Aluminum-Air Batteries Across Scales: A Multiscale Framework for Electrochemical Characterization, Materials Optimization, and Electric Vehicle Integration

Abstract

Aluminum-air (Al-air) batteries have emerged as a promising energy storage technology due to their exceptionally high theoretical energy density, material abundance, and potential for low-cost deployment. However, their practical implementation remains constrained by challenges related to electrochemical performance, parasitic corrosion, electrolyte management, and system-level integration. This thesis presents a comprehensive multiscale investigation of Al-air batteries, integrating techno-economic analysis, experimental characterization, data-driven modeling, and system-level simulation to evaluate their viability as advanced energy storage systems and range extenders in electric vehicles (EVs). The study begins with a techno-economic assessment of metal-air batteries in EV applications. A comparative framework is developed to evaluate the performance of Al-air systems relative to conventional lithium-ion batteries, incorporating key metrics such as gravimetric energy density, vehicle energy consumption, and cost. The analysis shows that Al-air batteries, with practical energy densities of approximately 700-900 Wh kg⁻¹, significantly outperform Li-ion systems (150-250 Wh kg⁻¹), offering strong potential for range extension. Simulated vehicle scenarios indicate that Al-air integration can extend driving range by a factor of 2-5, depending on system configuration and operating conditions. However, these benefits are offset by trade-offs related to system complexity, auxiliary components, and cost, highlighting the need for integrated evaluation frameworks. Experimental investigations are conducted to examine the electrochemical performance of Al-air batteries under varying electrolyte compositions and operating conditions. Using a novel galvanic generator-type Al-air system with a rotating electrode configuration, multiple prototype units provided by AlumaPower were evaluated. The rotating electrode design enhances mass transport, reduces passivation, and promotes uniform anodic dissolution, enabling improved discharge stability compared to conventional static systems. Systematic experiments reveal the alkaline electrolytes, particularly in the range of 6-8 M concentration, provide optimal performance by balancing ionic conductivity and electrochemical kinetics. Peak power densities exceeding 500 mW cm⁻² are achieved under controlled conditions, while discharge tests at moderate current densities (~80-100 mA cm⁻²) exhibit stable voltage profiles in the range of 1.0-1.2 V. The results further demonstrate that increasing electrolyte concentration beyond optimal levels accelerates parasitic corrosion and hydrogen evolution, leading to reduced efficiency and highlighting the importance of electrolyte optimization. To address the critical challenge of aluminum corrosion, a data-driven predictive modeling framework is developed. Artificial neural networks (ANNs) are trained on experimental datasets to model the relationship between electrolyte composition, temperature, and electrochemical variables with corrosion metrics such as corrosion potential (Ecorr) and corrosion current density (Icorr). The ANN models achieve high predictive accuracy, with coefficient of determination (R²) values exceeding 0.99, demonstrating their capability to capture complex nonlinear relationships in electrochemical systems. To further enhance system performance, genetic algorithms (GA) and multi-objective optimization (NSGA-II) are integrated with the ANN framework to identify optimal operating conditions. The optimization results reveal trade-offs between maximizing Ecorr and minimizing Icorr, enabling the identification of optimal electrolyte conditions that balance performance and degradation. This integrated modeling approach represents a significant advancement over conventional empirical methods by enabling predictive and systematic optimization of corrosion behavior. At the system level, the thesis develops a comprehensive modeling framework for integrating Al-air batteries within EV architectures. Using MATLAB and Simulink, a dual-energy storage system is implemented in which Al-air batteries function as range extenders for lithium-ion battery packs. The system incorporates experimentally informed battery models and employs state-of-charge (SOC)-based control strategies to manage power flow between energy sources. Simulations conducted under standard driving cycles, including UDDS, WLTP, and HWFET, demonstrate that Al-air integration can significantly mitigate SOC depletion and extend vehicle range, particularly in reduced-capacity Li-ion configurations (e.g., 50% and 35% baseline energy). The results highlight the importance of control strategy design, power limitations, and system configuration in achieving optimal performance. Collectively, the findings of this thesis establish a comprehensive framework linking electrochemical behavior, corrosion kinetics, and system-level performance of Al-air batteries. The integration of experimental characterization, data-driven modeling, and vehicle-level simulation provides new insights into the practical feasibility of Al-air systems and identifies key design and operational parameters governing their performance. The use of industrially relevant prototype systems further enhances the applicability of the research and bridges the gap between laboratory studies and real-world implementation. This work demonstrates that Al-air batteries, supported by optimized electrolyte conditions, predictive corrosion modeling, and intelligent system integration, represent a viable pathway for next-generation energy storage and EV range extension, while advancing both scientific understanding and practical development.