Category Archives: Research spotlight

Image of a chiral phonon in the moire cell

Chiral valley phonons and flat phonon bands in moiré materials

I. Maity, A. A. Mostofi and J. Lischner, Chiral valley phonons and flat phonon bands in moiré materials, Phys. Rev. B Letters 105, L041408 (2022).

Chirality–the characteristic of an object that can be distinguished from its mirror image–plays a fundamental role in understanding many phenomena in physics, chemistry, and biology. Recently, the study of chirality in phonons has attracted significant interest. The interaction of chiral phonons with other quasiparticles present in a crystal is relevant for understanding and controlling many electronic and optical phenomena.

A particularly promising platform for observing and manipulating chiral phonons are twisted bilayers of two-dimensional (2D) materials. Since the discovery of flat electronic bands in twisted bilayer graphene, such moiré materials have served as a rich playground for investigating the properties of correlated electrons, excitons and phonons. In this work, we demonstrate that phonons in moiré materials can also be chiral. Focusing on twisted WSe2 at twist angles close to 60 degress, we find two sets of emergent chiral moiré valley phonons that originate from symmetry breaking at the moiré scale. We also discover flat chiral phonon bands in the energy gap between the acoustic and optical phonon modes, which, similar to the flat electronic bands in this system, occur over a range of twist angles and without the requirement of a magic angle.

Our findings, which are expected to be generic for moiré systems composed of two-dimensional materials that break inversion symmetry, are relevant for understanding electron-phonon and exciton-phonon
scattering, and also for the design of phononic analogues of flat band electrons.

Bandstructure of a GNR

Electrically Induced Dirac Fermions in Graphene Nanoribbons

M. Pizzochero, N. V. Tepliakov, A. A. Mostofi and E. Kaxiras, Electrically Induced Dirac Fermions in Graphene Nanoribbons, Nano Letters 21, 9332 (2021).

Graphene nanoribbons (GNRs) are few-nanometer wide strips of graphene. Unlike graphene, graphene nanoribbons have a band-gap, while still maintaining high electron and hole mobility. This makes them promising candidates for beyond-silicon nanoelectronic devices.

In this work we used first-principles calculations and simpler tight-binding models to explore whether and to what degree the properties of graphene nanoribbons can be controlled by an electric field. This is of crucial importance for developing electronic devices based on graphene nanoribbons. We showed that applying an electric field across the width of the nanoribbon modulates the size of the band-gap, and can be used to turn a normally semiconducting nanoribbon into a semimetal. The electrically-induced semimetallic phase features zero-energy massless Dirac fermions, analogous to graphene. The electric field can be applied externally, i.e., by contacting the edges of the nanoribbon to source and drain electrodes, or internally, by incorporating doping elements of opposite polarity in the opposite edges of the nanoribbon. These findings generalize to other group-IV nanoribbons, including recently fabricated silicene nanoribbons, which hold great promise due to their potential for integration into current silicon electronics.

This work was part of the PhD work of Nikita Tepliakov and was in collaboration with colleagues Michele Pizzochero and Efthimios Kaxiras at Harvard University.

Computation of optical absorption spectra of molecular crystals: the case of the polymorphs of ROY

J. C. A. Prentice and A. A. Mostofi, Accurate and efficient computation of optical absorption spectra of molecular crystals: the case of the polymorphs of ROY, Journal of Chemical Theory and Computation 17, 5214 (2021)

Molecular crystals are a class of solids that have applications in organic LEDs, lasers, and solar cells, and optical absorption is among their most important properties. The optical absorption behaviour can be strongly affected by changes in the conformation or packing of the molecule, leading to phenomena such as colour polymorphism, where the same molecule can form different crystal structures that have different colours. 

In this work, we present a ‘crystal spectral warping’ method that allows the accurate and efficient prediction of optical absorption spectra of molecular crystals, including both long-range interactions, and the accuracy of hybrid TDDFT calculations. We have applied this approach to the well-known colour polymorphic system ROY, and show that our results give a better description of experimental observations than previous attempts.

This research was done by Joseph Prentice, who was a post-doctoral research associate funded by EPSRC (EP/P02209X/1).

Bandstructure of twisted bilayer graphene

Hartree theory calculations of quasiparticle properties in twisted bilayer graphene

Z. A. H. Goodwin, V. Vitale, X. Liang, A. A. Mostofi and J. Lischner, Hartree theory calculations of quasiparticle properties in twisted bilayer graphene, Electronic Structure 2, 034001 (2020)

Twisted bilayer graphene (tBLG) has generated a tremendous amount of interest in the condensed matter physics community (read our earlier research spotlights on the twist angle dependence of electron correlations in tBLG and the critical role of device geometry on the phase diagram of tBLG). Recent tunnelling experiments have found signatures corresponding to correlated insulator states as well as a number of other observations that cannot be understood within the framework of non-interacting electrons.

In a paper appearing in Physical Review B, we study the behaviour of interacting electrons in twisted bilayer graphene near the magic angle using atomistic Hartree theory. In particular, we calculate band structures and (local) densities of states in the metallic state as function of doping for a variety of twist angles. Our results explain several features of the experimentally observed tunnelling spectra. We also provide a simple parametrisation of the Hartree potential which enables Hartree band structures to be calculated without the need for self-consistency.

This work was part of Zachary Goodwin’s PhD in the Centre for Doctoral Training in Theory and Simulation of Materials. Zachary is supervised by Johannes Lischner and Arash Mostofi.

Transport eigenchannel between two carbon nanotubes

J. Chem. Phys. Special Topic Issue on Electronic Structure Software

Prentice et al., The ONETEP linear-scaling density functional theory program, J. Chem. Phys. 152, 174111 (2020)

Oliveira et al., The CECAM electronic structure library and the modular software development paradigm, J. Chem. Phys. 153, 024117 (2020)

In the summer of 2020 the Journal of Chemical Physics published a special issue on the topic of electronic structure software. This issue highlights the excellent ecosystem of electronic structure codes that are developed in the community and includes papers on some very well-established codes as well as more recent ones. It is these advances in electronic structure software, hand-in-hand with developments in computer architecture and compute capacity, that drive our progress towards a predictive, first-principles-based understanding of materials and molecular systems.

We have contributed to two papers in this special topic issue. The first paper is on the ONETEP software package, which is a linear-scaling density-functional theory code that is able to bring to bear the predictive power of first-principles simulations on systems that are typically well beyond the capabilities of conventional approaches, such as the plane-wave pseudopotential method. The second paper is about the CECAM Electronic Structure Library (ESL). The ESL was founded to catalyse a transition away from the traditional “monolithic” model of electronic structure software, towards a more modular structure that incorporates interoperable and reusable modules and libraries that are open-source and available to the general community. The aim is that this will lead to better optimisation, more agile code development, and lower barriers to entry for new developers. One of the other codes developed in our group, Wannier90, and the subject of an earlier research spotlight, is a paradigmatic example of the ESL philosophy. You can also read the special AIP Scilight on the ESL effort.

bandstructure from automated wannierisation

Automated high-throughput wannierisation

V. Vitale, G. Pizzi, A. Marrazzo, J. R. Yates, N. Marzari and A. A. Mostofi, Automated high-throughput wannierisation, npj Computational Materials 6, 66 (2020).

Maximally-localised Wannier functions (MLWFs) are routinely used to compute a wide range of advanced materials properties, including highly accurate band structures, gyrotropic effects, spin Hall effects, thermoelectric properties and electron-phonon interactions. They are also an important scale-bridging tool for transferring the accuracy and transferability of first-principles methods at atomic scales to tight-binding models that can be used for mesoscale simulations (read our review article in Reviews of Modern Physics for more details).

The impact of MLWFs on electronic structure research is exemplified by the widespread usage of the Wannier90 code for generating and using MLWFs, which is interfaced to almost every major electronic structure code and is used by thousands of researchers worldwide.

In a paper appearing in npj Computational Materials, we present a general automated high-throughput framework for computing MLWFs and using them to compute material properties. Our approach is based on the recently proposed selected columns of the density matrix algorithm. The process does not require any a priori chemical intuition about the system being studied, and is therefore ideally suited to high-throughput calculations. Our workflows have been implemented within the AiiDA framework, and the full dataset of ~200 materials as well as the full provenance of how to generate the data shown in this work are provided via a publicly available Virtual Machine that can be downloaded from the Materials Cloud. To the best of our knowledge, this is the first time that such a level of reproducibility has been made available in the field of first-principles electronic structure simulations.

This work was part of Dr Valerio Vitale’s postdoctoral work as part of the European Union’s Centre of Excellence E-CAM, supervised by Jonathan Yates and Arash Mostofi.

carbon nanotube polymer composite model

Atomistic QM/MM simulations of the strength of covalent interfaces in carbon nanotube-polymer composites

J. R. Golebiowski, J. R. Kermode, P. D. Haynes and A. A. Mostofi, Atomistic QM/MM simulations of the strength of covalent interfaces in carbon nanotube-polymer composites, Phys. Chem. Chem. Phys. 22, 12007 (2020).

There is a “tyranny of scales” associated with the theoretical prediction of complex processes in materials and molecular systems. Often there is a small “active region” that governs the physics and chemistry of the phenomenon being investigated, but this region interacts and is influenced by the rest of the system, the “environment”, which is generally much larger and structurally more complex. Carbon nanotube-polymer composites are one such material. They are strong and lightweight with applications in demanding areas of industry such as ballistic protection and aerospace. Properties of the interface between carbon nanotubes (CNT) and polymers are a crucial factor affecting critical mechanical failure in these materials and chemical functionalisation of the CNTs can be used to promote strong interfaces based on CNT-polymer covalent crosslinks.

Developing a fundamental understanding of interfacial failure in CNT-polymer composites is challenging because the bond-breaking processes at the polymer-CNT attachment point that initiate failure are quantum-mechanical in nature, yet the mechanisms by which stresses are transferred through the disordered polymer occur on length-scales far in excess of anything that can be simulated quantum-mechanically.

In a paper appearing in Phys. Chem. Chem. Phys., we address the multi-scale nature of critical failure using a concurrent hybrid method that couples a quantum mechanical description of the bond-breaking process at the interface to a classical description of the remainder of the system. Using this “QM/MM” approach, which we first proposed and validated in an earlier work, we investigated the effects of different interfacial chemistries on mechanical strength and mechanisms mechanical failure in model CNT-polyethylene composites by simulating a CNT pull-out experiment. The predictions from our simulations can help to guide further experimental work with the aim of accelerating the design of high-performance CNT-polymer composites.

This work was part of Jacek Golebiowski’s PhD in the Centre for Doctoral Training in Theory and Simulation of Materials. Jacek was supervised by Peter Haynes, James Kermode and Arash Mostofi.

twisted bilayer graphene moire pattern

Critical role of device geometry for the phase diagram of twisted bilayer graphene

Z. A. H. Goodwin, V. Vitale, F. Corsetti, D. K. Efetov, A. A. Mostofi and J. Lischner, Critical role of device geometry for the phase diagram of twisted bilayer graphene, Phys. Rev. B 101, 165110 (2020).

Twisted bilayer graphene (tBLG) is made of two stacked sheets of graphene that are rotated with respect to one another. The relative twist of the two layers produces a moiré pattern that forms a superlattice whose repeat length is much greater than the carbon-carbon bond length of the individual layers. For small twist angles around a “magic angle” of 1.1°, the behaviour of electrons in the moiré superlattice is fundamentally different to that of electrons in graphene or in untwisted bilayers, and tBLG is seen to exhibit unexpected insulating and superconducting states, in addition to the expected metallic state.

Crucially, in experiments, the tBLG is not free-standing, but is encapsulated by a thin dielectric medium which, in turn, has metallic contacts. This is extremely important, since it determines the way in which electrons in tBLG interact with one another. In a paper appearing in Physical Review B, we develop and use an electronic model of tBLG within such a device configuration. Our model captures the influence of the device geometry on the competition between phases. In particular, we show that the correlated insulator phase can be switched on or off by controlling the thickness of the dielectric layers separating tBLG from the metallic gates.

The competition between the different phases observed in tBLG makes it a rich problem, and understanding what influences it has the potential to advance our understanding of the physics of strongly-correlated electrons and unconventional superconductors.

This work was part of Zachary Goodwin’s PhD in the Centre for Doctoral Training in Theory and Simulation of Materials. Zachary is supervised by Johannes Lischner and Arash Mostofi.

Wannier function in diamond

Wannier90 as a community code: new features and applications

G. Pizzi et al., Wannier90 as a community code: new features and applications, J. Phys.: Condens. Matter 32, 165902 (2020)

Wannier functions provide a local real-space description of electronic structure in materials and molecular systems. They are used, for example, to understand chemical bonding and as a compact and highly computationally efficient basis in which to compute electronic properties (see our review article in Reviews of Modern Physics for more details). Indeed, they have become an indispensable element in the tool-kit of contemporary materials and molecular computational science.

The Wannier90 code, an open-source software program that is developed in our research group, is the most prevalent code for computing maximally-localised Wannier functions. It is a paradigmatic example of interoperable software and, at the present time, has been interfaced to over 25 electronic structure codes and post-processing software tools.

In a paper that appears in the Journal of Physics: Condensed Matter we describe the recent evolution of Wannier90 from a code developed by a small group of developers to one in which enhancements in functionality are contributed by a wide, global community of developers. The paper describes how this transition has been achieved and some of the software enhancements it has enabled for the latest v3.x release.

This work was part of Dr Valerio Vitale’s postdoctoral work as part of the European Union’s Centre of Excellence E-CAM, supervised by Jonathan Yates and Arash Mostofi.

pentacene molecule in p-terphenyl molecular crystal

Combining embedded mean-field theory with linear-scaling density-functional theory

J. C. A. Prentice, R. J. Charlton, A. A. Mostofi and P. D. Haynes, Combining embedded mean-field theory with linear-scaling density-functional theory, Journal of Chemical Theory and Computation 16, 354 (2020)

Quantum embedding methods are a vital tool in the study of complex molecular or extended systems, where the most interesting physics or chemistry is associated with a particular region, but the rest of the system still interacts significantly with this “active region”, in a way that can only be described using quantum mechanics. Using an accurate quantum mechanical method (e.g., hybrid density functional theory (DFT)) may be necessary to correctly predict the properties of the active region, but using this level of theory for the entire system may be prohibitively expensive. Instead, quantum embedding schemes enable the active region to be treated at a high level of theory, whilst simultaneously treating the environment at a lower, less computationally demanding level of theory (e.g. semi-local DFT). The result is a calculation that has the accuracy of the high-level theory, but at a fraction of the cost.

In a paper published in the Journal of Chemical Theory and Computation, we present a novel implementation of embedded mean field theory (EMFT) into the linear-scaling DFT code ONETEP, enabling efficient hybrid DFT-in-semi-local DFT quantum embedding calculations on large-scale systems. We have applied our implementation to a range of systems, including pentacene, pentacene-doped crystalline p-terphenyl, and metallocene-doped carbon nanotubes, and show it provides excellent results compared to full system hybrid DFT calculations and experiment. In particular, in pentacene-doped p-terphenyl, the ability to include a much larger p-terphenyl environment than previous calculations, whilst still using hybrid DFT for the pentacene dopant itself, gives much better agreement with experimental measurements of the shift in the S0 to S1 excitation due to the presence of the p-terphenyl host. This calculation contained nearly 3000 atoms, an order of magnitude larger than any calculation attempted before with EMFT.

This work was done as part of Dr Joseph Prentice’s postdoctoral research funded by the EPSRC (grant no. EP/P02209X/1) and Robert Charlton’s PhD in the Centre for Doctoral Training in Theory and Simulation of Materials. The project was supervised by Peter Haynes and Arash Mostofi.