To get regular updates on all our news and events direct to your inbox
Sign up to our newsletter here
PhD candidate Hongyang Dong and Finden research scientists have developed a regression CNN that performs full profile analysis of powder diffraction data yielding physicochemical information (scale factors, lattice parameters and crystallite size) from multiphase systems. This project was performed in collaboration with National Physical Laboratory, STFC Scientific Machine Learning Group and UCL Department of Chemistry.
The work has resulted in a paper “A deep convolutional neural network for real-time full profile analysis of big powder diffraction data” published in NPJ Computational Materials 7, 74 (2021).
We present Parameter Quantification Network (PQ-Net), a regression deep convolutional neural network providing quantitative analysis of powder X-ray diffraction patterns from multi-phase systems. The network is tested against simulated and experimental datasets of increasing complexity with the last one being an X-ray diffraction computed tomography dataset of a multi-phase Ni-Pd/CeO2-ZrO2/Al2O3 catalytic material system consisting of ca. 20,000 diffraction patterns. It is shown that the network predicts accurate scale factor, lattice parameter and crystallite size maps for all phases, which are comparable to those obtained through full profile analysis using the Rietveld method, also providing a reliable uncertainty measure on the results. The main advantage of PQ-Net is its ability to yield these results orders of magnitude faster showing its potential as a tool for real-time diffraction data analysis during in situ/operando experiments.
You can read the full paper at https://doi.org/10.1038/s41524-021-00542-4
Our Research and Development Lead Scientist Antony Vamvakeros gave a webinar on chemical tomography and neural networks on the 20th May hosted by the Neel Institut
Synchrotron X-ray chemical tomography methods combine a scattering or spectroscopic technique with a tomographic data acquisition approach. These non-destructive methods yield a cross-section of the studied sample where each pixel in the reconstructed images corresponds to a chemical signal (e.g. X-ray diffraction pattern or spectrum). These spatially-resolved signals most often reveal information that is lost in conventional bulk measurements. In this talk, he briefly introduces X-ray diffraction computed tomography (XRD-CT) and presents a few examples where we have applied such methods to track the evolving solid-state chemistry of complex functional materials and devices under operating conditions. In the second part of the webinar, he focuses on our latest technical advances regarding processing and analysis of these large and rich chemical imaging datasets using deep learning methods with neural networks.
Watch a recording of the webinar below:
A 5D diffraction imaging experiment (with 3D spatial, 1D time/imposed operating conditions and 1D scattering signal) was performed with a Ni-Pd/CeO2-ZrO2/Al2O3 catalyst. The catalyst was investigated during both activation and partial oxidation of methane (POX). The spatio-temporal resolved diffraction data allowed us to obtain unprecedented insight into the behaviour and fate of the various metal and metal oxide species and how this is affected by the heterogeneity across catalyst particles. We show firstly, how Pd promotion although facilitating Ni reduction, over time leads to formation of unstable Ni-Pd metallic alloy, rendering the impact of Pd beyond the initial reduction less important. Furthermore, in the core of the particles, where the metallic Ni is primarily supported on Al2O3, poor resistance towards coke deposition was observed. We identified that this preceded via the formation of an active yet metastable interstitial solid solution of Ni-C and led to the exclusive formation of graphitic carbon, the only polymorph of coke observed. In contrast, at the outermost part of the catalyst particle, where Ni is predominantly supported on CeO2-ZrO2, the graphite formation was mitigated but sintering of Ni crystallites was more severe.
Read the full article at https://doi.org/10.1039/D1TA01464A.
We are very pleased to be featured in ESRF Highlights 2020 – their annual round-up of the most exciting experiments carried out at the ESRF in 2020. Read more on page 164 at http://www.esrf.eu/Apache_files/Highlights/2020/index.html#/page/166
We are looking forward to working with the University of Sheffield on the new Faraday Industry Fellowship, a collaborative energy storage research project.
This is an innovative programme that strengthens ties between battery researchers working in industry and academia.
Each fellowship enables academics and industrialists to undertake a mutually beneficial, electrochemical energy storage research project that aims to solve a critical industrial problem and that has the potential for near- and longer-term benefit to the wider UK battery industry.
We will be working to deepen the understanding of new cathode materials. The aim is to fast track the best-performing high energy density cathodes to aid their early adoption by UK industry
Read all about it at https://faraday.ac.uk/ind-fellowship-feb2021/
A team of scientists from Finden Ltd (Dr Stephen Price, Dr Simon Jacques and Prof Andrew Beale) in collaboration with Infineum UK Ltd (Dr Nathan Hollingsworth, Dr Matthew Irving) have used IMAT to help understand where and how coking occurs on engine components. This understanding will help develop more durable, fuel efficient lubricants.
Read the case study at https://www.isis.stfc.ac.uk/Pages/IMAT-Coking-of-engine-components.aspx
From 1st December 2020 to 31st January 2021 scientists and engineers who work as researchers at a university or other research institution can apply for one of ten places at the CAROTS STARTUP SCHOOL. Everyone with an idea for a new Scientific Service Company based on an advanced analytical technique, for example at a large-scale research infrastructure such as a synchrotron or a neutron source or in collaboration with a university, is welcome to apply. A place at the STARTUP SCHOOL includes individual coaching sessions with some of Europe’s leading CEOs of Scientific Service Companies as well as a webinar programme teaching everything worth knowing to take the jump from scientist to entrepreneur. Participants will also get the opportunity to join a European network of likeminded people and successful scientific service companies. Online 1 to 1-coachings and monthly webinars will start in March 2021 through to June 2021.
More info: http://carots.eu/startup_school
Our scientists’ new work on finding a solution to the parallax problem in X-ray scattering/diffraction experiments has been published in a new paper, “DLSR: a solution to the parallax artefact in X‐ray diffraction computed tomography data,” published in the Journal of Applied Crystallography.
The work was performed in collaboration with the creator of the TOPAS software Alan Coelho, DESY, UCL Chemistry, ESRF and SciML.
A new tomographic reconstruction algorithm is presented, termed direct least‐squares reconstruction (DLSR), which solves the well known parallax problem in X‐ray‐scattering‐based experiments. The parallax artefact arises from relatively large samples where X‐rays, scattered from a scattering angle 2gθ, arrive at multiple detector elements. This phenomenon leads to loss of physico‐chemical information associated with diffraction peak shape and position (i.e. altering the calculated crystallite size and lattice parameter values, respectively) and is currently the major barrier to investigating samples and devices at the centimetre level (scale‐up problem). The accuracy of the DLSR algorithm has been tested against simulated and experimental X‐ray diffraction computed tomography data using the TOPAS software.
This will allow upscaling chemical tomography techniques to study large samples.
Read the full article at https://doi.org/10.1107/S1600576720013576
The research was carried out in collaboration with Dr Thomas Heenan from the Electrochemical Innovation Lab (EIL) at UCL Chemical Engineering, UCL Chemistry and ESRF.
The solid oxide fuel cell (SOFC) anode is often composed of nickel (Ni) and yttria-stabilized zirconia (YSZ). The yttria is added in small quantities (e.g., 8 mol %) to maintain the crystallographic structure throughout the operating temperatures (e.g., room-temperature to >800 °C). The YSZ skeleton provides a constraining structural support that inhibits degradation mechanisms such as Ni agglomeration and thermal expansion miss-match between the anode and electrolyte layers. Within this structure, the Ni is deposited in the oxide form and then reduced during start-up; however, exposure to oxygen (e.g., during gasket failure) readily re-oxidizes the Ni back to NiO, impeding electrochemical performance and introducing complex structural stresses. In this work, we correlate lab-based X-ray computed tomography using zone plate focusing optics, with X-ray synchrotron diffraction computed tomography to explore the crystal structure of a partially re-oxidized Ni/NiO-YSZ electrode. These state-of-the-art techniques expose several novel findings: non-isotropic YSZ lattice distributions; the presence of monoclinic zirconia around the oxidation boundary; and metallic strain complications in the presence of variable yttria content. This work provides evidence that the reduction–oxidation processes may destabilize the YSZ structure, producing monoclinic zirconia and microscopic YSZ strain, which has implications upon the electrode’s mechanical integrity and thus lifetime of the SOFC.
Read the article at https://www.mdpi.com/2073-4352/10/10/941
Our latest work on the characterisation of NMC electodes used in Li-ion batteries at the PCCP has been published in a new paper, “Exploring cycling induced crystallographic change in NMC with X-ray diffraction computed tomography” in the journal Physical Chemistry Chemical Physics.
The research was carried out in collaboration with the Electrochemical Innovation Lab (EIL) from the UCL Chemical Engineering, Johnson Matthey, the Faraday Institution, NREL, UCL Chemistry and ESRF.
This study presents the application of X-ray diffraction computed tomography for the first time to analyze the crystal dimensions of LiNi0.33Mn0.33Co0.33O2 electrodes cycled to 4.2 and 4.7 V in full cells with graphite as negative electrodes at 1 μm spatial resolution to determine the change in unit cell dimensions as a result of electrochemical cycling. The nature of the technique permits the spatial localization of the diffraction information in 3D and mapping of heterogeneities from the electrode to the particle level. An overall decrease of 0.4% and 0.6% was observed for the unit cell volume after 100 cycles for the electrodes cycled to 4.2 and 4.7 V. Additionally, focused ion beam-scanning electron microscope cross-sections indicate extensive particle cracking as a function of upper cut-off voltage, further confirming that severe cycling stresses exacerbate degradation. Finally, the technique facilitates the detection of parts of the electrode that have inhomogeneous lattice parameters that deviate from the bulk of the sample, further highlighting the effectiveness of the technique as a diagnostic tool, bridging the gap between crystal structure and electrochemical performance.
Read the full article at https://doi.org/10.1039/D0CP01851A
For further news and events follow us on Twitter at @FindenLtd