Tunable Mechanochemistry of Lithium Battery Electrodes - ACS Nano

Jun 2, 2017 - In the case of thin films deposited on substrates with dissimilar coefficients of thermal expansions (CTE), significant interface strain...
1 downloads 5 Views 2MB Size
Tunable Mechanochemistry of Lithium Battery Electrodes Nitin Muralidharan,†,‡ Casey N. Brock,†,‡ Adam P. Cohn,‡ Deanna Schauben,‡ Rachel E. Carter,‡ Landon Oakes,†,‡ D. Greg Walker,‡ and Cary L. Pint*,‡ †

Interdisciplinary Materials Science Program and ‡Department of Mechanical Engineering, Vanderbilt University, Nashville, Tennessee 37235, United States S Supporting Information *

ABSTRACT: The interplay between mechanical strains and battery electrochemistry, or the tunable mechanochemistry of batteries, remains an emerging research area with limited experimental progress. In this report, we demonstrate how elastic strains applied to vanadium pentoxide (V2O5), a widely studied cathode material for Li-ion batteries, can modulate the kinetics and energetics of lithium-ion intercalation. We utilize atomic layer deposition to coat V2O5 materials onto the surface of a shapememory superelastic NiTi alloy, which allows electrochemical assessment at a fixed and measurable level of elastic strain imposed on the V2O5, with strain state assessed through Raman spectroscopy and X-ray diffraction. Our results indicate modulation of electrochemical intercalation potentials by ∼40 mV and an increase of the diffusion coefficient of lithium ions by up to 2.5-times with elastic prestrains of 1) results in structurally irreversible phase transformations (γ and ω phases), which were avoided in our work.12,41 These phase transformations occurring during the Li+ ion intercalation process are accompanied by structural changes (strains) to the layered V2O5 crystal structure. Therefore, we expect that prestraining V2O5 can contribute to changes in the energetics of the intercalation reaction. Figure 4a shows the cyclic voltammograms of the unstrained (0%) and strained (−-0.35%, −0.5% based on ‘c’ direction) at a scan rate of 30 mV/s normalized to the maximum peak current. The anodic and cathodic peaks corresponding to the reversible transformations α ↔ ε and ε ↔ δ were observed to shift to lower voltages (also see Supporting Information, Figure S9). We attribute the shift in intercalation potentials to the property changes caused by the elastic strains on the film (also see Supporting Information, Figures S10 and S11). The effect of

ΔEeq(Strained) = −

ΔGStrained ΔU = − Strained nF nF

(2)

where n is the number of electrons transferred and F is the Faraday constant (Figure 4c). Simultaneous modulation of XRD peaks, Raman shifts, and electrochemical potentials is a result of changes to the total internal energy of the metal oxide films. In particular, only stored elastic strain in our system can alter the internal energy landscape of the V2O5 film.31,32 Ab initio simulations using DFT were performed to determine the 6246

DOI: 10.1021/acsnano.7b02404 ACS Nano 2017, 11, 6243−6251

Article

ACS Nano

Figure 4. (a) Normalized cyclic voltammograms of the unstrained and strained states of the surface V2O5 film at a scan rate of 30 mV/s. (b) Average equilibrium potential variation with imposed strain on the surface film. (c) Schematic representation of the energetics governing the electrochemistry of strained and unstrained states of V2O5. (d) Intercalation energy of the Li+ ion intercalation process arising from in-plane prestraining the V2O5 lattice to ±2% determined using DFT simulations.

Figure 5. (a) Diffusion coefficient ratio DStrained/DUnstrained variation with both ‘c’ strain and corresponding in-plane strain owing to Poisson ratio, v = 0.3. (b) Schematic representation of the enhancement in diffusion coefficient arising due to a prestrained V2O5 electrode.

change in intercalation energy, EI, between a prestrained state of V2O5 undergoing lithiation compared to an unstrained state (see Supporting Information, Figure S16). This intercalation energy is a measure of the change in total energy of the system when a single lithium atom is removed from BCC lithium metal and placed inside the V2O5 lattice. We can model applied strain by changing lattice parameters and assessing how this affects EI. A change in intercalation energy of ∼75 meV (∼75 mV) was calculated with an applied in-plane strain of +1.5% on the V2O5 lattice (Figure 4d), comparable to the measured change in potential of ∼40 mV at about ∼1.66% in-plane prestrain in Figure 4b. These data corroborate the fact that elastic strains can be an input control parameter to alter the energetics of Li+ ion intercalation electrodes. Additionally, this model can be used to predict changes in intercalation potentials over a wide range of applied in-plane strains.

An important parameter in the study of intercalation electrodes such as V2O5 is diffusion of the ions through the crystal lattice. By determining potential dependent ‘b’ values (see Supporting Information, Table S17) from the relationship, i = aνb, where “i” is the peak current, “ν” is the scan rate, “a” and “b” are adjustable parameters, we observe a combination of diffusion controlled and intercalation pseudocapacitive processes at these peak voltages.20,26,56 On the basis of Randle Sevcik analysis (see Supporting Information, Figure S18), the diffusion coefficient of Li+ ions (see Supporting Information, Table S19) during the different phase transformations was determined.12,57 Owing to the sub-100 nm thickness of these strained V2O5 films on NiTi alloy, diffusion coefficients in the order of 10−9 to 10−10 cm2 s−1 were observed for all the phase transformations studied in this work.58−61 The diffusion coefficient ratio between the strained and unstrained states of the V2O5 films is given in Figure 5a. Because of Poisson’s ratio 6247

DOI: 10.1021/acsnano.7b02404 ACS Nano 2017, 11, 6243−6251

Article

ACS Nano (assuming ∼0.3 in our case), a net positive volume change is expected in the film. An out of plane compressive strain results in in-plane tensile strains, which effectively increase the total volume of the crystal (Figure 5b). In accordance with this, the diffusion coefficient in the strained state is observed to increase to about 2.5-times when compared to the unstrained state. It can also be noted that as Li+ ion intercalation in V2O5 is accompanied by phase transformations, elastic prestrains can have an effect on the transformational strain6 produced due to diffusion of Li+ ions resulting in significant variation between the diffusion coefficient ratios. On the basis of previous reports using simulated studies of intercalation cathodes, it can be implied that in our system, the increase in diffusion coefficient could be due to the reduction of ion migration barriers, which can be achieved through imposed mechanical strains.34,35 The results presented here highlight the principle that tunable mechanochemistry can be a tool to modulate the energetics and kinetics during ion intercalation reactions in energy storage electrode materials. Going forward with this approach where V2O5 is only one of many host electrode options, strain engineering can be further applied to emerging research areas focused on storage using alternative ions (Na+ and K+)15,23,62 and multivalent ions (Mg2+ and Ca2+).63,64 The ion size of these alternative materials remains a challenge, but mechanical strains could offset the insertion barriers for these systems and be an important tool for moving beyond lithium. In addition, because strain engineering is a widely applied principle in solid-state semiconductor electronics, it could be integrated with solid-state battery architectures developed on thin film substrates as well as next-generation microbatteries. In any case where cutting-edge battery research efforts have focused efforts on analyzing mechanical strains as a byproduct of intercalation chemistry, our work provides a viewpoint of the tunable interplay between mechanics and chemistry of battery electrodes toward development of efficient energy storage devices.

Acetone (Aldrich) followed by Ethanol (Aldrich) and nanopure water (Millipore water purifier). The wires were then air-dried prior to further processing. This aging process was used to activate the superelastic/shapememory property of the NiTi alloy. Atomic Layer Deposition of Vanadium Oxide Films and Annealing Treatments. Vanadium oxide coatings were deposited on the aged, mechanically polished NiTi wires using a GemStar 6″ ALD system. The precursor, (98+%) Vanadium(V)tri-i-propoxy oxide (VTIP) with a chemical formula, VO(OC3H7)3 (Strem Chemicals), was preheated to 55 °C. The oxidizer used in this process was nanopure water (Millipore water purifier). Heating the precursor and oxidizer manifolds to 115 °C ensured that no condensation of the reagents occurred inside the manifolds. The carrier gas used was ultrahigh pure Argon and a reaction temperature of 200 °C was maintained throughout the process to ensure uniform deposition on the surface of the NiTi wire. VTIP and water pulses of 2 s each and long residence times of 20 s for both reactants in the reaction chamber were employed to achieve complete saturation of the precursors in the reaction chamber, similar to previous reports.43,48 The wires were subjected to repeated ALD cycles to achieve a thin conformal coating of vanadium oxide on the surface. The deposited oxide films were annealed at 450 °C for 30 min in a Lindberg Blue 1″ Quartz CVD tube furnace open to air to achieve crystalline coatings of vanadium pentoxide (V2O5) on the surface of the alloy. Surface Ellipsometry Analysis and Mass Estimations. To determine the thickness and mass of the oxide deposited per cycle, a silicon wafer (Diameter: 100 mm) was used along with the NiTi wires during the ALD process. Surface ellipsometry analysis was performed on these silicon wafers after annealing treatments using a JA Woollam M2000VI Spectroscopic Ellipsometer. The mass of the wafer before and after the ALD process post annealing treatments gives an estimation of the mass of oxide deposited per cycle. This provides a comparable estimate of thickness and mass of the thin vanadium pentoxide films deposited on NiTi wires. Straining the Surface Coating. The NiTi wires with the surface V2O5 coatings were subjected to axial tensile deformation up to 5% and 10% strains at a rate of 2 mm/min using an Instron 5944 mechanical testing system. Tensile deformations beyond 10% were not performed to avoid delamination of the coating from the wire surface. Electron Imaging. Microstructural characterization and energy dispersive spectroscopy (EDS) analysis were carried out using a Zeiss Merlin scanning electron microscope using a 5 kV acceleration voltage for imaging and 20 kV beam voltage for EDS elemental analysis and mapping. Raman Spectroscopy and X-ray Diffraction. Analysis of the “locked-in” elastic strains on the surface V2O5 coatings was performed using XRD and Raman spectroscopy. Raman spectroscopy was carried out using a Renishaw Raman microscope using 532 nm laser excitations. Statistical analysis of the strained surface coating was obtained through maps composed of >500 spots across the surface of the wire. The laser exposure time was 60 s at power of 10% to avoid thermal effects. Lorentzian fits were applied to the individual stretch modes to obtain peak positions. Crystallographic analysis to quantify the strain on the surface V2O5 coating was performed by obtaining Xray diffractograms using a Scintag XGEN 4000 using Cu Kα 1.542 Å. To maximize the X-ray counts, a 40 s exposure time per 0.02 degree increment was maintained throughout the measurement. Electrochemical Measurements. Electrochemical measurements and analysis were performed in a three-electrode configuration using a beaker-type cell. The V2O5 coating on NiTi wire was used as the working electrode with a platinum foil (Alfa Aesar 1 cm × 1 cm) as the counter electrode and a saturated calomel electrode (SCE) as the reference electrode. The electrolyte used in this study was freshly prepared 1 molar lithium perchlorate (LiClO4) (Aldrich) dissolved in propylene carbonate (PC) solvent (Aldrich). The electrochemical data were normalized to the immersed wire electrode area in the electrolyte. Cyclic voltammograms and galvanostatic charge−discharge curves were obtained for the unstrained and strained states of the surface V2O5 coating at various scan rates (10−100 mV/s) in the

CONCLUSION In summary, we provide insight into the mechanochemistry of battery electrodes or more generally the ability to modulate battery electrochemistry using tunable mechanical strains. We leverage ALD to deposit nanoscale coatings of V2O5 cathodes onto superelastic shape memory NiTi current collectors that are “locked in” at various strain states for electrochemical testing. Our findings indicate modulation of the intercalation potentials by ∼40 mV and improving the diffusion coefficient of lithium ions in V2O5 by around 2.5-times for applied strains