Since the project start in 2015, we have conducted one-to-one interviews with decision makers and technical R&D personnel to gain an even deeper understanding of industry needs and requirements to inform the development of the NOMAD Laboratory and our Sustainability Plan.
You can read the full set of interviews from the companies below:
- Dr. Suchismita Sanyal, Shell
- Dr. Ansgar Schäfer, BASF
- Dr. Pedro Martínez, Iberdrola
- Mr. Julio Bono Pérez, ISFOC
- Dr. Michael Krein, Lockheed Martin
- Dr. Victor Milman, BIOVIA
- Ms. Angela McKane, BP
- Dr. James Goddin, Granta Design Ltd
- Dr. Teodosio del Caño, Onyx Solar
- Dr. Yolanda de Miguel, TECNALIA
- Dr. Moritz to Baben, GTT-Technologies
- Dr. Sebastián Echeverri Restrepo, SKF Research & Technology
- Dr. Carina Arasa Cid, Elsevier
- Dr. Filippo Zuliani, TataSteel
- Dr. Antonio Sichirollo, Greenetica
- Dr. Sebastián Echeverri Restrepo, SKF Resesach & Technology
- Dr. Misbah Sarwar and Dr. Crispin Cooper, Johnson Matthey
Industry Case Studies
Overview: The NOMAD Laboratory CoE provides a huge amount of computational data already available in the NOMAD Repository, and also offers to perform new, high-quality calculations for materials where important information is missing in the NOMAD database. Here, we are carefully listening to suggestions from our industrial colleagues. An area of interest noted by various industrial colleagues is the storage, release and conversion of CO2 to valuable chemicals and fuels, such as methanol or syngas. A use case for this is being developed jointly by NOMAD groups at the Institut de Química Teòrica i Computacional de la Universitat de Barcelona (IQTCUB) and the Fritz Haber Institute of the Max Planck Society in Berlin.
Interest in the case of CO2 conversion has been shown in particular by I-deals, a company coordinating the Methanol fuel from CO2 (MefCO2) project, which aims to develop an innovative, green chemical production technology that contributes significantly to European objectives to decrease CO2 emissions and increase renewable energy usage. The four-year MefCO2 project has a budget of over 11 million Euros, and will wrap up in late 2018. The MefCO2 consortium are impressed by results from the IQTCUB group on CO2 sequestration and activation by transition metal carbides and from computational and experimental studies showing excellent CO2 conversion to methanol on Cu supported on MoC and on Mo2C.
Results to date: Applying the newly developed compressed-sensing and subgroup-discovery approaches, the team has identified physically interpretable ab initio descriptors for energy and structure of CO2 adsorbed at binary and ternary oxide surfaces. The descriptors include only properties of involved atomic species, bulk materials, and clean surfaces. We have shown that, contrary to the standard understanding, the O-C-O bending angle does not correlate well with the charge transferred to CO2 for the whole data set. However, the subgroup discovery identified a subset of surfaces for which this correlation is accurate. This subset is characterized by a more ionic character of the bonding between surface cations and oxygen.
The NOMAD team is working with BP on producing data and developing post-processing analytic tools for a tribological application relevant for lubricants R&D (and more generally, for any industrially-hosted modelling of hydrocarbon-based fluids). The produced data will be uploaded on the NOMAD repository, and all analytic tools will be made available by the NOMAD CoE.
Understanding and controlling the fundamental mechanism underlying lubrication is of very significant industrial interest. Confining hydrocarbons in high shear and pressure conditions produces a number of complex and intriguing phenomena. This study is based on the assumption that molecular dynamics simulations can yield valuable insight into these phenomena thus enabling a better fundamental understanding of the physical behaviour of lubricants. Predicting the viscosity of mixtures (complex lubricant compositions) on the basis of the results of MD calculation is also expected to require machine learning-based tools that industry R&D users would be unlikely to develop in isolation, but could easily take on using after appropriate training.
The NOMAD CoE is teaming up with the Finnish software company Nanolayers on applying machine learning techniques on materials data, to help produce new and better catalyst which make no use of “critical” transition metal elements. Key rare metals like palladium and platinum are found in many industrial catalysts, and replacing them with inexpensive, earth-abundant materials is of high economic and political interest.
Teaming up through its Aalto partner with a sister H2020-NMP project ("CritCat") that will produce Density Functional Theory data, NOMAD will support Nanolayers, a company specialized in the development of machine learning approaches for challenges in the physical sciences. The goal of this collaboration is to apply NOMAD’s tools in Nanolayer’s projects, exploring whether and how NOMAD’s infrastructure for storage and high throughput search, augmented by NOMAD’s machine learning-based analytic tools can offer solutions to Nanolayers’ clients’ problems.
Check out the recent related publication from the NOMAD team.
The NOMAD team is helping organic chemists to identify new molecules with optimised energy levels and light absorption for polymer solar cells (PSC). The INKA project sponsored by the Danish Innovation Fund aims at pushing the commercialisation of polymer solar cells by developing robust, cost-effective inks for the photoactive layer of the PCS.
The NOMAD researchers at DTU are performing an extensive computational screening of donor-acceptor polymers to map out their frontier energy levels and light absorption properties. Machine learning algorithms are used to identify trends and correlations in the data and for efficient property predictions. The database is now complete - available here. The team has published a paper where machine learning (after training on the database) was used to predict new molecules for polymer solar cells (Machine learning-based screening of complex molecules for polymer solar cells, The Journal of Chemical Physics 148, 241735 (2018); 10.1063/1.5023563). Experimentalists are now trying to synthesise some of the molecules that were identified in the computational study.
NOMAD is co-developing the Imeall set of tools for addressing the atomistic properties of grain boundaries in metals, starting from bcc Fe. This will provide a programmable application interface to interatomic potential calculators as well as a database of atomistic structures, thus helping research and development materials engineers working in metallurgy with a general resource that can be used to query the chemo-mechanical properties of grain boundaries.
Given the atomistic grain boundary structures stored in the NOMAD archive, the users will be able to define and mount a desired database subset, index its structures, and use the provided software tools to perform extended analysis of grain boundary-specific engineering properties, e.g. connected with plasticity and failure.
Industry-Relevant Use Cases
NOMAD will start soon a series of high-quality calculations for materials where important information is missing in the Repository and Archive. This follows the development of a novel thermal-conductivity computational approach which for the first time enables accurate calculations for materials from very low to very high thermal conductivity. Systematic calculations of heat-transport tensors for many materials will be started soon.
Computational 2D Materials Database: Electronic Structure of Transition-Metal Dichalcogenides and Oxides
NOMAD researchers at the Technical University of Denmark, Department of Physics (DTU) have conducted a systematic study of the electronic properties (band structures, absolute band off-sets, effective masses, dielectric functions) of around 50 different monolayer transition metal dichalcogenides and oxides. These materials constitute an important class of the family of two-dimensional (2D) materials which is expanding rapidly following the discovery of graphene. The calculated properties are stored in the open Computational Materials Repository (CMR) and will also be uploaded to the NOMAD Repository.
High-throughput Screening and Statistical Learning For the Design of Transparent Oxide Materials
NOMAD researchers at the Fritz Haber Institute of the Max Planck Society have been investigating transparent oxide metals. They have undertaken high-throughput screening of ternary and quaternary group-III oxides with the formula of (AlxGayInZ)2O3 as n-type (anode) transparent conducting oxides by the effective mass (for conductivity) and band gap (for optical transparency).
Evaluating the (dis)similarity of crystalline, disordered and molecular compounds
Evaluating the (dis)similarity of crystalline, disordered and molecular compounds is a critical step in the development of algorithms to navigate automatically the configuration space of complex materials. For instance, a structural similarity metric is crucial for classifying structural patterns, searching chemical databases for better compounds and materials, and driving the next generation of machine-learning techniques for predicting the stability and properties of molecules and materials. In the last few years, several strategies have been designed to compare local atomic structures. In particular, the smooth overlap of atomic positions (SOAP) has emerged as an elegant framework to obtain translation, rotation and permutation-invariant descriptors of the local structure of atomic neighbourhoods. NOMAD researchers at the University of Cambridge have shown how to combine these local descriptors to quantify the similarity of both whole molecular and bulk periodic structures, introducing powerful metrics that enable the navigation of alchemical and structural complexities within a unified framework.