Fabrication of Silica Microspheres (HB/A@SI-MNS) for Hafnium and Zirconium Recovery from Zirconyl Leach Liquor

This work describes the synthesis of silica microspheres using sodium silicate obtained as a byproduct in the production of Egyptian Rosetta zircon concentrate. The obtained mesoporous silica microspheres were further modified with aminopropyltriethoxy silane and 2,4-Dihydroxybenzaldehyde to produce Schiff’s base silica sorbent (HB/A@Si-MNS). HB/A@Si-MNS was used for the selective extraction of hafnium from zircon mineral leach liqueur. The fabrication process and surface properties of HB/A@Si-MNS were confirmed by the means of X-ray Fluorescence (XRF), scanning electron microscop (SEM), energy depressive X-ray (EDX), Fourier-transform infrared spectroscopy (FT-IR), and elemental analysis. The uptake behavior of HB/A@Si-MNS towards Zr(IV) and Hf(IV) ions were studied under different experimental conditions. Adsorption curves indicate that the uptake of Zr(IV) and Hf(IV) on HB/A@Si-MNS is a spontaneous, endothermic monolayer system controlled by intraparticle diffusion. Elution efficiencies were found to be 94% and 98% for Zr(IV) and Hf(IV), respectively. The regenerated HB/A@SI-MNS showed uptake capacity comparable to that of fresh ones over 3 cycles. The results of the extraction of Hf(IV) than Zr(IV) from Rosetta zircon concentrate show that HB/A@SI-MNS has a preferential selectivity towards Hf(IV) than Zr(IV). Therefore, the studied material may be promising for the selective separation of Hf(IV) from Zr(IV).


Introduction
Zirconium and hafnium coexist in nature as the mineral zircon, and less commonly as baddeleyite. The zircon mineral (Zr(Hf)SiO 4 ) is one of the most stable chemical compounds due to the strong bond between zirconia and silica [1,2]. This is the result of the high coordination of bisdisphenoid ZrO 8 in a tetragonal structure with the tetrahedral SiO 4 [3,4]. Several industrial methods are used to decompose zircon, including its fusion with caustic soda at 650 • C or with soda ash at 900 • C. Together, zirconium and hafnium have potential in numerous industries, including catalyst, pigments, and alloys. One amongst the foremost necessary uses of zirconium and hafnium is within the nuclear industry as cladding for the nuclear material in reactors [5][6][7]. High corrosion resistance and the low neutron cross section absorption of nuclear grade zirconium make this metal ideally suited to be used in highly robust environments. Hafnium, on the other hand, is also used in the nuclear industry, but due to its high neutron absorption coefficient, it is used as control rods in the nuclear process to control and regulate the nuclear reactions [7]. Ultra-pure Zr metal (Hf < 100 ppm) is needed to improve the efficiency of a nuclear reactor in the nuclear industry. The difficulty in separation and the natural abundance techniques makes hafnium a scarce commodity and very expensive [8,9]. There are considerable efforts to develop new and cheaper methods to extract zirconium and hafnium ions. Adsorption on organic resins or inorganic oxides is one of the most efficient methods. Inorganic adsorbents are characterized by good selectivity, low price, rapid kinetics, mechanical stability, and no swelling [10][11][12][13][14].
Recently, some articles have reported on the preparation and application of silica from various sources [15][16][17][18][19][20]. Many chelating moieties, such as amine, sulfonate, crown ether and phosphonic were immobilized on silica surfaces through a physical or chemical interaction between silica's network and functional groups. Sodium silicate is known as the main source for the preparation of commercially available silica. Sodium silicate was obtained as a waste product through the extraction of Zr(IV) and Hf(IV) from Egyptian Rosetta zircon concentrate. The present study seeks to determine the feasibility of zircon mineral processing in terms of the preparation of zirconium/hafnium oxychlorides, mesoporous silica fabrication and modification to separate Hf(IV) and Zr(IV)ions from a zircon mineral concentrate. To achieve this goal, the waste obtained from the Egyptian Rosetta zircon concentrate was treated to obtain pure sodium silicate. The silica obtained was characterized and was further modified with Schiff base functionality. The obtained silica (HB/A@Si-MNS) was used for the selective extraction of Hf(IV) from Zr(IV) ions at different experimental conditions of pH, time, concentration, and temperature.

Zircon Mineral Concentrates Processing
A type 316-L stainless steel crucible with dimensions of 15 cm in height and 10 cm in diameter was used as the reaction vessel. A sample of 250 g of Rosetta zircon concentrate (purity 97% as shown in Table 1) was mixed with 312.5 g of sodium hydroxide and charged to the reaction vessel. The reaction vessel was fed to the electric furnace and the reaction temperature was adjusted at 650 • C for 2 h. The fusion product resulted (sodium zirconate and silicate) was washed with hot distilled water (4 times of the product weight). The residual hydrous zirconium was dried before leaching in hot HCl (2:1 w/w) to produce Zr(Hf)OCl 2 ·8H 2 O. The latter was then allowed to crystallize in ice. The formed crystals were separated from the mother liquor by decantation and the pure crystals were redissolved in dilute HCl [4].

Recycling of Silicate Waste for Fabrication of Mesoporous Silica Microspheres
A sample of mesoporous silica was prepared in a one pot synthesis assembly process using the sodium silicate obtained from the above step in the presence of CTAB as a cationic structure-directing agent. The silicate solution (25% SiO 2 , 17% NaOH) was diluted using diluted CH 3 COOH to adjust the solution pH at 8. The surfactant and an amount of acetic acid equivalent to the hydroxide content of the sodium silicate solution were mixed at 25 • C and added to the sodium silicate solution. The mixture was stirred vigorously for 2 h at 55 • C. The formed silica microspheres (pH 8; CTAB/SiO 2 molar ratio of 0.45) were stirred for about 24 h. This allowed the formation of the channels framework under basic pH conditions, where both the silica precursor and the surfactant were primarily charged species. Silica was filtered off, washed with water and ethanol to remove the surfactant, and the as-synthesized products were calcined [20].

Synthesis of HB/A@Si-MNS Meso-sorbent
Two mL of 3-Aminopropyltriethoxysilane (APTS) were dissolved in 100 mL of distilled water acidified with acetic acid (pH 4). Two grams of activated silica (dried at 150 • C for 18 h) were added to the above solution (the hydrolyzed APTS) and stirred for 2 h at room temperature. The formed monoamine modified silica (MAMS) was washed repeatedly with water, ethanol, and acetone several times. The washed material was then dried in an oven at 120 • C for another 2 h. One gram of MAMS was added to 0.5 g of 2,4-dihydroxybenzaldehyde (HB) and dissolved in 20 mL methanol. The mixture was refluxed for 72 h at 75-80 • C. The product, referred to as HB/A@Si-MNS, was then filtered off, washed, and dried.

Characterization of Fabricated Materials
A major and trace elemental analysis of the zircon sample was performed using elemental analysis and X-ray fluorescence (Philips PW X-Unique II X-ray spectrometer, Philips Co., Amsterdam, The Netherlands). The Brunauer-Emmett-Teller (BET) surface area, pore volume, and average pore diameter of the materials were measured by a Quanta chrome NOVA automated gas sorption system using nitrogen as the adsorbate. A Fourier-transform infrared spectroscopy (FT-IR) (Nexeus-Nicolite-640-MSA FTIR, Waltham, MA, USA) and Environmental Scanning Electron Microscope (ESEM, EXL 130, Philips Co., Amsterdam, The Netherlands) were used to characterize the fabricated materials.

Adsorption Experiments
Stock solutions (1000 mg L −1 ) of Zr(IV) and Hf(IV) were prepared by dissolving 3.533 and 1.8 g of zirconium oxychloride octahydrate (ZrOCl 2 ·8H 2 O) and hafnium tetrachloride (HfCl 4 ), respectively, in a liter of 2 M HCl. The concentration of Zr(IV) and Hf(IV) ions in both synthetic and real sample mixtures were measured by Inductively Coupled Plasma Mass Spectrometry (Scan-16-ICP-MS, Thermo Finnigan, San Jose, CA, USA). Uptake experiments at different time intervals were done by placing 0.1 g of HB/A@Si-MNS in a round flask containing 100 mL of the metal ion solution (100 mg L −1 ) for the required interval time. The contents of the flask were shaken on a Vibromatic-384 shaker (Selecta, Barcelona, Spain) at 300 rpm and at room temperature. The residual concentrations of Zr(IV) or Hf(IV) ions were determined spectrophotometrically [21] and were used to calculate the uptake using Equation (1) [22].
where q is the uptake (mg g −1 ), C i and C f are the initial and final concentrations of the metal ion in adsorption media (mg L −1 ), V is the volume (L) and W is the HB/A@Si-MNS weight (g). Uptake experiments under controlled pH were carried out following the above procedures while keeping the equilibrating time at 30 min for the two metal ions. Diluted solutions of HCl and NaOH were used to adjust the pH of the medium. The effect of the initial metal ion concentrations on the adsorption was carried out at different temperatures (25,30,35 and 40 • C).

Characterization of HB/A@Si-MNS
The successful fabrication of silica particles, Figure 1, from Rosetta zircon concentrates (purity 97% as shown in Table 1  The spectra of the modified silica display similar characteristic bands along with new bands. The spectrum of the modified silica (MAMS and HB/A@Si-MNS) is characterized by the appearance of two weak bands at 2361 cm −1 and 2964 cm −1 that were assigned to ν(CH 2 ) stretching. In addition, there is another new band at 1440 cm −1 , which was assigned to ν(C-N). The spectrum showed that the amino group was covalently grafted to the silica successfully, as shown in Figure 1. The BET surface area, pore volume, and average pore diameter for the unmodified and modified silica were determined and given in Figure 3A. Samples showed nitrogen adsorption-desorption isotherms typically of type IV with H1 hysteresis loop. Samples exhibited a pronounced steep condensation step for relative pressures 0.76-0.95 arising from the condensation of the nitrogen inside the mesopores, indicating the good structural order of the fabricated materials. Moreover, the high surface area (266 cm 2 g −1 ) and large pore size (0.85 cm 3 g −1 ) with high pore opening (4.685 nm) of this material are excellent advantages of the fabricated of meso-adsorbent for capture Zr(IV) from Hf(IV) ions. A decrease in the surface area (111.5 cm 2 g −1 ), pore volume (0.44 cm 3 g −1 ), and pore diameter (3.20 nm) after the functionalization of fabricated cylindrical mesoporous silica with HB provides further evidence that the organic moieties were embedded inside the mesopore of the modified silica. This may be attributed to the partial blocking obtained by the modifying silica that hinders the adsorption of nitrogen molecules [23][24][25]. The scanning electron microscope (SEM) showed uniform particles with a well-defined spherical morphology with particle sizes ranging from 250 to 300 nm ( Figure 3B). Therefore, the combined characteristics of both the SEM and N 2 adsorption reveals the appearance of HB/A@Si-MNS as a uniform and cylindrical mesopores material. On the other hand, the energy depressive X-ray (EDX) of both non functionalized silica and HB/A@Si-MNS is characterized by nitrogen and carbon lines. Elemental analyses of HB/A@Si-MNS give 28.3% (C), 3.21% (H) and 2.98% (N). The content of (=N-) group was found to be 2.14 mmol g −1 which also confirms the success of the modification process.

Effect of pH on the Batch Adsorption Process
The adsorption of Zr(IV) and Hf(IV) by HB/A@Si-MNS as a function of pH was carried out over the pH range of 1-3 and 20 min and is shown Figure 4. The highest uptake value of 19 and 46 mg g −1 was observed at pH 2.8 for Zr(IV) and Hf(IV), respectively. The aqueous chemistry of zirconium and hafnium is extensively influenced by hydrolysis and polymerization reactions. The hydrolysis of the process depends on the solution pH and can exist even in very acidic solutions (Hf(OH) 3+ is known to exist even at pH below 1). Generally, the hydrolysis and polymerization reaction of zirconium or hafnium is represented by Equation (2). Chloride ions exhibit weak complexing action on zirconium and hafnium (Equation (3)) [26][27][28][29].

Adsorption Kinetics
Studying the effect of contact time on the uptake process of Zr(IV) and Hf(IV) is essential to determine the optimum operating conditions for the full-scale batch reactor. The uptake of Zr(IV) and Hf(IV) on HB/A@Si-MNS, at 25 • C and pH value of 2.8, were shown in Figure 5a. The equilibrium was attained within 20 min for both metal ions. The adsorption data were treated according to pseudo-first (Equation (4)) and pseudo-second-order (Equation (5)) kinetic models [30,31].
where q e and q t refer to the adsorbed amount of Zr(IV) or Hf(IV) (mg g −1 ) at equilibrium and at time t (min), respectively. k 1 and k 2 are the overall rate constants of pseudo-first and pseudo-second order, respectively. Intercepts and slopes of the straight-line plots of Figure 5b,c were used to calculate the maximum theoretical capacity (q 1 and q 2 ), k 2 and k 1 as given in Table 2. Obviously, a satisfactory agreement was obtained between calculated q 2 of the pseudo-second model rather than that of the pseudo-first-order (q 1 ). This implies that the adsorption process proceeds according to the pseudo-second-order kinetics and depends upon the metal ion and active site concentrations.  In order to show the effect of the intraparticle diffusion (IPD) on the rate of the adsorption process, the plots between the amount of Zr(IV) and Hf(IV) adsorbed at different time intervals (q t ) and the square root of time (t 0.5 ) were carried out according to IPD model [22][23][24][25].
where k ip is the intraparticle diffusion rate constant (mmol g −1 min −0.5 ) and I is a constant equals intercept and is proportional to the thickness of the boundary layer. A higher value of I reflects the pronounced effect of the external film diffusion. The results gave straight lines for Zr(IV) and Hf(IV) ions, indicating that the adsorption process is controlled by IPD (Figure 4c and Table 2). The negative value of I indicates the limited effect of the boundary layer on the rate of adsorption.

Adsorption Isotherms and Thermodynamic Studies
Various adsorption models were applied to the experimental adsorption data to verify the adsorption parameters driving the adsorption reaction. The adsorption of Zr(IV) and Hf(IV) by HB/A@Si-MNS as a function of equilibrium metal ions concentration at different temperatures were shown in Figure 6. The obtained adsorption curves indicate that the uptake of Zr(IV) and Hf(IV) increased as the initial concentration increases. Also, the uptake was found to increase with increased temperature. The maximum uptake value of Zr(IV) was increased from 19 to 33 mg g −1 with increasing temperature from 25-40 • C, respectively. The maximum uptake value of Hf(IV) was increased from 46 to 64 mg g −1 with increasing temperature from 25-40 • C, respectively. The linear form of the Langmuir model is given by Equation (7) [32].
The values of the Langmuir parameters, Q max and K L , were calculated from the slopes and intercepts of the straight lines ( Figure 6) and were reported in Table 3 Langmuir adsorption isotherm can be expressed in terms of a dimensionless constant called the equilibrium parameter (R L ) [24,25]: On application adsorption data to Equation (8), R L values were found between zero and one. This implies that the adsorption of Zr(IV) and Hf(IV) on HB/A@Si-MNS are favored under the conditions used in this study.  The Freundlich model is another wide adsorption isotherm that represents the relation between solute concentration and the adsorbent surface affinity. The Freundlich isotherm is defined by (Equation (9)) [31,33].
The plot of logq e versus logC e shows that the adsorption of Zr(IV) and Hf(IV) on HB/A@Si-MNS do not follow the Freundlich isotherm as shown in Figure 7a. The correlation coefficient values, R 2 , obtained from the Freundlich isotherm were less than those obtained from the Langmuir plots, indicating that the Langmuir model is the more valid isotherm to describe the descriptions of the adsorption process. Thermodynamic parameters were calculated from the linear form of the Van't Hoff equation [34].
where (R) and T are the universal gas constant (8.314 J mol −1 K −1 ) and absolute temperature (K), respectively. The values of ∆H • and ∆S • , Table 4, were calculated from the intercepts and slopes of the straight line plots of lnK L vs. 1/T, Figure 7b. The Gibbs free energy (∆G • ) for each temperature was calculated using Equation (11).
The values of ∆G • and T∆S • were given in Table 4. The observed increase in the negative values of ∆G • with increasing temperature implies that the adsorption becomes more favorable at a higher temperature. The positive values of ∆H • indicate an endothermic adsorption process. The data given in Table 4 also shows that | ∆H • | < | − T∆S • |, this indicates that the adsorption of Zr(IV) and Hf(IV) on HB/A@Si-MNS are dominated by entropic rather than enthalpic changes.

Elution and Regeneration
Sorption/desorption experiments were carried out using the batch method. The metal ions loaded with HB/A@SI-MNS (0.1 g) was washed with distilled water and then subjected to the elution process using 1 M HNO 3 (100 mL) for the different interval time. The concentration of the released metal ions was determined by ICP-MS. After the elution process the HB/A@SI-MNS was washed repeatedly with distilled water (until pH 7), then it was reloaded with metal ions. Both Zr(IV) and Hf(IV) were eluted successfully using 1 M HNO 3 . The elution efficiencies were found to be 94% and 98% for Zr(IV) and Hf(IV), respectively. The regenerated HB/A@SI-MNS showed an uptake capacity comparable to that of fresh ones over three cycles.

Selective Separation of Hf(IV) from Zr(IV) Metal Ions
The Selectivity of HB/A@SI-MNS towards Zr(IV) and Hf(IV) was studied from the synthetic mixture as well as from real matrix samples obtained from the Rosetta zircon concentrate. The selective separation experiments were studied using the batch method. The separating factor (S) was calculated using the following equation [1,2].
The values of separation factor between Hf(IV) and Zr(IV) were found to be (2.7) and (3.5) from Rosetta zircon concentrate and a synthetic mixture, respectively. This indicates that HB/A@SI-MNS has a preferential selectivity towards Hf(IV) than Zr(IV) at low Hf(IV) concentration. Therefore, the studied material may be promising for the effective separation of Hf(IV) from Zr(IV).

Conclusions
In this work, we reported the synthesis of an adsorbent based on the functionalization of mesoporous silica surfaces by HB. HB/A@Si-MNS adsorbent showed an affinity towards Zr(IV) and Hf(IV) with maximum uptake values of 33 and 64 mg g −1 , respectively. Kinetic and thermodynamic parameters of uptake process were reported and show a pseudo-second-order, spontaneous endothermic adsorption process. The regeneration of the loaded HB/A@Si-MNS was carried out successfully using 1 M HNO 3 . The separation factor between Hf(IV) and Zr(IV) from the Rosetta zircon concentrate and synthetic mixtures showed that HB/A@Si-MNS has a preferential selectivity towards Hf(IV) than Zr(IV). The studied adsorbent may be promising for the effective separation of Hf(IV) from Zr(IV).