Skip to main content
NCBI home page
ACS AuthorChoice logoLink to ACS AuthorChoice
. 2023 May 22;15(22):26660–26669. doi: 10.1021/acsami.3c02770

Correlating Nanoscale Structures with Electrochemical Properties of Solid Electrolyte Interphases in Solid-State Battery Electrodes

Jimin Oh †,‡, Gun Park †, Hongjun Kim †, Sujung Kim §, Dong Ok Shin ‡, Kwang Man Kim ‡, Hye Ryung Byon §, Young-Gi Lee ‡, Seungbum Hong †,*
PMCID: PMC10252843  PMID: 37212378

Abstract

graphic file with name am3c02770_0007.jpg

Here, we investigate the nonlinear relationship between the content of solid electrolytes in composite electrodes and the irreversible capacity via the degree of nanoscale uniformity of the surface morphology and chemical composition of the solid electrolyte interphase (SEI) layer. Using electrochemical strain microscopy (ESM) and X-ray photoelectron spectroscopy (XPS), changes of the chemical composition and morphology (Li and F distribution) in SEI layers on the electrodes as a function of solid electrolyte contents are analyzed. As a result, we find that the solid electrolyte content affects the variation of the SEI layer thickness and chemical distributions of Li and F ions in the SEI layer, which, in turn, influence the Coulombic efficiency. This correlation determines the composition of the composite electrode surface that can maximize the physical and chemical uniformity of the solid electrolyte on the electrode, which is a key parameter to increase electrochemical performance in solid-state batteries.

Keywords: solid electrolyte interphase (SEI), nanoscale morphology, chemical composition, composite electrode, solid-state battery

1. Introduction

The technology of lithium-ion batteries has evolved to include applications ranging from small home appliances and smartphones with capacities of tens of watt-hours (Wh) to electric vehicles (EVs) and energy storage systems (ESSs) with capacities of tens of kilowatt-hours (kWh) or more.1,2

However, for large-scale applications, superior safety characteristics are required. To address this, solid-state lithium-ion batteries using solid electrolytes have been researched for several decades. Solid electrolytes offer advantages such as nonflammability and a wide electrochemical window of stability by, e.g., suppressing the formation of lithium dendrites, making it possible to use a lithium metal electrode and achieve high energy density cells.3

Various types of solid-state batteries have been reported, such as those that use oxide,4,5 sulfide,6−8 solid polymer,9,10 and solid–liquid composite electrolytes.11−13 Despite their advantages, the low ionic conductivity of solid electrolytes and high interfacial resistance between active materials and solid electrolytes have limited their application.

To overcome the limitations of solid-state batteries, researchers need analysis techniques that can quantify ion mobility and diffusion coefficients and visualize local ion concentrations and distributions in solid electrolytes.14−17 Among the various analysis techniques, atomic force microscopy (AFM) can visualize surface morphologies and mechanical, chemical, and electrical properties of electrodes at the nanoscale, making it a valuable tool for the characterization of solid-state battery components such as solid electrolytes, electrodes, and interfacial layers.17−21

Electrochemical strain microscopy (ESM), one of the functional AFM modes, has been recently developed to visualize the surface lattice variation induced by the electric field between the tip and the sample. ESM, the electrically biased contact AFM mode, and its derivative techniques are utilized for the analysis of local ion concentrations and lattice variations on the unit cell down to the nanometer level.22−28 For example, Alikin et al. probed the local diffusion coefficients and ion concentrations of particles of lithium-ion battery cathode materials, LiMn2O4, with different sizes.29 Their quantitative analysis of the ionic mobility and concentration unveiled the dynamic behavior of ion migration and relaxation and the change in ion concentration profiles.

It should be noted that the analysis of the surface distribution of chemical elements and their concentrations after electrochemical reactions provides powerful insights into the aging/degradation mechanisms and capacity properties of cells.30,31 Using the ESM method, Wang et al. revealed the movement of ions between the grain and grain boundaries at different sites of a sodium (Na) superionic conductor (NASICON)-structured Li1.5Al0.5Ge1.5(PO4)3 (LAGP) solid electrolyte.32 Furthermore, Strelcov et al. have developed functional electrochemical probing methods applicable to liquid–solid interfaces and have suggested optimal scanning probe microscopy (SPM) techniques for a multitude of complex cross-coupled physical and chemical phenomena in liquid-free environments.33 However, there has been no report on elucidating the importance of physical and chemical uniformity of the solid electrolyte interphase in contact with the electrode for stable cycling of the solid-state battery.

Here, we visualize the surface morphology and Li-ion distribution of solid electrolyte interphase (SEI) layers in composite electrodes varying the ratio of natural graphite and solid electrolyte. The mapping of the lithium-ion distributions on the surface of the SEI layers is of specific interest in practical surface analyses of electrode degradation in batteries.16,28,34,35 In addition, information about the chemical concentration and distribution of elements in SEI layers on the electrodes is required to determine the presence and extent of unnecessary and undesirable electrochemical reactions. As such, the chemical distributions and morphology characterizations of the SEI layers provide insights into the mechanisms governing ion mobility in solid-state batteries, of which information can be used to develop new materials and strategies for improving the energy density and safety of lithium-ion batteries. This correlation determines the composition of the composite electrode surface that can maximize the physical and chemical uniformity of the solid electrolyte on the electrode, which is a key parameter to increase electrochemical performance in solid-state batteries.

2. Results and Discussion

2.1. Visualization of Li-Ion Distribution

To visualize Li distribution on a composite electrode, we conducted ESM imaging of a conventional natural graphite (NG) electrode as a reference. Figure 1 shows the electrochemical impedance spectroscopy (EIS) results, topography and ESM images, and the relationship between ESM and charge transfer resistance of conventional NG electrodes.

Figure 1.

Figure 1

Open in a new tab

(a) EIS results of NG electrodes with different DOLs, (b) topography/ESM images, and (c) plot of amplitude and charge transfer resistance at different DOLs using conventional NG electrodes.

The electrodes consist of NG/SBR-CMC/Super-P with five different degrees of lithiation (DOL) states where DOL20, DOL40, DOL60, DOL80, and DOL100 represent 20, 40, 60, 80, and 100% of lithiation, respectively (Figure S1). Five different states of samples were prepared and lithiated at a 0.1C rate (37.2 mAh gNG–1). Each charge transfer resistance was observed by EIS.

The surface morphology shows granular-shaped NG particles with a diameter of 10 ± 1.35 μm. In the height images, most NG surfaces are covered with various sizes of turtle-shell-shaped agglomerates that are derived from electrolyte decompositions under 0.1 V vs Li/Li+, depending on DOL.

ESM images show the spatial distribution of Li-ions on the surface of Li-ion battery materials.18,36,37 Therefore, the average ESM amplitude is expected to linearly scale with DOL.38 As DOL increases, Li distribution (bright region in the amplitude map) is observed from the boundary of NG particles (DOL20) to all areas on the electrode (DOL100). However, the amplitude increased in a nonlinear fashion in Figure 1c. In order to understand this discrepancy, we analyzed the charge transfer resistance as a function of DOL.

From the EIS results, we observed that the critical DOL is 60, below which the charge transfer resistance decreases as the lithiation increases and above which it saturates at a low value. As such, in the low DOL regime, the amount of Li intercalation will be hindered by high charge transfer resistance, whereas in the high DOL regime above 60, the Li intercalation will exponentially increase due to the low charge transfer resistance.

The ESM amplitude image of DOL20 showed a bright region near NG particles due to the high charge transfer resistance, while that of DOL100 showed a uniform bright contrast over the whole region on the electrode due to the low resistance. The above observation explains the drastic change of the slope in ESM amplitude where DOL is 60 and the corresponding variation of EIS results. Therefore, one can correlate the Li-ion distribution on the electrodes at the nanoscale with the electrochemical reactions at a much larger scale.

2.2. Analysis of Solid Electrolyte Effects

Solid electrolytes affect the lithium-ion diffusivity and storage capacity in composite electrodes, which, in turn, determine the entire electrochemical properties of solid-state batteries. Most of the SEI is known to be formed at the first cycle under 1 V through the mixing of the decomposed lithium salt and solvent.39 However, the formation mechanism of SEI in the vicinity of a composite electrode mixed with a solid electrolyte has remained largely unknown.

To investigate the effect of a solid electrolyte on the composite electrode, we fabricated composite electrodes composed of NG/LSTP/SBR-CMC/Super-P with different NG/LSTP ratios. Figure 2 shows lithiation–delithiation specific capacities and height images of pristine and cycled electrodes as a function of LSTP content. Table 1 summarizes the values of lithiation–delithiation specific capacities, irreversible specific capacities, and the Coulombic efficiency for the composite electrode samples.

Figure 2.

Figure 2

Open in a new tab

(a) Plot of the specific capacity of composite electrodes after lithiation and delithiation as a function of LSTP content. (b) AFM topography images of pristine and cycled electrodes as a function of LSTP content. The red dashed lines represent the boundary of regions higher than a certain threshold (2 μm) for LSTP12 and LSTP24.

Table 1. Electrochemical Properties of Composite Electrodes after Cycling at 0.1C in the First Cyclea.

sample average lithiation specific capacity ± deviation [mAh·g–1] average delithiation specific capacity ± deviation [mAh·g–1] average irreversible specific capacity ± deviation [mAh·g–1] Coulombic efficiency ± deviation [%]
ref (LSTP0) 355 ± 36 326 ± 29 29 ± 7 91.7 ± 1.74
LSTP12 406 ± 149 110 ± 53 296 ± 96 26.5 ± 8.64
LSTP24 633 ± 42 184 ± 57 448 ± 33 29.1 ± 9.67
LSTP30 660 ± 58 331 ± 51 330 ± 49 49.9 ± 3.16
LSTP50 952 ± 251 257 ± 131 695 ± 120 26.1 ± 8.94
LSTP70 2155 ± 223 359 ± 35 1796 ± 188 16.7 ±0.10
Open in a new tab
a

5 Different cells were prepared for each sample case.

As the LSTP content increases, the lithiation specific capacity showed an overall increasing trend when lithiated, whereas it fluctuated up and down and saturated at the level of theoretical specific capacity of NG for more than 30 wt % of the LSTP content when delithiated (as shown in Figures 2a and S2).

The reason for the overall increase in the apparent lithiation specific capacity is the significant increase of the specific surface area and the reduction decomposition of LSTP during the first lithiation. The specific surface area of the electrode increases with the addition of LSTP from eq 1

2.2. 1

where Mg and MLSTP denote mass fractions of NG and LSTP, respectively. βg and βLSTP are the specific surface areas of NG (10 m2 g–1) and LSTP (60–70 m2 g–1), respectively. The specific surface area with a higher LSTP amount in the electrode induces more electrochemical reactions, resulting in higher electron consumption. The contribution to this excess capacity from the increase of the specific surface area is because of the improved lithiation/delithiation kinetics. However, the reason why this is not the case for the 2nd and 3rd lithiation/delithiation stages is because the SEI layer is formed after the 1st cycle, where the interface between the SEI layer and electrolyte is more important than the specific surface area of the electrode. In addition, we found that the SEI formation peak in lithiation near 0.13 V in the ref (LSTP0) was broadened in LSTP12 and LSTP30 by the decomposition of LSTP (Figure S3).

Figure 2b shows topographical maps in pristine and cycled electrodes as a function of LSTP contents up to 30 wt %. NG particles are clearly seen in all of the pristine electrodes. After the initial cycle, NG particles remained well distinguished in the LSTP0 and LSTP30 electrodes, while they became obscure in LSTP12 and LSTP24. The lower Coulombic efficiency, the ratio of delithiation and lithiation specific capacity, induces a significant change of surface topography. Also, delithiation capacity characteristics affect the deposition surface pattern near the active material on the electrodes.

In order to understand the change of topography in Figure 2b for cycled composite electrodes of LSTP12 and LSTP24, we analyzed the contour of pixels higher than a certain threshold in the height image and colored it red. Clusters of white regions surrounded by red contours could be identified, which we presume to be nonuniformly deposited lithium compounds. The compounds consume excessive lithium ions and electrons, which leads to an increase of irreversible specific capacity and a decrease of Coulombic efficiency.

The reason we did not analyze the topography images of composite electrodes with LSTP contents of more than 30 wt % is because NG particles were mostly covered with LSTP, making our comparison between pristine and cycled electrodes difficult (Figure S4).

2.3. Surface Morphology and Li-Ion Distribution on SEI Layers

ESM imaging was conducted to understand the microscopic origin of the topographic change induced by a sequence of lithiation and delithiation as a function of LSTP content. When conducting ESM, one needs to decouple the electromechanical response from the electrostatic effects induced between the cantilever and samples by voltage-modulated force spectroscopy.40−43 First, we apply a bias voltage to the tip, which makes the electric field penetrate near the sample surface down to a depth of 50 to 100 nm, where most of the SEI layer is concentrated. Additionally, a perfect SEI is a pure ionic conductor, which leads to electronic charge injection and trapping in the SEI layer. This can induce a large electrostatic effect in the acquired ESM signal.

However, a thick and imperfect SEI layer can be considered a mixed conductor because it displays a low Coulombic efficiency. Therefore, the ESM signal from a mixed conductor can be attributed mostly to the Vegard strain because the injected charges will have a very short trapping time on the surface.44 As such, we expect that our ESM signal is coming from the Vegard strain, which is proportional to the Li-ion concentration. In addition, we are aware of the fact that for a more accurate analysis, we need to measure the effective Vegard coefficient β. However, even without the knowledge of β, our analysis still holds true because we made a relative comparison as a function of the solid electrolyte content.

In order to prove our claim, we conducted Pearson correlation mapping between topography and ESM signals at the same position. In the case of capacitive artifacts in ESM images, there is often found a strong linear relationship between the topography and ESM signal. As such, the Pearson correlation plot is useful in checking the presence of capacitive artifacts.

Another approach is to measure the second or higher contact resonance Eigen mode with the reduction of signal contributions from nonlocal electrostatic interactions between the sample and cantilever.42 An alternative way has been suggested by Han et al.,43 where they removed the electrostatic effects by increasing the contact force that, in turn, increased the contact stiffness with a minimal effect on the dynamic stiffness for the first contact resonance, which was proven by the cantilever dynamics modeling. As such, we choose the high contact force (∼250 nN) to reduce the electrostatic effects between the sample and the cantilever.

In Figure 3, height and ESM amplitude images were compared side by side, and their correlation was analyzed by visualization of a Pearson correlation map. Images were obtained at 10 different positions of each sample, and the correlation map was obtained at 5 different positions from selected height and amplitude images to show the reproducibility of our methods (Figure S5). Pearson correlation plots in Figure 3 did not show a perfect linear correlation but a complicated pattern between the height and ESM signal, indicating the suppression of the capacitive artifact in the ESM images.

Figure 3.

Figure 3

Open in a new tab

Heights, amplitudes, correlation maps, and line profiles of (a) ref. (LSTP0), (b) LSTP12, (c) LSTP24, and (d) LSTP30 electrodes.

Line profiles of both height and ESM amplitude are shown to intuitively understand the correlation captured in the correlation map analysis. The x and y axes represent the height and ESM amplitude where the values were varied between 0 and 255 because we used 8-bit images to represent them in separate topography and ESM images. As such, 0 in height is −2.5 μm and 255 means 2.5 μm on the x-axis, and 0 in ESM amplitude means 0 pm and 255 means 400 pm.

In the case of the reference NG electrode, NG particles were clearly revealed in the topography image. In the meantime, the ESM amplitude image showed an overall dark contrast with a faint resemblance of topographic features. Furthermore, bright regions were observed near some of the NG particle interfaces.

Based on these two images, a Pearson correlation map is obtained to elucidate the relationship between height and ESM amplitude. The result shows, statistically, a narrow deviation of the amplitude with a low value in a wide range of heights. In the meantime, the height distribution was concentrated at a height slightly larger than the middle point.

The narrowly distributed and low ESM amplitude, regardless of the widely spread-out height, implies that a thin active SEI layer uniformly covers the NG particles over the region of interest. Moreover, the concentrated distribution of height indicates that the surface of NG particles is relatively flat, as they occupy the majority of the surface where the flat NG particles formed by the cold press during the fabrication are presented.

For further understanding of the correlation, we extracted a profile along an arbitrary line over 20 μm distance in both height and ESM amplitude images (see Figure 3a). As a result, we could see a big valley in the height profile with an overall V-shape, representing the junction between two NG particles, which means the presence of a gap between NG particles. At the position near 10 μm, a boundary between the two NG particles was clearly observed.

However, the ESM amplitude fluctuated between 20 and 50 pm regardless of the position along the line profile. The line profile analysis supports our finding via Pearson correlation map analysis, where the ESM amplitude shows a statistically narrow deviation over a wide range of heights.

In the case of LSTP12, NG particles were not clearly identified in the topography images. The adsorbates appeared to be unevenly covering the surface area of NG particles. In the meantime, the ESM amplitude showed an overall gray contrast with some noise and dark contrast along some valleys coinciding with the part of the boundaries in the topography. However, there was no strong correlation between the ESM amplitude and topography images.

Our finding suggests that with the addition of LSTP, lithiation occurs more uniformly on the surface regardless of the local topographic variation. The correlation map shows a wide deviation of the amplitude in a narrower range of height than in Figure 3a. In the meantime, height data was distributed around the middle point. One possible way to understand this pattern in the correlation map is to presume that the SEI layer fully covered the valleys between NG particles, which results in a relatively smooth and flat surface, whereas the active volume of Li ions varies more drastically. Furthermore, a weakly negative correlation between height and ESM amplitude was observed, which may explain why we observed a dark contrast along some in the topography image.

In order to elucidate the correlation further, we also extracted a profile along an arbitrary line over a 20 μm distance in both height and ESM amplitude images (see Figure 3b). As a result, the amplitude shows a high value where the height is relatively low (see red arrows), while the amplitude shows a low value where the height is relatively high (see black arrows). This negative relationship means that the SEI layer is preferentially deposited near lower-height regions, as shown in line profiles. The line profile analysis supports our correlation results where the ESM amplitude shows a wide deviation over a relatively narrow distribution of height.

In the case of LSTP24, some NG particles are clearly observed. The turtle-shell-shaped adsorbates are also clearly revealed on the NG surfaces. In the meantime, the ESM amplitude image showed an overall bright contrast without resemblance to the topography image, indicating that the SEI layer covers uniformly over the electrode. The image correlation map shows a wider distribution of the ESM amplitude in a narrower range of height when compared with LSTP12 in Figure 3b.

Moreover, the weakly negative relationship between height and ESM amplitude is also observed. However, the slope of the negative relationship is steeper in LSTP24 (purple color dotted line) than in LSTP12 (orange color dotted line). This can be explained by the fact that the electrode is more fully covered in the case of LSTP24 when compared with LSTP12, leading to less variation in the surface height but more variation of the active Li volume, which in turn results in more variation in the ESM amplitude in the valleys between NG particles for the fully covered case.

The widely spread-out ESM amplitude at the middle point, regardless of the narrowly distributed height, implies that the thick SEI layer covers both NG particles and a valley of their particles. In order to understand the correlation further, we conducted a profile analysis along an arbitrary line over a 20 μm distance in both height and ESM amplitude images (see Figure 3c).

As a result, we could see a small valley in the height profile. However, the ESM amplitude fluctuated between 40 and 90 pm regardless of the position along the line profile. Similar to the result of LSTP12, the amplitude shows a high value where the height is low (see red arrows), while the amplitude reveals a low value where the height is high (see black arrows). The line profile analysis also supports our correlation results.

In the case of the LSTP30 electrode, NG particles were clearly observed in the topography image. In the meantime, the ESM amplitude image contains overall dark contrast with a weak resemblance of topographic features. Furthermore, the brightest regions were observed on the surface of NG particles. Based on these two images, image correlation maps are obtained to confirm the relationship between height and ESM amplitude. The result shows a narrow deviation of the amplitude in a wide range of heights. In the meantime, the height distribution was concentrated at a height slightly larger than the middle region.

The narrowly distributed ESM amplitude, regardless of the spread-out height, implies that the thin SEI layer homogeneously covers the NG particles over the region of interest. Furthermore, the concentrated distribution of height indicates that the surface of NG particles is relatively flat, as the surface is occupied with SEI compounds. For a further understanding of the correlation, we extracted a profile along a line over the same distance in both height and ESM amplitude (see Figure 3d).

As a result, we could see a big valley in the height profile with an overall V-shape, similar to the result shown in Figure 3a. The ESM amplitude fluctuated between 10 and 40 pm regardless of the position along the line profile. The line profile analysis confirms our finding via correlation analysis, where the ESM amplitude shows a narrow deviation over a wide range of heights.

In order to further analyze the SEI layer on the electrodes, we measured the impedance of pristine and single-cycled states (Figure S6) using EIS and acquired the force–distance curves on the surface of cycled electrodes (Figure S7) for LSPT0 (reference) and LSTP30 samples. The EIS results revealed a semicircular pattern in both the LSTP0 and LSTP30 pristine electrodes with similar levels of charge transfer resistance (∼85 Ω for LSTP0 and ∼70 Ω for LSTP30). However, after 1 cycle, the charge transfer resistance in LSTP0 reduced to 65 Ω due to the generation of a highly ionic conductive SEI layer, while the charge transfer resistance in LSTP30 increased to ∼150 Ω.

The EIS findings indicate a higher irreversible capacity in the LSTP30 electrode compared to the reference electrodes, which is supported by AFM indentation measurements of the solid electrolyte interphase (SEI) layer thickness. It is evident that the SEI layer in the reference electrode is 45 nm thick with two layers, while the SEI layer in the LSTP30 electrode is 298 nm thick with three layers (see Figure S7), which strongly suggests that the increased irreversible capacity in the LSTP30 electrodes can be attributed to the enlarged thickness and increased charge transfer resistance of the SEI layer.

2.4. In-Depth Chemical Composition Analysis of the SEI Layer

We analyzed the height and ESM amplitude at the nanoscale and discovered a correlation between the distribution of Li ions in the solid electrolyte interphase (SEI) layer and its overall morphology as a function of LSTP content. To gain a more comprehensive understanding of the significant changes in the SEI morphology in LSTP24 and LSTP30, we utilized depth-resolved X-ray photoelectron spectroscopy (XPS) experiments with Ar ion etching to identify the chemical compounds in the SEI layers. Additionally, we included scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDS) results of pristine and cycled electrodes on the surface to determine the surface chemical compositions (see Figures S8 and S9).

Among the four different LSTP composite electrodes, we chose LSTP24 and LSTP30 that showed distinct differences in both height and ESM amplitude distributions. Here, we assumed that different chemical compounds might form in the SEI layer as we increase the LSTP amount in the composite electrode, which may account for the drastic change in both morphology and Li-ion distribution. In such a case, we expect to see peaks with different binding energies.

However, as shown in Figure S10, we observed that the peak positions in both LSTP24 and LSTP30 were identical in all binding energy ranges as a function of depth. Based on these results, we concluded that these changes occur even with the same chemical compounds in the SEI layer.

Therefore, we checked whether the ratio of each chemical compound could determine the morphology and ion distribution change in the SEI layer by analyzing its elemental ratio using XPS. Specifically, we analyzed the ratio of all four chemical elements (Li, O, F, and C) in the SEI layer as a function of etching time (see Figure 4).

Figure 4.

Figure 4

Open in a new tab

Atomic fractions of four elements of cycled electrodes in (a) LSTP24 and (b) LSTP30 in depth directions from XPS measurements.

The initial portion (38%) of Li in LSTP24 is the same as that in LSTP30. However, as the depth increases, the Li portion tends to increase and then decrease after passing through the maximum (43%) in LSTP24. On the other hand, in LSTP30, as the depth increases, the Li portion tends to decrease and then saturates after passing through the minimum (36%).

Since the ESM signal mostly reflects the ionic information near the surface region, the higher ESM amplitude for LSTP24 in comparison to LSTP30 in Figure 3c,d can be explained by the fact that the integration of Li ions near the surface of LSTP24 is higher than that of LSTP30.

In the case of F, its ratio directly represents the amount of LiF compounds in the SEI layer; hence, an excellent marker for its distribution. The fluorine element gradually decreased in LSTP30, while it fluctuated up to 20% and down to 10% in LSTP24. Based on our observation, LSTP24 shows a higher intensity deviation compared to LSTP30, leading to a more nonuniform passivation layer consisting mainly of LiF.

In addition, the distribution of oxygen and carbon elements can be linked to various compounds such as C=O, CO32–, Li2O, TiO2, Li2CO3, and carbonaceous species, as shown in Figure S10.

Table S1 summarizes the distribution of Li and F elements along the depth direction. The range of concentration was calculated by subtracting the minimum from the maximum and then normalized by the surface concentration. LSTP24 showed a larger normalized range (0.106) than LSTP30 (0.059), even though the surface concentration was quite similar.

The larger range can lead to a more nonuniform Li distribution along the depth direction in the SEI layer. Similarly, LSTP24 showed a higher normalized range of fluorine (0.615) compared to LSTP30 (0.498), leading to a more nonuniform distribution of the fluorine element similar to Li.

By comparing the above results with irreversible specific capacity and the Coulombic efficiency in Table 1, we found that the nanoscale fluctuation of Li and F components in the SEI layer is linked with the macroscopic irreversible capacity and Coulombic efficiency.

In order to understand our results, we came up with a hypothetical model, as depicted in Figure 5. The morphology and chemical homogeneity of the SEI layer is affected by the component in direct contact with the liquid electrolyte. As such, for LSTP0 and LSTP30, the component in direct contact is either NG or LSTP, respectively.

Figure 5.

Figure 5

Open in a new tab

Schematic diagrams of SEI layers formed on the electrode depending on the solid electrolyte contents. Homogeneous depositions show a wide height and a narrow amplitude, whereas inhomogeneous depositions show a narrow height and a wide amplitude.

The SEI properties (morphology and homogeneity) in conventional NG electrodes are confirmed with different degrees of lithiation. In particular, when the LSTP is almost uniformly covering the NG in LSTP30, the thickness of the SEI on the NG increases by the LSTP, but the increased SEI can also have homogeneity due to the uniform distribution of the LSTP.

Due to the electrochemical heterogeneity of NG and LSTP, it was confirmed that a stable SEI was formed in the electrode dominated by either NG or LSTP. On the other hand, for LSTP12 and LSTP24, the components in direct contact are both NG and LSTP. As a result, lithium-ion flux is affected by the distribution of LSTP. The interface between NG and LSTP accelerates the electrochemical reaction, which is exposed to the electrolyte in the cases of LSTP12 and LSTP24.

3. Conclusions

We analyzed solid electrolyte interphase (SEI) layers derived from the electrochemical effects of solid electrolytes on graphite-based composite electrodes. The decrease of the initial Coulombic efficiency after 1 cycle affects the surface morphology and chemical distribution significantly, depending on the contents of the solid electrolyte. Along with these results, we visualized Li-ion distribution in SEI layers by DART mode ESM and characterized both Li and F ions using XPS. We found a strong correlation between morphological variations on the electrode, Li and F distribution in the SEI layer, and Coulombic efficiency. Especially, our results provide important information about the nanoscale structure and property evolution of the SEI layer formed by competition between different Li sources, which will be useful for designing (all) solid-state electrolyte systems.

Among composite electrodes with different ratios of the NG and electrolyte, LSTP30 showed the most homogeneous physical and chemical SEI properties with a higher Coulombic efficiency, which attests to the fact that the solid electrolyte homogeneity has a strong impact on the performance, which is why LSTP30 works better than LSTP12 and LSTP24.

Furthermore, our findings provide insights into the physical and chemical properties of composite electrodes that can maximize the uniformity of the SEI layer and hence minimize irreversible capacity, which is critical to increase the charge–discharge performance of solid-state batteries.

4. Experimental Section

4.1. Fabrication of Composite Electrodes

All of the preparations were conducted in a dry room (dew point < −50 °C) and prepared via a tape casting method using a doctor blade with electrode slurry. The natural graphite (NG, BTR) (16 μm) and lithium silicon titanium phosphate (Jeong Kwan display) (D50 = 250 nm, ion conductivity = 8 × 10–4 S cm–1) were employed for an active material and a lithium-ion conductor, respectively. Carboxymethyl cellulose (CMC, Dai-Ichi Kogyo Seiyaku) and styrene-butadiene rubber (SBR, ZEON) were prepared as a binder with a weight ratio of 1:4. The CMC–SBR binder was ionized in distilled water, while its content was fixed at 2 wt % of the composite electrodes. Carbon black (Super-P, TIMCAL), an electron conductive material, was fixed at 0.5 wt%.

The weight ratio of NG and LSTP in composite electrodes was 97.5:0 (LSTP0), 85.5:12 (LSTP12), 73.5:24 (LSTP24), 67.5:30 (LSTP30), 47.5:50 (LSTP50) and 27.5:70 (LSTP70), respectively. The electrode slurry prepared under the above conditions was tape-cast on an 11 μm Cu film and dried at 100 °C for 12 h to remove the solvent and finally to obtain the electrode. After drying, the electrode was pressed with a line pressure of 1000 kgf. The loading level and the electrode thickness were determined to be about 2.27–2.79 mg cm–2 and 22–24 μm, respectively.

4.2. Preparation and Electrochemical Test of Cells

Coin cells (2032) for charge–discharge tests were fabricated by sequentially superimposing the composite electrode, a porous polyethylene separator (Celgard), and finally, the lithium metal (Honjo). The liquid electrolyte solution (400 μL) of 1.15 M LiPF6 dissolved in a 3:7 v/v mixture of ethylene carbonate and ethyl methyl carbonate was supplied by Enchem Co. Ltd.

Charge and discharge tests of the coin cells were carried out in the constant current mode using a galvanostatic cycler (TOSCAT 3000, Toyo Systems) in a voltage range of 0.01-1 V with a 0.1 C-rate (equivalent to 37.2 mA gNG)–1. Cell impedance was measured in the frequency range of 10–1–106 Hz with an amplitude of 10 mV using a frequency response analyzer (VSP, Biologic). The coin cells were operated at 27 °C.

4.3. Surface Characterizations

After electrochemical reactions of the coin cells, the cells were disassembled, and the composite electrodes were rinsed with the dimethyl carbonate solvent, followed by drying under vacuum for 12 h to remove the residual electrolyte species. In order to analyze the spatial distribution of components in electrode samples, we used a commercial atomic force microscope (AFM, MFP-3D Origin, Asylum research, Oxford Instruments). All AFM analyses in this study were conducted at 25 °C under ambient air conditions. A conductive cantilever with a Cr/Pt overall-coated silicon tip (ContE-G, Budget sensors) was used for each measurement.

The cantilever has a nominal spring constant of 0.2 N m–1 and a tip resonance frequency of 13 kHz for observing very small amplitude in amorphous SEI layers. The typical scan rate used for the measurements was 1 Hz, and the scan angle was 0°. The AFM topography images were obtained simultaneously during each AFM imaging mode. ESM imaging was performed to measure the electromechanical response of each sample.

In order to magnify the sensitivity of the measurement, we used dual-ac resonance tracking (DART) mode for ESM mapping. The AC drive voltage (2 V) with a contact resonance frequency from 70 to 90 kHz was applied via the AFM tip while the tip loading force was 250 nN. We conducted each experiment at least 10 times to ensure repeatability.

Scanning electron microscopy (SEM, EC SNE-4500M) and energy-dispersive X-ray spectroscopy (EDS, Bruker XFlash 640H Mini) were used for surface structural analysis and elemental mapping, respectively. The X-ray photoelectron spectroscopy (XPS) tests (Al Kα, Thermo VG Scientific) were performed for surface chemical analysis of cycled electrodes. After scanning of survey and surface measurements, a depth profile was carried out from 2 keV of an Ar+ ion beam for every 30 s. All spectra were calibrated by setting the C 1s photoemission peak for sp2-hybridized carbon to 284.5 eV.

4.4. Statistical Analysis

We obtained the Pearson correlation of the height (x) and the amplitude (y) image signals, measured at a fixed DC (0 V) and AC bias (2 V), using the ImageJ program. The dots in the Pearson correlation were plotted for each pixel of the two images, with their positions determined by the pixel values of the two pixels at identical coordinates. The values along both axes range from 0 to 255, representing the pixel values constituting the two images involved in the correlation process.

Acknowledgments

This work was supported by the National Research Foundation (NRF) funded by the Korean Ministry of Science & ICT (2020M3H4A3081880 and 2022K1A4A7A04095892) and the KAIST-funded Global Singularity Research Program for 2021, 2022, and 2023.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsami.3c02770.

  • Lithiation profiles of NG electrodes; voltage and specific capacity, as a function of LSTP contents; current–voltage profile; height maps of pristine and 1 cycled electrode in LSTP50 and LSTP70; height and amplitude at different 10 positions, and correlation maps at selected 5 positions from the height and amplitude images of 1 cycled electrodes; EIS results; force-indentation curves; SEM/EDS results; photoelectron signals during argon etching in C 1s, F 1s, O 1s, and Li 1s signals; depth distributions in lithium and fluorine (PDF)

The authors declare no competing financial interest.

Supplementary Material

am3c02770_si_001.pdf (1.4MB, pdf)

References

  1. Liu J.; Bao Z.; Cui Y.; Dufek E. J.; Goodenough J. B.; Khalifah P.; Li Q.; Liaw B. Y.; Liu P.; Manthiram A.; Meng Y. S.; Subramanian V. R.; Toney M. F.; Viswanathan V. V.; Whittingham M. S.; Xiao J.; Xu W.; Yang J.; Yang X.-Q.; Zhang J.-G. Pathways for Practical High-Energy Long-Cycling Lithium Metal Batteries. Nat. Energy 2019, 4, 180–186. 10.1038/s41560-019-0338-x. [DOI] [Google Scholar]
  2. Li W.; Erickson E. M.; Manthiram A. High-Nickel Layered Oxide Cathodes for Lithium-based Automotive Batteries. Nat. Energy 2020, 5, 26–34. 10.1038/s41560-019-0513-0. [DOI] [Google Scholar]
  3. Cheng X.-B.; Zhao C.-Z.; Yao Y.-X.; Liu H.; Zhang Q. Recent Advances in Energy Chemistry between Solid-State Electrolyte and Safe Lithium-Metal Anodes. Chem 2019, 5, 74–96. 10.1016/j.chempr.2018.12.002. [DOI] [Google Scholar]
  4. Meier K.; Laino T.; Curioni A. Solid-State Electrolytes: Revealing the Mechanisms of Li-Ion Conduction in Tetragonal and Cubic LLZO by First-Principles Calculations. J. Phys. Chem. C 2014, 118, 6668–6679. 10.1021/jp5002463. [DOI] [Google Scholar]
  5. Shin D. O.; Oh K.; Kim K. M.; Park K. Y.; Lee B.; Lee Y. G.; Kang K. Synergistic Multi-Doping Effects on the Li7La3Zr2O12 Solid Electrolyte for Fast Lithium Ion Conduction. Sci. Rep. 2015, 5, 18053 10.1038/srep18053. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Jung Y. S.; Oh D. Y.; Nam Y. J.; Park K. H. Issues and Challenges for Bulk-Type All-Solid-State Rechargeable Lithium Batteries using Sulfide Solid Electrolytes. Isr. J. Chem. 2015, 55, 472–485. 10.1002/ijch.201400112. [DOI] [Google Scholar]
  7. Kato Y.; Hori S.; Saito T.; Suzuki K.; Hirayama M.; Mitsui A.; Yonemura M.; Iba H.; Kanno R. High-Power All-Solid-State Batteries using Sulfide Superionic Conductors. Nat. Energy 2016, 1, 16030 10.1038/nenergy.2016.30. [DOI] [Google Scholar]
  8. Seino Y.; Takada K.; Kim B.-C.; Zhang L.; Ohta N.; Wada H.; Osada M.; Sasaki T. Synthesis of Phosphorous Sulfide Solid Electrolyte and All-Solid-State Lithium Batteries with Graphite Electrode. Solid State Ion. 2005, 176, 2389–2393. 10.1016/j.ssi.2005.03.026. [DOI] [Google Scholar]
  9. Fu K. K.; Gong Y.; Dai J.; Gong A.; Han X.; Yao Y.; Wang C.; Wang Y.; Chen Y.; Yan C.; Li Y.; Wachsman E. D.; Hu L. Flexible, Solid-State, Ion-Conducting Membrane with 3D Garnet Nanofiber Networks for Lithium Batteries. Proc. Natl. Acad. Sci. U.S.A. 2016, 113, 7094–7099. 10.1073/pnas.1600422113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Manuel Stephan A.; Nahm K. S. Review on Composite Polymer Electrolytes for Lithium Batteries. Polymer 2006, 47, 5952–5964. 10.1016/j.polymer.2006.05.069. [DOI] [Google Scholar]
  11. Baskoro F.; Wong H. Q.; Yen H.-J. Strategic Structural Design of a Gel Polymer Electrolyte toward a High Efficiency Lithium-Ion Battery. ACS Appl. Energy Mater. 2019, 2, 3937–3971. 10.1021/acsaem.9b00295. [DOI] [Google Scholar]
  12. Pitawala J.; Navarra M. A.; Scrosati B.; Jacobsson P.; Matic A. Structure and Properties of Li-Ion Conducting Polymer Gel Electrolytes based on Ionic Liquids of the Pyrrolidinium cation and the Bis(trifluoromethanesulfonyl)imide anion. J. Power Sources 2014, 245, 830–835. 10.1016/j.jpowsour.2013.07.045. [DOI] [Google Scholar]
  13. Zalewska A.; Dumińska J.; Langwald N.; Syzdek J.; Zawadzki M. Preparation and Performance of Gel Polymer Electrolytes Doped with Ionic Liquids and Surface-Modified Inorganic Fillers. Electrochim. Acta 2014, 121, 337–344. 10.1016/j.electacta.2013.12.135. [DOI] [Google Scholar]
  14. Gannepalli A.; Yablon D. G.; Tsou A. H.; Proksch R. Corrigendum: Mapping Nanoscale Elasticity and Dissipation using Dual Frequency Contact Resonance AFM. Nanotechnology 2013, 24, 159501 10.1088/0957-4484/24/15/159501. [DOI] [PubMed] [Google Scholar]
  15. Wen R.; Hong M.; Byon H. R. In Situ AFM Imaging of Li–O2 Electrochemical Reaction on Highly Oriented Pyrolytic Graphite with Ether-Based Electrolyte. J. Am. Chem. Soc. 2013, 135, 10870–10876. 10.1021/ja405188g. [DOI] [PubMed] [Google Scholar]
  16. Jeong S.-K.; Inaba M.; Abe T.; Ogumi Z. Surface Film Formation on Graphite Negative Electrode in Lithium-Ion Batteries: AFM Study in an Ethylene Carbonate-Based Solution. J. Electrochem. Soc. 2001, 148, A989 10.1149/1.1387981. [DOI] [Google Scholar]
  17. Lang S.-Y.; Shi Y.; Hu X.-C.; Yan H.-J.; Wen R.; Wan L.-J. Recent Progress in the application of In Situ Atomic Force Microscopy for Rechargeable Batteries. Curr. Opin. Electrochem. 2019, 17, 134–142. 10.1016/j.coelec.2019.05.004. [DOI] [Google Scholar]
  18. Balke N.; Jesse S.; Morozovska A. N.; Eliseev E.; Chung D. W.; Kim Y.; Adamczyk L.; Garcia R. E.; Dudney N.; Kalinin S. V. Nanoscale Mapping of Ion Diffusion in a Lithium-Ion Battery Cathode. Nat. Nanotechnol. 2010, 5, 749–754. 10.1038/nnano.2010.174. [DOI] [PubMed] [Google Scholar]
  19. Ramdon S.; Bhushan B.; Nagpure S. C. In Situ Electrochemical Studies of Lithium-Ion Battery Cathodes using Atomic Force Microscopy. J. Power Sources 2014, 249, 373–384. 10.1016/j.jpowsour.2013.10.099. [DOI] [Google Scholar]
  20. Zhang J.; Yang X.; Wang R.; Dong W.; Lu W.; Wu X.; Wang X.; Li H.; Chen L. Influences of Additives on the Formation of a Solid Electrolyte Interphase on MnO Electrode Studied by Atomic Force Microscopy and Force Spectroscopy. J. Phys. Chem. C 2014, 118, 20756–20762. 10.1021/jp503953n. [DOI] [Google Scholar]
  21. Kim H.; Oh J.; Park G.; Jetybayeva A.; Kim J.; Lee Y.-G.; Hong S. Visualization of Functional Components in a Lithium Silicon Titanium Phosphate–Natural Graphite Composite Anode. ACS Appl. Energy Mater. 2020, 3, 3253–3261. 10.1021/acsaem.9b02045. [DOI] [Google Scholar]
  22. Alikin D. O.; Ievlev A. V.; Luchkin S. Y.; Turygin A. P.; Shur V. Y.; Kalinin S. V.; Kholkin A. L. Characterization of LiMn2O4 Cathodes by Electrochemical Strain Microscopy. Appl. Phys. Lett. 2016, 108, 113106 10.1063/1.4943944. [DOI] [Google Scholar]
  23. Amanieu H.-Y.; Thai H. N. M.; Luchkin S. Y.; Rosato D.; Lupascu D. C.; Keip M.-A.; Schröder J.; Kholkin A. L. Electrochemical Strain Microscopy Time Spectroscopy: Model and Experiment on LiMn2O4. J. Appl. Phys. 2015, 118, 055101 10.1063/1.4927747. [DOI] [Google Scholar]
  24. Bradler S.; Schirmeisen A.; Roling B. Piezoresponse Force and Electrochemical Strain Microscopy in Dual AC Resonance Tracking Mode: Analysis of Tracking Errors. J. Appl. Phys. 2018, 123, 035106 10.1063/1.5004472. [DOI] [Google Scholar]
  25. Morozovska A. N.; Eliseev E. A.; Balke N.; Kalinin S. V. Local Probing of Ionic Diffusion by Electrochemical Strain Microscopy: Spatial Resolution and Signal Formation Mechanisms. J. Appl. Phys. 2010, 108, 053712 10.1063/1.3460637. [DOI] [Google Scholar]
  26. Proksch R. Electrochemical Strain Microscopy of Silica Glasses. J. Appl. Phys. 2014, 116, 066804 10.1063/1.4891349. [DOI] [Google Scholar]
  27. Zhang G.; Liu W.; Chen J.; Shen S. Nonlinear Electrochemomechanical Modelling of Electrochemical Strain Microscopy Mmaging. Nanotechnology 2020, 31, 315704 10.1088/1361-6528/ab898c. [DOI] [PubMed] [Google Scholar]
  28. Arruda T. M.; Kumar A.; Kalinin S. V.; Jesse S. Mapping Irreversible Electrochemical Processes on the Nanoscale: Ionic Phenomena in Li Ion Conductive Glass Ceramics. Nano Lett. 2011, 11, 4161–4167. 10.1021/nl202039v. [DOI] [PubMed] [Google Scholar]
  29. Alikin D. O.; Romanyuk K. N.; Slautin B. N.; Rosato D.; Shur V. Y.; Kholkin A. L. Quantitative Characterization of the Ionic Mobility and Concentration in Li-battery Cathodes via Low Frequency Electrochemical Strain Microscopy. Nanoscale 2018, 10, 2503–2511. 10.1039/C7NR08001H. [DOI] [PubMed] [Google Scholar]
  30. Giridharagopal R.; Flagg L. Q.; Harrison J. S.; Ziffer M. E.; Onorato J.; Luscombe C. K.; Ginger D. S. Electrochemical Strain Microscopy Probes Morphology-Induced Variations in Ion Uptake and Performance in Organic Electrochemical Transistors. Nat. Mater. 2017, 16, 737–742. 10.1038/nmat4918. [DOI] [PubMed] [Google Scholar]
  31. Xie H.; Alonso J. A.; Li Y.; Fernández-Díaz M. T.; Goodenough J. B. Lithium Distribution in Aluminum-Free Cubic Li7La3Zr2O12. Chem. Mater. 2011, 23, 3587–3589. 10.1021/cm201671k. [DOI] [Google Scholar]
  32. Wang Z.; Kotobuki M.; Lu L.; Zeng K. Nanoscale Characterization of Solid Electrolyte by Scanning Probe Microscopy Techniques. Electrochim. Acta 2020, 334, 135553 10.1016/j.electacta.2019.135553. [DOI] [Google Scholar]
  33. Strelcov E.; Yang S. M.; Jesse S.; Balke N.; Vasudevan R. K.; Kalinin S. V. Solid-State Electrochemistry on the Nanometer and Atomic Scales: the Scanning Probe Microscopy Approach. Nanoscale 2016, 8, 13838–13858. 10.1039/C6NR01524G. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Busche M. R.; Drossel T.; Leichtweiss T.; Weber D. A.; Falk M.; Schneider M.; Reich M. L.; Sommer H.; Adelhelm P.; Janek J. Dynamic Formation of a Solid-Liquid Electrolyte Interphase and its Consequences for Hybrid-Battery Concepts. Nat. Chem. 2016, 8, 426–434. 10.1038/nchem.2470. [DOI] [PubMed] [Google Scholar]
  35. Campbell C.; Lee Y. M.; Cho K. Y.; Lee Y. G.; Lee B.; Phatak C.; Hong S. Effect of Nanopatterning on Mechanical Properties of Lithium Anode. Sci. Rep. 2018, 8, 2514 10.1038/s41598-018-20773-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Kalinin S.; Balke N.; Jesse S.; Tselev A.; Kumar A.; Arruda T. M.; Guo S.; Proksch R. Li-Ion Dynamics and Reactivity on the Nanoscale. Mater. Today 2011, 14, 548–558. 10.1016/S1369-7021(11)70280-2. [DOI] [Google Scholar]
  37. Park G.; Kim H.; Oh J.; Choi Y.; Ovchinnikova O. S.; Min S.; Lee Y.-G.; Hong S. Quantitative Measurement of Li-Ion Concentration and Diffusivity in Solid-State Electrolyte. ACS Appl. Energy Mater. 2021, 4, 784–790. 10.1021/acsaem.0c02660. [DOI] [Google Scholar]
  38. Jetybayeva A.; Schön N.; Oh J.; Kim J.; Kim H.; Park G.; Lee Y.-G.; Eichel R.-A.; Kleiner K.; Hausen F.; Hong S. Unraveling the State of Charge-Dependent Electronic and Ionic Structure–Property Relationships in NCM622 Cells by Multiscale Characterization. ACS Appl. Energy Mater. 2022, 5, 1731–1742. 10.1021/acsaem.1c03173. [DOI] [Google Scholar]
  39. An S. J.; Li J.; Daniel C.; Mohanty D.; Nagpure S.; Wood D. L. III The State of Understanding of the Lithium-Ion-Battery Graphite Solid Electrolyte Interphase (SEI) and its Relationship to Formation Cycling. Carbon 2016, 105, 52–76. 10.1016/j.carbon.2016.04.008. [DOI] [Google Scholar]
  40. Hong S. Single Frequency Vertical Piezoresponse Force Microscopy. J. Appl. Phys. 2021, 129, 051101 10.1063/5.0038744. [DOI] [Google Scholar]
  41. Kim S.; No K.; Hong S. Visualization of Ion Transport in Nafion using Electrochemical Strain Microscopy. Chem. Commun. 2016, 52, 831–834. 10.1039/C5CC07412F. [DOI] [PubMed] [Google Scholar]
  42. Bradler S.; Schirmeisen A.; Roling B. Amplitude Quantification in Contact-Resonance-based Voltage-Modulated Force Spectroscopy. J. Appl. Phys. 2017, 122, 065106 10.1063/1.4998435. [DOI] [Google Scholar]
  43. Han Y.; Ryu J.; Oh C.; Kim S.; Schiffer A.; Killgore J. P.; Hong S. Deciphering Osteoconductive Surface Charge Effects in Sintered Hydroxyapatite via Piezoresponse Force Microscopy. J. Appl. Phys. 2021, 129, 094902 10.1063/5.0028525. [DOI] [Google Scholar]
  44. Schön N.; Schierholz R.; Jesse S.; Yu S.; Eichel R.-A.; Balke N.; Hausen F. Signal Origin of Electrochemical Strain Microscopy and Link to Local Chemical Distribution in Solid State Electrolytes. Small Methods 2021, 5, 2001279 10.1002/smtd.202001279. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

am3c02770_si_001.pdf (1.4MB, pdf)

Articles from ACS Applied Materials & Interfaces are provided here courtesy of American Chemical Society

RESOURCES