Multiscale computational understanding and growth of 2D materials: a review
PDF variant
Uncorrected proof PDF. Canonical article text and stable pagination: 2020momeni-npj-computat-multiscale-computational (papers/Momeni_2D_review_NPJ_CompMat_2020.pdf).
Summary¶
The uncorrected proof PDF papers/Momeni_2D_review_NPJ_CompMat_proof.pdf corresponds to the npj Computational Materials review DOI 10.1038/s41524-020-0280-2, “Multiscale computational understanding and growth of 2D materials.” The extract opens by noting graphene’s 2004 isolation and the subsequent expansion of layered semiconductors and heterostructures, then states the review’s aim: to overview theoretical, computational, and machine-learning tools across length and time scales to assist design and synthesis of 2D materials beyond graphene. It announces three methodological tiers: (i) nanoscale atomistic simulations including DFT and MD with empirical and reactive potentials; (ii) mesoscale methods such as phase-field modeling; and (iii) macroscale continuum approaches coupling thermal and chemical transport. It further promises discussion of machine learning combined with computation and experiments to link structures and properties and to guide discovery, plus an outlook on computational approaches to synthesis and growth.
Methods¶
As a review, the article surveys heterogeneous literature rather than a single protocol. The Introduction in the extract contrasts top-down approaches (mechanical and liquid-phase exfoliation) with bottom-up CVD/ALD routes, notes sensitivity of morphology to thermodynamic and kinetic conditions (heat/mass transfer, reaction kinetics, adsorption, nucleation), and lists multiscale challenges: quantum and atomistic activation energies for migration, FEM mass transport, substrate defects, wrinkling, van der Waals interactions at mesoscales, and growth kinetics specific to atomically thin films, including flexural phonon modeling issues for 2D materials. The following section heading ATOMISTIC COMPUTATIONAL METHODS begins with First-principles calculations, describing DFT’s role in comparing chemical potentials of polymorphs and analyzing kinetic pathways via transient structures; it notes standard PBE-GGA usage and limitations such as underestimated reaction barriers from delocalization error, and introduces “above-hull” energy as a practical stability screen. Detailed numerical settings for any one cited study remain in primary references; follow 2020momeni-npj-computat-multiscale-computational for the finalized sectioning.
Findings¶
The proof excerpt argues that developing design tools for 2D synthesis requires bridging very wide length and time scales—for example, reactive force-field calculations of surface migration barriers coupled to finite-element mass transport—and that useful multiscale models must be efficient, accurate, and able to capture multi-physical links among growth conditions, morphology, and properties, with an eventual goal of guiding reactor design for uniform large-area films. Because the excerpt shown is only the opening and the start of atomistic methods, later sections on mesoscale approaches and machine-learning integrations are partially represented here; full positioning of ReaxFF alongside classical potentials and continuum solvers appears in the published PDF summarized on 2020momeni-npj-computat-multiscale-computational. Final figure and equation numbering may differ between uncorrected proof and issue PDF.
Limitations¶
Uncorrected proofs can differ from the final Nature Partner Journal PDF in typography and minor editorial fixes. Use the VOR ingest for citation locators.
Relevance to group¶
Van Duin co-authorship ties the review’s atomistic sections—including reactive potentials—to Penn State’s broader 2D materials modeling ecosystem.