Revisiting the Solid-State Synthesis of Alkali–Tantalum(V) Oxyfluorides

Abstract

The solid-state synthesis of alkali–tantalum(V) oxyfluorides KTa₂O₅F and CsTa₂O₅F was revisited with the aim of streamlining their preparation as single-phase polycrystalline solids. Alkali fluorides (KF, CsF) and trifluoroacetates (KH(CF₃COO)₂, CsH(CF₃COO)₂) and tantalum oxide (Ta₂O₅) were used as precursors. Reaction temperatures were optimized by means of thermal analysis coupled with powder X-ray diffraction. Phase-pure KTa₂O₅F was obtained by heating stoichiometric mixtures of KF + Ta₂O₅ and KH(CF₃COO)₂ + Ta₂O₅ at 900–1000 °C in alumina crucibles under ambient atmosphere. Similar conditions were employed to synthesize phase-pure CsTa₂O₅F from a stoichiometric mixture of CsF + Ta₂O₅ heated to 800 °C. On the other hand, the preparation of CsTa₂O₅F from a mixture of CsH(CF₃COO)₂ + Ta₂O₅ required an excess of the trifluoroacetate precursor to obtain the targeted oxyfluoride as the sole crystalline phase. Results presented herein demonstrate that mixed-metal oxyfluorides previously thought to be synthetically challenging may be synthesized via facile solid-state reactions without the need for specialized containers and stringent conditions. In addition to streamlined synthetic routes to alkali– tantalum(V) oxyfluorides, a neutron powder diffraction study of the crystal structure of KTa₂O₅F is presented to fill a gap in crystallographic databases commonly accessed by materials and solid-state chemists.

1. Introduction

Alkali–tantalum(V) oxyfluorides KTa₂O₅F and CsTa₂O₅F were first reported by Magneli and Nord in 1965 and Babel et al. in 1967, respectively [1,2]. Their crystal structures are shown in Figure 1. KTa₂O₅F exhibits a tetragonal tungsten bronze structure (space group P4/mbm, see Figure 1a). It consists of a framework of corner-sharing Ta(O/F)₆ octahedra and potassium ions located in channels that extend along the c axis. The crystal structure of CsTa₂O₅F is a cubic pyrochlore that features a compact arrangement of corner-sharing Ta(O/F)₆ octahedra (space group Fd$\stackrel{-}{3}$m, see Figure 1b). The anionic substructures of KTa₂O₅F and CsTa₂O₅F exhibit occupational disorder of oxygen and fluorine. Both oxyfluorides were reported several decades ago and have been explored as ionic conductors along with their alkali–niobium(V) counterparts [3,4]. Despite these facts, two fundamental gaps persist when it comes to their synthesis and structural depiction. With regard to their preparation, a comprehensive review of the literature presented in

Table 1. Review of Synthetic Routes to KTa₂O₅F and CsTa₂O₅F.

Alkali Metal	Synthesis Method, Reactants, and Conditions	Structural Analysis	Ref.
K	Solid-state reaction of KF and Ta2O5 at 900 °C for 24 h in evacuated Pt tubes	PXRD (data not shown)	[1]
K	Alkali ion exchange of RbTa2O5F a in molten KNO3 under Ar atmosphere	SCXRD (data not shown) b	[3]
K	Electrolysis of K3TaOF6 c at 750 °C in molten NaCl–KCl under Ar atmosphere	IR (data not shown)	[5]
Cs	Solid-state reaction of CsHF2 and Ta precursor at 600–1000 °C under inert atmosphere	PXRD (data not shown) PDF No. 01–075–5921	[2]
Cs	Electrolysis of K3TaOF6 d and Cs3TaOF6 d at 650–850 °C in molten CsCl under Ar atmosphere	SCXRD and PXRD	[6,7,8]
Cs	Hydrolysis of H2TaF7 e in CsOH(aq) at 170 °C for 12 h under hydrothermal conditions	Rietveld analysis of PXRD PDF No. 01–078–9470	[9]
Cs	Hydrothermal reaction of BaO, TiO2, Ta2O5, and CsOH at 575 °C and 20,000 psi for 6 d	PXRD	[10]

^a RbTa₂O₅F was synthesized via solid-state reaction of RbF and Ta₂O₅ at 750 °C for 48 h under Ar. ^b Results from SCXRD analysis were given for KTa₂O₅⋅H₂O (PDF No. 01–071–1037). ^c K₃TaOF₆ was synthesized via solid-state reaction of K₂TaF₇, KF, and Ta₂O₅ at 810 °C for 7 h under Ar. A glassy carbon crucible was used as the reaction container. ^d K₃TaOF₆ (Cs₃TaOF₆) was synthesized by melting K₂TaF₇ (Cs₂TaF₇), KF (Cs), and Ta₂O₅ under Ar. A glassy carbon crucible was used as the reaction container. ^e H₂TaF₇ was prepared by dissolving Ta₂O₅ in HF at 120 °C for 12 h under hydrothermal conditions.

demonstrates the absence of straightforward solid-state routes to KTa₂O₅F and CsTa₂O₅F. Synthetic strategies utilized so far involve solid-state reactions requiring specialized reaction containers, molten salts, and inert atmosphere [1,2,3,5]. A combination of these experimental configurations and conditions was employed for the preparation of KTa₂O₅F and CsTa₂O₅F as byproducts from the electrolysis of K₃TaOF₆ and Cs₃TaOF₆, respectively [5,6,7,8]. Unlike KTa₂O₅F, CsTa₂O₅F has also been prepared using solution-based approaches involving hydrothermal reactions. However, these required the use of corrosive solvents (HF_(aq)) [9] or extreme hydrothermal conditions (575 °C, 20,000 psi) [10]. To the best of our knowledge, a straightforward solid-state route to KTa₂O₅F and CsTa₂O₅F is yet to be reported; that is, one in which phase-pure oxyfluorides are obtained upon heating a reaction mixture in a regular alumina crucible under ambient atmosphere. Additionally, a gap exists in the structural characterization of the tetragonal tungsten bronze KTa₂O₅F. As shown in Table 1, previous investigations that reported the preparation of this oxyfluoride did not conduct quantitative analysis of X-ray diffraction data. As a result, there are no entries for KTa₂O₅F in either the International Center for Diffraction Data Powder Diffraction File (ICDD–PDF) or in the Inorganic Crystal Structure Database (ICSD). For completeness, we note that crystal structures displayed in Figure 1 were built by replacing Ta⁵⁺ for Nb⁵⁺ (r_Ta⁵⁺ = r_Nb⁵⁺ = 0.64 Å) [11] in tetragonal KNb₂O₅F (isomorphous to KTa₂O₅F) and cubic CsNb₂O₅F (isomorphous to CsTa₂O₅F) [12,13,14]. The corresponding structural parameters are given in the Supporting Information (Table S1).

The primary aim of this study was to streamline the solid-state synthesis of alkali–tantalum(V) oxyfluorides KTa₂O₅F and CsTa₂O₅F. Alkali fluorides and trifluoroacetates and tantalum(V) oxide were used as metal precursors. Thermal analysis and Rietveld analysis of powder diffraction data were employed to identify reaction conditions that yielded single-phase oxyfluorides under ambient atmosphere. Additionally, the crystal structure of KTa₂O₅F was quantitatively probed using neutron powder diffraction. Results presented herein complement those recently reported by our group on the solid-state preparation of alkali–niobium(V) oxyfluorides KNb₂O₅F and CsNb₂O₅F [15].

2. Results and Discussion

Reactions between alkali fluorides and trifluoroacetates and Ta₂O₅ were first carried out in a TGA–DTA analyzer under flowing synthetic air with the goal of determining reaction temperatures for ambient synthesis. Besides providing us an accurate thermal profile of each reaction, this experimental configuration enabled precise control over reaction atmosphere. Such control was important because of the hygroscopic nature of both the reactants and the target products, which could impact reaction outcomes. Results from reactivity studies conducted in the TGA–DTA analyzer are summarized in Figure 2, where reaction thermal profiles and PXRD patterns of the corresponding products are shown. Reaction between KF and Ta₂O₅ resulted in the crystallization of KTa₂O₅F at 779 °C, as indicated by an exothermic peak in the DTA trace (Figure 2a). All diffraction maxima in the PXRD pattern were indexed to that oxyfluoride phase; no secondary crystalline phases were observed (Figure 2b). Similar results were obtained upon reacting KH(tfa)₂ and Ta₂O₅. The only difference was that no exothermic peak indicating crystallization of KTa₂O₅F was clearly observable in the DTA trace (Figure 2c), despite the fact that the oxyfluoride was the only crystalline phase obtained according to PXRD (Figure 2d). In the case of CsTa₂O₅F, crystallization from a CsF + Ta₂O₅ mixture occurred at 595 °C (Figure 2e). The resulting solid consisted of the targeted oxyfluoride phase as well as a secondary crystalline phase whose diffraction maxima could not be indexed (Figure 2f). The same crystallization temperature was extracted from the thermal profile of the CsH(tfa)₂ + Ta₂O₅ mixture (Figure 2g). However, in such a case, the resulting solid comprised CsTa₂O₅F, Ta₂O₅ (PDF No. 01–079–1375), and an unindexed phase (Figure 2h). With these results in hand, we chose 800 and 600 °C as the starting reaction temperatures for the synthesis of KTa₂O₅F and CsTa₂O₅F, respectively.

Synthesis optimization under ambient atmosphere was conducted by heating reaction mixtures in a box furnace, as described in the Experimental section. Attempts at synthesizing single-phase oxyfluorides began with KTa₂O₅F as the target phase. Results from successful attempts are given in Figure 3, while those from control experiments are given in the Supporting Information (Figure S1). Phase-pure KTa₂O₅F was obtained upon heating KF + Ta₂O₅ and KH(tfa)₂ + Ta₂O₅ mixtures at 1000 °C and 900 °C, respectively; using lower reaction temperatures led to the formation of a minor secondary phase (Figure S1). The phase-purity of the polycrystalline solids was demonstrated via Rietveld analysis of PXRD data, which showed that all diffraction maxima belong to KTa₂O₅F (Figure 3a,b). Structural refinements were conducted as described in the Experimental section, with the only difference being that atomic coordinates were refined for metal sites only; coordinates of oxygen/fluorine sites were fixed at their initial values. The resulting structural parameters are given in the Supporting Information (Table S2). Refined atomic coordinates of metal sites showed minimal variation relative to their initial values. On the other hand, occupancies of the K1 and K2 sites changed significantly; these values went from 0.5 and 1.0, respectively, to 0.89 and 0.8 (KF precursor) and to 0.99 and 0.75 (KH(tfa)₂ precursor). Similar changes were observed when refining the isomorphous oxyfluoride KNb₂O₅F and were attributed to the impact of differing synthetic conditions on potassium disorder [15]. Despite all diffraction maxima being indexed to the targeted oxyfluoride phase, our attention was drawn to the numerous intensity misfits. Attempts to remediate these misfits by refining the atomic coordinates of the oxygen/fluorine sites led to unreasonable values for tantalum–oxygen/fluorine distances. On this basis, and considering that there were no entries for KTa₂O₅F in crystallographic databases, we decided to carry out neutron diffraction studies. Room-temperature TOF–NPD data was collected for a KTa₂O₅F sample prepared at 1000 °C using KF as the potassium precursor. Next, Rietveld analysis was conducted to fit the starting structural model of KTa₂O₅F to the resulting NPD pattern. A full refinement was carried out, as described in the Experimental section. This included both metal and oxygen/fluorine atomic positions; only oxygen/fluorine site occupancies were fixed. Results from this analysis are given in Figure 3c and the corresponding refined structural parameters are provided in

Table 2. Structural Parameters of KTa₂O₅F from Rietveld Analysis of NPD Data.

Lattice Constants	Atom a	x	y	z	f	Uiso (Å2) b
a = 12.5798(4) Å	K1 (2a)	0	0	0	0.72 (3)	0.8 (3)
c = 3.95703(17) Å	K2 (4g)	0.1732 (4)	0.6732 (4)	0	0.89 (2)	0.47 (11)
	Ta1 (8j)	0.0750 (19)	0.20819 (19)	0.5	1	0.91 (4)
	Ta2 (2c)	0	0.5	0.5	1	1.12 (9)
	O1/F1 (2d)	0	0.5	0	0.833/0.167	2.56 (15)
	O2/F2 (8j)	0.1418 (3)	0.0690 (3)	0.5	0.833/0.167	1.48 (6)
	O3/F3 (8j)	0.3447 (3)	0.0023 (3)	0.5	0.833/0.167	1.712 (7)
	O4/F4 (8i)	0.0776 (3)	0.2094 (3)	0	0.833/0.167	1.53 (6)
	O5/F5 (4h)	0.2888 (2)	0.7888 (2)	0.5	0.833/0.167	0.85 (7)

^a Wyckoff positions are given in parentheses. ^b Given as 100 × U.

. Unlike what we observed when analyzing PXRD data, the experimental and calculated patterns were in excellent agreement. Further, the refined atomic positions translated into meaningful metal–oxygen/fluorine distances (K–O/F: 2.80–3.32 Å; Ta–O/F: 1.94–1.99 Å). Therefore, we concluded that the tetragonal tungsten bronze model of KTa₂O₅F provides an accurate structural description of this material.

Having successfully synthesized KTa₂O₅F using both potassium fluoride and trifluoroacetate precursors, we moved forward aiming to synthesize single-phase CsTa₂O₅F. Results from the most successful attempts are given in Figure 4, while those from control experiments are given in the Supporting Information (Figure S2). Phase-pure CsTa₂O₅F was obtained upon heating CsF + Ta₂O₅ at 800 °C; using lower reaction temperatures resulted in unreacted Ta₂O₅ as a secondary phase (Figure S2). Rietveld analysis of the corresponding PXRD pattern showed that all diffraction could be indexed to the pyrochlore phase CsTa₂O₅F (PDF No. 01–078–9470, Figure 4a). The resulting structural parameters are given in the Supporting Information (Table S3). Refined atomic coordinates of the oxygen/fluorine site showed minimal variation relative to their initial values. The preparation of single-phase CsTa₂O₅F using a stoichiometric CsH(tfa)₂ + Ta₂O₅ mixture proved more challenging. According to Rietveld analysis, conducting the reaction at 800 °C led to a mixture of CsTa₂O₅F (≈96 wt.%) and unreacted Ta₂O₅ (≈4 wt.%) (Figure 4b). Attempts to get rid of that secondary phase by lowering the reaction temperature to 700 °C were unsuccessful (Figure S2). Reaction temperatures above 800 °C were not explored because results from thermal analysis showed that unreacted Ta₂O₅ was present even after heating up to 1000 °C (vide supra). On this basis, we decided to tune the composition of the reaction mixture by adding an excess of CsH(tfa)₂ (Cs:Ta molar ratio = 1.1:2). Our hypothesis was that the inability to achieve a quantitative reaction between CsH(tfa)₂ and Ta₂O₅ was due to the fast volatilization of the former, which melts at ≈120 °C [16]. Heating the non-stoichiometric mixture to 800 °C led to the formation of CsTa₂O₅F as the only crystalline phase observable via Rietveld analysis of PXRD (Figure 4c). However, careful observation of the experimental pattern and of the difference curve in the angular range 26.5–32.0° showed that the background was noticeably more intense than for solids prepared using stoichiometric mixtures. Thus, although CsTa₂O₅F was the only observable crystalline phase, the presence of a non-negligible amount of amorphous (or nanocrystalline) matter could not be ruled out.

3. Materials and Methods

Synthesis of KH(tfa)₂ and CsH(tfa)₂ Precursors. Polycrystalline KH(tfa)₂ and CsH(tfa)₂ (tfa = CF₃COO) were synthesized via solvent evaporation [15,17,18]. K₂CO₃ (99%), Cs₂CO₃ (99.9%), and anhydrous CF₃COOH (tfaH, 99%) were purchased from Sigma-Aldrich (USA) and used as received. Double-deionized water was used as a cosolvent. 1 mmol of the corresponding metal carbonate (K₂CO₃ or Cs₂CO₃), 3 mL of double-deionized water, and 3 mL of tfaH were mixed in a 50 mL two-neck round-bottom flask. The resulting colorless transparent solution was immersed in a sand bath and heated at 65 °C for 24 h (potassium) and 48 h (cesium) under a constant flow of 140 mL min⁻¹ of N₂. White polycrystalline solids were thus obtained.

Synthesis of KTa₂O₅F and CsTa₂O₅F. KTa₂O₅F and CsTa₂O₅F were synthesized via solid-state reaction under air (relative humidity 25–65%). Starting materials included KF (99%), CsF (99%), and Ta₂O₅ (99.99%) purchased from Sigma-Aldrich (St. Louis, MO, USA) and used as received, and KH(tfa)₂ and CsH(tfa)₂ prepared in-house (vide supra). KF + Ta₂O₅, KH(tfa)₂ + Ta₂O₅, CsF + Ta₂O₅, and CsH(tfa)₂ + Ta₂O₅ reaction mixtures were prepared by mixing and grinding stoichiometric amounts of starting materials in an agate mortar in a nitrogen-filled glovebox. ≈100 mg of each mixture were transferred to 5 mL alumina crucibles, covered with alumina disks, removed from the glovebox, and quickly placed inside a box furnace at 100 °C. Reaction mixtures were then heated under air to the target reaction temperature (T_reaction, 600–1000 °C) at a rate of 10 °C min⁻¹ for 1 h. After heating, the furnace was allowed to cool to ≈400 °C, and the crucibles were removed. A total of five heating cycles were carried out, with intermediate grindings between each cycle performed under air. Polycrystalline white powders were thus obtained. Phase-pure KTa₂O₅F and CsTa₂O₅F were obtained upon heating KF + Ta₂O₅ and KH(tfa)₂ + Ta₂O₅ mixtures at 1000 °C and 900 °C, respectively, and CsF + Ta₂O₅ at 800 °C.

Thermal Analysis (TGA–DTA). Thermogravimetric (TGA) and differential thermal (DTA) analyses of reaction mixtures KF + Ta₂O₅, KH(tfa)₂ + Ta₂O₅, CsF + Ta₂O₅, and CsH(tfa)₂ + Ta₂O₅ were conducted under dry synthetic air (100 mL min⁻¹) using a TGA–DTA analyzer (SDT2960, TA Instruments, New Castle, DE, USA). ≈35–65 mg of sample were placed in an alumina crucible, held at 35 °C for 10 min, and then ramped to 1000 °C at a rate of 10 °C min⁻¹. Heating was stopped once the temperature reached 1000 °C and samples were allowed to cool to room temperature.

Powder X-ray Diffraction (PXRD). PXRD patterns were collected using a D2 Phaser diffractometer operated at 30 kV and 10 mA (Bruker Corporation, Billerica, MA, USA). Cu Kα radiation (λ = 1.5418 Å) was employed. A nickel filter was used to remove Cu Kβ radiation. Diffractograms were collected in the 10–60° 2θ range using a step size of 0.012° and a step time of 0.4 s, unless otherwise noted.

Time-of-Flight Neutron Powder Diffraction (TOF–NPD). Neutron diffraction data were collected using the POWGEN diffractometer of the Spallation Neutron Source at Oak Ridge National Laboratory (Oak Ridge, TN, USA). ≈700 mg of KTa₂O₅F were loaded into a 6 mm diameter vanadium can. A TOF–NPD pattern was collected at 300 K with the high-resolution setting and a center wavelength of 0.8 Å in the 3.6–176 ms time window. Raw diffraction data was processed using POWGEN’s autoreduction software.

Rietveld Analysis. Rietveld analysis [19,20] of PXRD and NPD data was conducted using GSAS-II (version 5.6.4) [21]. Initial structural parameters of KTa₂O₅F and CsTa₂O₅F are given in the Supporting Information (Table S1). In the case of X-ray diffraction data, the following parameters were refined: (1) scale factor and sample displacement; (2) background, which was modeled using a shifted Chebyschev polynomial function; (3) lattice constants; (4) atomic coordinates of metal atoms when allowed by space-group symmetry; (5) one isotropic atomic displacement parameter (U^iso) for each type of atom in the structure, constrained to U_iso^K = 1.2 × U_iso^Ta and U_iso^O/F = 1.5 × U_iso^Ta in KTa₂O₅F and to U_iso^O/F = 1.5 × U_iso^Cs = 1.5 × U_iso^Ta in CsTa₂O₅F; (6) occupancy factors (f) of K1 and K2 sites constrained to 2 × f_K1 + 4 × f_K2 = 5; and (7) crystallite size and microstrain. Occupancy factors of the oxygen/fluorine sites were fixed. In the case of neutron diffraction data, the following parameters were refined for KTa₂O₅F: (1) scale factor; (2) background, which was modeled using a logarithmic interpolation function; (3) instrument parameters, including diffractometer constant and contributions to peak profile; (4) lattice constants; (5) atomic positions when allowed by space-group symmetry; (6) an isotropic displacement parameter for each site in the structure; (7) occupancy factors (f) of K1 and K2 sites constrained to 2 × f_K1 + 4 × f_K2 = 5; and (8) crystallite size and microstrain. Occupancy factors of the oxygen/fluorine sites were fixed. The quality of the refined structural models was assessed using the difference between the observed and calculated intensities divided by the standard uncertainty of the observed intensities (Δ(I)/σ(I)) and R_w residuals. Crystal structures were visualized using VESTA (version 3.90.5a) [22].

4. Conclusions

In conclusion, results presented in this article demonstrate that KTa₂O₅F and CsTa₂O₅F oxyfluoride phases are accessible by traditional solid-state reaction methods using alkali fluorides and trifluoroacetates as alkali and fluorine sources. Specialized containers and/or stringent conditions like those employed in previous investigations are not needed to prepare single-phase KTa₂O₅F and CsTa₂O₅F. This result is in line with that previously reported by our group for alkali–niobium(V) oxyfluorides KNb₂O₅F and CsNb₂O₅F [15]. Therefore, the preparation of other oxyfluorides belonging to these composition spaces is worth revisiting with a focus on streamlining experimental conditions. In addition to synthetic insight, this article provides a set of high-quality structural parameters for the tetragonal tungsten bronze KTa₂O₅F, thus filling a gap in crystallographic databases commonly accessed by materials and solid-state chemists. There is room to further improve the structural depiction of KTa₂O₅F and CsTa₂O₅F, particularly with regard to local ordering in mixed-occupancy sites present in cationic (K/vacancy) and anionic substructures (O/F) [23].