Recovery of Vanadium (V) Oxyanions by a Magnetic Macroporous Copolymer Nanocomposite Sorbent

: An amino-functionalized magnetic macroporous copolymer of glycidyl methacrylate (GM) and ethylene glycol (E) dimethacrylate (m-poly(GME)-deta) was synthesized, fully characterized, and used to investigate the adsorption of vanadium (V) oxyanions from aqueous solutions (Ci = 0.5 mM) in a batch system at room temperature (298 K). Pseudo-ﬁrst-order (PFO), pseudo-second-order (PSO), Elovich, and intra-particle diffusion (IPD) models were used to analyze the kinetic data. The study showed that sorption is rapid, i.e., the sorption half-time is approximately one minute. Initially, the sorption process primarily involved surface sorbent particles, and it was best described by the PSO model. However, after saturation of the surface active sites is attained, the sorption rate decreases signiﬁcantly because of limitations of the diffusion rate, which is then primarily controlled by the IPD process. The sorption process is favorable in the pH range of 3–6 due to the strong electrostatic interactions between the absorption centers of copolymer and vanadium (V) oxyanions. In the stated pH range, deta absorption centers with two and three protonated N atoms are in equilibrium as studied by quantum chemical modeling. Among V(V) species present in diluted aqueous media, the adsorption of H 2 VO 4 − ions dominates.


Introduction
Vanadium occurs in nature in the form of 65 different minerals, mostly vanadinite, descloisite, carnotite, and patronite. Natural sources of vanadium in the atmosphere include dust (crushed soil particles), marine aerosol and volcanic emissions, anthropogenic industrial products, and fossil fuel byproducts. Sorption of vanadium (V) from aqua solution onto natural soil colloids depends on pH and ionic strength conditions [1]. Vanadium is widely used as a steel additive and catalyst in the metallurgical and chemical industries [2], and while being found in traces in nature, it represents a significant raw material. Various processes for the extraction of vanadium from the complex metallic mixture are in usage nowadays [3][4][5][6], but only a few are environmentally friendly.
Several studies showed that amino-functionalized GM-based copolymers possess a significant affinity for the sorption metal ions [7,8]. Moreover, these polymers are characterized by thermal and chemical stability [9,10]. Embedding magnetic particles in the polymer matrix of GMA further extends their functionality, simplifying the removal process by applying a magnetic field [11,12]. It was shown that the metal ion sorption process on the surface of macroporous polymer nanocomposites depends on various parameters, such as the properties of the sorbent (chemical structure, porosity, specific surface area, particle size), ion properties (chemical nature and charge), and sorption conditions, such as ion concentrations in solution, pH, temperature, presence of other competing species, and contact time. In this study, an amino-functionalized magnetic macroporous copolymer of GM and E dimethacrylate synthesized by in situ suspension copolymerization in the presence of Fe 3 O 4 (m) nanoparticles coated with 3-aminopropyltrimethoxysilane (APTMS) through a silanization process, functionalized with diethylene triamine (deta), m-Si-poly(GME)-deta, was fully characterized in terms of its structural properties and tested as a sorbent of vanadium (V) oxyanions.

Materials and Methods
The tested sorbent m-Si-poly(GME)-deta was synthesized as described in previous works [13]. Pore size distributions were determined by a high-pressure mercury intrusion porosimeter Carlo Erba Porosimeter 2000 (Washington, DC, USA, software Milestone 200). The SEM/EDX analysis of sorbent was performed on a JEOL JSM-6610LV SEM instrument (JEOL Inc., Peabody, MA, USA) equipped with EDX detector (X-Max Large Area Analytical Silicon Drift connected with INCA Energy 350 Microanalysis System, Oxford Instruments Nanoanalysis, High Wycombe, UK) after sputtering with gold in order to enable the imaging and to prevent thermal damage of the specimen.
The thermogravimetric analysis (TGA) (30-700 • C range) was performed on SDT Q600 TGA/DSC instrument (TA Instruments, New Castle, DE, USA). Heating rates were 20 • C/min using less than 10 mg sample mass. The furnace atmosphere consisted of dry nitrogen at a flow rate of 100 cm 3 /min. Further, m-Si-poly(GME)-deta was used to investigate the sorption of vanadium (V) oxyanions from diluted aqueous solutions (C 0 = 25 ppm) in a batch system at room temperature (298 K). The oxyanion concentrations in a solution were determined by inductively coupled plasma optical emission spectrometry (ICP-OES) (model Thermo Scientific iCAP 6500, Waltham, MA, USA). The PFO, PSO, Elovich, and IPD models were used to analyze the kinetic data.
The sorption capacity, i.e., the amount of vanadium (V) sorbed onto a unit mass of magnetic sorbent, Q, was calculated from: where C i and C are initial and concentration in the sample after sorption, V is the solution volume, and m is the sorbent mass. Kinetic data were analyzed with the PFO (Equation (2)), PSO (Equation (3)), Elovich (Equation (4)), and IPD (Equation (5)): where Q e (mg/g) and Q t (mg/g) are the amounts of sorbed ions at equilibrium and at time t, dQ t dt a e × e −b eQ t where a e is initial sorption rate (mg/g·min or mmol/g·min) and b e is a parameter determined by the surface coverage that depends on the activation energy during chemisorption (g/mg or g/mmol) [14]: where Q e (mg/g) and Q t (mg/g) are the amounts of sorbed ions at equilibrium and at time t, respectively, α e = Q t /Qe is fractional attainment of equilibrium, K p is the film diffusion rate constant, D F is a parameter of the linear driving force, k id is the intraparticle diffusion coefficient (µmol/g·min 0.5 ), t is the time, and C id is a constant, related to the thickness of the boundary layer.
For equilibrium data analysis, the Langmuir (Equation (6)), Freundlich (Equation (7)), and Tempkin (Equation (8)) isotherm models were used [15]: lnQ e = lnK F + 1 n lnC e (7) where Q max (mg/g) is the Langmuir maximum adsorption capacity, K L (L/mg) is the Langmuir isotherm constant, K F (mg/g)(g/L) 1/n and n are the Freundlich constant and isotherm exponent, respectively, K T (L/mg) is the Tempkin isotherm equilibrium binding constant, b T (mg/L) is the Tempkin isotherm constant, R is the gas constant (8.314 J/mol K), and T is the absolute temperature (K). The efficiency of the adsorption process can be predicted by the dimensionless Langmuir equilibrium parameter R L : The value of R L indicates the type of the isotherm either to be unfavorable (R L > 1), linear (R L = 1), favorable (0 < R L < 1) or irreversible (R L = 0).
The quantum chemical calculations were performed using the Gaussian09 software package. (Gaussian, Inc.). Four possible dimers of the studied systems were optimized by using the B3LYP functional, 6-311+g** basis set for non-metals, lanl2dz basis set for vanadium, and solvation model based on density (SMD) for water solvent. At the same level of theory, the interaction energies between vanadium (V) oxyanions and detaOH absorbent in the optimized dimer structures were calculated. DetaOH represents the chemical species formed in the reaction of deta amine and 1,2-epoxyethane, used as a model for the adsorption centers in the copolymer.

Structural Characterization
The porosity parameters for sample m-Si-poly(GME)-deta, i.e., the values of specific pore volume, V s , and pore diameter corresponding to half of pore volume, D V/2 , were read from the pore size distribution curves determined by mercury porosimetry, while the specific surface area, S s,Hg , was calculated as the sum of incremental specific surface areas from the pore size distribution curves as described in the literature [16]. Calculated values are presented in Table 1, along with the data for different samples of the macroporous copolymer of glycidyl methacrylate and ethylene glycol dimethacrylate: magnetic nonfunctionalized specimen m-poly(GME), magnetic sample grafted with diethylene triamine m-poly(GME)-deta, obtained in literature [17], and magnetic silanized non-functionalized m-Si-poly(GME). The silanization of magnetite leads to the sample with 2.7 times larger pore diameter and an almost 50% reduction in S s,Hg values compared to the non-silanized sample (D V/2 = 100 nm and S s,Hg = 67 m 2 /g for m-poly(GME) [17]). As illustrated in Table 1, amino-functionalization had no influence on porosity parameters.
Characterization of the m-Si-poly(GME)-deta sorbent by SEM microscopy (Figure 1) clearly indicated the presence of spherical particles, very smooth and uniform in morphology. As expected, the most intense peaks in the EDX spectrum ( Figure 2) originate from elements C and O. The presence of Fe peaks clearly indicates the incorporation of magnetite. A stronger Fe peak was detected on the surface of the sample particles and the Si peak (which confirms the silanization of magnetite particles). The N peak indicates the functionalization of amino groups and is predominantly oriented on the inner surface of the particles (~14 w/w % at the cross-section and 4.7 w/w % at the particle surface). The peak for Au in the SEM/EDX spectrum originates from Au coating of the m-Si-poly(GME)-deta. As one can see, TGA thermograms for both samples exhibited a multiple-step degradation pattern. The primary degradation mechanism of the samples is depolymerization to monomers and oligomers initiated by scission of weak linkages (i.e., H-H bonds, double bonds in backbone pendant groups and random chain scission) and ester decomposition that simultaneously occur [18]. Besides depolymerization and ester decomposition, the thermal degradation of m-Si-poly(GME)-deta also involves the elimination of amine groups. In addition, an initial loss of about 6 wt. % below 100 • C might be assigned to the loss of physically retained or occluded solvents [19,20]. The weight loss in the temperature range of 30-190 • C was mainly from the elimination of molecular water adsorbed on the copolymer surface.
The TGA curves were presented in Figure 3. The main parameters evaluated on the basis of these curves are presented in Table 2.  Since the boiling point of APTMS is 194 • C, physically adsorbed APTMS is completely desorbed to 300 • C [21]. The main decomposition of chemically bonded APTMS occurred up to 470 • C and 560 • C for m-Si-poly(GME) and m-Si-poly(GME)-deta, respectively. This is in agreement with the literature data showing that the thermal decomposition of grafted silane takes place at above 450 • C and the breaking of the C-Si bond starts at 450-510 • C [21,22]. A significant loss in both samples was observed during the second degradation step. The residue values (7 and 11 wt % for m-Si-poly(GME) and m-Si-poly(GME)-deta, respectively) are in agreement with the magnetite content incorporated in the investigated samples.

Vanadium Sorption
In this study, the sorption capacity of m-Si-poly(GME)-deta in diluted vanadium solution (0.5 mM) as a function of concentration, contact time, and pH was tested.
The experimental isotherm data are shown in Figure 4. Sorption was rapid. The sorption half time, t 1/2 (the time required to reach 50% of the total sorption capacity) was approximately 1 min and the sorbent saturation Q max was achieved after 120 min. The maximum sorption capacity was 28.7 µmol/g.

Effect of pH
Vanadium speciation in an aqueous system is a complex function of concentration, pH, redox potential (oxidation/reduction potential, i.e., E h ), and solution chemistry ( Figure 5). Vanadium exists in the +3, +4, and +5 oxidation states in natural waters. Vanadate species [V(V)] are thermodynamically stable in oxic conditions, while V(IV) is stable under suboxic conditions, and V(III) is found in anoxic environments [23].   In dilute solutions at neutral pH, the dominant species of V(V) are the phosphate-like mononuclear vanadate oxyanions H n VO 4 n−3 . The species VO 2+ dominates under acidic conditions, and VO 4 3− occurs in extremely alkaline conditions. Vanadic acid, H 3 VO 4 0 , is a minor constituent within the narrow pH range 3-4. Vanadium has a notable tendency to oligomerize, forming species containing up to 10 atoms of V [23]. The mass action, charge, and mass balance equations [24] used to generate Figures 5-7 were solved using the computer program written in Mathcad by one of the authors (A.O.). Relevant thermodynamic data were taken from literature [25]. The sorption process of V(V) ions is dependent on pH of the equilibrium solution. The influence of pH on the adsorption of V(V) was investigated in the pH range of 2.5-11.0 with a contact time of 60 min and V(V) concentration fixed at 0.5 mM in a batch system, at room temperature (298 K). The results are shown in Figure 7. It was observed that the adsorption percentage of V(V) was significantly high in the aqueous solution pH range from 3.0 to 6.0, and maximum adsorption for the V(V) was obtained at pH 5.0. When the pH was further increased from 6.0 to 8.0, the adsorption percentage decreased.
When the pH of the solution is higher than 6.0, the decreased adsorption efficiency of V(V) ions might result from the other V(V) oxidation states that form at high pH, such as V 3 O 9 3− , HVO 4 2− , and HV 2 O 7 3− , and which might affect the adsorption capacity on the m-Si-poly(GME)-deta. Thus, pH 5.0 was adopted for further studies. Similar behavior has been reported for V(V) adsorption on chitosan-zirconium (IV) [26], biomass [27], and other oxyanions on similar sorbents [17,[28][29][30][31].

Sorption Isotherms
The experimental isotherm data (Figure 8a) have been fitted with Langmuir, Freundlich, and Tempkin equations (Figure 8b-d). Isotherm parameters and regression data are presented in Table 3. Table 3. Isotherm parameters and regression data for V(V) sorption on m-Si-poly(GME)-deta at 298 K.

Isotherm Models Parameters V(V)
Langmuir Q max (mg/g) 11.23 Sorption capacity increases with higher initial concentration. Freundlich isotherms could be identified as the most suitable for modeling the equilibrium sorption behavior and the process could be described as multilayer sorption of the presence a number of vacant active sites. Besides, as the value of the parameter was calculated as n f = 1.18, it is indicative of a chemisorption process [32]. The values of the Freundlich constants and R L value suggest that the sorption process of oxyanion V(V) is favored by the magnetic sorbent m-Si-poly(GME)-deta.

Kinetic Models of Sorption
Sorption kinetic data were analyzed using the surface-reaction (PFO and PSO), Elovich, and particle diffusion-based (IPD) kinetic models to investigate the controlling mechanism of V(V) sorption by m-Si-poly(GME)-deta (Figure 9). Kinetic parameters calculated from these models are presented in Table 4, where Q t denotes the amount of sorbed metal ions at time t, Q e the amount of sorbed metal ions at equilibrium, k 1 the PFO rate constant, k 2 the PSO rate constant, b e the Elovich parameter determined by the surface coverage, k id the IPD coefficient, and C id the intercept of the IPD plot.  Table 4. PFO, PSO, Elovich and IPD kinetic parameters V(V) ion sorption using m-Si-poly(GME)-deta. The sorption of oxyanions on m-Si-PGME-deta is initially very fast. The sorption half time, t 1/2 (the time required to reach 50% of the total sorption capacity), was approximately t 1/2 < 2 min and the sorbent saturation Q max is achieved after 40 min. The PSO model is applicable only during the initial stage when the sorption process is film diffusioncontrolled because oxyanions are bound on the surface of the sorbent particles. As the saturation rate of active surface sites gradually increases, the process becomes interparticlediffusion controlled and data are better described by the IPD model.

Quantum Chemical Modeling
To understand the absorption profile of copolymer as a function of pH value, quantumchemical calculations were performed on several model systems. Model systems contain two chemical species. The first species is an absorbent (detaOH absorption center of the copolymer, Figure 10) while the second one is vanadium (V) oxyanion. Absorption is monitored in the pH region from 2 to 9, in which the three absorption centers (detaOH-H 3 3+ , detaOH-H 2 2+ , and detaOH-H + ) are in equilibrium [33].  Figure 11). However, the VO 2 + cation is the most common vanadium (V) species in solution with pH ≤ 3 ( Figure 5), the absorption of vanadium species originates from the vanadic acid (H 3 VO 4 ), which is present in appreciable quantities. H 3 VO 4 interacts with the detaOH-H 3 3+ absorption center with an interaction energy of −55.9 kcal/mol ( Table 5). The interaction between VO 2+ ion and detaOH-H 3 3+ would be repulsive due to the same charge of both species.  In the pH range from 3 to 8, the negatively charged H 2 VO 4 − ion becomes the most abundant vanadium (V) oxyanion in the solution. Due to their opposite charges, H 2 VO 4 − ion builds much stronger interactions with detaOH-H 3 3+ center (−311.6 kcal/mol) than H 3 VO 4 (−55.9 kcal/mol). The second detaOH absorption center (detaOH-H 2 2+ ) occurs at pH greater than 2, and it is the main absorption center in the pH range from 5 to 9. Due to the less positive charge, detaOH-H 2 2+ builds weaker interactions with H 2 VO 4 − ion (−203.4 kcal/mol) than detaOH-H 3 3+ center (−311.6 kcal/mol). According to calculations, it should be expected that the absorption increases up to pH = 5, and after that, it should decline. As the detaOH-H 2 2+ and detaOH-H 3 3+ species are in equilibrium up to pH = 6.5 (when the detaOH-H 3 3+ species disappears), the increase in absorption up to pH = 6 is understandable when the maximum of absorption is monitored. As detaOH-H 2 2+ completely replaces detaOH-H 3 3+ species in solution at pH greater than 7, a decrease in absorption is expected.
In order to complete the absorption profile of copolymer, another ionic species of vanadium (V) should be considered. Namely, the HVO 4 2− ion appears at pH = 7 (Figure 10), which has a higher negative charge than the H 2 Figure 11). Therefore, the interaction energy is significantly lower (−44.8 kcal/mol) than the interaction energy in detaOH-H 2 2+ . . . H 2 VO 4 − system (−203.4 kcal/mol). The transition of H ions could explain the decrease in adsorption at pH values greater than 7.
The appearance of detaOH-H + species at pH greater than 7 ( Figure 10) also contributes to a decrease in absorption. However, this contribution is probably negligible due to the equilibrium of detaOH-H 2 2+ and detaOH-H + species.

Conclusions
Two different macroporous glycidyl methacrylate and ethylene glycol dimethacrylate copolymer sorbents were synthesized: magnetic silanized non-functionalized and magnetic silanized grafted with diethylene triamine (deta). The sorbent particles are spherical as well as chemically and thermally stable. Pore size and surface functionalization were efficiently controlled during the synthesis. The sorption kinetics for V(V) oxyanions adhered to the PSO model, implying an important role of chemisorption in the sorption process, with the evident influence of the intraparticle diffusion. Major V(V) species that are adsorbed from the diluted aqueous solution are H 2 VO 4 − ions. The process is favorable at the acidic pH values, while the adsorption in alkaline solution is negligible. In acidic solutions, the electrostatic interactions between the absorption centers of copolymer and vanadium (V) oxyanions are the strongest. The absorption centers in this medium have the highest possible positive charge (+3 and +2), located on N atoms of the deta group. The increase of pH value leads to a reduction in the positive charge of the surface adsorption centers, thereby reducing the driving (electrostatic) force between the copolymer and oxyanions.