Dr. Sukanya Ghosh

Present Research

My present research is primarily based on magnetism and spintronics in low dimensions. My recent studies
involve a deep understanding of novel and exotic properties arising in quantum materials. For example, I am
interested in phenomena related to spin-orbit coupling, electron correlation effects, topological spin textures, magnetic transitions, performing first-principles calculations in conjunction with Monte Carlo and atomistic spin dynamics simulations. So far, my research on magnetism is focused on two-dimensional (2D) van-der-Waals (vdW) materials. The 2D vdW magnets I have explored till now are CrI3 and FenGeTe2 (n=3, 4, 5). CrI3 is a magnetic semiconductor with a low transition temperature (TC ) ∼ 45-65 K. While the FenGeTe2 systems are metallic with TC close to the room temperature.

a. CrI3
i) My research on 2D magnetic semiconductor CrI3 mainly involves how to tune its physical properties by
applying external perturbations, e.g., external electric field, mechanical strain and charge doping. I investigated the magnetoelectric and magnetostructural effects present in the monolayer and bilayer CrI3.

ii) My study unveils the importance of spin-orbit coupling present in CrI 3 which governs the magnetic ground state attained by this system, causing the asymmetry in magnetic transition induced by electron doping, not with hole doping. I proposed a technique, i.e., the application of a moderate compressive vertical strain ( ∼ 5%) lifts such asymmetry by causing magnetic transition from antiferromagnetic (AFM) to ferromagnetic (FM) for both hole and electron doping.

iii) I investigated the reciprocal-space spin textures exhibited by these systems and the microscopic mechanisms responsible for such features. My study finds the in-plane spins of CrI 3 bilayer in AFM configuration shows Rashba-type texture due to the presence of internal electric field caused by the breaking of inversion symmetry.
iv) My work on CrI3 are published in the following articles:Physica B: Condensed Matter 570, 166 (2019);
Applied Physics Letters 116, 086101 (2020); Nanoscale 13, 9391, 2021; Journal of Physics and Chemistry of Solids 173, 111100 (2023); arXiv 2206.10777 [cond-mat.mtrl-sci] (2022).

b. FenGeTe2
i ) My research on 2D metallic magnet, Fe5GeTe2 was motivated by peculiar experimental results on magnetism and structures. After careful electronic structure calculations, extraction of interatomic exchange interactions and Monte Carlo simulations, I found a reconstructed structural form of the monolayer and its consequence on magnetic exchange interactions and resulting magnetic structures. The presence of a special Fe species was verified which gives rise to significant non-collinear magnetic configurations. This result is published in J. Phys. Chem. Lett. 13, 4877 (2022).

ii) I have done a systematic study of FenGeTe2 (n =3,4,5) to elucidate the role of electron correlation in
determining electronic structure, magnetic interactions, spectral properties, effective mass and Curie
temperatures. Accurate calculations of various types of exchange interactions are computed with standard
density functional theory (DFT), DFT+U and DFT+DMFT methods. My research has found that dynamical
mean field theory (DMFT) yields results very in a very good agreement with the experimental ones. These
results are published in the following paper: npj Computational Materials 9, 86 (2023).

iii) My research on 2D Fe5GeTe2 showed how to tune ferromagnetic transition or T C in this system beyond room-temperature by substitutional doping with Ni. Here I have mainly investigated the interplay of structure, magnetism and electron correlation in achieving ferromagnetism beyond room temperature. My study finds with 20% Ni doping, it is possible to achieve TC ∼ 400 K. The magnetic properties of Fe5−xNixGeTe2 (x: doping concentration) monolayer agree quite well with the experimental observations of bulk Fe 5−xNixGeTe2 . This work is presently available in preprint form: arXiv 2305.04366v1 (2023).

iv) In experimental collaboration with Chalmers University of Technology, Sweden, I investigated room-
temperature spintronic properties of Fe5GeTe2 investigated at interface with graphene, which could be used as a spin-valve device. Performing first-principles calculations in combination with Monte Carlo simulations I found significantly canted Fe magnetic moments in Fe5GeTe2 at T~300 K, in agreement with experimental observations. This results are published in Advanced Materials 35, 2209113 (2023). 


PhD Research

During PhD, I have worked in the area of density functional theory calculations, applied to problems in
materials science. I worked on diverse projects; what most of them had in common was the desire
to find a ‘descriptor’ for various properties of materials. Descriptors are quantities that correlate well
with properties of materials, but are much quicker to evaluate than performing a density functional
theory calculation or experiment. Theses descriptors can thus be used to rapidly screen a large number
of systems and shortlist a few candidates that are most likely to possess target properties that fall within
a desired range that is determined by a specific application that one has in mind.

Though the use of descriptors in materials science has a long history, nowadays they have attracted
fresh attention because of the possibility of developing them using machine learning (ML) techniques.
Though I did not actually use any ML in my thesis work, the approach used was influenced
by the philosophy and techniques that can be found in recent ML studies.

In one of my work, I showed that it is possible to develop simple yet successful descriptors for the
architecture of self-assembled host-guest structures of organic molecules on surfaces. These descriptors
depend only on the form and geometry of the organic molecules in the gas phase, and hence can be
evaluated at almost zero computational cost. Interestingly, I showed that it was possible to develop
such descriptors by training on extremely small data sets. Given the small size of the data sets, no ML
was necessary, and simple regression techniques sufficed, combined with intuition. The form of the
descriptor for the host molecules seems rather non-intuitive, and the mathematical form for this
descriptor brilliantly. Formulating suitable descriptors for the host and guest molecules I could
successfully predict the resulting self-assembled patterns formed by them. This work was done in
collaboration with the STM experimental group of Prof K George Thomas, and was published in
‘Chem. Mater. 2017, 29, 17, 7170’ as a cover article.

I found a nice descriptor for the electronic structure of such self-assembled architectures. The preprint
of this work is available.

I have formulated descriptors to predict the efficacy of nanocatalysts deposited on doped oxide
substrates. I studied how the charge, morphology and catalytic activity of deposited gold
clusters/nanoparticles can be tuned by doping the oxide substrate with electron donors/acceptors. Here I
found first one descriptor that depended on both electronegativities and atomic sizes, which could
predict which elements would be the best dopants. This work was published in J. Phys. Chem. C 2019,
123, 32, 19794–19805. However, this descriptor could not handle changes in dopant concentration. So,
in subsequent work, I showed that the support work function (which changes not just as a function of
dopant element but also as a function of dopant concentration) could be a good predictor of the
morphology and charge of the deposited gold clusters. Since the work function of the bare cluster is
much quicker to calculate than any properties of deposited clusters, this is an important result. This
work is published in J. Chem. Phys. 2020, 152, 144704.

I worked with experimental collaborators in the group of Sylvie Rousset and Jerome Lagoute at the
Universite de Paris-Diderot, studying the electronic properties of the self-assembly of the acceptor-type
molecule TCNQ on graphene. This work was published in npj 2D Mater Appl 2019, 3, 5. This project
was sponsored by Indo-French Centre for the Promotion of Advanced Research (IFCPAR/CEFIPRA).