Excitations in Complex Environments: Multiphysics embedding for large scale electronic structure

Quantum mechanical simulations from first principles are today used hand in hand with experiments to guide the design of new materials or biomolecules as they provide a very accurate description of the electrons that determine all the observable properties of the materials. With the advent of first principles quantum methods where the computational effort increases linearly with the number of atoms we have the capability to simulate complex materials at the forefront of research such as nanostructures (e.g. in fuel cell catalysts or electronic devices) and entire biomolecules (as needed in drug design or studies of components of the living cell). The UK-developed ONETEP program is the leading linear-scaling first principles quantum code, due to its new generation of theory that retains the full level of accuracy of conventional cubic-scaling first principles quantum methods. ONETEP has a wide and growing international user base not just within academia, but within industry (via the commercial version of the code distributed by BIOVIA). The code was developed from the beginning using modern software engineering principles with the aim of portability and high scalability to modern supercomputing platforms and user-friendly interactive input and output.

The present project aims to develop in ONETEP the capabilities for a whole new level of simulation. It will expand the regime of applicability of the code from the ground state to excited states; it will provide much more accurate approximations for the electrons (hybrid and range-separated exchange correlation functionals) and finally it will dispense with the monolithic single-theory description of the entire system by allowing to seamlessly combine different levels of theory that match the different parts of complex materials systems. A multitude of grand-challenge problems will become accessible to accurate simulation with these developments: examples include light energy harvesting in biomolecules such as chlorophyl, new materials for flexible and cheap organic photovoltaics, new types of lasers/masers. In all these problems there are different levels of complexity as the photoactive site and its environment are clearly distinct, thus the multilevel description will be indispensable.

This project is the flagship project of the CCP9 materials simulation community and has received overwhelming support with several members of the CCP9 consortium offering to be early adopters of our developments. The ONETEP code will become freely available to all UK academics (via free membership of CCP9, which is open to the whole UK academic community) and as a result it is expected to be accessible to all the materials, chemistry and biomolecular simulation communities. We will further promote the dissemination of the code via a dedicated masterclass (open to both academic and industrial users), and a European CECAM/Psi-k workshop. The new developments will also be disseminated to industry through their exposure within the BIOVIA Materials Studio graphical user interface via which the ONETEP code is marketed to industrial customers.

ONETEP has both a growing base of academic users and is a code that is commercially licensed to industry by Dassault Systèmes BIOVIA, a company that is committed to enhancing innovation and accelerating product development through scientific research and development. It has over 2,000 customers including 30 of the top 35 pharma & biotech companies, 7 of the top 10 chemicals companies and the top 5 aerospace companies. In particular, it has an established position in the field of materials modelling and simulation currently based around their Materials Studio package. The potential integration of ONETEP into their Discovery Studio package for life-sciences simulation is a particularly exciting development that will lead to greater impact of the code in pharma and biotech sector. The developments within this project will be made available to users across industrial and government sectors through BIOVIA. As the importance of theory and simulation of materials grows within industry, so the development of new methodologies with expanded capabilities will contribute to the UK economy through the creation of jobs and improved products.

One of the principal goals in developing ONETEP has been to maximise its economic impact by expanding the scale and scope of the simulations that are possible: for example, its capability to model much larger systems than traditional methods has made first-principles methods applicable to problems in nanotechnology and biology, such as modelling entire nanocrystals and proteins. In this proposal, we plan to further extend its capabilities to the simulation of electronically-excited states of large systems in complex environments. This will provide a simulation tool that the electronic structure community has identified as a priority to tackle high-impact research challenges including functional organic devices (photovoltaics and MASERs) and photoprotective molecules.

The existing strong links to industry, through collaborations and through BIOVIA, mean that the simulations enabled will accelerate the development of more competitive technologies and thus drive economic growth. Both emerging technologies, such as organic photovoltaics and MASERs, and established ones, such as dyes and sun-cream, stand to benefit from improved fundamental understanding of atomistic mechanisms of light-matter interaction. Understanding and improving upon light-harvesting in biological systems could lead to more productive crops via genetic modification, while understanding photodegradation and photoprotective mechanisms in organic compounds could lead to longer-lasting, higher-quality pharmaceutical products.

The team has a strong track record in training researchers to find successful strategies for simulating complex systems. In recent years, we have developed a training event in which the majority of the time is not spent giving lectures and running toy tutorial problems. Instead, "Masterclasses" are held in which a small number of attendees are invited to work for a week with experienced ONETEP developers and users on their own research problem. This event is open both to academic users (from theoretical and experimental groups) and to industrial partners (identified through our partner BIOVIA). Training events are complemented by online support and documentation on the ONETEP wiki (www.onetep.org) and ONETEP email forums. Members of the team will also give webinars on the latest developments for BIOVIA users and talks at BIOVIA International User Meetings.

Finally, Training PDRAs and PhD students in advanced electronic structure methods and high-quality software engineering prepares them well for R&D jobs in industry, in which there is an identified skills gap. Many former members of our groups are now working in software/HPC, including two for MathWorks, one for Materials Design, one for QuantumWise, one at the Argonne Leadership Computing Facility, and one at the A*STAR Institute of HPC.