ARTICLES PUBLISHED ONLINE: 14 MARCH 2010 | DOI: 10.1038/NMAT2725
High-performance lithium-ion anodes using a hierarchical bottom-up approach A. Magasinki1 , P. Dixon1 , B. Hertzberg1 , A. Kvit2 , J. Ayala3 and G. Yushin1,4 * Si-based Li-ion battery anodes have recently received great attention, as they offer specific capacity an order of magnitude beyond that of conventional graphite. The applications of this transformative technology require synthesis routes capable of producing safe and easy-to-handle anode particles with low volume changes and stable performance during battery operation. Herein, we report a large-scale hierarchical bottom-up assembly route for the formation of Si on the nanoscale—containing rigid and robust spheres with irregular channels for rapid access of Li ions into the particle bulk. Large Si volume changes on Li insertion and extraction are accommodated by the particle’s internal porosity. Reversible capacities over five times higher than that of the state-of-the-art anodes (1,950 mA h g−1 ) and stable performance are attained. The synthesis process is simple, low-cost, safe and broadly applicable, providing new avenues for the rational engineering of electrode materials with enhanced conductivity and power.
T
he development of lightweight, long-lasting Li-ion batteries is of great technological importance for critical applications, such as low- or zero-emission hybrid electrical and electrical vehicles, energy-efficient cargo ships and locomotives, aerospace and power-grid applications1–4 . Increasing the specific capacity of Li-ion battery anodes is considered an attractive route to lower battery weight, volume and ideally cost5–8 . The low discharge potential and high theoretical capacity of silicon have triggered significant research efforts on Si-based anodes9 . One of the challenges associated with the use of high-capacity Si anodes is how to increase capacity retention with battery cycling9,10 . In Li-ion battery anodes, particles of the active material are held together using a polymeric binder. High-capacity Si anodes show large volume changes during Li insertion and extraction, leading to electrode failure10 . Lithiated Si anode particles contract, losing electrical contact with each other and the current collector on Li extraction, which leads to low power capability and rapid loss of capacity. In recent efforts11–15 , improvements of the binder properties have been proposed to improve Si anode performance. As an alternative approach, the use of binder-free electrodes consisting of Si nanowires grown directly on conductive steel substrates has recently been proposed5 . The small diameter of these nanowires (10–200 nm) allows for rapid transport of Li ions and improves the cycling stability of the anodes5,16,17 . However, the use of nanowires in the near term is unlikely, because of their prohibitively high synthesis costs for industry as well as the present technological difficulties, such as the weak adhesion of wires to the substrate and the lack of sufficient control over the pore volume, wire length, diameter and uniformity. There are also challenges involved in assembling the nanowire electrodes into cylindrical cells, in minimizing the presence of impurities and catalyst particles in nanowire electrodes and, most importantly, in the health hazards of nanowires18,19 . The use of Si–C composites to circumvent the limitations of pure Si power has been investigated for many years20–28 . A comprehensive summary of these efforts is available29 . Conventional composites,
commonly prepared by pyrolysis20,27 , mechanical mixing and milling21–24 , or some combination of these two25,26 , consist of Si particles embedded in a dense carbon matrix. However, the large volume changes in Si on Li insertion can be accommodated by carbon only to a limited degree, thus offering only limited stability and capacity enhancements29 . Porous Si–C composite particles may be a viable solution to overcome the limitations of non-porous Si–C materials. Preexisting pores will provide the volume needed for Si expansion and allow for fast transport of Li ions, and C will allow the improved solid/electrolyte interface formation30 , structural integrity31 and high electrical conductivity31 . If properly engineered, the porous Si–C composite particles can conserve size and shape on cycling, which is critical for industrial applications because commercial battery cells have very little volume available for anode expansion. However, the low-cost synthesis of such composites is not a trivial task. Although large composite particle sizes are attractive for industrial applications because of their higher electrical conductivity and safety, the uniform deposition of Si nanoparticles inside the narrow pores of large C particles is challenging, because of the limited precursor diffusion paths inside the narrow channels. If the particle size of Si is limited to 10–30 nm for the improved power characteristics and stability32 , the porous, low-cost carbon substrate powder should have interconnected pores in the range of ∼34–102 nm in diameter to accommodate the ∼70% changes in diameter of the Si particles. Such carbons are not available commercially. The use of templated carbons with uniform channels of the desired size for infiltration of active material33 may be attractive as a proof-of-concept for performance enhancements, but it greatly increases the production cost and thus is not practical. The bottom-up assembly offers an attractive route for low-cost synthesis of composite (nano)materials34,35 . In nature, self-assembly allows composite systems to attain new functionality unavailable in individual components36 . It is known that small nanoparticles tend to form either ordered or disordered agglomerates if the interactive potential exceeds the energy of the thermal vibrations35 .
1 School
of Materials Science and Engineering, Georgia Institute of Technology, Atlanta, Georgia 30326, USA, 2 Materials Science Center & Materials Science Department, University of Wisconsin-Madison, Madison, Wisconsin 53706, USA, 3 Superior Graphite, Chicago, Illinois 60606, USA, 4 Streamline Nanotechnologies Inc., Atlanta, Georgia 30326, USA. *e-mail:
[email protected].
NATURE MATERIALS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturematerials
© 2010 Macmillan Publishers Limited. All rights reserved.
1
NATURE MATERIALS DOI: 10.1038/NMAT2725
ARTICLES a
b
c
Annealed carbon black Si CVD
C CVD
Si nanoparticles’ assembly on the surface of annealed carbon black
Assembly of Si-coated carbon black particles into rigid spherical granules
Figure 1 | Schematic of Si–C nanocomposite granule formation through hierarchical bottom-up assembly. a–c, Annealed carbon-black dendritic particles (a) are coated by Si nanoparticles (b) and then assembled into rigid spheres with open interconnected internal channels during C deposition (c).
Larger particles or nanoparticle assemblies can then be formed into spherical granules of the desired size through the application of a sacrificial liquid binder, which forms permanent bonds on drying or sintering37 . This low-cost process, known as granulation or balling, is often used in the pharmaceutical and food industries37 , but it remains uncommon for energy-storage applications. Here we harness the bottom-up assembly to develop Si–C porous composite powders with high capacity, stable performance and particle size comparable to that of milled graphite (15–30 µm). This powder size is commonly used in the production of Li-ion battery electrodes and does not possess the same inhalation hazard as nanoparticles do18 . We demonstrate that our bottom-up approach overcomes existing challenges and allows for the fabrication of high-capacity stable composite anodes with rapid charging capability. We anticipate that our strategy will provide avenues for the rational design of low-cost nanoparticles containing functional composites. We have designed a chemical vapour deposition (CVD) synthesis process for Si nanoparticles on the surface of carbonblack nanoparticles, branched into short chains during hightemperature pre-annealing (Fig. 1a). The 0.5–1 µm multibranched nanocomposite (Fig. 1b) is then self-assembled into large porous spherical granules during the atmospheric pressure CVD deposition of C (Fig. 1c). The Si CVD deposition time as well as the pressure and temperature in the deposition system38 determines the size of the deposited Si nanoparticles. The size of the branches in the nanocarbon and the size of the deposited Si nanoparticles determine the pore size in the synthesized composite. The diameter of the composite particles is influenced by the carbon CVD process parameters and by the size of the initial branched carbon particles. Therefore, the process allows control over the particle size, pore size and composition of the composite. We have selected annealed and chained carbon-black particles as substrates for Si sphere assembly because of their open structure, very low apparent density and high specific surface area (SSA; ∼80 m2 g−1 ), providing multiple accessible sites for Si deposition. The ultralow cost of carbon black (∼10–20% of purified natural graphite) and large production volume (∼9 times higher than that of natural graphite) help maintain the low cost of the synthesized composite. Impurities in carbon electrodes are detrimental to battery operation39 , contributing to parasitic side reactions, gassing in the cells, self-discharge and degradation of the shelf life of the cells. The annealing of carbon black at temperatures above ∼2,000 ◦ C results in graphitization, linkage of neighbouring particles and, most importantly, a very high degree of purification (>99.9%), promoting consistent properties40 , which are required for Li-ion battery systems and are unattainable by means of chemical purification with acids39 . Such effects are similar to what is observed during high-temperature annealing of other carbon nanoparticles, such as carbon nanotubes and nanodiamond41 . 2
The deposition of Si nanoparticles was conducted by means of low-pressure (∼1 torr) silane (SiH4 ) decomposition at 500 ◦ C for 1 h. Transmission electron microscopy (TEM) revealed the nanoparticles to possess a spherical shape 10–30 nm in diameter, having been deposited on the surface of the annealed carbon black (Fig. 2a), as shown in Fig. 1a. Most nanoparticles showed amorphous microstructure or were highly disordered (Fig. 2b,c). The nanoparticles densely coated the carbon surface, frequently attached to the edges of the graphitic structures. Once a stable nucleus is formed, growth can occur through adsorption of gas species on the nucleus surface38 . The spherical shape of the particles minimized the contact area between the Si and carbon-black surfaces, probably because of the high interfacial energy between Si and the flat faces of the graphitized carbonblack particles, which possess a negligible concentration of surface functionalities. The low synthesis temperature may have minimized the surface mobility of Si atoms and contributed to the smooth morphology of the Si surface. No impurities were detected in the sample by energy-dispersive X-ray spectroscopy (EDX; Fig. 2d). In wet granulation, a liquid binder wets small primary particles as they are agitated, causing them to self-assemble into the larger spheres by a combination of viscous and capillary forces37 . The drying or annealing process transforms the binder into a dense solid that preserves the shape of the granules37 . For electrode particles, this solid should ideally have high electrical conductivity, high mechanical stability and high permeability to Li ions. Graphitic carbon shows a unique combination of these attributes. To prevent the oxidation of Si nanoparticles, we selected a hydrocarbon as a carbon-precursor binder for granulation. In a conventional wet granulation process, a liquid binder is typically introduced as droplets. It penetrates into the pores of the powder surface, forming initial nuclei, which grow over time. If the droplet size is relatively small, the nucleation will occur by distribution of the drops on the surface of the particles and subsequent particle coalescence37 . The process is similar to melt agglomeration42 , where the binder melts and the melt-coated particles stick together to form granules. However, it is commonly difficult to achieve the uniform binder deposition required for the controlled and uniform formation of granules37 . Therefore, in contrast to the previous studies we introduced a binder in a gaseous form. Simultaneous granulation and carbon deposition processes were carried out by means of decomposition of propylene (C3 H6 ) at 700 ◦ C for 30 min. This higher temperature step causes significant crystallization in the deposited Si nanoparticles. X-ray diffraction (XRD) analysis of the produced samples showed the average grain size of Si nanoparticles to be ∼30 nm (Fig. 2e). TEM studies confirmed the crystalline structure of Si nanoparticles after exposure to 700 ◦ C (Fig. 2f).
NATURE MATERIALS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturematerials © 2010 Macmillan Publishers Limited. All rights reserved.
NATURE MATERIALS DOI: 10.1038/NMAT2725 a
ARTICLES
b
c Si nanoparticles
Amorphous Si
C
3.34 Å
C
200 nm
50 nm
d
e Intensity (a.u.)
Intensity (a.u.)
1
3.13 Å Si
C
Si After
C
O Cu 0
C
f
Si
Si
C
10 nm
Crystalline Si nanoparticle
Before
2 4 3 X-ray energy (eV)
5
30
20
40 2θ (°)
50
60
10 nm
Figure 2 | Structure of the C–Si nanocomposite synthesized through Si CVD on the annealed carbon black. a–c, TEM micrographs recorded at different magnifications. The black arrows in a point to spherical amorphous Si nanoparticles, and the white arrows point to the edges of the graphitized carbon-black backbone chain. The size of the inset in a is 800 nm × 800 nm. The high-resolution TEM micrograph (c) shows the highly ordered graphitic structure of the carbon-black surface with a (002) interplanar spacing of 3.34 Å and the amorphous structure of Si. d, EDX spectrum of the composite showing the C and Si Kα lines, O and Cu sample holder lines. e, XRD spectra of Si-coated carbon black before and after C deposition at 700 ◦ C for 30 min. f, High-resolution TEM micrograph of a Si nanoparticle crystallized after the exposure to 700 ◦ C. a
b
cc
1 µm
10 µm
100 µm
Experimental 20
30
40
Diameter (µm)
50
80 70 60 50 40 30 20 10 0
Before After 0
0.2 0.4 0.6 0.8 P/P0
h Surface area (m2 g¬1)
Boltzmann fit
100 nm
g N2 adsorbed (cm3 g¬1)
f Number of granules (a.u.)
e
10
d
1.0
35 30 25 20 15 10 5 0
Before
After 10
100 Pore width (Å)
1,000
100 nm
Figure 3 | Structure of the C–Si nanocomposite spherical granules self-assembled during C CVD on the Si-decorated annealed carbon black. a–d, SEM micrographs recorded at different magnifications. The white arrows in d point to C-coated Si nanoparticles visible on the surface of the granules. e, Cumulative size distribution of spherical granules synthesized at 700 ◦ C. f, N2 sorption isotherms on the surface Si-decorated annealed carbon black before and after C CVD. g, Barrett–Joyner–Halenda cumulative SSA of Si-decorated annealed carbon black before and after C CVD. h, High-magnification SEM micrograph of the surface of spherical granules produced during C coating of pure carbon black shown for comparison with d.
Scanning electron microscopy (SEM) micrographs show the spherical granules formed in the course of C deposition (Fig. 3a,b). The particle surface is rough, with surface asperities of ∼0.5–1 µm (Fig. 3b,c). In spite of C coating, small Si nanoparticles are visible on the surface (Fig. 3d). The diameter of the produced spheres ranged from ∼15 to 35 µm and showed a narrow particle size distribution with an average diameter of ∼26 µm (Fig. 3e). The sphere size distribution is controlled by the granulation process conditions37 (Supplementary Fig. S1) and can be optimized for specific
applications. Propylene decomposition takes place through several intermediate steps43 . The hydrocarbon products of intermediate steps of C3 H6 decomposition are known to form larger-molecularweight compounds, including toluene, ethylbenzene, styrene, naphthalene, biphenyl and others, which adsorb on the surface of the substrates during the CVD reaction43 . In the adsorbed state, they act as a liquid agglomeration binder before their final transformation into carbon. In our initial experiments, we introduced vibration to the sample tube to agitate the nanoparticles.
NATURE MATERIALS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturematerials
© 2010 Macmillan Publishers Limited. All rights reserved.
3
NATURE MATERIALS DOI: 10.1038/NMAT2725
1C
C/20
2,000
95
C¬Si granules
1,600
90
1,200 800 Theoretical for graphite
400 0
10
20
30
40 50 60 Cycle number
c
70
80
dQ/dV (mA h g¬1 V¬1)
d 20,000 15,000 10,000 5,000 0 ¬5,000 ¬10,000 ¬15,000 ¬20,000
90
80 100
2.0 1.5 1.0 0.5 0 ¬0.5
2.0 1.5 1.0 0.5 0 ¬0.5
0.49 V
Li insertion (alloy)
0.21 V 0.06 V 1 cycle 0.5
10 µm
1.0
Voltage (V)
1C C/20
8C
1C C/20
8C 0
500
1,000
1,500
2,000
Annealed CB 8C 1C C/20
2.0 1.5 1.0 0.5 0 ¬0.5
C/20
8C 1C 50 100 150 Capacity (mA h g¬1)
0
Li extraction (dealloy)
2¬5 cycles
0
C¬Si granules
Capacity (mA h g¬1)
Voltage (V)
0
85
b Voltage (V)
100
Voltage (V)
2,800 2,400
Coulombic efficiency (%)
a
Reversible deintercalation capacity (mA h g¬1)
ARTICLES
Graphite
8C
8C 0
200
1C C/20
1C C/20
100 200 300 Capacity (mA h g¬1)
Figure 4 | Electrochemical performance of the C–Si bottom-up-assembled nanocomposite spherical granules. a, Reversible Li deintercalation capacity and Coulombic efficiency of the C–Si granule electrode versus cycle number in comparison with the theoretical capacity of graphite. b, Galvanostatic charge–discharge profiles of the C–Si granule electrode at rates of ∼C/20, 1C and 8C in comparison with that of the annealed carbon-black and commercial graphite-based electrodes between 0 and 1.1 V. c, Differential capacity curves of the C–Si granule electrode in the potential window of 0–1.1 V collected at a rate of 0.025 mV s−1 . All electrochemical measurements (a–c) were carried out at room temperature in two-electrode 2016 coin-type half-cells. d, SEM micrograph of a C–Si nanocomposite granule after electrochemical cycling.
However, further experiments proved that because of the very low density of our powder, it was unnecessary. All of the granules reported here (Fig. 3) were synthesized without artificial agitation in a simple horizontal tube furnace. The granulation bottom-up assembly is known to preserve most of the surface area of the primary particles37 . Indeed, N2 gas sorption measurements (Fig. 3f) showed that the decrease of the SSA after C deposition was rather modest—the Brunauer–Emmett–Teller SSA of the Si–C self-assembled granules was ∼24 m2 g−1 , which is close to that of the Si-decorated carbon black (∼33 m2 g−1 ). The pore size distribution of the spherical particles shows the presence of 30–100-nm pores (Fig. 3g). These pores are also visible on the SEM micrographs (Fig. 3c,d) and are available to accommodate the expansion of Si during the Si–Li electrochemical alloy formation. Carbon coating of the surface of the pure carbon black also results in the formation of porous granules with larger surface features and no 10–30-nm spherical particles visible (Fig. 3h). Coin cells (2016) with a metallic Li counter electrode were used to evaluate the electrochemical performance of materials. The specific reversible deintercalation capacity of the sample with an estimated ∼50 wt% of Si reached ∼1,950 mA h g−1 at C/20 (Fig. 4a). This gravimetric capacity is over five times higher than that of the theoretical capacity of graphite, six times that of high-performance graphitic anodes and is nearly 18 times that of the annealed carbon black (Fig. 4a,b and Supplementary Fig. S2). The specific capacity of the Si nanoparticles alone was estimated as ∼3,670 mA h g−1 at C/20, which is the highest value ever reported for nanoparticles. It approaches the theoretical capacity of Si (∼4,200 mA h g−1 if Li22 Si5 is achieved). Such a high specific capacity value indicates high accessibility of the active Si for Li insertion in the designed composite architecture (Figs 1–3). The overall carbon contribution was estimated as ∼230 mA h g−1 (115 mA h/0.5 g). The volumetric capacity was determined to be 1,270 mA h cm−3 at C/20, which is higher than ∼620 mA h cm−3 for graphitic anodes. However, further optimization of the electrode porosity and the formation 4
of bulk electrodes (Supplementary Fig. S3) may further increase the volumetric capacity values. The irreversible capacity losses in the first cycle (Fig. 4a) are related to the solid–electrolyte interphase formation and, in contrast to carbon black (Fig. 4b, Supplementary Fig. S2), are rather modest (∼15%) because of the high electrode capacity (Fig. 4a,b). Whereas Si anodes are known to suffer from sluggish kinetics, our self-assembled electrodes demonstrated outstanding high rate capability. The specific capacity of the composite anodes at the fast discharge rates of 1C and 8C was 1,590 and 870 mA h g−1 , respectively, which is ∼82 and 45% of that at C/20 (Fig. 4a,b). Even graphites with high Li diffusion coefficients and low overall capacity could not match such capacity retention at 8C rate (2.98 A g−1 ) and showed a deintercalation capacity of ∼40 mA h g−1 , which is 13% of the C/20 specific capacity (Fig. 4b). For the same specific current value (2.98 A g− ), our C–Si electrodes showed capacity in excess of 1,500 mA h g−1 , which is over 37 times higher. Clearly, in spite of the large particle size (Fig. 3a,b), Li ions can rapidly reach the active anode material within each granule. The differential capacity curves show broad lithiation (Li insertion) peaks at 0.21 and 0.06 V, and a narrower delithiation (Li extraction) peak at 0.5 V (Fig. 4c). The C delithiation peaks commonly observed at 0.2 V are too small to be visible, because of the very small contribution of carbon to the overall anode capacity. A delithiation peak at 0.3 V often reported in both micrometre-scale Si powder15 and Si nanowires5,17 cells was not present. The increase in the 0.5 V peak height after the first cycle indicates improvement in Li extraction kinetics. The formation of an amorphous Si–Li alloy on the insertion of Li into crystalline Si in the first cycle began at 0.1 V, in agreement with previous studies on nanowires5,17 . Subsequent cycles showed an extra lithiation peak at 0.21 V, which corresponds to higher-voltage lithiation of amorphous Si–Li phase15 . The pores available in the composite granules for Si expansion during Li insertion (Fig. 3c,d,f,g) also allowed for efficient and
NATURE MATERIALS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturematerials © 2010 Macmillan Publishers Limited. All rights reserved.
NATURE MATERIALS DOI: 10.1038/NMAT2725 stable anode performance (Fig. 4a). The SEM studies of the anode particles after high-speed mixing, calendaring and cycling demonstrated exceptional robustness of the granules (Fig. 4d). The tap density of the Si–C powder was estimated to be 0.49 g cm−3 , which is lower than that of the graphite (1.3 g cm−3 ), but higher than that of annealed carbon black (0.22 g cm−3 ). The Si nanopowder (10–30 nm) alone is expected to have an even lower tap density. The observed high capacity (Fig. 4a) combined with excellent sample stability and high rate capability is unprecedented in Si–C composite powders. In contrast to many photonic44,45 , electronic46,47 or membrane48 applications, where a high degree of order is typically required, bottom-up-assembled electrode granules may benefit from a disorder in their structure. If the path of Li ions is blocked or impeded in a narrow channel by an expanded Si–Li alloy particle or by an area of unevenly formed solid–electrolyte interphase, the interconnected aperiodic porous network would allow for the redirection of the ion traffic, maintaining rapid charging capability for these granules. Therefore, the disorder in granules (Fig. 1c) may enhance the functionality of the composite anode, as it does in some photonic crystals49 and catalytic structures50 . Thus, the applications of a hierarchical bottom-up assembly method for the rational design of nanocomposite powders offer exceptional potential in energy-storage applications. Whereas nanoparticles or nanowhiskers alone are known to possess inhalation and often explosion risks18,19 , poor flow and handling37 and challenges in metering and control37 , the Si–C nanocomposite granules show greatly improved handling, reduced dustiness, which minimizes losses, increased bulk density and other positive attributes. A similar strategy for hierarchical granule formation could be used to enhance the electrical conductivity and handling of cathode nanoparticles.
Methods Synthesis. Si deposition onto the annealed carbon black (∼0.5 g) was carried out at ∼1 torr in a horizontal tube furnace (inner-tube diameter ∼28 mm) heated to 500 ◦ C. High-purity 5% SiH4 in a He precursor gas mixture (Airgas) was introduced at a flow rate of 50 sccm for 1 h. Carbon deposition was carried out at atmospheric pressure in a horizontal tube furnace (inner-tube diameter ∼20 mm) heated to 700 ◦ C. High-purity C3 H6 precursor gas (99.5%, Airgas) was introduced at a flow rate of 50 sccm for 30 min. A bubbler filled with mineral oil was placed at the exhaust to minimize the backflow of air into the system. Before and after the Si and C deposition experiments, the system was purged with high-purity Ar (99.99%, Airgas) at a flow rate of 50 sccm. The samples were taken out of the furnace at temperatures below 50 ◦ C. Characterization. The wt% of the deposited Si was estimated using a high-precision analytical balance (AND GH-120, A&D Company) as well as by measuring the mass change during C and Si oxidation in air using a thermogravimetric analysis system (Q50, TA Instruments). TEM observations were carried out using a Philips CM200UT microscope (Philips) that has a 0.18 nm point resolution, operates at 200 kV and is equipped with an EDX detector. High-resolution TEM measurements were done on a Titan S/TEM microscope (FEI) equipped with a SuperTwin objective lens with a Cs of 1.2 mm. The microscope has a point resolution of 0.245 nm, an information limit of 0.08 nm and operates at 200 kV. XRD studies were carried out using a Panalytical X’Pert PRO Alpha-1 diffraction system (Panalytical) equipped with an incident beam monochromator. The system used only the Kα1 component of Cu radiation, improving the overall quality of the collected powder diffraction data. An accelerating voltage of 45 kV, current of 40 mA, 2θ-step of 0.03 and a hold time of 200 s were selected. X’Pert HighScore Plus software (Panalytical) was used for spectral analysis. The nitrogen adsorption and desorption isotherms were collected at 77 K in the range of relative pressures of 0.0002–0.99P/P0 using a TriStar II 3020 (V1.03) surface area and porosity measurement system (Micromeritics) and used for measurements of the SSA and pore size distribution in the 2–100 nm range. After drying the powder under a vacuum at 80 ◦ C for at least 12 h, 50–100 mg of each powder sample was degassed under a N2 gas flow at 300 ◦ C for at least 2 h before weighting and gas sorption measurements. The SSA and pore size distribution were calculated using the Brunauer–Emmett–Teller and Barrett–Joyner–Halenda methods, respectively, using Micromeritics DataMaster software. The relative pressure range of P/P0 from 0.05 to 0.3 was used for multipoint Brunauer–Emmett–Teller calculations. Ultrahigh-purity gases (99.99%, Airgas) were used for all experiments. SEM studies were carried out using a LEO
ARTICLES 1530 SEM microscope (LEO, now Nano Technology Systems Division of Carl Zeiss SMT). An in-lens secondary-electron detector was used for the studies, most of which were carried out using an accelerating voltage of 5 kV and a working distance of 2–5 mm. Electrochemistry. Working electrodes were prepared by casting slurry containing an active material (C–Si composite granules or graphite or annealed carbon black), a polyvinylidene fluoride binder (pure 9305 (Kureha) for carbon electrodes and 9305 with 10 wt% addition of polyacrylic acid for Si-containing electrodes; 20 wt% of the binder was used for annealed carbon black and for C–Si composite granules and 10 wt% for graphite) and N -methyl-2-pyrrolidone on an 18 µm Cu foil (Fukuda). The electrodes were calendared and degassed in vacuum at 70 ◦ C for at least 2 h inside an Ar-filled glove box (
