Cation-ordering in low-temperature niobium-rich NbWO bronzes: New anodes for high rate Li-ion batteries

doi:10.21203/rs.3.rs-3693278/v1

Download PDF

Article

Cation-ordering in low-temperature niobium-rich NbWO bronzes: New anodes for high rate Li-ion batteries

https://doi.org/10.21203/rs.3.rs-3693278/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Niobium tungsten oxide phases, as anodes for lithium-ion batteries, have gained considerable attention due to their high volumetric charge storage densities at high rates. Here we report the microwave-assisted solution-based synthesis and characterization of two new niobium tungsten bronze structures, NbWO_5.5 and β−Nb₂WO₈, which adopt a simple tetragonal tungsten bronze (TTB) structure and a TTB with √2×√2 superstructure, respectively. These novel TTB structures were synthesized at temperatures below 900°C for Nb:W ratios of 1–3, the latter composition (Nb₃WO_10.5) resulting in a √2×√2 TTB closely related to β−Nb₂WO₈. Nb:W ≥ 4 compositions result in two-phase behaviour forming Nb₂O₅ and Nb₃WO_10.5, while W-rich bronzes (Nb/W < 1) exhibiting local domains of WO₃ within the NbWO_5.5 lattice. Through comprehensive analysis using X-ray and neutron diffraction and scanning transmission electron microscopy - energy dispersive spectroscopy (STEM-EDS) we observed cation ordering in the Nb-rich bronzes at both short and long length scales. The microwave synthesis method results in NbWO microspheres with a unique, microporous structure, where primary particles are interconnected by amorphous NbWO bridges. Notably, these NbWO bronzes, with the highest Nb content and thus specific energy density of all known NbWO bronzes, exhibited high-rate capabilities and long cycle lives, positioning them as promising candidates for energy storage applications. Our findings underscore the potential of the microwave-assisted solution method for synthesizing complex oxide materials, with significant implications for the development of advanced functional materials across diverse applications.

Physical sciences/Materials science/Materials for energy and catalysis/Batteries

Physical sciences/Chemistry/Chemical synthesis/Microwave chemistry

Physical sciences/Chemistry/Inorganic chemistry/Solid-state chemistry

The need for fast charging, high-rate Li-ion batteries stems from the ever-increasing demand for efficient and convenient energy storage solutions. Whether it is a smartphone, laptop, or electric vehicle, the ability to rapidly cycle the battery is essential for uninterrupted productivity and enhanced user experience. Furthermore, fast charging batteries, coupled with the relevant infrastructure, are crucial for the widespread adoption of electric vehicles, as they help reduce time at charging stations, reducing the land needed for charging stations, helping to promote the transition to clean, sustainable transport. Thus, fast charging Li-ion batteries emerge as important solutions to meet the demands of modern society, motivating the search for novel electrode materials that can enable rapid charging while maintaining long-term battery performance and stability.

Niobium-based metal oxides such as niobium tungsten oxides (NbWOs) are promising electrode materials for fast charging batteries.^1,2 The high rate performance of NbWOs is attributed to the topochemical charge storage mechanism that exhibits (i) minimal, highly reversible structural changes during the lithiation and delithiation cycles, (ii) high Li-ion diffusion coefficients due to the presence of tunnels that allow fast-ion transport, and (iii) high electronic conductivity upon lithiation.

Niobium tungsten oxides have been extensively studied by Wadsley and other groups since the early 1960s due to their rich crystal chemistry.^3–9 NbWOs can be represented as a Nb₂O₅-WO₃ pseudo-binary system, their structure being derived from the simple ReO₃ type structure. WO₃ crystallises in a distorted ReO₃ structure that comprises a primitive cubic arrangement of corner sharing WO₆ octahedra. Isomorphous substitution of a fraction of hexavalent W-ion in the lattice with a pentavalent Nb-ion results in a net negative charge. The charge neutrality in these oxides can be retained by either removing the O-ions or by generating new cation sites. The removal of a row of O-ions results in the condensation of the structure with a shear motion, converting some corner-sharing octahedra to edge-sharing octahedra (Fig. 1a). If this occurs in two dimensions, then it results in blocks and the resulting structure is known as a crystallographic shear (CS) or block structure. Multiple phases have been identified with different block sizes that depend on the Nb:W ratio.¹⁰ Additionally, new cation sites can be generated by a 45° rotation of a 2×2 block of octahedra (orange circle in Fig. 1b) relative to the surrounding octahedra. This rotation affects the eight surrounding square channels forming four trigonal and four pentagonal channels (PCs). A larger unit cell is needed to describe the resulting “tetragonal tungsten bronze” (TTB) structure (Fig. 1b).¹¹ These transformations occur within the lattice along two dimensions and the lattice remains unaffected in the third dimension. The formation of different bronze phases is dictated by the arrangement of PCs in the lattice and the extent of occupancy of the PCs by the cations (and associated oxygen anions).¹²

The Nb₂O₅/WO₃ (or NbWO) phase diagram first constructed by Roth and Waring in 1965 still captures all the different block and bronze phases synthesized at elevated temperatures.¹³ To more readily compare the results obtained in this study, the original phase diagram is replotted in terms of the molar ratios of Nb and W (Fig. S1). The phase diagram comprises a series of solid “line phases” or binary oxides with fixed ratios of unitary oxides, NbO_2.5:WO₃ with decreasing Nb:W ratio along the x-axis. The phases with Nb:W ratio greater than 0.667 form crystallographic shear/block structures, while phases with Nb:W ratios between 0.667–0.2 form the tungsten bronze structures. Materials with compositions that deviate from those of the line phases form distinct biphasic mixtures of line phases, or in some cases, intergrowth phases.^14,15,16 The synthesis temperature also affects the resulting phase, leading to wide structural and compositional variability. For example, a bronze phase with a 2:1 Nb:W ratio exists between 1000–1115°C, but above this temperature, it phase-separates to form block (Nb:W = 16:5) and bronze (Nb:W = 8:9) phases.¹³

The phase diagram is based on materials calcined at temperatures ranging from 1000 to 1500°C as NbWOs are generally synthesized using solid-state methods, and lower calcination temperatures often result in incomplete reactions of the precursors. Fiegel et al. attempted to construct an equilibrium diagram below 900°C, but observed the coexistence of precursor phases that did not completely react even after 30 days.¹⁷ The poor reactivity is due to the high energetic barrier of the intra and inter-particle diffusion of highly charged cations across the particles of the oxides precursors, which are generally in the nm - µm range. The poor ionic transport, coupled with similar thermodynamic stability and structures of these complex phases, can lead to the formation of islands of different phases with varying O/M ratio within a single crystallite, further complicating the characterization process. For example, recent microscopy work by Krumeich and co-workers demonstrated the formation of multiple domains of bronze phases Nb₈W₉O₄₇, Nb₄W₇O₃₁ and disordered NbWO within a single crystallite of the material synthesized at 1250°C.¹⁶

The formation of multiple domains within a single crystallite and the poor reactivity at lower temperatures raise two crucial questions: (i) Can stable phases of NbWO form at temperatures below 1000°C? and (ii) How can we overcome the significant energy barrier hindering the migration of Nb and W cations across particles to achieve a homogeneous distribution within the lattice? One potential solution lies in employing solution-based methods, which can address the cation migration issue observed in solid-state approaches. Precipitation of NbWO from corresponding salt precursors through hydrolysis offers better control over the synthesis conditions. However, the hydrolysis rate for Nb is notably faster than that of W, resulting in the hydrolysis of Nb-ions first, forming Nb₂O₅ islands, followed by the precipitation of WO₃. This sequential process leads to inhomogeneity, impeding the uniform distribution of cations in the final product. Wang et al. recently used a solution-based method to synthesize Nb₁₈W₁₉O₉₃ at 850°C; however, no new phases were synthesized in their study.¹⁸

In this study, we employed an innovative microwave-assisted solvothermal technique to synthesize NbWO phases with varying Nb/W ratios at temperatures below 1000°C that were hitherto inaccessible. The use of microwaves results in instant precipitation of the precursor in which the need for long-range metal ion diffusion is eliminated, as the cations are evenly dispersed within the precursor.

To explore the NbWO phases achievable below 1000°C, we explored distinct Nb:W ratios from 1/5 to 5. Two new phases, termed Nb1 (NbWO_5.5; Nb1 denoting a material with an Nb:W ratio of 1) and Nb2 (Nb₂WO₈) with cation ordered TTB structures were identified via powder X-ray diffraction (PXRD). Nb2 has a different structure than the metastable bronze phase Nb₂WO₈^13,19 and thus we label our new Nb2 polymorph β-Nb₂WO₈ and the more well established material, α-Nb₂WO₈. A related phase, Nb3 (Nb₃WO_10.5) was also identified with a structure similar to Nb2. Scanning transmission electron microscopy (STEM) based high-angle annular dark field (HAADF) imaging and energy-dispersive X-ray spectroscopy (EDS) were used to establish the cation ordering in these structures, the combined TEM and Rietveld analysis results revealing that Nb and W occupy distinct sites in the lattice, leading to the formation of supercell structures. The structures exhibited high-rate capabilities when cycled vs. Li, which we discuss in terms of their crystal structures and unusual particle microstructures for this class of materials.

Nb1: NbWO_5.5

A series of amorphous “NbWO” precursors with Nb/W ratios of 1/5, 1/4, 1/3, 1/2, 1, 2, 3, 4, and 5 were prepared by coprecipitation of NbCl₅ and WCl₆ from an ethanol solution, while constant stirring in a closed vessel under pressure and microwave irradiation, the latter resulting in local, instantaneous, and uniform heating of the sample. The samples were then calcined at various temperatures to monitor the crystallisation process. All the precursors had similar morphologies and crystallized at similar temperatures, and thus to analyse this in more detail, we first focus on Nb1.

Variable temperature (VT) PXRD was performed on Nb1 precursor to identify the optimal temperature for crystalline NbWO formation, a crystalline phase emerging at 800–900°C, the structure remaining unaltered upon cooling to room temperature (Fig. 1c). The NbWO precursor particles exhibit spherical morphology with diameters ranging from 1 to 5 µm (Fig. 1d,e). During crystallization, the NbWO retains its spherical shape while forming secondary agglomerates composed of nano-sized primary particles (Fig. 1f,g). SEM-EDS elemental mapping images (Fig. S2) and X-ray fluorescence spectroscopy reveal a homogeneous distribution of niobium, and tungsten, within each spherical particle, and an Nb:W ratio that is essentially identical to the quantities of Nb and W salts utilized during synthesis (Table S1).

All reflections observed in the PXRD pattern of Nb1 were successfully indexed to a unit cell with cell parameters a₁ = b₁ = 12.2223 Å, c = 3.9381 Å, and space group P4/mbm, which corresponds to a simple TTB structure. A joint time of flight (TOF) neutron and X-ray diffraction Rietveld refinement was then performed using the structure of the lead tungsten bronze (Pb_0.26WO₃) ²⁰ as a model (Fig. 2a,b). The Nb and W cations were initially equally distributed between two octahedral (2c (0, 0.5, 0.5) and 8j (x, y, 0.5)) and one pentagonal (4h (x, 0.5 + x, 0.5)) cation site (Fig. 2c). In addition to the existing oxygen positions in the Pb_0.26WO₃ structural model, an additional oxygen atom (4g (x, 0.5 + x, 0)) was placed below the cation in the pentagonal cation site to form a 7 coordinated pentagonal bipyramidal (PB) site. The cation positions and their occupancies were refined to investigate any preferential occupancy by Nb and W of the octahedral and PB cation sites, the final refined structure (Fig. 2c) indicating that the Nb and W ions were distributed across all octahedral sites, with a slight preference for W in the 2c site. The PB site displayed a low occupancy that could only be accounted for with a partial (1/3) occupancy of this site by Nb only, to maintain the overall Nb:W ratio of 1, with the additional 4g oxygen site being associated with a similarly low occupancy. All the atoms were then refined together and the refined atomic coordinates and their occupancies are given in Table S2.

Diffraction studies provide long-range ordering information for the crystal structure of Nb1 but provide limited insight into the local structure, including cation ordering and atomic scale defects. STEM helps to overcome these limitations and complements the X-ray and neutron diffraction study of the order-disorder phenomena. The layer thickness of the metal oxide polyhedra is 3.9 Å, which corresponds to the c parameter (Fig. 2c). Atomic-resolution, STEM-HAADF imaging of the a-b plane captures the -M-O-M-O- tunnels in this plane (Fig. 2d) and the overlay with the structure model (one unit cell) indicates the atomic positions of Nb and W. A relatively large area STEM-HAADF image of a Nb1 crystal along the c-axis and the corresponding FFT is shown in Fig. S3(a,d). The FFT-derived lattice parameter (a ~ 13 Å) closely matches the value obtained from PXRD data. The intensity of atomic columns in the STEM-HAADF image is associated with the -M-O-M-O- strings throughout the sample thickness. An analysis of the PCs across the HAADF image in Fig. S4a indicates that most PCs show weak contrast (intensity), while the PCs showing contrast also exhibit different intensities (Fig. S4b, highlighted in green and pink circles). The contrast from HAADF images depends on the Z number of the atomic columns. The observed variance in the contrast from the PCs can be due to different cation occupancies, or different cation species. The cationic columns were then manually color-coded, revealing the random occupation of PCs by cations, resulting in the formation of a simple tetragonal tungsten bronze (TTB) structure as depicted in Fig. S4. Based on an analysis of 255 unit cells (15×15) containing a total of 900 PCs (Table S3), 350 PCs are either fully or partially occupied, resulting in approximately one-third of the PCs being filled similar to the occupancy observed in the Rietveld refinement. The varying intensity of the atomic columns in a HAADF image can be attributed to variations in the occupancy of a single element in the lattice position and/or different elements occupying the same lattice position.²¹ Recent advancements in aberration-corrected (S)TEM combined with EDS have enabled atomic-resolution elemental mapping of a crystal structure, ²² and we use this method here to confirm the preferential occupancy of Nb and W on the octahedral and PB sites obtained from the XRD refinements. High-resolution STEM-HAADF images and the corresponding EDS elemental maps of Nb1 along the c-axis are shown in Fig. 2(d-g). In the atomic resolution EDS map, the bright spots highlighted by cyan circles correspond to the 2c octahedral site, which, on the basis of the XRD refinements, is preferentially occupied by W atoms. The spots indicated by the yellow circles correspond to the octahedral site 8j that is partially filled by W atoms, suggesting a shared occupancy of W and Nb atoms. The magenta circles correspond to the PB site 4h, which appears bright in the Nb-map but is missing in the W-map, confirming that the site is occupied only by Nb atoms. This observation is further supported by the overlay EDS map in Fig. 2g.

Exploration of the Nb-rich NbWO phases

NbWOs were synthesised with Nb:W ratios of 2, 3, 4, and 5 using identical synthesis procedures to that used for Nb1 and were calcined at 800°C. All the materials (Nb2-Nb5) exhibited PXRD reflections similar to Nb1, the Nb-rich phases Nb4 and Nb5 (Fig. S5), displaying additional reflections corresponding to the low temperature bronze-like phase T-Nb₂O₅.²³ This suggests that NbWO phases with Nb:W ≥ 4 synthesized at 800°C are unstable as TTBs, and the excess Nb crystallizes as Nb₂O₅. To investigate this further, the simple TTB phases (Nb1), Nb2 and Nb3, were analysed using high-resolution synchrotron XRD and NPD (Fig. 3). While the major reflections in the PXRD pattern of Nb2 and Nb3 could be indexed to the Nb1 simple TTB unit cell, a new set of reflections marked ‘#’ was seen, which grew sharper with an increase in the Nb:W ratio from 2 to 3 and could be indexed (hk0, h + k = odd) to a larger tetragonal cell (space group P4) with lattice parameters a = b = 17.3787 Å, c = 3.9437 Å, representing a supercell (√2a˳×√2a˳) of the Nb1 cell (Fig. 3b). The hk0 family of reflections, which undergo a change due to the supercell formation, are shaded in green, their intensity increasing with increasing the Nb:W ratio, while those marked in pink indicate reflections whose intensity diminishes.

To further validate the formation of a supercell, we examined the TOF neutron diffraction data of Nb1, Nb2, and Nb3 (Fig. 3c). The reflections at approximately 3.9 Å (c-parameter) in all three diffraction patterns correspond to the 001 reflection in both cells. The patterns of Nb2 and Nb3 displayed at peak at 17.57 Å which is absent in Nb1 and is assigned to the 010 reflection corresponding to the a parameter of the supercell. The 010 peak in the pattern of Nb3 was sharper than that of Nb2, suggesting the formation of larger domains of the supercell structure with increasing Nb concentration in the lattice. The formation of the supercell was also evident from the position of the 210 reflections at 5.53 Å and 7.83 Å in the diffraction patterns of simple (Nb1) TTB and the supercell (Nb2 and Nb3) structure, respectively.

Crystal structure of Nb2

The crystal structure of Nb2 was refined with PXRD data using the √2a˳×√2a˳ TTB superstructure reported by Li and co-workers for Na₃Nb₁₂O₃₁F as a model.²⁴ To fix the atoms in special positions 4j and 4k with z = 0 and 0.5 respectively, a higher symmetry space group P 4/m instead of P4 was chosen. In addition, two extra oxygen atoms were placed above the cations in the two sets of PCs (purple and orange circles within the green unit cell in Fig. 3b) to create PB sites in the model structure. The Nb and W ions were initially distributed equally among the five octahedral and two PB sites. Refining the occupancies and position parameters of the cation revealed that the two sets of octahedral sites in the centre of the unit cell (cyan in Fig. 4c) exhibited high electron density and were thus filled with W, while the other octahedral sites (magenta in Fig. 4c) were occupied by Nb. The occupancy of the cations in the four PB sites surrounding the central octahedral site (purple rings in Fig. 3b) refined to a negligible value. Therefore, the PB cation and the corresponding oxygen sites were left unoccupied, and the other PB site was filled with Nb to maintain the stoichiometry determined by EDS and XRF measurements (Table S1). The positions of the O atoms were refined through a combined PXRD and NPD refinement (Fig. 4a,b) to obtain the final structure of Nb2 (Fig. 4c). The resulting structural parameters are reported in Table S4. Since the experimental data exhibited broader supercell reflections (hk0, h + k = odd) compared to the calculated pattern (Fig. S6a), selective broadening was incorporated to this class of reflections resulting in a significant improvement of the fit (Fig. S6b). In comparison to the Nb1 phase, the Nb2 phase exhibited an ordered arrangement of the empty and filled PCs.

Antiphase boundaries (APB) and structural evolutions of the bronzes

The observation of selective broadening in a PXRD pattern often indicates the presence of lattice defects resulting from the formation of a supercell. The HR(S)TEM image of Nb2 in Fig. S6c reveals the presence of antiphase boundaries (APBs), marked by a green dashed line that separates the two ordered domains. A schematic model in Fig. S6d shows how they originate from disorder with 2₁ screw along the a-axis. The unit cell shaded in red transposes with the yellow-shaded unit cell during the 2₁ screw dislocation at the APB. These boundaries introduce a discontinuity in the ordering of vacant PCs along the [100] and [010] directions thereby broadening the supercell reflections corresponding the 120, 230 and 140 planes (Fig. S6d). To examine the cation ordering along the -M-O-M-O- channels in the ordered and APB regions, STEM-HAADF images and EDS maps were captured along the c-axis (Fig. 4d-g). The image also features an overlay of the structure model (one-unit cell) illustrating the atomic positions of Nb and W. The FFT analysis (Fig. S3e) of a relatively large-area STEM-HAADF image (Fig. S3b) yielded a lattice parameter of a ~ 18 Å, which closely matched the diffraction result (17.4 Å).

The XRD, neutron, and TEM results unequivocally establish Nb2 as a supercell of Nb1. In the Nb1 phase, the O:M ratio is 2.75, and approximately one-third of the PB sites are randomly occupied (Fig. S4d). As the Nb content increases, the O:M ratio decreases, and the excess Nb ions progressively occupy the vacant PB cation sites. In the Nb2 phase, the O/M reduces to 2.37, and half of the PB sites are now filled. To avoid cation repulsion, four out of the eight PCs surrounding the 2x2 central block of WO₆ octahedra remain unoccupied. This ordering of the occupied (purple squares, Fig. S18b) and vacant PCs (yellow squares, Fig. S18b) results in the formation of the supercell, as discussed in more detail in the SI.

Crystal structure of Nb3

The PXRD pattern of Nb3 closely resembles that of Nb2, exhibiting sharper supercell reflections and a few additional weak reflections (indicated by red asterisks in Fig. S7). While most of these reflections can be attributed to the supercell with lattice parameters a₃ = b₃ = 17.4166 Å and c₃ = 3.9454 Å, two of them remain unindexed (highlighted by blue arrows in Fig. S7 and listed in Table S5) suggesting the presence of impurities in the sample. These unindexed peaks do not correspond to any of the known NbWO phases. Moreover, the additional reflections observed in the PXRD pattern of Nb3 are poorly fit using the structure model obtained for the Nb2 structure, even when accounting for the change in Nb:W ratio to 3. The FFT (Fig. S3f) analysis of the STEM-HAADF image (Fig. S3c) of Nb3 along the c-axis exhibits similarities to Nb2 rather than Nb1, and the measured a lattice parameter is ~ 18 Å which is comparable to Nb2. The combined findings from diffraction and STEM suggest that an increase in the Nb content from Nb1 to Nb2 leads to the formation of a supercell, and the supercell structure is retained even with a further increase in Nb content from Nb2 to Nb3. The sharper supercell reflections observed in the diffraction pattern of Nb3 as compared to Nb2 (Fig. 3a,c) indicate a lower occurrence of defects in Nb3.

Thermodynamic Stability

To investigate the thermodynamic stability, the newly synthesized NbWOs were recalcined at 800°C for 30 days and their PXRD patterns are shown in Fig. S8. Interestingly, the Nb1 phase remained unchanged during the prolonged heating, indicating its thermodynamic stability. However, in the case of Nb2, Rietveld refinement revealed that ~ 71.5% of the material retained its structure while the remaining portion decomposed into the known bronze phase α−Nb₂WO₈. In Nb3, only 7.6% of the material retained its structure while the majority decomposed into Nb₁₄W₃O₄₄ (33%) and α−Nb₂WO₈ (59.4%). The phase composition of the decomposed products also correlated with the observed morphological changes: Nb1 maintained its spherical morphology with an increase in the size of the primary particles (Fig. S9), while Nb2 only partially retained its spherical morphology, with the remaining portion crystallizing as micron-sized rods, consistent with typical NbWO bronze and block phases. (Fig. S10). Nb3 was completely converted into micron-sized particles (Fig. S11).

The stability of the Nb1 phase was observed up to 1000°C, decomposing into Nb₁₂W₁₁O₆₃, Nb₈W₉O₄₇ and Nb₁₄W₃O₄₄ phases when calcined at 1200°C for 12 h (Fig. S12). The decomposition was accompanied by a transformation into rod-like structures as seen by SEM (Fig. S13). The spherical secondary particles were only observed at temperatures below 900°C, and the transformation to rod-like structures was always accompanied by decomposition.

W-rich NbWO phases

Finally, to explore the W-rich phases, various NbWOs with Nb/W ratios of less than 1 were synthesized. Samples W2, W3, W4 and W5, with Nb:W ratios of 1/2, 1/3, 1/4 and 1/5, were precipitated and calcined at 800°C following the procedure identical to that used for Nb1. The PXRD patterns of all W-rich phases exhibit reflections corresponding to the Nb1 phase, the excess W crystallizing as WO₃ (Fig. S14), the concentration of WO₃ progressively increasing from W2 to W5. STEM-HAADF images of the W-rich phases reveal TTB lattices similar to the Nb1 phase with domains of cubic close-packed cations. Additionally, within the lattice, the excess W atoms form domains that crystallize in the ReO₃-type structure (Fig. S15).

Electrochemical performance of the low temperature phases

The electrochemical performance of the NbWO bronzes was evaluated in half cells to access their potential as energy storage materials. The cells were cycled within a specific potential range of 3–1.3 V (Fig. 5a-c). These materials demonstrated remarkable high-rate capabilities, with a reversible capacity of 160, 155 and 110 mAh/g at the end of 1000 cycles at high rates (10C), for Nb3, Nb2, and Nb1, respectively (Fig. 5d-f). Although both Nb2 and Nb3 phases possessed a supercell TTB structure, the Nb3 phase demonstrated a higher capacity compared to Nb2 largely due to its lower molar mass.

The gravimetric capacity of all the three bronzes reported here are higher than the low capacity of 80 mAh/g reported for the α-Nb₂WO₈ with a mass loading of ~ 0.8 mg cm^− 2 cycling at 3 A/g (a rate that is lower than 10 C).²⁵ The increase in capacity of the new bronzes can be attributed to the presence of interconnected vacant pentagonal channels that run across the particle helping in the Li ion transport and storage (see SI for further discussion).

Focused ion beam (FIB) cross-section of the microspheres (Fig. S16) show primary particles interconnected by NbWO bridges forming porous structure which likely contributes to the excellent rate performance. FFT analysis show that the primary particles are crystalline (Fig. S17b and d), while the NbWO bridges linking them are amorphous (Fig. S17c).

We have investigated the low-temperature region of the NbWO phase diagram and have identified new NbWO bronzes in regions of the phase diagram where block phases form at higher temperatures. Leveraging a microwave-assisted solution method, we successfully synthesized two new TTB structures, Nb1 and Nb2 – as part of a series of NbWO bronzes (Nb_xW_yO_2.5x+3y with 1 ≤ x/y ≥ 3). This innovative synthesis method enabled precise atomic-scale mixing and instantaneous precipitation, eliminating the reliance on ion migration within the solid reactants and reducing the synthesis temperature. Importantly, these phases have not been observed when using traditional synthesis techniques such as solid-state and coprecipitation methods. The crystal structures of Nb1 and Nb2 were determined through PXRD and neutron diffraction, the results corroborating the local structures revealed by STEM, the latter identifying cation ordering at both macroscopic and microscopic length scales.

The Nb content exerted a significant influence on cation ordering within the a-b plane and along the c-axis. As the Nb content increased, the occupancy of PB sites increased correspondingly. In β−Nb₂WO₈, the excess Nb content compared to Nb1 resulted in the ordering of 50% of the vacant PCs surrounding the central 2×2 WO₆ corner-sharing octahedra within the unit cell. Further increasing the Nb content led to the formation of Nb3 with fewer antiphase boundaries in the particles. The microwave synthesis of W-rich bronzes yielded localized domains of WO₃ rather than W-rich bronzes. These new findings underscore the intricate and complex nature of the crystal structure exhibited by the NbWO bronze phases and emphasize the importance of employing a combination of techniques to fully determine their chemical and structural properties.

The materials exhibit excellent electrochemical performance, with high-rate capabilities cycled at 10C up to 1000 cycles. β−Nb₂WO₈ shows considerably enhanced electrochemical performance over the known bronze phase α−Nb₂WO₈, a material which is synthesized at higher temperature, but contains fully filled PB sites. This research highlights the potential of microwave-assisted solution methods for synthesizing complex oxide materials, offering exciting prospects for the development of advanced functional materials across various fields. By overcoming the limitations of conventional synthesis approaches, this methodology opens avenues for exploring metastable phases and low temperature phase diagrams, potentially leading to materials with new properties.

Acknowledgements: S.N. acknowledges support from the Royal Society (United Kingdom) and Science and Engineering Research Board (Government of India) for the award of Newton-Bhabha International Fellowship (NIF/R1/180075). C.P.G., S.N., and B.W. acknowledge support from CBMM for partial funding of this work. We would like to thank the Science and Technology Facilities Council (STFC) for access to the Polaris and GEM beamline at the ISIS neutron source. We thank Diamond Light Source for access and support in use of the electron Physical Science Imaging Centre (Instrument E01 and proposal numbers MG32597 and MG28816) that contributed to the STEM-EDS results presented here. S.N. and A.M. thank Energy Transitions@Cambridge, Isaac Newton Trust and EPRG for funding the STEM measurements at Oxford. We thank EPSRC Underpinning Multi-User Equipment call (EP/P030467/1) for TEM imaging. S.V. acknowledges funding from the Cambridge Commonwealth European and International Trust and Royal Society (RP/R1/180147). C.D., A.M., S.V., and W.S., acknowledges Faraday Institution via the degradation project (FIRG011, FIRG020) and SOLBAT. BW acknowledges the ERC (via Consolidator Grant (Mighty 866005)) and an Advanced Grant (BATNMR 835073), and the support from EPSRC and Cambridge Graphene Centre.

Competing interests: C.P.G. is the co-founder and a major share-holder of a company that seeks to develop fast charging solutions using Nb-based anode chemistries; S.N. is currently an employee of the same company (Nyobolt). S.N. and C.P.G., via Nyobolt ltd., have filed a UK patent application (GB2218984.9) covering the materials and high-rate energy storage application described in this manuscript.

Author contributions: S.N. and C.P.G. conceived the idea. S.N. designed the experiments, synthesis, and characterization of the materials, as well as the execution of electrochemistry experiments. A.M. conducted the FIB-SEM work, while the TEM of FIB lamella was performed by both A.M. and S.N. A.M., S.N., and M.D. carried out STEM-EDS experiments. W.S. was responsible for the STEM-HAADF imaging. S.V. and C.D. contributed to discussions, provided valuable suggestions, aided in figure preparation, and contributed to the analysis of TEM data. B.W. contributed to the material synthesis process. S.N. drafted the original manuscript and prepared the figures. S.N., A.M., and C.P.G. reviewed and edited the manuscript incorporating contributions from all co-authors.

Griffith, K. J., Wiaderek, K. M., Cibin, G., Marbella, L. E. & Grey, C. P. Niobium tungsten oxides for high-rate lithium-ion energy storage. Nature 559, 556–563 (2018).
Koçer, C. P., Griffith, K. J., Grey, C. P. & Morris, A. J. Lithium diffusion in niobium tungsten oxide shear structures. Chem. Mater. 32, 3980–3989 (2020).
Roth, R. S. & Wadsley, A. D. Multiple phase formation in the binary system Nb₂O₅–WO₃. I. Preparation and identification of phases. Acta Crystallogr. 19, 26–32 (1965).
Roth, R. S. & Wadsley, A. D. Multiple phase formation in the binary system Nb₂O₅–WO₃ . II. The structure of the monoclinic phases WNb₁₂O₃₃ and W₅Nb₁₆O₅₅. Acta Crystallogr. 19, 32–38 (1965).
Roth, R. S. & Wadsley, A. D. Multiple phase formation in the binary system Nb₂O₅–WO₃. III. The structures of the tetragonal phases W₃Nb₁₄O₄₄ and W₈Nb₁₈O₆₉. Acta Crystallogr. 19, 38–42 (1965).
Roth, R. S. & Wadsley, A. D. Multiple phase formation in the binary system Nb₂O₅–WO₃. IV. The block principle. Acta Crystallogr. 19, 42–47 (1965).
Andersson, S., Mumme, W. G. & Wadsley, A. D. Multiple phase formation in the binary system Nb₂O₅ –WO₃. The structure of W₄Nb₂₆O₇₇, an ordered intergrowth of the adjoining compounds WNb₁₂O₃₃ and W₃Nb₁₄O₄₄. Acta Crystallogr. 21, 802–808 (1966).
Allpress, J. G., Sanders, J. V. & Wadsley, A. D. Multiple phase formation in the binary system Nb₂O₅–WO₃ VI. Electron microscopic observation and evaluation of non-periodic shear structures. Acta Crystallogr. Sect. B Struct. Crystallogr. Cryst. Chem. 25, 1156–1164 (1969).
Allpress, J. G. & Wadsley, A. D. Multiple phase formation in the binary system Nb₂O₅-WO₃. VII. Intergrowth of H-Nb₂O₅ and WNb₁₂O₃₃. J. Solid State Chem. 1, 28–38 (1969).
Bevan, D. J. M. & Hagenmuller, P. Non-stoichiometric compounds: tungsten bronzes, vanadium bronzes and related compounds. in Comprehensive Inorganic Chemistry (Pergamon Press: Oxford, 1973).
Hyde, B. G. & O’Keeffe, M. Relations between the DO₉(ReO₃) structure type and some `bronze’ and `tunnel’ structures. Acta Crystallogr. Sect. A 29, 243–248 (1973).
Krumeich, F. The complex crystal chemistry of niobium tungsten oxides. Chem. Mater. 34, 911–934 (2022).
Roth, R. S. & Waring, J. L. Phase equilibria as related to crystal structure in the system niobium pentoxide-tungsten trioxide. J. Res. Natl. Bur. Stand. Sect. A Phys. Chem. 70A, 281 (1966).
Krumeich, F. Bronze‐type niobium tungsten oxides. Wiley Analytical Science Magazine.
Wörle, M. & Krumeich, F. On the structural complexity of tetragonal tungsten bronze type niobium tungsten oxides. Zeitschrift fur Anorg. und Allg. Chemie 647, 98–106 (2021).
Krumeich, F. Intergrowth of niobium tungsten oxides of the tetragonal tungsten bronze type. Zeitschrift fur Naturforsch. - Sect. B J. Chem. Sci. 75, 913–919 (2020).
Fiegel, L. J., Mohanty, G. P. & Healy, J. H. Equilibrium diagram of the system Nb₂O₅-WO₃. J. Chem. Eng. Data 7, 365–369 (1964).
Wang, L. et al. To effectively drive the conversion of sulfur with electroactive niobium tungsten oxide microspheres for lithium−sulfur battery. Nano Energy 77, 105173 (2020).
Lundberg, M. The crystal structure of Nb₂WO₈. Acta Chem. Scand. 26, 2932–2940 (1972).
Triantafyllou, S. T., Christidis, P. C. & Lioutas, C. B. An X-ray and electron diffraction study of the intergrowth tungsten bronze Pb_0.26WO₃. J. Solid State Chem. 134, 344–348 (1997).
Nellist, P. D. & Pennycook, S. J. The principles and interpretation of annular dark-field Z-contrast imaging. Advances in Imaging and Electron Physics vol. 113 (2000).
Iijima, S., Yang, W., Matsumura, S. & Ohnishi, I. Atomic resolution imaging of cation ordering in niobium–tungsten complex oxides. Commun. Mater. 2, 1–9 (2021).
Tamura, S., Kato, K. & Goto, M. Single crystals of T‐Nb₂O₅ obtained by slow cooling method under high pressures. ZAAC ‐ J. Inorg. Gen. Chem. 410, 313–315 (1974).
Li, D. X. The Crystal Structure of Na₃Nb₁₂O₃₁F. J. Solid State Chem. 73, 1–4 (1988).
Yao, W. et al. Structural insights into the lithium ion storage behaviors of niobium tungsten double oxides. Chem. Mater. 34, 388–398 (2022).
Smith, R. I. et al. The upgraded Polaris powder diffractometer at the ISIS neutron source. Rev. Sci. Instrum. 90, (2019).
Williams, W. G., Ibberson, R. M., Day, P. & Enderby, J. E. GEM - General Materials Diffractometer at ISIS. Phys. B Condens. Matter 241–243, 234–236 (1997).
Rodríguez-Carvajal, J. FULLPROF: a program for Rietveld refinement and pattern matching analysis, abstracts of the satellite meeting on powder diffraction of the XVth congress of the International Union of Crystallography. at (1990).
Reyntjens, S. & Puers, R. A review of focused ion beam applications in microsystem technology. J. Micromechanics Microengineering 32, (2021).
Stephenson, N. C. A structural investigation of some stable phases in the region Nb₂O₅.WO₃–WO₃. Acta Crystallogr. Sect. B Struct. Crystallogr. Cryst. Chem. 24, 637–653 (1968).
Krumeich, F., Wörle, M. & Hussain, A. Superstructure and twinning in the tetragonal tungsten bronze-type phase Nb₇W₁₀O₄₇. J. Solid State Chem. 149, 428–433 (2000).
Iijima, S. & Allpress, J. G. Structural studies by high-resolution electron microscopy- tetragonal tungsten bronze-type structures in the system Nb₂O₅-WO₃. Acta Crystallogr. Sect. A 22–29 (1974).
Krumeich, F., Hussain, A., Bartsch, C. & Gruehn, R. Zur präparation und struktur gemischtvalenter niob‐wolframoxide der zusammensetzung (Nb,W)₁₇O₄₇. ZAAC ‐ J. Inorg. Gen. Chem. 621, 799–806 (1995).
Krumeich, F. On the arrangement of pentagonal columns in tetragonal tungsten bronze-type Nb₁₈W₁₆O₉₃. Crystals 11, (2021).
Goldschmidt, H. J. An X-ray investigation of systems between niobium pentoxide and certain additional oxides. Metallurgia 62, 214 (1960).
Kovba, L. M. & Trunov, V. K. On binary oxides containing tungsten, tantalum and niobium. Dokl. Akad. Nauk SSSR 147, 622–624 (1962).
Mohanty, G. P. & Fiegel, L. J. A crystallographic study of Nb₂O₅·3WO₃. Acta Crystallogr. 17, 454–454 (1964).
Yu, J., Han, Y., Jong, H., Jong, H. I. & Ra, G. Two-step hydrothermal synthetic method of niobium-tungsten complex oxide and its adsorption of methylene blue. Inorganica Chim. Acta 507, 119562 (2020).
Dong, S., Wang, Y., Chen, C., Shen, L. & Zhang, X. Niobium tungsten oxide in a green water-in-salt electrolyte enables ultra-stable aqueous lithium-ion capacitors. Nano-Micro Lett. 12, 1–9 (2020).
Krumeich, F. Oxidation products of Nb₇W₁₀O₄₇: A transmission electron microscopy study. J. Solid State Chem. 119, 420–427 (1995).
Krumeich, F. Order and disorder in niobium tungsten oxides of the tetragonal tungsten bronze type. Acta Crystallogr. Sect. B Struct. Sci. 54, 240–249 (1998).
Frevel, L. K. & Rinn, H. W. Powder Diffraction Standards for Niobium Pentoxide and Tantalum Pentoxide. Anal. Chem. 27, 1329–1330 (1955).
Weissman, J. G., Ko, E. I., Wynblatt, P. & Howe, J. M. High-resolution electron microscopy and image simulation of TT-,T-, and H-niobia and model silica-supported niobium surface oxides. Chem. Mater. 1, 187–193 (1989).
Ko, E. I. & Weissman, J. G. Structures of niobium pentoxide and their implications on chemical behavior. Catal. Today 8, 27–36 (1990).
Kato, K. & Tamura, S. Die Kristallstruktur von T-Nb₂O₅. Acta Crystallogr. Sect. B Struct. Crystallogr. Cryst. Chem. 31, 673–677 (1975).
Laves, F., Petter, W. & Wulf, H. Die Kristallstruktur von ζ-Nb2O5. Naturwissenschaften 51, 633–634 (1964).
Kato, K. & Tamura, S. Structure refinement of H-Nb₂O₅. Acta Crystallogr. Sect. B Struct. Crystallogr. Cryst. Chem. 32, 764–767 (1976).
Griffith, K. J., Forse, A. C., Griffin, J. M. & Grey, C. P. High-rate intercalation without nanostructuring in metastable Nb₂O₅ bronze phases. J. Am. Chem. Soc. 138, 8888–8899 (2016).
Lundberg, M. & Sundberg, M. New tetragonal tungsten-bronze-related phases in the pseudobinary system KNb₃O₈-Nb₃O₇F studied by high resolution electron microscopy and X-ray powder diffraction. J. Less-Common Met. 137, 163–179 (1988).
Solid, O. F. High resolution electron microscopy NaNbO₃- Nb₂O₅- WO₃ system. J. Solid State Chem. 84, 23–38 (1990).
Sayagués, M. J., Hutchison, J. L. & Krumeich, F. A new niobium tungsten oxide as a result of an in situ reaction in a gas reaction cell microscope. J. Solid State Chem. 143, 33–40 (1999).

Synthesis and structure characterization. The microspheres of NbWO bronze phases were synthesized using a solvothermal route using Anton Paar Multiwave Pro microwave reactor. In a typical synthesis, NbCl₅ (Alfa Aesar, 99%) and WCl₆ (Acros Organics, 99.9+ %) salts were dissolved in ethanol in the desired ratio, forming a clear solution. The solution was transferred into a Teflon vessel along with a magnetic stirrer bit placed in an alumina jacket. The solution is then solvothermally treated under constant stirring at 150 °C for 2h in the microwave reactor. The precipitate was filtered, washed three times in ethanol, and then dried at 65 °C for 12 h. The precursor was then calcined in an alumina crucible at 800 °C for 12 h to obtain the desired phase.

X-ray diffraction. Powder X-ray diffraction (PXRD) was performed using a Malvern Panalytical Empyrean powder diffractometer equipped with an X’celerator Scientific detector. The patterns were recorded in a reflection geometry using non-monochromated Cu Kα radiation (λ = 1.5406 Å + 1.5444 Å). A nickel filter was used to filter the Kβ radiation prior to the detector. Variable temperature (VT) measurements were performed with a XRK900 (Anton Paar) reactor chamber as an attachment. The amorphous precursor sample was placed on a MACOR disc, heated from room temperature to 900 °C at a heating rate of 10 °C/min, and VT-PXRD patterns (10 – 90° 2θ, step size 0.017° 2θ, counting time 0.13 s step^-1) were recorded from 500 - 900 °C for every 100 °C, with an interval of 90 min between each step. For Rietveld refinement of Nb1, Nb2, and Nb3, the PXRD data were collected at room temperature. The samples were tightly packed on a glass disc placed on a spinning stage, and the data were recorded over a 2-100° 2θ range (step size of 0.008° 2θ, counting time 457.2 s step^-1).

Neutron diffraction. In presence of the heavy scatterers Nb and W, the contribution of O atoms is negligible to the PXRD data. Thus, neutron diffraction was used to aid in the refinement of the position and occupancies of O, since all three elements have a different neutron scattering lengths (b_Nb = 7.054 b_W = 4.86 fm, and b_O = 5.803 fm). The neutron powder diffraction measurements were performed at room temperature on the time-of-flight powder diffractometers POLARIS²⁶ and GEM²⁷ at the ISIS pulsed spallation neutron source in the United Kingdom. Samples were loaded into 6 mm diameter vanadium cans. Data was collected from diffraction banks 1-5 of the GEM for Nb1 and Nb2, and Polaris for Nb3 phases. The data from bank 1 and 4 were used for analysis due to their appropriate d-spacing range (~1.5 – 25 Å, and ~1-3.7 Å). The final structures were determined using a combined Rietveld refinement of PXRD and neutron diffraction data using the FULLPROF suite of programs.²⁸

Scanning electron microscopy - Surface morphology characterization was conducted using a scanning electron microscope (SEM, Tescan MIRA3) operated at 5 kV with a working distance of 6mm. For EDS maps, the operating voltage was 30 kV with a working distance of 15 mm. The images were recorded using the MiraTC software. EDS data were collected using the Aztec software from Oxford instruments. The sample was prepared by sprinkling the ground sample on to a SEM stub with a carbon tape.

X-ray fluorescence measurements - A benchtop X-ray Analytical Microscope (XRF) (HORIBA Jobin Yvon XGT-7000 V) at Diamond Light Source was used to measure the X- Ray Fluorescence point spectra on the samples to determine the composition. A 100 µm beam size and 30 s acquisition time was used to collect the data. The 13 mm pellets prepared using 20-25 mg of the sample with ~60 mg of cellulose were used for XRF measurements. The overestimated W values in the samples were corrected using Nb₁₄W₃O₄₄ as a reference.

Transmission electron microscopy – TEM was used to probe two types of samples: (i) a lamella prepared via ion-beam milling, and (ii) particles dispersed on a lacey-carbon TEM grid. The lamellae were prepared using a focused ion beam scanning electron microscope (FIB-SEM), FEI Helios NanoLab with a Ga source following the procedure described by Reyntjens and Puers.²⁹ TEM images and selected-area electron diffraction (SAED) patterns from these lamellae were acquired via a transmission electron microscope - Thermo Scientific (FEI) Talos F200X G2 operating at 200 kV.

The second set of particle-based samples was prepared by grinding the oxides with ethanol in a mortar and pestle. The suspension was drop casted on a lacey-carbon Cu-based TEM grid. Multiple particles from different batches of samples were studied to ensure their reproducibility. A JEOL ARM200F (Cs-corrected) transmission electron microscope fitted with large solid angle dual EDX detectors (JEOL Centurio) was used in the (S)TEM mode at 200 kV to image and acquire high-resolution EDS data from the NbWO oxide particles.

The TEM grids with the ground particles were loaded on a beryllium double-tilt holder to find particles close to the required zone axis (along the c-axis of the crystal unless explicitly stated). Then α and β tilts of the holder were altered, often iteratively, to align the particles along the zone axis with the help of Kikuchi lines. Atomic resolution STEM-HAADF and EDS datasets were acquired under low-dose conditions to reduce electron beam damage on the specimen: dwell time = 0.01 ms, spot size = 6 C, aperture size = 20 μm, and electron beam current = 18 pA. Multi-frame STEM-EDS data acquisition was carried out due to the low electron beam current and hence the relatively low intensities of the X-rays emitted from the specimen. Each dataset typically consisted of 500-1000 frames. The STEM-HAADF images were analysed using Gatan Microscopy Suite Software and ImageJ.

Electrochemical measurements - The electrochemical performance of the NbWOs were tested in 2032 stainless-steel (ss) coin-type cell assembly with ss conical spring, two 0.5 mm thick ss spacer disks, and a glass microfibre separator. The electrode had an active material:conductive carbon(super P):binder (PVDF) mass ratio of 8:1:1, with an active mass loading of ~2 mg cm^-2 and an area of 1.27 cm². The galvanostatic charge and discharge testing was carried out using a Biologic battery cycler model MPG2 in the voltage range from 3.0 to 1.3 V versus Li⁺/Li separated by a glass fibre separator. 1 M Lithium hexafluorophosphate (LiPF₆) solution in ethylene carbonate (EC)/ dimethyl carbonate (DMC) = 50/50 (v/v) was used as the electrolyte.

CCDC-2291053 (for NbWO_5.5), -2291054 (for Nb₂WO₈) contain supplementary crystallographic data for this paper. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/ data_request/cif.

Yes there is potential Competing Interest. C.P.G. is the co-founder and a major share-holder of a company that seeks to develop fast charging solutions using Nb-based anode chemistries; S.N. is currently an employee of the same company (Nyobolt). S.N. and C.P.G., via Nyobolt ltd., have filed a UK patent application (GB2218984.9) covering the materials and high-rate energy storage application described in this manuscript.

NbWOBronzeSI.docx
Supplementary Information and Extended data

Download PDF

Version 1

posted

You are reading this latest preprint version

Cation-ordering in low-temperature niobium-rich NbWO bronzes: New anodes for high rate Li-ion batteries

Status:

Version 1

Abstract

Figures

Introduction

Results

Nb1: NbWO_5.5

Exploration of the Nb-rich NbWO phases

Crystal structure of Nb2

Antiphase boundaries (APB) and structural evolutions of the bronzes

Crystal structure of Nb3

Thermodynamic Stability

W-rich NbWO phases

Electrochemical performance of the low temperature phases

Conclusions

Declarations

References

METHODS

Additional Declarations

Supplementary Files

Status:

Version 1

Cation-ordering in low-temperature niobium-rich NbWO bronzes: New anodes for high rate Li-ion batteries

Status:

Version 1

Abstract

Figures

Introduction

Results

Nb1: NbWO5.5

Exploration of the Nb-rich NbWO phases

Crystal structure of Nb2

Antiphase boundaries (APB) and structural evolutions of the bronzes

Crystal structure of Nb3

Thermodynamic Stability

W-rich NbWO phases

Electrochemical performance of the low temperature phases

Conclusions

Declarations

References

METHODS

Additional Declarations

Supplementary Files

Status:

Version 1

Nb1: NbWO_5.5