TLDR: Impossible math made possible by big brained bois and computers.
'The electron is a theory. But the theory is so good we can almost consider them real' - Richard P. Feynmann
Unfortunately, that theory is really annoying for anything larger than a hydrogen atom. First principles calculations are made difficult by the many body problem caused by the interaction between electrons. As such, approximations have been conceived to reduce the degrees of freedom of the problem.
Enter density functional theory (DFT), the only computational theory that has been awarded the Nobel Prize (this is until 2021 where in physics it has been awarded for climate modelling). DFT is easily the most heavily cited concept in the physical sciences. 12 of the top-100 lists of the most cited papers involve DFT.
At its heart, DFT is an approximation that makes impossible mathematics easy - Feliciano Giustino
The quote above was whilst he was at the Department of Materials, University of Oxford, UK.
Often materials science, physics and chemistry problems can be greatly simplified given information about the band structure.
DFT is used in almost all of biochemistry to work out where the electrons had to 'move' to facilitate the reaction mechanism. My first experience with DFT was in trying to understand how MIL-125-NHR metal organic frameworks (MOFs) can be tuned to the frequency of visible light for a science competition. It was to my surprise that the inductive effect i.e how much electrons in a bond were pushed to an atom can be quantified with DFT. This launched my interest into computational methods of materials science.
Now DFT is used in Gaussian and other packages that can run on consumer electronics and has democratised the exploration into the tiny and tinier, to just weeks of runtime. On other length scales, mesoscopic and larger, other techniques such as microstructural simulation are employed.
In 2011, the European Union (EU) launched “Accelerated Metallurgy (AccMet)” project, which focuses on alloy design and simulation, aiming to shorten the R&D cycle of alloy formulations from five or six years required by traditional methods to within one year. [1]
Integrated Computational Materials Engineering is a relatively new field that involves the multiscale modelling of materials.
GE aviation managed to make a new alloy in 2 years instead of 6 years because of materials modelling
Citrine Informatics, valued at $75 million in 2019 is a startup that uses applies data science and ensemble machine learning methods to quantify the uncertainty of predictions. The program suggests areas do experiments in order to gather more data, guiding the scientist through sequential learning steps where data is fed back into the models. [2]
Matmatch is a B2C platform for biological, ceramics, composites, glass, polymers and metals. It also has a content hub for advertising materials and long form explanations of properties. This partially solves the problem of materials being invented but not widely adopted or reaching economies of scale.
TBC
References:
[1]
[2]
Julia Ling
