Electrical Double Layer Formation at Intercalation Cathode–Organic Electrolyte Interfaces During Initial Lithium‐Ion Battery Reactions

Information on the cathode/organic–electrolyte interface structure provides clues regarding the rate and reversibility of lithium intercalation reactions in lithium‐ion batteries. Herein, structural changes within the LiCoO2 electrode, throughout the interphase region, and in the LiPF6/propylene carbonate electrolyte are observed concurrently by in situ neutron reflectometry. The formation of an electrical double layer (EDL) during the early stages of charging and discharging is investigated and compared with that at an intercalation‐inactive Nb:SrTiO3 electrode. At the intercalation‐inactive interface between Nb:SrTiO3 and the electrolyte, a voltage‐dependent ionic distribution corresponding to the EDL forms on the electrolyte side without the formation of a cathode/electrolyte interphase (CEI) layer. In contrast, at the intercalation‐active LiCoO2/electrolyte interface, a CEI layer forms immediately after cell construction, and the ionic distribution in the electrolyte formed outside the CEI layer scarcely changes upon voltage application. The CEI/electrolyte interface is shielded from potential changes by the electronically insulating CEI; therefore, structural changes in the EDL are restricted. This supports the prevailing understanding that the CEI layer defines the rates of solvation/de‐solvation and adsorption/desorption reactions of lithium ions.


Introduction
Lithium-ion batteries operate via electrochemical intercalation, whereby lithium ions are inserted into and extracted from gaps in nanostructures of the active material surfaces and interphases observed in situ differ significantly from those determined via ex situ (postmortem) studies. [8]3b,c] A remaining challenge in the research on the interfacial region of batteries is to elucidate the structure of the electric double layer (EDL) formed on the electrolyte side of the interface and to understand its influence on the elementary processes of lithium intercalation.The EDL comprises the distribution of charged ions that form in an interfacial region to mitigate the potential difference between different materials. [9]The lithium ions in the EDL exhibit activities and solvent-coordinated structures different from those in the bulk electrolyte, and these differences impact the rates of diffusion, de-solvation, and adsorption of solvated lithium ions during lithium intercalation. [10]However, few experimental techniques can directly probe EDL structures because most methods for nanoscale measurement require high-vacuum conditions.Furthermore, the majority of experimental studies have been conducted on metal electrode-liquid electrolyte interfaces. [7,11]Thus, EDL research has been dominated by studies using computational methods such as molecular dynamics, [12] which could not give direct evidences regarding the interfacial structures.
Neutron reflectometry (NR) is a powerful technique that enables nanoscopic observation of solid-liquid interfaces as a function of depth. [13]Neutrons are capable of probing electrochemical interfaces because of their ability to penetrate the substrate.Moreover, lithium can be sensitively detected by neutrons because the neutron scattering length of lithium is comparable to that of the other elements in the electrode, whereas it is small and nearly undetectable by X-rays.The depth profile of the scatteringlength density (SLD) obtained by analyzing the reflectivity profile reveals the ionic distribution from inside the cathode, through the interface, and into the electrolyte.Despite these analytical advantages, previous studies on electrode-electrolyte interfaces using NR have focused solely on solid-state interphases formed on intercalation electrode surfaces. [14]n this study, we report the interfacial structural changes at oxide electrode-organic electrolyte interfaces characterized by in situ NR.Atomically flat samples of intercalationinactive Nb:SrTiO 3 (100) substrates and intercalation-active LiCoO 2 (104) epitaxial films were used as model electrodes.The Nb:SrTiO 3 (100) substrate provides a simplified electrochemical reaction field for investigating EDL formation by eliminating lithium intercalation and interphase-layer formation.Subsequently, we discuss the interfacial structures formed on the intercalation electrode during initial charge-discharge processes based on a comparative analysis of the interfacial structures at the Nb:SrTiO 3 (100)/electrolyte and LiCoO 2 (104)/electrolyte interfaces.

Fabrication of Nb:SrTiO 3 (100) Flat Surface
A Nb:SrTiO 3 (100) substrate (20 × 20 × 5 mm) was etched in a buffered HF solution, rinsed with distilled water, and an- nealed at 1000 °C under vacuum to produce an atomically flat surface. [15] Step-terrace structures were confirmed in the atomic force microscope (AFM) image of the etched Nb:SrTiO 3 (100) surface, as shown in Figure 1c.Thus, an atomically flat Nb:SrTiO 3 (100) surface has been successfully prepared by etching.The etched Nb:SrTiO 3 (100) substrate provides an electrochemical reaction field suitable for investigating the nanometer-scale EDL that is expected to form on the liquidelectrolyte side of the electrochemical interface.

Structural Changes in Nb:SrTiO 3 /Electrolyte Model Interface
The etched Nb:SrTiO 3 sample, which was stored in the absence of air prior to in situ observations, was examined at 50-80 Pa in a vacuum-sealed cell (Figure 2a).Neutron beams are incident through the Al window; they are reflected by the sample surface and pass out through the Al window.The electrochemical cell for in situ measurements comprises a Nb:SrTiO 3 cathode, a Li anode, and 1 m LiPF 6 in propylene carbonate (PC) as the electrolyte.Neutrons in the in situ NR cell are incident on the substrate side of the film and reflected at the electrochemical interface (Figure 2b). [16]Figure 2c,d   considering the interfacial layer (see Section S2, Supporting Information).After cell construction, the interfacial layer has a thickness of 3.0 nm and nSLD of 5.30 × 10 −4 nm −2 .The 0.8 nm interfacial roughness between Nb:SrTiO 3 and the interfacial layer is the same as that of the pristine surface.This agreement confirms the absence of compositional and morphological changes on the Nb:SrTiO 3 surface upon contact with the electrolyte.In contrast, the roughness between the interfacial layer and electrolyte is 2.3 nm.The large roughness relative to thickness leads to a gradual nSLD slope on the electrolyte side, as shown in the inset of Figure 2d.This difference is attributed to the change in ionic distribution caused by the formation of an EDL.No solid surface-layer is detected upon increasing the cell voltage from 3.0 to 4.2 V. Minimal oxidative decomposition of electrolyte species occurs on the Nb:SrTiO 3 surface.The interfacial layer expands slightly to 4.2 nm upon applying a voltage of 4.2 V and is accompanied by an increase in the nSLD to 5.45 × 10 −4 nm −2 .These changes indicate that the ionic distribution in the interfacial layer depends on the electrode potential.
According to the Gouy-Chapman model, [9] a distribution of charged ions forms in a 2 nm region from the Nb:SrTiO 3 surface in 1.0 Mmm LiPF 6 PC electrolyte (see Section S3, Supporting Information).The gradient regions observed in the nSLD profile are wider than expected.This may be associated with ion interactions that are not considered in the Gouy-Chapman model. [17]To investigate the chemical composition of the interfacial layer, we evaluated the nSLD values of electrolytes with different proportions of Li + , PF 6 − , and deuterated PC based on their assumed molecular volumes (see Section S4, Supporting Information).Figure 3 contains a color map of the nSLD in the form of a pseudo-ternary diagram as a function of fractional Li + , PF 6 − , and deuterated PC concentrations.The nSLD value decreases with increasing Li + concentration in contrast to its behavior as a function of PF 6 − or PC concentration.This result indicates a positive charge imbalance in the interfacial layer when the Nb:SrTiO 3 electrode is in contact with the LiPF 6 PC electrolyte.In contrast, the increase in nSLD at 4.2 V corresponds to a counteracting Lipoor environment.This change in ionic distribution is reasonable because the negative charge density at the Nb:SrTiO 3 surface should be decreased by applying a positive bias.The foregoing results demonstrate that the formation of and structural changes within the EDL can be detected on a scale of a few nanometers by in situ NR analysis using an atomically flat Nb:SrTiO 3 electrode.

Fabrication and Electrochemical Properties of LiCoO 2 /Electrolyte Model Interfaces
We next investigated the interfacial phenomena between the LiCoO 2 cathode and LiPF 6 PC electrolyte.3a,18] The SrRuO 3 (100) film served as a current collector to support the electronic conduction between the LiCoO 2 film and Nb:SrTiO 3 (100) substrate. [18]igure 4 presents the X-ray diffraction (XRD) patterns, results of XRR analyses, and charge-discharge properties of the LiCoO 2 film synthesized on SrRuO 3 (100)/Nb:SrTiO 3 (100).The 104, 02-8, and 10-8 diffraction peaks of layered rock-salt-type LiCoO 2 are observed along the SrTiO 3 <100>, <010>, and <011> directions, respectively (Figure 4a-c).These findings confirm the epitaxial growth of LiCoO 2 (104).The two-layer model provides adequate fitting to the XRR spectrum (Figure 4d).The refined thickness, SLD, and roughness values are summarized in Table S3 (Supporting Information).The thickness and SLD of the LiCoO 2 layer are 25.6 and 38.1 × 10 −4 nm −2 , respectively.The LiCoO 2 film exhibits a very flat surface with 0.5 nm roughness.A 2032-type coin cell was constructed using a LiCoO 2 (104) film as the cathode, Li metal as the anode, 1 Mm LiPF 6 in PC as the electrolyte, and a glass fiber separator.Plateau regions are evident at ca. 3.9 V in the charge-discharge curves (Figure 4f), which correspond to the two-phase reaction of rhomboidal LiCoO 2 and Li 1-x CoO 2 (0 ≤ x ≤ 0.3). [19]The initial charge capacity of 216 mA h g −1 is considerably greater than the theoretical capacity at an upper cut-off voltage of 4.2 V (137 mAh g −1 ). [20]3a] The initial discharge capacity is 134 mA h g −1 , which is close to the theoretical value.The discharge capacity decreases to 127 mA h g −1 over the 30th cycle.The capacity decrease at the (104) surface is consistent with that previously reported for a (104)-oriented polycrystalline LiCoO 2 film. [21]Thus, a LiCoO 2 (104) film with a very flat surface is deemed to be suitable for investigating structural changes at the cathode-electrolyte interface.

Structural Changes of LiCoO 2 /Electrolyte Model Interfaces
Figure 5 presents the NR fitting results and nSLD profiles for the LiCoO 2 (104) film and LiCoO 2 (104)/1 Mm LiPF 6 PC interface.NR spectra were collected under vacuum, after cell assembly, and after charging at 4.2 V and discharging at 3.2 V during the first cycle.Similarly to Nb:SrTiO 3 (100), no surface-layer model (LiCoO 2 /SrRuO 3 /Nb:SrTiO 3 ) corresponds to the reflectivity spectrum collected under vacuum.The thick- ness, nSLD, and surface roughness of the pristine LiCoO 2 film are 23.8 nm, 3.87 × 10 −4 nm −2 , and 0.5 nm, respectively (see Section S6, Supporting Information).In contrast to the Nb:SrTiO 3 /electrolyte interface, the double interfacial-layer model (SrRuO 3 /LiCoO 2 /inner layer/outer layer) corresponds to the in situ NR spectra of the LiCoO 2 (104)/electrolyte interface.The details of the analytical procedure are provided in Section S6 (Supporting Information).The refined thickness, roughness, and nSLD are listed in Table 2.The nSLD of LiCoO 2 increases to 4.15 × 10 −4 nm −2 after charging at 4.2 V and reversibly decreases to 3.80 × 10 −4 nm −2 after discharging at 3.0 V.This behavior suggests that reversible lithium (de)intercalation occurs, which is consistent with the initial discharge capacity observed in the charge-discharge curves (Figure 4f) and reinforces the validity of the double interfacial-layer model.
Hereafter, we focus on the interfacial layers.Following cell construction, a 4.2 nm-thick inner layer with an nSLD of 2.53 × 10 −4 nm −2 forms on the LiCoO 2 surface.The interfacial roughness between LiCoO 2 and the inner layer is 1.6 nm, which is larger than the surface roughness in a vacuum.The thickness of the LiCoO 2 layer decreases to 21.1 nm compared with that in vacuum (23.8 nm).These results indicate that the LiCoO 2 surface decomposes by reacting with the electrolyte species to form an interphase.Previous research has shown that intercalation electrode surfaces chemically interact with impurities such as HF and H 3 PO 4 in the electrolyte generating LiF (2.30 × 10 −4 nm −2 ) and Li 3 PO x F y (≈2.8 × 10 −4 nm −2 ), [22] whose nSLDs are close to the observed value (2.53 × 10 −4 nm −2 ) for the inner layer formed after cell construction.Thus, the inner layer can be identified as  [23] owing to oxidative decomposition and/or nucleophilic reactions of the electrolyte at extreme potentials. [24]These species have nSLD values exceeding 4 × 10 −4 nm −2 (4.67 × 10 −4 , 5.69 × 10 −4 , and 4.87 × 10 −4 nm −2 for C 2 D 3 O 3 Li, C 4 D 4 O 6 Li 2 , and C 5 D 6 O 6 Li 2 , respectively).Thus, organic components form in the CEI layer during charging.However, the nSLD decreases to 2.54 × 10 −4 nm −2 at 3.2 V, indicating that the organic CEI species are unstable during the initial discharge process.The decomposition of CEI species during discharging has been detected using ex situ NMR. [25] discernible, 4.2 nm outer layer is observed between the inner layer and the electrolyte after cell construction.The interfacial roughness between the outer layer and the electrolyte is 3.4 nm, which approximately equals the outer layer thickness.This property produces an nSLD slope within the outer layer resembling that observed at the Nb:SrTiO 3 /electrolyte interface.This outer layer corresponds to the EDL.Its 5.68 × 10 −4 nm −2 nSLD value is greater than that of the EDL at the Nb:SrTiO 3 /electrolyte interface (5.30 × 10 −4 nm −2 ).This condition creates a Li + -poor imbalance on the electrolyte-side interface to mitigate the potential difference.The nSLD value scarcely changes with the cell voltage during charging and discharging, indicating that the potential distribution at the interface between the CEI and electrolyte remains unchanged.The electrode is apparently shielded from voltage change by the electronically insulating CEI, which suppresses direct contact between LiCoO 2 and the liquid electrolyte.This result indicates that the electrolyte-side structure is determined by the CEI composition.
To confirm the previous conclusions, we investigated the effects of LiPF 6 concentration on the interfacial structures in the inner and outer layers.Figure 6 presents the NR fitting results and nSLD profiles for the LiCoO 2 (104)/0.1 m LiPF 6 PC interface.As with the LiCoO 2 (104)/1 m LiPF 6 PC interface, the double interfacial-layer model (SrRuO 3 /LiCoO 2 /inner layer/outer layer) results correspond closely to those of the calculated in situ NR spectra.All refined parameters are provided in Section S7 (Supporting Information).The nSLD of LiCoO 2 is 3.87 × 10 −4 nm −2 at cell construction, increases to 4.15 × 10 −4 nm −2 upon charging at 4.2 V, and reversibly decreases to 3.85 × 10 −4 nm −2 upon discharging at 3.0 V.This sequence of events confirms reversible lithium (de)intercalation at the LiCoO 2 (104)/0.1&nbsp;mm LiPF 6 PC interface.The inner layer exhibits thicknesses of 2.0 to 2.9 nm during the first-cycle charging and discharging, respectively.The nSLD value of the inner layer increases from 3.68 × 10 −4 to 4.36 × 10 −4 nm −2 during the first charge and decreases to 3.07 × 10 −4 nm −2 during the first discharge.The nSLD values under all conditions are greater than those observed at the LiCoO 2 (104)/1 mm LiPF 6 PC interface.In contrast, the inner-layer thicknesses of 2.0 to 2.9 nm are smaller than those observed at the LiCoO 2 (104)/1 m LiPF 6 PC interface.These results indicate that small quantities of inorganic CEI species with low nSLD values form at the LiCoO 2 (104)/0.1 mm LiPF 6 PC interface.This behavior may be due to the relatively small amount of HF and H 3 PO 4 impurities generated in 0.1 mm LiPF 6 PC.The CEI structure depends on the concentrations of LiPF 6 and residual water in the electrolyte.The 6.5 nm-thick outer layer has an nSLD of 5.98 × 10 −4 nm −2 during cell construction, which is consistent with the EDL formed between the inner CEI layer and bulk electrolyte.The nSLD value is greater than that observed at the CEI layer/1 m LiPF 6 PC interface (5.68 × 10 −4 nm −2 ).These differences are associated with changes in the chemical compositions and structures of the CEI layer (3.68 × 10 −4 nm −2 ) and bulk electrolyte (5.81 × 10 −4 nm −2 ) owing to the low LiPF 6 concentration in the electrolyte.The outer layer shows no significant nSLD changes during charging and discharging.This behavior may be due to the shielding of LiCoO 2 from potential change by the CEI layer, as observed at the LiCoO 2 (104)/1 m m LiPF 6 PC interface.The 6.9 and 7.1 nm outerlayer thicknesses upon charging and discharging, respectively, are slightly smaller than those observed at the LiCoO 2 (104)/1 m m LiPF 6 PC interface (8.3 and 8.8 nm, respectively).The observed EDL thicknesses do not agree with the Gouy-Chapman model, wherein the EDL region expands with decreasing salt concentration (see Section S3, Supporting Information).

Proposed Interfacial Models
Figure 7 shows a schematic of the structural changes at the intercalation-inactive Nb:SrTiO 3 /LiPF 6 PC and intercalationactive LiCoO 2 /LiPF 6 PC model interfaces based on our in situ NR analyses.The interfacial structure changes are summarized as follows: (a) Nb:SrTiO 3 /1 m m LiPF 6 PC interface: Nb:SrTiO 3 scarcely reacts with electrolyte species upon cell construction, and no electrolyte decomposition occurs at 3.0 to 4.2 V during the initial cycle, which results in no CEI formation at the interface.An ionic distribution corresponding to the EDL forms at the electrolyte-side interface.As the negative charge on the electrode surface decreases at high potentials, the Li + concentration in the EDL decreases.
(b) LiCoO 2 /1 m LiPF 6 PC interface: LiCoO 2 reacts with HF and/or H 3 PO 4 impurities in the electrolyte to form a CEI layer comprising mainly inorganic components after cell construction.At high potentials, organic components in the CEI layer increase owing to the decomposition of the solvent.In this case, the ionic distribution in the EDL remains unchanged because the CEI layer screens it from electrode potential change.
(c) LiCoO 2 /0.1 m m LiPF 6 PC interface: Although a CEI layer also is formed by chemical reactions between LiCoO 2 and impurities in the electrolyte, the low LiPF 6 concentration creates a thinner CEI, and the ratio of inorganic to organic species is smaller in the CEI.At high potentials, the organic components in the CEI increase, and the ionic distribution in the EDL remains unchanged.
The intercalation electrode surface is highly reactive toward acidic impurities and facilitates the decomposition of the electrolyte leading to the formation of a CEI layer.An EDL forms at the electrolyte-side interface to cancel the potential difference between the CEI layer and the liquid electrolyte but does not form between the intercalation electrode and the liquid electrolyte.1c] The ionic conductivity of the CEI layer is also reported to affect the intercalation reaction rate. [26]In contrast, little information is available regarding Li transfer at the CEI/organic electrolyte interface, where adsorption/desorption and de-solvation/solvation reactions of Li + occur.Our NR analyses demonstrate that a potential change in the intercalation electrode has little effect on the structure of the EDL during charging and discharging because the CEI layer is electronically insulating.Experimental and molecular dynamics studies on metal/organic electrolyte interfaces showed that the EDL regions exhibit structures different from those in the bulk region, such as different lithium concentrations, cation/anion ratios, and coordination numbers of solvents, which depend on the electrode potentials. [10,27]More anion species are located at the electrode surface at higher potentials, resulting in low concentrations of lithium ions, which limits the lithium (de)intercalation at the cathode surfaces.Thus, the restriction of the EDL structure change is another role of the CEI when designing a cathodeelectrolyte interface with a high rate capability.Direct observation of interfacial structures by in situ NR during cycling, which is capable of simultaneously probing the chemical composition and providing morphological information, is a powerful method for elucidating the detailed mechanism of battery reactions.

Conclusion
Nb:SrTiO 3 (100) and LiCoO 2 (104) surfaces with less than 1 nm roughness were fabricated to provide model electrochemical interfaces in LiPF 6 -containing PC electrolytes.In situ NR analyses of these interfaces revealed concurrent interfacial structural changes within the cathode, through the interface, and into the electrolyte.Both Nb:SrTiO 3 /LiPF 6 PC and LiCoO 2 /LiPF 6 PC interfaces show an interfacial layer with a sloped ionic distribution on the electrolyte side comprising an EDL.The interfacial layer formed at the Nb:SrTiO 3 /LiPF 6 PC interface exhibits a change in ionic distribution with an applied voltage owing to the potential change at the Nb:SrTiO 3 surface.In contrast, the LiCoO 2 /LiPF 6 PC interface shows no significant changes in ionic distribution during charging and discharging.A solid CEI layer forms inside the EDL at the LiCoO 2 /LiPF 6 PC interface via chemical and electrochemical reactions of electrolyte species with the LiCoO 2 surface.No CEI layer is formed at the Nb:SrTiO 3 /LiPF 6 PC interface.These results indicate that the electronically insulating, ionically conducting CEI shields the electrolyte-side interface from the potential change in LiCoO 2 , which restricts changes in the ionic distribution.Because an ionic distribution forms on the electrolyte side to mitigate the potential difference, the properties of the CEI layer have a crucial impact on the ionic distribution, which directly influences the adsorption/desorption and desolvation/solvation rates of Li + during lithium intercalation.Our results here reveal this further role of the CEI, in addition to its established abilities to suppress side reactions and provide ionic conduction pathways.

Experimental Section
Sample Preparation: To obtain an atomically flat surface, the Nb:SrTiO 3 (100) substrates (size: 10 × 10 × 0.5 mm [Nb-doped 0.5 wt.%] and 20 × 20 × 5 mm [Nb-doped 0.2 wt.%], Crystal Base Co., Ltd.) were etched in a hydrofluoric acid buffer solution.A buffered HF solution was prepared by mixing a 50 wt.% hydrofluoric acid (Kanto Chemical Co., Ltd.) and 40 wt.% ammonium fluoride solution prepared by dissolving ammonium fluoride (Kanto Chemical Co., Ltd.) in distilled water at a weight ratio of 1:7.Nb:SrTiO 3 (100) substrates were immersed in BHF for 120 s, rinsed with distilled water, and annealed at 1000 °C in vacuum. [15]SrRuO 3 and LiCoO 2 were deposited on Nb:SrTiO 3 (100) using a PLD system consisting of a 248 nm KrF excimer laser (Lambda Physik, COMPex201) and vacuum chamber (PLAD-312, AOV Inc.).The SrRuO 3 (100) film was deposited as a buffer layer on Nb:SrTiO 3 (100) substrate to improve the electronic conduction between the LiCoO 2 and the Nb:SrTiO 3 (100) substrate. [18]he synthetic conditions for each layer were the same as those used previously. [28]tructural Characterization: The substrate and deposited films were investigated by XRD (ATX-G, Rigaku Inc.) with Cu K 1 ( = 1.541Å) radiation.The lattice orientation was determined using out-of-plane and inplane 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. [29]The surface morphology of the etched substrate was investigated using AFM (JSPM-5200, JEOL) in the tapping mode.
Charge-Discharge Measurements: Coin cells of 2032-type were constructed with the LiCoO 2 (104) film as the cathode, lithium metal as the anode, 1 and 0.1 Mm of LiPF 6 in PC as the electrolyte, and a glass fiber separator.The electrolyte was prepared by dissolving LiPF 6 (Kishida Chemical Co., Ltd.) in PC (Kishida Chemical Co., Ltd.) using a cool stirrer (Scinics Corp.).The dimensions of the LiCoO 2 (104) film were 8 × 8 mm.Constant current charge-discharge tests were conducted in the range of 3.0-4.2V using a potentio-galvanostat (TOSCAT, Toyo system).The current density was set to 1C.
In situ NR: NR measurements were performed using SOFIA (BL-16, Materials and Life Science Facility, Japan Proton Accelerator Research Complex, Tokai, Japan), which is a time-of-flight reflectometer. [16,30]R measurements using pristine samples were performed in a vacuum using a custom-made chamber to prevent chemical reactions with air species.14b,16] The PC solvent (Wako) was deuterated (98%) to prevent incoherent scattering (80.27 and 2.05 for 1 H and 2 H, respectively).Before the LiPF 6 salt was dissolved, the residual water in the solvent was removed using molecular sieves (MS, 4A 1/16, Kanto Chemical Co., Ltd.) that were dried at 400 °C for 4 h.After natural cooling, the material was placed in a side box of a glove box (GB).The NR spectra were collected using the single-frame mode at 0.3°, 0.6°, and 1.3°w ith a footprint of 15 × 15 mm.NR spectra were collected at the opencircuit voltage (OCV) after assembly, charged at 4.2 V, and discharged at 3.2 V in the first 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 Motofit software, [29] different numbers of layers were investigated to determine the interfacial structures (see Section S5, Supporting Information).The nSLDs were estimated from the composition and theoretical density of the solid materials and the composition and molar volumes of each species for the liquid electrolytes to consider the components in the layers (see details in Section S4, Supporting Information).
Figure1ashows X-ray reflectometry (XRR) spectra of the unetched and etched Nb:SrTiO 3 (100) substrates collected in the air.The structural model, including the surface layer, provided adequate fitting results for both spectra.The refined thickness, X-ray scattering length density (SLD), and roughness values are summarized in TableS1(Supporting Information) in the Supporting Information, and the calculated SLD profiles are shown in Figure1b.The unetched surface is covered by a 2 nm-thick surface layer with an SLD of 31.4 × 10 −4 nm −2 .This depth corresponds to the Kiessig fringes observed in the XRR spectrum.The SLD is considerably less than that of Nb:SrTiO 3 (39.4× 10 −4 nm −2 ) and indicates the presence of surface impurities of TiO x (31.6-34.4× 10 −4 nm −2 ), SrCO 3 (27.3× 10 −4 nm −2 ), and Sr(OH) 2 (28.1 × 10 −4 nm −2 ) formed by reaction with air.In contrast, the surface layer of the etched surface has a greater SLD of 38.6 × 10 −4 nm −2 , which is close to that of the substrate.The 0.3 nm surface roughness is comparable to the unit cell size of Nb:SrTiO 3 (a = 0.3905 nm).

Figure 2 .
Figure 2. Photographs of cells used for NR measurements under a) vacuum and b) electrochemical conditions.c) Observed and calculated NR spectra,and d) refined nSLD profiles of etched Nb:SrTiO 3 (100) in vacuum, after cell construction (3.1 V), and at 4.2 V.The cell was fabricated using a Li counter electrode and 1 m LiPF 6 PC electrolyte.The in situ NR spectra were calculated using a model consisting of a SrTiO 3 /interfacial layer/ LiPF 6 PC electrolyte.

Figure 3 .
Figure 3. Color map of the nSLD in the form of a pseudo ternary diagram of Li + cation, PF 6 − anion, and solvent deuterated PC.Calculation details are presented in Section S4(Supporting Information).

Figure 5 .
Figure 5.In situ NR analyses at LiCoO 2 /1 m LiPF 6 PC interfaces.a) Fitting models, b) observed and calculated NR spectra, and c) refined nSLD profiles in vacuum, after cell construction (3.0 V), and at 4.2 and 3.0 V during the first cycle.

Figure 6 .
Figure 6.In situ NR analyses at LiCoO 2 /0.1 m LiPF 6 PC interfaces.a) Fitting model, b) observed and calculated NR spectra, and c) refined nSLD profiles after cell construction (3.0 V), at 4.2 and 3.0 V during the first cycle.

Table 1 .
Refined thickness, nSLD, and roughness of etched Nb:SrTiO 3 (100) in vacuum, after cell construction (3.1 V), and at 4.2 V.The cell was fabricated using a Li counter electrode and 1 m LiPF 6 PC electrolyte.

Table 2 .
Refined thickness, nSLDs, and roughness of a LiCoO 2 (104) film in contact with 1 m LiPF 6 PC determined by NR analysis using a doubleinterfacial layer model.