Federated Learning in Small-Cell Networks: Stochastic Geometry-Based Analysis on the Required Base Station Density.

Recently, federated learning (FL) has been receiving great attention as an effective machine learning method to avoid the security issue in raw data collection, as well as to distribute the computing load to edge devices. However, even though wireless communication is an essential component for implementing FL in edge networks, there have been few works that analyze the effect of wireless networks on FL. In this paper, we investigate FL in small-cell networks where multiple base stations (BSs) and users are located according to a homogeneous Poisson point process (PPP) with different densities. We comprehensively analyze the effects of geographic node deployment on the model aggregation in FL on the basis of stochastic geometry-based analysis. We derive the closed-form expressions of coverage probability with tractable approximations and discuss the minimum required BS density for achieving a target model aggregation rate in small-cell networks. Our analysis and simulation results provide insightful information for understanding the behaviors of FL in small-cell networks; these can be exploited as a guideline for designing the network facilitating wireless FL.

Combining ReaxFF Simulations and Experiments to Evaluate the Structure-Property Characteristics of Polymeric Binders in Si-Based Li-Ion Batteries.

High energy capacity silicon (Si) anodes in Li-ion batteries incorporate polymeric binders to improve cycle life, which is otherwise limited by large volume and stress fluctuations during charging/discharging cycles. Several properties of the polymeric binder play a role in achieving optimal battery performance, including interfacial adhesion strength, mechanical elasticity, and lithium-ion conduction rate. In this work, we utilize atomistic simulations with the ReaxFF force field and complementary experiments to investigate how these properties dictate the performance of Si/binder anodes. We study three C/N/H-based polymer binders with varying structures (pyrolyzed polyacrylonitrile (PPAN), polyacrylonitrile (PAN), and polyaniline (PANI)) to determine how the structure-property characteristics of the binder affect performance. The Si/binder adhesion analysis reveals some counter-intuitive results: although an individual PANI chain has a stronger affinity to Si compared to PPAN, the PANI bulk binds weaker to the Si surface. Interfacial structural analyses from simulations of the bulk phase show that PANI chains have poor stacking at the interface, while PPAN chains exhibit dense and highly ordered stacking behavior, leading to stronger adhesion. PPAN also has a lower Young's modulus compared to PANI and PAN owing to its ordered and less entangled bulk structure. This added elasticity better accommodates volume changes associated with cycling, making it a more suitable candidate for Si anodes. Finally, both simulations and experimental measurements of Li-ion diffusion rates show higher Li mobility through PPAN than PAN and PANI because the ordered stacking of PPAN chains creates channels that are favorable for Li diffusion to the Si surface. Galvanostatic charge-discharge cycling experiments show that PPAN is indeed a highly promising binder for Si anodes in Li-ion batteries, retaining a capacity of ∼1400 mAh g-1 for 150 cycles. This work demonstrates that the orientation and structure of the polymer at and near the interface are essential for optimizing binder performance as well as showcases the initial steps for binder evaluation, selection, and application for electrodes in Li-ion batteries.

Investigating the Compatibility of TTMSP and FEC Electrolyte Additives for LiNi0.5Mn0.3Co0.2O2 (NMC)-Silicon Lithium-Ion Batteries.

This study examines the compatibility of multielectrolyte additives for NMC-silicon lithium-ion batteries. Research studies with Si-based anodes have shown stable reversible cycling using electrolytes containing fluoroethylene carbonate (FEC). At the same time, the electrolyte additive, tris(trimethylsilyl) phosphite (TTMSP), has shown to improve the electrochemical performance of nickel-rich layered cathodes, such as LiNi0.5Mn0.3Co0.2O2 (NMC). However, the combination of these electrolyte additives for the realization of a full-cell NMC-Si lithium-ion battery has not been previously explored. Changes in the electrochemical performance (capacity retention, internal cell resistance, and electrochemical impedance) in half-cells are studied as the ratio of TTMSP and FEC is tuned. At the optimal TTMSP/FEC ratio of 0.33 (T1F3), the NMC-Si full-cells achieve a 2× longer cycle life when compared to the FEC-rich (T0F4) electrolyte. Moreover, T1F3 full-cells demonstrate 1.5 mAh/cm2 areal capacities and high-capacity retention (25% more than T0F4). A detailed investigation of the electrode-electrolyte interfaces is conducted by using time-of-flight secondary ion mass spectroscopy (ToF-SIMS) and X-ray photoelectron spectroscopy (XPS). The chemical species depth profiles and elemental analysis illustrate adequate hydrogen fluoride (HF) scavenging. These results demonstrate the synergistic effects of electrolyte additives in minimizing the capacity degradation in NMC-Si full-cells by effectively stabilizing the electrode-electrolyte interfaces.