Electrochemical Modification of Negative Thermal Expansion

Aug 22, 2018 - School of Chemistry, UNSW Sydney, Sydney , New South Wales 2052 , ... and thermal treatment can be used to synthesize new compounds...
0 downloads 0 Views 2MB Size
Article Cite This: Inorg. Chem. XXXX, XXX, XXX−XXX

pubs.acs.org/IC

Electrochemical Modification of Negative Thermal Expansion Materials in the TaxNb1−xVO5 Series Sunny Wang, Damian Goonetilleke, and Neeraj Sharma* School of Chemistry, UNSW Sydney, Sydney, New South Wales 2052, Australia

Downloaded via EASTERN KENTUCKY UNIV on August 23, 2018 at 03:59:12 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

S Supporting Information *

ABSTRACT: Electrochemical processes transfer charge carriers to and from electrodes, e.g., Li+ ions are inserted into anodes during discharge and extracted during charge in a Li half-cell. Using an electrode that features negative thermal expansion (NTE) properties in an electrochemical cell allows a means to study the interaction of the charge carrier with an NTE material and potentially modify or tune its NTE properties. This work examines the NTE properties of TaxNb1−xVO5 (x = 1, 0.9, 0.75, 0.5, 0.25) and the effect of Li+/Na+/K+ electrochemical discharge of TaVO5-based electrodes. Sodium discharge was found to drastically alter NTE properties with 25% Na+ discharged electrodes exhibiting a linear volumetric coefficient of thermal expansion of −5.75 ± 0.20 × 10−5 Å3/°C between 350 and 500 °C, one of the largest reported for any NTE system. Furthermore, at higher temperatures, the Na+- and K+-discharged and heated electrodes generate new phases, suggesting that a combination of electrochemical discharge and thermal treatment can be used to synthesize new compounds. This work lays the foundation for the concept of using electrochemical discharge followed by subsequent thermal treatments to modify the physical properties of a compound.

1. INTRODUCTION Negative thermal expansion (NTE) materials exhibit net contraction upon heating. This unusual characteristic has made these materials a target of great interest due to the possibility of modifying their thermal expansion properties to create zero thermal expansion materials or materials with tunable thermal expansion properties.1−7 Materials which do not exhibit significant volume change as a function of temperature are desirable for application in various electronic and coating technologies where thermal stability of properties, such as refractive index or dielectric constant, is beneficial for consistent performance.2,4,8 Since isotropic NTE over wide temperature ranges was initially reported in cubic ZrV2xPxO79 and ZrW2O8,10 open framework ceramic oxides have been among the most studied subset of NTE materials. The structure of these materials typically consists of large networks of polyhedra (e.g., MO4 tetrahedra and MO6 octahedra where M is a metal or combination of metals) with prominent interstitial or crystallographic voids. The mechanism of NTE in such materials has been suggested to arise through transverse oxygen vibrations coupled with cooperative lattice movements.1,11,12 A pervasive yet simplified approach to consider these vibrations is the rigid unit modes (RUMs) model which has been applied extensively to study NTE behavior in materials such as ZrW2O8.11−14 Such a model assumes polyhedron distortions (anisotropic changes in bond lengths within the polyhedra during heating or cooling) to be energetically unfavorable in comparison to cooperative polyhedron rotations due to the strength of metal−oxygen bonds.12 Upon heating, transverse vibrations of © XXXX American Chemical Society

M−O−M bonds at the polyhedron hinges induces a concerted tilting of multiple polyhedron units. The subsequent shrinkage of the crystallographic voids allows the lattice to achieve a more compact configuration. If the concerted polyhedron movements responsible for NTE can be sterically manipulated, for example by the occupancy of the crystallographic voids by ions, then in principle, it may be possible to modify NTE. Such an approach has been explored in organometallic frameworks. For example, hydrated Prussian Blue structures exhibit marginally reduced NTE and solvent-assisted intercalation of Li+ into ScF3 allows the effective tuning of thermal expansion from NTE to positive thermal expansion (PTE).5,15 It is postulated that such a method should also be applicable to ceramic oxides using electrochemical intercalation much like in lithium or sodium ion batteries.16 This work examines the NTE properties of the isostructural TaxNb1−xVO5 series of compounds and the effect of Li+/Na+/ K+ discharge into TaVO5. The end-member species TaVO5 and NbVO5 are known to exhibit NTE, with reported volumetric coefficients of thermal expansion (CTEs) being −8.92 × 10−6 and −6.63 × 10−6 Å3/°C, respectively, between 20 and 600 °C.17,18 In this temperature range, these materials adopt an orthorhombic crystal structure with space group Pnma.19 The lattice consists of TaO6 and NbO6 octahedra alongside VO4 tetrahedra connected by corner sharing oxygens which bind large pentagonal tunnels along the (010) axis (see Figure S1). Received: May 10, 2018

A

DOI: 10.1021/acs.inorgchem.8b01280 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

lattice parameters for intermediate Ta:Nb solid solution compositions.22 The results are consistent with fractional occupancies obtained from Rietveld refinement and semiquantitative analysis via ICP-MS (Tables S1 and S2). The NTE properties of intermediate Ta x Nb 1−x VO 5 compositions have not been characterized in the literature. In laboratory-based variable-temperature (VT) experiments, the absolute 2θ shift and sample displacement parameters were calibrated internally and externally using the platinum heating plate as the reference.23 The TaxNb1−xVO5 compounds exhibit anisotropic NTE and the coefficient of thermal expansion (CTE) calculated between 100 and 600 °C show that the extent of NTE decreases approximately linearly with niobium doping (Figure 1b). Examination of each unit cell axis shows that this is largely accounted for by the contraction along the b and c axes illustrating an anisotropic contribution of the unit cell axes to the NTE across the series (Table S3). It is noted that the CTE of pristine TaVO5 obtained from the flat plate XRD measurements in the laboratory (−1.11 ± 0.03 × 10−5 Å3/°C) is statistically identical to that of TaVO5 electrodes from capillary-based synchrotron XRD measurements (−1.14 ± 0.04 × 10−5 Å3/°C). This suggests that the addition of carbon black and polyvinylidene fluoride (PVDF) in electrodes does not affect the NTE properties of TaVO5. NPD data was also collected on the pristine TaVO5 with illustration of the refined structure with thermal ellipsoids shown in Figure S1 and the corresponding atomic parameters shown in Table S4. The shape of the ellipsoids may indicate the polyhedral movements within the structure. 2.2. Structural Change of TaVO5 Electrodes Following Electrochemical Discharge. Electrochemical discharge curves in Li+/Na+/K+ cells are shown in Figure S4 and the structural changes upon discharge were found to be similar across the TaxNb1−xVO5 series, see Figure S5. We focus on the parent TaVO5 material for full characterization. Ex situ synchrotron XRD data on extracted TaVO5 electrodes at various states of discharge in Li+/Na+/K+ half cells are shown in Figure 2. In this work, an n% discharged electrode refers to a TaVO5 electrode that has been discharged to n% of its first discharge capacity and then extracted from the cell. The first discharge capacities of TaVO5 electrodes in Li+, Na+ and K+ half cells were 406, 213, and 195 mAh/g, respectively. Note that electrochemical charging was not considered in this work. This work focuses on the first electrochemical discharge or first electrochemical insertion process with a TaVO5 electrode, in other words where the Li + , Na + , and K + from the corresponding electrolyte or metal electrode are inserted or reacted with the TaVO5 electrode. This reaction is allowed to proceed to various states of completion, where completion can be considered as 100% discharged. Note the charge process which is to remove the inserted Li+, Na+, and K+ from the inserted TaVO5 electrode is not examined in this work because the first electrochemical charge differs from the first electrochemical discharge due to irreversible reactions and typically the formation of surface layers; see Figure S6. Note, thermal studies can be performed for the discharged−charged electrodes that are extracted at various percent of charge, but whether this follows the same thermal evolution as the corresponding percent of discharge is unlikely due to the irreversible nature of the electrochemistry. However, such studies are slated for future work. Previous intercalation studies on NbVO5 found that the active material undergoes amorphization upon lithiation.21 At

The presence of large pentagonal voids with readily reducible vanadium(V) makes TaxNb1−xVO5 a promising host lattice for ion insertion reactions and have been examined by Amarilla and co-workers in the context of insertion electrodes for lithium ion batteries.20,21 First, discharge capacities in TaxNb1−xVO5 compounds were reported to increase with Nb substitution ranging from 135 mAh/g for TaVO5 to 233 mAh/g for NbVO5 electrodes when cycled between 1.5 and 4 V. Significant capacity loss occurs in latter cycles which was attributed to the structural degradation of the electrodes.

2. RESULTS AND DISCUSSION 2.1. Synthesis and XRD Studies on TaxNb1−xVO5 Materials. The sintering temperatures required for materials in the TaxNb1−xVO5 series decrease successively with increasing niobium content. TaVO5 and Ta0.9Nb0.1VO5 can be synthesized at 800 °C, but compounds such as Ta0.25Nb0.75VO5 form TaxNb9−xVO25 impurities when sintered above 700 °C (Figure S2). The intention of this work was to synthesize the TaxNb1−xVO5 series for the electrochemical− thermal studies using the same method. Structural characterization of all the compounds synthesized can be found in the Supporting Information. Figure 1a depicts the unit cell volume across the TaxNb1−xVO5 series obtained from Rietveld analysis (see Figure S3). Substitution of niobium onto the tantalum site was confirmed via Vegard’s Law which predicts a linear shift in

Figure 1. (a) Unit cell volume of the TaxNb1−xVO5 series at ambient temperature (R2 = 0.966 for the line of best fit) and (b) CTEs of the TaxNb1−xVO5 series between 100 and 600 °C. See Table S5 for refined values. B

DOI: 10.1021/acs.inorgchem.8b01280 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 3. Rietveld refined fit using the TaVO5 structural model with XRD data of the 25% Na+ discharged TaVO5 electrode, WR = 14.898%.

Overall, it appears that Li+ discharge has a larger incidence of amorphization after first discharge compared to that of Na+ and K+ discharge which may be correlated to ion size. 2.3. VT Studies on 100% K+ Discharged TaVO5 Electrodes. As previously mentioned, the parent TaVO5 material exhibits no change of the CTE between 100 and 600 °C when incorporated into an electrode, i.e., with the addition of PVDF and carbon black components (see Figure 1b). The high-temperature behavior of the 100% discharged electrode in a potassium half-cell is shown in Figure 4 with the Figure 2. XRD patterns of TaVO5 electrodes at (a) 100% K+, (b) 25% Li+, and (c) 25% Na+ discharge. Fourier density maps calculated from Rietveld analysis using the TaVO5 model are inserted in panels a and c. The dashed red lines mark the expected positions of reflections corresponding to TaVO5 at room temperature.

the 25% lithium discharged state, TaVO5 also becomes partially amorphous or nanocrystalline (see Figure 2b). This change is further exasperated upon complete discharge (see Figure S7a). However, in contrast to lithium discharged samples, ex situ XRD patterns of potassium discharged samples exhibit no change even upon 100% discharge. Fourier difference maps calculated from the Rietveld model and the synchrotron XRD data also exhibit little difference to the pristine TaVO5 model. With the exception of rare zero-strain electrodes,24,25 ion insertion into the NTE voids (if this does occur) should be accompanied by structural changes which can be detected by shifting reflections, changing intensities of existing reflections, the appearance of new reflections, or a combination of these changes. As such, the results indicate K+ insertion during discharge if it occurs at all is minimal. Similar to lithium discharged samples, complete discharge of sodium-based half cells results in significant amorphization of TaVO5 electrodes. However, some TaVO5-type reflections still appear to be present (see Figure S7b). Interestingly, at the 25% sodium discharged state, some degree of crystallinity is retained alongside the formation of several new reflections which may indicate a coexistence of multiple phases. A notable change is the increase in the relative intensity of what appears to be the (011) TaVO5-type reflection (* in Figure 2c). The Rietveld refined fit using the TaVO5 structural model (Figure 3) clearly depicts the presence of at least one further phase.

Figure 4. 100% K+ discharged electrode. (a) Unit cell volume (25− 400 °C) and (b) selected XRD patterns (25−800 °C).

volume evolution of the TaVO5 electrode between 25 and 400 °C in Figure 4a. The CTE between 25 and 250 °C (−1.46 ± 0.07 × 10−5 Å3/°C) is statistically identical to that of the uninserted TaVO5 (−1.43 ± 0.02 × 10−5 Å3/°C). This is consistent with the diffraction data of the 100% K+ electrode which remains unchanged compared to the unmodified TaVO5 electrode (see Figure S8). These observations suggest that there is no potassium insertion into the TaVO5 structure. However, above 300 °C, there is a transition to PTE which is accompanied by a significant loss of reflection intensity. Complete amorphization is achieved upon heating between 500 and 550 °C but is followed by the formation of several new reflections which remain even after quenching to room temperature. Attempts to identify the new phases yielded no matches across all ICDD and ICSD databases, indicating the products formed have yet to be reported. This transformation C

DOI: 10.1021/acs.inorgchem.8b01280 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

accompanied by a decline in the TaVO5 reflection intensity. The strong NTE exhibited by the electrode in this region is of particular interest as the calculated volumetric CTE of −5.75 ± 0.20 × 10−5 Å3/°C is nearly twice that observed for ZrW2O810 and over 5 times larger than the parent material. This magnitude of contraction is comparable to the strongest reported NTE materials; see Table 1. It remains to be seen whether the NTE behavior is reversible upon cooling as further heating beyond the NTE temperature range results in an irreversible phase transition. However, the results demonstrate sodium discharge can drastically alter the NTE properties of TaVO5. It is possible that a lower degree of sodium insertion may lead to less erratic phase transitions and this is slated for future work as is the evaluation of the reversibility of the strong NTE within the 300−500 °C range. Further heating past 500 °C, results in the complete amorphization of the electrode which is followed by the formation of new reflections above 600 °C. Search and match algorithms were used to identify two potential phases present: a dominant phase that is isostructural to CaTa0.67V0.33O233 and a rutile-like phase of Ta0.5V0.5O234 (Figure 6). Both phases themselves are interesting, owing to the low oxidation states of tantalum and vanadium. The CaTa0.67V0.33O2-type phase was interpreted to be NaxTa0.67V0.33O2 due to the presence of sodium, but no reports of its synthesis are found in the literature. It remains to be seen whether these structures are correct especially given the low oxidation states of tantalum and vanadium in the proposed structures. However, the ascollected peak positions show a good correlation with expected reflection positions from these models. As with the potassium discharged electrodes, the phase transitions are not found to be reversible upon cooling to room temperature. Therefore, the formation of new phases upon heating discharged electrodes may suggest a combined electrochemical−thermal approach allows for synthetic pathways to new materials. In order to obtain initial indications of whether the measured NTE between 300 and 500 °C is reversible, a VT laboratory XRD experiment was performed. Data were analyzed based on the 111 TaVO5-type reflection as shown in Figure S9 of an electrode lower than the 25% Na discharged state. The behavior on heating from 300 to 500 °C for this reflection does not repeat on cooling from 500 to 300 °C. It provisionally appears that heat treatments after electrochemical discharge are not reversible on cooling (within the same structure type) at the very least at this level of discharge and using the laboratory XRD setup. Furthermore, there is evidence of positive thermal expansion on heating and contraction on cooling with this setup. The differences noted in the thermal evolution can be related to the conditions of the experiment, where an open system is used, exposure to air occurs and the atmosphere is N2 which differs from a capillary sealed in an Ar glovebox. Nonetheless, the different thermal evolution on heating and cooling under these conditions may indicate that the observed NTE within a capillary on heating from 300 to 500 °C may not be reversible on cooling and will be explored in detail in future work.

can be attributed to sintering reactions between any surface deposited potassium or electrolyte decomposition products with the TaVO5 electrode, and the interaction of these products with the PVDF binder and carbon black and their thermal evolution. It should be noted that the newly formed diffraction peaks exhibit some peak broadening compared to the original TaVO5 reflections which indicate a smaller particle size distribution. 2.4. VT Studies on 25% Na+ Discharged TaVO5. Unlike the K+ discharged electrodes, significant phase transitions occur in the 25% sodium discharged electrodes during thermal treatment. Between 25 and 150 °C, the modeled TaVO5-type phase exhibits strong PTE with a CTE of 6.01 ± 1.01 × 10−5 Å3/°C (see Figure 5) although a second unmodeled phase is

Figure 5. 25% Na+ discharged electrode with (a) refined unit cell volumes (25−500 °C), (b) selected XRD patterns (25−500 °C), and (c) Rietveld refined fit of the TaVO5 structural model to the 400 °C data set, WR = 10.125%. *, reflections corresponding to an additional phase; however, the intensity of these reflections remains constant throughout the VT experiment.

3. CONCLUSIONS High-purity TaxNb1−xVO5 compounds were synthesized via solid state heating with niobium substitution onto the tantalum site demonstrated via X-ray diffraction, Rietveld refinement and elemental analysis (ICP-MS). The NTE properties of intermediate oxides were quantified for the first time with all

clearly present. Further heating above 200 °C results in the transformation of the multiple-phase electrode into a single TaVO5-type phase and a correlated evolution back to NTE between 300 and 500 °C (see Figure 5c for a Rietveld fit of the 400 °C data set showing an effectively pure TaVO5-type phase). The NTE period between 300 and 500 °C is D

DOI: 10.1021/acs.inorgchem.8b01280 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry Table 1. Comparison of CTEs Reported for Various Strong NTE Systemsa material

CTE (αv)

temperature range (°C)

isotropic

ref

Na discharged TaVO5 AlPO4-17 Sc2(WO4)3 Y2W3O12 ZrW2O8 [Mn0.96Fe0.043]Zn0.5Ge0.5N CuO nanoparticles MOF-5 Cd(CN)2

−5.75(20) × 10−5 −3.51 × 10−5 -6.60 × 10−6 −2.10 × 10−5 −2.60 × 10−5 −7.50 × 10−5 −1.10 × 10−4 −3.93(3) × 10−5 −1.005(3) × 10−4

350−500