Investigating electron dynamics and radiation transport in solid-density plasmas using X-ray FELs

Lead Research Organisation: University of Oxford
Department Name: Oxford Physics

Abstract

The advent of 4th generation light sources - X-ray free-electron lasers (FELs) - is revolutionising the way we investigate matter in extreme conditions by providing ultra-bright, femtosecond, nearly monochromatic X-rays at tuneable photon energies from the XUV to the hard X-ray spectral region. When focused to micron-sized spots, intensities exceeding 10^17 W/cm^3 can be generated at X-ray wavelengths for the first time. We showed recently that such high intensities are sufficient to heat solid systems to temperatures of several million Kelvin within a few tens of femtoseconds, i.e., to temperature and density conditions similar to those found half way into the centre of the Sun, thus paving the way to novel investigations of extreme states of matter of broad interest to astrophysics, planetary science, inertial confinement fusion research, and national security applications.

Alongside generating hot-dense plasmas, intense X-ray interactions with matter give rise to a well-controlled source of non-thermal (hot) electrons which are generated either directly by photoionization, or via inner-shell atomic recombination processes such as Auger decay. Because ponderomotive energies are negligible at X-ray wavelengths, these are the only 'hot' electrons generated during the irradiation, leading to a non-thermal electron distribution that can be controlled directly by modulating the X-ray wavelength and intensity, and is also far simpler to model theoretically than hot electrons produced in intense optical laser-plasma interactions.

In this project we aim to use these unique characteristics of X-ray FEL pulses to experimentally create a tailored non-thermal electron distribution within a hot-dense plasma, and track its evolution and equilibration dynamics on ultra-fast timescales. These measurements will not only provide some of the first measurements of electron-electron collisionality in strongly-coupled systems, but will also more broadly assess the validity of the Coulomb Logarithm framework commonly used to model a wide range of electron interaction processes, including bremsstrahlung emission, conductivity, thermal transport and stopping power.

Importantly, we note that the irradiation of solid samples with intense X-ray light allows us to reach the temperature-density conditions corresponding to the radiation/convection zone boundary of the Sun. By using our recently developed spectroscopic techniques we aim to investigate radiation transport and the opacity of low and mid-Z elements in these extreme conditions, and determine whether the opacity can help address the outstanding disagreement between solar models and the internal structure of the Sun determined by helio-seismic observations. Accurate independent measurements of the Fe opacity in this regime are particularly of interest given the recent experimental results from Bailey et al. (Nature 517, 56, 2015), showing a significant deviation in the experimental opacity from that predicted by plasma opacity models for lower density plasmas.

Planned Impact

Our proposed research project will focus on investigating the fundamental aspects of ultra-fast electron dynamics in high-energy density systems and, as explained in the academic beneficiaries section, will primarily impact various fields of academic research. There are, however, several other areas where we envisage our work will have a significant impact.

Firstly, we note that X-ray-matter interactions and the evolution of high-density plasmas on femtosecond timescales is of particular importance to research in structural biology using coherent diffraction imaging techniques on FELs, including serial femtosecond crystallography, X-ray holography and single-molecule imaging, since electron collisional dynamics is one of the most important processes contributing to damage on femtosecond timescales. Here we will investigate some of these dynamical processes for the first time, which will provide unique new insight into electron damage processes and help improve and develop new computational models to be used across the biological imaging community. The capability to image biological systems on ultra-short time scales, and how they fold or interact with light, is of great practical interest not only to structural biologists but also the medical and pharmaceutical industries more broadly, and the successful development of appropriate techniques and the accurate, predictive computational modelling of samples irradiated by intense X-rays will be key to fully exploit the range of future opportunities. We envisage parts of this project could become eventually a standalone computational package with a broad user base and, with limited additional support, be made commercially viable, aimed at a range of public and private enterprise.

An understanding of the physics of dense plasmas is further of considerable interest and importance to MoD and AWE in the context of the CTBT, and the current nuclear security environment that relies on simulation, and thus the underpinning physics models. The Oxford group has strong links with AWE via the OxCHEDS institute (the Oxford University Centre for High Energy Density Science) and will continue to engage with AWE to ensure they are aware in a timely manner of the results of the proposed experiments.

Importantly, we are keen to promote our research more widely making full use of opportunities for public engagement, including writing for the Oxford Science Blog, working with press offices at Oxford and from abroad to produce press releases publicising our research results as widely as possible, and actively engaging with the popular scientific press (our work has in the past been picked up often and has featured in, among others, New Scientist, La Recherche, National Geographic and Scientific American) and with national broadcasters.

Finally, the proposed research will provide a truly exceptional opportunity for young researchers (both at the postdoctoral and graduate student level) to work on billion-dollar flagship science facilities worldwide, and to collaborate internationally with a wide range of world-leading researchers, universities and laboratories. The training of scientists on the newly developed X-ray FEL facilities will not only provide unique opportunities for personal and professional growth, but is also strategically important to enable the UK community more widely to make full use of future opportunities arising from the large investments the UK has in the European XFEL project: alongside direct contributions to the project the UK is (co-)funding dedicated beam lines for serial femtosecond crystallography (SFX) and high energy density-science (HED) via the HiBEF consortium.

Publications

10 25 50
 
Description We have developed a novel way to investigate the electronic structure of dense plasmas using resonant inelastic x-ray scattering (PRL 125, 195001).

We have developed a novel approach to using resonant spectroscopy to investigate ultra-fast collisional ionization rates (PRL 120 055002).

We have developed a novel approach to measuring the x-ray opacity of matter in extreme conditions using a bright x-ray FEL pulse. By irradiating a series of thin foils we are able to reconstruct both he opacity and emissivity of radiating systems using the equations for radiative transfer (PRL 119 085001).

We have concluded the first stage of software development of a non-LTE atomic kinetics code with configuration-level atomic data which can run on mid-Z elements (CCFLY). Developments to integrate non-thermal electron processes are ongoing.

We have developed and benchmarked an optimized a spectroscopic software code to compute emission and absorption spectra from the output of the atomic kinetics module on-the-fly. First application of these results have now been published (Scientific Reports 8 6276).

We have developed new deep-learning tools to accelerate simulations by over a billion times. This approach, called DENSE, provides fast emulators that are sufficiently accurate for use in data statistics to treat inverse problem instabilities, and will have a transformative effect on our modelling and data-interpretation capabilities.
Exploitation Route Our scattering approach opens new avenues in the study of electronic structure of extreme states of matter, and provides new opportunities to measure temperatures in shocked systems with very high signal levels, suitable for use in dynamic diffraction experiments.

Our opacity method provides a new way to investigate opacities without requiring broad-band x-ray sources.

We have developed one of the best-validated and fastest atomic kinetics packages capable of treating time-dependent external x-ray fields. This development is already being applied to the interpretation of several experimental datasets in collaboration with our international partners.
Sectors Aerospace, Defence and Marine,Digital/Communication/Information Technologies (including Software),Education,Security and Diplomacy,Other

 
Description Our software development and work on computational tools and technique, including those based on machine learning, have led to the founding of an Oxford University spinout company, Machine Discovery Ltd. This company is already successfully licensing Oxford University IP, generated from research funded from this grant, to industrial companies and to educational institutions around the world. The company has raised over £1.5m in private funding to date, has opened a subsidiary company in California, and employs 8 people on a full time basis as of early 2023, with a further 4 employed on a part-time basis. The company plans to double in size over the next 12-18 months, substantially expanding its core R&D presence in Oxfordshire.
First Year Of Impact 2019
Sector Aerospace, Defence and Marine,Digital/Communication/Information Technologies (including Software),Education,Electronics,Energy,Financial Services, and Management Consultancy,Manufacturing, including Industrial Biotechology
Impact Types Economic

 
Description Data-driven discovery science using ultra-bright x-ray light-sources
Amount £463,621 (GBP)
Organisation The Royal Society 
Sector Charity/Non Profit
Country United Kingdom
Start 10/2019 
End 10/2022
 
Description Exploring quantum high energy density matter via Resonant Inelastic X-ray Scattering
Amount £51,135 (GBP)
Funding ID 2444668 
Organisation Science and Technologies Facilities Council (STFC) 
Sector Public
Country United Kingdom
Start 09/2020 
End 09/2024
 
Description International Exchanges Grant, The Royal Society
Amount £11,900 (GBP)
Funding ID IE161574 
Organisation The Royal Society 
Sector Charity/Non Profit
Country United Kingdom
Start 04/2017 
End 03/2019
 
Description Investigating matter in extreme conditions via self-backlighting x-ray spectroscopy on the National Ignition Facility
Amount £100,281 (GBP)
Organisation The Royal Society 
Sector Charity/Non Profit
Country United Kingdom
Start 03/2018 
End 03/2021
 
Description Novel optimization framework for real-time automated radiation therapy
Amount £59,140 (GBP)
Funding ID ST/S002197/1 
Organisation Science and Technologies Facilities Council (STFC) 
Sector Public
Country United Kingdom
Start 10/2018 
End 03/2019
 
Description European Cooperation in Science and Technology (COST) Action CA17126 - Towards understanding and modelling intense electronic excitation 
Organisation Queen's University Belfast
Country United Kingdom 
Sector Academic/University 
PI Contribution This is an EU-funded action to organise meetings and fund collaboration across the continent. It started in 2019 and will last for 4 years. Our team has presented research at the first action meeting, and is building a collaborative network.
Collaborator Contribution Obtained funding that will support all meeting costs and travel for exchanges between collaborating partners. Importantly, this includes students who are entirely excluded from our EPSRC grant so this will be the main funding stream to enable them to interact with partners across Europe and integrate in the research.
Impact No research outcomes yet.
Start Year 2019
 
Description European Cooperation in Science and Technology (COST) Action CA17126 - Towards understanding and modelling intense electronic excitation 
Organisation Technical University of Madrid
Country Spain 
Sector Academic/University 
PI Contribution This is an EU-funded action to organise meetings and fund collaboration across the continent. It started in 2019 and will last for 4 years. Our team has presented research at the first action meeting, and is building a collaborative network.
Collaborator Contribution Obtained funding that will support all meeting costs and travel for exchanges between collaborating partners. Importantly, this includes students who are entirely excluded from our EPSRC grant so this will be the main funding stream to enable them to interact with partners across Europe and integrate in the research.
Impact No research outcomes yet.
Start Year 2019
 
Title DENSE - deep neural architecture search 
Description Computer simulations are invaluable tools for scientific discovery. However, accurate simulations are often expensive and slow to execute, which limits their applicability to extensive parameter exploration, online diagnostics, and uncertainty quantification. A promising route to accelerate simulations is to build fast emulators, however, to date these have shown limited accuracy, severely restricting applications. DENSE presents a robust method to build fast and accurate emulators to accelerate simulations based on neural architecture search. The method successfully accelerates simulations by up to 70 million times in 8 scientific cases including astrophysics, high energy density physics, inertial confinement fusion, magnetic fusion energy, and seismology, using the same super-architecture, algorithm, and hyperparameters. The constructed DENSE emulators achieve 14 times lower loss function values than non-deep learning techniques and obtain better results than manually-designed deep neural networks in most cases. The emulators are sufficiently accurate to solve inverse problems, taking mere seconds compared with hours or days when using simulations. Our approach also inherently provides emulator uncertainty estimation, adding further confidence in their use. The work will accelerate development of accurate online diagnostics, allow more extensive parameters exploration, and enable new, previously unfeasible computational discovery. 
IP Reference  
Protection Patent application published
Year Protection Granted 2020
Licensed Yes
Impact Achieved acceleration in various modelling application including that of climate aerosols of up to 2 billion times.
 
Title Maleo - an optimization framework for distributed computing 
Description We have developed an optimization software framework Maleo that integrates with any simulation software to run on local and/or remote computers simultaneously. The software enables convenient and fast adaptation of advanced optimization algorithms to complex simulation software problems, and automates most tasks that need to be performed for the statistical analysis of data. 
IP Reference  
Protection Patent application published
Year Protection Granted 2018
Licensed Yes
Impact Enables advanced computational and statistical tasks to be undertaken by scientists and researchers without requiring them to have to know the detailed computational algorithms needed for optimization and Bayesian statistical sampling. Speeds up and automates simulation-based discovery while at the same time providing a robust treatment of uncertainties in yielding predictions. Technology has been licensed to a University startup, and is already used in several University-based and industrial settings in Europe and the US.
 
Title ?-torch: differentiable scientific computing library 
Description Physics-informed learning has shown to have a better generalization than learning without physical priors. However, training physics-informed deep neural networks requires some aspect of physical simulations to be written in a differentiable manner. Unfortunately, some operations and functionals commonly used in physical simulations are scattered, hard to integrate, and lack higher order derivatives which are needed in physical simulations. In this work, we present ?-torch, a library of differentiable functionals for scientific simulations. Example functionals are a root finder and an initial value problem solver, among others. The gradient of functionals in ?-torch are written based on their analytical expression to improve numerical stability and reduce memory requirements. ?-torch also provides second and higher order derivatives of the functionals which are rarely available in existing packages. We show two applications of this library in optimizing parameters in physics simulations. 
Type Of Technology Software 
Year Produced 2021 
Open Source License? Yes  
Impact Key component of the Differentiable DFT development. 
URL https://arxiv.org/abs/2010.01921
 
Title CCFLY 
Description Highly-optimized non-LTE collisional radiative code for intense x-ray interactions with matter 
Type Of Technology Software 
Year Produced 2018 
Impact Enables the full treatment of mid-Z elements at the level of atomic configurations on small-scale hardware. 
 
Title DQC: A Python program package for differentiable quantum chemistry 
Description Fully differentiable density functional theory package for computational quantum chemistry. 
Type Of Technology Software 
Year Produced 2021 
Open Source License? Yes  
Impact Core package for two papers to date (Physical Review Letters 127, 126403 (2021), Journal of Chemical Physics 156, 084801 (2022)). 
URL https://github.com/diffqc/dqc
 
Title Deep Emulator Network SEarch (DENSE) 
Description A new approach to building high-fidelity deep-learning emulators for complex simulation software packages. 
Type Of Technology Software 
Year Produced 2020 
Impact In the process of being licensed via the commercial arm of Oxford University. Extensively used in research and in the process of being submitted for publication in peer-review literature. 
 
Title Maleo: optimisation framework in distributed systems 
Description Optimisation framework for treating inverse problems making full use of distributed computing systems. 
Type Of Technology Software 
Year Produced 2018 
Impact Work in progress. Several papers have been submitted for publication making use of this tool, and the software package has been filed for patent protection (Patent Application GB1804154.1). The capabilities allowed us to successfully apply for STFC funding to apply methods developed for treating plasma physics problems to the automatisation of treatment planning in radiation therapy (STFC grant #ST/S002197/1). 
 
Title Software for "Building high accuracy emulators for scientific simulations with deep neural architecture search" 
Description This is the code and datasets for "Building high accuracy emulators for scientific simulations with deep neural architecture search". 
Type Of Technology Software 
Year Produced 2020 
Open Source License? Yes  
URL https://zenodo.org/record/3782843
 
Title Software for "Building high accuracy emulators for scientific simulations with deep neural architecture search" 
Description This is the code and datasets for "Building high accuracy emulators for scientific simulations with deep neural architecture search". 
Type Of Technology Software 
Year Produced 2020 
Open Source License? Yes  
URL https://zenodo.org/record/3782842
 
Company Name MACHINE DISCOVERY LIMITED 
Description Development and commercialisation of quantum physics, acceleration and optimisation software for intelligent computational R&D. 
Year Established 2019 
Impact License contracts of Oxford-generated IP to both industry and research institutions.
Website https://machine-discovery.com