Thermal‐Induced Structure Evolution at the Interface between Cathode and Solid‐State Electrolyte

The interfaces between the electrode and solid‐state electrolyte play a decisive role in the performance of all‐solid‐state batteries. For example, the formation of the interphase between cathode and solid‐state electrolyte can affect interfacial impedance and thus the rate capability. Herein, the thermal stability at the solid–solid interface between LiMn2O4 cathode and garnet electrolyte LLZTO via combined in situ techniques is studied, including in situ X‐Ray diffraction, in situ transmission electron microscopy, and in situ Raman. The dynamic process of interfacial reaction at different scales is elucidated. Starting from 300 °C, Mn ions from LiMn2O4 would migrate into the solid‐state electrolyte, accompanied with the formation of LiMn3O4 interphase. As the temperature increases to 500 °C, the LiMn3O4 interphase transforms to MnO structure which hinders Li‐ion transportation and therefore increases the interfacial impedance. Although both LiMn2O4 and LLZTO could withstand continuous heating, their interface is inherently thermally unstable at relatively low temperature, which requires special attention during thermal treatment for practical fabrication. Findings provide mechanistic insights into the interfacial reaction which serves as a guidance for the design and manufacturing of all‐solid‐state batteries.


Introduction
[12][13][14][15] The nature of the solid-state electrolyte (SSE) determines the kinetic properties of batteries.Among the three main types of SSEs, polymer, inorganic oxide and sulfide electrolytes, oxide SSEs exhibit excellent compatibility with commercial oxide cathode materials, but their large interfacial resistances remain the major challenge. [16][19] During the fabrication of ASSBs, either oxide SSEs or cathodes require to be synthesized at high temperature.For example, it requires 1200 °C for Li 7 La 3 Zr 2 O 12 (LLZO) [20] and 1300 °C for Li 3x La 2/3Àx TiO 3 (LLTO) [21] synthesis.24] The fabrication of the intimate interface between SSE and cathode usually requires a solid-state sintering process, with indispensable ionic diffusion to form the compact interface and increase Li-ion conductivity. [25,26]Unfortunately, the interphases with low ion conductivity would also form under high temperature, leading to restricted Li ion transportation. [27,28] similar problem also happens when ASSBs are heated.[29] Considering the above issues, it is important to understand how thermal treatment affects the microstructure of the interface between cathode and SSE.
It is well known that the structure and chemistry at the interface are different from those in the bulk.As lattice symmetry and periodicity break at the interface between cathode and SSE, reactions under heating are prone to occur due to defects formation and consequent charge carrier redistribution. [30][33] For example, Kim et al. investigated the interfacial reaction between LiCoO 2 and LLZO through electron diffraction method and found the formation of La 2 CoO 4 structure when heating at 700 °C. [34]Zhang et al. studied the reaction between LiNi 1/3 Co 1/3 Mn 1/3 O 2 and LLZO under high-temperature using X-Ray diffraction (XRD) method and found the formation of LaNiO 3 phase under 600 °C. [35]However, the correlation between the chemical dynamics and phase evolution is still not well understood. [36]39][40][41][42] In this work, we probe the structural variation at the interface between cathode LiMn 2 O 4 (LMO) and solid-electrolyte Li 6.5 La 3 Zr 1.5 Ta 0.5 O 12 (LLZTO) during the heating process, utilizing in situ transmission electron microscopy (TEM) combined with other in situ analytical techniques.LiMn 2 O 4 is a spinel cathode with a practical voltage from 3.0 to 4.4 V, which is cobalt free and less expensive compared with cobalt-contained cathode materials. [43]LLZTO is a garnet-type solid-state electrolyte which possesses a high ionic conductivity of 10 À4 Scm À1 and excellent thermal stability. [20]We have applied in situ XRD, in situ TEM, and in situ Raman to elucidate the dynamic process of interfacial reaction at different scales, as illustrated in Figure 1a.In situ TEM experiment reveals ion migration and phase transition with increased temperature.Electron energy loss spectroscopy (EELS) results clarify the ion diffusion behavior, and the consequent phase variation at interface is confirmed by high-resolution TEM (HRTEM) analysis.The formation of new phases at relatively low temperature is also verified by in situ XRD and Raman.Electrochemical impedance spectroscopy (EIS) is performed to understand how thermal treatment affects the interfacial resistance.We emphasize that the combination of multiple in situ techniques is essential to study the interfacial structural transitions, which further provides mechanistic understanding of the structure-property correlation for ASSBs.

Results and Discussions
LLZTO was synthesized with a solid-state reaction process and hot pressed into a pellet.LMO thin films were deposited on LLZTO pellet at 650 °C by pulsed laser deposition (PLD) method.XRD pattern shows a mixture of garnet phase LLZTO and LiMn 2 O 4 structure in Figure 1b.Elemental mappings with scanning electron microscopy (SEM)-energy-dispersive spectrometry (EDS) verify the uniform distribution of Mn, as shown in Figure S1 (Supporting Information).To study the interface structure at atomic resolution, cross-sectional TEM samples were prepared with a focused ion beam (FIB) lift-out method.As shown in the HRTEM image of Figure S2 (Supporting Information), a clear boundary between LMO and LLZTO can be directly identified by their difference in contrast.The LMO thin film is polycrystalline and has a bright contrast.From a selected-area electron diffraction (SAED) pattern in Figure S3 (Supporting Information), both LMO and LLZTO phases are identified.The scanning transmission electron microscopy (STEM)-EDS mappings show a homogeneous elemental distribution in both LMO film and LLZTO substrate in Figure 1c.The thickness of LMO thin film is measured to be around 250 nm.In addition, the depth-dependent X-ray photoemission spectroscopy (XPS) results confirm the thickness of LMO film as well (Figure S4, Supporting Information), with the signal of Mn 3d vanishing at around 250 nm depth.The above results suggest that a pure, polycrystalline LMO has been successfully deposited on LLZTO pellet.
The in situ XRD heating experiment from room temperature (RT) to 550 °C is first carried out to study thermal-induced phase evolution of LMO/LLZTO interface.The deposited pellet is placed in a vacuum heating chamber where X-Ray could detect the exposed LiMn 2 O 4 thin film.Figure 2a shows contour plot of the diffraction patterns in function of the heating temperature.It is observed that the peaks retain their existence with elevated temperature for both LMO and LLZTO, indicating the intrinsic thermal stability of the two structures.However, (113) and (004) peaks of LiMn 2 O 4 slightly varied with temperature increase, with the details plotted and further analyzed in Figure 2b,c.As shown in Figure 2b, the (113) peak of LiMn 2 O 4 at RT is identified at 2θ = 37°.The (113) peak stays unchanged until a shoulder peak occurs from 400 °C.The signal around 44°consists of three adjacent peaks: (004) for LiMn 2 O 4 , (220) for Mn 3 O 4 , and (313) for MnO, respectively.We attribute the existence of Mn 3 O 4 and MnO to the surface of LiMn 2 O 4 that reacts with air; hence, they will not be discussed in the following analysis.No obvious change could be detected from RT to 200 °C, whereas an abrupt (004) plane expansion occurs at 300 °C (Figure 2c).This lattice expansion implies that interfacial reaction occurs at the temperature as low as 300 °C and the reaction is also observed in TEM experiment which will be discussed later.Above 400 °C, a new phase emerges with the occurrence of peaks at 36.8°and 43.9°, which is identified as LiMn 3 O 4 by its (002) and ( 220) lattice planes.The formation of LiMn 3 O 4 phase suggests Mn ion migration during continuous heating, which leads to structure reconstruction and new phase generation.
In situ TEM heating experiment is performed to investigate the structure transition at microscale.The FIB-processed specimen is loaded on a DENS heating chip, as shown in the SEM image of Figure S5 (Supporting Information).To avoid the beam-induced damage, we darken the electron beam during heating and only collect the data at RT, 300, 400, and 500 °C, respectively.As shown in the STEM-high-angle-annular darkfield (HAADF) images of Figure 3a-d, we observed a clear boundary between LMO and LLZTO, which becomes blurred with temperature increase.HRTEM images at the side of LMO are also acquired to probe the variation of atomic structures (Figure 3e-h and S7).Below 300 °C, we could only see lattice fringes of LiMn 2 O 4 .As the sample is heated to 400 °C, we started to observe the formation of LiMn 3 O 4 .At 500 °C, MnO phase is identified as well as LiMn 3 O 4 .The observation of LiMn 3 O 4 phase is consistent with the XRD result in Figure 2c.However, the XRD spectrum picked up the information from the whole thin film while the HRTEM can detect the local structural changes.We identified the MnO phase which was not retrieved from XRD results.Therefore, the interphase transition route from LiMn 2 O 4 to LiMn 3 O 4 and MnO is confirmed, with the accumulation of Mn and loss of Li elements.Given that the phase evolution could be induced by ion diffusion, we then performed STEM-EELS mappings at the interface to track the elemental migrations.The distributions of Li, Mn, Ta, and La at different temperatures are plotted in Figure 3i-l, respectively.The EELS mappings at RT confirm a homogeneous distribution of Li, Mn, Ta, and La, as revealed by STEM-EDS in Figure 1c but in higher spatial resolution.In addition, we also distinguished the signal of lithium from its K edge at 58 eV, which is unavailable in EDS mappings.It is noted that we observe a dark contrast of Li distribution at the LLZTO side, which could be caused by the diffraction effect of STEM-EELS: the higher Z (LLZTO) region can elastically scatter more electrons to a higher collection angle which causes a lower signal from LLZTO (Figure S6, Supporting Information). [44,45]At 300 °C, a Mn-rich layer can be found at the surface of LiMn 2 O 4 with a thickness of around 5 nm.Correspondingly, we observed that the (003) lattice of LiMn 2 O 4 slightly changed at the same temperature (Figure 2c).These results suggest a LiMn 2þx O 4 phase formed at 300 °C due to the blocked diffusion of Mn ions at the interface.When the temperature is raised to 400 °C, further accumulation of Mn prompts the formation of LiMn 3 O 4 phase with a thickness of 10 nm, as identified in Figure 2c and 3g.Mn ion easily diffuses at the cathode side but is hard to cross the interface, which eventually leads to the Mn aggregation at the interface.This phenomenon should be attributed to the space-charge layer (SCL) between LMO and LLZTO, where the electric field could mitigate Figure 3.In situ TEM investigation of LMO/LLZTO interface during heating.a-d) HAADF images of LMO/LLZTO interface at RT, 300, 400, and 500 °C, respectively.e-h) HRTEM images for LMO/LLZTO interface at marked regions in (a-d), respectively.i-l) STEM-EELS mapping for Li, Mn, Ta, and La at marked regions in (a-d), respectively.
Mn ion motion until higher temperature endows it with larger kinetic energy to overcome this energy barrier. [46]At 500 °C, the layer of Mn aggregation raises to a thickness of 15 nm and the dissolution of Mn ion into LLZTO side, up to 20 nm, is also observed.At the same time, this Mn-rich layer also becomes Li deficient, because the most lithium sites were occupied by Mn.As a result, a MnO phase is finally formed at the interface, which results from the different kinetic behaviors of Li and Mn ions during thermal treatment.
The migration tendency of Li and Mn ions can be quantitatively analyzed from the results of STEM-EELS mappings.As shown in Figure 4a, we integrate the signals of Li and Mn crossing the interface.It is observed that with the increase of temperature, the concentration profile of lithium becomes smoother across the interface while the concentration of Mn accumulates.The EELS spectra of Mn-L 2,3 edges and the O-K edge crossing the interface are shown in Figure 4b,c, respectively, marked with the temperature and positions of "A", "B", "C," and "D".The Mn-L 2,3 edges are sensitive to the electronic structure variations and therefore could be utilized to analyze valence information at different temperatures and regions. [47]It is noticed that the prepeak of O at around 530 eV undergoes significant change.Specifically, the prepeak intensity at interface gradually decreases with elevated temperature, from RT-A, 300-B, 400-B to 500-C.Also, with continuous heating, the reduced O prepeak character progressively spreads into the bulk of LMO, implying a thicker reaction layer.The decreased intensity of O pre-edge represents reduced valence of 3d metal ions, which is attributed to weakened hybridization between metal and oxygen ions. [48,49]esides O-K edge, the variation of Mn ion valence could be revealed by the change of Mn-L 2,3 edges as well.It is noticed that the L 3 peak shifts to lower energy either with increased temperature at interface or from interface to LMO bulk.Since lower L 3 peak energy indicates reduced valence for Mn ions, the valence information of Mn revealed by its L 3 peak is consistent with O-K edge.Moreover, EELS spectra could be further analyzed quantitatively by calculating the white-line ratio of Mn-L 2,3 edges and prepeak intensity of O-K edge.The white-line ratio is the area ratio of L 3 and L 2 peaks, which is negatively correlated with Mn valence; however, the prepeak intensity of O-K edge is positively correlated. [50]The calculation details and results are shown in Figure S8 and Table S1 (Supporting Information).The relation between Mn valence and these quantified values are plotted in  phase.MnO structure in position 500-C indicates its þ2 valence.Since positions 400-C, 500-C, and 500-D share similar quantified values, it is suggested that valence of Mn ions inside LLZTO is þ2 as well.Hence, the connection between valence variation and ion migration is clear via quantitative analysis.Position A at each temperature represents bulk LiMn 2 O 4 , the spectrum of which does not change during heating process.As temperature is increased to 300 °C, Mn ions start to accumulate at cathode surface (300-B), leading to the reduction of Mn element.When it is raised to 400 °C, the Mn ion aggregation at the interface reduces Mn to a lower valence (400-B) and this phenomenon is more severe for 500 °C (500-B and C).Also, Mn ions would dissolve into LLZTO from 400 °C.The spectra of Mn in position 400-C and 500-D suggest that these Mn ions possess the lowest valence.
As high temperature and high voltage could both lead to interfacial degradation between cathode and SSE, their difference remains incompletely understood.In order to compare the distinct effects of thermal and electrochemical instability, we then conducted a study to investigate the interfacial variation under the influence of high voltage.After maintaining the cathode at a high voltage of 4.5 V for 60 min, a relaxation process lasting more than 10 h is carried out before TEM sample preparation.This extended relaxation ensures a uniform distribution of Li ions within the cathode, so that the interfacial structural change is exclusively attributed to the electrochemical reaction rather than inhomogeneous Li ion distribution.As shown in Figure S9a (Supporting Information), an interface between cathode and SSE is clearly observed.Furthermore, the HRTEM image in Figure S9b (Supporting Information) demonstrates the presence of an interphase layer of around 6 nm between LiMn 2 O 4 and LLZTO.This interphase is identified as Mn 3 O 4 , a degraded structure commonly associated with cycled LiMn 2 O 4 in previous reports. [51]To uncover the origin of phase transition under highvoltage conditions, we then performed EELS mappings to examine the element redistribution across the interface (Figure S9c, Supporting Information).As a result, a Li-deficient layer measuring 5-6 nm is observed at the interface, which is also corroborated by the integration of Li and Mn signals, as shown in Figure S9d (Supporting Information).Given that high voltage could trigger the decomposition of oxides, we propose that the observed phase transition stems from oxygen loss, leading to structure change from oxygen-rich LiMn 2 O 4 to oxygen-poor Mn 3 O 4 . [52]he oxygen loss is also evident in the reduced valence of Mn. Figure S10a (Supporting Information) illustrates the O-K and Mn-L 2,3 spectra ranging from the interface to cathode bulk.The reduced intensity of O prepeak and the shift of the Mn signal toward lower energy both indicate a reduction of Mn ion at the interface, as described in Figure 4b,c.Furthermore, Figure S10b (Supporting Information) presents the quantification results of O prepeak intensity and Mn white-line ratio, revealing a gradual decrease of Mn valence from the bulk to interface over a distance of around 6 nm.The progressive reduction of Mn at the interface also suggests an incomplete transition from LiMn 2 O 4 to Mn 3 O 4 , which implies that the electrochemical-induced cathode degradation is confined to the interface without massive migration of Mn ions.The distinct influence of thermal and high-voltage instability is thus revealed by comparing their interfacial variations.Higher temperature could activate the motion of Mn ions, leading to severe structure damage extending from the interface to cathode bulk.In contrast, the structure change is confined to the interface under high voltage which does not reach the same magnitude as those observed under high-temperature conditions.
As in situ TEM experiment reveals structure evolution and valence change of LMO/LLZTO interface, and in situ Raman is performed to study heat-induced variation of Mn--O bond.Raman spectra from LMO/LLZTO at varied temperatures and pure LLZTO are plotted in Figure 5a.At RT, the spectrum of LMO/LLZTO shows two LMO peaks at around 500 and 620 cm À1 , which are superimposed on the pure LLZTO signal.The two peaks represent F 2g and A 1g vibration modes of [MnO 6 ] octahedron in LMO structure, respectively. [53]Also, it is noticed that the small peaks exist between 650 and 700 cm À1 on Zr-O peak shoulder could be ascribed to the surface Mn 3 O 4 or MnO that is previously detected in XRD. [54,55]The change of Mn--O bonding state during heating would be reflected on their Raman peak shift.The two peaks of LMO remain unchanged from RT to 300 °C; however, they both shift to lower value from 300 to 550 °C.The vibrational frequency is useful to reveal Mn--O bonding following the equation.
where v, c, f, and u are wave number, velocity of light, average force constant, and effective mass, respectively. [56]The average force constant f is negatively related with Mn--O bond length; thus, the peak shift indicates a weaker bonding of Mn and O ions. [57]Actually, the increased Mn--O bond length is in accordance with the observed phase transition from LiMn 2 O 4 to LiMn 3 O 4 structure where the Mn-O distance increases from 2.001 to 2.075 Å. [58] Therefore, in situ Raman method provides another way to explore the interfacial alteration via lattice vibration detection, in which weakened Mn--O bonding at higher temperature is in agreement with in situ XRD and TEM results.
Li-deficient interfacial phases can form during heating as reported in many previous studies, which hinders the transportation of Li ions, leading to increased interfacial resistance eventually. [59]The electrochemical impedance spectrum (EIS) is carried out for LMO/LLZTO sample after heated at different temperatures.As shown in Figure 5b, Au electrodes are deposited on LMO/LLZTO system for EIS test and the experimental details are described in the Experimental Section.The as-synthesized sample shows mainly two half circles: highfrequency part representing the bulk resistance of LLZTO and low-frequency part representing the interfacial resistance, as reported in literature. [60]After the specimen is heated up to 300-500 °C, the half circle at lower frequency starts to increase and merge with the other half circle.To quantitatively analyze the alteration of interfacial impedance, an equivalent circuit is used to fit the EIS spectra (Figure 5b). [61]The fit values of resistance are plotted in Figure 5c, showing the influence of temperature on interfacial impedance.Resistance of LLZTO bulk is nearly unchanged, whereas the interfacial resistance increases with elevated temperature.Given that the interphase gradually changes from Li-ion conductor to Li-deficient species with lower  conductivity, the increment of interfacial impedance is attributed to the interfacial structure transition. [62,63]echanistic insights into thermal-induced interfacial evolution are proposed in Figure 6, showing the interplay of SCL, chemical potential of Mn (μ Mn ), Mn ion distribution (C Mn ), and interphase transition.At RT, SCL naturally forms at the boundary between cathode and SSE, which is mainly due to the rebalance of interfacial chemical potential and consequent charge carrier redistribution. [64]The SCL raises the chemical potential of Mn than that in the cathode bulk, serving as a barrier to mitigate the migration of Mn from Mn-rich cathode to Mndeficient LLZTO side. [65,66]Although the reaction between LiMn 2 O 4 and LLZTO is thermodynamically favorable, the SCL could prevent the reaction at RT. [67] As a result, a sharp and intact interface generates which is verified by HAADF images and EELS mappings.When the temperature is raised to 300 °C, higher kinetics of Mn ion favors its tendency to dissolve into SSE.However, the SCL could prevent Mn ion motion, leading to its segregation at the interface that reduces its valence as well.Meanwhile, the reduction of Mn ion requires consumption of electrons inside the SCL.Therefore, the chemical potential of Mn gradually decreases at the interface due to the weakened SCL, which lowers the barrier for Mn ion to cross the interface.Consequently, a temperature of 400 °C is high enough to enable Mn ion cross the damaged SCL and dissolve into SSE.Continuous Mn ion migration from cathode bulk to the interface finally leads to a phase transition from LiMn 2 O 4 to LiMn 3 O 4 .Further, a higher temperature of 500 °C exacerbates the increment of Mn at the interface, which eventually causes the formation of MnO phase.Moreover, there is no barrier for Mn ion to cross the interface due to the totally broken SCL at this temperature, and severe Mn dissolution into SSE is found as a result.
Since the reaction prefers to happen at interface, trace amounts of products at relatively low temperature cannot be easily detected by conventional XRD method.Nonetheless, in situ TEM results confirm that the reaction between two phases would occur even at a low temperature of 300 °C, which results in increased interfacial resistance.With continuous heating, the surface structure of LiMn 2 O 4 gradually transforms to LiMn 3 O 4 and eventually to MnO.These structures with the Li sites occupied by Mn ions would hinder Li ion transportation for LiMn 2 O 4 cathode, as illustrated in previous reports. [68]Therefore, Li ion diffusion is blocked due to the formed interphases with lower ion conductivity.These structural variations finally lead to larger interfacial resistance with increased temperature, confirming the instable character of LiMn 2 O 4 and LLZTO interface.Therefore, further studies to improve the interfacial stability, including cathode modification and artificial layer addition, are required to improve the performance of ASSBs. [69]

Conclusion
To summarize, structural evolution under heating for LiMn 2 O 4 and LLZTO interface is investigated via combined in situ methods.Although the formed SCL could inhibit the reaction between LMn 2 O 4 and LLZTO at RT, high temperature would activate the interface and cause irreversible phase variation, which is more severe than high voltage-induced structure degradation.
We found that the interphase transition from LiMn 2 O 4 to LiMn 3 O 4 and then MnO is attributed to two reasons: Mn ion migration due to higher energy and its aggregation at solid-state interface.Consequently, Mn ions would occupy the original Li sites of LiMn 2 O 4 , thus lowering interfacial ionic conductivity.In addition, the larger kinetic energy of Mn ion also facilities its dissolution into LLZTO side, which may change the element composition of LLZTO surface and influence interfacial kinetics.Our results indicate that the interface between cathode and solid electrolyte has lower thermal stability than both cathode and solid electrolyte.Attention needs to be paid during synthesis and thermal treatment.We believe our work uncovers the correlation between structural transition and ionic transportation for cathode and SSE interface, which would inspire further studies for the development of thermostable all-solid-state batteries.

Experimental Section
Thin-Film Preparation: The synthesis process of LLZTO pellet is described elsewhere. [70]LiMn 2 O 4 thin film was deposited on LLZTO pellet by PLD method.The wavelength of XeCl excimer laser was 308 nm for film deposition.O 2 atmosphere of 3.6 kPa was maintained during deposition process at 650 °C for 50 min, after which the specimen temperature was cooled down to RT at a rate of 20 °C min À1 .
In Situ XRD Characterization: The structure evolution was detected by in situ XRD method in an X-Ray diffractometer (Rigaku Smart) at different temperature.The specimen was placed on a vacuum heating holder before XRD test.After the specimen was raised to target temperature (RT, 100, 200, 300, 400, 500, and 550 °C) and maintained for 30 min, scans of 10°-70°(2theta) was applied at a rate of 5°min À1 .To enhance the signal quality of interested peaks and observe peak variation, additional scans with a slower rate of 2°min À1 between 36.0 and 37.8°and 43.4 and 45.2°were applied.
In Situ TEM Experiment: FIB (FEI Helios 600i) was used to fabricate the cross-sectional TEM samples.A lamellar sample of 1 μm thickness was extracted from the specimen by conventional FIB lift-out method.After the lamellar sample was transferred to a TEM heating chip (DENS solutions, wildfire), the thickness of cross section was thinned to around 100 nm using 2-30 kV Ga ion beam.In situ TEM experiment was carried out in a JEOL JEM F200 microscope at 200 kV with a DENS heating holder.HRTEM, HAADF, and STEM-EDS/EELS characterizations were obtained after the specimen was heated to target temperature (RT, 300, 400, and 500 °C) and maintained for 30 min.
XPS and Raman Spectroscopy Characterization: X-Ray photoemission spectroscopy (PHI 5000 VersaProbe III) was used to analyze the depth distribution of Mn and La.Ar ion with 3 kV accelerating voltage was applied to etch specimen at a rate of 1 nm min À1 .Spectra of Mn and La signal were acquired at every 50 nm ranging from 630 to 840 eV.Raman spectra were collected using a laser with 532 nm wavelength (LabRam-HR/VV).After the specimen was heated and preserved at targeted temperature for 30 min in an argon-protected glovebox, it was transferred to Raman system to collect the spectra ranging from 300 to 900 cm À1 .
EIS Measurements and High-Voltage Retention: The as-synthesized LMO/LLZTO specimens were heated to target temperature (300, 400, and 500 °C) and maintained for 30 min in Ar atmosphere.Au films were deposited on both sides of the pellet as blocking electrodes using a sputter coater to form a configuration of Au/LMO/LLZTO/Au, which was ready for EIS test.EIS measurements were performed in a frequency ranging from 100 kHz to 10 mHz using the Bio-Logic SP200 workstation with a current amplitude of 200 nA.To maintain the cathode at a high voltage of 4.5 V, the LLZTO pellet with LMO thin film was placed in a Swagelok-type cell with Li metal as anode for testing.After maintaining the cell at 4.5 V for

Figure 1 .
Figure 1.a) Schematics of the in situ techniques used in this work including XRD, TEM, and Raman.b) XRD pattern of LMO/LLZTO sample.c) STEM-EDS element mapping of Pt, Mn, La, O, Ta, and Zr from a cross-sectional LMO/LLZTO sample.

Figure 2 .
Figure 2. In situ XRD results of LMO/LLZTO obtained during the heating process from RT to 550 °C.a) Contour maps of temperature-dependent in situ XRD patterns at 2θ from 15°to 65°.The arrows inside mark target peaks.b) Fine features of LiMn 2 O 4 (113) and (004) reflections in function of temperature.c) Variation of lattice parameters of different phases during heating.
Figure 4d.Since position A at each temperature represents original LiMn 2 O 4 phase and shares similar white-line ratios and O prepeak intensities, their valence of Mn is marked as þ3.5.Positions 400-B and 500-B are identified as LiMn 3 O 4 structure revealed by HRTEM results and their valence is þ2.33.It is noticed that Mn valence in position 300-B is in the midst of position 300-A and 400-B, which is in accordance with its LiMn 2þx O 4

Figure 4 .
Figure 4. EELS analysis of LMO/LLZTO interface during in situ heating process.a) Profiles of Mn and Li across the LMO/LLZTO interface at varied temperatures.b,c) EELS spectra of O-K edge and Mn-L 2,3 edge, respectively, taken from marked regions in (a).d) Plot of Mn white-line ratio and O prepeak intensity at different states and their relation with valence state of Mn ions.

Figure 5 .
Figure 5. Raman and EIS results.a) Raman spectra of pure LLZTO and LMO/LLZO specimen at varied temperatures.b) EIS test of Au/LMO/LLZTO/Au at varied temperatures.c) Resistance of LLZTO and interface at different temperatures fit by the equivalent circuit model.

Figure 6 .
Figure 6.Schematic diagram of interfacial reaction mechanism.The dashed brown and red lines represent the distributions of the chemical potential and concentration for Mn ion at RT, respectively.