Non-ergodic dynamics and topological-sector fluctuations in layered high-temperature superconductors

Lead Research Organisation: University of Bristol
Department Name: Physics

Abstract

At low enough temperatures, the constituent electrons of certain materials flow as a single body with zero electrical resistance. This is called superconductivity. The behaviour was first measured in solid mercury, which superconducts at around -270C and is therefore classed as a low-temperature superconductor. Certain copper-oxide-based materials, however, can superconduct at much higher temperatures: up to -130C. These materials therefore belong to the separate group known as high-temperature superconductors. This group of materials have extremely complex multi-layered crystal structures that are difficult to model, meaning that a theory of high-temperature superconductivity remains one of the major unsolved problems in condensed-matter physics.

At any given temperature, a superconductor will either be in its normal or superconducting state. Recent experiments on copper-oxide-based materials measured large fluctuations in their electrical resistances at the transition temperature between these two states. The large fluctuations are a result of the complex structures of the materials: a theoretical model for this phenomenon will therefore uncover details of these structures and drive the research community towards a complete theory of high-temperature superconductivity. This will lead to advances in the myriad engineering applications of superconductivity, which include superconductor-based quantum computing, magnetic resonance imaging, particle confinement in synchrotrons such as the Large Hadron Collider, plasma confinement in fusion reactors, and superconducting quantum interference devices used for high-precision magnetic measurements in medicine and further afield.

Planned Impact

In the short term, this research will mostly impact academic researchers, but also the public through my continued commitment to public engagement and outreach. Descriptions of my research to young people facing socio-economic disadvantage inspires them to remain in STEM education, which will benefit both their lives and careers, and also the wider public through a better educated workforce and the pooling of talent from different demographics.

Further public engagement will also benefit the wider public more directly. We invest in scientific research to drive long-term economic growth, but equally important is the cultural enrichment of science and science education in creating a well-informed and inspired public. This improves the quality of life and creative output of those touched by enrichment, and also helps to create a workforce that is educated, highly skilled and inspired to improve. Society as a whole will therefore benefit from my research and public engagement as this fosters both the economic competitiveness of the UK and economic performance on a global level.

In the long term, this research will contribute to a theory of high-temperature superconductivity. It is of course too soon to tell if such a theory could lead to the engineering of room-temperature superconducting devices. If this were to prove to be the case, however, it is feasible that we could undergo an energy revolution in which we could generate clean energy and store and transport it with almost perfect efficiency. The project therefore also addresses both Energy Storage and Materials for Energy Applications. Superconductors may also lead to the first accessible quantum computer, which would constitute an information revolution. The high magnetic fields expelled by superconducting materials are currently used in magnetic resonance imaging, nuclear magnetic resonance, and particle accelerators such as the Large Hadron Collider. Superconducting Josephson junctions are used to form superconducting quantum interference devices (SQUIDs), which we can use for high-precision magnetic measurements in many different systems, from detecting neural activity in the brain and small magnetic fields in the heart, to oil prospecting, mineral exploration and even the detection of gravitational waves. New high-temperature superconducting devices could make all of these applications more cost effective, which will benefit industrial organisations such as Oxford Instruments and Siemens, and society as a whole as the applications become more widely available. The research will therefore also contribute to Quantum Physics for New Quantum Technologies.

Further to this, topological physics could lead to quantum technologies that we cannot currently predict, but it is again feasible that such devices may contribute to energy and information revolutions. Finally, irreversible Markov chain Monte Carlo algorithms may be the next big step change in the Bayesian modelling of large data sets in computational statistics. Big Data is becoming increasingly relevant in our data-driven world; such step changes could bring large-scale changes to our analysis of the ever-larger data sets to which we are gaining access. Properly regulated, this could vastly improve the quality of life of the human race by increasing the efficiency of everyday systems. The specific algorithms I propose also have applications in Soft Matter Physics and Biophysics, which are both rapidly growing fields of applied and fundamental science research.

Publications

10 25 50
publication icon
Faulkner MF (2018) All-atom computations with irreversible Markov chains. in The Journal of chemical physics

publication icon
Höllmer P (2020) JeLLyFysh-Version1.0 - a Python application for all-atom event-chain Monte Carlo in Computer Physics Communications

publication icon
Livingstone S (2019) Kinetic energy choice in Hamiltonian/hybrid Monte Carlo in Biometrika

 
Description PLEASE NOTE that I was not able to add my two most articles to the Publications sections. This may be because they are preprints under peer review.

Simulating electrolytes using event-chain Monte Carlo is possible: has so far led to two papers, along with an accompanying Python application for numerically stable (with zero tuning parameters) electrolyte simulation of 3D soft-matter systems using event-chain Monte Carlo. This application is called JeLLyFysh and is hosted publicly on Github at https://github.com/jellyfysh/JeLLyFysh

With my statistics colleagues, I published a paper that details a Hamiltonian Monte Carlo algorithm (a type of statistical-sampling algorithm) that is numerically stable (with one tuning parameter) in the tails of both light- and heavy-tailed probability distributions in Bayesian computation. We have named this the super-relativistic Monte Carlo algorithm, and, since the start of this grant, have created an accompanying Python application called super-aLby. We have extended super-aLby such that it applies to statistical-physics models in soft-matter physics and may publish this if we have time. super-aLby is hosted publicly on Github at https://github.com/michaelfaulkner/super-aLby

JeLLyFysh and super-aLby are important because the current industry standard for the simulation of models of electrolytes and other soft-matter fluids (the molecular dynamics algorithm) is numerically unstable unless it advances very slowly through configuration space; in contrast, JeLLyFysh and super-aLby are both numerically stable (one with zero tuning parameters, the other with one tuning parameter) and converge on the target distribution with machine precision. In the case of JeLLyFysh, it is completely stable and can therefore advance through configuration space using much larger (effective) time steps, allowing access to the long timescales of rare-event physics (we aim to simulate rare protein-folding events).

Much progress has also been made on strongly nonergodic autocorrelations at the superconducting transition in the superconducting film. I discovered a general symmetry breaking at the Berezinskii-Kosterlitz-Thouless (BKT) phase transition. This explains the low-temperature superconducting state, resolving a 50-year-old paradox between theory and experiment. A single-author paper outlining the result is currently under peer review at Nature Communications. This paper has led to the understanding that the nonergodic autocorrelations must be due to a critical slowing down at the BKT transition, as the experiments have all the hallmarks of critical slowing down, and this phenomenon is a consequence (at the transition) of the low-temperature symmetry breaking. This is now the focus of my future work. In addition, I'm currently writing a review paper outlining the general symmetry breaking in more detail, and how it relates to the topological nonergodicity of the low-temperature phase that I'd previously discovered and which was cited (by the superconductivity experimentalists) as a base explanation for the nonergodic autocorrelations. The experimentalists and I are currently devising experiments to characterise the critical slowing down in the superconducting films - this would be an excellent complement to the simulations that I plan with a PhD student.

The symmetry-breaking result also predicts a memory timescales in colloidal films and 'XY' magnetic films, the latter of which may be useful in engineering, e.g., magnetic memory storage devices. With the relevant specialist experimentalists, I am currently discussing experiments to demonstrate the symmetry breaking in magnetic and colloidal films. The code for this project is my open-source Fortran-Python application, which is hosted publicly on Github at https://github.com/michaelfaulkner/xy-type-models

My collaborator in Bayesian computation (a subset of data science) and I have written an article on Sampling algorithms in statistical physics. This elucidates statistical physics for the data scientist, and also provides some important proofs of results in statistics and statistical physics. The article also outlines the importance of symmetry breaking and critical slowing down in physics, and which modern algorithms circumvent the problems posed by the latter phenomenon. This has led to ideas regarding critical slowing down in epidemic modelling - harnessing our unique partnership of shared knowledge. We plan to do this in parallel with the physics application (of critical slowing down) explained above. I have also had three separate invitations to talk about this work, including by the Applied Probability section of the Royal Statistical Society.

Please note that progress was slowed by the PI's (me) significant health problems, including invasive hospital treatment in 2022. My employer and I arranged, however, a manageable reduced-hours contract - this has undoubtedly led to the very pleasing results that meet the principal goals of the fellowship.
Exploitation Route Our Python application for ECMC electrolyte simulation is open source. We expect it to be both widely used and contributed to in/by the soft-matter community. We plan to eventually integrate models of proteins and polyelectrolyte batteries, to provide the technology to electrochemistry and biophysics; this is a long-term goal, however.

super-aLby is open source. We expect super-aLby (and indeed the super-relativistic Monte Carlo algorithm) to be used in the data-science community. It also has the potential to be used widely in the soft-matter community.

The symmetry breaking has inspired plans for experiments on the other 2D materials (e.g., magnetic and colloidal films) and potentially on engineering (e.g., magnetic-film memory storage). The results have also inspired the experimental and theoretical projects on critical slowing down. Moreover, in conjunction with the paper on Sampling algorithms in statistical physics, this has inspired ideas regarding critical slowing down in epidemic modelling. These two seemingly unrelated sub-projects have therefore led to multiple exciting and relevant projects across the interface of the two fields.
Sectors Chemicals,Digital/Communication/Information Technologies (including Software),Electronics,Energy,Manufacturing, including Industrial Biotechology,Pharmaceuticals and Medical Biotechnology

URL https://github.com/michaelfaulkner
 
Description We created an open-source Python package for machine-precision electrolyte simulation, which we expect to be used in industry (as well as in academia).
First Year Of Impact 2019
Impact Types Cultural

 
Title Event-chain Monte Carlo in soft-matter physics 
Description We designed an algorithm for event-chain Monte Carlo simulation of atomistic fluid models in statistical physics, biophysics and electrochemistry. We expect this to challenge the state of the art (molecular dynamics) in rare-event physics such as protein folding as its total numerical stability results in it reaching much longer timescales than molecular dynamics, which is forced to crawl through configuration space due to stability issues. We created an accompanying Python-C application called JeLLyFysh. This is hosted publicly on Github at https://github.com/jellyfysh/JeLLyFysh 
Type Of Material Improvements to research infrastructure 
Year Produced 2018 
Provided To Others? Yes  
Impact We created an accompanying Python-C application called JeLLyFysh. This is hosted publicly on Github at https://github.com/jellyfysh/JeLLyFysh We have also published two papers on the subject (https://doi.org/10.1063/1.5036638 and https://doi.org/10.1016/j.cpc.2020.107168) which have a total of 27 citations to date. 
URL https://doi.org/10.1063/1.5036638
 
Description Team JeLLyFysh 
Organisation ESPCI ParisTech
Country France 
Sector Academic/University 
PI Contribution I am the sole member of my EPSRC research team. I worked on Project JeLLyFysh with Prof. Werner Krauth of ENS and his doctoral student Liang Qin, along with Philipp Hoellmer who worked as an intern and Master's student for a year, at / linked to ENS. And Prof. Tony Maggs of ESPCI is the fifth member of our team. We all worked equally on the project, when available. I was unfortunately ill from October 2018--October 2019. Philipp Hoellmer joined us in the summer of 2018. We wrote two papers together and created an open-source Python package for machine-precision electrolyte simulation, which we called JeLLyFysh.
Collaborator Contribution I am the sole member of my EPSRC research team. I worked on Project JeLLyFysh with Prof. Werner Krauth of ENS and his doctoral student Liang Qin, along with Philipp Hoellmer who worked as an intern and Master's student for a year, at / linked to ENS. And Prof. Tony Maggs of ESPCI is the fifth member of our team. We all worked equally on the project, when available. I was unfortunately ill from October 2018--October 2019. Philipp Hoellmer joined us in the summer of 2018. We wrote two papers together and created an open-source Python package for machine-precision electrolyte simulation, which we called JeLLyFysh.
Impact We wrote two papers together and created an open-source Python package for machine-precision electrolyte simulation, which we called JeLLyFysh. This project covers soft-matter physics, biophysics, electrochemistry, and physical chemistry.
Start Year 2017
 
Description Team JeLLyFysh 
Organisation École Normale Supérieure, Paris
Country France 
Sector Academic/University 
PI Contribution I am the sole member of my EPSRC research team. I worked on Project JeLLyFysh with Prof. Werner Krauth of ENS and his doctoral student Liang Qin, along with Philipp Hoellmer who worked as an intern and Master's student for a year, at / linked to ENS. And Prof. Tony Maggs of ESPCI is the fifth member of our team. We all worked equally on the project, when available. I was unfortunately ill from October 2018--October 2019. Philipp Hoellmer joined us in the summer of 2018. We wrote two papers together and created an open-source Python package for machine-precision electrolyte simulation, which we called JeLLyFysh.
Collaborator Contribution I am the sole member of my EPSRC research team. I worked on Project JeLLyFysh with Prof. Werner Krauth of ENS and his doctoral student Liang Qin, along with Philipp Hoellmer who worked as an intern and Master's student for a year, at / linked to ENS. And Prof. Tony Maggs of ESPCI is the fifth member of our team. We all worked equally on the project, when available. I was unfortunately ill from October 2018--October 2019. Philipp Hoellmer joined us in the summer of 2018. We wrote two papers together and created an open-source Python package for machine-precision electrolyte simulation, which we called JeLLyFysh.
Impact We wrote two papers together and created an open-source Python package for machine-precision electrolyte simulation, which we called JeLLyFysh. This project covers soft-matter physics, biophysics, electrochemistry, and physical chemistry.
Start Year 2017
 
Title JeLLyFysh - a Python application for all-atom event-chain Monte Carlo 
Description JeLLyFysh is an open-source Python-C application for the event-chain Monte Carlo (an event-driven irreversible Markov-chain Monte Carlo algorithm) simulation of classical N-body simulations in statistical mechanics, biophysics and electrochemistry. The application's architecture mirrors the mathematical formulation of event-chain Monte Carlo. Local potentials, long-range Coulomb interactions and multi-body bending potentials are covered, as well as bounding potentials and cell systems including the cell-veto algorithm. 
Type Of Technology Software 
Year Produced 2019 
Open Source License? Yes  
Impact Its two accompanying papers (https://doi.org/10.1063/1.5036638 and https://doi.org/10.1016/j.cpc.2020.107168) have a total of 27 citations to date. 
URL https://github.com/jellyfysh/JeLLyFysh
 
Title super-aLby -- a Python application for super-relativistic Monte Carlo in Bayesian computation and statistical physics 
Description super-aLby is an open-source Python application that implements the super-relativistic Monte Carlo algorithm for the simulation of Bayesian probability models and classical N-body soft-matter models in statistical physics. For a closely connected discussion of kinetic-energy choice in Hamiltonian/hybrid Monte Carlo, see https://doi.org/10.1093/biomet/asz013, where we first introduced super-relativistic Monte Carlo (though we did not name it). 
Type Of Technology Software 
Year Produced 2021 
Open Source License? Yes  
Impact Its accompanying paper (https://doi.org/10.1093/biomet/asz013) has a total of 38 citations to date. 
URL https://github.com/michaelfaulkner/super-aLby
 
Title xy-type-models 
Description xy-type-models is an open-source Fortran-Python application that implements the event-chain and Metropolis Monte Carlo algorithms for the simulation of two-dimensional XY-type models in statistical physics. 
Type Of Technology Software 
Year Produced 2021 
Open Source License? Yes  
Impact This software was used in https://doi.org/10.1103/PhysRevB.91.155412 and https://doi.org/10.1088/1361-648X/aa523f, which have a total of 15 citations to date. It is also currently being used in two further papers associated with this grant. 
URL https://github.com/michaelfaulkner/xy-type-models