Diffusion‐Optimized Long Lifespan 4.6 V LiCoO2: Homogenizing Cycled Bulk‐To‐Surface Li Concentration with Reduced Structure Stress

Abstract Increasing the charging cut‐off voltage (e.g., 4.6 V) to extract more Li ions are pushing the LiCoO2 (LCO) cathode to achieve a higher energy density. However, an inhomogeneous cycled bulk‐to‐surface Li distribution, which is closely associated with the enhanced extracted Li ions, is usually ignored, and severely restricts the design of long lifespan high voltage LCO. Here, a strategy by constructing an artificial solid–solid Li diffusion environment on LCO's surface is proposed to achieve a homogeneous bulk‐to‐surface Li distribution upon cycling. The diffusion optimized LCO not only shows a highly reversible capacity of 212 mA h g−1 but also an ultrahigh capacity retention of 80% over 600 cycles at 4.6 V. Combined in situ X‐ray diffraction measurements and stress‐evolution simulation analysis, it is revealed that the superior 4.6 V long‐cycled stability is ascribed to a reduced structure stress leaded by the homogeneous bulk‐to‐surface Li diffusion. This work broadens approaches for the design of highly stable layered oxide cathodes with low ion‐storage structure stress.

(LGTP) are introduced to the surface of LCO material to change the environment of Li + transport.Firstly, LiNO3, Ga(NO3)3, C12H28O4Ti, NH4H2PO4 with the molar ratio Li:Ga:Ti:P of 1.5:0.5:1.5:3 were first dissolved in ethanol, citric acid was then added to the above solution and mixed thoroughly for 5 h at 50 °C to obtain LGTP sol solutions.Next, LCO powder was mixed with LGTP sol solutions with the amount of LGTP being 0.3 wt%, 0.5 wt% and 0.7 wt% of LCO powder, respectively.The mixtures were stirred constantly at 80 °C until the solvents have evaporated.The thus obtained precursor gels were dried at 130 °C for 12 h in air, and then calcined at 900 °C for 10 h to achieve the surface of solid electrolyte.

Preparation LGTP samples
In order to further compare the Li+ diffusion ability of LCO and surface LGTP, we synthesized pure LGTP material through sol-gel process, and the specific synthesis method is as follows: stoichiometric amounts of LiNO3, Ga(NO3)3, C12H28O4Ti and NH4H2PO4 were used.
LiNO3, Ga(NO3)3, C12H28O4Ti, NH4H2PO4 with the molar ratio Li:Ga:Ti:P of 1.5:0.5:1.5:3 were first dissolved in ethanol, citric acid was then added to the above solution and mixed thoroughly for 5 h at 50 °C to obtain LGTP sol solutions.The gel was dried at a temperature of 80°C and calcined at 700°C in air to obtain LGTP material.

Laboratory characterizations
The crystal structures of synthetic samples were analyzed using X-ray powder diffraction (XRD) on a Persee XD2 diffractometer with Cu Kα radiation and angle coverage of 10°-80°.
The element contents and dissolved Co contents in electrolyte after cycles were detected using an inductively coupled plasma (ICP) spectrometer.Sample surface element valence states and compositions were measured by X-ray photoelectron spectroscopy (XPS) using a Thermo Fisher ESCALAB XI+.Scanning electron microscopy (SEM) characterization of the samples were carried out by using a Gemini 300 equipment.Sample elemental mapping images and microstructural change were obtained by high resolution transmission electron microscopy (HRTEM) on a JEM-F200.The relationship between Li + diffusion and structural stability was evaluated by ex-situ electron spin resonance (EPR) spectroscopy (EMXplus).For ex-suit testing, cathode materials were charged to 4.6 V, washed with dimethyl carbonate (DMC) and then scraped off from the current collectors.STEM (Spectra 300) equipped with a field emission gun at 300 kV was utilized to confirm the atomic arrangement and structure of the samples.GPA method is proposed to extract the 2D relative strain maps from HAADF-STEM image and analyze the internal and external stress of highly delithiated LCO.

In-situ XRD and in-situ Raman
In-situ XRD was performed on a Bruker D8-ADVANCE X-ray diffractometer using a Cu target under 45 kV and 40 mA.The data were collected at 2θ 3° per minute and the 2θ angle range is 10°-50°.In order to allow X-ray to penetrate the in-situ cell, Be metal was designed as a diffraction window.For the in-situ Raman (on Horiba LabRam HR Evolution) measurements, a thin quartz window (0.5 mm thick) is fixed on the top of the battery and a small hole is punched in the center of the separator/lithium foil to collect laser and Raman signals.The charge/discharge current density is 27.4 mA g -1 in a voltage range of 3.0-4.6V for the in-situ XRD and in-situ Raman tests.

In-situ differential electrochemical mass spectrometry (DEMS)
Differential electrochemical mass spectrometry experiments were carried out on a PM-DEMS instrument in a customized battery case (ECC-DEMS).A slurry consisting of 80 wt.% active materials, 10 wt.% polyvinylidene fluoride (PVDF) binder, and 10 wt.% conductive carbon was coated on a 12-mm-diameter aluminum foil collector.Lithium metal was used for the anode electrode.Glass fiber was used as the separator.The cell was charged/discharged at 3.0-4.6V with a current density of 27.4 mA g -1 .High-purity argon (99.999 %) was used as the carrier at a flow rate of 0.7 sccm.

Ex-situ soft X-ray absorption spectroscopy (sXAS)
The sXAS data were collected on the beamline BL20A1 in National Synchrotron Radiation Research Center (NSRRC).The beamline covers the spectral range from 60 eV to 1250 eV, with an average energy resolving power of 5000.sXAS can be detected to depths of several nanometers in total electron yield (TEY) mode and hundreds of nanometers in total fluorescence yield (TFY) mode.Thus, the TEY model is surface sensitive, while the TFY model is used to evaluate the electronic structure of the bulk material.

Electrochemical measurements
CR2032 coin cells with liquid electrolyte were assembled using an electrolyte composed of 1.2 M LiFP6 in ethylene carbonate (EC) and diethyl carbonate (DEC), glass fiber separators, and Li as the counter electrode, in high-purity argon atmosphere.LCO and DO-LCO active materials, super P conductive additive, and poly (vinylidene fluoride) were mixed in 80:10:10 ratio to obtain a uniform slurry with a thickness of 150 μm, which was then dried at 120 ℃ for 12 h in a vacuum oven, and the mass loading of both active substance is 1.4 mg/cm 2 approximately.Galvanostatic charge-discharge and Galvanostatic Intermittent Titration Technique (GITT) tests were conducted on an automatic galvanostat (NEWARE) in the voltage window of 3.0-4.6V at various current densities (1 C=274 mA g -1 ).The cyclic voltammetry (CV) measurement was employed through an electrochemical workstation (PGSTAT302N, Autolab).

COMSOL simulation
The finite element analysis is carried out by the commercial software COMSOL       XRD was firstly used to evaluate the layered crystal structure of LCO and diffusion optimized electrodes (Figure S2-S5).All samples can be described using a well-developed layered structure with an emblematical group space of R-3m (No. 166).In addition, the Rietveld refinement calculated crystallographic parameters are summarized in Table S2-S5.It can be seen from the refinement results that the lattice parameter c gradually increases with the increase of solid electrolyte content, mainly because Ga 3+ /Ti 4+ with small radius are doped in the transition metal (TM) layer shortening the TM layer distance during the high temperature sintering process.In addition, we also performed XPS analysis for two elements to further verify that Ti 4+ and Ga 3+ were successfully doped into the material (Figure S6).The resistance and the Li + diffusion coefficient of LCO and LGTP materials was obtained through electrochemical impedance spectroscopy (EIS).The specific impedance spectroscopy (Figure S8) and the calculation results of the Li + diffusion coefficient are as follows: According to Equation (1): where T, F, and R, are absolute temperature, Faraday's constant, and gas constant, respectively.A and C are the area of electrode and Li ion molar concentration, respectively.
Warburg factor (σ) can be determined by Equation (2): (2) According to formula 2, the calculated σ value of LCO and LGTP are 186 and 115, so the Li + diffusion coefficient (DLi+) corresponding to LCO and LGTP are 2.7*10^-12 cm 2 s -1 and 7.4*10^-13 cm 2 s -1 , respectively.S2] The diffusion coefficient of Li + in LGTP is smaller than that of LCO, which indicates that the LGTP layer with slow solid-solid diffusion environment close to the bulk can achieve a homogeneous bulkto-surface Li distribution upon cycling.However, the removal of Li + in the material is accompanied by the oxidation of Co 3+ , so the distribution of Li + on the bulk and surface can be verified by the content of Co 3+ and Co 4+ .S4] In order to better show the surface of LCO and DO-LCO materials, the surface of the two materials were etched at 20 nm and 200 nm, respectively, corresponding to the cathode electrolyte interface and LGTP on the surface.It is obvious that the strong decrease of the relative satellite area (from 31.41% to 29.33%) observed in Figure S15, S16 and Table S6, together with the strong broadening of the main peak, can be attributed to an oxidation process of Co 3+ at surface of LCO and DO-LCO.The above results indicate that the surface of LCO has a higher content of Co 4+ than that of DO-LCO.It is worth noting that as the etching depth increases (surface to bulk), the area of satellite peaks in LCO gradually increases, indicating that the surface and bulk exhibit a significantly inconsistent oxidation state of Co.However, the area of satellite peaks on the surface and bulk of DO-LCO changes relatively small, further confirming the uniform distribution of Li + .In order to investigate the effect of Li + diffusion optimization on the electrochemical behavior of materials, all prepared materials were subjected to electrochemical performance testing.The initial discharge capacity is 212, 212, 212 mA h g -1 accompanied by capacity retention rates of 92%, 93% and 90% at 1 C for 0.3% DO-LCO, 0.5% DO-LCO and 0.7% DO-LCO samples, respectively.Notably, the capacity retention rate of 0.3% DO-LCO, 0.5% DO-LCO and 0.7% DO-LCO material is still as high as 75%, 80% and 67%, accompanied by the attenuation of LCO to only 50% at 1 C after 600 cycles, which suggests that diffusion optimization homogenizes Li + distribution, significantly improving the long cycle life of the material.Moreover, the average voltage attenuation rate at 1 C of 0.3% DO-LCO, 0.5% DO-LCO and 0.7% DO-LCO samples is only 0.018% (0.12 mV per cycle), 0.017% (0. Multiphysics.A three-dimensional electric field and ion distribution model based on COMSOL was established to elucidate the distribution of ions in different structures (LCO and DO-LCO), coupled with solid mechanics to calculate stresses due to uneven delithium and insertion to predict particle damage.The model uses a 4 μm cube to simulate nanoparticles and a comparison sample of modified particles covered with LGTP (0.1 μm).The model is considered as a multi-physics coupling of the Nernst-Planck equation for the mass transport of all particles.The exchange current density at the electrode surface satisfies the Butler-Volmer equation with an exchange current density of 100 A m -2 , a transfer coefficient of 0.5 and the electrolyte current density is 1 mA cm -2 .

Figure S1 .
Figure S1.The STEM image for the 4.6 V-charged LCO.

Figure S2 .
Figure S2.XRD data and corresponding refinement results for LCO sample.

Figure S3 .
Figure S3.XRD data and corresponding refinement results for 0.3% DO-LCO sample.

Figure S4 .
Figure S4.XRD data and corresponding refinement results for 0.5% DO-LCO sample.

Figure S5 .
Figure S5.XRD data and corresponding refinement results for 0.7% DO-LCO sample.
Figure S7.The XRD pattern for the LGTP material.XRD results show that LGTP has been successfully synthesized.

Figure S10 .
Figure S10.Formation energy of the reaction for LCO and LGTP with HF.In order to better clarify the relationship between solid electrolyte and the excellent electrochemical properties associated with structural integrity, the reaction energies between LCO/solid electrolyte and HF in the electrolyte are calculated.The results show that the binding energy between solid electrolyte and HF is smaller than that of LCO, indicating thatLGTP can prevent the corrosion of the electrolyte, thus inhibiting the irreversible phase transition and structural collapse during the long cycles.

Figure S11 .
Figure S11.The TEM image of LCO.

Figure S12 .
Figure S12.The TEM image of DO-LCO.It can be seen from FigureS11that LCO exhibits relatively regular lattice stripes accompanied by (104) crystal planes with a crystal plane spacing of 0.208 nm.Notably, it clearly confirms that the LGTP coating layer (about 200 nm) is well coated onto LCO in FigureS12.Besides, two kinds of different lattice fringes, the interplanar spacing values of 0.480 and 0.208 nm belong to the (-111) plane of LGTP (pristine rhombohedral phase LGTP coating layer transformed into a spinel-like phase after the high-temperature sintering)[S3]  and the (104) plane of LCO, respectively, were distinguished.

Figure S13 .
Figure S13.EDS overlay plot and single element distribution of Co and Ti elements in DO-LCO sample.Energy dispersive spectroscopy (EDS) mapping was performed to obtain elemental distribution.Co is relatively evenly distributed in the bulk, while Ti elements are enriched on the surface of DO-LCO accompanied by a dense layer of about 100 nm.The EDS results well demonstrate our expected tailored solid electrolyte surface lattice structure.

Figure S14 .
Figure S14.The scanning electron microscope (SEM) images of LCO and DO-LCO.The morphologies of LCO and DO-LCO, both of which are all single crystal particles

Figure S15 .
Figure S15.The Co 2p XPS for LCO at different etching depth.

Figure S16 .
Figure S16.The Co 2p XPS peak for DO-LCO at different etching depth.We conducted XPS etching of Li ions and Co ions on the 4.6 V charged state LCO and DO-LCO to further explore the in-depth chemical state of the electrodes.Due to the low concentration of Li + in 4.6 V highly delithiated state, XPS cannot detect the peak of Li 1s.

Figure S17 .
Figure S17.The CV curve of 2-4 cycles for LCO an DO-LCO materials.

Figure S18 .
Figure S18.The long cycle and average voltage performance at 1 C of LCO, 0.3% DO-LCO and 0.7% DO-LCO materials.

Figure S19 .
Figure S19.The variation of discharge curves upon cycling at 1 C for 0.3% DO-LCO (a) and 11 mV per cycle) and 0.026% (0.18 mV per cycle) much lower than LCO of 0.035% (0.23 mV per cycle), further demonstrating that optimized Li + diffusion can suppress voltage attenuation.By evaluating various aspects of electrochemical performance of the prepared materials, we selected 0.5% DO-LCO material with high capacity, excellent cyclic stability and low voltage decay rate as representative material for subsequent electrochemical testing and structure characterization.

Figure S20 .
Figure S20.The long cycle performance at 10 C of LCO and 0.5% DO-LCO materials.

Figure S21 .
Figure S21.(a) The GITT results of LCO and DO-LCO.The figure includes the calculated diffusion coefficient.(b) Overpotential obtained from GITT results for LCO and DO-LCO.

Figure S22 .
Figure S22.The in-situ XRD results of LCO during the initial charge and discharge processes and changes in cell parameters a and c obtained by refinement.

Figure S23 .
Figure S23.The in-situ XRD results of DO-LCO during the initial charge and discharge processes and changes in cell parameters a and c obtained by refinement.

Figure S24 .
Figure S24.The simulation diagram of Li + concentration distribution and structure stress simulation diagram for LCO and DO-LCO at the bulk and surface with COMSOL software at 100 cycles.

Figure S26 .
Figure S26.Comparison of ex-situ EPR results of LCO (a) and DO-LCO (b) at different states.

Figure S27 .
Figure S27.The fitted O 1s (a, b) and F 1s (c, d) XPS patterns for LCO at pristine and 4.6 Vcharged states, the insert lateral histogram shows the calculated CEI thickness and components.

Figure S28 .
Figure S28.The fitted O 1s (a, b) and F 1s (c, d) XPS patterns for DO-LCO at pristine and 4.6 V-charged states, the insert lateral histogram shows the calculated CEI thickness and components.According to the content of lattice oxygen, the thickness of cathode electrolyte interphase (CEI) is calculated by the formula D=ln(Ia/Ib)λadscosθetfactor.Where Ia/Ib represents the ratio of the pristine/soaked lattice oxygen atom percentage; λads represents the average free path of photoelectrons passing through the whole CEI film at the lattice oxygen peak, which is 2.84 nm; θe is the emission angle and the value is 0; tfactor refers to the correction factor of surface facial mask on spherical particles with a value of 0.67.Casa XPS software was used for fitting and quantitative analysis of the test data.The spectral data energy correction is to correct the peak potential energy value corresponding to C-C bond (conductive additive) in the C 1s spectrum to 284.8 eV.Based on the O 1s and F 1s XPS spectra, the ratio of different components in the original/soaked electrodes was obtained by the area of fitting.To more intuitively compare the CEI content of the pristine and soaked electrodes, and exclude the influence caused by the difference of the original electrode surface state, we carried out a quantitative analysis of each peak.Then, we obtained the relative percentage of CEI

Figure S29 .
Figure S29.The in-situ Raman results of LCO during the initial charge and discharge processes and selected in-situ Raman data from the 800-1000 cm -1 .

Figure S30 .
Figure S30.The in-situ Raman results of DO-LCO during the initial charge and discharge processes and selected in-situ Raman data from the 800-1000 cm -1 .

Figure S31 .
Figure S31.Contents of transition metal Co ions in electrolytes for LCO and DO-LCO materials after 100 cycles.The dissolved Co contents in electrolyte after cycles were detected using ICP spectrometer.Both electrolytes were collected after the LCO and DO-LCO cathodes are charged/discharged 100 cycles.The Co content is 80 ppm in electrolyte for LCO, while only 15 ppm of the Co content is in electrolyte for DO-LCO, further proving that the diffusion optimization of Li + effectively inhibits the dissolution of Co (Figure S31).

Figure S32 .
Figure S32.HRTEM data for LCO electrode collected after 100 cycles, the twisted lattice is highlighted by red circle.

Figure S33 .
Figure S33.HRTEM data for DO-LCO electrode collected after 100 cycles, the clear lattice is highlighted by red rectangle.

Figure S34 .
Figure S34.The SEM morphology of LCO material with different number of cycles.

Figure S35 .
Figure S35.The SEM morphology of DO-LCO material with different number of cycles.

Figure S36 .
Figure S36.The cross-sectional SEM images of LCO material after 100 cycles.

Figure S37 .
Figure S37.The cross-sectional SEM images of DO-LCO material after 100 cycles.

Table S2 .
The refined crystal structure parameters of LCO sample using X-ray diffraction.

Table S3 .
The refined crystal structure parameters of 0.3% DO-LCO sample using X-ray

Table S4 .
The refined crystal structure parameters of 0.5% DO-LCO sample using X-ray

Table S6 .
Satellite Relative Area (%) of the Main Peak from Co 2p3/2 Spectra for LCO andLGTP-LCO at different etching depth.

Table S7 .
Comparison of various electrochemical properties (initial capacity and capacity retention) of LCO at different current densities for hundreds of cycles in literatures.

Table S9 .
The Atomic Percentages (%) of each component in the state of pristine and charge to 4.6 V from O 1s and F 1s XPS spectra of LCO and DO-LCO cathodes.