Deciphering borophene growth pathways with data-driven simulations

Colin Bousige, Jean Furstoss, Julien Lam, Pierre Mignon · 2026

Read Deciphering borophene growth pathways with data-driven simulations by Colin Bousige, Jean Furstoss, Julien Lam, Pierre Mignon (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-2606-03886.txt.

Introduction

Bousige and co-authors simulate borophene growth to find windows for chosen phases. Making borophene to order is hard, since many polymorphs compete as layers form. A reactive learned potential drives the runs. Grand-canonical Monte Carlo runs on silver surfaces, both the 111 and 100 faces. A data-driven sorting step tracks growth from early nuclei to full layers.

Growth maps over heat and pressure come out of the runs. Vacancy motifs, phase mixing, and seed shape govern which polymorph wins. The runs match lab trends, with two phases on top and a heat-linked rivalry. Kinetic paths link short-lived nuclei to long-range order. Conditions that suppress rival motifs are named. Growers get workable windows for targeted phases. The frame extends to phase control in other thin materials.

Catalog record

CREATORColin Bousige, Jean Furstoss, Julien Lam, Pierre Mignon
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE39 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2606.03886

Browse the full library · About the Tower of Babel

Full text

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

Deciphering borophene growth pathways with data-driven simulations Colin Bousige, Jean Furstoss, Julien Lam, Pierre Mignon Abstract Deterministic synthesis of borophene remains challenging because many polymorphs compete during nucleation and growth. Here we combine a reactive machine-learned interatomic potential with grand-canonical Monte Carlo simulations and data-driven structural classification to track borophene formation from early nuclei to extended layers on Ag(111) and Ag(100). We build temperature-pressure substrate growth maps and resolve how vacancy motifs, phase intermixing and seed structure govern polymorph selection. The simulations reproduce key experimental trends, including the prevalence of β 12 \beta_{12} / χ 3 \chi_{3} phases and their temperature-dependent competition, while revealing kinetic pathways that connect metastable nuclei to long-range order. We identify conditions that suppress competing motifs and promote targeted phases, providing actionable synthesis windows. These results establish a predictive framework for directing borophene growth and, more broadly, for controlling polymorphism in low-dimensional materials by coupling atomistic simulation with machine-learning-enabled phase recognition. The advent of borophene, a one-atom-thick two-dimensional boron sheet, has marked a milestone in materials science Mannix et al. 2015 ; Feng et al. 2016 . This material, characterized by its rich polymorphism Tang and Ismail-Beigi 2007 ; Mannix et al. 2015 ; Feng et al. 2016 ; Wu et al. 2012 ; Penev et al. 2012 , introduces a metallic 2D system with partial stability. Its attributes position borophene as a candidate for next-generation technologies, including high-performance, flexible Mortazavi et al. 2017 ; Pang et al. 2016 ; Zhang et al. 2017a , and transparent Lherbier et al. 2016 electronics, as well as optoelectronic devices Huang et al. 2017 ; Lian et al. 2020 and high-capacity ionic batteries Liu et al. 2016 ; Jiang et al. 2016 ; Li et al. 2017 ; Liang et al. 2017 ; Rao et al. 2017 ; Zhang et al. 2017b . Borophene’s structural diversity enables precise tuning of its plasmonic, electronic, thermal, and mechanical properties. A growing body of literature now offers comprehensive reviews on borophene’s progress and potential Mannix et al. 2018 ; Hou et al. 2020 ; Kaneti et al. 2021 ; Ou et al. 2021 ; Kumar et al. 2025 ; Wang et al. 2024 ; Wang et al. 2019 ; Adekoya et al. 2024 ; Gupta et al. 2024 ; Innis et al. 2025 . To date, bottom-up synthesis methods have emerged as the most reliable approach for producing monolayer borophene Innis et al. 2025 . Experimental efforts have identified eleven distinct polymorphs, each influenced by synthesis parameters such as temperature and substrate type and orientation Mannix et al. 2015 ; Feng et al. 2016 ; Li et al. 2018 ; Vinogradov et al. 2019 ; Zhong et al. 2017a ; Zhong et al. 2017b ; Kiraly et al. 2019 ; Mazaheri et al. 2021 ; Wu et al. 2019 ; Wu et al. 2022 ; Tai et al. 2015 ; Sutter and Sutter 2021 ; Radatović et al. 2022 ; Omambac et al. 2021 ; Cuxart et al. 2021 ; Sheng et al. 2019 ; Wu et al. 2025 ; Innis et al. 2025 . To systematically investigate these polymorphs, we compiled a dataset of 21 distinct borophene allotropes (including all those that have been observed experimentally, as well as others), as shown in Fig. 1 . In addition, phase intermixing is frequently observed Liu et al. 2018 ; Li et al. 2023 ; Kiraly et al. 2019 ; Wang et al. 2020 ; Wu et al. 2025 , underscoring the complexity of the borophene system. Yet, this very complexity also hints at the possibility of selectively synthesizing specific polymorphs tailored to desired properties – provided the optimal experimental conditions are identified. Consequently, unraveling the growth mechanisms of borophene has become a critical priority, as achieving deterministic synthesis is essential to harnessing its full potential across a spectrum of applications. Figure 1: Allotrope identification. a . Snapshots of the 21 different allotropes included in the training dataset in this study. In direct application of our classification model, the colored squares identify the vacancies with their corresponding allotrope. The same colors will be used throughout the article. The nomenclature ( \alpha , \beta , \chi , and \delta ) refers to the coordination of the boron atoms, while the subscript is only used to distinguish between different allotropes (except for the \delta allotropes where it indicates the coordination). We selected these allotropes because (1) they include all those experimentally observed on silver Innis et al. 2025 ( \alpha , \beta_{8} , \beta_{10} , \beta_{12} , and \chi_{3} ) as well as those synthesized on copper, gold, and aluminum ( \alpha_{1} , \alpha_{5} , \chi_{3} ), and (2) they cover the full range of vacancy densities \nu (from 0 for \delta_{6} to 1/3 for \delta_{3} ), a critical parameter for borophene properties. Some other allotropes are punctually seen during our long growth simulations (e.g. \alpha_{7} ), so we also included them in the training dataset to be able to identify them during the growth process. b . Dimensionality reduction of the LMBTR descriptors for the different allotropes in the training set through a two-components t-SNE projection. Each point corresponds to a vacancy in the structures obtained during the growth process, and the color corresponds to the allotrope it belongs to. Even with a 2-components t-SNE projection, the different allotropes are quite well separated in the descriptor space. We use a 155-components PCA projection for the actual analysis, which allows for an even better separation. Experimentally, there have been no systematic studies of borophene growth allowing identifying early steps in the expansion mechanism or possible precursors leading to particular allotropes – or to phase mixing of specific allotropes. This is further complicated by the fact that the borophene growth mechanism is substrate-dependent: it was shown that the dose of atomic boron required to observe nanoscale borophene islands on Au(111) is an order of magnitude greater than that required to form a monolayer on Ag(111) Kiraly et al. 2019 . It was interpreted as an indication that boron dissolves into the bulk in Au(111) – leading to a surface segregation growth mechanism – while remaining on the surface in Ag(111). This is consistent with the lower enthalpy of formation of gold borides compared to silver borides, although it is positive in both cases Niessen and De Boer 1981 . Although gas-phase photoelectron spectroscopy combined with theoretical atomistic calculations predicted several planar structures stabilized by the presence of hexagonal vacancies for B 30 , B 36 , B {}_{37}^{-} or B {}_{38}^{-} clusters Sergeeva et al. 2014 ; Li et al. 2023 ; Kiran et al. 2005 ; Li et al. 2014 ; Chen et al. 2017 ; Piazza et al. 2014 , there is no experimental evidence for their existence on metal substrates. Only a few studies have investigated the growth and nucleation of borophene on metal substrates Liu et al. 2013 ; Xu et al. 2016 ; Shu et al. 2016 ; Yu et al. 2026 , and most have focused on the very early stages of growth. It was shown from both ab initio molecular dynamics (MD) Shu et al. 2016 and static DFT calculations Xu et al. 2016 that the presence of hexagonal

Read the complete text