Binder-Free, Thin-Film Ceramic-Coated Separators for Improved Safety of Lithium-Ion Batteries.

Separators play a crucial role in ensuring the safety of lithium-ion batteries (LIBs). Commercial polyolefin-based separators such as polyethylene (PE) still possess serious safety risks under abuse conditions because of their poor thermal stability. In this work, a novel type of binder-free, thin ceramic-coated separators with superior safety characteristics is demonstrated. A thin layer of alumina (Al2O3) is coated on commercial PE separators using the electron-beam physical vapor deposition (EB-PVD) technique. Scanning electron microscopy (SEM), contact angle, impedance spectroscopy, and adhesion test techniques were employed to evaluate structure-property correlations. When compared to commercial slurry-coated separators, the EB-PVD-coated separators display (i) higher thermal stability, (ii) stronger ceramic-polymer adhesion, and (iii) competitive electrochemical performance of full LIB cells. Thermal stability, in terms of improved shutdown and breakdown characteristics of the separator, was studied using the in situ impedance technique up to 190 °C. In addition, the improved adhesion of the ceramic layer deposited on the PE separator was studied following the tape adhesion strength test. We prove that the thin (binder-free) ceramic layer coated by EB-PVD is far more effective in improving separator safety than those made using the conventional thick slurry coating.

Introduction

Since the commercialization by Sony in 1991, lithium-ion batteries (LIBs) have played an increasingly important role as power sources for portable devices including consumer electronics and electric vehicles, because of their high operating voltage, high energy density, long cycle life, and high efficiency.[1] However, battery safety under abuse conditions still needs to be improved, especially for automobile and aviation applications.[2] The most common types of abuse include mechanical, thermal, and electrochemical, all of which eventually result in thermal runaway and battery fire/explosion.[3]

One of the most critical components for battery safety is the separator, which is a thin, porous membrane that physically separates the cathode from the anode. Properties of separators play an important role in determining the thermal response of batteries during an abuse event. Commercial state-of-the-art polyolefin-based LIB separators comprise of PE, polypropylene (PP), or hybrids of PE and PP. Although PE and PP-based separators offer excellent mechanical properties, they are susceptible to thermal failure because of their relatively low transition temperatures (135 °C for PE and 165 °C for PP).[4] In addition, polyolefin-based separators generally display poor wetting properties with carbonate-based electrolytes used in LIBs.[5]

Most commercial separators display shutdown behavior during an abuse event as a safety feature.[6] When the cell temperature rises near the melting point of the polymeric material, the pores inside the separator collapse, resulting in a significant increase in cell impedance ceasing the battery operation. This is known as separator shutdown. As the cell temperature continues to rise, shrinkage and breakage of the separator occur, causing direct contact between the cathode and the anode. Such a process is referred to as separator breakdown. This event is characterized by a sharp reduction of impedance, as the electrolyte can flow freely. Separator breakdown and the resulting internal short speed up the thermal runaway process. Therefore, it is desired that the shutdown temperature and breakdown temperature of separators are as far apart as possible in order to avoid or delay the thermal runaway process.

Commercial tri-layer PP/PE/PP separators take advantage of the difference in the melting point of PP and PE, using PE as the shutdown layer and PP to protect structural integrity.[7] Unfortunately, such protection is only effective below the melting point of PP. Commercially available ceramic-coated separators have also gained popularity, where ceramic particles with binders are slurry-coated on polymer membranes. Although the micron-thick ceramic coating has been proven effective in improving the thermal stability of separators to a certain extent, it adds extra weight, volume, and processing time, and the effectiveness of the protection is still limited by the thermal stability of the polymeric binder used.[8,9] More importantly, delamination and slippage of the ceramic layer could occur due to poor binding to the polymer separator, which was identified as one of the key contributing factors to battery failure in the Samsung Galaxy Note 7 incident.[10,11] Other approaches under development include electrospinning or fabricating alumina- or alumina/phenolphthalein polyetherketone-based, porous ceramic membranes.[12−15] Despite the good thermal stability of such separators, their practical use in LIBs is limited because of their poor mechanical properties.

Physical vapor deposition techniques such as electron beam physical vapor deposition (EB-PVD), magnetron sputtering, pulsed laser deposition (PLD), and so forth have been widely applied to deposit binder-free, ceramic (inorganic) thin films with better thickness and morphology control than the slurry-coating technique.[16−21] Among the several thin film deposition methods, EB-PVD is a fast (2 nm/s) and scalable process that produces dense, uniform ceramic layer, and need no post fabrication conditioning. It employs an electron beam (EB) source that can evaporate a target at a very high rate (≈2 nm/s) and deposit on a fixed large surface area or roll-to-roll fabrication required for large-scale battery manufacturing. In the present work, we evaluate the properties of a novel separator fabricated by depositing a thin layer (≈100 nm) of Al2O3 ceramic on the PE separator (ENTEK) using EB-PVD equipment with a deposition area >200 cm2 and with a fast deposition rate ≈2 nm/s. Al2O3 has been selected for coating because of its low cost, high thermal stability, and good electrolyte wettability.[22,24] The goal was to improve the thermal stability of the separators beyond the capability of commercial thick (1–2 μm) Al2O3 with a binder by a thin (≈100 nm) Al2O3 binder-free coating without compromising the electrochemical performance. The performance of such EB-PVD-coated Al2O3-PE (100 nm) separators is compared with three other commercially available separators [PE, commercial 1–2 μm Al2O3-PE with the binder, and Al2O3 coated polyimide (PI) separators]. Thus, the novelty lies in development of a binder-free and thin-film ceramic coating on the battery separator that not only improved thermal performance of the resultant separator but also either maintained or improved the electrochemical performance in a full Li-ion cell.

Results and Discussion

Morphology

The PE separator (Figure a) has a uniformly interconnected fiber network with a pore size of around 300 nm. Figure b shows the surface morphology of the EB-PVD-coated Al2O3-PE separator. The ceramic layer completely covers the PE separator surface. The coating appears uniform and dense. However, cracks of around 200 nm wide are present. The size of pores in the PE separator is several nanometers to micrometers, and thus, a crack with 100 nm coating is expected. Though a crack-free Al2O3 can be achieved at higher Al2O3 thickness and with post deposition thermal conditioning, a thick layer may negatively impact electrochemical performances by blocking Li+ transport, as Al2O3 is an insulating material. The advantage of thin-film Al2O3 together with discontinuity in the Al2O3 layer is to facilitate lithium-ion transport and still improve thermal stability while maintaining or improving electrochemical performances. The ceramic layer on the commercial Al2O3-PE slurry-coated separator (Figure c) consists of uniformly distributed, but loosely connected Al2O3 spherical particles.[23] The PI separator (Figure d) is composed of a polyimide fiber matrix with Al2O3 particles on top and within. The PI fibers appear to be 1–2 μm in diameter with pores as large as 2 μm across. This is in stark contrast to the PE separators, where the diameter of fibers and pore sizes are much smaller. The Al2O3 particles have irregular shapes, wide particle size distribution from 200 nm to 5 μm, and seem to preferably reside inside the pores of the PI matrix. The distinct morphology of the Al2O3 coating on commercial Al2O3-PE and PI separators may stem from the manufacturing process, which is proprietary. Table S1 summarizes other physical properties of all separators including separator thickness, weights, and volume. The increase in the weight and thickness of the PE separator with the Al2O3 coating by EB-PVD has also been calculated and compared to that of commercial slurry-coated separators (using the density of Al2O3 as 3.95 g/cm3). Also, the presence of Al2O3 ceramic coating by EB-PVD on the surface of the PE layer was verified from the energy-dispersive spectroscopy (EDS, Genesis 2000) and has been shown in Figure S2.

Figure 1

Morphology of (a) PE, (b) EB-PVD-coated Al2O3-PE (Al2O3, 100 nm), (c) commercial Al2O3-PE (Al2O3, 1–2 μm), and (d) Al2O3 coated PI separators, respectively.

Contact Angle and Electrolyte Uptake

As seen in Figure a, a high contact angle between the liquid electrolyte and separator surface of 45.2° was observed for the pristine PE separator. It is well known that polyolefin-based separators have poor wettability toward polar organic liquid electrolytes because of their non-polar nature.[24] In contrast, the contact angle on the EB-PVD-coated Al2O3-PE separator was 33.7°. Therefore, the surface wetting properties were significantly improved because of the presence of the inorganic Al2O3 layer. This improvement in wettability could be attributed to the strong interactions between the oxygen-containing groups in Al2O3 on the coated layer and the carbonate solvents of the electrolytes.[24,25] Similarly, the contact angle for the commercial Al2O3-PE and Al2O3-PI separators was 32.5 and 29.2°, respectively. Although the morphology of the Al2O3 layer also affects the wetting properties, the differences in the contact angles were small among the three coated separators but large between coated and non-coated separators, indicating that the ceramic (Al2O3) material property is the dominating factor in determining separator wettability.

Figure 2

(a) Contact angle images and (b) electrolyte uptake measurement of the PE, EB-PVD coated Al2O3-PE (Al2O3, 100 nm), commercial Al2O3-PE (Al2O3, 1–2 μm), and Al2O3-coated PI separators.

Electrolyte uptake is largely determined by the porosity and wettability of separators. The electrolyte uptake for the PE separator is 51.6% (Figure b). In contrast, the uptake for EB-PVD-coated Al2O3-PE, commercial Al2O3-PE, and Al2O3-PI separators is 57.6, 60.9, and 62.4%, respectively. In theory, the EB-PVD-coated Al2O3-PE (100 nm) and PE separators should have identical porosity, ignoring the very thin ceramic layer, but EB-PVD-coated Al2O3-PE (100 nm) has 5–6% more electrolyte uptake. This increase in the electrolyte uptake suggests that the ceramic (Al2O3) layer increases electrolyte retention possibly because of a combination of physical and chemical effects. Similar increase in the electrolyte uptake behavior has been seen for both the commercial Al2O3-PE and the Al2O3-PI separators. The highest electrolyte uptake of Al2O3-PI can be attributed to larger porosity. According to the product specification, the porosities of Al2O3-PI and pristine PE separators are 51 and 43%, respectively.

Thermal Stability

The thermal stability of separators is critical to the safety of batteries as it is the separator that keeps cathode and anode physically/electrically separated, as dimensional loss in the separator will allow electrodes to come in physical/electrical contact that will lead to cell shorting. As seen from Figure , the pristine PE separator shows the highest shrinkage of about 20% at 150 °C and 70% at 200 °C. The EB-PVD coated Al2O3-PE separator showed no shrinkage at 150 °C and shrank by 10% at 200 °C. In contrast, the commercial Al2O3-PE separator shrank by about 10% at 150 °C and 30% at 200 °C. No shrinkage was observed for the Al2O3-PI separators at 200 °C, as polyimide is thermally stable up to 350 °C. Optical images of all separators are shown in Figure S3. Therefore, ceramic coating improves the thermal stability of PE separators regardless of the coating method. However, the EB-PVD-coated Al2O3-PE separator showed much smaller shrinkage despite the Al2O3 layer in this method being only 1/20 of the thickness of the Al2O3 layer on the slurry-coated commercial Al2O3-PE separator. This could be attributed to the binders used in the ceramic layer of commercial Al2O3-PE, which melts at around ∼170 °C. On the contrary, the ceramic layer of the EB-PVD coated Al2O3-PE is free of any binder or organic component. Therefore, the polymer network is mechanically supported by the rigid ceramic layer at higher temperatures where large shrinkage would occur without coating.

Figure 3

Thermal shrinkage vs temperature plot of PE, EB-PVD-coated Al2O3-PE (Al2O3, 100 nm), commercial Al2O3-PE (Al2O3, 1–2 μm), and Al2O3-coated PI separators at measured at 130, 150, and 200 °C, respectively.

Separator Shutdown and Breakdown (In Situ Impedance Measurement)

To further characterize the thermal stability (separator shutdown and breakdown) of the separators, in situ impedance was measured from 100 to 190 °C (Figure ). The measurement was conducted till 190 °C because of the limitation of existing equipment (Tenney environmental chamber). The technique provides a more accurate understanding of the thermal stability of separators in the presence of a liquid electrolyte (real-life application).

Figure 4

Shutdown and breakdown behavior (in situ impedance measurement) of pristine PE, EB-PVD-coated Al2O3-PE (Al2O3, 100 nm), commercial Al2O3-PE (Al2O3, 1–2 μm), and Al2O3-coated PI separators soaked in a standard liquid electrolyte.

Figure shows the impedance evolution of SS/separator + 1 M LiPF6 (EC/DMC/EMC) electrolyte/SS cells heated from 100 to 190 °C. The impedance of all separators soaked in the standard liquid electrolyte is largely constant up to 150 °C as there are no significant changes in the pore structure of the separators. For the pristine PE separator, the impedance substantially increased at around 160 °C, marking the onset of separator shutdown; the impedance reached a maximum value at 166 °C and then dropped sharply at around 170 °C. It should be noted that since the temperature of the oven ramps at a finite rate, we expect that the actual temperature of the cells lags behind the nominal oven temperature. The initial increase in impedance was due to the pore closure of the PE separators. With further increase of temperature, the PE membrane started melting and lost its integrity, resulting in macroscopic holes (separator breakdown) in the membrane and therefore a sudden drop in impedance. The cell was opened, and the separator was subjected to visual inspection. In Figure S4a, a large hole can be seen in the PE separator which was the result of excessive shrinkage that ripped the separator. For the EB-PVD-coated Al2O3-PE separator, the impedance remained stable until 190 °C, showing no shutdown nor breakdown behavior, which indicates that the polymeric structure of the separator remained largely stable throughout the measurement.[23] Upon post-mortem inspection, no breakage or ripping was observed (Figure S4b), and the ceramic layer looks uniform and pristine. The absence of separator breakdown was further confirmed by testing the insulation of the separator using multimeter probes. For the commercial Al2O3-PE, the separator still displays breakdown behavior starting at 170 °C, although the impedance drop is more gradual. A window of 20 °C between the shutdown temperature and breakdown temperature can be attributed to the enhancement of thermal stability by the coated commercial Al2O3 layer, but the enhancement is marginal. Post-mortem inspection indicated that the ceramic (Al2O3) layer (whitish) has shrunk to a smaller area (circled) at the center of the separator (Figure S4c). The periphery of the separator appears to be uncoated. A small rip was also observed in the uncoated area, which was further confirmed with the probe test. Finally, the Al2O3-coated PI separators showed no impedance/visual change (Figure S4d), which is expected as polyimide is thermally stable up to at least 350 °C (Figure S5).[26] Therefore, the impedance of the Al2O3-coated PI cell provides a baseline of how the impedance of the electrolyte evolves when there is no morphology change of the separators.

The morphology of the separators after in situ impedance measurement was characterized by scanning electron microscopy (SEM) (Figure ). The pristine PE separator looks completely melted with no porosity. No obvious morphology changes were observed for the ceramic layer of the EB-PVD-coated Al2O3-PE separator. Although the polymeric structure below cannot be observed, the absence of separator shutdown suggests that pore closure should be minimal. Also, there was no sign of delamination of the ceramic coating from the polymer layer for the EB-PVD-coated Al2O3-PE. However, cracks are visible which is likely due to strains induced by the shrinkage of the PE upon cooling. In contrast, the ceramic particles in the commercial Al2O3-PE separator agglomerated, leaving some area uncoated. The PE layer underneath was dense, indicating complete melting. The agglomeration of Al2O3 particles was likely due to the melting of the PVDF-binders (around 170 °C).

Figure 5

Post-mortem SEM morphology of PE, EB-PVD-coated Al2O3-PE (Al2O3, 100 nm), commercial Al2O3-PE (Al2O3, 1–2 μm), and Al2O3-coated PI separators.

Al2O3-PE Adhesion Test

To compare the adhesion between the ceramic (Al2O3) coatings with the polymer (PE) membranes, a simple tape peeling test was performed (Figure ). Briefly, a piece of scotch tape was pressed on the separator with the same force and then peeled off normally. For the commercial Al2O3-PE slurry-coated separator, the entire ceramic (Al2O3) layer was peeled off the separator and stuck to the tape (Figure c), indicating that the adhesion between the ceramic layer and the polymer is weak. In contrast, for the EB-PVD coated Al2O3-PE, there was no change in the morphology before and after conducting the peel test. No damage can be seen on the ceramic coating on the separator, and no ceramic was observed on the tape side (Figure d). This suggests that the ceramic layer is strongly bonded to the polymer membrane, which is another key reason for the enhanced structural integrity of the separators developed in the present investigation under thermally abuse conditions. The adhesion strength in the case of Al2O3-coated PI separators is also weak (Figure S6).

Figure 6

Morphology of the junction area (a,b) and tape side (c,d) after the peeling test of commercial Al2O3-PE (a,c) and EB-PVD-coated Al2O3-PE (b,d).

Electrochemical Performance

The impedance of SS/separator/SS cells was measured to characterize ion transport through the separator soaked in liquid electrolytes (Figure a). The impedance of the soaked separator corresponds to the intercept value on the real axis (x-axis). All the ceramic-coated separators (EB-PVD-coated Al2O3-PE, commercial Al2O3-PE, and Al2O3-coated PI separator) showed slightly higher impedance than pristine PE, probably due to the high coverage of the ceramic layer and the inactive nature of Al2O3.

Figure 7

(a) Impedance of symmetric cells (electrolyte impedance) with stainless steel electrodes; (b) Electrochemical performance of NMC111/graphite full cells using PE, EB-PVD-coated Al2O3-PE (Al2O3, 100 nm), commercial Al2O3-PE (Al2O3, 1–2 μm), and Al2O3-coated PI separators measured at room temperature (22 °C).

Cycling stability of NMC111/graphite full cells using different separators was characterized to evaluate the effect of ceramic coating on electrochemical performance (Figure b). All cells were cycled at 0.05 C (formation) followed by 0.5 rate (Figure S7). Cells with the EB-PVD-coated Al2O3-PE separator exhibit comparable electrochemical performance with that of all other commercial separators (PE, commercial Al2O3-PE, and Al2O3-coated PI). The capacity fade and low coulombic efficiency (CE) observed for initial cycles are due to the presence of hydrofluoric acid (HF) in the electrolyte.[27] The cells with the Al2O3-coated PI separator showed lower capacity fading, which may be attributed to the Al2O3 coating ceramic layer as Al2O3 is known to scavenge HF.[28] We suspect that the limited amount of ceramic materials in EB-PVD-coated Al2O3-PE separator (Al2O3, 100 nm) was not as effective in scavenging HF. Nonetheless, the ceramic (Al2O3) coating through the EB-PVD method is competitive to commercial Al2O3-PE (slurry coated). The cycling performance difference may not entirely depend on HF scavenging by Al2O3. Other factors like type of binder, percentage of binder, type of Al2O3, and so forth used to fabricate the Al2O3 layer in commercial Al2O3/PE and Al2O3/PI separators along with porosity and tortuosity differences may play a role in determining the cycling performance. More work needed to be done for further understanding. Post cycling, the EB-PVD-coated Al2O3-PE separator (Al2O3, 100 nm) cell was disassembled and the separator was observed under SEM (Figure S8). No delamination was observed, and the Al2O3 coating remained stable without much degradation.

Enhanced Volumetric/Gravimetric Capacity

Thicker separator results in a larger cell weight/volume, which leads to lower specific energy and energy density. The Al2O3 layer (Al2O3, 1–2 μm) with a binder on the commercial Al2O3-PE separators adds 30% of volume and nearly 60% of weight to the PE separators. In contrast, the Al2O3 layer (Al2O3, 100 nm) without any binder on the EB-PVD coated Al2O3-PE separator adds only 1.6% of volume and 15% of the weight. Using a calculation based on NMC111/graphite chemistry, it is estimated that the energy density of cells using EB-PVD-coated Al2O3-PE (100 nm) will be 3% higher than cells using the commercial Al2O3-PE (1–2 μm) slurry-coated separator.

This analysis demonstrates that the EB-PVD coated Al2O3-PE fabrication approach has multiple advantages especially improved thermal and mechanical stabilities (safety) and increased volumetric/gravimetric densities and thus particularly suitable for applications (electronics, electric vehicles, defense equipment, and electric vehicles for aerospace and space), where safety and better energy densities are key desirables.

Conclusions

In this study, binder-free thin-film ceramic-coated separators were prepared using EB-PVD for rechargeable LIBs. With only 100 nm of coating, EB-PVD-coated Al2O3-PE exhibits superior thermal stability, higher electrolyte uptake, and stronger ceramic-polymer adhesion than commercial slurry-coated Al2O3-PE. The improvement in the thermal stability was also confirmed in the presence of the standard liquid electrolyte by an in situ impedance measurement technique that showed negligible shutdown and breakdown behavior for the EB-PVD coated Al2O3-PE with stable impedance up to 190 °C. The electrochemical cycling of full cells (NMC111/graphite electrodes) with the EB-PVD-coated Al2O3-PE separator has also been compared with other commercial separators and has shown comparable performance at room temperature. Therefore, EB-PVD-coated Al2O3-PE separators can be an attractive option for advanced LIBs with enhanced safety, higher energy densities, and no compromise to electrochemical properties. Determining the effect of temperature on cell cycling, electrolyte additives, and so forth on these coated separators will be an important focus of our future work.

Experimental Section

Commercial Separators and Preparation of Ceramic-Coated Separators by EB-PVD

Al2O3 powder (Aldrich) was pressed into pellets using a hydraulic press before evaporation by EB-PVD so that Al2O3 powder does not fly inside the vacuum chamber. A 12 μm thick PE separator (ENTEK) was used as a substrate to deposit Al2O3 layers (100 nm) on each side of the separator using EB-PVD. The Al2O3 pellet was placed in a graphite crucible. The chamber was evacuated to a base pressure of <10–6 Torr. A deposition rate of 1.0–2.0 nm/s was used. The deposition rate was controlled by the power of the electron beam (EB) and the distance between the crucible and the substrate. The performance (thermal, mechanical, electrical, and electrochemical) of the binder-free, thin-film Al2O3-coated PE separator developed and being reported herein was compared with the commercial thick-film Al2O3-PE slurry-coated separator and Al2O3-PI separators. PE separators are obtained from ENTEK. Polyimide-based separators were from Jiangxi Advanced Nanofiber S&T Co., and the commercial thick-film Al2O3-PE slurry-coated separator were obtained from an internal sources.

Surface morphologies of the separators were investigated using high-resolution scanning electron microscopy (HRSEM, S-4800, Hitachi High Technologies). The samples were coated with 6 nm of gold using a sputtering system (Hummer 6.2 sputter system). For post-mortem analysis, all separators were removed from tested cells and rinsed with dimethyl carbonate (DMC) to remove the residual electrolyte before SEM observation.

Linear Thermal Shrinkage

To determine the thermal shrinkage of the separators, circular pieces (17.05 mm diameter) of the separators were annealed at different temperatures for 2 h, and their resultant diameters were measured before and after annealing. Based upon the diameter, the area of the separators was calculated, where S1 and S2 represent the area of the separator before and after annealing in air. The shrinkage was calculated using the following equation

Electrolyte Uptake

Separators were punched with an 18 mm punch and weighed (W1). They were then immersed in standard liquid electrolyte [1 M lithium hexafluorophosphate (LiPF6) in ethylene carbonate: dimethyl carbonate: ethyl methyl carbonate (EC/DMC/EMC) (1:1:1–v/v/v) with 2 wt % vinylene carbonate (VC) electrolyte] for 24 h. The excess electrolyte solution on its surface was removed by leaving the separators on the surface of the vials for 1 h before measuring the final weights (W2). The initial weight of all separators was measured and summarized in Figure S1. The liquid uptakes of the separators were measured in a glovebox filled with argon according to the following equation

Separator Wettability

Contact angle measurement was used to evaluate the interaction between liquid electrolytes and separators. The experimental setup consists of a syringe with a needle whose tip was fixed at a few millimeters above the separator. The separators were mounted on a glass slide, and consecutive images of the separator were taken 30 s after dropping the standard electrolyte. The volume of the electrolyte dropped on the separator was estimated to be 12.1 μL. The contact angles were then measured using the ImageJ software.

Separator Shutdown and Breakdown

Separators were assembled in coin cells with a copious amount of the standard electrolyte. Stainless steel plates were employed as current collectors. Polytetrafluoroethylene gaskets were employed, and a clamp setup (Figure ) was used to counter the pressure inside the coin cell because the test temperature is above the boiling temperature of some electrolyte components. The cell temperature was raised at 1 °C/min from 100 to 190 °C in an environmental chamber, while the cell impedance was monitored using a Solartron 1260/1287 potentiostat.

Figure 8

Fixture for in situ impedance measurement on separators with a standard liquid electrolyte to determine shutdown and breakdown behavior of separators.

Electrode Preparation

Coin cells were made with Li[Ni0.33Mn0.33Co0.33]O2 (NMC111) as the positive electrode (cathode) and graphite as a negative electrode (anode). NMC111 powder was obtained from MTI Corporation (particle size, D50 = 9.0–12.0 μm). The electrode formulation was 94 wt % active material, 3 wt % poly(vinylidene fluoride) (PVDF) binder (MTI), and 3 wt % super P conductive carbon. Electrode materials and n-methyl pyrrolidone solvent (NMP) were mixed in a planetary mixer (Thinky, USA) at 2000 rpm for 10 min. The resulting ink was coated onto aluminum foil using an automatic thick-film coating machine (MTI) and dried at 120 °C in a vacuum oven overnight. The cathode had a mass loading of 22.4 mg/cm2, corresponding to an areal capacity of about 3.1 mA h/cm2.

The anode was composed of graphite (MTI, TB-17), carbon black, oxalic acid (EMD Millipore Corporation), and PVDF binder at a weight ratio of 85:5:0.5:9.5. A similar slurry coating process was used to prepare the anode film on copper current collectors. The anode had a mass loading of 12.8 mg/cm2, corresponding to an areal capacity of about 3.4 mA h/cm2.

Cell Assembly and Electrochemical Measurements

2032 coin-type cells with standard liquid electrolyte, graphite anode, and NMC111 cathode were assembled in an argon-filled glovebox (O2, H2O < 1 ppm). 75 μL of a standard electrolyte was added for each cell. All LIB cells were conditioned at room temperature (22 °C) by cycling at C/20 for 3 cycles and then cycled at C/2 between 3.0 and 4.3 V without hold periods. To determine the conductivity of all separators, AC impedance data were also obtained for stainless steel/separators/stainless steel (SS/separators/SS) symmetric cells with standard liquid electrolyte over a frequency range 0.1 Hz to 1 MHz.