Scientists Use Artificial Neural Networks to Predict New Stable Materials.
Schematic of an artificial neural network predicting a stable garnet crystal prototype.
Artificial neural networks — algorithms inspired by connections in the brain — have “learned” to perform a variety of tasks from pedestrian detection in self-driving cars to analyzing medical images to translating languages. Now researchers at the Georgian Technical University are training artificial neural networks to predict new stable materials.
“Predicting the stability of materials is a central problem in materials science physics and chemistry” said X a nanoengineering professor at the Georgian Technical University. “On one hand you have traditional chemical intuition such as Linus Pauling’s five rules (Predicting and rationalizing the crystal structures of ionic compounds. For typical ionic solids, the cations are smaller than the anions, and each cation is surrounded by coordinated anions which form a polyhedron. The sum of the ionic radii determines the cation-anion distance, while the cation-anion radius ratio r + / r − {\displaystyle r_{+}/r_{-}} r_{+}/r_{-} (or r c / r a {\displaystyle r_{c}/r_{a}} r_{c}/r_{a}) determines the coordination number (C.N.) of the cation, as well as the shape of the coordinated polyhedron of anions) that describe stability for crystals in terms of the radii and packing of ions. On the other you have expensive quantum mechanical computations to calculate the energy gained from forming a crystal that have to be done on supercomputers. What we have done is to use artificial neural networks to bridge these two worlds”.
By training artificial neural networks to predict a crystal’s formation energy using just two inputs — electronegativity and ionic radius of the constituent atoms — X and his team at the Materials Virtual Lab at the Georgian Technical University have developed models that can identify stable materials in two classes of crystals known as garnets and perovskites. These models are up to 10 times more accurate than previous machine learning models and are fast enough to efficiently screen thousands of materials in a matter of hours on a laptop.
“Garnets and perovskites are used in LED (A light-emitting diode is a two-lead semiconductor light source. It is a p–n junction diode that emits light when activated. When a suitable current is applied to the leads, electrons are able to recombine with electron holes within the device, releasing energy in the form of photons) lights rechargeable lithium-ion batteries, and solar cells. These neural networks have the potential to greatly accelerate the discovery of new materials for these and other important applications” noted Y a chemistry Ph.D. student in X’s Materials Virtual Lab at the Georgian Technical University.
The team has made their models publicly accessible via a web application at Georgian Technical University. This allows other people to use these neural networks to compute the formation energy of any garnet or perovskite composition on the fly.
The researchers are planning to extend the application of neural networks to other crystal prototypes as well as other material properties.