I actually think this is a great example of why AI is so important. This is a relatively simple combination of ions and solvents that has this remarkable ability and yet it has taken us decades to discover it.
1
1
13
Replies
Why? Because we have not had the ability to predict the properties of electrolyte solutions and so must rely on empirical trial and error to discover new promising electrolytes. It is the same story with lithium ion batteries.
1
0
5
Graphite intercalation of lithium was discovered back in the 70s but it took two decades to commercialise because we had to stumble on the right solvent to allow reversible intercalation.
1
0
1
How can AI address this? Parrinello is already showing how neural networks can massively accelerate simulation of Haber-Bosch:
This is so exciting: We can now just see what happens at a molecular scale for incredibly important processes!
0
4
4
1
0
1
And we can now also simulate simple lithium electrolytes with unprecedented speed and accuracy thanks to equivariant neural network potentials
So pleased to get this preprint out. Feel like we’ve finally worked out how to do something I’ve been trying to do for 13 years since the start of my PhD: Build an accurate continuum solvent model of ion-ion interactions in solution.
5
11
82
1
0
2
These simulations allow should allow us to actually see what is happening at the molecular scale and so provide direct physical insight and predictivity.
1
0
2
This approach should enable us to begin screening electrolytes computationally instead of empirically, which should be much faster and enable us to build up a databases of electrolyte solution properties that can serve as training data for a new generation of models.
1
0
2
There is almost certainly a vast number of ion-solvent combinations that have remarkable properties like the ones mentioned above waiting to be discovered if only we had a more efficient and scalable means of searching for them.
0
0
1