Hydrothermal synthesis of titanate nanotubes from TiO₂ nanorods prepared via a molten salt flux method as an effective adsorbent for strontium ion recovery

Abstract

Hydrated titanate nanotubes (TNTs) were hydrothermally synthesized at 160 °C over reaction times of 6–72 h from molten salt TiO₂ nanorods (NRs). Most of the TiO₂ NRs were transformed into tubular structure within 24–72 h. The samples synthesized over short reaction times (6–24 h) formed admixtures of TNT and untransformed TiO₂ NR residues. Strontium ion (Sr²⁺) adsorption by the as-prepared samples was quantified. The surface area of the TNTs increased the Sr²⁺ ion adsorption relative to that of the TiO₂ NRs. The mechanism underlying Sr²⁺ adsorption relied on an ion exchange reaction between Sr²⁺ ions in the stock solution and Na⁺ ions in an interlayer of the TNTs. TEM, EDAX, and XAFS analysis confirmed that Sr²⁺ adsorption and Na⁺ release occurred at the interlayer of the TNT-2D. The maximum adsorption capacity of the TNTs was calculated using the Langmuir equation. TNT (TNT-2D) sample synthesized over 48 h displayed the highest adsorption capacity (113.6 mg g⁻¹), with a Sr²⁺ uptake having a nearly 99% efficiency.

---

Introduction

Over the past few decades, titanate nanostructures have received significant attention due to their physicochemical properties, including their nanostructures, porosities, specific surface areas, and economical one-step synthesis. These physicochemical properties render titanate nanotubes (TNTs) useful for diverse applications in, for example, Li-ion batteries,^1–3 photo-catalysts,^4,5 hydrogen storage devices,^6,7 solar batteries,^8,9 ion exchange membranes,^10–14 and gas sensors.¹⁵ Additionally, TNTs may be used as adsorbents to remove and recover¹⁶ both organic and inorganic materials^17,18 such as hazardous metal ions (e.g., Pb²⁺ and Cd²⁺), heavy metal cations (e.g., Ni²⁺, Cr²⁺, Cu²⁺, and Sr²⁺), dyes, and radioactive isotopes.^19,20

Seawater contains abundant mineral resources, including strontium, a rare earth metal used widely in CRT screens, ceramic ferrite magnets, glass, small engines, and fireworks. The extraction of Sr²⁺ from seawater is effective, beneficial, and eco-friendly compared to conventional land mining of mineral ores. Sr²⁺ recovery from seawater has been extensively explored due to the abundance of Sr²⁺ in seawater and the demands of industry. The Sr²⁺ concentration in seawater is ∼7 mg L⁻¹.²¹ Despite extensive research into heavy and toxic metal removal, few studies have focused on Sr²⁺ recovery. Moon et al. and Ahmadi et al. reported the removal of radioactive Sr²⁺ using PAN–potassium titanate composite ion-exchanger beads and MnO₂–ZrO₂ nanocomposite material-type sorbents, respectively.^16,22 Jeong et al. investigated Sr²⁺ extraction from seawater using a chemical precipitation and membrane filtration method.²³ Zhang et al. group has synthesized highly efficient and selective ion exchanged titanate nanosheets with large amount of interlayer water and large surface area.²⁴ They discussed about the selectivity and acid-induced phase regeneration of titanates; valence, hardness and radius of cations are the main factors that affect the selectivity of ion exchange. The cation with higher valence, lower hardness, and smaller radius is more preferred to ion exchange. Recently, Hong et al. successfully developed alginate microspheres as a low-cost adsorbent for Sr²⁺(II) recovery from seawater.²⁵ Basnet et al. group has synthesized porous, amorphous tungsten trioxide submicrometer rods for adsorption of toxic dye molecules. Large specific surface area and strong electrostatic interaction of negative surface charged WO₃ makes promising practical remediate of cationic MB pollutants.²⁶

TNTs are a promising candidate adsorbent because of their high surface area (abundant hydroxyl groups) and ion exchange ability.^27–29 Generally, commercial P25 nanoparticles and TiO₂ nano-sols are used as titanate precursors to synthesize TNTs. In our previous study we have synthesized titanate nanotubes from P25 hydrothermal process follows a 3D → 2D → 1D mechanism. Titanate nanotubes with small sized and large specific surface area were tested for adsorption and desorption of Sr²⁺ ions in real seawater and co-existing matrix ions medium and evaluate the better strontium efficiency of 97 mg g⁻¹ because of their large surface area.³⁰ In this study, for the first time, molten salt rutile TiO₂ NRs were used as precursors for TNTs. Molten salt flux synthesis was accomplished using a mixture of highly reactive molten salts.³¹ Liu et al. reported the production of large quantities of high-single-crystalline TiO₂ nanowires using the molten salt method.³² The molten salt flux method significantly improved the one-dimension nanowire yield while maintaining the high crystal quality.

We investigated, for the first time, the formation of titanate nanotubes via a hydrothermal route from TiO₂ NRs, and we described Sr²⁺ adsorption from a Sr²⁺ stock solution using TNTs synthesized over different hydrothermal reaction times. We assessed the correlation between the surface area and Sr²⁺ uptake by the TNT samples prepared over different hydrothermal reaction times. The Langmuir isotherm model for monolayer adsorption and a pseudo-second-order kinetics model for chemisorption fit well to the Sr²⁺ adsorption curves obtained from TNT-2D. The effects of the Sr²⁺ concentration on the Sr²⁺ uptake from the TNT-2D adsorbents were evaluated, and the maximum adsorption capacities of the TNT samples were calculated using the Langmuir equation.

Results and discussion

In a typical synthesis, the molten-salt flux method was used to produce a large quantity of high-quality single crystalline TiO₂ nanorods. In this synthesis route, commercial P25 (Degussa) was used as a TiO₂ precursor. TiO₂ nanoparticles (P25) were mixed with NaCl and Na₂HPO₄ at a 1 : 4 : 1 ratio and heated above the melting point of the salt mixture to achieve a high crystallinity with 90% yield (ESI, Table S1†). FE-SEM images confirmed that the one-dimension TiO₂ nanorods exhibited a tetragonal morphology (Fig. 2a). The average diameter of the TiO₂ nanorods was 100–200 nm, and the average length was 1.5–5 μm. The mechanism underlying the formation of the titanate nanotubes from molten salt TiO₂ nanorods (depending on the HT reaction time period) is illustrated in Fig. 1. During the hydrothermal synthesis, the outermost surface of a TiO₂ NR directly contacted the alkali (10 M NaOH), in which TiO₂ atoms were dissolved. The dissolved TiO₂ atoms tended to cluster in the alkali solution under constant thermal excitation to form intermediate nanosheet-like layered structures. The intermediate nanosheet layers were unstable but symmetric on both sides due to the thermochemical environment.³³ The high reaction temperature and hydrogen deficiency (due to water molecule formation) on the top surface of the interlayer produced unequal surface tensions on opposite sides of the nanosheets, which tended to bend the surface and form a folded tube-like structure.³⁴

Fig. 1 Formation of TNTs from a molten salt NR solution. TNTs were synthesized using an aqueous alkali solution (10 M NaOH) and a hydrothermal reaction at 160 °C over (a) 6 h, (b) 12 h, (c) 24 h (1D), (d) 48 h (2D) and (e) 72 h (3D).

The TNTs were synthesized over reaction times of 24 (1D), 48 (2D), and 72 (3D) h. FE-SEM images confirmed the folding mechanism by which the TNTs were formed at various synthesis times (ESI, Fig. S1†). The changes in morphology and crystal structure during the preparation process were assessed by synthesizing intermediate materials over short reaction times (6 and 12 h, (b) and (c) in Fig. 2 and the ESI, Fig. S1†). Compared to the TNT sample synthesized over 12 h, the TNT-6 h sample formed a short tubular structure, and only a small portion of TiO₂ NRs were transformed (ESI, Fig. S2†). FE-SEM image of the TNT-1D sample revealed a tubular morphology of ca. 18–25 nm diameter, whereas the TNT-3D sample exhibited a solid rod-like structure ca. 70–150 nm in diameter (ESI, Fig. S2c and S1e, respectively†). Solid rod-like nanostructures formed in the TNT-3D samples due to continuous dehydration of the intermediate nanosheets. As discussed earlier, the hydrogen deficiency due to water molecule formation created unequal surface tensions on either side of the nanosheets, resulting in bending of the top layer and the formation of tight multi-layered rolls with huge diameters.³⁵

Fig. 2 FE-SEM images of (a) the TiO₂ nanorods (NRs), and TNTs prepared by hydrothermal reactions over (b) 6 h, (c) 12 h, (d) 24 h (1D), (e) 48 h (2D), and (f) 72 h (3D).

The XRD patterns collected from the TNTs and nanorods are shown in Fig. 3. The major peaks corresponding to the molten salt TiO₂ NRs arose from the tetragonal rutile titanium oxide (JCPDS 89-4202). The XRD pattern obtained from TNT-2D revealed peaks characteristic of hydrogen hydrate titanate at (2θ) 10.7°, 24°, 28°, and 48° (JCPDS no. 72-0148), which were differentiated from the titanate and rutile peaks (NRs) of the TNT-1D. No significant phase transformations (from rutile to the titanate phase) occurred over the 6 and 12 h reaction times. After hydrothermal synthesis over 72 h (i.e., 160-3D), most of the tubular structures formed tight scrolls and were transformed into solid cylindrical nanorods. The XRD pattern of TNT-3D revealed that the diffraction peaks at 9.7° and 28.5° were shifted to higher angles, new peaks appeared with increased intensity, indicating that the crystallinity of the TNT-3D sample improved for longer hydrothermal reaction times.^36,37

Fig. 3 XRD patterns obtained from the (a) TiO₂ nanorods (NRs) and TNTs prepared by hydrothermal reactions over (b) 6 h, (c) 12 h, (d) 24 h, (e) 48 h, and (f) 72 h at 160 °C. The corresponding phase peaks (R, rutile and T, titanate) are shown.

Fig. 4 shows the pore volume distribution (BJH desorption) and nitrogen adsorption–desorption isotherms of the TNT samples synthesized over different hydrothermal times (labelled 1D, 2D, and 3D in Fig. 2). The BJH desorption curve indicated the pore size distribution of TNT samples (Fig. 4A). The small peak at 3 nm corresponded to the inner diameter of TNTs, whereas the broad peak at ca. 23.6 nm originated from the intra-pore between the agglomerated nanotubes. The N₂ adsorption–desorption hysteresis curve was characteristic of a mesoporous material (2–50 nm, Fig. 4B). The corresponding hysteresis loops were a function of the HT reaction time at a relatively high pressure range of 0.8–1.0. The TNT-3D sample displayed a relatively narrow and short hysteresis loop (0.85–1.0), whereas the hysteresis loops of the TNT-1D and TNT-2D samples were longer and broader. This difference arose from the pore size distribution of the TiO₂ NR and TNT samples. The specific surface area and Sr²⁺ adsorption analyses results are presented in Fig. 5. Among the as-synthesized samples, the TNT-1D sample displayed the largest surface area (235 m² g⁻¹). The surface area of the TNT samples decreased as the HT reaction time increased. The pore size distribution and BJH adsorption–desorption curves obtained from the TiO₂ NRs (9.97 m² g⁻¹) are shown in the ESI, Fig. S3.†

Fig. 4 (A) BJH pore size distribution and (B) nitrogen adsorption–desorption isotherm collected from 1D, 2D, and 3D TNTs.

Fig. 5 Correlation curves obtained from Sr²⁺ adsorption curves and specific BET surface area of the TiO₂ (P25), TiO₂ NRs, and as-prepared samples ([Sr]₀ = 10 and 100 mg L⁻¹, [samples] = 1 g L⁻¹, contact time = 30 min).

We tested the synthesized nanomaterials as adsorbents for Sr²⁺ ion recovery from water. Twenty milligram samples were used to adsorb Sr²⁺ from 10 and 100 mg L⁻¹ of Sr²⁺ aqueous solutions at room temperature (Fig. 5). Reference TiO₂ (P25) and TiO₂ NR samples showed negligible Sr²⁺ adsorption. At 10 mg L⁻¹, all TNT samples, regardless of the hydrothermal reaction duration, exhibited a high adsorption capacity of ∼10 mg g⁻¹, indicating adsorption of most Sr²⁺ ions. The adsorption of 100 mg L⁻¹ Sr²⁺ differed markedly among the TNTs. TNT-2D showed the highest Sr²⁺ uptake of 100 mg g⁻¹, slightly higher than that of TNT-1D. Moreover, TNT-3D had a Sr²⁺ adsorption capacity of 40 mg g⁻¹, considerably lower than those of TNT-1D and 2D. Indeed, the Sr²⁺ adsorption capacity increased with increasing surface area. The fact that TNT-1D, which had the largest surface area, did not exhibit the highest Sr²⁺ uptake may have been due to the greater quantity of untransformed NR residue present in the sample, as these components were nearly inactive with respect to Sr²⁺ adsorption (Fig. S2c†). The significantly lower surface area of the TNT-3D sample resulted in inefficient Sr²⁺ adsorption. The surface area plays an important role in ion-exchange adsorption. In our previous work, we synthesized the small sized titanate nanotubes (TiNT-S) from P-25 (Degussa) and evaluated the better strontium efficiency of 97 mg g⁻¹ because of their large surface area.³⁰ The large surface area of the TiNT samples arising from the tubular structure makes them very promising candidates for the adsorption of target elements. In another literature, Kasap et al.,³⁸ group also synthesized titanate nanotubes with ca. 165 m² g⁻¹ (4 nm inner and 12 nm outer diameters) of specific surface area via sonochemical method and showed the maximum adsorption capacity of 66.72 mg g⁻¹ which was the lower result as comparing with our previous³⁰ and current reported work. The above mentioned literatures explained the importance of the specific surface area of TNT adsorbents over ion adsorption process. In our study we found out the effect of synthesis reaction time. Thus, the mesoporous tubular TNT-2D sample represented an optimal Sr²⁺ adsorbent due to its high porosity and large specific surface area (189 m² g⁻¹).

Sr²⁺ adsorption occurred rapidly, as equilibrium was reached within 10 min (Fig. 6). Indeed, >90% of Sr²⁺ was removed by TNT-2D within 5 min, indicating almost instantaneously rapid adsorption and equilibration after mixing. These results suggested that the reaction occurred on the surface of the adsorbents. This adsorption kinetics fit well to a pseudo-second-order kinetic model (eqn (1)):

where, q_t and q_e are the quantities of ions adsorbed (mg g⁻¹) at time t and at equilibrium time (min), respectively, and k₂ is the pseudo-second-order rate constant [g (mg min)⁻¹]. A linear fit, calculated using a pseudo-second-order kinetic equation, yielded the obtained kinetic parameters summarized in Fig. 6. The correlation coefficient (R²) was 0.9983, and the q_e value of 97.13 mg g⁻¹, calculated from the model fit, was very close to that derived experimentally. This suggested that the reaction occurred on the surfaces of the TNT-2D sample, and the overall rate of adsorption was determined by chemisorption kinetics rather than mass transport.³⁹

Fig. 6 Time profile curve associated with Sr²⁺ adsorption onto the TNT-2D samples ([Sr]₀ = 100 mg L⁻¹, [TNTs-2D] = 1 g L⁻¹).

The adsorption capacity of TNT-2D was evaluated using a 10–400 mg L⁻¹, Sr²⁺ solution with a fixed contact time of 30 min (as necessitated by the fast reaction kinetics). Sr²⁺ uptake increased proportionally with the Sr²⁺ concentration and plateaued above 100 mg L⁻¹ (Fig. 7). A variety of equilibrium models could be used to determine the adsorption behaviour of an adsorbent. The equilibrium data were obtained from Langmuir and Freundlich model fits to the Sr²⁺ adsorption profiles. The Langmuir isotherm model relates to monolayer sorption onto the homogeneous surface of an adsorbent, as follows (eqn (2)):

where q_e (mg g⁻¹) and C_e (mg L⁻¹) are the equilibrium adsorption capacity and equilibrium Sr²⁺ concentration, respectively. q_m (mg g⁻¹) is the maximum monolayer adsorption capacity, and K_L is the Langmuir constant related to the free energy of adsorption. The Freundlich isotherm model is considered to be a semi-empirical equation, as follows (eqn (3)):

q_e = K_fC_e^1/n, (3)

where K_f (mg g⁻¹) is the Freundlich constant related to the adsorption capacity of the adsorbent, and n is the heterogeneity factor indicative of the adsorption intensity of the adsorbent. The inset of Fig. 7 shows that Sr²⁺ adsorption could be fit to the Langmuir model with a correlation coefficient (R²) of 0.9976, whereas the R² value of the Freundlich model fit was 0.642. These results suggested that Sr²⁺ adsorption proceeded to form monolayer coverage. The entire surface of the TNT-2D sample could be modelled as offering a single adsorption capacity, and no detrimental interactions were observed among the adsorbed Sr²⁺ ions. As the ion-exchange mechanism of the TNTs involved Na ions located in the interlayers of the layered lattice structure, it was reasonable that the entire surface of the TNTs provided a homogeneous adsorption capacity. The maximum monolayer adsorption capacity (q_m) of the TNT-2D was calculated to be 113.6 mg g⁻¹. In the representative test, the effect of Ca²⁺ cation on the adsorption capacity of TNT-2D was investigated. Fig. S4 (ESI†) shows the adsorption uptake of strontium ion from the mixed cations stock solution and particular error bar mentioned consecutive batch uptake results. The Sr stock concentration was fixed at 10 mg L⁻¹ over varying concentrations of Ca stock solutions. The adsorption curve of strontium ion has declined exponentially over the increment of Ca concentration. The results are shown clearly Ca²⁺ significantly hindered Sr²⁺ sorption performance because of the high concentration of Ca ions than that of Sr ions. Generally, it is well known that Sr²⁺ and Ca²⁺ have similar chemical behaviours. The hardness values and ionic radii of Sr²⁺ and Ca²⁺ are known to be 16.3 (hard) and 19.7 (hard) and 1 and 1.18, respectively.²⁴ That is, as for Sr²⁺ and Ca²⁺ with the same valence, Sr²⁺ has priority over Ca²⁺ in the aspect of hardness, whereas Ca²⁺ is favourable to be exchanged from the perspective of ionic radius. Consequently, the significant competition between Sr²⁺ and Ca²⁺ in the ion-exchange reaction could be ascribed to the very similar chemical properties of both cations. The equimolar analysis has been performed additionally to check the adsorption ability of TNT-2D in both mixed cations solution (Fig. S4,† inset). The equimolar condition of both the cations Ca²⁺ and Sr²⁺ have obtain by mixing equal molar concentration (mmol g⁻¹) of each cations into DI water. The ICP data for equimolar analysis has demonstrated linear uptakes of both the anions Ca²⁺ and Sr²⁺ for low concentration (mmol g⁻¹) but further increment in concentration shows a sudden decline of Sr²⁺ uptake which is not mentioned in. Fig. S4† (inset) confirms the less adsorption capability of Sr²⁺ ions uptake in the presence of equimolar concentration of both Sr²⁺ and Ca²⁺ cations.

Fig. 7 Sr²⁺ uptake at equilibrium onto the TNT-2D sample as a function of the initial concentration. The inset shows the corresponding Langmuir plot ([TNTs-2D] = 1 g L⁻¹, contact time = 30 min).

The solution pH varies the surface charge of metal oxide adsorbents in the suspension, which directly influences the electrostatic interaction between adsorbates and adsorbents, resulting in the change of adsorption behaviour. In our previous study,³⁰ the isoelectric point (IEP) was measured to be at a pH of approximately 3.5. Because strontium ions exist in water as Sr²⁺ (divalent cations), it is easily expected that a negatively charged surface facilitates the adsorption of Sr²⁺. The variation of Sr²⁺ uptake at different pH values was well correlated to that of the surface charge. In addition, H⁺ is the smallest cation and thus may possess the highest priority in the ion exchange. This means that H⁺ can compete with Sr²⁺ for the ion exchange on TNT along with affecting electrostatic behaviour.²⁴

The calcination effect on the Sr²⁺ adsorption by the hydrate TNT-2D was tested by heating the adsorbent material to 300–500 °C. The adsorption kinetics associated with the Sr²⁺ uptake by the samples calcined at different temperatures were determined under equilibrium Sr²⁺ uptake conditions as a function of the stirring time (Fig. S5†). The calcined samples exhibited slower Sr²⁺ adsorption saturation compared to the non-calcined samples. Adsorption by the unheated TNT-2D was considerably faster than that of the calcined samples, reaching equilibrium within the first 30 min of contact with the Sr²⁺ solution. These results indicated that the hydrate phase of the TNTs supported the adsorption process. The presence of intercalated hydroxide molecules (OH) in the hydrate titanate was responsible for ion exchange in the aqueous solution.

The XRD patterns measured before and after Sr²⁺ ion adsorption are shown in Fig. 8. A peak shift toward lower 2θ values (19.5° and 27.5°) was observed in the diffraction pattern of the Sr²⁺ exchanged product. Cation sorption (onto the TNTs) also resulted in a substantial decrease in the diffraction intensity in the (310) plane of the TNT-2D (JCPDS card no. 72-0148).^40,41 These results indicated significant uptake of Sr²⁺ by the TNT adsorbent. The instantaneous uptake of Sr²⁺ was facilitated by ion exchange with an interlayer of Na ions at the structure of titanate nanotube (TNT-2D). These results were confirmed by TEM image mapping and energy dispersive spectroscopy (EDS) of the adsorbed TNT-2D (Fig. 9).

Fig. 8 XRD patterns obtained from the TNT-2D (a) and Sr²⁺-adsorbed TNT-2D (b) samples. Peak shifts toward lower angles were observed at 19.5° and 27.5°.

Fig. 9 TEM images before (A) and after (B) Sr²⁺ adsorption onto a TNT-2D sample, and (C) TEM image mapping with an EDAX image of a TNT-2D sample after Sr²⁺ adsorption.

The Sr²⁺ adsorption by the TNT adsorbent was further examined using a local structure-determining probe, XAFS. Ti K-edge X-ray absorption near-edge structure (XANES) spectra (Fig. 10A) revealed no structural differences between the samples before or after Sr²⁺ adsorption, with the exception of a slight decrease in the oscillation amplitude after the edge (4988–5008 eV). A similar decrease was observed for the peak at 2.05–3.05 Å in the extended X-ray absorption fine structure (EXAFS) of the Sr-adsorbed TNTs (Fig. 10B). As the TNT peak at 2.05–3.05 Å consisted of one Ti–Ti interaction at 3.0617 Å, one Ti–Na interaction at 3.1168 Å, and five Ti–Ti interactions at 3.1212–3.2058 Å, the scattering signals could not be separated; however, FEFF calculations were used to investigate the effects of Sr²⁺ replacement on the position of the Na ions.⁴² Outside of any structural disorder, the scattering signal from the Sr²⁺ ions was stronger than that from Na²⁺ ions; thus, the intensity of the peak at 2.05–3.05 Å should have been enhanced by Sr²⁺ adsorption onto the TNTs (ESI, Fig. S6†). The decrease in the peak in the experimentally derived Fourier transform corresponded to greater structural disorder among the Ti–Sr²⁺ bonds compared with the Ti–Na bonds.

Fig. 10 Ti K-edge XANES (A) and k₂-weighted Fourier transforms (B) of Ti K-edge XAFS functions, for the TNT and Sr-adsorbed TNT samples.

Experimental

Chemicals and reagents

All chemicals were of analytical grade. Na₂HPO₄ and strontium standard solutions (Sr-100) were purchased from Kanto chemicals (Japan), NaCl and calcium chloride dihydrate from JUNSEI Chemicals (Japan), and P25 (Degussa) is commercially available. Strontium chloride hexahydrate (Sigma-Aldrich) was used to prepare standard stock solution by dissolving in deionized water (DI, CBNU, pH 7). A 10 M NaOH (Samchun, South Korea) solution in DI water was prepared.

Synthesis of the TNTs from TiO₂ nanorods

TiO₂ nanorods were synthesized using the molten salt flux method. In a typical preparation procedure, P25 (Degussa), NaCl, and Na₂HPO₄ at a 1 : 4 : 1 ratio (by weight percentage) were ground with a mortar and pestle for 1 h to prepare a homogeneous mixture. This mixture was transferred into an aluminum crucible and calcined in a box furnace at 825 °C for 8 h. At the ambient temperature of the furnace crucible, the mixture was washed and filtered with excess boiled DI water. The filtrate was dried overnight at 80 °C in an oven. To synthesize the TNTs, 0.7 g of the TiO₂ nanorod powder were added to 10 M NaOH under continuous stirring. After stirring for 1 h, the solution was transferred into a 120 mL Teflon-lined autoclave to a 60% fill ratio and maintained at 160 °C for 6, 12, 24 (1D), 48 (2D), or 72 (3D) h. The product was then collected from the solution by centrifugation. The product was washed with distilled water several times to restore the pH to 7 and dried in an oven at 80 °C for 8 h. The yield of the as-prepared samples is listed in Table S1 (ESI†).

Characterizations

An X-ray diffraction (XRD) structural analysis was performed using a PANalytical X'pert Pro MPD diffractometer equipped with a Cu K_α radiation source (wavelength K_α1 = 1.540598 Å and K_α2 = 1.544426 Å) operated at 40 kV, 30 mA, and at a scan rate of 0.03° 2θ s⁻¹ and 2θ over an angular range of 5–80°. Scanning electron microscopy (SEM) observations were carried out using a field-emission scanning electron microscope (FESEM) (SUPRA 40VP, Carl Zeiss, Germany) equipped with an X-ray energy dispersive spectrometer (EDS). The Brunauer–Emmett–Teller (BET) specific surface area (S_BET) of the powders was calculated from the nitrogen adsorption isotherms measured using a Micromeritics ASAP 2010 nitrogen adsorption apparatus (USA). The desorption isotherm was used to determine the pore size distribution according to the Barrett–Joyner–Halenda (BJH) method, assuming a cylindrical pore model. The Sr²⁺ recovery analysis was carried out using a Profile Plus highly dispersed ICP (Teledyne Leeman Labs) instrument. Transmission electron microscopy (TEM) was performed with a JEOL JEM-3100F transmission electron microscope operating at 200 kV. The sample for TEM was prepared by placing a drop of the sample suspension in ethanol on a standard carbon-coated copper grid.

X-ray absorption fine structure (XAFS) experiments were carried out using the 7D beamline of Pohang Accelerator Laboratory (PLS-II, 3.0 GeV). The synchrotron radiation was monochromatized using Si (111) double crystal monochromators. At room temperature, the spectra of the Ti K-edge (E₀ = 4966 eV) were collected in transmission mode. The incident beam was detuned by 30% at the Ti K-edge to minimize contamination by higher harmonics. The intensities of the incident and transmitted beams were monitored using separate He-filled and N₂-filled IC SPEC ionization chambers, respectively. To minimize ions, the spectrum was measured under a helium atmosphere. The ATHENA program in the IFEFFIT software suite was used to analyze the local structure of the Ti in the TNTs and Sr-adsorbed TNTs.⁴³

Batch adsorption test

Batch experiments were performed to examine the adsorption of Sr²⁺ ions by the titanate adsorbents. The adsorption analysis was performed using an inductively coupled plasma (ICP) instrument. Sr²⁺ recovery analysis was performed by dispersing 20 mg of the titanate nanotube powder in 20 mL of the Sr²⁺ chloride stock solution over 30 min. The mixture was then allowed to settle at room temperature, and 10 mL were filtered using a syringe filter system for ICP analysis.

Sr²⁺ uptake by the titanate nanomaterials was calculated using the following equation:

where q_e denotes Sr²⁺ uptake (mg g⁻¹) by the TNTs, and C_o (mg L⁻¹) and C_e (mg L⁻¹) indicate the initial and equilibrium concentration of Sr²⁺, respectively. V denotes the solution volume (L) and m is the mass of the TNT adsorbent (g).

Conclusions

This work demonstrated for first time the formation and hydrothermal synthesis of TNTs from a molten salt TiO₂ NR precursor. Sr²⁺ uptake by the TNTs from a stock aqueous solution was measured. A short reaction time for hydrothermal synthesis yielded an admixture of untransformed products that displayed poor adsorption performances. The TNT samples prepared using long reaction times (3D) exhibited smaller surface areas and limited Sr²⁺ adsorption. Our findings confirmed that the TNTs hydrothermally synthesized over 48 h (2D) assumed uniform tubular structures with a large surface area that enhanced the Sr²⁺ uptake rate and the adsorption capacity. The titanate nanotube-2D sample provided 90% adsorption after 5 min, and 30 min was the optimal contact time for obtaining efficient and maximal Sr²⁺ adsorption. The Langmuir isotherm model fit well to the Sr²⁺ adsorption of the TNT-2D sample. The maximum adsorption capacity of Sr²⁺ by the TNT-2D sample, calculated using the Langmuir equation, was 113.6 mg g⁻¹. The TNT-2D sample showed the greatest Sr²⁺ uptake from 50–100 mg L⁻¹ Sr²⁺ stock solutions. Thus, TNTs were prepared from a molten salt solution of TiO₂ NRs and hydrothermal synthesis as a promising absorbent material for Sr²⁺ recovery. The high specific surface area and tubular structures of the TNTs distinguish them as effective adsorbents for heavy metal recovery applications.

Acknowledgements

This research was supported by the Basic Science Research Programs through the National Research Foundation of Korea (NRF), funded by the Ministry of Education, Science and Technology (2012R1A6A3A04038530), the Korea Ministry of Environment (MOE) as Public Technology Program, based on Environmental Policy (2014000160001), and the Basic Research Project (GP2016-023, 16-3224) of the Korea Institute of Geoscience and Mineral Resources (KIGAM), funded by the Ministry of Science, ICT and Future Planning of Korea.

References

1. J. R. Li, Z. L. Tang and Z. T. Zhang, Chem. Mater., 2005, 17, 5848–5855.
2. H. Zhang, X. P. Gao, G. R. Li, T. Y. Yan and H. Y. Zhu, Electrochim. Acta, 2008, 53, 7061–7068.
3. Y. Lan, X. P. Gao, H. Y. Zhu, Z. F. Zheng, T. Y. Yan, F. Wu, S. P. Ringer and D. Y. Song, Adv. Funct. Mater., 2005, 15, 1310–1318.
4. L. Torrente-Murciano, A. A. Lapkin, D. V. Bavykin, F. C. Walsh and K. Wilson, J. Catal., 2007, 245, 272–278.
5. D. V. Bavykin, J. M. Friedrich and F. C. Walsh, Adv. Mater., 2006, 18, 2807–2824.
6. D. V. Bavykin, A. A. Lapkin, P. K. Plucinski, J. M. Friedrich and F. C. Walsh, J. Phys. Chem. B, 2005, 109, 19422–19427.
7. S. H. Lim, J. Luo, Z. Zhong, W. Ji and J. Lin, Inorg. Chem., 2005, 44, 4124–4126.
8. G. S. Kim, H. K. Seo, V. P. Godble, Y. S. Kim, O. B. Yang and H. S. Shin, Electrochem. Commun., 2006, 8, 961–966.
9. E. Enache-Pommer, J. E. Boercker and E. S. Aydil, Appl. Phys. Lett., 2007, 91, 123116.
10. X. H. Feng, L. M. Zhai, W. F. Tan, F. Liu and J. Z. He, Environ. Pollut., 2007, 147, 366–373.
11. Y. Cho, S. Jang, Y. Kim, S. Komarneni and S. Kim, J. Porous Mater., 2011, 18, 125–131.
12. C. K. Lee, C. C. Wang, L. C. Juang, M. D. Lyu, S. H. Hung and S.-S. Liu, Colloids Surf., 2008, 317, 164–173.
13. X. Sun and Y. Li, Chem.–Eur. J., 2003, 9, 2229–2238.
14. R. Ma, T. Sasaki and Y. Bando, Chem. Commun., 2005, 948–950.
15. C. H. Han, D. W. Hong, I. J. Kim, J. Gwak, S. D. Han and K. C. Singh, Sens. Actuators, B, 2007, 128, 320–325.
16. J. Moon, K. Kim, C. H. Jung, Y. G. Shul and E. H. Lee, J. Radioanal. Nucl. Chem., 2000, 246, 299–307.
17. D. J. Yang, Z. F. Zheng, H. Y. Zhu, H. W. Liu and X. P. Gao, Adv. Mater., 2008, 20, 2777–2781.
18. V. A. Avramenko, V. V. Zheleznov, E. V. Kaplun, T. A. Sokol'nitskaya and A. A. Yukhkam, Radiochemistry, 2001, 43, 433–436.
19. S. Yusan and S. Erenturk, World J. Nucl. Sci. Technol., 2011, 1, 6–12.
20. G. Tian, J. Geng, Y. Jin, C. Wang, S. Li, Z. Chen, H. Wang, Y. Zhao and S. Li, J. Hazard. Mater., 2011, 190, 442–450.
21. E. E. Angino, G. K. Billings and N. Andersen, Chem. Geol., 1966, 1, 145–153.
22. S. J. Ahmadi, N. Akbari, Z. S. Yekta, Md. H. Mashhadizadeh and M. Hosseinpour, Korean J. Chem. Eng., 2015, 32, 478–485.
23. S. Ki Jeong and C. Sik Ju, Korean J. Chem. Eng., 2002, 19, 93–98.
24. N. Li, L. Zhang, Y. Chen, M. Fang, J. Zhang and H. Wang, Adv. Funct. Mater., 2012, 22, 835–841.
25. H. J. Hong, J. Ryu, I. S. Park, T. Ryu, K. S. Chung and B. G. Kim, J. Environ. Manage., 2016, 165, 263–270.
26. P. Basnet and Y. Zhao, J. Mater. Chem. A, 2014, 2, 911–914.
27. L. Xiong, C. Chen, Q. Chen and J. R. Ni, J. Hazard. Mater., 2011, 189, 741–748.
28. A. J. Du, D. D. Sun and J. O. Leckie, J. Hazard. Mater., 2011, 187, 401–406.
29. Y. C. Chen, S. L. Lo and J. Kuo, Colloids Surf., A, 2010, 361, 126–131.
30. J. Ryu, S. Kim, H. J. Hong, J. Hong, M. Kim, T. Ryu, I. S. Park, K. S. Chung, J. S. Jang and B. G. Kim, Chem. Eng. J., 2016, 304, 503–510.
31. R. H. Arendt, J. H. Rosolowski and J. W. Szymaszek, Mater. Res. Bull., 1979, 14, 703–709.
32. B. Liu, H. M. Chen, C. Liu, S. C. Andrews, C. Hahn and P. Yang, J. Am. Chem. Soc., 2013, 135, 9995–9998.
33. S. Zhang, L. M. Peng, Q. Chen, G. H. Du, G. Dawson and W. Z. Zhou, Phys. Rev. Lett., 2003, 91, 25.
34. D. Yang, Z. Zheng, H. Liu, H. Zhu, X. Ke, Y. Xu, D. Wu and Y. Sun, J. Phys. Chem. C, 2008, 112, 16275–16280.
35. A. Kukovecz, M. Hodos, E. Horvath, G. Radnoczi, Z. Konya and I. Kiricsi, J. Phys. Chem. B, 2005, 109, 17781–17783.
36. R. A. Zarate, S. Fuentes, A. L. Cabrera and V. M. Fuenzalida, J. Cryst. Growth, 2008, 310, 3630–3637.
37. E. Horvvath, A. Kukovecz, Z. Konya and I. Kiricsi, Chem. Mater., 2007, 19, 927–931.
38. S. Kasap, S. Piskin and H. Tel, Radiochim. Acta, 2012, 100, 925–929.
39. F. C. Wu, R. L. Tseng, S. C. Huang and R. S. Juang, Chem. Eng. J., 2009, 151, 1–9.
40. J. Huang, Y. Cao, Z. Deng and H. Tong, J. Solid State Chem., 2011, 184, 712–719.
41. H. Zhang, G. R. Li, L. P. An, T. Y. Yan, X. P. Gao and H. Y. Zhu, J. Phys. Chem. C, 2007, 111, 6143–6148.
42. J. J. Rehr, J. J. Kas, M. P. Prange, A. P. Sorni, Y. Takimoto and F. Vila, Ab initio theory and calculations of X-ray spectra, C. R. Phys., 2008, 10, 548–559.
43. B. Ravel and M. Newville, ATHENA, ARTEMIS, HEPHAESTUS: Data Analysis for X-Ray Absorption Spectroscopy Using IFEFFIT, J. Synchrotron Radiat., 2005, 12, 537–541.

---

Footnotes

† Electronic supplementary information (ESI) available: Table (S1), FESEM images (Fig. S1), FE-SEM images (Fig. S2), BJH and N₂ adsorption and desorption isotherms (Fig. S3), k₂-weighted Fourier transform of the theoretical scattering patterns (Fig. S4), calcination effect (Fig. S5). See DOI: 10.1039/c6ra14769k

‡ Equal contribution.

---