Machine Learning Assisted Reconstruction of Local Electronic Structure of Non-Uniformly Strained MoS 2

Soumyadip Hazra, Sraboni Dey, Arijit Kayal, Narendra Shah, Renjith Nadarajan, Joy Mitra · 2026

Read Machine Learning Assisted Reconstruction of Local Electronic Structure of Non-Uniformly Strained MoS 2 by Soumyadip Hazra, Sraboni Dey, Arijit Kayal, Narendra Shah, Renjith Nadarajan, Joy Mitra (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-29298.txt.

Introduction

Hazra and co-authors map how uneven strain changes the electronic structure of monolayer MoS2. Wrinkles and nanobubbles are common in real devices. The team combines density functional theory with a recurrent neural network. Strain maps come from microscopy topography and Raman data. Biaxial bending strain proved far stronger than uniaxial or in-plane strain. A 0.35 percent strain cut the band gap by about 22 percent. Predicted gap maps matched light-emission measurements. The framework extends to other 2D materials.

Catalog record

CREATORSoumyadip Hazra, Sraboni Dey, Arijit Kayal, Narendra Shah, Renjith Nadarajan, Joy Mitra
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE51468
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.29298

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Machine Learning Assisted Reconstruction of Local Electronic Structure of Non-Uniformly Strained MoS 2 Soumyadip Hazra, Sraboni Dey, Arijit Kayal, Narendra Shah, Renjith Nadarajan, Joy Mitra Abstract Wrinkles and nanobubbles are an integral and often unavoidable part of integrating 2D van der Waals semiconductors into actual device architectures. Despite their ubiquitous nature, quantitative correlation between such spatially non-uniform strain and modifications to the local electronic structure remains challenging. Here, density functional theory is combined with a recurrent neural network to reconstruct the local electronic structure of monolayer MoS 2 from strain maps derived from atomic force microscopy (AFM) topography and Raman spectral maps. The analysis reveals that biaxial bending induced strain is significantly more effective than both uniaxial bending or in-plane strain in modifying electronic and dielectric properties. A ∼ \sim 0.35% strain induced by biaxial bending results in ∼ \sim 22% reduction in band gap and ∼ \sim 7% increase in dielectric constant, compared to a ∼ \sim 5% reduction in band gap and ∼ \sim 1% increase in dielectric constant under comparable uniaxial bending. The modified band structure reveals band edge states that concentrate charge in regions of high curvature or strain. While conductive AFM measurements indicate increased local conductance (carrier density) at wrinkles and nanobubbles, the spatial band gap maps predicted by the model are validated against experimental photoluminescence peak energy maps. The results indicate that strained features like wrinkles and nanobubbles commonly present in real devices influence the band gap, carrier distribution, and dielectric response, which favourably affects electrical transport in such systems. The framework developed here can be readily extended to other 2D materials and heterostructures, offering a computationally efficient route for studying and exploiting strain effects. keywords TMDC, MoS 2 , Biaxial Strain, Neural Network, Optical Spectroscopy † † affiliation: CAMRIE, Indian Institute of Science Education and Research Thiruvananthapuram, Kerala 695551, India † † affiliation: School of Physics, Indian Institute of Science Education and Research Thiruvananthapuram, Kerala 695551, India † † email: [email protected] † † affiliation: School of Physics, Indian Institute of Science Education and Research Thiruvananthapuram, Kerala 695551, India † † affiliation: J. Heyrovsky Institute of Physical Chemistry, Czech Academy of Sciences, Dolejskova 2155/3, 18200 Prague 8, Czech Republic † † affiliation: School of Physics, Indian Institute of Science Education and Research Thiruvananthapuram, Kerala 695551, India † † affiliation: School of Physics, Indian Institute of Science Education and Research Thiruvananthapuram, Kerala 695551, India † † affiliation: School of Physics, Indian Institute of Science Education and Research Thiruvananthapuram, Kerala 695551, India † † email: [email protected] † † affiliation: School of Physics, Indian Institute of Science Education and Research Thiruvananthapuram, Kerala 695551, India † † affiliation: CAMRIE, Indian Institute of Science Education and Research Thiruvananthapuram, Kerala 695551, India † † suppinfo: T 1 Introduction Precise control of carrier density and mobility in semiconductors is central to the development of high-performance, energy-efficient electronic and optoelectronic devices. The limitations of conventional doping strategies used in bulk semiconductors have been mitigated in two-dimensional electron gas (2DEG) systems, where quantum confinement and spatial separation of the active channel from dopants enhance mobility, along with high gate tunability. Such architectures underpin modern nanoscale transistors and quantum devices capable of high-frequency operation with ultralow power consumption Radisavljevic et al. 2011 . The isolation of graphene Novoselov et al. 2004 catalysed the emergence of a broader class of van der Waals (vdW) materials in which charge carriers are intrinsically confined to atomic-scale thicknesses. Reduced dielectric screening, strong Coulomb interactions, and symmetry-driven band structures in these systems gave rise to a wealth of emergent quantum phenomena, including strongly bound excitons Liang et al. 2014 ; Sajjad et al. 2018 , valley polarisation Kayal et al. 2022 ; Ho et al. 2020 ; Gopalan et al. 2025 ; Zhang et al. 2022 ; Kumar et al. 2025 , excitonic condensation Gao et al. 2023 ; Moon et al. 2025 , and correlated insulating and superconducting phases Dong et al. 2022 ; Jamwal et al. 2026 that have no direct bulk counterparts. Within the 2D materials family, transition metal dichalcogenides (TMDCs) are particularly compelling due to their semiconducting nature, strong in-plane covalent bonding, and weak interlayer van der Waals interactions, which enable both mechanical exfoliation and controlled growth in atomically thin layers Wang et al. 2020 ; Kang et al. 2022 . Molybdenum disulfide (MoS 2 ), a prototypical TMDC, exhibits polymorphism (1T, 1H, 2H, and 3R) Sam et al. 2020 and a pronounced thickness-dependent band structure, transforming from an indirect to a direct bandgap semiconductor at the monolayer limit Kuc et al. 2011 ; Yun et al. 2012 . Here, strain engineering has emerged as a powerful controllable tool to modulate the local electronic and phononic structure and the ensuing electrical, optical and opto-electronic properties, crucially without introducing chemical dopants. Experimental results for uniaxially and biaxially stretched monolayer (ML) MoS 2 show that the bandgap energy ( E_{g} ) decreases \sim 45-70 meV/% of uniaxial strain Conley et al. 2013 ; He et al. 2013 ; Wang et al. 2022 and by \sim 100 meV/% for biaxial strain Plechinger et al. 2015 ; Mohammad Tabatabaei et al. 2013 ; Yu et al. 2015 ; Li et al. 2015 ; López-Suárez et al. 2016 . However, actual deformations in 2D flakes immobilised on substrates are more complex, with local biaxial bending of the lattice evidenced by nanobubbles, wrinkles, and other buckled structures that nucleate spatially non-uniform strain across the flakes. Although the kinetics of formation of these deformations are directed by the specifics of the fabrication process, their equilibrium configuration is primarily governed by a competitive interplay between the material’s adhesion with the substrate and its bending modulus. The effect of such non-uniform strain has been explored experimentally using elastomeric substrates Castellanos-Gomez et al. 2013 , nanoindentation Huang et al. 2011 , patterned substrates Kayal et al. 2023 ; Yanev et al. 2024 ; Yang et al. 2025 ; Wang et al. 2019 , etc., resulting in substantial modification to the physical and chemical properties of the flakes. These high-strain deformations also lead to local doping, phonon quenching, and modifications to the dielectric environment, which collectively contribute to improved carrier transport Ng et al. 2022 ; Kayal et al. 2023 and enhanced catalytic properties Nadarajan et al. 2024 in ML MOS 2 . Recent developments have also demonstrated strain-tunable TMDC electronics, where controlled deformation enables local band alignment Ai et al. 2025 , exciton funnelling Moon et al. 2020 ; Qian et al. 2025 , piezoresistive switching Manzeli et al. 2015 ; Nakamura 2018 , and dynamically reconfigurable device behavior. In comparison, first-principles calculation based investigations on changes in electronic properties in non-uniformly strained 2D TMDC systems, due to local biaxial bending, have received limited attention Yu et al. 2016 ; González et al. 2018 ; Gupta et al. 2022 ; Gupta et al. 2025 . Here, the influence of local biaxial strain on the structural, electronic, and dielectric properties of ML MoS 2 is investigated using density functional theory (DFT) calculations Kohn and Sham 1965 . The deformations are simulated by bending the atomically thin lattice into a three-dimensional Gaussian feature, mimicking experimentally observed surface topographies. The relaxed atomic lattice is used to estimate local

Read the complete text