Fast Lithium Intercalation Mechanism on Surface‐Modified Cathodes for Lithium‐Ion Batteries

Enhancing the understanding of fast lithium intercalation on cathode surfaces modified by oxides is crucial for the development of electrode materials that offer high‐power and long‐life operation. Herein, lithium transfer is elucidated by directly observing the structural changes within the cathode, through the interface, and into the electrolyte using in situ neutron reflectometry (NR). Two films are studied—a Li2ZrO3‐modified and an unmodified LiCoO2 film—and it is found that the modified film exhibits a superior rate capability. In situ NR studies indicate that the surface modification facilitates the formation of a dense cathode–electrolyte interphase (CEI), primarily composed of inorganic species. In contrast, the unmodified surface is covered by a relatively sparse and electrolyte‐impregnated CEI. These structural observations suggest that lithium desolvation during intercalation primarily occurs on the CEI and LiCoO2 surfaces for the modified and unmodified films, respectively. Fast desolvation of lithium on the CEI may contribute to the superior rate capability of the surface‐modified cathodes. This suggests a mechanism of fast intercalation achieved by surface modification of low ionically conductive oxides. Simultaneous chemical composition and morphological information is a powerful way to elucidate the dynamics at cathode/liquid electrolyte interfaces suitable for high‐power operation.


DOI: 10.1002/aenm.202302402
battery technology for electric vehicles and plug-in hybrid electric vehicles. [1,2]enerally, the reaction potentials and current densities of lithium deintercalation and intercalation determine power characteristics.The reaction potential is intrinsically related to the change in Gibbs energy of the cathode material and kinetical changes depending on the charge and mass transfer resistances of the lithium (de)intercalation.Charge and mass transfer reactions are often limited owing to the formation of phases with low ionic conductivity at the cathode-organic electrolyte interface.This is caused by irreversible surface crystal structure changes associated with atomic displacements and/or dissolution of transition-metal ions [3,4] and the decomposition of electrolyte species to produce cathodeelectrolyte interphases (CEIs). [5,6]he cycle stability and rate capability of lithium battery cathodes have been significantly improved by modifying the cathode surface with oxide materials such as ZrO 2 , [6][7][8] MgO, [9,10] Al 2 O 3 , [5,11,12] AlPO 4 , [13] Li 2 ZrO 3 , [14,15] and Li 3 PO 4 . [16]These modified layers reduce the direct contact of the cathode with the electrolyte, suppressing the excessive formation of low-conductivity phases during charge-discharge cycles. [9,14,16]The stabilization of this interface is a widely accepted mechanism that explains the superior cycle stability of surface-modified cathodes compared to unmodified cathodes.However, most modification materials possess very low electronic and ionic conductivities, which raises the question of why surface modification also improves the rate capability at the interface.Some studies have proposed that the components of the modification layer change via (electro-)chemical reactions with electrolyte species, leading to the formation of an ionically conductive CEI layer. [5,6,13]This highlights the importance of understanding interfacial structural changes under battery operation to gain a further understanding the reaction rate of lithium (de)intercalation at the modified cathode-organic electrolyte interface.However, structural changes in each interfacial region have been observed using different methods, and their impacts on the interfacial properties have been discussed independently.24] Furthermore, changes in the interfacial structure on the electrolyte side have rarely been investigated, although these changes should affect the lithium transfer, desolvation, and adsorption processes during lithium intercalation.To acquire a unified understanding of interfacial phenomena, the changes in the interfacial structures occurring from the electrode side to the electrolyte side during battery reactions have to be investigated using a single technique.
Neutron reflectometry (NR) is a powerful technique that enables a nanoscopic observation of solid-liquid interfaces as a function of depth. [25]Neutrons can probe electrochemical interfaces due to their strong penetrating power through the substrate.Moreover, lithium can be sensitively detected by neutrons because the neutron scattering length of lithium is comparable with the other elements in the electrode, whereas it is small and nearly undetectable by X-rays.The depth profile of the scattering length density (SLD) obtained by analyzing the reflectivity profile can reveal the ionic distribution from inside the cathode, through the interface, and into the electrolyte.[28] In this study, we report the interfacial structural changes at LiCoO 2 -organic electrolyte interfaces, characterized by in situ NR, using a LiCoO 2 (104) epitaxial film model electrode.The LiCoO 2 (104) film, fabricated on SrRuO 3 (100)/Nb:SrTiO 3 (100) by pulsed laser deposition (PLD), exhibited a surface roughness of <1 nm, thus providing a simplified reaction field. [3,22,29,30]The SrRuO 3 layer acted as a current collector for investigating the rate capability of lithium intercalation quantitatively. [16,22]Significant improvements in the rate capability of the LiCoO 2 (104) film can be achieved through surface modification with a fewnanometer-thick Li 2 ZrO 3 layer. [15]Studies utilizing electrochemical impedance spectroscopy have shown that this surface modification reduces the activation energy required for lithium-ion transfer. [15]However, the role of surface modification in facilitating the fast lithium transfer remains unclear due to the lack of structural information.
To address this, we observed the interfacial structural changes directly at both the unmodified LiCoO 2 -electrolyte interface and the Li 2 ZrO 3 -modified LiCoO 2 -electrolyte interface, utilizing depth profiles of neutron SLD (nSLD) taken by NR.The model interfaces constructed with the atomically flat cathode allowed analysis of the chemical species and morphology at the interface on the order of a few nanometers.An interfacial layer on the electrolyte side of the interface was formed during the initial cycling, and its effect on the rate capability of lithium intercalation is discussed.

Effect of Surface Modification on Lithium Intercalation Rate
Figure 2a,b depicts the constant-current charge-discharge curves of the unmodified and Li 2 ZrO 3 -modified LiCoO 2 films collected at different discharge current densities.The retentions of the discharge capacity and average voltage are shown in Figure 2c,d.The unmodified and modified films delivered similar discharge capacities of ≈125 mAh g −1 under 100 C (300 μA cm −2 ), corresponding to the reversible lithium (de)intercalation of LiCoO 2 .The high-capacity retention is ascribed to the short transfer distance of electrons and lithium ions within the 26 nm thick film electrodes.For the unmodified film, the discharge voltage of V ave during the first cycle was 3.53 V.The V ave values gradually decreased to 3.49 V by the fifth discharge, primarily caused by voltage lowering at the end of the discharge.In subsequent cycles, the V ave values increased as the current density decreased.Conversely, the modified film exhibited a higher discharge voltage of 3.80 V with no significant decrease during the five cycles under 100 C. XRD analyses confirmed no substantial changes in the crystal structures between the unmodified and modified films in the as-grown condition.Moreover, a highly reversible intercalation capacity during twenty charge-discharge cycles revealed  S1, Supporting Information).Thus, the Li 2 ZrO 3 -modified LiCoO 2 /electrolyte interface exhibited superior lithium intercalation reaction rates compared with the unmodified film/electrolyte interface.

Factors Associated with Intercalation Rate
Figure 3a,b presents the Nyquist plots of electrochemical impedance spectroscopy (EIS) for unmodified and Li 2 ZrO 3modified LiCoO 2 (104) films collected at various discharge voltages during the second cycle.For both film types, a distorted semicircle is evident at 4.1, 4.0, and 3.9 V, with its diameter increasing from 4.0 to 3.9 V.At 3.7 and 3.5 V, impedance values in low-frequency regions below 10 Hz show a drastic increase, while the portion of the semicircle observed above 10 Hz remains relatively unchanged with voltage variations.This suggests that the distorted semicircles above 3.9 V comprise two components with close time constants.Accordingly, we refined the Nyquist plots using an equivalent circuit with two RC components of R LF CPE LF and R HF CPE HF .This equivalent circuit is depicted in Figure 3c.R sol , Z w , and CPE int correspond to the electrolyte resistance, Warburg impedance, and interfacial capacitance, respectively.Figure 4 shows the variations in the R LF and R HF values refined for both the unmodified and modified films during the second charge-discharge cycle.For both films, the R LF value decreases from 3.9 to 4.1 V during charge and then increases reversibly from 4.1 to 3.9 V during discharge.As the (de)intercalation rate changes with the Li + concentration in the cathode material (states of (dis)charge [33] ), R LF can be associated with the interfacial resistance of lithium (de)intercalation at the cathode-electrolyte interface.In contrast to R LF , the R HF value remains relatively stable with changes in the cell voltage for both films.Considering the minor resistances in the electrochemical processes included in the electrolyte and at the lithium anode (Figure S2, Supporting Information), the cathode side may include another resistive component with a high time constant, such as lithium transfer in the CEI layer (R CEI ) and the electronic contact resistance between the Nb:SrTiO 3 substrate and the current collector (R e ).From reflectometry analyses, the total thickness of the CEI layers formed on both LiCoO 2 films is estimated to be less than 10 nm (see Sections 2.4 and 2.5 for details).If the ionic conductivity of a Li 2 ZrO 3 -based CEI layer is assumed to be comparable to Li 2 ZrO 3 (10 −7 S cm −1 [34] ), the resistance of lithium transfer in the thin CEI layer is calculated to be ≈15 Ω cm 2 .This results in a polarization of 5 mV under 100 C operation.R HF could not be identified as R CEI .Hence, the electronic contact resistance R e could be the primary resistive component of the semicircles at high frequencies.Overall, the R LF values were higher than the R HF under all conditions for both films.This implies that the lithium intercalation at the cathode interface is the rate-determining step in the thin-film batteries.
Notably, the R LF and R HF values of the modified film are marginally higher than those of the unmodified film in the second cycle.This does not explain the faster lithium intercalation depicted in Figure 2. To address this discrepancy, we calculated the IR drops (ΔV) using the total resistance (R sol + R LF + R HF ) analyzed by EIS and compared them to the values obtained at the onset of discharge in the charge-discharge curves (refer to Figure S3, Supporting Information).For the modified film, the ΔV value at 100 C was estimated to be 0.34 V, higher than the observed value in the discharge curve (0.18 V).Conversely, the unmodified film's ΔV value (0.29 V) is lower than that observed in the discharge curve (0.46 V).The EIS measurement, which uses a small AC voltage, can detect interfacial phenomena in a near-equilibrium state, whereas the constantcurrent discharge measurement detects phenomena highly biased in the discharge direction.This behavior is similar to that reported for the interfacial resistance between the lithium anode and the electrolyte, which can vary depending on the eval-uation by EIS or DC measurements. [35]Electrochemical reactions are mainly classified into charge transfer and mass transfer (diffusion) processes. [33,36]The reaction rate is limited by the charge transfer at low current density operation because the redox species are supplied sufficiently.At high current density operation, the reaction rate is restricted by the mass transfer of redox species. [37]The CEI structures can change in situ in the (de)intercalation reaction field, which may also differ when AC and DC signals are applied.Structural changes in CEI under high current density operation may affect the lithium-ion transfer rate, resulting in different interfacial resistances evaluated by AC and DC methods.
The cathode-CEI-electrolyte interface is believed to encompass several elementary processes of lithium intercalation: i) adsorption and desolvation of solvated lithium at the CEI surface, ii) lithium transfer in the CEI layer, and iii) lithium intercalation accompanied by electron transfer at the cathode-CEI interface. [38]he rates of processes (i) and (ii) are likely dependent on the chemical composition and structure of the CEI layer, which are altered by surface modification.The rate of the lithium intercalation process (iii) relies on the reaction rate constant and the lithium concentrations at the interfaces of the cathode and electrolyte sides. [39,40]We can postulate that there is no significant deterioration of the LiCoO 2 surface for the unmodified and modified films, as the charge-discharge capacities remain consistent during the cycling in the voltage regions of 4.2 and 3.0 V (Figure 2c).This suggests that the rate of the lithium intercalation process primarily depends on the electrolyte side interface.Regardless of which process is the rate-limiting step at the interface, the structures of the interfacial layer are believed to determine the intercalation reaction resistance.Given that the resistive component changes under lithium intercalation, the structures of the CEI and liquid electrolyte formed at the interface during lithium intercalation could be altered by the surface modification of Li 2 ZrO 3 , which can provide an interfacial reaction field for fast intercalation.

CEI Components Analyzed by Ex Situ Neutron Reflectometry
To gain a deeper understanding of the effect of the CEI layer on rate capability, we investigated the structural changes at the elec- trode surface using ex situ NR and X-ray photoemission spectroscopy (XPS).Figure 5 presents the NR spectrum fitting results and nSLD profiles of the unmodified and Li 2 ZrO 3 -modified LiCoO 2 (104) films after the third discharge.In comparison with the pristine films (Figure 1d), a double-surface layer model comprising layer1/layer2/LiCoO 2 /SrRuO 3 /SrTiO 3 fits closely with the reflectivity curve for both films (refer to Figure S4, Supporting Information).Table 2 summarizes the refined thickness, nSLD, and roughness of the LiCoO 2 and surface layers.Both the unmodified and modified LiCoO 2 layers show a similar nSLD (≈3.60 × 10 −4 nm −2 ) and roughness (≈0.5 nm).This indicates no significant structural differences between the unmodified and modified LiCoO 2 .For the unmodified LiCoO 2 , the thicknesses (and nSLDs) were 2.2 nm (3.19 × 10 −4 nm −2 ) and 1.5 nm (2.18 × 10 −4 nm −2 ) for the outer and inner layers, respectively.The modified LiCoO 2 was coated with comparable outer and inner layers, each with a thickness of 1.5 nm and an nSLD of 3.09 × 10 −4 and 2.42 × 10 −4 nm −2 , respectively.
The surface layer in the pristine condition is reportedly replaced by CEI species, primarily formed by the decomposition of electrolyte species. [41,42]For alkyl carbonate-based electrolytes with LiPF 6 salt, lithium-containing fluorides, (fluoro)phosphates oxide, carbonates, alkoxides, and poly(ethylene carbonate) are thought to be products of oxidative decomposition and/or nucleophilic reactions. [43]XPS analyses confirmed the formation of LiF, Li 3 PO x F y , Li 2 CO 3 , LiOH, CH 2 OCO 2 Li, and Li 2 ZrO x F y , which can be attributed to decomposition products of the LiPF 6 containing ethylene carbonate (EC)-diethyl carbonate (DEC) electrolyte (refer to Figure S5, Supporting Information).Considering the nSLD values of the CEI components (see Table S1,S4, Supporting Information), the outer surface layer primarily comprises Li 2 CO 3 (3.37 × 10 −4 nm −2 ) and CH 2 OCO 2 Li (3.01 × 10 −4 nm −2 ),and the bottom surface layer consists mainly of LiF (2.30 × 10 −4 nm −2 ), Li 3 PO x F y (≈2.8 × 10 −4 nm −2 ), and LiOH (0.06 × 10 −4 nm −2 ).The inner layer on the modified surface shows a slightly higher nSLD than the unmodified surface, possibly due to Li 2 ZrO 3 -based components such as Li 2 ZrO x F y (≈3.3 × 10 −4 nm −2 ).The double-surface layer model aligns with prior studies suggesting that inorganic CEI species initially form at cathode surfaces, followed by stacking organic CEI species on the surface. [44,45]oth NR and XPS analyses detected no significant alterations in the components or thicknesses of the CEI layers between the unmodified and modified LiCoO 2 surfaces.During the ex situ analyses, the cathode surface was washed using the DEC solvent and dried in Ar to eliminate the solvent.These procedures could potentially alter the structures of the CEI, making identifying the differences challenging.No CEI layer was observed in ex situ STEM images of the unmodified LiCoO 2 surface covered by SiO 2 and Ti protection layers, which also demonstrates the difficulty of observing CEIs formed in the initial process without damage in ex situ measurements (refer to Figure S1, Supporting Information).Therefore, we performed in situ NR analyses to eliminate any influences from the sampling processes to explore the electrode-electrolyte interface during the initial reaction process.The thickness and roughness of the LiCoO 2 layer (≈25 and 1 nm, respectively) remain stable during the charge and discharge cycles for both films, exhibiting only minor changes.These findings suggest that the reversible lithium (de)intercalation occurs without degrading the LiCoO 2 phase.This is consistent with the high-capacity retention observed in the charge-discharge curves (Figure 2), reinforcing the validity of the double-interfacial-layer model.

Inner Layer Structures
We now focus on the interfacial layers.Following cell construction, an inner layer of thickness 2.6 nm was formed on the unmodified LiCoO 2 surface.The nSLD and roughness at the LiCoO 2 -inner-layer interface were 4.42 × 10 −4 nm −2 and 0.9 nm, respectively.No significant thickness, nSLD, or roughness changes were observed during the charge and discharge processes.This suggests that the inner layer, which forms after immersion in the electrolyte, [22] remains electrochemically stable during subsequent processes.Previous research has shown that surface species such as LiOH and Li 2 CO 3 chemically interact with impurities like HF and H 3 PO 4 in the electrolyte, generating LiF and Li 3 PO 4 , thus forming an inorganic CEI layer. [46]However, the observed nSLD of 4.53 × 10 −4 nm −2 after the third discharge was considerably higher than the theoretical values for LiF (2.30 × 10 −4 nm −2 ) and Li 3 PO 4 (2.83 × 10 −4 nm −2 ).The nSLD value of the electrolyte, 5.67 × 10 −4 nm −2 , could inflate the overall nSLD value, as the electrolyte might fill the cavities within the CEI layer.Consequently, the inner layer formed on the unmodified surface likely comprised both inorganic CEI species and liquid electrolyte.The inner layer observed by ex situ NR exhibited an nSLD value of 2.18 × 10 −4 nm −2 , similar to inorganic CEI species.In ex situ NR, the cell was disassembled, the cathode sample was washed gently with DEC, and any remaining electrolyte in the CEI layer was replaced by DEC, which was subsequently evaporated during the drying process.This ensures that the primarily detectable species in the inner CEI from ex situ NR and XPS should be insoluble inorganics.The combined analyses using both ex situ and in situ NR reveals that the components of the interfacial layer under electrochemical conditions differ from those detected by conventional ex situ measurements.At the Li 2 ZrO 3 -modified LiCoO 2 surface, the inner layer exhibited thickness and nSLD values of 2.6 nm and 2.96 × 10 −4 nm −2 , respectively, following cell construction.These values exhibit minimal change after the third discharge (3.2 nm and 2.91 × 10 −4 nm −2 ), which indicates that the inner CEI forms after soaking in the electrolyte, facilitated by chemical reactions between Li 2 ZrO 3 and the electrolyte species.The nSLD values are significantly lower than those observed for the unmodified LiCoO 2 and correspond well with the inner CEI values at the modified LiCoO 2 detected from ex situ NR (2.42 × 10 −4 nm −2 ).This observation suggests that the modified Li 2 ZrO 3 layer promotes the formation of a dense, inorganic CEI with relatively few cavities.This dense CEI then facilitates a lower permeation rate of the liquid electrolyte into the inner layer under electrochemical conditions.The enhancement of inorganic CEI growth through surface modification has also been reported on a ZrO 2 -modified LiCoO 2 surface. [6]hus, in situ NR analyses reveal that surface modification affects both the density of the CEI layer and the amount of liquid electrolyte impregnation.These alterations could change the interfacial diffusion process of lithium ions during intercalation.Such structural information proves challenging to detect using ex situ NR and XPS analyses.

Outer Layer Structures
A discernible outer layer, over 5 nm thick, is observed after cell construction in unmodified and Li 2 ZrO 3 -modified LiCoO 2 .This outer layer persists during charging and discharging in both types of LiCoO 2 .The interfacial roughness between the outer layer and the electrolyte approximates the outer layer thickness, resulting in an nSLD gradient within the outer layer regions.Intriguingly, the measured thickness and roughness are considerably greater than those observed via ex situ NR.Under electrochemical conditions, the outer layer exhibits distinct components and/or morphologies differing from those observed under dry conditions.The outer layer at the unmodified LiCoO 2 surface presents an nSLD value of 4.81 × 10 −4 nm −2 before cycling.The nSLD increases to 5.50 × 10 −4 nm −2 at a charged state of 4.2 V and decreases to 4.67 × 10 −4 nm −2 at a discharged state of 3.2 V.This suggests a shift in the chemical composition of the outer layer during the charge and discharge reactions.In contrast, the outer layer at the modified LiCoO 2 site demonstrates a lower pre-cycle nSLD of 3.46 × 10 −4 nm −2 compared to the unmodified LiCoO 2 .
The nSLD marginally increases to 3.69 × 10 −4 nm −2 at 4.2 V and remains relatively unchanged at 3.2 V. Similar to the inner layer, changes in the interfacial structure within the outer layer differ between the unmodified and modified LiCoO 2 .Three possible interfacial models are proposed to account for the sloped nSLD profiles of the outer layers: i) a CEI layer with a composition gradient, ii) a distribution of ionic concentration within the electrolyte, and iii) a depth variation in the ratio of solid CEI to liquid electrolyte.These models can be analyzed as follows.First, it has been reported that inorganic CEI species initially form at cathode surfaces, followed by an accumulation of organic CEI species stacked on the surface. [44,45]Organic CEI species, such as lithium ethylene decarbonate (CD 2 OCO 2 Li) 2 (nSLD = 5.83 × 10 −4 nm −2 ), lithium ethoxide (C 2 D 5 OLi) (nSLD = 4.78 × 10 −4 nm −2 ), and lithium ethyl carbonate (C 2 D 5 OCO 2 Li) (nSLD = 4.27 × 10 −4 nm −2 ), exhibit relatively high nSLDs exceeding 4 × 10 −4 nm −2 in comparison to inorganic CEI species (see Table S3,S7, Supporting Information).The gradual increase in organic to inorganic CEI species ratio could correspond to the sloped nSLD.In this scenario, the sloped nSLD profiles could be obtained via ex situ NR due to the low solubility of inorganic and organic CEI species in DEC used in the washing process.However, the outer layers detected by ex situ NR demonstrate considerably smaller thicknesses without any gradient in nSLD for both films.This discrepancy between in situ and ex situ NR results suggests that the CEI gradient model may not fully explain the observed profiles.
Second, the ionic distribution within the electrolyte typically forms at the electrode-electrolyte interface (electric double layer) as ions move to offset potential differences when the electrode contacts the electrolyte.Traditional electrochemical models predict the thickness of this ionic distribution to be less than 1 nm for organic electrolytes in lithium-ion batteries.However, recent reports suggest that experimental thicknesses in highly concentrated solutions considerably exceed these predictions. [47]This indicates the formation of an ionic distribution of Li + and PF 6 − and their solvated structures with d-EC and d-DEC at the electrolyte interface, potentially leading to a gradient in nSLD.The electric double layer is only observed in electrode-electrolyte systems, explaining the absence of a gradient in the dried state as seen in ex situ NR.To validate this model, we evaluated the nSLD values of electrolytes with different ratios of Li + , PF 6 − , d-EC, and d-DEC using their molecular volumes (see details in Figure S7,S9, Supporting Information).While the nSLD value decreases with an increase in Li + concentration, it does not reach the lowest nSLDs of the outer layers (3.46 × 10 −4 nm −2 ) observed for the modified film.Hence, models (i) and (ii), which assume pure CEI or liquid layers, do not satisfactorily explain the in situ and ex situ NR results.
Third, we assumed an outer CEI layer with cavities filled with liquid electrolytes in the model (iii).The volumetric ratio of the CEI to the liquid electrolyte should decrease with distance from the electrode surface, as an electrochemical reaction at the electrode surface forms the CEI.Because the liquid electrolyte has a higher nSLD than organic and inorganic CEI species, the nSLD value of the outer layer increases from the surface to the electrolyte bulk.An outer CEI layer with many cavities is considered physically unstable due to the weak interaction between the CEIs.Consequently, some outer CEI layers may detach from the surface during the DEC washing process in ex situ NR.Alternatively, the CEI layer may densify as the DEC solvent evaporates from the cavities during drying.These factors can account for the relatively small thicknesses and flat nSLD profiles of the outer layer, as observed by ex situ NR.We can thus conclude that, among the three models, model (iii) provides the most reasonable explanations for the in situ and ex situ NR results.
Building on the model (iii), we explore the structural transformations in the outer layer during charging and discharging processes.For the unmodified LiCoO 2 , the nSLD increases when charged to 4.2 V.The organic CEIs develop at high voltages due to oxidative decomposition and/or nucleophilic reactions of the electrolyte species. [48]This increase in nSLD suggests the formation of organic CEI species within the outer layer without a corresponding increase in thickness.Lithium ethylene decarbonate (CD 2 OCO 2 Li) 2 , with a value of 5.83 × 10 −4 nm −2 , is a potential candidate.A decrease in nSLD at the discharged 3.2 V indicates the decomposition and/or removal of the organic species, as reported by PM-FTIR, [19] AFM, [49] XRR, [50] and XPS [51] analyses.The Li 2 ZrO 3 -modified LiCoO 2 exhibits a lower nSLD than the unmodified variant, suggesting a higher inorganic-to-organic CEI ratio and liquid electrolyte in the outer layer.Notably, the outer layer demonstrates no significant changes in nSLD and thickness during charging and discharging, mimicking the behav-ior of the inner layer.The interfacial structures at the modified LiCoO 2 establish a certain chemical composition and morphology during the initial stages of charging and discharging.The dense inner CEI obstructs electronic contact between the LiCoO 2 and liquid electrolyte, which could hinder the further decomposition of electrolyte species to expand the CEI layers.Conversely, the less dense inner CEI formed at the unmodified LiCoO 2 leads to the growth and decomposition of the outer CEI during cycling.

Lithium Transfer Mechanism at Interfaces
Figure 8 presents a schematic representation of interfacial structures for unmodified and Li 2 ZrO 3 -modified LiCoO 2 films under electrochemical conditions based on our in situ and ex situ NR and ex situ XPS measurements.For both types of LiCoO 2 , the interfacial structures are divided into inner and outer parts, depending on the distance from the LiCoO 2 surface to the electrolyte bulk.The inner and outer layers primarily consist of inorganic CEIs and a mix of organic CEI and electrolytes, respectively.The surface modification with Li 2 ZrO 3 facilitates the formation of inorganic CEI species, resulting in a dense inorganic CEI in the inner layer and a relatively high ratio of inorganic-to-organic CEIs in the outer layer.
The CEI layer thickness is evaluated to be <10 nm at both interfaces.Assuming an ionic conductivity of 10 −7 S cm −1 for the CEI layer, [52] the lithium transfer resistance in the CEI layer is calculated to be ≈15 Ω cm 2 .Hence, lithium migration in the CEI layer may not be a rate-determining step during lithium intercalation.Based on these findings regarding the interfacial structures, we can predict the lithium transfer processes during lithium intercalation as follows: (a) Unmodified LiCoO 2 : The cavities in the inner and outer layers are impregnated with the liquid electrolyte.Lithium ions can diffuse in electrolyte, which offers considerably higher lithium conductivity than CEI components.Lithium ions move quickly through the interfacial layers from the electrolyte to reach LiCoO 2 .The ions are then desolvated to be absorbed in the LiCoO 2 surface, followed by intercalation into the LiCoO 2 lattice.
(b) Li 2 ZrO 3 -modified LiCoO 2 : Lithium ions diffuse through the liquid electrolyte, impregnated in the outer layer and reaching the inner surface.Most lithium ions must be desolvated to diffuse into the inner layer because of the high density of the inner CEI.The desolvated lithium ions intercalate into the LiCoO 2 lattice through the solid-solid interface between the LiCoO 2 and inner CEI.
Our electrochemical investigations clarified the superior rate capability of lithium intercalation at the modified LiCoO 2electrolyte interface.The relatively slow lithium diffusion in the dense inner layer can be inferred from the interfacial structures.This inference redirects our attention to the differences in the lithium desolvation process.Desolvation primarily occurs at the inner CEI and LiCoO 2 surfaces for the modified and unmodified LiCoO 2 , respectively.The ability of the lithium-ion to become desolvated inherently depends on the solvent species and solvated structures in the electrolyte bulk.In addition, it should also be affected by the adsorption ability of desolvated Li on the solid surface.Recently, Wang et al. proposed that a partly desolvated lithium-ion is formed on SEI, drastically reducing desolvation energy. [53]Conversely, delighted Li 1−x CoO 2 has negative adsorption energy for EC dissociative adsorption on surface oxygen sites. [54]The EC adsorption reduces the oxygen sites at LiCoO 2 to absorb the desolvated lithium-ion.
From these models, we can postulate that desolvation occurs faster at the inner CEI surface than at the LiCoO 2 surface.Although the lithium-ion is desolvated on the inner CEI formed at the unmodified LiCoO 2 , the small contact area between the CEIs does not provide sufficient pathways for lithium to transfer to LiCoO 2 .This explains the superior rate capability of lithium intercalation at the modified LiCoO 2 -electrolyte interface.We should also consider the possibility that the interfacial structures in the inner and outer layers undergo significant change under high current operations, which cannot be tracked in situ NR.Nevertheless, these changes should be associated with the interfacial structures at the onset of cycling.Our results highlight a previously undefined role of surface modification in facilitating the interfacial structures for fast desolvation of lithium ions, leading to a higher lithium intercalation rate.This finding will inform new design concepts for surface modification materials.
In situ NR analyses have revealed that both surfaces are covered by inner and outer layers, with compositions and thick-nesses differing from those observed from ex situ XRR and XPS.This variation is primarily caused by impregnating the liquid electrolyte into the cavities within the CEIs, potentially leading to delamination and/or densification of the CEIs during the washing and drying process in ex situ experiments.
The inner layer at the modified surface is composed of inorganic species, facilitating the desolvation of lithium during intercalation on the inner CEI.Conversely, the unmodified surface is covered by a relatively sparse and electrolyte-impregnated CEI, and the desolvation mainly proceeds on the LiCoO 2 surface.Hence, the superior rate capability of the modified LiCoO 2 may originate from faster desolvation of lithium on the CEI compared to the LiCoO 2 surfaces.These findings suggest that surface modification can enhance the desolvation rate, as well as stabilize the electrode surface, providing a revised design principle for the cathode-liquid-electrolyte interface suitable for high-power operation.Our results have shown that direct observation of the interfacial structures during cycling by in situ NR, capable of simultaneously detecting chemical composition and morphological information, is a powerful method for elucidating the detailed mechanism of battery reactions.

Experimental Section
Sample Preparation and Characterization: SrRuO 3 , LiCoO 2 , and Li 2 ZrO 3 films were deposited on Nb:SrTiO 3 (100) substrate (size: 10 × 10 × 0.5 mm [Nb-doped 0.5 wt.%] and 20 × 20 × 5 mm [Nbdoped 0.2 wt.%], Crystal Base Co., Ltd.) using a PLD system consisting of a KrF excimer laser (Lambda Physik, COMPex201) with a wavelength of 248 nm and a vacuum chamber (PLAD-312, AOV Inc.).The SrRuO 3 (100) film as a buffer layer was deposited on the Nb:SrTiO 3 (100) substrate to improve the electronic conduction between the LiCoO 2 and the Nb:SrTiO 3 (100) substrate. [55]The thickness of the Li 2 ZrO 3 was controlled to be ≈2 nm to prevent excessive formation of interphases with poor ionic conductivity. [6,7]The synthetic conditions for each layer were the same as those used previously. [15]The deposited films were investigated by XRD (ATX-G, Rigaku Inc.) with Cu K 1 ( = 1.541Å) radiation.The lattice orientation was determined by out-of-plane and in-plane XRD techniques.XRR spectra were plotted as a function of the scattering vector, Q z = (4sin)/, where  is the incident angle, and  is the wavelength of the X-ray (0.1541 nm).The film thickness, roughness, and SLD from the XRR spectra were fitted through the Motofit software. [56]harge-Discharge Measurements: 2032-type coin cells were constructed with the unmodified and modified LiCoO 2 (104) film as the cathode, lithium metal as the anode, 1 mol dm −3 of LiPF 6 in EC-DEC (3:7 v/v) as the electrolyte, and a polypropylene separator.The LiCoO 2 (104) film size was 8 × 8 mm.An excess amount of lithium foil with a relatively large roughness was used to ensure that the reaction rate on the anode side was sufficiently faster than that on the cathode side.Constant-current chargedischarge tests were conducted in the 3.0-4.2V range using a potentiogalvanostat (TOSCAT, Toyo system).The current densities used were 1 C (3 μA cm −2 ) for charging and 1, 3, 5, 10, 20, and 100 C for discharging.The rate capability test was conducted for three or five cycles, each from 100 to 1 C, to minimize cycle degradation.
Electrochemical Impedance Spectroscopy Measurements: EIS was performed with coin cells during the second cycle in the frequency range of 5 MHz-5 mHz and an applied voltage amplitude of 10 mV (VMP-300 Potentiostat, BioLogic).The cells were charged and discharged to setting voltages of 3.5, 3.7, 4.0, and 4.1 V with a constant current of 1 C at 25 °C and held for 15 min at each voltage prior to the EIS measurements.
Ex Situ Neutron Reflectometry and X-Ray Photoemission Spectroscopy: Components of the CEI layers formed on the unmodified and modified LiCoO 2 (104) films were analyzed by ex situ neutron reflectometry (NR) and X-ray photoemission spectroscopy (XPS).[59] The charge-discharge cycles were performed with the same conditions for in situ NR.The coin cells were dissembled in an Arfilled glovebox, and the samples were washed with DEC.The ex situ NR was performed in a vacuum using a self-made chamber to prevent chemical reactions of the CEI with air species.The NR spectra were collected using the single-frame mode at 0.3°, 0.6°, and 1.3°with a footprint of 10 × 10 mm, and the individual data points were combined.C 1s, O 1s, F 1s, P 2p, and Zr 3d XPS spectra were conducted using monochromatic Al K (1486.6 eV) as the X-ray source (ESCA1700R, Ulvac-phi Inc.).An airtight sample holder was used to transfer samples between the glove box and the XPS chamber.Peak fitting of the detected spectra was performed using the CasaXPS software. [60]All spectra were recorded with a pass energy of 23.5 eV, and the binding energy was calibrated using the C 1s peak of carbon contamination at 284.8 eV.
In Situ Neutron Reflectometry: An electrochemical cell composed of the unmodified and modified LiCoO 2 /SrRuO 3 /SrTiO 3 cathode, a Li anode, and 1 mol dm −3 of LiPF 6 containing EC+DEC (3:7) as the electrolyte. [23,59]he EC and DEC solvents were deuterated (98%) to prevent incoherent scattering (80.27 and 2.05 for 1 H and 2 H, respectively).In situ neutron reflectometry (NR) spectra were conducted on SOFIA using the single-frame mode with a footprint of 15 × 15 mm.NR spectra were collected at open circuit voltage after assembling, charged at 4.2 V, and discharged at 3.2 V at the third cycle.The cell voltages were changed by linear sweep voltammetry at 1 mV s −1 between the setting voltages and were fixed during the NR measurements.In spectrum fitting using the Motofit software, [56] different numbers of layers were investigated to determine the interfacial structures (see Figure S5, Supporting Information).The nSLDs were estimated from composition and theoretical density for solid materials and composition and mole volumes of each species for liquid electrolytes to consider the components in the layers (see details in Supporting Information).

Figure 2 .
Figure 2. Charge-discharge curves of a) unmodified LiCoO 2 and b) Li 2 ZrO 3 -modified LiCoO 2 at different discharge current densities from 100 C (300 μA cm −2 ) to 1 C (3 μA cm −2 ).The charge current density was fixed to be 1 C. c) Variations of discharge capacity and d) average discharge voltage obtained from in (a,b).Variation in discharge capacity observed for modified LiCoO 2 at 100 C operation was due to the time resolution of the charge-discharge equipment.

Figure 3 .
Figure 3. Nyquist plots of a) unmodified LiCoO 2 and b) Li 2 ZrO 3 -modified LiCoO 2 films after being discharged to different voltages during the second cycle.c) Equivalent circuit is used for spectrum fitting.

Figure 4 .
Figure 4. Variations of a) R LF and b) R HF values refined for unmodified and Li 2 ZrO 3 -modified LiCoO 2 at the second charging and discharging.

Figure 5 .
Figure 5. a) Observed and calculated NR spectra of the unmodified LiCoO 2 and the Li 2 ZrO 3 -modified LiCoO 2 after the third discharge and b) refined nSLD profiles of the unmodified and Li 2 ZrO 3 -modified film.The ex situ NR spectra were calculated by the double-surface layer model of SrRuO 3 /LiCoO 2 /layer2/layer1.

Figures 6 and 7
Figures 6 and 7 present the in situ NR fitting results and nSLD profiles of the unmodified and Li 2 ZrO 3 -modified LiCoO 2 (104) film after cell construction and at charged and discharged states during the third cycle.In line with the ex situ NR analysis, the double-interfacial layer model (SrRuO 3 /LiCoO 2 /inner layer/outer layer/electrolyte) corresponds closely with the reflectivity spectra of both samples (see Figures S6 and S7, Supporting Information).The refined thickness, nSLD, and roughness of each layer are summarized in Table3.At the charged state of 4.2 V, the unmodified and modified LiCoO 2 layers exhibit nSLD values of 3.88 × 10 −4 and 3.83 × 10 −4 nm −2 , respectively, which exceed those observed after cell construction.This corresponds to the formation of a lithium-deficient phase, Li 1−x CoO 2 , likely due to lithium deintercalation during the charging process, as natural lithium (7Li:6Li = 92.5:7.5)possesses a negative coherent scattering length for neutrons (−1.9 fm).The nSLD values decrease reversibly at the discharge state, confirming the lithium intercalation into Li 1−x CoO 2 .The thickness and roughness of the LiCoO 2 layer (≈25 and 1 nm, respectively) remain stable during the charge and discharge cycles for both films, exhibiting only minor changes.These findings suggest that the reversible lithium (de)intercalation occurs without degrading the LiCoO 2 phase.This is consistent with the high-capacity retention observed in the charge-discharge curves (Figure2), reinforcing the validity of the double-interfacial-layer model.

Figure 6 .
Figure 6.a) Observed and simulated NR spectra and b) refined nSLD profiles of the unmodified LiCoO 2 film after cell construction (3.3 V) and after charged to 4.2 V and discharged to 3.2 V at the third cycle.The in situ NR spectra were calculated by the four-layer model of SrRuO 3 /LiCoO 2 /inner layer/outer layer.

Figure 7 .
Figure 7. a) Observed and simulated NR spectra and b) refined nSLD profiles of the Li 2 ZrO 3 -modified LiCoO 2 film after cell construction (2.1 V) and after charged to 4.2 V and discharged to 3.2 V at the third cycle.The in situ NR spectra were calculated by the four-layer model of SrRuO 3 /LiCoO 2 /inner layer/outer layer.

Table 1 .
Refined thickness, SLD, and roughness of unmodified and Li 2 ZrO 3 -modified LiCoO 2 .The XRR spectra and fitting model are shown in Figure 1d.

Table 2 .
Refined thickness, SLD, and roughness of unmodified and Li 2 ZrO 3 modified LiCoO 2 after the 3rd discharge.The ex situ NR spectra and fitting model are shown in Figure 5a.

Table 3 .
Refined thickness, SLD, and roughness of unmodified and Li 2 ZrO 3 -modified LiCoO 2 before the cycle and after the 3rd charging and discharging.The in situ NR spectra and fitting model are shown in Figures 6a,7a.