Five-Membered Loop-Initiated Nucleation of Unlayered Graphene Sheets in a Carbon Melt, with Diffusion Barrier Possibilities

Abstract

A subset of isotopically “presolar” carbon particles extracted from the Murchison meteorite contain signatures expected of particles formed in the atmosphere of red giant stars. Some of these micron-size particles have spherical cores that show diffraction rings from atom-thick graphene, possibly formed by solidification of liquid carbon at low pressure. Electron phase-contrast transmission electron microscopy (TEM) images suggest that these cores originate from supercooled carbon droplets that formed graphene sheets on randomly oriented 5-membered loops. In addition to presolar data, laboratory synthesis in an “evaporating carbon oven” creates similar core-rim structures by slow cooling of carbon vapor. In research studies, it was shown that 5-membered loops are essential to the initiation of carbon nanotubes on catalyst particles. In addition to offering this experimental context, we present density functional theory (VASP) computer simulations suggesting that 5-member loops are more likely than 6-member loops in a solidifying carbon melt. These things suggest that 5-member loops compete effectively as nucleation seeds for explaining the faceted pentacones inferred from TEM images of presolar particle cores. In that context, pent-first nucleation (along with the crowding of growing sheets by nearby liquid atoms) may reduce the chances of graphite layer formation and lead to unprecedented diffusion barrier properties for this composite material.

A Semi-empirical Model for the Containerless Solidification of Liquid Carbon

“…Independent semiempirical modeling also saw a high preference of carbyne chains for containerless liquid carbon . Another DFT study of liquid carbon at low pressures also noticed a preference for pent-first vs hex-first loop formation during nucleation …”

“…Independent DFT studies have shown that setting b = 5 is likely a more plausible value as low-pressure carbon melts tend toward a "pentagon first mechanism". 10,44 The value for b can also change when considering other factors which affect growth such as edge curvature and shape changes.…”

“…It is also consistent with the preference for carbyne chains at low pressures when studying liquid carbon using DFT 10 and the experimental/modeling evidence for pent-first nucleation. 44 From our nucleation and growth models, we were able to study the graphene sheet number density, mass-weighted average, and fraction crystalline. Our analytical models showed good agreement with values determined from lab-grown and presolar cores.…”

“…This will give a quadratic in time given by n ( t ) = b + a b t + true( a t 2 true) 2 The nominal value for b is usually considered to be 6 as the carbon hex-loop’s tight binding make it a great candidate as a nucleation seed. Independent DFT studies have shown that setting b = 5 is likely a more plausible value as low-pressure carbon melts tend toward a “pentagon first mechanism”. , The value for b can also change when considering other factors which affect growth such as edge curvature and shape changes.…”

“…The growth model follows a single fixed Arrhenius activation energy model, 43 starting with our independent DFT and experimental work predictions of 5-and 6-atom critical nucleus rings. 44 Our models use data from electron microscope study of presolar and lab-grown "cores", DFT predictions from other works, and data obtained from our own molecular dynamics studies of liquid carbon when employing the LCBOP semiempirical potential. 45 This work is complementary to work done by Goldman et al, 10 which looked at high-pressure cases of the carbon triple point where we are studying the low-pressure case.…”