The dynamics of nanomaterials and light metal deuterides studied by means of coherent inelastic neutron scattering measurements and model simulations

Lead Research Organisation: University of Salford
Department Name: Unlisted

Abstract

Neutron inelastic scattering is a very powerful technique for studying the dynamics of atoms in solids. For incoherent scatterers such as hydrogen, the measurement yields a vibrational density of states. For coherent scattering materials, the experiment yields information about the relative motion (both in phase and frequency) of pairs of atoms in the unit cell. With a single crystal sample, measured for instance on a Triple Axis Spectrometer, the measurement yields phonon dispersion curves which can be directly related to the forces between atoms in the crystal. However, for coherently scattering polycrystalline materials, the superposition of phonon frequencies in different directions yields a complex 2-D contour plot in Q and omega ( related to momentum transfer and energy transfer in the collision with the neutron) and to date, there have only been a handful of experiments attempting to unravel this complex picture. However, with the development of powerful computers, of ab initio simulation of atomic structures and interactions and flexible software tools that can provide complete models of the lattice dynamics for any material, it becomes practical to approach the problem from the other point of view, by simulating the coherent inelastic scattering from a polycrystal (poly-CINS) and seeing how well it matches the experimental data. We have developed this method so as to analyse the poly-CINS from graphite as this material is not available in true single crystalline form, and have demonstrated that the existing models are not adequate descriptions of its lattice dynamics. It is particularly interesting to apply this method to the more complex polymorphs of carbon. For instance, our model calculations demonstrate that poly-CINS data for single walled carbon nanotubes should yield a value for Young's Modulus along the tube axis. Indeed, the method will be particularly interesting to apply to nanomaterials in general where the material has a structure on the nanoscale. We have already obtained interesting data on natural graphite, carbon nanohorns and carbon fibres which all show unexpected structure in the inelastic scattering that will require considerable modeling effort to interpret. We will also apply the technique to the interpretation of scattering from light complex hydrides which are being investigated as possible hydrogen storage materials for use on cars. These materials are not available as single crystals and other methods of investigating the lattice dynamics are complicated because of the nature of the H-H bonding which existing techniques cannot easily unravel. Using deuterated polycrystals, we should be able to interpret the inelastic scattering to validate the ab initio calculations and hence the enthalpy of formation at finite temperaturesA major objective is to make the method available to neutron scatterers internationally. We have already set up an extensive group of collaborators who will provide samples for measurement and will collaborate in the data interpretation. The software is being developed in collaboration with Prof. Fultz's group at Caltech, which has recently started working towards the same objective. We plan a workshop in the first year to establish the protocol for these collaborations and a further workshop in the final year to introduce the method to the international neutron scattering centres.

Publications

10 25 50
publication icon
Buchan J (2015) Molecular dynamics simulation of radiation damage cascades in diamond in Journal of Applied Physics

publication icon
Garba MT (2010) Parallel Computational Modelling of Inelastic Neutron Scattering in Multi-Mode and Multi-Core Architecture in 12th IEEE Int Conf. on High Performance Computing and Communications (2010)

publication icon
Garba MT (2013) GPU Acceleration for Hermitian Eigen Systems in Lecture Notes in Computer Science (Springer)

 
Description The gold standard for being able to accurately determine the forces between atoms in crystalline solids is to obtain dispersion curves using coherent inelastic neutron scattering. To do this, one would traditionally have to grow a single crystal large enough to measure the dispersion curves on a triple axis spectrometer. However, most materials are polycrystalline in nature and do not easily lend themselves to the growth of such large, highly ordered single crystals, and so information about the spatial arrangement of forces in powder samples is usually not possible to precisely determine.

This project has resulted in a method by which dispersion curves can be extracted from powder samples - the poly-CINS technique for analysing coherent inelastic neutron scattering data from powders. This method, which relies on a hybrid computational simulation and experimental analysis approach, uses the selection rules of coherent scattering to identify key features in theoretical data. These simulated features are matched with dispersion curves from the model and used to identify equivalent features in the experimental data. This iterative method can then be used to improve model fits, determine semi-empirical or ab initio model predictive capability and general suitability and to build new models from experimental data.
As well as the method itself, the project has also developed two pieces of user-friendly and powerful software - Scatter (the simulation code built into the popular GULP modelling software) an PreFiT, a profile refinement and data analysis suite. In order to do this effectively, considerable work was undertaken on massive parallelism of the simulation code, although this method is several orders of magnitude more efficient than alternative approaches (requiring fits with tens and hundreds of data points, rahter than tens or hundreds of thousand) being explored elsewhere.
Exploitation Route The project developed a new spectroscopic method for analysing previously inaccessible experimental data, which has the potential to open up the detailed spatial mapping of forces in powders in a way only previously available for single crystal samples.

International neutron sources (SNS, ISIS, HZB, ANSTO and JPARC) have shown interest in the method and discussions with these central facilities are currently underway aimed at popularising, exploiting and further developing this innovative spectroscopic method.
Sectors Aerospace, Defence and Marine,Construction,Digital/Communication/Information Technologies (including Software),Electronics,Energy,Manufacturing, including Industrial Biotechology,Pharmaceuticals and Medical Biotechnology

 
Description The spectroscopic method developed in this project is now being integrated into powder spectrometer work being conducted at the ISIS facility (and elsewhere). The plan is to further extend the method to produce hardware-integrated software that is capable of approaching the analysis of powder systems in a more automated way. This new method, which is still being introduced to physicists and materials scientists at national facilities such as ISIS, SNS, JPARC and the ESS, has the potential to provide powerful new insight into the dynamics of powder samples, where previously the only other option has been to rely on phonon density of states or so called 'full spectrum' fitting methods which lack both the efficiency and feature specificity that poy-CINS provides. Key fundamental energy based materials currently being examined using this technique include key energy related materials such as nuclear graphites and magnesium hydride, a promising hydrogen storage material.
First Year Of Impact 2014
Sector Energy
Impact Types Economic

 
Description ICase with AWE
Amount £135,186 (GBP)
Organisation Atomic Weapons Establishment 
Sector Private
Country United Kingdom
Start 08/2008 
End 07/2012
 
Description Icase with AWE
Amount £130,102 (GBP)
Organisation Atomic Weapons Establishment 
Sector Private
Country United Kingdom
Start 10/2012 
End 01/2016
 
Description Collaboration on advanced software parallelisation methods 
Organisation NVIDIA
Country Global 
Sector Private 
PI Contribution This Salford led collaboration involved the implementation and testing (using PRACE tier-1 supercomputer time and NVidea in-kind support via prototype TESLA box provision) of GPU enabled, muliti-core algorithmic skeletons for poly-CINS calculation implementation.
Collaborator Contribution NCCC (academic responsible - H. Gonzalez-Velez) provided core tier one access (with D. L. Roach) and NVidea providing a (then) prototype TESLA box (GPU card) to test GPU enabling. Work done by Gonzalez-Velez and Garba (still based at RGU, being supervised by HGV and DLR) on full parallelisation of code was then tested using NCCC equipment.
Impact Two publications. Two pieces of software. Multi-disciplinary - computer science and computational physics (neutron)
Start Year 2011
 
Description Collaboration on advanced software parallelisation methods 
Organisation University College Cork
Country Ireland 
Sector Academic/University 
PI Contribution This Salford led collaboration involved the implementation and testing (using PRACE tier-1 supercomputer time and NVidea in-kind support via prototype TESLA box provision) of GPU enabled, muliti-core algorithmic skeletons for poly-CINS calculation implementation.
Collaborator Contribution NCCC (academic responsible - H. Gonzalez-Velez) provided core tier one access (with D. L. Roach) and NVidea providing a (then) prototype TESLA box (GPU card) to test GPU enabling. Work done by Gonzalez-Velez and Garba (still based at RGU, being supervised by HGV and DLR) on full parallelisation of code was then tested using NCCC equipment.
Impact Two publications. Two pieces of software. Multi-disciplinary - computer science and computational physics (neutron)
Start Year 2011
 
Description Collaboration on fitting and software development 
Organisation Curtin University
Country Australia 
Sector Academic/University 
PI Contribution This Salford - led collaboration was responsible for the building, development and testing of the SCATTER and PREFIT software packages. Salford team designed, wrote and developed both software packages (in collaboration with partners), and applied the software to the development of the poly-CINS spectroscopic method.
Collaborator Contribution Curtin - 1 PhD student (and time provided by collaborating academic J. D. Gale) on fitting and model construction and development for the magnesium hydride system (and backplane methods developed to make this system accessible) RGU - 1 PhD student (and time provided by collaborating academic H. Gonzalez-Velez) on large scale parallelisation (multi-core, GPU and cloud enabled) of the SCATTER code and GULP's internal eigensolve routines.
Impact Multi-disciplinary (computational chemistry, materials science, physics (neutron) and computer science (advanced parallelisation and visualisation)
 
Description Collaboration on fitting and software development 
Organisation Robert Gordon University
Country United Kingdom 
Sector Academic/University 
PI Contribution This Salford - led collaboration was responsible for the building, development and testing of the SCATTER and PREFIT software packages. Salford team designed, wrote and developed both software packages (in collaboration with partners), and applied the software to the development of the poly-CINS spectroscopic method.
Collaborator Contribution Curtin - 1 PhD student (and time provided by collaborating academic J. D. Gale) on fitting and model construction and development for the magnesium hydride system (and backplane methods developed to make this system accessible) RGU - 1 PhD student (and time provided by collaborating academic H. Gonzalez-Velez) on large scale parallelisation (multi-core, GPU and cloud enabled) of the SCATTER code and GULP's internal eigensolve routines.
Impact Multi-disciplinary (computational chemistry, materials science, physics (neutron) and computer science (advanced parallelisation and visualisation)
 
Description Collaboration on software engineering and novel spectroscopic methods 
Organisation California Institute of Technology
Country United States 
Sector Academic/University 
PI Contribution D. L. Roach invited to participate and contribute to the DANSE consortium (NSF funded project). Roach invited to speak and work with Caltech scientists on the poly-CINS methodology and software methods
Collaborator Contribution Insight into powder spectroscopic methods developed at Salford invited to, and contributed to, national workshop on computational simulation modelling (NSF funded meeting)
Impact Report - workflows for computational scattering science (2013) (contributing author)
Start Year 2009
 
Title PreFiT - A poly-CINS analysis and workflow toolbox 
Description PreFiT (written by Garba and Roach) was designed to provide labour saving automation of the time-consuming processes used in poly-CINS. Using PreFiT, a user may view and compare experimental and theoretical neutron scattering data, perform edge detection, build fitting files for use in GULP and many other powerful and time-saving activities (including edge detection, signal processing and intensity filtering functionality). Months of manual analysis can now be accomplished in hours - all packaged in a user friendly and intuitive visual interface. 
Type Of Technology Software 
Year Produced 2014 
Impact None yet. 
 
Title SCATTER 
Description SCATTER is the poly-CINS simulation kernel for the method. Embedded into the GULP software, it uses all the functionality of this popular lattice/molecular dynamics code to general one-phonon S(Q,w) for coherent and incoherent scattering in powders. Options include a variety of output and pre-analysis options to assist the user to generate appropriate neutron scattering data for comparision and fitting to experimental data. Interfaces with the PreFiT analysis toolbox. 
Type Of Technology Software 
Year Produced 2013 
Open Source License? Yes  
Impact None yet.