Direct measurement of the diamond nucleation energy landscape during plasma-enhanced chemical vapor deposition (Gebbie et al., PNAS 2018) showed that diamond critical nuclei are metastable, hydrogen-terminated diamondoid molecules containing between 22 and 26 carbon atoms and no bulk atoms, with a barrier four orders of magnitude below capillary estimates. The microscopic origin of this abrupt critical-size behavior has not been identified: it is unknown whether the nucleus–plasma interfacial energy drops sharply with size, whether the surface bonding energy carries the size dependence, or whether the landscape genuinely requires a nonclassical framework with non-Boltzmann subcritical cluster statistics. The problem is to determine this origin experimentally or by quantitative theory validated against the measured landscape, including the role of hydrogen termination in stabilizing the molecular critical nucleus.

Background

Nucleation is the stochastic formation of a thermodynamically stable phase inside a metastable one and is the rate-limiting step of most phase transitions. Classical nucleation theory (CNT) describes it with a capillary free-energy landscape: a cluster of size nn carries a bulk volume term and a surface (interfacial-energy) term, so that subcritical clusters are populated according to a Boltzmann distribution ∼e−ΔG(n)/kBT\sim e^{-\Delta G(n)/k_B T} and growth becomes barrierless above a critical size. Diamond is grown industrially by plasma-enhanced chemical vapor deposition (PECVD), in which a hydrogen-rich microwave plasma supersaturated with carbon radicals condenses diamond on a hot substrate; atomic hydrogen simultaneously etches carbon and terminates dangling surface bonds. Diamondoids are nanometer-scale, hydrogen-terminated hydrocarbon molecules (adamantane C10H16-family cages up to pentamantane C26H32) whose carbon skeletons are exact molecular fragments of the diamond lattice. Because homogeneous diamond nucleation barriers estimated from bulk properties (cleavage surface energies of 1–3 J·m−2^{-2}) exceed 1,000 kBTk_B T, classical capillary reasoning predicts nucleation should essentially never occur, yet CVD diamond nucleates readily — a longstanding paradox, sharpened by calculations (Badziag et al., 1990) showing that hydrogen-terminated nanodiamond clusters can be more stable than hydrogen-terminated graphite below about 100 carbon atoms. Gebbie et al. (2018) measured the first molecular-scale nucleation energy landscape: they covalently anchored monolayers of size-graded diamondoid molecules ('protonuclei', 10–26 carbon atoms) on silicon, grew diamond by PECVD under low supersaturation, and counted the resulting particle densities by electron microscopy. Growth probability grew exponentially with seed size, with an abrupt break between the largest subcritical seed (tetramantane, 22 carbon atoms) and postcritical pentamantane (26 carbon atoms), locating the critical nucleus between 22 and 26 carbon atoms (about 0.8 nm); the fitted nucleation barrier is on the order of 10−20^{-20} J (a few kBTk_B T) with an effective diamond–plasma interfacial energy of 12 ± 6 mJ·m−2^{-2} at 750 °C — more than four orders of magnitude below the capillary estimate. The critical nuclei contain no bulk atoms at all: every atom is a surface atom of a hydrogen-terminated diamondoid cage, which is inconsistent with the capillary picture of a bulk-like core. Within the energetic model, at least one surface term must therefore depend explicitly on cluster size. Two competing explanations were posed: (i) the nucleus–plasma interfacial energy abruptly decreases between 22 and 26 carbon atoms (size-dependent surface energies of small clusters are also predicted by lattice-model studies of prenucleation clusters, Legg and De Yoreo 2016); or (ii) the surface bonding energy itself changes with size. Hydrogen termination is central because it strongly reduces diamond surface energies (Hong and Chou 1998) and stabilizes sp3^3 clusters, but its role in the critical nucleus and its evolution during the subsequent 'bulk transformation' step — the conversion of the molecular cage atoms into bulk diamond — remained undetermined. On the theory side, the two-parameter extension of CNT by Lutsko and Durán-Olivencia (2015), a fluctuating-hydrodynamic stochastic model that generalizes CNT beyond a single cluster-size variable, predicts that subcritical cluster distributions can depart from the Boltzmann population statistics assumed in classical analysis, and the implications for the measured diamond landscape were, per the source paper, under active investigation. The source work proposed a two-step mechanism reconciling the classical and nonclassical features — metastable hydrogen-terminated diamondoid nucleation followed by bulk transformation — and suggested, without demonstration, its generality to other vapor-deposited semiconductors such as silicon.

Problem Statement

Identify the microscopic origin of the abrupt critical-size behavior measured in the diamond CVD nucleation landscape: determine why the growth probability breaks sharply between molecular nuclei of 22 and 26 carbon atoms and why the barrier is four orders of magnitude below capillary estimates. Specifically, establish — by experiment, by quantitative theory validated against the measured landscape (exponential size dependence, critical size between 22 and 26 carbon atoms, barrier of order 10−20^{-20} J, effective interfacial energy 12 ± 6 mJ·m−2^{-2}), or by their combination — whether the critical barrier arises from a nucleus–plasma interfacial energy that decreases sharply with cluster size, from a size-dependent surface bonding energy, from genuinely nonclassical behavior such as multistep nucleation through metastable hydrogen-terminated molecular clusters with subcritical populations departing from the Boltzmann distribution assumed in classical analysis, or from some combination of these. This includes elucidating the role of hydrogen termination in stabilizing the molecular critical nucleus and how termination evolves during the subsequent bulk transformation of the cage atoms into diamond. A complete answer should state what carries the size dependence — which physical quantity or quantities, or which nonclassical departure from the Boltzmann subcritical statistics — with a validated account of the measured landscape; demonstrating the generality of the two-step molecular-precursor mechanism to other vapor-deposited materials such as silicon, suggested but not demonstrated by the source, is an explicit secondary deliverable that extends, but is not required for, closure of the core question.

References

  1. Matthew A. Gebbie, Hitoshi Ishiwata, Patrick J. McQuade, Václav Petrák, Andrew Taylor, Christopher Freiwald, Jeremy E. P. Dahl, Robert M. K. Carlson, Andrey A. Fokin, Peter R. Schreiner, Zhi-Xun Shen, Miloš Nesládek, Nicholas A. Melosh, 'Experimental measurement of the diamond nucleation landscape reveals classical and nonclassical features', Proc. Natl. Acad. Sci. USA 115, 8284–8289 (2018). doi:10.1073/pnas.1803654115, https://doi.org/10.1073/pnas.1803654115
  2. James F. Lutsko, Miguel A. Durán-Olivencia, 'A two-parameter extension of classical nucleation theory', J. Phys.: Condens. Matter 27, 235101 (2015). doi:10.1088/0953-8984/27/23/235101, https://doi.org/10.1088/0953-8984/27/23/235101
  3. Benjamin A. Legg, James J. De Yoreo, 'The energetics of prenucleation clusters in lattice solutions', J. Chem. Phys. 145, 211921 (2016). doi:10.1063/1.4964489, https://doi.org/10.1063/1.4964489
  4. Suklyun Hong, M. Y. Chou, 'Effect of hydrogen on the surface-energy anisotropy of diamond and silicon', Phys. Rev. B 57, 6262–6265 (1998). doi:10.1103/PhysRevB.57.6262, https://doi.org/10.1103/PhysRevB.57.6262
  5. P. Badziag, W. S. Verwoerd, W. P. Ellis, N. R. Greiner, 'Nanometre-sized diamonds are more stable than graphite', Nature 343, 244–245 (1990). doi:10.1038/343244a0, https://doi.org/10.1038/343244a0
  6. Sulgiye Park, Iwnetim I. Abate, Jin Liu, Chenxu Wang, Jeremy E. P. Dahl, Robert M. K. Carlson, Liuxiang Yang, Vitali B. Prakapenka, Eran Greenberg, Thomas P. Devereaux, Chunjing Jia, Rodney C. Ewing, Wendy L. Mao, Yu Lin, 'Facile diamond synthesis from lower diamondoids', Science Advances 6, eaay9405 (2020). doi:10.1126/sciadv.aay9405, https://doi.org/10.1126/sciadv.aay9405
  7. Christopher P. Ender, Jiaxu Liang, Jonas Friebel, Todd Zapata, Manfred Wagner, Anna Ermakova, Tanja Weil, 'Mechanistic insights of seeded diamond growth from molecular precursors', Diamond and Related Materials 122, 108796 (2022). doi:10.1016/j.diamond.2021.108796, https://doi.org/10.1016/j.diamond.2021.108796
  8. Raffaella Salerno, Biagio Pede, Matteo Mastellone, Valerio Serpente, Veronica Valentini, Alessandro Bellucci, Daniele M. Trucchi, Fabio Domenici, Massimo Tomellini, Riccardo Polini, 'Etching Kinetics of Nanodiamond Seeds in the Early Stages of CVD Diamond Growth', ACS Omega 8, 25496–25505 (2023). doi:10.1021/acsomega.3c03080, https://doi.org/10.1021/acsomega.3c03080
  9. Ta Tang, Sulgiye Park, Thomas Peter Devereaux, Yu Lin, Chunjing Jia, 'Molecular geometry specific Monte Carlo simulation of the efficacy of diamond crystal formation from diamondoids', Communications Chemistry 7, 194 (2024). doi:10.1038/s42004-024-01261-9, https://doi.org/10.1038/s42004-024-01261-9
  10. Maria Luce Frezzotti, 'Diamond growth from organic compounds in hydrous fluids deep within the Earth', Nature Communications 10, 4952 (2019). doi:10.1038/s41467-019-12984-y, https://doi.org/10.1038/s41467-019-12984-y
  11. Pedro A. Santos-Flórez, Howard Yanxon, Byungkyun Kang, Yansun Yao, Qiang Zhu, 'Size-Dependent Nucleation in Crystal Phase Transition from Machine Learning Metadynamics', Physical Review Letters 129, 185701 (2022). doi:10.1103/PhysRevLett.129.185701, https://doi.org/10.1103/PhysRevLett.129.185701