Lithium-ion batteries are currently the best option to power both electronic devices and electric vehicles. They are the ones that best meet the envelope of the criteria required of a battery: safety, energy density, weight, volume, sufficient life cycle and recyclability, all at a minimum cost. However, experts say that fluoride ion batteries are better, but the ideal materials that conduct these ions have not yet been found. A team from the University of North Carolina seems to have the solution to this problem by applying the use of machine learning as speeds up the material selection process and stimulates research.
The materials used to make today’s lithium-ion batteries are expensive and often hard to come by. Scaling its production to the levels required by the industry’s transition to fully electric cars requires a new approach to avoid both shortages and predictable price increases in the future.
Theoretically systems based on fluoride ions They are ideal to be used as batteries both to power consumer electronics and electric vehicles. Fluoride offers tangible advantages over lithium, as it is lightweight and highly stable. It is cheaper than lithium and cobalt, metals used in today’s lithium-ion batteries. In addition, scientists say that fluoride ion batteries can offer a higher energy density than their lithium-ion equivalents. They can multiply it by seven, which could translate into much greater autonomy than we currently know.
Taking into account all these theoretical advantages, fluoride ion batteries (FIB) meet the minimum criteria of the aforementioned envelope better than lithium batteries. Why are these batteries still not viable? The problem is that there is a big gap between theory and practice.
The fluoride ion (F-) is the form in which fluorine normally occurs in an aqueous solution. In a fluoride ion battery, electricity is generated by sending fluoride ions from one electrode to the other through a fluoride ion conducting electrolyte. Research on fluoride ion batteries is still are in their early stages. The first samples of rechargeable fluoride batteries appeared in 2011. Their biggest handicap is that scientists have not yet found the perfect conducting materials for fluoride ions. More specifically, those with high ionic conductivity lack stability and vice versa.
Jack Sundberg and his team at Scott Warren’s lab at the University of North Carolina have devised a new method to speed up the search for the best conducting materials for fluoride. The team uses machine learning algorithms and supercomputers to quickly and accurately predict how easily fluoride ions move in any crystal containing them.
The database managed by the team at the University of North Carolina has selected 10,000 fluoride-containing candidates from the 140,000 known materials. They chose 300 at random and made precise calculations to check the fluoride-carrying capacity of each material. This phase took about a week for each material. The results were used to train the algorithm machine learningmanaging to speed up the calculations to only one hour per material.
Research has already produced some candidate materials that are described as “better conductors than those used in lithium ion batteries”. An example is a fluoride-containing zinc-titanium compound, ZnTiF6. It is cheap, highly conductive to fluoride, and shows promising characteristics in other areas. However, it is not the only one, and the team has already patented the most promising compositions.
Solid Electrolyte Fluoride Ion Batteries
Having a solid starting point like this should speed up the development of fluoride ion batteries. Although it is too early to tell, this type of chemistry could be even better if implemented in future solid electrolyte batteries. One of the main advantages of these batteries is that they cannot catch fireincreasing its safety and allowing engineers to avoid having to create systems to prevent overheating.
Another challenge of these batteries is that they only work at high temperatures. Fluoride ions are useful conductors, that is, they move from one polarized electrode to another, but only when the solid-state electrolyte gets hot enough. This makes fluoride ion technology impractical for many consumer applications.