Research

OMDB, Material Informatics

Using Machine Learning to study materials like never before
Image is not available

The organic materials database is an open access electronic structure database for 3-dimensional organic crystals, developed and hosted Nordita. It provides tools for search queries based on data-mining and machine learning techniques. The universal features provided on our web interface facilitate the design of novel functional organic materials with a wide-range of applications.

ODD-FREQUENCY

Unravelling new hidden dynamical orders
Image is not available

Most theories of superconductivity consider the limit of equal-time pairing, however, electrons may also be paired at unequal times which leads to superconducting gaps that can be odd in time or, equivalently, frequency.

MATERIAL INFORMATICS

Finding new materials with Machine Learning
Image is not available

Machine learning has entered the field of quantum matter with applications covering quantum materials and the many-body problem. Interpretable and computationally-efficient machine learning models are able to capture the structure-property relationship in materials science.

OPTICS - DIRAC MEDIA

Detecting dark matter with Dirac media
Image is not available

Dirac materials have a wide range of applications and have even been proposed as a sensor material for dark matter particles. The energy range accessible by the small gap in massive Dirac cones provides the sensitivity to search in the sub-MeV mass scale. The small gap filters out the background thermal noise while still capturing excitations due to dark matter

OPTICS - DIRAC MEDIA

Harvesting the power of light in Dirac media
Image is not available

The linearity of the spectrum around the Dirac points leads to unusual physics. Using external electromagnetic fields, the response of Dirac media can generate currents of nontrivial harmonic composition.

previous arrowprevious arrow
next arrownext arrow
Slider

Research Areas

Material Informatics

Over the past decades there has been an an enormous increase of computational power and a rapid development of experimental techniques. Both developments, together with the great advancements of data storage capacities, have initiated the application of methods taken from computer and data science into the research of functional quantum materials and quantum many-body physics. For example, interpretable and computationally-efficient machine learning models are able to capture the structure-property relationship in materials science opening the path towards an efficient computer based materials design.

In supervised learning, large data sets, e.g., of ab initio calculations, provide the necessary training examples. The trained models facilitate high-throughput screening of materials by reducing the search space. Additionally, the models enable dynamic simulation on longer timescales than traditionally feasible. Unsupervised clustering approaches using structural similarity metrics allow for a new way of exploring the large chemical space. In case of the many body problem, machine learning architectures provide versatile wavefunctions that lead to accurate results and prove to be more flexible than traditional methods.

Work in the group has focused on developing data mining and machine learning techniques to investigate and analyse the calculated properties of organic materials. This is closely linked with the development of the Organic Materials Database (OMDB).

  • OMDB: The Organic Materials Database (OMDB) contains DFT data for over 25,000 organic crystals that have been synthesised and provides the data set for the group’s material informatics work.
  • Magnetism: Magnetic properties are of interest from both fundamental and application perspectives; atomistic spin dynamics protocols are being used to extend the DFT data of the OMDB.
  • Machine learning: machine learning techniques are being developed to predict potentially interesting materials based on the data contained in the OMDB.

The distinguishing feature of a quantum material is that a classical macroscopic description is unsatisfactory and the bulk properties can only be obtained by invoking quantum mechanics at higher energies and larger length scales than usually required for describing the fundamental interactions between isolated electrons and atoms. Examples of large scale quantum effects include phenomena such as superconductivity and superfluidity and quantum effects are responsible for the unique features that occur in materials identified as topological insulators and spin ices among others. Dynamic quantum matter refers to any type of matter that requires both a quantum mechanical and time-dependent treatment.

Most traditional theoretical approaches to describing materials assume that the system in question is held in equilibrium, which simplifies calculations by removing all time-dependence from the system. Dynamical aspects have to be considered when one or more of external driving (excitation), internal dynamics, inherent fluctuations or dissipation to an environment are included. In addition, other systems that include or allow unusual behaviour under reordering or variation of time labels can also require dynamical factors to be included in any discussion.

Excitation is often realised in experiments, either as continuous or pulsed protocols, and will be crucial in the vast majority of potential applications. Internal dynamics are always present and energy relaxation protocols are used to find the ground state in equilibrium studies. Dissipation can be difficult to include since this is often experimentally uncontrolled; a Markovian heat bath is an analytically tractable example. Analysing the quantum or thermal fluctuations that occur near a phase transition help understand which interactions are important (e.g.: via renormalisation group), but may also have their own signatures.

Research in the group considers three realisations of dynamical quantum matter:

  • Dynamical multiferroicity: The Dzyaloshiksii-Moriya interaction leads to an effective polarisation from a static spin spiral; exploiting the duality between electricity and magnetism in Maxwell’s equations, a complementary effect – Dynamical Multiferroicity – is possible.

  • Pumped Dirac materials: Pump-probe experiments have demonstrated that Dirac materials may be driven into transient excited states with two chemical potentials, one for the electrons and one for the holes, which effectively offers control of the strength of the Coulomb interaction.

  • Odd frequency superconductivity: Most theories of superconductivity consider the limit of equal-time pairing, however, electrons may also be paired at unequal times which leads to superconducting gaps that can be odd in time or, equivalently, frequency.