Ensemble Monte Carlo and DFPT simulation data for charge transport and gas sensing in monolayer MoS2

Technical details Everything is uploaded as a single archive. It extracts to one folder with three parts: emc/ (transport outputs, the 17-point self-consistent device sweep, and the C++ drivers and parameter headers that produced them), dfpt/ (the first-principles calculations, namely electron-phonon couplings at 34 wavevectors, the Wannier interpolation and Boltzmann-solution files, the K-Q valley separation study, and the adsorbate relaxations for NO2, NH3 and O2), and scripts/ (the Python scripts that turn the outputs into the figures of the paper).

Data files are plain text, whitespace-delimited, each with a # header naming its columns and units. The binary files are Quantum ESPRESSO dynamical matrices and matrix-element files, and some large outputs are gzipped. Both are documented in the READMEs.

Reading the data needs only a text viewer or Python with NumPy. Rerunning the decks needs the ViennaEMC framework, and repeating the first-principles work needs Quantum ESPRESSO 7.3.1 with the PseudoDojo norm-conserving PBE pseudopotentials, EPW 5.8 and Wannier90 3.1. The charge-transfer analysis also uses the Bader program of Henkelman et al. A README at the top level documents every file, with further READMEs in dfpt/couplings/ and dfpt/inputs/.

The framework and example decks are at github.com/ViennaTools/ViennaEMC. Further details The record contains the authors' own simulation data, with one exception. Two small files hold points extracted from the published figures of Liu et al. (2014) [ doi.org/10.1021/nn5015215 ] and Late et al. (2013) [ doi.org/10.1021/nn400026u ], so that the comparison figures can be reproduced, and those measurements belong to the cited publications.

Other comparison data used in the paper (Ahn 2017 [ doi.org/10.1038/s41598-017-04350-z ], Kumar 2025 [ doi.org/10.1016/j.chphi.2025.100847 ] and others) are not redistributed. Quantum ESPRESSO wavefunction scratch, charge-density cubes and the frozen-strain deformation-potential runs are excluded, being regenerable from the decks and not needed to reproduce the provided data.