The Nuts and Bolts of FirstPrinciples Simulation 27

The Nuts and Bolts of First-Principles Simulation 27: Upcoming Developments and the Future for DFT Durham, 6 th-13 th December 2001 Nuts and Bolts 2001 CASTEP Developers’ Group with support from the ESF k Network Lecture 27: Developments and the Future

Outline Brief history of DFT r Prospects for the next 5 years u New functionality u Scientific challenges u New research fields r Beyond the next 5 years r Nuts and Bolts 2001 Lecture 27: Developments and the Future� 2

Brief History of DFT 1964 Hohenberg-Kohn theorem - foundation of density functional theory. r 1965 Kohn-Sham equations - practical method for applying density functional theory. r r THEN NOTHING until Nuts and Bolts 2001 Lecture 27: Developments and the Future� 3

Brief History of DFT (all dates are approximate) 1981 First density functional theory calculations r 1985 Car-Parrinello method r 1989 Conjugate gradients techniques r 1991 Implementation on parallel computers r 1992 Chemists begin to adopt DFT - useful GGAs r 1996 Residual minimisation, DIIS, density mixing r 1998 Nobel prize for chemistry awarded to Walter Kohn r Nuts and Bolts 2001 Lecture 27: Developments and the Future� 4

New Functionality r Linear response - relevant to many properties: u NMR u g-tensor of electron u Dielectric properties u Phonons u Born effective charges u Elastic constants u Anything that can be phrased in terms of a perturbed external potential Nuts and Bolts 2001 Lecture 27: Developments and the Future� 5

New functionality r Phonons can be used to calculate vibrational free energies: u Phase stability u Crystal morphology u Defect processes Nuts and Bolts 2001 Lecture 27: Developments and the Future� 6

BUT Plane waves allow fast algorithm development. When this is combined with a modular program design built from a full specification document adding new functionality should be fast, efficient (many operations will be available already) and as painless as possible. CASTEP should be ideal for developing new functionality, for studying novel problems and for taking to new scientific fields. Nuts and Bolts 2001 Lecture 27: Developments and the Future� 7

Scientific challenges Finite temperatures u Free energies, entropy u Phase transitions u. . p Larger systems u Embedding, QM/MM approaches, . . . p Longer timescales p Complex phase spaces r Nuts and Bolts 2001 Lecture 27: Developments and the Future� 8

New research fields Biology r Materials science r Nuts and Bolts 2001 Lecture 27: Developments and the Future� 9

Beyond the next 5 years Linear scaling approaches r Embedding r Nuts and Bolts 2001 Lecture 27: Developments and the Future� 10

Thanks Workshop organisers - Stewart Clark, Phil Lindan, Walter Temmerman r Workshop sponsors - Accelrys, ESF-STRUC Psi-k Programme r CASTEP Developers Group - Stewart Clark, Phil Hasnip, Phil Lindan, Chris Pickard, Matt Probert and Matt Segall r Nuts and Bolts 2001 Lecture 27: Developments and the Future� 11