As strategic components for energy storage and electric mobility, batteries are increasingly sought after. Initiatives are being advanced to enhance their performance and ensure the sustainability and reliability of their materials. Artificial intelligence systems are playing a crucial role in both respects.
Future batteries face a dual challenge: managing the ever-increasing generation of energy from renewable sources and powering a growing fleet of electric vehicles.
Regarding energy, «we are on course to add, over the next five years, more renewable capacity than that installed since the first commercial renewable energy power plant was built more than a century ago», according to the International Energy Agency in its Renewables 2023 report. An overwhelming 95% of this new capacity, amounting to nearly 3,700 GW, will originate from intermittent sources such as solar and wind.
Moreover, future batteries will need to address the increasing burden on the electric grid caused by distributed generation, playing a pivotal role in maintaining network stability and reliability amidst rising demand.
In the realm of electric mobility, the International Energy Agency, in its recently released Global EV Outlook 2024, predicts that – globally, for the year 2024 – over one in five cars sold will be electric, with a surge in demand expected over the coming decade. By 2030, it is forecasted that electric vehicles will constitute over two-thirds of global car sales [source: Rocky Mountain Institute].
In this context, the demand for batteries is expected to rise sharply, with the global market projected to approach nearly $424 billion by 2030, nearly quadrupling its value since 2021 [source: Statista]. This robust demand necessitates an adequate supply of solutions, thus underscoring the need for reliable raw material sourcing.
Currently, lithium-ion batteries are the predominant choice, especially in e-mobility. The principal sourcing challenges are as follows: 75% of the world’s lithium production is sourced from Australia and Chile; for cobalt, used in Li-Ion batteries, 70% is sourced from the Democratic Republic of Congo. In both instances, processing is mainly conducted in China, as specified by the IEA in its contribution to Grid-scale storage.
Research has been steadfast in pursuing solutions that cater to multiple needs by exploring alternative chemistries that are equally effective for both energy storage and electric mobility. In this pursuit, artificial intelligence techniques are increasingly utilized to enhance the precision of these discoveries, leading to encouraging outcomes.
TAKEAWAYS
What is a battery?
A battery is an electrochemical device that transforms chemical energy into electrical energy through a redox reaction. It is composed of several electrochemical cells, each featuring three primary components: the anode, the cathode, and the electrolyte.
The anode (negative pole) is the terminal that releases electrons to an external circuit during the electrochemical reaction, whereas the cathode (positive pole) receives electrons from the external circuit. The electrolyte acts as the ionic conductor that facilitates the movement of negative and positive ions. It can be either liquid or solid; in the liquid form, it typically consists of a solvent (such as water) in which a salt, acid, or base is dissolved.
Lithium-ion batteries are the most commonly employed across various sectors, from portable electronic devices to electric vehicles. Their appeal lies in their impressive energy density per unit of mass and volume, surpassing other electrical energy storage systems. Additionally, Li-ion batteries offer an excellent weight-to-power ratio, high energy efficiency, long lifespan, and robust performance at high temperatures.
Their already notable growth is expected to accelerate dramatically with the expansion of the electric vehicle market. By 2030, global demand is estimated to increase from approximately 700 GWh in 2022 to around 4.7 TWh [source: McKinsey]. Consequently, Li-ion batteries are anticipated to dominate the automotive sector in the near future.
Over the past century, batteries have consistently been at the center of ongoing innovation, driven by ongoing research. To appreciate the level of interest in batteries, one might consider the volume of patent applications filed with the European Patent Office in 2023. The sector covering electric machines, household appliances, and energy noted a 12.2% increase from 2022, with cleantech inventions, particularly batteries, experiencing a notable surge of 28%.
New chemicals for future batteries, facilitated by artificial intelligence
The pursuit of future batteries focuses research on improving performance and finding economically sustainable alternatives to essential materials such as lithium, nickel, and cobalt. Notably, cobalt has experienced dramatic price volatility; it plummeted from $60,000 per ton to $32,000 in 2023 due to market surplus, though it is projected to rise by 2027 [source: Reuters].
Artificial intelligence techniques have proven instrumental in devising novel solutions. In research led by Austin Sendek, Professor of Materials Science and Engineering at Stanford University, the lithium-boron-sulfur (LBS) lithium-ion electrolyte was discovered, showcasing considerable stability. Over seven years, Sendek and his team analyzed lithium-containing materials to identify viable candidates for solid electrolytes. Among more than 12,000 known lithium-containing substances, the AI algorithms identified 21 promising candidates. Subsequent simulations honed their focus to four new compounds showing substantial potential for solid-state batteries. This research was conducted in 2020.
Further studies in 2024 by the Pacific Northwest National Laboratory and Microsoft illustrated the acceleration of outcomes via combined AI and high-performance computing. The Microsoft-PNNL study specifically sought possible solid-state electrolytes for battery applications.
As detailed in “Accelerating computational materials discovery with artificial intelligence and cloud high-performance computing: from large-scale screening to experimental validation“, the integration of machine learning and traditional physics-based models enabled the team to sift through over 32 million candidates. The outcomes were remarkable, with around half a million potentially stable materials identified within just 80 hours.
This method allowed for the filtration of materials based on stability, reducing the list to fewer than 600,000 candidates. Further AI-driven analyses selected candidates with the requisite electrical and chemical properties for batteries while discarding rare, toxic, or costly materials. Though still in preliminary stages, the promise is vast and could lead to the creation of new, more sustainable chemical formulations. According to Microsoft, this approach might eventually facilitate the manufacture of future batteries containing 70% less lithium than current Li-Ion batteries.
Robotics, digital twins, and machine learning propel future batteries
Developing new chemistries is a primary focus of research into future batteries, with ongoing refinement of methods to identify promising raw materials. An example of this is the collaborative effort by the University of Michigan and Samsung’s Advanced Materials Lab, showcased in the recent article “Better battery manufacturing: Robotic lab vets new reaction design strategy“. This project aimed to blend unconventional ingredients to forge chemical compounds for easier-to-produce batteries via robotic laboratories. The researchers implemented a strategy that more reliably purifies materials than traditional sequential reaction methods. They noted:
«If the chemical bonds in intermediate compounds are tough to break, they may not fully react with other ingredients. Incomplete reactions can leave intermediates as unwanted impurities in the final material»
Robotic arms were employed to manipulate ingredients and operate lab equipment to assess the purity of the produced materials. Concurrently, each experiment’s outcomes were documented, creating a database that refined results based on recipe efficacy. These experiments demonstrated that the new formulas enhanced material purity by up to 80%.
From research to industry, the case of SES AI stands out as the first company globally to supply an automotive manufacturer with advanced battery prototypes featuring a novel technology in collaboration with GM, Hyundai, and Honda, as highlighted in The Wall Street Journal’s “The EV Battery of Your Dreams Is Coming“. Specializing in lithium metal cell batteries, SES AI is developing two platforms: one based on battery digital twins and another on AI. The first platform enhances traceability, management, and monitoring of the battery’s status, encapsulating all pertinent data from the material’s source mines through to battery production and end-of-life recycling. The second platform involves training machine learning models to devise new materials and processes.
Glimpses of Futures
In the quest to design the future batteries, there is an increasing focus on more readily available and sustainable raw materials such as sodium and potassium. Notably, research conducted by PNNL and Microsoft has highlighted potential candidates for a new electrolyte, akin to the current one but capable of substituting lithium with sodium. It is crucial to recognize that sodium, while being one of the most abundant elements on Earth, not only shares characteristics with lithium but also offers significant economic advantages essential for advancing battery technology during the energy transition.
Anticipating future scenarios, we apply the STEPS matrix to assess the potential impacts of seeking more dependable and accessible materials for batteries across social, technological, economic, political, and sustainability dimensions.
S – SOCIAL: a pivotal factor for batteries concerns the materials they are composed of, including copper, graphite, nickel, lithium, aluminium, manganese, and cobalt. The extraction of these materials, whether direct or indirect, frequently leads to social and economic impacts on local communities, such as loss of agricultural land and degradation of habitats. The shift towards battery chemistries aimed at minimizing environmental and social impacts is clear. For instance, Swedish developer Northvolt has recently introduced a sodium-ion battery for energy storage systems that surpasses its counterparts in energy density and «is safer, more cost-effective, and sustainable compared to conventional chemistries such as nickel-manganese-cobalt (NMC) or lithium iron phosphate (LFP). It utilizes abundant minerals like iron and sodium, which are widely available in global markets».
T – TECHNOLOGICAL: the development of future batteries is centred on technological innovations that enhance performance, safety, stability, and longevity. Solid-state batteries, distinguished by their solid electrolytes instead of liquid, are particularly promising. These batteries provide a higher energy density than traditional lithium-ion batteries, are lighter, and can be charged more rapidly. The exploration of alternative elements like potassium and magnesium is also intensifying. A notable breakthrough includes a team from Tohoku University in Japan developing a new cathode material for magnesium rechargeable batteries that maintains efficient charging and discharging even at low temperatures.
E – ECONOMIC: the surging demand for energy storage and electric vehicles led to significant battery price increases in 2022, followed by a 14% decline in 2023, setting a record low of $139/kWh. This reduction was driven by decreased costs of raw materials and components, alongside expanded production capacity [source: BloombergNEF]. According to a recent study by Nikhil Bhandari, co-head of Asia-Pacific Natural Resources and Clean Energy Research at Goldman Sachs, battery prices are projected to fall nearly 40% between 2023 and 2025.
P – POLITICAL: the strategic role of batteries in the energy transition has secured substantial backing through pivotal political decisions. The U.S. Inflation Reduction Act and Bipartisan Infrastructure Law have greatly accelerated the development of renewable energy facilities and battery manufacturing centers. Last year, the U.S. invested $239 billion in cleantech solutions, marking a 38% increase from 2022 [source: Clean Investment Monitor]. The European Union has been fostering the development of a competitive European battery industry since 2017, culminating in the launch of the first gigafactory in the EU three years later. As the Net-Zero Industry Act (NZIA) awaits approval, the EU has initiated BATT4EU, a co-programmed public-private partnership under Horizon Europe, designed to ensure the effective allocation and utilization of funds dedicated to battery research.
S – SUSTAINABILITY: a decisive challenge for next-generation batteries is reducing environmental impacts and enhancing the circularity of end-of-life materials. In this context, the European Parliament and Council enacted Regulation (EU) 2023/1542 D in 2023, mandating the collection, recycling, and reuse of batteries in Europe, ensuring that batteries marketed in the EU contain only limited amounts of necessary hazardous substances. Gigafactories are being designed to incorporate recycling facilities for battery components and materials at their end of life. The transition towards using more sustainable and accessible raw materials like sodium, magnesium, and potassium as alternatives to more impactful substances is a trend being pursued with increasing vigor.
