Xin Xu1*

1Fudan University, Shanghai, 20048, China

EXTENDED ABSTRACT: In materials science, trial-and-error has long been used. Thus, huge experimental resources and time are often required in order to develop new materials with desired properties. On the other hand, thousands of materials have already been synthesized over the years, while, for most of them, there is still no thorough assessment. Nowadays, machine-learning based computational design is becoming a new design paradigm in materials science. For the targeted property, candidate materials can be screened out effectively by computational simulations alone and there is no need of any real materials synthesis and actual measurement of their properties. Computational simulation has now been an indispensable part of material genomics engineering, producing a large number of data and dramatically speeding up the process of materials design. As an exciting vision made in a 2016 Nature paper, titled "Can artificial intelligence create the next wonder materials?"[1], "We are now seeing a real convergence of what experimentalists want and what theorists can deliver", which, however, imposes a great challenge to the computational simulations. "The truth is, some errors come with the theory itself: we may never be able to correct them", as also stated in the Nature paper [1]. Density functional theory (DPT) is now the most popular computational method in materials science. As the new generation functionals, doubly hybrid approximations (DHAs) have been shown to improve over conventional functionals with unprecedented accuracy in describing various kinds of chemical interactions and properties for finite molecules [2,3]. Recently, we realized the periodic implementations of DHAs [ 4,5] and demonstrated that the accuracy of DHAs achieved for finite molecules can be transferred to extended materials. Thus, DHAs hold the promise to create the big data with high quality, which is of vital importance for the success of machine-learning based materials design.
Keywords: material genomics engineering; density functional theory; doubly hybrid functional
REFERENCES

[1] N. Nosengo, Nature, 533, (2016) 23-25;

 [2] Y. Zhang, X. Xu, and W.A. Goddard, Proc. Natl. Acad. Sci. USA, 106, (2009), 4963-4968; 

[3] I. Y. Zhang, X. Xu, Y. Jung, and W.A. Goddard, Proc. Natl. Acad. Sci. USA, 108, (2011), 19896-19900;

 [4] B. Chen, X. Xu, J. Chem. Theory Comput. 16 (2020), 4271-4285;

 [5] Y Wang, Y. Li, J. Chen, I. Y. Zhang, X. Xu, JACS Au, 1, (2021), 543-549.