Article pubs.acs.org/IC
Structure Determination and Time-Resolved Raman Spectroscopy of Yttrium Ion Exchange into Microporous Titanosilicate ETS‑4 Aaron J. Celestian,*,† Caleb J. Chappell,‡ Melinda J. Rucks,§ and Pauline Norris∥ †
Department Department § Department ∥ Department ‡
of of of of
Mineral Sciences, Natural History Museum of Los Angeles, Los Angeles, California 90007, United States Geology and Enivronmental Earth Science, Miami University, Oxford, Ohio 45056, United States Geoscience, Stony Brook University, Stony Brook, New York 11794, United States Chemistry, Western Kentucky University, Bowling Green, Kentucky 42101, United States
S Supporting Information *
ABSTRACT: The ion exchange of yttrium, one of the five most critical rare-earth elements as outlined by the U.S. Department of Energy, into ETS-4 is a dynamic, multistep ion exchange process. The ion exchange process was followed using in situ time-resolved Raman spectroscopy, and the crystal structures of the pre-exchange and post-exchange forms were determined by single-crystal X-ray diffraction. In situ Raman spectroscopy is an ideal tool for this type of study, as it measures the spectral changes that are a result of molecular geometry changes at fast time intervals, even where symmetry and unit volume changes are minimally detected by X-ray diffraction. By tracking the stepwise changes in the peak positions and intensities in the spectra, where we focused primarily on the strong spectral features corresponding to titania quantum wires and three-membered-ring bending and breathing modes, we constructed molecular models to explain the changes in the Raman spectrum during ion exchange. The multistep ion exchange process started with rapid absorption of Y into the Na2 site, causing titania quantum wires to kink. After this initial uptake, the exchange process slowed, likely caused by hydration coordination changes within the channels. Next, Y exchange accelerated again, during which time the Y site moved closer to the framework O2−. Crystal structures of the maximal Y exchanged ETS-4 material were determined and confirmed the splitting of the Y site. Inductively coupled plasma optical emission spectroscopy was also used to quantify the extent of Y exchange and to measure if there were indications of titania leaching from the framework.
■
INTRODUCTION Ion exchange in zeolites and related microporous materials have broad-ranging applications, including water purification,1 gas separation,2 catalysis,3 and heavy-cation sequestration4 from aqueous solutions. Some of the most important and as yet outstanding areas of research in cation exchange in nanoporous materials are determining the molecular driving forces and finding ways to model the chemical dynamics of this spontaneous processes. Another area that needs to be addressed is determining the crystal structures of intermediate structural states where ingoing cations change, stress, or deform the crystal structure. These areas of research are now being addressed with new in situ experimental techniques, and our study contributes to that effort. In addition, we are investigating the process of how rare-earth elements (specifically yttrium) are exchanged into the nanoporous structure. A better understanding of rare-earth-element (REE) ion exchange would have potential applications in the design of future REE catalysts, materials for novel gas separation technology, and pharmaceuticals. One particularly beneficial application is the possible use of microporous minerals and materials in selective sequestration of rare-earth elements from aqueous solutions.5 The rare-earth elements (Sc, Y, and the lanthanides)6 are used in a large number of technological applications,7−9 where Y is among the five most critical REEs (the others are Nd, Eu, Tb, Dy) according to the U.S. Department of Energy.10 © XXXX American Chemical Society
The complex nature of REE separation starts at mineral ore separation11 followed by element separation. Current methods of REE separation are summarized elsewhere.12−17 The use of REEs in technology and chemical processing is increasing,7 and there is a relative scarcity of economic ore deposits in the United States. The Molycorp Mountain Pass Mine in California was the only commercial REE mine in the United States, where the mineral bastnäsite was the primary REE ore mineral (currently closed to mining operations as of November 2015). These factors have driven interest in further studies of the crystal chemistry and ion exchange mechanisms of REE sequestration as a possible way of developing ionselective absorbents, or enhanced catalytic materials, for the mining, recycling, and environmental industries. The similar physiochemical properties of the REEs (e.g., electronic configuration, similar ionic radii, coordination geometry in crystals, hydration spheres in aqueous solutions, and valence states18) have led to difficulty in their separation. The focus of this study is to investigate ion exchange mechanisms of REEs, specifically Y, into the ETS-4 structure. The goal is to better understand the ion exchange mechanisms of high charge density cations into known functional microporous materials for future materials science applications. Received: July 13, 2016
A
DOI: 10.1021/acs.inorgchem.6b01627 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry Zorite19 (Na6Ti5[Si12O34](O,OH)5·11H2O) is a sodium titanium silicate mineral found at the Lovozero Massif, Kola Peninsula, Russia. The zorite microporous framework is topologically identical with the Ca-rich haineaultite (the Ca analogue form), chivruaiite (Na and Ca members), and the synthetic zorite analogue Na-ETS-4 (Na member, Englehard Titanium Silicate). The ion-selective separation properties of the Na-ETS-4 member have been well studied since its initial development at the Englehard Coorporation.2,20−22 The efficiency and affinity of Na-ETS-4 to selectively absorb Sr2+, which promotes selective gas separation and catalysis, have led to many experimental investigations into its ion exchange properties.23−28 However, there are very few studies of the exchange of high-valent cations into this material, nor have there been detailed studies of the time-dependent changes during ion exchange in ETS-4. The work presented in this study helps to explain how high-valent cations diffuse into porous titanium silicates and has broader application to materials design research.
■
for all experiments in this study. Calibration of laser frequency and spot position were performed using a polystyrene standard, and the spectrometer was calibrated using the excitation lines from a neon lamp. Exposures for the time-resolved experiments were a total of 30 s using the average of three sequential 10 s exposures. The microscope and laser settings were held constant for data acquisitions using a 50× (0.5 N.A.) long-working-distance objective in orthoscopic geometry (50 μm laser slit), and the spot size of the laser at the sample was estimated to be 2.1 μm. A total of 400 spectra were collected for each experiment, each spectrum was smoothed using a Savitzky− Golay algorithm (5 points, 2.411 cm−1), and both the original data and the processed data were saved as separate ASCII files. Peak fitting was performed with the Magic Plot software. The advantage of using this software is the flexibility of peak shape profile usage and the ability to batch process the fitting procedure for all spectra. To maintain software stability during the peak-fitting process (using a Gaussian profile), the fitting was performed in stages where the spectra were fit in three discrete regions (Figure 1). The regions
EXPERIMENTAL SECTION
Synthesis of Na-ETS-4. The sodium form of ETS-4 was synthesized as follows. In a 250 mL polypropylene graduated bottle were placed 4.7 mL of deionized H2O and 1.35 mL of C12H28O4Ti; the mixture was stirred with a magnetic stir plate,. Next, 10 mL of hydrogen peroxide was added to the stirred mixture, and then 30 mL of deionized H2O was quickly added. After H2O was added, 2.75 mL of 10 M NaOH was added. This mixture was allowed to age for 1 h prior to loading in autoclaves. A 12 mL portion of this gel was placed in each 23 mL Teflon-lined Parr autoclave. In each autoclave was placed 1 mL of AS-40 (40 wt % amorphous silica). Finally, 1 mL of H2O was placed in each autoclave, which was then sealed and placed in an oven at 210 °C for 2 days (48 h). X-ray powder diffraction confirmed a pure product of Na-ETS-4. Crystal sizes averaged 5 × 10 × 200 μm. Inductively Coupled Plasma Optical Emission Spectroscopy. As-synthesized and ion-exchanged samples of Na-ETS-4 were weighed (∼30 mg), and each sample was placed in its own platinum dish. Fluxing and wetting agents, approximately 1 g of a lithium tetraborate−lithium carbonate mixture purchased from Claisse and 0.05 g of ammonium iodide purchased from MP Biomedicals, were placed in each dish. Samples were heated in a furnace at 1050 °C for 5 min. The mixtures were cooled and then dissolved in 10% trace metal grade hydrochloric acid purchased from Fisher Scientific. Samples were subsequently analyzed using an ICAP 6500 ICP-OES system from ThermoScientific. The amounts of extraframework cations in the Y-exchanged Na-ETS-4 material (Y-ETS-4) were measured from three separate analyses to be 7.40 ppm for Na and 13.51 ppm for Y, resulting in a near 50% Y exchange. No trace metals were detected from the Na-ETS-4 analysis. From the ICP-OES data, approximate chemical formulas for Na-ETS-4 and Y-ETS-4 were determined to be Na6Ti5[Si12O34](O,OH)5·xH2O and YNa3Ti5[Si12O34](O,OH)5·xH2O, respectively. The H2O and OH contents were not determined by ICP. Ion Exchange. Ion exchange solutions were prepared at room temperature and ambient pressure. The 0.001 M YNO3 solution was made in deionized H2O. The low REE concentrations were used to slow the ion exchange process. A vacuum environmental cell29 was used for the in situ Raman spectroscopy experiments. Suitable single crystals for single-crystal X-ray diffraction experiments (approximately 30−50 crystals) of Na-ETS-4 were placed inside the cell. The flow rate of the exchanging solution was held constant for the duration of the experiment using a Masterflex peristaltic pump set to 0.05 mL/min. After ion exchange, a suitable single crystal was selected for singlecrystal XRD. Raman Spectroscopy. A Thermo DXR dispersive Raman microscope equipped with a 780 nm diode laser (14 mW at the sample), a Peltier-cooled CCD detector, and a high-resolution diffraction grating (≃1.2 cm−1 spectral resolution between 50 and 1800 cm−1) was used
Figure 1. Results of peak fitting of ETS-4. Peaks are labeled v1−v20, with each fitted curve being given under the spectrum. The dashed red line is the fit to the data (solid blue line). Fitted regions are delineated by vertical dashed black lines and are labeled A−C. For the timeresolved Raman spectroscopy data, the v1, v2, v11, and v18−v20 peaks were too weak to be fitted precisely for all data sets and therefore have been removed from that analysis. are as follows: region A is dominated by Si−O−Ti bends, region B is dominated by pyramidal O−Ti−O and Ti−O−Ti bending, and region C is dominated by octahedral and pyramidal Ti−O stretching modes. Beyond region C the peaks are dominated by Si−O stretching modes.30−33 Single-Crystal X-ray Diffraction. Single-crystal X-ray diffraction (SCXRD) data collection was performed on a Bruker AXS Quasar diffractometer using an APEX-II CCD; structure solution and refinement details for Y-ETS-4 are summarized in Table 1, and unit cells of other ETS-4 materials are compared in Table 2. For these diffraction experiments, a suitable single crystal was found using a polarized light microscope, where the crystal had good (not undulatory or diffuse) extinction and was transparent in planepolarized light. The poor scattering statistics of this crystal are due in part to the small size of the crystal and the amount of disorder in the framework and extraframework atomic sites. As a result, we were not able to resolve many of the expected H2O positions, and we acknowledge that our estimates of interstitial H2O content and bond valence sums of extraframework cations are likely underestimated. Crystal structures for the as-synthesized Na-ETS-4 were also determined and were in good agreement with those in published work.34,35 B
DOI: 10.1021/acs.inorgchem.6b01627 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry Table 1. Summary of SCXRD Data Collection and Structure Refinement temp/K cryst syst space group a, Å b, Å c, Å α, deg β, deg γ, deg V, Å 3 Z ρcalcd, g/cm3 F(000) cryst size, mm3 radiation 2θ range for data collection, deg index ranges no. of rflns collected no. of indep rflns no. of data/restraints/params goodness of fit on F2 final R indexes (I ≥ 2σ(I)) final R indexes (all data) largest diff peak/hole, e Å−3
298 orthorhombic Cmmm 7.1628(3) 23.1657(9) 6.9267(2) 90 90 90 1149.36(7) 1 2.064 697.0 0.128 × 0.025 × 0.014 Mo Kα (λ = 0.71073 Å) 3.516−61.922 −10 ≤ h ≤ 10, −33 ≤ k ≤ 33, −10 ≤ l ≤ 9 15287 1076 (Rint = 0.0413, Rσ = 0.0170) 1076/0/87 1.188 R1 = 0.0561, wR2 = 0.1754 R1 = 0.0629, wR2 = 0.1821 0.94/−1.20
Table 2. Comparison of Zorite/ETS-4 Unit Cellsa Na-zorite Pb-zorite Sr-zorite Na-ETS-4 Sr-ETS-4 Y-ETS-4 a
a (Å)
b (Å)
c (Å)
ref
7.238(4) 7.161(3) 7.23810(33) 7.1751(11) 7.2259(13) 7.1628(3)
23.241(7) 23.22(1) 23.1962(12) 23.2272(4) 23.1900(5) 23.1657(9)
6.955(4) 6.980(3) 6.96517(31) 6.9727(6) 6.9699(13) 6.9267(2)
34 41 37 42 42 current study
Transformed to the unit cell setting in this study.
■
RESULTS AND DISCUSSION Single-Crystal X-ray Diffraction. The structure of Y-ETS-4 (Figures 2−5) consists of chains of TiO6 octahedra that link 8MR (eight-membered rings) of SiO4 tetrahedra. These rings are further linked by disordered five-coordinated TiO5 square pyramids. The secondary building units (SBUs) in Y-ETS-4 are 8MR, 7MR, 6MR, and 3MR. Each of these structural aspects of Y-ETS-4 are described in the following sections. The TiO6 chains (Figure 3) are oriented parallel to the [100] direction. The kink angle of the chains is 138.8° before ion exchange, which forces the O2− in the TiO6 square plane to be adjacent to another TiO6 square plane O2− (2.68 Å). After Y ion exchange, the kink angle of the TiO6 chains is slightly smaller at 137.6°. These titania chains have been previously referred to as quantum wires, and these wires contain natural defects caused by localized charge-transfer transitions.36 The exchange of a high-valence Y3+ into the crystal structure that is bonded to the O2− in the quantum wires could cause greater defects (with an increase in charge transfer from the REE) in the wire structure as the kink angle in the chains decreases. Four of the SiO4 tetrahedra in the 8MR are disordered with ring dimensions of 7.12, 6.92, and 6.12 Å (Figure 4). The disorder of the SiO4 groups accommodates the disorder of the TiO5 linking groups, where the SiO4 arrangement is either up/up (UU) or down/down (DD). These SiO4 8MRs are
Figure 2. Crystal structure of Y-ETS-4 as viewed down [100]. Titania (TiO6, TiO5) are shown as light blue polyhedra, SiO4 as dark blue tetrahedra, Na+ in yellow, and Y3+ in green; unbound O2− atoms in red are modeled as H2O. Si−O 8MRs are seen in the plane of the figure, as well as 6MRs occupied by Na1. H atom sites were not determined and are not shown.
joined by TiO6 chains to form a 3MR SBU containing two TiO6 octahedra and one SiO4 tetrahedron. Before ion exchange, the bond angle in the 3MR of Ti−OSi is 135°, and after Y ion exchange the angle is 136.6°, accounting for the shorter Y2−O distances of 2.4 and 2.6 Å to the O2− of the TiO6 group. A third structural feature is the elliptical 7MR (Figure 3) consisting of two TiO6 groups, two ordered SiO4 groups (parts of the 3MR described above), one TiO5, and two disordered SiO4 groups. The ordered sites of the SiO4 8MR form the building units for the 6MR consisting of two TiO6 groups and four SiO4 groups (Figure 4). The disordered Na1 site is located in this 6MR window and is bound to six framework O2−. The Y2 sites are located in the window of the 7MR and are primarily C
DOI: 10.1021/acs.inorgchem.6b01627 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry
Figure 4. Silica 8MRs and mixed polyhedra 6MRs. The H2O sites in Na-ETS-4 (labeled as Zow, transparent red circles, in the 8MR of this viewing direction) were searched for after Y exchange but were not found in the Y-ETS-4 structure. The arrow next to Zow indicates the possible direction of H2O migration to hydrate the Y2 site. Transparent yellow circles indicate the position of the Na2 site in Na-ETS-4 prior to ion exchange. Arrows at the Na2/Y2 site illustrate the proposed direction of Y2 splitting and movement direction as indicated by Raman spectroscopy. H atom sites were not determined and are not shown.
Figure 3. Crystal structure of Y-ETS-4 as viewed down [001]. See Figure 2 for the color scheme. The 7MRs are occupied by Y3+, and the titania chains are seen in the plane of the figure oriented parallel to [100]. The disordered SiO4 groups (around the Si2 site) are also shown bound to the disordered TiO5. H atom sites were not determined and are not shown.
bound to five framework O2− and to several disordered H2O molecules (Figure 4). The Y cation preferentially exchanges into the Na2 site, and this is in close agreement with the previously reported Sr site for Sr-ETS-4.37 Attempts were made to model Y into the Na1 site, but least-squares refinements would not converge, and so further efforts to forcefully model Y into the Na1 site were abandoned. For the penultimate crystal structure refinements, Na1 and the split Y2 site (Y2A, Y2B, Y2C) occupancies were freely refined. Attempts were made to forcefully charge balance the crystal structure by fixing site occupancies so that their total charge sum was equal to 2, but this resulted in a worse overall fit to the data with Rwp > 20. The sum of the Y2 occupancies was 0.115(5), and this is only slightly less than the expected occupancy (0.16) for full Y exchange into the Na2 site. We are aware of the indication of OH groups on the apical O7 site of the TiO5 polyhedra after exchange (Ti2−O7 bond lengths before and after exchange are 1.67 and 1.96 Å, respectively), and the bond valence sum decreased from 4.17 to 3.63 vu. The change in bond length and bond valence sums may indicate the presence of OH at the O7 site and may affect the final proposed chemical formula and Y2 occupancies. The increased H component, coupled with the decreased H2O, could be a result of hydrolysis caused by Y38 in the channels during exchange. However, we do not have direct measurements of the chemical occupancy for the H sites (or pH of the
solutions) and therefore have decided to omit the influence of H chemical occupancy on the Y2 or Na1 values during the structure refinements. The final crystal structure of modeled atomic positions should be independent of the modeling (or calculation) of the H component. The ultimate goal of this study was to locate the positions of the Y and Na sites, and future work will be needed to determine the occupancies and orientations of the OH and H2O sites. Given the higher charge, larger electronegativity, and smaller ionic radius of Y3+ (as compared to Sr2+37), there might exist enhanced catalytic and selective gas separation capabilities (at ambient or elevated temperatures) of REE exchange titanium silicates. Further work is needed to confirm this hypothesis and the functional role of REEs in ETS-4. Events in the Raman Spectra. To better understand the dynamics of the ion exchange process, time-resolved Raman spectroscopy was used to monitor the molecular changes in situ. Even though all possible Raman-active modes for Na-ETS-4 could not be distinguished in the collected Raman spectrum due to significant peak broadening and peak overlap (Figures 6 and 7), we were able to track the movement of major peaks corresponding to titania, silica, 3MR, and chain vibrations. On the basis of past molecular vibrational interpretations of Raman mode analyses (Table 3) and how those vibrational modes changed as Y exchanges into the crystal structure, a model of D
DOI: 10.1021/acs.inorgchem.6b01627 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry
Figure 7. Evolution of peak positions as a function of time during the Y exchange into Na-ETS-4. Plots are shown as difference Raman shifts relative to their original starting position (see Table 3). Lines are denoted with their vibrational mode label and their original starting position in cm−1. Major changes in the peak evolution paths are shown as events 0−4 (see text for descriptions). Figure 5. Local bonding geometry of the Y2 (only Y2C shown for clarity) and Na1 sites and the structure of the 3MR (Ti1−O3−Si1− O3−Ti1). H atom sites were not determined and are not shown.
of the initial v8 peak and v9 peak (see Table 3) to higher wavenumbers, and minor changes of the v10 peak to lower wavenumbers (see Table 3). These changes indicate distortions in the octahedral geometry of the TiO6 groups and may be interpreted as the TiO6 octahedra bending closer together as the Ti−O−Ti bending angle of the TiO6 chains becomes more acute and the internal O−Ti−O bends distorting as the octahedra move closer, forcing a charge repulsion between the Ti of adjacent octahedra (Figure 7). The square plane of the TiO6 octahedra is shown to be lengthening, as indicated by the decrease in the v10 peak (Figure 7). The movement of the v10 peak is small in comparison to other major changes occurring. Between Events 1 and 2. The processes occurring from minutes 4 to 9 (Figure 7) are indicated by small changes in the v10 peak, while all other peaks showed very small changes (
