SOLID Application of Titanium Dioxide for Zirconium Ions Adsorption and Separation from a Multicomponent Mixture

This work studies the adsorption of zirconium ions by the mesoporous titanium dioxide and by sodium-modified mesoporous titanium dioxide. Experimental maximal adsorption values of zirconium ions by H-TiO 2 and Na-TiO 2 were found to be 64 mg/g and 109.5 mg/g respectively. This process depends on the interaction time, the equilibrium concentration of zirconium ions, and the acidity of the solution. Adsorption kinetics fit well into the diffusion kinetic model and indicate several stages of zirconium ions adsorption. Equilibrium adsorption of zirconium ions is well described by Langmuir's adsorption theory for both adsorbents. The results obtained by inductively coupled plasma mass spectrometry showed that the investigated adsorbent selectively adsorb zirconium ions from the mixture with strontium and yttrium ions in the range of solution pH = 0 - 1. The percentage of maximum extraction of zirconium ions is 86.61 % for H-TiO 2 and 94.11 % for Na-TiO 2 . This fact is extremely valuable for nuclear forensics or the determination of 90 Sr in low activity background samples.


Introduction
The relevance of studies of the adsorption of zirconium ions from aqueous solutions and its selective separation from strontium and yttrium can be explained by multiple factors. One of them is the determination of the low activity of 90 Sr in environmental samples. Radioactive 90 Sr is a very harmful beta-emitter that accumulates in bones or fish scales [1] and may, significantly harm living organisms in which it accumulates due to the long halflife. Therefore, the maximum allowable levels of 90 Sr in the environment are extremely low [2], and their analysis is very difficult to perform. 90 Sr is a pure beta-emitter. Gamma spectrometry is unable to determine the amount of 90 Sr because it does not produce gamma lines. Its daughter nuclide 90 Y also does not emit gamma rays. 90 Y β-particles have higher maximum energy than βparticles of 90 Sr (2.281 MeV versus 0.546 MeV, respectively) and completely shield strontium. The amount of 90 Sr can be approximately determined based on Vavilov-Cherenkov radiation generated by yttrium βparticles [3]. Another technique of environmental 90 Sr measurement is the separation of 90 Sr from yttrium by the oxalate precipitation and subsequent determination of 90 Sr amount based on the freshly formed daughter 90 Y. This technique is used in the determination of the amount of 90 Sr in aquatic ecosystems of the Chornobyl Exclusion Zone [1]. However, the accuracy of the oxalate technique is not very high and does not allow determining 90 Sr in low-level environmental samples. In this case, inductively coupled plasma mass spectrometry (ICP-MS) may be the only method of 90 Sr analysis. Several leading laboratories worldwide [4][5][6] use this method to analyze 90 Sr in the environment. The effectiveness of the ICP-MS method compared to other highly sensitive modern methods is presented in fig. 1.
This method of analysis is very sensitive (Fig. 1), but it diagnoses 90 Sr, 90 Y, and 90 Zr as a single peak at the mass of 90amu. The isotopes of 90 Sr, 90 Y, and 90 Zr are spectroscopic interferences and overlap each other [4][5][6]. 90 Zr, the "granddaughter" of 90 Sr, hurts the mass spectrometry of 90 Sr samples, as described in [5]. Therefore, it is necessary to separate zirconium and strontium cations.
Nuclear forensic analysis is another application of zirconium adsorption and chemical separation from 90 Sr and 90 Y in combination with the ICP-MS analysis [7]. The isotopic ratio, measured using ICP-MS, serves as a radio chronometer and can be used to determine the date of fabrication of a 90 Sr -90 Y radioactive source. The half-life of 90 Y is much shorter (64 hours) than that of 90 Sr (28.8 years). On elapse of time equal to several half-lives of yttrium, the amount of 90 Y becomes constant [8]. Therefore, for nuclear forensics, the important parameters are the number of 90 Sr and its "granddaughter" 90 Zr nuclei, i. e., the 90 Zr/ 90 Sr ratio, thus methods of chemical separation of these elements become the main direction of scientific research in this field.
Adsorbents based on mesoporous TiO2 showed high adsorption activity toward many cations, such as strontium, barium, zinc, cobalt, and yttrium [9][10][11][12][13][14][15]. The uniqueness of the properties of adsorbents based on mesoporous TiO2 is due to the developed surface area and a large number of adsorption centers on it in combination with the predominantly crystalline structure of TiO2. These adsorbents are resistant to aggressive environments and thermally stable. The resistance of mesoporous TiO2 to the action of acids is a very valuable characteristic, as it greatly simplifies the subsequent mass spectrometry of the studied solutions. The introduction of sodium cations into the structure of mesoporous TiO2 leads to functional changes in the surface (for example, an increase in the number of active adsorption sites or shift of the point of zero charges), increasing the adsorption capacity of Na-TiO2 compared to H-TiO2 relative to heavy metal cations [12]. While the general laws of strontium ion adsorption process by adsorbents based on titanium dioxide are well studied, no surveys have been conducted of the adsorption of zirconium ions by these adsorbents yet.
This work will study the adsorption of zirconium ions by H-TiO2 and Na-TiO2, as well as the possibility of separation of 90 Sr and 90 Zr isotopes by investigated adsorbents for subsequent analysis using inductively coupled plasma mass spectrometry.

Synthesis of H-TiO2 and Na-TiO2 adsorbents
Experimental samples of anatase modification titanium dioxide with protonated and sodium-modified surface were obtained by sol-gel method, using a solution of titanium aqua complex [Ti(OH2)6] 3+ •3Clas a precursor [12]. The 0,1M solution of titanium precursor was heated at 60 ℃ for 60 minutes to prepare titanium dioxide with a protonated surface (H-TiO2). Globular particles of TiO2 with a diameter of 4 -5 nm were formed in the reaction medium, due to the hydrolysis of the precursor, and condensation of the molecules Ti(OH)4•2H2O. The 10 % NaOH solution was added dropwise to the dispersion to deoxidize the resulting product. At pH ~ 5.3 the dispersion thickened rapidly because of gel formation. The hydrogel with a pH of ~ 7.0 was washed with distilled water from the adsorbed impurities of Na + , Clions. Then, it was dried for 4 hours, at a temperature of 140℃.
Titanium dioxide with a sodium surface was obtained by contacting globular TiO2 particles with 10 % NaOH solution. The TiO2 dispersion was kept in an alkali medium (pH ~ 12) for 3 hours at room temperature, then the dispersion was washed with distilled water. Washing and decantation of the dispersion were repeated at least 5 times, and after establishing the pH ~ 7, it was dried at a temperature of 140 ℃. The main textural characteristics of the adsorbents are shown in Table 1.

Batch adsorption studies
Recommendations for the separation of elements can be made after studying the general patterns of the process of adsorption of the studied elements. Studies of the adsorption capacity of H-TiO2 and Na-TiO2 toward zirconium cations were performed under batch conditions. The weight of the adsorbent was 50 mg, and the volume of the ZrOCl2 aqueous solution was 5 ml. The influence of time interaction, the equilibrium concentration of the zirconium ions, and the acidity of the solution was investigated. Initial and residual concentrations of Table 1 The textural characteristics of investigated samples, according to [12] Sample SBET (m 2 g -1 ) Smicro (m 2 g -1 ) Smeso, (m 2 g -1 ) Vp, (cm 3 g -1 ) Vmicro (cm 3 g -1 ) Vmeso (cm 3  zirconium ions were determined using direct complexometric titration in a strongly acidic medium with Xylenol Orange as an indicator [16]. Adsorption values were calculated using (eq.1) as follows: where qeis the amount of adsorbate uptake, mg/g; Co and Ceare initial and residual concentrations of adsorbate, mg/L; V -is solution volume, L; m -is mass of adsorbent, g [17].
The influence of solution acidity on zirconium cations adsorption by investigated adsorbents was studied under the same conditions, however, the volume of the solution was 10 ml (5 ml of ZrOCl2 solution and 5 ml of the medium) and the mass of the adsorbent remained unchanged.
Separation factor αSr, Zr of strontium and zirconium was calculated according to [7]: where Intensity in [count per second] was experimentally measured value by ICP-MS of corresponding isotopes after adsorption. Nonlinear approximation of experimental results was processed in the "Solver add-in" to Microsoft Office Excel. The chi-squared test is highly recommended for the nonlinear method to confirm the best fitting isotherm for a given adsorption system. High χ 2 indicates high divergence between the experiment and the model. The chi-square value χ 2 and coefficient of determination R 2 were calculated using equations (5) and (6). In general, the recommendations from the publication [20] were used in the analysis of experimental adsorption results.

ICP-MS analysis of separation of investigated cations
The study of the possibility of strontium, yttrium, and zirconium ions separation using the investigated adsorbents was performed under batch conditions. The 88 Sr, 89 Y, 90,91 Zr isotopes were analyzed by the inductively coupled plasma mass spectrometer (ICP-MS) [4][5][6] 'Element 2' with argon plasma, located in the Nuclear Forensics laboratory of KINR, Kyiv. The study was performed with a mixture of the high purity standards solutions (High Purity Standards, USA) of the corresponding elements.
The results were measured in the number of [CPS] pulses "count per second" relative to m/z.
As has been said above, 90 Y does not play a significant role in determining the date of fabrication of a 90 Sr -90 Y radioactive source. This is due to the fact, that the half-life of 90 Y is much shorter (64 hours) than the half-life of 90 Sr (28.8 years) and its amount throughout the life of 90 Sr remains constant. This statement can be illustrated by formulas (7) -(9): Where λ 2 , − Decay constant of 90 Y; λ 1 , − decay constant of 90 Sr. The number of daughter nuclei is related to the number of parent nuclei by the equation (7): After several half-lives, an equilibrium establishes and the amount of 90Y can be calculated: Stable 89 Y, in contrast to 90 Y, plays a significant role, because it was very often used as a carrier for radioactive 90 Sr and 90 Y. This was because, the separation of 90 Sr from a mixture of fission radionuclides, during fabrication of a 90 Sr -90 Y radioactive source, was carried out by the reaction of precipitation of strontium and yttrium with oxalates [1]. Stable 89 Y in the result 90 Sr-90 Y source may be at a high concentration. It can affect the selective adsorption of Sr 2+ or Zr 4+ because the cations Y 3+ are also adsorbed on the surface of titanium dioxide. Therefore, a mixture of isotopes of 88 Sr, 90 Zr, 91 Zr, and 89 Y was subjected to the analysis by ICP-MS. A great advantage of ICP-MS is its multi-element capability, which allows multiple elements to be measured simultaneously in a single analysis. However, interferences need to be controlled. Internal standardization is usually employed to correct for changes in instrument operating conditions and sample-specific matrix effects. According to recommendations, which were given in publication [5], the internal standard will have a similar mass and ionization potential to the analyte ( 88 Sr, 89 Y, 90 Zr, 91 Zr). Therefore, 103 Rh was chosen, as an internal standard. External calibration of the ICP-MS was performed using calibration Standard A, containing known concentrations of the elements. The error due to the formation of polyatomic ions 89 Y 1 H in plasma was determined using Standard 1. Standard 1 is a mixture of about 10 ng 88 Sr and 89 Y. Standard solution 2 (Standard 2) was a mixture of 88 Sr, 89 Y, 90 Zr, 91 Zr, which was divided into 5 equal parts, four of them were mixed with adsorbents, and the fifth was left as a control solution. In addition, Standard 3 was pure 2 % HNO3 'Optima'.
Adsorption and separation of strontium, yttrium, and zirconium ions were performed in batch conditions for an hour. Mass of adsorbents was 100 mg, volume of solutions -10 ml. After adsorption, the adsorbents were separated from the solution by filtration. The slow-filtering paper with dense narrow pores for finest deposits was selected for this purpose. Selective adsorption of zirconium and the possibility of separation of zirconium from strontium and yttrium were evaluated by analyzing the isotopes 88 Sr, 89 Y, 90 Zr, 91 Zr, and 103 Rh by ICP-MS at least three times.

Adsorption of zirconium ions by Н-ТіО2 and Na-ТіО2
The results of the dependence of the zirconium adsorption values by Н-ТіО2 and Na-ТіО2 on agitation time are shown in Fig. 2. The results, which are shown in Fig. 3 and Table 2, indicate that the diffusion model best describes the process of adsorption of zirconium ions by the investigated adsorbents. According to the results of the analysis, the zirconium ions adsorption by H-TiO2 and Na-TiO2 occurs in several stages: diffusion from solution to the surface of the adsorbents (Step 1 'bulk transport'); diffusion on the adsorbent's surface (Step 2 'film diffusion') and diffusion deep into the pores of the adsorbent (Step 3 'intraparticle diffusion'). The highest diffusion coefficients (kipd) were Step 3 and has a value of 19 mg/g; 20.7 mg/g, and 52 mg/g, respectively.

Equilibrium adsorption of zirconium ions by H-TiO2 and Na-TiO2 and influence of solution acidity on the adsorption process
The results of equilibrium studies of zirconium ions adsorption by H-TiO2 and Na-TiO2 are shown in Fig. 4 and Fig. 5. The application of the Langmuir and Freundlich theories to the experimental results is given in Table 3.
The experimental isotherms of zirconium ions adsorption by the studied adsorbents are better described by Langmuir's theory than by Freundlich's theory for both adsorbents. However, the divergence between the experimental results and calculated by Langmuir theory (χ 2 ) for Na-TiO2 are higher. The main differences between the experimental values and the theoretically calculated ones for zirconium ions adsorption by the Na-TiO2 are in the region of low equilibrium concentrations of the adsorbate. This can be explained by the very active surface of the adsorbent. In our opinion, this is the reason for rather high values of χ 2 . The values of the maximum adsorption of zirconium ions calculated according to Langmuir theory are in good agreement with the experimentally determined ones (Table 3) and are lower than the values of adsorption of strontium ions by the H-TiO2 and Na-TiO2 [12]. The maximum adsorption values   of the strontium ions in the neutral medium are 78.4 mg · g -1 by H-TiO2 and 208.4 mg · g -1 by Na-TiO2 and the maximum adsorption of zirconium ions are 61.4 mg · g -1 by H-TiO2 and 109.5 mg · g -1 by Na-TiO2. The reason for the lower adsorption of zirconium ions compared to strontium ions in aqueous solutions with a neutral pH may be the high tendency of zirconium cations to hydrolysis [24,25]. The plot, shown in Fig. 6 indicates that the weakest adsorption of zirconium ions occurs in a neutral medium. The tendency of zirconium cations to hydrolyze [20,21] determines the large size of hydrolyzed zirconium ions and their low mobility in solution and, accordingly, insignificant adsorption.
The results presented in Fig. 6 are in good agreement with the literature data on the mobility and size of zirconium ions in solution, shown in Table 4. For comparison, the table shows the values of the mobility of strontium cations, as well. The concentration of the corresponding ions was 10 -5 mole/L; studies were performed using HCl, NH4OH, and 100-fold excess of electrolyte NH4Cl.

Selective adsorption of zirconium ions from the mixture
Schematically, the experiment on the adsorption separation of strontium, yttrium, and zirconium ions is shown in Fig. 7.
The results presented in Tables 5 and 5 (a) indicate the selective adsorption of zirconium ions from a mixture of strontium and yttrium ions. This process is illustrated in Figs. 8 (a) and (b). The results in Table 5 (a) indicate that the amount of strontium, yttrium and internal rhodium ions was not changed after interaction with adsorbents. This fact evidences the absence of adsorption of these elements under experimental conditions. Analysis by ICP-MS showed that the error due to the formation of 89 Y 1 H polyatomic ions is 1210 counts per second at a mass of 90amu. In addition, the number of contaminants of 88 Sr and 90 Zr in the used solvent was recorded. If we take into account all these contributors to the mass peak of 90amu and subtract the background number of counts from 90 Zr, the percentage of extracted 90 Zr will be 86.61 % for H-TiO2 and 94.11 % for Na-TiO2. Figures 8 (a) and (b) illustrate the selective adsorption of zirconium ions by H-TiO2 and Na-TiO2 from nitric acid with HF micro-impurities in the absence of adsorption of strontium and yttrium ions. The amount of strontium ions even slightly increases due to the 88 Sr contaminated micro impurities ( Fig. 8 (a)).
Selective zirconium ion adsorption is observed in the range of solutions acidity pH = 0 -1. This range of solution acidities is below the pHpzc of the studied adsorbents ( Table 1). The number of adsorption centers (≡ТіО-) able to bind cations is slightly reduced and the surface of H-TiO2 or Na-TiO2 has an overall positive charge in the experimental conditions, although some areas are still able to bind cations on the adsorbent surface. The overall positive surface charge of H-TiO2 or Na-TiO2 becomes a Coulomb barrier for the adsorption of relatively large, positively charged ions of divalent strontium and trivalent yttrium. Therefore, strontium and yttrium ions were not adsorbed by these adsorbents in the 0.32M HNO3 medium. It should be noted, that at pH values closer to a   Fig. 8 (a). ICP-MS spectra of a solution containing about 10 ng of strontium, yttrium, and zirconium ions, before and after adsorption by H-TiO2. Fig. 8 (b). ICP-MS spectra of a solution containing about 10 ng of strontium, yttrium, and zirconium ions, before and after adsorption by Na-TiO2.
neutral or alkaline environment, the adsorption of strontium ions by these adsorbents is extremely intense [9][10][11][12][13][14]. Ionic radius of Zr 4+ is smaller than the corresponding value of Sr 2+ or Y 3+ (0.84 Å compared to 1.02 Å for Y 3+ and 1.26 Å for Sr 2+ ions) [26]. Zirconium loses its hydrate shell under such acidity of the solution. This fact allows zirconium ions to neglect electrostatic repulsion and interact with residual cation-bonding adsorption cites of the adsorbents. On the other hand, the mobility of zirconium ions is much lower than the mobility of strontium ions [25]. Therefore, after approaching the surface of TiO2, zirconium generally remains in the adsorbed state.
The obtained results are promising for the separation of 90 Sr and 90 Zr isotopes. The absence of adsorption of strontium and yttrium ions under such extreme solution acidity indicates that zirconium ions can be selectively adsorbed even with an excess of cations of alkaline earth elements and elements of the YRE group. The resistance of TiO2 to the acidic media allows direct one-stage separation of zirconium ions without additional procedures for the separation of alkaline earth elements and lanthanides, which can shift the initial ratio of 90 Sr/ 90 Zr [27][28][29]. This fact is extremely valuable for nuclear forensics or the determination of 90 Sr in low activity background samples.

Conclusion
The adsorption of zirconium ions by titanium dioxide has been studied. The dependence of this process on the agitation time, the equilibrium concentration of the zirconium ions, and the acidity of the solution have been established.
The kinetics of zirconium ions adsorption by H-TiO2 and Na-TiO2 is stepwise and fitting well with the diffusion kinetic model.
Experimental maximum values of adsorption of zirconium ions by the studied adsorbents are 61.4 mg·g -1 for H-TiO2 and 109.5 mg·g -1 for Na-TiO2. These values are in good agreement with the calculated values of the maximum adsorption according to Langmuir's theory.