increase the productivity of the well. 19 Produced water represents the chief effluent for companies involved in the oil sector. 20 Several metal ions and anions are found in produced waters. Their amount depends on the producing field, particularly with regard to age and geological formation. 21 The concentration of most metal ions is usually much above the levels found in seawater in an uncontaminated environment, 22 even more than a thousand times higher. The toxic and harmful effects of the produced waters are due to the nature and concentration of the various components present, 23 e.g., cadmium, arsenic, copper, lead and mercury.
In recent years, several ternary oxides of the SiO 2 /M x O y /N z O w type have been synthesized by the sol-gel process, among them ZnO/TiO 2 /SiO 2 , 24 34 (v) adsorption of effluent dyes (SiO 2 /Al 2 O 3 /TiO 2 ). 35 The use of these ternary oxides in the petroleum area is unprecedented and has several promising applications ranging from the control of inorganic contaminants in oil and oil products to the treatment of water effluents in refineries and produced waters in onshore and offshore platforms. Thus, in this work, we describe a detailed study on adsorption of standard Ca II , Cd II , and Cu II and a real sample of produced water by a silicon-aluminumzirconium ternary oxide (SiO 2 /Al 2 O 3 /ZrO 2 ) prepared by the sol-gel process. This study stands out, in the future, the use of ternary adsorbents in the oil industry, where ZrO 2 has more basic properties than Al 2 O 3 and SiO 2 , which allows opportunity to adsorb a greater range of ions.

Experimental
Properties of the SiO 2 /Al 2 O 3 /ZrO 2 ternary oxide The detailed study of synthesis and characterization of SiO 2 /Al 2 O 3 /ZrO 2 ternary oxide (SiAlZr-A) has been published elsewhere. 29 In this work, the SiAlZr-A material will simply be called SiAlZr.
The main properties of the SiAlZr ternary oxide are: (i) chemical composition: 34.9 wt.% of ZrO 2 and 4.7 wt.% of Al 2 O 3 in the silica matrix; (ii) Brunauer-Emmett-Teller (BET) specific surface area (S BET ), 349 m 2 g −1 ; (iii) mean average pore size, 33.4 Å, according to the Barrett-Joyner-Halenda (BJH) method, indicating that this ternary oxide is predominantly mesoporous; (iv) mean pore volume, 0.021 cm −3 g −1 . According to X-ray photoelectron spectroscopy (XPS) and scanning electron microscopy with energy dispersive spectroscopy (SEM-EDS) data, the synthesized materials are homogeneous. This material presents good thermal stability according to thermogravimetric-differential thermal analysis (TGA-DTA) data. From X-ray diffraction (XRD) analysis, crystalline phases in the ternary oxide were found only after heating at 1073 K. CO 2 -TPD (temperature programmed desorption) showed the presence of weak and strong basic sites for the material, and NH 3 -TPD pointed to the presence of weak acid sites only, but pyridine adsorption detected both Lewis (LAS) and Brønsted (BAS) acidic sites. Due to these characteristics, 29 the material is promising for the application described in this paper.

Preparation of solutions
Cu II 1.2000 g of high purity metallic copper metal (Sigma-Aldrich, Saint Louis, Missouri, USA; > 99.995%) was weighed on an analytical balance. The metal was transferred to a beaker and 20.0 mL of concentrated nitric acid were added. The resulting solution was evaporated to dryness, followed by dissolution in ultrapure water. After cooling to 298.15 K, the solution was quantitatively transferred to a 1000 mL volumetric flask and the volume was made up with ultrapure water. , a primary standard, were weighed on an analytical balance. This reagent was transferred to a beaker and 20.0 mL of 4 mol L −1 HCl were added to dissolve it. The resulting solution was heated for the total expulsion of the carbon dioxide (CO 2 ). After cooling to 298.15 K, the solution was quantitatively transferred to a 1000 mL volumetric flask and the volume was made up with ultrapure water.

Sample of produced water
The real sample chosen for the application of this material was a produced water from the Almirante Barroso Maritime Terminal (TEBA), located in São Sebastião (São Paulo State, Brazil), provided for this study by the Applied Electrochemistry Laboratory of the Metallurgical and Materials Engineering Program (PEMM-COPPE-UFRJ). Its main physico-chemical properties are pH, 7.4; electrical conductivity, 7.86 μS cm −1 ; oils and waxes, 10 mg L −1 . Its chemical characterization is shown in Table 1. 37 Adsorption assays

At different pH
The pH of the solution exerts a profound influence on adsorption process, since pH variations impact not only on protonation of functional groups of adsorbent, but also on the degree of ionization of the metal ions. 38 The pH range of assays was chosen between 2-6 in order to avoid the precipitation of ions as metal hydroxides, which occurs at a pH above 6. 38,39 The increase in adsorption capacity at pH > 6 may result from a combination of adsorption and precipitation on the surface of the adsorbent. 40 The adsorption isotherms for Ca II , Cd II and Cu II ions in aqueous medium were performed by batch technique. In this study, the concentration of the metal ion was firstly kept constant and the pH was varied. For each metallic ion, five aliquots of the adsorbent were weighed, each containing approximately 100 mg. The aliquots were transferred to flasks containing 30 mL of a solution consisting of 5 mL of a specific buffer solution for each pH (ranging from pH 2 to 6) and 25 mL of the metal ion solution. The flasks were placed in a thermostatic bath (Nova Ética Dubnoff) and stirred for 24 h at a controlled temperature of 298.16 K.
After the experiments, the system was transferred to Falcon-type tubes and centrifuged (Biomedics 1970). The supernatant was analyzed by atomic absorption spectrometry (Shimadzu AA6800) with an automatic sampler and an air-acetylene flame. The amount of metal ion adsorbed on the adsorbent surface was determined by the difference between the amount present in the supernatant and the amount initially present in the original solution.
The equilibrium time is of fundamental importance to guarantee the minimum time necessary to promote an effective contact between the adsorbent and the metal ion solution. For this reason, the experiments were run for 24 h. As in most of studies involving metal ions adsorption, temperature was set up at 298.16 K.

At different metal ion concentration
In these experiments, the pH was maintained constant, and the concentration of the metal ions was varied by diluting the standard solution with the appropriate amount of ultrapure water. The batch procedure and the analytical protocol follow the same procedures described earlier.
The experimental data obtained from this assay were fitted to non-linear models of Langmuir, Freundlich, single, and dual-site hybrid Langmuir-Freundlich. 26 The Langmuir model is described by equation 1, where Q e is the amount adsorbed (mg g −1 ); C eq is the concentration of the adsorbate in the solution at equilibrium (mg L −1 ); b is the maximum adsorption capacity (mg g −1 ) and K L refers to the equilibrium constant of Langmuir (L mg −1 ), which is related to the energy of adsorption. This model predicts a fixed number of energetically homogeneous sites, in which the interaction with the adsorbate occurs, but there are no interactions between the adsorbed species. Therefore, when the saturation occurs, a monolayer of the adsorbate is formed on the surface of the adsorbent. 41,42 The separation factor (R L ) is a dimensionless constant that can express essential features of the Langmuir isotherm and is defined by the equation 2, where C 0 (mg L −1 ) is the initial adsorbate concentration and K L is the Langmuir constant. The value of R L indicates the shape of the isotherms to be either unfavorable (R L > 1), linear (R L = 1), favorable (0 < R L < 1) or irreversible (R L = 0). 43 (1) The Freundlich model is represented by equation 3, where K F refers to the equilibrium constant of Freundlich (mg g −1 ) (L g −1 ) and n is the empirical constant of Freundlich (dimensionless), referring to the heterogeneity of adsorption sites. According to this model, the amount of the analyte adsorbed is proportional to its concentration in the solute. Thus, it assumes that, theoretically, the adsorption is infinite. Also, it predicts that the adsorption may occur via multiple layers rather than a monolayer. 41,42 Q e = K F × C eq 1/n Single-site Langmuir-Freundlich model is originated from the combination of Langmuir and Freundlich models, and it is represented by equation 4. This model assumes the existence of energetically homogeneous sites. When the adsorbate is in low concentration, it is reduced to the Freundlich model, while when in high concentration, it assumes a monolayer adsorption characteristic of the Langmuir model. On the other hand, the dual-site Langmuir-Freundlich model, described by equation 5, assumes the existence of adsorption sites with different energies. According to this model, the adsorbate interacts with sites with low energy when it is at low concentrations, while, at high concentrations, interacts also with high energy. 41,42 (4) In equations 4 and 5, K is the equilibrium constant (L mg −1 ) representing adsorbent-adsorbate affinity, and the subscript numbers 1 and 2 refer to the binding sites.
Tests with a real sample of produced water Approximately 300 mg of the adsorbent were weighed and transferred to flasks containing 90 mL of a solution consisting of 15 mL of the buffer solution (pH 5 or 6) and 75 mL of the produced water. The experimental conditions were the same described earlier. Each sample was weighed before and after the adsorption assay. The adsorbent material after the test was sent for X-ray fluorescence by wavelength dispersion (WDXRF) analysis (AxiosMax model from Panalytical). Samples were prepared by melting a 1:10 m/m dry sample to flux ratio at 1323.15 K in a Vulcan machine. The flux was composed of lithium borates (66 wt.% Li 2 B 4 O 7 , and 34 wt.% LiBO 2 ). The results were expressed in wt.% as oxides and normalized to 100 wt.%. Loss by calcination was determined by roasting weighed samples in a furnace at 1273.15 K for 16 h. After cooling, they were weighed on an analytical balance.

Effect of pH
The adsorption isotherms are a useful way to study the metal ion removal capacity by different solid adsorbents. The adsorption isotherms allow the development of the adsorption curve equations, obtained by the mass balance after performing batch tests varying the pH and concentration of each ion studied (Ca II , Cd II and Cu II ions).
The amount of metal ion adsorbed on the adsorbent surface was determined by the difference between the amount present after the batch and the initial amount. The calculation of adsorption equilibrium in batch assays was evaluated using equation 6: where q e is the amount of solute adsorbed per unit mass of adsorbent, in equilibrium (mg g −1 ), C o is the initial concentration of adsorbate in the liquid phase (mg L −1 ), C e is the final concentration of adsorbate or equilibrium concentration in the liquid phase (mg L −1 ), V is the volume of sample used (L) and M is mass of the adsorbent material (g). 44 Initially, the best pH was chosen for the adsorption of each ion. For this purpose, the adsorption isotherms were made, in which the adsorbed quantities (Q e ) of Ca II , Cd II , and Cu II , were plotted on the ordinate axis, while the abscissa axis is plotted the pH of the solution. The isotherms are represented in Figure 1. Through this graphical analysis, it was possible to select the pH that maximizes the adsorption of each ion. For Ca II and Cd II ions, the highest efficiency was found at pH 6, while for Cu II ions the best result was at pH 5. The pH assays show that the factors that affect the adsorption characteristics of adsorbents are: (i) the competition of H + ions with metal ions for adsorption at low pH; (ii) maximum adsorption at pH 5-6; (iii) the precipitation of hydroxylated species in adsorbents at pH above 6. 40 The studies were not carried out at pH < 2, as it occurs the protonation of the silanol group of matrix generating the SiOH 2 + species, leading to the dissolution of the SiAlZr material. 45 Although there is apparently a smaller amount of Cu II in solution at pH 6, this does not mean a greater adsorption, instead, at this pH copper precipitated as Cu(OH) 2 , which reduced the amount available for adsorption. This can be attributed to additional restrictions on the mass transfer imposed by the formation/deposition of suspended solids. For this reason, this point was disregarded in the analysis of this ion. This problem did not take place in the case of Cd II and Ca II as their hydroxides only precipitate at higher pH. 46 Effect of the initial concentration of the metal ion Figure 2 shows the isotherms for Ca II , Cd II and Cu II , in which the quantities of each ion adsorbed per unit of adsorbent mass (Q e ) are plotted on the ordinate axis, while the equilibrium concentration of the metal ions in solution (C eq ) is plotted on the abscissa.
With the data resulting from the batch test, the parameters could be estimated by adjusting theoretical curves to the isotherm data. This adaptation was performed based on the non-linear models of Freundlich, Langmuir, single and dual-site hybrid Langmuir-Freundlich equations. The most appropriate model was selected by comparing the values of the correlation coefficients (R 2 ), the root mean square error (RMSE), and the theoretical maximum adsorption capacity (b) close to the experimental one for the SiO 2 /Al 2 O 3 /ZrO 2 adsorbent and each ion studied. The adjusted models can be found in Supplementary Information, while the calculated parameters for each model are shown in Table 2.
The experimental maximum adsorption capacity of the ions was 2.28 mg g −1 for Ca II , 9.89 mg g −1 for Cd II , and 14.88 mg g −1 for Cu II . Therefore, for adsorption, the best adjustments were obtained by using both the single and dual-site Langmuir-Freundlich models for all metal ions. The dual-site Langmuir-Freundlich model predicts the existence of two adsorption sites with different affinities towards the analyte, determined by the parameters K 1 and K 2 (L mg −1 ). However, it can be observed that approximately 80% of the Ca II adsorption and around 95% of the Cd II and Cu II adsorption occur in the lower affinity sites (K 1 ). Therefore, this explains the goodness of the fit provided by the single-site Langmuir-Freundlich model as well. 26 Beyond that, as shown in Table 2, R L values regarding the Langmuir features for all ions studied are in the range of 0-1, which indicates their favorable adsorption on SiO 2 /Al 2 O 3 /ZrO 2 adsorbent. It is important to mention that the R L values in Table 2 were calculated using the first initial concentrations of each ion, and lower R L values are  obtained in higher initial concentrations. Once the degree of favorability is generally related to the irreversibility of the system and it tends towards zero as the C 0 approaches equilibrium, it can be assumed that the ions adsorption on SiO 2 /Al 2 O 3 /ZrO 2 adsorbent becomes more suitable at high concentrations. 47 The surface sites for the ions adsorption are represented by Si−OH, Al−OH and Zr−OH. Considering that the Al 2 O 3 percentage in the material is very low, it is likely that the adsorptions occur mainly in the Si−OH and Zr−OH sites. Additionally, since zirconium oxide has a more basic character than silanols, Zr−OH might be considered the binding site responsible for the most metal ions adsorption. 26,29 Table 3 shows the comparison of adsorptive capacities between the material studied and others found in the literature, at each optimum pH. Analyzing these data, our adsorbent is one of the least effective, but this comparison is not trivial, as the studies are carried out under different conditions. Another important detail is that our study was introductory and that there was no chemical modification over the adsorbent surface, as occurred in other studies, these alterations can increase the specificity for a determined ion, causing an increase in its adsorptive capacity. 8 In future works this chemical modification will be studied.
Application in a real sample of produced water In this case, two different pH were studied to better assess the influence of the acidity. Table 4 shows the amount of adsorbed ions in the batch assays using the produced water as a real sample. This table shows the mass composition of each chemical species according to WDXRF data, including the components of the SiO 2 /Al 2 O 3 / ZrO 2 adsorbent.
As expected, the introduction of species from the sample of produced water reduced the original levels of Si, Al and Zr. This effect decreased in the order Al > Zr >> Si. It was not possible to define whether phosphorus entry occurred via adsorption at exchange sites or is simply an effect of the deposition of metal phosphates precipitates (such as FePO 4 ) on the adsorbent surface at the pH of the experiments. However, the significant increase in the amount of chloride in the adsorbent after the tests suggests the insertion of foreign anions in anion exchange sites (basic OH -) since the metals detected in Table 4 do not form insoluble chlorides. 53 This same effect was noticed, although in a lower degree, in the case of the sulfate ion. The increase in the percentage of calcium (and magnesium) after the experiment at pH 6 correlates with the higher adsorption at that pH in assays with Ca II standard solution. The adsorption of other cations, especially sodium ion (due to its high concentration in the produced water) also increased with the increase in pH, reflecting the exchange on cationic sites (O-H + → O-Na + ). This suggests that the prepared materials present acidic and basic exchange sites, which is a very useful property for treating effluents containing a variety of components such as produced waters.

Conclusions
The adsorption capacity of SiO 2 /Al 2 O 3 /ZrO 2 decreased in the order Cu II > Cd II >> Ca II . The model that best describes the adsorption isotherms was the dual-site Langmuir-Freundlich. However, since the adsorption is very low in the sites with high energy, the single-site Langmuir-Freundlich provided a good fit as well. The site responsible for the most ion adsorption was assigned to the Zr−OH group. In tests with a sample of produced water, the SiO 2 /Al 2 O 3 /ZrO 2 material proved to be quite promising as an adsorbent for alkali and alkali-earth metal ions in high concentrations. The ability to retain anions leads to the presence of both acidic and basic exchange sites. It must be emphasized that the assays with the produced water were carried out in a single condition. The synthesized adsorbent has potential to be used for removal of inorganic ions from the produced water at weakly acidic pH. The optimization of the process can broaden the prospects of this new adsorbent for removing metal ions from complex effluents.

Supplementary Information
Supplementary information is available free of charge at http://jbcs.sbq.org.br as PDF file.