You might already be using ChatGPT to ask for a recipe, clarify a doubt, prepare a lesson, or understand a complex topic. However, the power of generative AI extends far beyond these everyday applications.
Across laboratories, tech hubs, and industries worldwide, we are employing AI to tackle some of the planet’s most pressing challenges: generating energy more efficiently, reducing transportation emissions, and designing materials with a lower environmental impact. AI is helping us envision solutions that humans alone cannot achieve.
In my field of expertise, artificial intelligence is shaping the conceptual materials that will construct a sustainable future.
Materials Under Extreme Conditions
One of the significant challenges in the energy transition is finding materials capable of withstanding extreme conditions. This is evident, for example, in concentrated solar power, an interesting renewable alternative that stores heat in molten salts to generate electricity even when there’s no sun.
Las torres de energía solar de concentración de Ivanpah. Es una de las mayores plantas termosolares del mundo, situada en el desierto de Mojave (California, Estados Unidos).
Today, nitrate salts operating up to 560 °C are used, but molten chlorides capable of working at 780 °C and storing more energy are being researched. The issue is that at this temperature, they become highly corrosive, and conventional metals cannot withstand these conditions for long periods.
A similar situation occurs in the aerospace sector. Aircraft turbine temperatures are extremely high: the higher the temperature, the greater the efficiency, less fuel is consumed, and fewer emissions are generated.
However, surpassing 1000 °C requires materials that do not deform or oxidize under extreme conditions. In both cases –turbines and corrosive salts– we need metals that do not yet exist in the market.
High-Entropy Alloys
This is where high-entropy alloys come into play, a type of material that has revolutionized metallurgy since 2004.
Unlike traditional alloys, formed around a primary element like iron or nickel, these combine five or more elements in similar proportions. This change opens up an almost infinite design space, covering a vast number of possible combinations across the periodic table, resulting in diverse properties such as high mechanical strength at high temperatures or excellent performance against corrosion and oxidation.
AI allows for the rapid evaluation of millions of compositions, identifying the most promising ones and reducing the time needed to transition from a hypothesis to a real material.
For instance, if we aim for an alloy capable of withstanding molten chlorides at 780 °C in thermal power plants or maintain mechanical resistance above 1000 °C in an aircraft turbine, AI can sift through the options and highlight the most viable candidates.
Dependency on China for Batteries
AI is also making strides in battery materials, an area crucial for the growing electrification of transportation and deployment of energy storage technologies.
The increasing electrification of transportation and the rollout of energy storage technologies heavily rely on critical materials like lithium, cobalt, nickel, and graphite used in lithium-ion batteries.
However, these elements not only pose environmental and technological challenges but also geopolitical ones.
The supply chain is heavily concentrated, with China controlling a significant portion of the global processing and refining of many strategic materials, as well as growing influence over cobalt extraction in Central Africa. This dependence limits the technological autonomy of regions like Europe and exposes them to trade tensions or export restrictions, as has already happened with other key metals like gallium and germanium.
Developing new compositions that reduce or eliminate the need for these elements has become a scientific and industrial priority.
New materials for electrodes or solid electrolytes, more abundant, recyclable, and with a lower environmental footprint, are essential to ensure more sustainable, accessible, and resilient batteries amid international context fluctuations.
AI Alone
Applying AI to material design opens a promising avenue for identifying alternatives with functional properties similar to those currently used, but made from safer, more local, and sustainable elements.
The challenge lies in exploring this vast combination space. With traditional methods based on trial and error, it would be slow and costly.
The Food for Artificial Intelligences
Designing with artificial intelligence does not mean pressing a button and obtaining the perfect answer. It depends on the available data. Generating reliable information through experiments, standardizing databases, and sharing results between research centers is now a priority.
Without this foundation, AI models cannot learn or produce solid predictions.
When experimental data is unavailable, another essential tool is computational simulation. Physical and chemical models allow anticipating how a material would behave under certain conditions, generating synthetic data that feed AI algorithms.
Discoveries Guided by Design
This approach aligns with what is known as ‘design-driven discovery,’ or design-guided discovery.
Unlike the traditional logic –called ‘material-driven’– where one starts with an existing material and observes its properties, the design-driven approach begins by defining which properties are needed (e.g., corrosion resistance at 780 °C for solar panels) and then searches for the chemical combinations capable of fulfilling them.
Instead of discovering what a material can do, we directly design the one we need to perform that function. And artificial intelligence is the tool that makes this shift possible.
AI is a strategic ally if we want the world to be sustainable.
