Enhanced Cycling and Rate Capability by Epitaxially Matched Conductive Cubic TiO Coating on LiCoO2 Cathode Films.

Layered lithium transition-metal oxides, such as LiCoO2 and its doped and lithium-rich analogues, have become the most attractive cathode material for current lithium-ion batteries due to their excellent power and energy densities. However, parasitic reactions at the cathode-electrolyte interface, such as metal-ion dissolution and electrolyte degradation, instigate major safety and performance issues. Although metal oxide coatings can enhance the chemical and structural stability, their insulating nature and lattice mismatch with the adjacent cathode material can act as a physical barrier for ion transport, which increases the charge-transfer resistance across the interface and impedes cell performance at high rates. Here, epitaxial engineering is applied to stabilize a cubic (100)-oriented TiO layer on top of single (104)-oriented LiCoO2 thin films to study the effect of a conductive coating on the electrochemical performance. Lattice matching between the (104) LiCoO2 surface facets and the (100) TiO plane enables the formation of the titanium mono-oxide phase, which dramatically enhances the cycling stability as well as the rate capability of LiCoO2. This cubic TiO coating enhances the preservation of the phase and structural stability across the (104) LiCoO2 surface. The results suggest a more stable Co3+ oxidation state, which not only limits the cobalt-ion dissolution into the electrolyte but also suppresses the catalytic degradation of the liquid electrolyte. Furthermore, the high c-rate performance combined with high Columbic efficiency indicates that interstitial sites in the cubic TiO lattice offer facile pathways for fast lithium-ion transport.

Introduction

As society is transiting toward sustainable and emission-free mobility, demands have risen for batteries with improved energy and power capacities, combined with enhanced safety and extended cycle life. Therefore, layered lithium transition-metal oxides (such as LiCoO2 and its doped and lithium-rich analogues), exhibiting excellent power and energy densities, have become the best cathode material for current lithium-ion batteries used in a wide variety of applications ranging from thin-film micro-batteries to large-scale traction batteries.[1] However, the electrochemical performance of these layered transition-metal oxides is intrinsically limited by the lattice distortions from a hexagonal phase to a monoclinic phase during repetitive charge–discharge cycling and the detrimental anionic reactions under overcharging conditions.[2−5] As overcharging introduces lithium vacancies in the two-dimensional transition-metal layers, it leads to an order/disorder phase transition.[6,7] This continuous phase transition causes a gradual oxygen loss from the cathode material, which in turn deteriorates the electrochemically active phases and corresponding kinetic properties over extensive cycling.[8] This loss of oxygen changes the valence state of the cobalt atoms from Co3+ to Co2+ at the cathode–electrolyte interface, resulting in parasitic reactions, such as metal-ion dissolution and electrolyte degradation,[9,10] which instigate major safety and performance issues.[7,11,12]

To overcome these inherent issues in LiCoO2-type cathode materials, various strategies have been explored, including nanostructuring,[13,14] crystal engineering,[15−17] cation doping,[18] surface coating,[19,20] and electrolyte additives.[21,22] Thus far, the application of a thin-film coating on the surface of the active cathode material has turned out to be the most successful strategy to limit the cathode–electrolyte reactivity and to enhance the cycle life of the battery.[23,24] These metal oxide coatings (e.g., Al2O3, TiO2, ZrO2, or SiO2) are believed to form a solid solution interphase layer (e.g., LiAlCo1–O2 for Al2O3 coating) at the surface of the layered transition-metal oxide, which will passivate the cathode surface and enhance the interfacial kinetics for lithium transport.[25,26] Furthermore, it was also suggested that such coatings mitigate the order/disorder transition and oxygen losses from the cathode lattice as well as other parasitic reactions, such as HF evolution due to electrolyte decomposition and the irreversible growth of solid electrolyte interphase (SEI) layers.[5,19,27−29] However, the exact nature and contribution of these surface coatings are not fully understood and still under debate.[24]

Another strategy to enhance the cycle life and safety of lithium-ion batteries is to combine successful LiCoO2-type cathode materials with a solid electrolyte. Although solid electrolytes are inherently safer as compared to liquid electrolytes, their limited stability against such high-voltage cathode materials remains a major bottleneck.[30−32] Furthermore, the combination of anisotropic lithium transport in such layered LiCoO2-type materials, slow lithium diffusion within many solid electrolytes, and the incremental formation of a space charge layer across the cathode–electrolyte solid–solid interface limits such all-solid-state battery concepts for usage in large-scale and high-power applications.[32−34] Interestingly, previous studies have shown that a heterogeneous interfacial coating at the solid–solid cathode–electrolyte interface can suppress the mixed electronic state and improve the lithium diffusion behavior.[32,35−38]

Although metal oxide coatings on cathode materials enhance the chemical and structural stability of liquid- and solid-electrolyte-based battery systems, these coatings are intrinsically electrically insulating (e.g., Al2O3, TiO2, ZrO2, or SiO2) and commonly exhibit a lattice mismatch with the adjacent cathode material. This lattice mismatch between the crystal structures of the cathode and coating materials can act as a physical barrier for ion transport, which increases the charge-transfer resistance across the interface and impedes the overall cell performance.[33,39−41]

Epitaxial engineering can be used to control the crystal orientation within thin-film model systems, which enables a unique insight into the relation between electrochemistry and crystal directionality of such chemically complex interfaces, not obtainable in single crystals or polycrystalline samples. Previous thin-film studies have demonstrated the optimization of lithium transport for epitaxial LiCoO2-type cathode systems by controlling the specific crystal orientation of the facets [e.g., (001), (110), or (104)] at the interface toward the electrolyte.[42−45] Although these crystal facets exhibit the lowest surface energy among all nonpolar and electrochemical active facets, the multiple electronic states at the surface govern their reaction kinetics with the electrolyte and dominate their electrochemical performance.[38,46,47] Therefore, metal oxide coatings still play an important role in improving the electrochemical performance of LiCoO2-type thin-film model systems. Many thin-film studies have investigated the effect of an additional amorphous or polycrystalline coating; however, only a few studies have explored the impact of an epitaxially matched layer (ZrO2, BaTiO3) on the surface of a highly crystalline LiCoO2 thin film.[24,48] Therefore, a large knowledge gap currently exists in how the alignment of the crystal structures across such epitaxial coating–cathode interface can enhance lithium transport while preventing parasitic reactions with the adjacent electrolyte.

In this study, we focus on the epitaxial engineering of an ionic and electronic conducting TiO (x = 0.7–1.25) coating at the interface with a single-oriented LiCoO2 thin film. The cubic TiO phase consists of approximately 15% randomly arranged cation and anionic vacancies in a defect rock salt structure and typically exhibits metallic behavior.[49−52] Furthermore, such oxygen vacancies in oxide materials have shown enhanced lithium transport and storage properties.[47,53−55] The cubic titanium mono-oxide (TiO) is normally stabilized under high temperature and pressure conditions,[50] while in our case, the lattice matching with the underlying (104) LiCoO2 crystal facet has enabled us to realize a full epitaxial thin-film model system. Detailed analysis in a half cell against lithium metal has shown dramatic enhancement of the cycling stability as well as the rate capability. The improved electrochemical performance and preservation of the phase and structural stability across the (104) LiCoO2 surface suggest a more stable Co3+ oxidation state, which not only limits the cobalt-ion dissolution into the electrolyte but also suppresses the catalytic degradation of the liquid electrolyte. The high c-rate performance combined with high Columbic efficiency indicates that interstitial sites in the cubic TiO lattice offer facile pathways for fast lithium-ion transport.

Experimental Section

The LiCoO2 thin films are deposited on metallic SrRuO3 buffer layers, which enhances the electrical transport toward the underlying conducting substrates.[43] The LiCoO2 and SrRuO3 layers were grown by pulsed laser deposition (PLD) on Nb-doped (0.5 wt %) single-crystalline SrTiO3 [(100), (110), or (111)] substrates from sintered LiCoO2 (40% excess of lithium) and SrRuO3 targets, using a KrF excimer laser operating at 248 nm. The Nb–SrTiO3 substrates were pre-annealed at 950 °C for 1.5 h in an oxygen flow of 150 mL/min. The laser energy fluence was 2.3 J cm–2 for the growth of both LiCoO2 and SrRuO3. The temperature (and oxygen pressure) during SrRuO3 and LiCoO2 growth were, respectively, 600 °C (0.13 mbar) and 800 °C (0.23 mbar). In total, 3600 pulses at 2 Hz were used to deposit the SrRuO3 layer and 7200 pulses at 20 Hz for the LiCoO2 layer. Under these growth conditions, the thicknesses of the SrRuO3 and LiCoO2 layers are, respectively, about 60 and 120 nm.[56] After deposition, the films were cooled down to room temperature under 1 bar oxygen pressure at 10 °C min–1 to enhance the oxidation level. For the deposition of an additional TiO thin film, a sintered anatase TiO2 target was ablated for 1800 pulses at 5 Hz, resulting in a layer thickness of about 30 nm, which is similar to thicknesses of surface coatings in previous LiCoO2 studies.[57−59] The cubic TiO thin film was realized on a (104)-oriented LiCoO2 film at a temperature of 550 °C and 0.01 mbar argon pressure. After deposition, the thin films were cooled down to room temperature under 0.01 mbar argon pressure at 10 °C min–1 to maintain the oxidation level in the deposited layers.

The crystal structure, surface morphology, and thin-film thickness were investigated by X-ray diffraction (XRD, PANalytical X’Pert PRO), Raman spectroscopy (WiTec), and atomic force microscopy (AFM, Bruker ICON Dimension Microscope). The lithium content in the LiCoO2 thin films was measured using inductively coupled plasma–optical emission spectroscopy (ICP–OES, PerkinElmer 8300dv) in which the plasma was inspected vertically with respect to the detector using the radial viewing mode.

High-resolution XRD data were collected post cycling using a Bruker D8 Discover diffractometer with a high brilliance microfocus Cu rotating anode generator (TXS, 2.5 kW), hybrid Montel optics (parallel-focusing), a channel-cut two-bounce Ge(220) monochromator (ACC2), a 1 mm diameter circular pinhole beam collimator, and an EIGER2 R 500 K area detector. A coupled scan was performed by operating the detector in the conventional 0D mode with a small region of interest (13 × 61 pixels, pixel size 75 × 75 μm2). No secondary optics in the diffracted beam paths between sample and detector were installed. During scans, the intensity of the incident beam is automatically adjusted by a multilevel rotary absorber to optimize the effective dynamic range of each measurement and to avoid detector saturation. Raw data for the symmetrical and asymmetrical reciprocal space maps (RSMs) of various Bragg reflections were collected via sets of high-resolution rocking curves for which the detector was operated in the stationary 2D snapshot mode in a grazing-exit configuration. In our unique lab-based XRD system, the combination of high beam collimation to a virtual point source in combination with a large area detector enables us to reconstruct the full 3D reciprocal space around the chosen Bragg reflections. Angular to reciprocal space conversion was done according to He.[60] Subsequently, the array of 3D reciprocal space coordinates with the corresponding scattering intensities was binned and interpolated onto an orthogonal and equidistant 3D matrix with voxel size q, q, q = 0.0005 Å–1. The effective resolution is furthermore limited by the convolution of the aforementioned voxel size and the experimental accuracy. An ideal, asymmetrical RSM contains in-plane and out-of-plane information about the substrate and all thin-film layers simultaneously, that is, in one picture. In coplanar geometry, the (103) and (113) reflections of the cubic SrTiO3(001) substrate are commonly investigated, but we confirmed that the adjacent regions in reciprocal space do not contain (strong) reflections of both LiCoO2 and TiO. The selection of a suitable region in reciprocal space is complicated by the dissimilar crystal structures of the substrate and the layers of interest. While Nb/SrTiO3 and SrRuO3 are perovskites, LiCoO2 has a hexagonal layered structure, and TiO has a defect rock salt structure.

For electrochemical characterization, the films were transferred to an argon atmosphere glovebox (<0.1 ppm H2O and O2) and placed on a hotplate for ∼10 min at 125 °C to remove any water content. Subsequently, they were positioned in an electrochemical EC-Ref cell (EL-CELL) and combined with a glass fiber separator of 1 mm thickness, 100 μL electrolyte with 1 M LiPF6 in 1:1 ethylene carbonate/dimethyl carbonate (EC/DMC), and a lithium metal anode. The electrochemical measurements were performed at room temperature using a BioLogic VMP-300 system in a two-electrode setup. Galvanostatic charge–discharge cycling was performed within 3.5–4.2 V range to avoid any contributions of structural or phase transitions on the cycle performance for higher voltages, as previously demonstrated for LiCoO2 cathodes with and without surface coatings.[3,4,7,8,20,28,61,62] After each charge step, the cell was put to rest for 10 min followed by the discharge step. Cyclic voltammetry was carried out between 3.6 and 4.2 V at various scan rates. Electrochemical impedance spectroscopy (EIS) was performed within 1 MHz to 0.1 Hz range with a 10 mV AC perturbation using a BioLogic VMP-300 system.

Results and Discussion

Structural Characterization

The epitaxial relationship between the LiCoO2 films and the Nb/strontium titanate (STO) substrates is shown by the out-of-plane XRD measurements in Figure a. Besides the high-intensity sharp peaks from the Nb/STO substrates, only peaks belonging to the single crystalline layered LiCoO2 (R3m) phase can be observed. For (100)-oriented Nb/STO substrates, the epitaxial LiCoO2 layer exhibits a (104) orientation, while for (110)-oriented Nb/STO substrates, the LiCoO2 layer aligns to the (110) orientation. In contrast, using (111)-oriented Nb/STO substrates provides a good match with the in-plane triangular structure of the layered LiCoO2 phase, resulting in alignment of the LiCoO2 layer in the (001) orientation. The successful synthesis of single-crystalline LiCoO2 thin films with specific orientations is in good agreement with observations in previous studies.[42−44] Although no peaks for other LiCoO2 phases could be observed in the XRD measurements, some contributions of the spinel LiCoO2 phase (Fd3̅m) could still be present as it is nearly indistinguishable from the layered LiCoO2 phase (R3m). For example, the (100) planes of spinel LiCoO2 are equivalent to the (104) planes of layered LiCoO2, while the (110) planes are also similar in both phases. This is due to a similar oxygen anion framework with ABC-packing which is present in both spinel and layered hexagonal phases. Since LiCoO2 films were grown at 800 °C, the presence of the Co3O4 phase and lithium deficiency cannot be excluded initially. Therefore, Raman analysis was performed to further distinguish any undesired phases and impurities in the thin films. Figure b shows the Raman spectra of the LiCoO2 films grown in either the (104) or (001) orientation. Factor group analysis suggests distinct vibrational modes for spinel and layered LiCoO2 phases.[63,64] The spinel LiCoO2 (Fd3̅m) phase exhibits four Raman active bands, attributed to, respectively, A1g, Eg, and two F2g modes. On the other hand, the layered LiCoO2 (R3m) phase exhibits only two Raman active bands, attributed to, respectively, A1g and Eg modes. The high-frequency mode is assigned to Co–O stretching in CoO6 octahedra, whereas the low-frequency mode is assigned to O–Co–O bending.[63,65]Figure b shows that in the spectra for both oriented LiCoO2 films, only two strong Raman bands were observed at 484 and 594 cm–1, which confirms that the LiCoO2 films grown by PLD exhibit the layered structure with space group R3m. The distinct increase in the intensity of the A1g mode for the (001)-oriented LiCoO2 film suggests an enhancement in Co–O stretching as expected for the epitaxial in-plane alignment of the CoO2 layers. The additional minor peak at around 659 cm–1 suggests the presence of some traces of the Co3O4 impurity phase.[66] ICP–OES analysis was performed to quantify the lithium to cobalt ratio in the LiCoO2 films, which was 0.982 ± 0.06 and 0.940 ± 0.06 for, respectively, the (104)-oriented and (001)-oriented LiCoO2 films. This confirms that the excess of lithium in the initial target is able to compensate for the loss of lithium during the PLD process caused by the volatility of lithium atoms.

Figure 1

Out-of-plane XRD pattern of LiCoO2 epitaxial thin films grown on SrRuO3-buffered Nb–SrTiO3 (Nb/STO) substrates with different crystal orientations: (100), (110), and (111). Nb/SrTiO3 and SrRuO3 peaks are indicated by □ and * symbols, respectively. (b) Raman spectra of LCO thin films epitaxially grown in (104) and (001) orientations.

To investigate the effect of a conductive coating, a cubic TiO film was subsequently in situ deposited on top of the (104)-oriented LiCoO2 film. The epitaxial relation between the TiO and LiCoO2 layers was evaluated by out-of-plane XRD analysis, as shown in Figure . The cubic (100)-oriented TiO layer was epitaxially stabilized on the (104)-oriented LiCoO2 layer as only two distinct diffraction peaks at 43.2 and 94.5°, ascribed to (200) and (400) reflections of the cubic TiO phase, respectively, can be observed besides the characteristic peaks for the (104)-oriented LiCoO2 phase on top of a SrRuO3-buffered Nb/STO(100) substrate. The absence of any additional diffraction peaks confirms that the cubic TiO phase is completely stabilized and free from anatase or rutile phases. The epitaxial growth of cubic TiO (where a = 4.18 Å, Ti–Ti = 4.18 Å, O–O = 4.18 Å) is facilitated by the underlying cubic surface arrangement of lithium and cobalt atoms (Li–Co = 4.06 Å) in the (104) plane of LiCoO2 with a 45° in-plane rotation with respect to the (200) plane of TiO, as schematically shown in Figure . The lattice strain across the (200)TiO//(104)LiCoO2 interface is 2.9%, which is similar to the epitaxial match at the underlying (104)LiCoO2//(100)SrRuO3 interface (where Li–Co = 4.06 Å and Sr–Sr = 3.91 Å) exhibiting a lattice strain of 3.7%. The surface morphology of the noncoated and TiO-coated (104)-oriented LiCoO2 thin films were characterized by AFM, as shown in Figure a,b respectively. The (104)-oriented LiCoO2 films exhibit a morphology with trenches in between the cubic structures (rms: ∼20 nm), which is in good agreement with previous studies.[42−44] The subsequently deposited TiO layer forms into 20–30 nm-sized grains on top of the (104)-oriented LiCoO2 surface. The TiO layer induces an increase in the surface roughness on top of the cubic structures, from a smooth (104)-oriented LiCoO2 surface (rms: ∼0.3 nm) to a grainy (200)-oriented TiO surface (rms: ∼4.0 nm); see Supporting Information, Figure S1. The structural ordering within the full TiO–LiCoO2–SrRuO3–Nb/SrTiO3 stack was investigated by scanning electron microscopy (SEM) analysis, as shown in Figure c. The layer thicknesses of the SrRuO3 transport layer and the TiO coating layer were confirmed to be, respectively, 60 and 30 nm. The surface morphology of the (104)-oriented LiCoO2 cathode layer was clearly observed, and an average layer thickness of 120 nm was determined, in good agreement with the LiCoO2 growth rate previously determined for low surface roughness (001)-oriented LiCoO2 layers (not shown).

Figure 2

Out-of-plane XRD pattern of epitaxially stabilized (100)-oriented cubic TiO grown on top of a (104)-oriented LiCoO2 thin film. For comparison, the reference diffraction patterns for cubic TiO, anatase TiO2, and rutile TiO2 are also included. Nb/SrTiO3 and SrRuO3 peaks are indicated by □ and * symbols, respectively.

Figure 3

Schematic (a) side view and (b) top view of the atomic arrangement across the epitaxial interface between the (200) TiO and (104) LiCoO2 planes.

Figure 4

AFM analysis of the surface morphology of (a) bare (104)-oriented LiCoO2 thin film and (b) TiO-coated (104)-oriented LiCoO2 thin film. (c) SEM analysis of cross-section showing the structural ordering within the full TiO–LiCoO2–SrRuO3–Nb/SrTiO3 stack.

Electrochemical Characterization

The impact of the applied TiO coating on the electrochemical performance of the LiCoO2 cathode thin film was investigated through charge–discharge cycling under galvanostatic conditions. The discharge profiles during prolonged cycling at 5 C (keeping charge rate fixed to 1 C rate) of the TiO-coated and noncoated LiCoO2 films are shown in Figure a,b respectively. The presence of the characteristic voltage plateau at ∼3.9 V for both LiCoO2 films confirms that the TiO coating remains electrochemically inactive within the used voltage range (3.5–4.2 V). Interestingly, the noncoated LiCoO2 film shows a slightly higher capacity with a shorter voltage plateau as compared to the TiO-coated LiCoO2 film. This can be attributed to an interfacial pseudo capacitance arising from electrolyte reactivity with the bare LiCoO2 surface.[11] The capacity of the noncoated LiCoO2 film reduces significantly faster during cycling than that of the TiO-coated LiCoO2 most likely due to the continuous cation dissolution into the electrolyte as well as the formation of a SEI layer at the LiCoO2 interface with the liquid electrolyte. The difference in cycle life between both cases can be clearly observed in Figure a, which shows that the capacity of TiO-coated LiCoO2 is still 92% after 200 cycles, in sharp contrast with 67% for the noncoated LiCoO2.

Figure 5

Charge–discharge analysis of (a) TiO-coated and (b) noncoated (104)-oriented LCO films during prolonged cycling at rates of 1 C (during charge) and 5 C (during discharge).

Figure 6

Electrochemical analysis of (top) rate capability at 5 C and (bottom) cycle life of the discharge capacity at various rates determined for TiO-coated and noncoated (104)-oriented LCO films.

The applied cubic (100)-oriented TiO layer is expected to provide an enhancement of the surface passivation (to minimize the SEI formation and structural degradation) as well the ionic transport of lithium between the liquid electrolyte and LiCoO2 cathode. The effect of the TiO coating on the high rate capability of the test cells was studied between 2 and 40 C, as shown in Figure b. As observed before, the TiO-coated and noncoated LiCoO2 films show very similar capacities when cycled at 2 and 5 C. However, as the C rate increases from 10 to 40 C, the capacity of noncoated LiCoO2 drops significantly (5 C = 2.31 μA h/cm2, 10 C = 1.56 μA h/cm2, 20 C = 1.07 μA h/cm2) and becomes almost zero at 40 C. This poor rate and cycle performance is typical for bare LiCoO2 surfaces and can be attributed to surface degradation and unfavorable electrode–electrolyte reactions caused by high oxygen reduction and evolution reactions manifested by the presence of high and intermediate spin Co3+ at the (104) LiCoO2 surface.[67] These reactions cause the formation of a SEI layer, as schematically shown in Figure a, which increases the interfacial resistance during subsequent charge–discharge cycles and limits the rate and cycling performance. In contrast, the TiO-coated LiCoO2 films show remarkable enhanced cycling and rate performance (5 C = 2.33 μA h/cm2, 10 C = 2.25 μA h/cm2, 20 C = 2.10 μA h/cm2, and 40 C = 1.95 μA h cm2) as compared to noncoated LiCoO2. Although the capacity also decreases with increasing C rate, the overall reduction is rather limited as compared to that in noncoated LiCoO2 and a significant amount of initial capacity is recovered by final cycling at 2 C. Furthermore, the cubic (100)-oriented TiO layer clearly exhibits good lithium-ion transport behavior, as was expected for the open cubic structure providing a large number of interstitial sites for lithium diffusion.[50−52] In comparison, the commonly used metal oxide coatings (e.g., TiO2, Al2O3) do not provide such high rate capability due to the close-packed nature of the crystal structures with lower lithium diffusivity.[19,48,68,69] Therefore, the improved electrochemical performance is attributed to the combination of effective stabilization of the Co3+ state at the LiCoO2 surface with the optimized pathway through the TiO layer for lithium diffusion, as schematically shown in Figure b. The effect of the TiO coating on the electrochemical performance was investigated by EIS analysis for (104)-oriented LiCoO2 (LCO) thin-film electrodes with and without a TiO coating in the discharge state after the 1st and 50th charge–discharge cycle at 5 C in 3.5–4.2 V range; see Figure c. The noncoated LCO film exhibits a much higher initial resistance after the 1st cycle as compared to the TiO-coated LCO film, which suggests a reaction between the liquid electrolyte and LiCoO2 surface leading to the formation of a CEI layer. During subsequent cycling, the noncoated LCO shows a dramatic increase in resistance, as shown after the 50th cycle, which is in good agreement with the known surface degradation and unfavorable electrode–electrolyte reactions for noncoated LiCoO2 surfaces responsible for the poor rate and cycle performance. However, the TiO-coated LCO shows a very minimal change in resistance after 50 charge–discharge cycles, indicating enhanced surface stability and optimal lithium diffusion through the TiO coating, enabling very stable, reversible (de)lithiation of the LiCoO2 electrode. Further EIS measurements and modeling will be performed in a future study to explore the detailed electrochemistry in this material system.

Figure 7

Schematic showing mechanistic insights of surface passivation and lithium transport across electrode–electrolyte interface in (a) noncoated and (b) TiO-coated (104)-oriented LCO thin-film electrodes. (c) EIS analysis of (104)-oriented LCO thin-film electrodes with and without a TiO coating in the discharge state after the 1st and 50th charge–discharge cycle at 5 C in 3.5–4.2 V range.

The excellent cycle life performance of the TiO-coated LiCoO2 film was investigated in detail during prolonged cycling at a 5 C rate; see Figure a. The condition of the LiCoO2 cathode was evaluated through cyclic voltammetry prior to the cycle life test as well as after 500 cycles, as shown in Figure b,c, respectively. The presence of the characteristic cathodic peak at 3.95 V and the anodic peak at 3.87 V suggests that the LiCoO2 film remains close to its pristine condition. When the scan rate increases from 0.1 to 1.0 mV/s in subsequent cycles, the characteristic peaks shift due to enhanced polarization associated with charge–discharge cycling. After 500 charge–discharge cycles, the successive CV measurements show similar cathodic and anodic peak positions as compared to the initial CV measurement indicating good reversibility and cycling stability of the TiO-coated LiCoO2 electrode. At the end of 900 cycles, the TiO-coated LiCoO2 film still exhibited ∼80% (∼1.95 μA h cm–2) of its initial capacity, while the Coulombic efficiency remained above 99.9% throughout the complete cycle life analysis. Selected charge–discharge profiles for the 100th, 500th, and 900th cycles are shown in Supporting Information, Figure S2. After cycling, the cell was disassembled in a glovebox, the liquid electrolyte was removed, and the TiO-coated LiCoO2 film was analyzed using high-resolution XRD and SEM analysis to confirm the structural integrity throughout prolonged cycling. SEM analysis showed a surface morphology after extensive cycling (Supporting Information, Figure S3) very similar to the pristine surface of the as-deposited sample (Figure c), indicating preservation of the good contact between all layers without any exfoliation or crack formation.

Figure 8

(a) Cycle life analysis of capacity and Coulombic efficiency during prolonged cycling at 5 C for TiO-coated (104)-oriented LCO thin films. Cyclic voltammogram of TiO-coated LiCoO2 thin films (b) prior to cycle life test and (c) after 500 charge–discharge cycles.

High-resolution XRD data were collected post cycling to determine the structural integrity of all layers after extensive charge–discharge cycling, as well as to confirm the epitaxial relation between the layers and the underlying substrate. Here, we present a RSM measured in skew-geometry, that is, non-coplanar, with ψ = 45° sample tilt, around the (022) reflection of the substrate; see Figure a. The processed 3D data set is integrated along the q (also known as K) axis to produce a 2D out-of-plane q–q RSM (also known as an H-L map). Whereas a polycrystalline thin film with randomly oriented grains would show a continuous circular arc around the center of the reciprocal space, that is, a projection of a Debye–Scherrer ring; here, we observe well-defined, narrow spots for all three thin-film layers, SrRuO3, LiCoO2, and TiO, reflecting low mosaic tilt. Additionally, in-plane q–q RMSs (also known as H–K maps) were made via projection and piecewise integration along the q axis and analyzed to further confirm the in-plane epitaxial nature of all samples; see Figure b,c. Textured films with preferential out-of-plane orientation, but without in-plane order, would show a ring feature in such a plot, which clearly we do not observe. Instead, we find well-defined, symmetrical spots attesting to minimal mosaic twist. Thus, we can exclude with certainty that these thin films are polycrystalline. Furthermore, we measured phi scans of suitable, that is, non-coplanar, Nb/SrTiO3, LiCoO2, and TiO reflections and observed the expected fourfold symmetry dictated by the substrate (Figure ). The narrow shape of the peaks and the absence of intensity in between confirms that the layers are grown in full registry with the substrate. We note that the epitaxial relation between LiCoO2 and Nb/SrTiO3 observed here confirms earlier studies by Nishio et al.[44] Furthermore, out of plane XRD analysis (Figure S4) shows that no additional or impurity phases were detected in cycled TiO-coated LiCoO2 film. However, the possible formation of a very thin fully oxidized TiO2 layer at the (100) TiO surface in contact with the liquid electrolyte cannot be excluded. These post cycling XRD results unequivocally confirm that the SrRuO3, LiCoO2, and TiO layers are epitaxially related to the underlying Nb/SrTiO3 substrate, and this high degree of structural order is preserved over about a thousand charge and discharge cycles.

Figure 9

XRD after extensive electrochemical cycling (900 cycles for 3.5–4.2 V range at 5 C). Non-coplanar RSMs measured in skew geometry (ψ = 45°) around the (220) reflection of the Nb/SrTiO3 (100) substrate. (a) Out-of-plane (H–L) projection displaying only the four expected sharp reflections for Nb/SrTiO3(220), SrRuO3(220)pc, LiCoO2(018), and TiO(220), confirming the epitaxial nature of the layer system with very low out-of-plane crystallographic tilt. (b,c) In-plane projection (H–K) of LiCoO2(018) reflection and TiO(220), respectively. The well-defined circular spots attest to a minimal in-plane mosaic twist. (d) Extracted line profile along q showing reflections corresponding to TiO(200), LiCoO2(104), SrRuO3(002), and Nb/SrTiO3(200).

Figure 10

XRD after extensive electrochemical cycling (900 cycles for 3.5–4.2 V range at 5 C). Phi scans of LiCoO2(018) and TiO(220) reflections shown in the RSMs, revealing the expected 4-fold symmetry of the epitaxial layer system in full registry with the single crystalline Nb–SrTiO3 substrate, confirming minimal mosaic twist.

Conclusions

Epitaxial engineering is applied to stabilize a cubic (100)-oriented TiO layer on top of single (104)-oriented LiCoO2 thin films to study the effect of an ionically and electronically conductive coating. Lattice matching between the (104) LiCoO2 surface facets and the (100) TiO plane is crucial to enable the formation of the titanium mono-oxide phase. The application of a cubic TiO coating dramatically enhances the electrochemical performance of LiCoO2 thin films. This enhanced electrochemical performance combined with post cycling XRD analysis clearly indicates that such cubic TiO coating enhances the preservation of the phase and structural stability across the (104) LiCoO2 surface. The results suggest a more stable Co3+ oxidation state, which not only limits the cobalt-ion dissolution into the electrolyte but also suppresses the catalytic degradation of the liquid electrolyte.[9,67] Furthermore, the high c-rate performance combined with high Columbic efficiency indicates that interstitial sites in the cubic TiO lattice offer facile pathways for fast lithium-ion transport.

Detailed analysis in a half cell against lithium metal has shown a dramatic enhancement of the cycling stability as well as the rate capability. The results indicate that TiO remains electrochemically inactive within the used voltage window (3.5–4.2 V) and does not contribute to the overall capacity. However, operando spectroscopic studies are required to probe the role of the cubic TiO coating, and a possible gradient of the O/Ti ratio,[70] when exposed to a liquid electrolyte and its interface to the buried LiCoO2 electrode during consecutive lithium (de)intercalations to build in-depth understanding about the interfacial reactions within such lithium-ion batteries.