Silicene Flowers: A Dual Stabilized Silicon Building Block for High-Performance Lithium Battery Anodes Xinghao Zhang,†,‡ Xiongying Qiu,† Debin Kong,† Lu Zhou,† Zihao Li,† Xianglong Li,*,†,‡ and Linjie Zhi*,†,‡ †
CAS Key Laboratory of Nanosystem and Hierarchical Fabrication, CAS Center for Excellence in Nanoscience, National Center for Nanoscience and Technology, Beijing 100190, PR China ‡ University of Chinese Academy of Sciences, Beijing 100049, PR China S Supporting Information *
ABSTRACT: Nanostructuring is a transformative way to improve the structure stability of high capacity silicon for lithium batteries. Yet, the interface instability issue remains and even propagates in the existing nanostructured silicon building blocks. Here we demonstrate an intrinsically dual stabilized silicon building block, namely silicene flowers, to simultaneously address the structure and interface stability issues. These original Si building blocks as lithium battery anodes exhibit extraordinary combined performance including high gravimetric capacity (2000 mAh g−1 at 800 mA g−1), high volumetric capacity (1799 mAh cm−3), remarkable rate capability (950 mAh g−1 at 8 A g−1), and excellent cycling stability (1100 mA h g−1 at 2000 mA g−1 over 600 cycles). Paired with a conventional cathode, the fabricated full cells deliver extraordinarily high specific energy and energy density (543 Wh kgca−1 and 1257 Wh Lca−1, respectively) based on the cathode and anode, which are 152% and 239% of their commercial counterparts using graphite anodes. Coupled with a simple, cost-effective, scalable synthesis approach, this silicon building block offers a horizon for the development of highperformance batteries. KEYWORDS: silicene flower, silica fume, building block, silicon anode, lithium-ion battery
T
Si, since the strain in such nanostructured Si can be easily relaxed due to its small size and the available surrounding free space and consequently the degradation of the mechanical integrity of Si can be minimized; various nanostructured Si paradigms, such as zero-dimensional (0D) nanoparticles and nanospheres,15−24 one-dimensional (1D) nanowires and nanotubes,7,11,25−29 as well as three-dimensional (3D) porous nanostructures,30−35 have long been adopted as the basic building blocks of Si anodes, exhibiting very high capacity and good cycling stability. As primary Si building blocks (especially
he ever-growing demands of portable electronic devices, electric vehicles, smart grids, and renewable energy utilizations have stimulated a surge in developing high power and energy density lithium-ion batteries (LIBs).1−3 The key to fabricate such energy storage devices is to exploit high-performance cathode and anode materials.4−6 Silicon (Si) is a promising anode material7,8 for next-generation LIBs, having the highest known theoretical capacity, which is about ten times that of commercial graphite anodes. However, silicon anodes suffer from severe capacity fading upon cycling; this can be attributed to two critical factors including the structural degradation9 and interfacial instability10−12 caused by its large volume change (∼300%) during lithiation and delithiation. The nanostructuring13,14 has thus far been shown to be a transformative way to improve the structural stability of © 2017 American Chemical Society
Received: June 5, 2017 Accepted: July 10, 2017 Published: July 10, 2017 7476
DOI: 10.1021/acsnano.7b03942 ACS Nano 2017, 11, 7476−7484
Article
www.acsnano.org
Article
ACS Nano
Figure 1. Inspiration from disassembling a Si thin film into its 2D building blocks and fabrication of silicene flowers (SF). (a) Schematic showing the nominal disassembly of a Si thin film with stable cycling upon lithiation and delithiation, into 2D Si building blocks inheriting limited variable SEI area in the lateral surface and infinite nearly invariable SEI area on the planar surface. (b) Specific variable SEI area of 0D, 1D, 2D, and 3D Si building blocks versus a/t ratio of the 2D Si building block. (c) Schematic of the production process of SF involving the magnesiothemic reduction of silica fume at 850 °C and higher. This SF building block features high tap density, three-dimensional electron/ lithium ion transport channels, reduced lithium ion diffusion length, and more importantly limited variable SEI area as annotated. (d) Photograph of silica fume. (e) TEM image of silica fume. (f) Photographs of SF, Si microparticle (Alfa Aesar), Si nanoparticle (ST-NANO), and Si nanoparticle (Alfa Aesar) of the same weight (10 g) from left to right.
commercial applications, specifically in terms of gravimetric capacity, volumetric capacity, rate capability, and also cycling stability at the same time.14,15,17 We believe that the situation associated with synthetic complexity and/or unsatisfactory performance holds inevitably for other emerging electrode materials subject to large volume changes. Tracing the early studies of silicon anodes, Si thin films especially down to 50 nm in thickness,9,37 if ignoring such an industrially unacceptable material loading, have been reported exhibiting very high capacity and stable cycling over hundreds of cycles, even without the introduction of a second phase such as carbon. In principle, this scenario can be associated with its intrinsic structural stability and more importantly, distinct interfacial characteristic where an infinitely large nearly invariable SEI area on the planar surface and a very limited variable SEI area on the lateral surface are involved upon charge/discharge cycling since preferential lithiation of Si at its free surfaces and consequently volume expansion perpendicular to the planar surface.38 Given the disassembly of a Si thin film into numerous two-dimensional (2D) square-shaped Si nanoplate units with a side length of a and a thickness of t and/or
with a material characteristic size smaller than the critical value36) can decently survive without mechanical fracture during lithiation and delithiation, a prolonged cycling has inevitably necessitated an elaborate combination of these Si units with a second phase (e.g., carbon) as well as the subsequent assembly, so as to stabilize the solid electrolyte interphase (SEI) on the surface of individual Si building blocks and also strengthen the interfacial electrical contacts between Si building blocks. Toward the structural and interfacial stabilization, a variety of sophisticated material design/ preparation/combination concepts have thus been extensively developed, resulting in significant improvements in some properties of silicon anodes. Despite these impressive advances, there remain two critical challenges of Si anodes to overcome. First, the preparation, combination and assembly of the existing Si building blocks always reply heavily on costly Si, extremely hazardous silicon precursors, complicated equipment, and/or complex and time-consuming procedures, which have severely impeded practical implementation.13 Second, based upon the existing nanostructured Si building blocks, the resulted silicon anodes are still far from meeting the requirements for 7477
DOI: 10.1021/acsnano.7b03942 ACS Nano 2017, 11, 7476−7484
Article
ACS Nano
Figure 2. Morphological, structural, and componential characterization of SF. (a, b) SEM images. (c, d) TEM images. (e) Photograph of a hydrangea flower. (f−j) Magnified TEM images. (k) High-resolution TEM image. (l) SAED pattern. (m, n) Scanning transmission electron microscopy (STEM) and elemental mapping images. (o) Raman spectrum of SF superimposed with that of silicon nanoparticle. (p) XRD pattern comparing with that of silica fume. (q) Si 1s XPS spectrum comparing with that of silica fume. (r) Nitrogen adsorption−desorption isotherms. (s) Pore size distribution.
ment of such rolled-up tube-like 2D Si still requires the presence of carbon in both sides of Si nanomembranes. The magnesiothermic reduction of silicon dioxide (SiO2), usually at temperatures near the melting point (650 °C) of magnesium (Mg), is recently considered as an attractive, cost-effective, and scalable route to synthesize nanostructured silicon including 2D Si nanostructures.48−52 Yet, regardless of the use of a template, the as-synthesized 2D Si nanostructures are intrinsically composed of particulate silicon, which rather incur their ineluctable dependence on a second phase when being used as anode materials. In this communication, we develop an intrinsically dual stabilized Si building block (namely silicene flowers (SF)), which comprises inherently interconnected silicene nanoplates with different spatial orientations thereby forming a threedimensional flower-like ensemble. This building block holds several advantages (Figure 1c). First, every silicene nanoplates mimick the behavior of a planar Si thin film in respective spatial orientations upon cycling, which allows for better accommodation of the volume change of Si ensuring the structural stability and more importantly, leads to the formation of limited variable SEI area improving the interfacial stability. Second, silicene nanoplates possess shortened lithium ion diffusion length and electron transport paths at the building block scale, and the inherently interconnected nature of every silicene nanoplates further renders the creation of three-dimensional robust electron and lithium ion transport channels between the building blocks throughout the whole SF. Third, the flower-like configuration of silicene nanoplates endows the material with high tap density, which is critical for practical applications.
their subunits (Figure 1a), individual nominally disassembled Si nanoplates can be envisioned to possess similar structural and interfacial features. While the variable SEI surface area per unit mass of these 2D Si nanoplates (defined as specific variable surface area (Ssv)) can be estimated to be constant at different a/t ratios, it is found that the specific variable SEI surface areas of 0D, 1D, and 3D Si nanostructures are always larger than that of 2D Si nanoplates as well as this difference increases with increasing a/t ratios (Figure 1b, Calculations). From this viewpoint, while the utilizations of 0D, 1D, and 3D Si nanostructures unavoidably require additional interfacial stabilization in view of their large specific variable SEI surface areas, 2D Si nanostructures such as planar Si nanoplates may be an intriguing and viable class of building blocks for Si anodes since their limited variable SEI surface area. Yet, such 2D nanostructures are not available commercially. Depositing silicon and/or silicon precursors on specific surfaces under stringent conditions can create silicene39−42 (the silicon analogue of graphene with a buckled honeycomb structure) and also other silicon nanosheets.43 Although these layered silicon materials are theoretically assessed possessing high lithium storage capability compared with other layered analogues such as phosphorene and group IV monochalcogenides,44−46 there are few studies devoted to their demonstration as 2D building blocks for Si anodes mostly because their production is not a trivial task. Peeling off radio frequency-sputtered carbonsandwiched Si thin films on a sacrificial layer to obtain rolledup Si nanomembrane-containing pieces is a direct way to produce 2D Si building blocks.47 Yet, the scalability of the thinfilm peeling process is limited, and the performance enhance7478
DOI: 10.1021/acsnano.7b03942 ACS Nano 2017, 11, 7476−7484
Article
ACS Nano
Figure 3. Electrochemical characterization of SF anodes. (a) Typical CV curve of a SF electrode. (b) Galvanostatic charge/discharge voltage profiles of the SF and Si nanoparticle electrodes at 0.05 C (1 C = 4000 mA g−1). (c) Voltage profiles of the SF electrodes at different rates. (d) Gravimetric capacity versus cycle plots at different current rates. (e) Coulombic efficiency versus cycle plots of the SF electrodes at different rates, as well as the Si nanoparticle and Si nanowire electrodes at 0.5 C. (f) Volumetric capacity of the SF electrodes at different rates, comparing with some representative Si anodes reported in the literatures as noted. (g) Cycling performance of the SF and control electrodes over 600 cycles at 0.5 C (2000 mA g−1). The areal mass loading of SF, Si microparticle, Si nanoparticle, and Si nanowire electrodes is 1.5, 0.9, 0.5, and 0.21 mg cm−2. The theoretical capacity of a graphite electrode is shown as a green dashed line. The Coulombic efficiency of SF is plotted on the secondary y-axis. All the capacity values reported are on the basis of active materials.
or higher) for the preparation of SF (Figure 1c,f, Figure S1,2). The morphological, structural and componential features of the SF produced at 850 °C are summarized in Figure 2. It is clearly observed that the SF has a flower-like architecture with diameters ranging from ∼1 to 10 μm and each SF is composed of numerous nanoplates with different geometries and spatial orientations (Figure 2a,b,c,d), resembling hydrangea flowers in nature (Figure 2e). The magnified transmission electron microscope (TEM) images (Figure 2f,g,h,i,j) further reveal that the nanoplates are inherently interconnected in an either overlapping or bridging manner with each other, thereby forming interpenetrating pores (around or smaller than 20 nm) throughout the whole SF. The high-resolution TEM and elemental mapping images, selected area electron diffraction (SAED) and powder X-ray diffraction (XRD) patterns, as well as X-ray photoelectron spectroscopy (XPS) analyses (Figure 2k,l,m,n,p,q) all verify that the obtained nanoplates are comprised of crystalline silicon, while the similarity to those of sp3 crystalline silicon can be associated with the multilayer nature of silicene nanoplates and more importantly, the presence of strong interlayer covalent bonding between the silicene layers.54 It is worth noting that, different from those of full sp3 crystalline Si (e.g., Si nanoparticles and Si nanowires) and silicon microparticle with one nearly symmetric peak at 515−520 cm−1 (Figure S3), the Raman spectra of SF are always
When being directly used as anode materials, the SF with high mass loadings exhibits extraordinary combined performance, including high gravimetric capacity (2000 mAh g−1 at 800 mA g−1), high volumetric capacity (1799 mAh cm−3 that is more than three times that of graphite anodes), remarkable rate capability (950 mAh g−1 at 8 A g−1), and excellent cycling stability (1100 mA h g−1 at 2000 mA g−1 over 600 cycles). As a proof-of-concept demonstration, a full cell (LCO/SF) using the SF anode and a conventional lithium cobalt oxide (LCO) cathode has successfully delivered high specific energy (543 Wh kgca−1) and energy density (1257 Wh Lca−1) based on both the cathode and anode, which are far beyond those of its commercial counterparts using graphite electrodes (357 Wh kgca−1 and 526 Wh Lca−1). We anticipate that this Si building block will offers a horizon for the rational design and construction of high-capacity but large-volume-change energy storage electrode materials toward a viable and high-performance lithium battery.
RESULTS AND DISCUSSION We have employed silica fume,53 an industrial byproduct in producing silicon and ferrosilicon alloys having a very high content (typically, above 90 wt %) of SiO2 nanoparticles (Figure 1d,e), as the starting material, and harnessed the magnesiothermic reduction at elevated temperatures (850 °C 7479
DOI: 10.1021/acsnano.7b03942 ACS Nano 2017, 11, 7476−7484
Article
ACS Nano
Figure 4. Electrochemical performance of full cells using the SF anodes and conventional LCO cathodes. (a) Schematic of a LCO/SF full cell configuration with its LCO/graphite counterpart. (b) Typical voltage profiles of the LCO/SF full cell with a LCO/graphite full cell as well as a LCO/Li half cell. (c) Reversible capacity and Coulombic efficiency versus cycle plots of the LCO/SF full cell and a LCO/Li half cell. (d) Rate capability of LCO/SF and LCO/graphite full cells. (e) Comparison of specific energy and energy density of the LCO/SF and LCO/graphite full cells on the basis of both the cathode and anode. (f) A LCO/SF pouch cell-powered “NCNST” pattern consisting of five different colors of light emitting diodes (LEDs) (the initials of National Center for Nanoscience and Technology).
characterized by an intense peak (E2g mode) located at 515 cm−1 reflecting the honeycomb nature of the 2D Si nanoplate lattice, along with a broad shoulder (A1g mode) at a lower frequency (Figure 2o, Figure S4) indicating the mixed sp2−sp3 feature of the honeycomb Si lattice. This observation is in good agreement with the case of silicene obtained by other methods,39 implying the successful production of silicene nanoplates based on our approach. Atomic force microscope (AFM) further reveals that the silicene nanoplates possess thicknesses between 3.6−9.6 nm presenting an average value of 7.2 nm, thus disclosing their multilayer nature (Figure S5). Brunauer−Emmett−Teller (BET) nitrogen adsorption−desorption measurements (Figure 2r) show the specific surface area of SF to be ∼44 m2 g−1; this value is lower than those estimated theoretically for square-shaped Si nanoplates (e.g., 100 nm in side length and 3.6−9.6 nm in thickness), which can be ascribed to the possible stacking between some silicene nanoplates as observed in TEM images (Figure 2i, Figure S6). Furthermore, Barrett−Joyner−Halenda (BJH) analysis (Figure 2s) indicates the presence of significant pores less than 20 nm in size, consistent with the above TEM observation. All the above evidenced thin, inherently interconnecting, 3D spatially oriented and flower-like features of our silicene nanoplates are
crucial for their use as the next-generation building block of Si anodes as discussed below. To investigate the electrochemical properties of the SF and also control samples as LIB anode materials, unless otherwise specified (Figure S7), the electrodes were prepared by casting slurry containing active material, conductive additive, and sodium alginate binder at a mass ratio of 8:1:1, and coin-type half cells (CR2032) fabricated by using lithium foil as the counter electrode, with the results shown in Figure 3. It can be observed that the SF shows similar cyclic voltammetry (CV) peaks (Figure 3a) and galvanostatic voltage profiles (Figure 3b) to those of other Si building blocks.15,25 The Coulombic efficiency of the first cycle is 74% for our SF because the initial SEI formation consumes some lithium, although this disadvantage can be diminished by prelithiation.30,50 In Figure 3c and d, we prove that the SF demonstrates high gravimetric capacity and capacity retention at charge/discharge rates ranging from 400 to 8000 mA g−1. For example, the average reversible capacity of the SF over 150 cycles reaches 2000 mAh g−1 at a rate of 800 mA g−1; even at a much higher rate of 8000 mA g−1, the gravimetric capacity is still above 950 mAh g−1 and remains highly stable over 150 cycles, clearly reflecting the structural effectiveness and stability of SF. As exhibited in 7480
DOI: 10.1021/acsnano.7b03942 ACS Nano 2017, 11, 7476−7484
Article
ACS Nano
gLCO−1 and high Coulombic efficiency over only initial 3 cycles (Figure 4c), comparable to the case of LCO/Li half cells. Capacity retention tests at various current rates from 0.2 to 3 C evidently reveal the superior power capability of the LCO/SF full cell relative to the LCO/graphite full cell. More importantly, the specific energy and energy density of the LCO/SF full cell are estimated to be 543 Wh kgca−1 and 1257 Wh Lca−1 when considering both the cathode and anode, which are 52% and 139% increase relative to those of the commercial LCO/graphite full cell (357 Wh kgca−1 and 526 Wh Lca−1), respectively, stamping great potential of the SF for nextgeneration battery applications. In addition, a “NCNST” pattern consisting of five different colors of light emitting diodes (LEDs) has been successfully powered using an aspaired LCO/SF pouch cell device (Figure 4f). In general, the magnesiothemic reduction of SiO2 was carried out routinely at 650−700 °C under atmospheric pressure,18,24,49−52 which results in nanostructured particulate silicon, along with a significant amount of magnesium silicide (Mg2Si) and unreacted SiO2 as byproducts, at a low production yield. Some previous studies48,55 showed both lowering the reaction pressure and/or raising the reaction temperature could accelerate the reaction kinetics and thus improve the yield, yet without substantially tailoring the particulate morphology and structure. In our case, the magnesiothemic reduction of silica fume at elevated temperatures (850 °C and higher) has successfully resulted in the formation of silicene flowers at a high production yield (Figure S1, S9). Although the detailed formation mechanism must be investigated in the future study, this can be mainly associated with the synergistic effect of the relatively small size of SiO2 nanoparticles in silica fume and the elevated reaction temperatures. First, a relatively small size of SiO2 particles (e.g., ∼150 nm in the case of silica fume) allows for the simultaneous formation of considerable amounts of neighboring silicon atoms as the component of silicene nanoplates, due to their easily accessible surface and shortened solid-state diffusion length for Mg. This particle size effect has further been verified by using other sizes of SiO2 particles. An excessively small size of SiO2 particles (e.g., 20 nm in diameter) cannot ensure the formation of sufficient amounts of neighboring silicon atoms for silicene nanoplates, and a larger size of SiO2 particles (e.g., 1.5 μm) significantly retards the diffusion of Mg and suppresses the simultaneous formation of neighboring silicon atoms, both of which result in particulate silicon even if at an elevated temperature of 850 °C (Figure S10, S11). Second, the elevated temperatures enhance the reaction kinetics by accelerating the vapor-phase and solid-state diffusivity of Mg, which not only improves the yield significantly, and also greatly facilitates the simultaneous formation and fusion of neighboring silicon atoms toward silicene nanoplates and their ensemble. Otherwise, the magnesiothemic reduction of silica fume at temperatures below 850 °C leads to the formation of particulate silicon as the majority (Figure S1, S4). Furthermore, different from those obtained at lower temperatures, the negligible presence of Mg2Si and SiO2 in the unwashed samples obtained at elevated temperatures further validates the high-yield conversion of SiO2 in silica fume to Si in silicene flowers (Figure S12). The produced SF possesses the distinct structure, interface, and architecture characteristics, permitting the exceptional lithium storage performance achieved. Although being transformed into amorphous silicon upon cycling (Figure S13), the SF nanoplates remain intact even after cycling as disclosed by
Figure 3e, the SF impressively shows a rapid increase of stabilized Coulombic efficiency greater than 99% after only 5 cycles, depicting the interfacial stabilization. By comparison, the Si nanoparticles require more than 100 cycles to reach such a level of Coulombic efficiency and the Si nanowires cannot approach 99% Coulombic efficiency even after 150 cycles. This significantly enhanced Coulombic efficiency of SF confirms that flower-like silicene nanoplates possess limited variable SEI area upon cycling as discussed above (Figure 1b), which fundamentally restrains the irreversible consumption of lithium and confers the interfacial stabilization. Benefiting from the high gravimetric capacity achieved and the high tap/packing density of the material, the volumetric capacity of SF anodes, a critical parameter dominating the material deployment, is shown to be extraordinarily high. Specifically, considering the electrode volume, the SF delivers a volumetric capacity of 1799 mAh cm−3 at the rate of 400 mA g−1, which is more than three times that (∼500 mAh cm−3) of commercial graphite anodes. Even at high rates of 800, 2000, 4000, and 8000 mA g−1, the volumetric capacities of 1578, 1267, 887, and 728 mAh cm−3 are delivered, respectively, which are among or even surpass the best of the reported Si anodes (Figure 3f).12,17−19,22,24−26,28,31 Moreover, the long-term cycling performance of SF at a rate of 2000 mA g−1 was further examined and compared with other Si building blocks (Figure 3g). Either commercially available silicon microparticles and nanoparticles or emerging silicon nanowires show different degrees of capacity fading over the measured cycles. In contrast, the SF, even at a much higher mass loading level, exhibits excellent cycling stability and almost invariably achieves a gravimetric capacity of about 1100 mAh g−1 over 600 cycles, three times that of commercial graphite anodes. This is also the case in the SF obtained at temperatures of 950 and 1050 °C (Figure S8). The high capacity, high rate capability, and stable cycling of SF at high mass loadings, without the need of complex design/preparation/combination/ assembly engineering, can be mainly attributed to two synergistic effects mainly involving the material shape and interconnectivity. First, as other shapes of silicon (e.g., Si nanoparticles, Si nanowires) feature relatively large specific variable SEI surface areas upon cycling (Figure 1b) and require additional interfacial stabilization, the structural distinction of thin silicene nanoplates in the SF allows accommodating the volume change of Si specifically without introducing substantial variable SEI area, thus promoting the simultaneous stabilization of the material structure and interface. Second, different from separated Si nanoparticles and/or Si nanowires, the inherently interconnecting characteristic of 3D spatially orientated silicene nanoplates enables the creation of robust and efficient transport channels for both electrons and lithium ions from/to each silicene nanoplate of the SF. Considering the advantages identified above of the SF, a prototype full-cell device was built to demonstrate the commercial viability of a high-energy-density LIB system based on the SF anode and a conventional commercialized lithium cobalt oxide (LCO) cathode. The fabricated LCO/SF full cell was compared with the LCO/graphite full cell in the same configuration (Figure 4a) over the potential range from 2.7 to 4.35 V. Figure 4b plots the voltage profiles of the LCO/ SF and LCO/graphite full cells as well as a LCO/Li half-cell, showing the charge and discharge processes with average voltages of 3.70, 3.83, and 3.92 V, respectively. At a current rate of 0.5 C (1 C = 274 mA gLCO−1), the fabricated LCO/SF full cell shows stable cycling at a gravimetric capacity of ∼150 mAh 7481
DOI: 10.1021/acsnano.7b03942 ACS Nano 2017, 11, 7476−7484
Article
ACS Nano
(FEI Tecnai G2 20 STWIN and Tecnai G2 F20 U-TWIN). Raman spectra were collected using a Renishaw inVia Raman microscope with a laser wavelength of 514.5 nm. X-ray photoelectron spectroscopy (XPS) measurements were performed on an ESCALAB250Xi apparatus with an Al Kα X-ray source. The X-ray diffraction (XRD) instrument type was D/MAX-TTRIII (CBO). Nitrogen adsorption/ desorption isotherms were measured at 77 K with an ASAP 2020 physisorption analyzer. The Brunauer−Emmett−Teller (BET) method and Barrett−Joyner−Halenda (BJH) model were utilized to estimate the specific surface area and pore size distribution, respectively. Electrochemical Characterization. Working electrodes were fabricated by casting slurry on copper (Cu) current collectors with active materials (silicene flowers), conductive additive (Super P, Alfa Aesar), and sodium alginate binder (Alfa Aesar) at a desirable mass ratio. All the electrodes were degassed in vacuum at 60 °C for at least 2 h before use. The typical mass loading for silicene flowers electrodes is 0.9−1.5 mg cm−2. 2032 coin type half cells were assembled in an argon-filled glovebox (
