The development of a scalable, economical production technique for monoand few-layer graphene (M-/FLG) is a key requirement to exploit its unique properties for applications. Catalytic chemical vapour deposition (CVD) has emerged as one of the most promising and versatile methods for M-/FLG growth. The generic principle of catalytic, rather than pyrolytic, CVD is to expose a catalyst template to a gaseous precursor at temperatures/conditions for which the precursor preferentially dissociates on the catalyst. Hence, the catalyst is key to M-/FLG formation, in particular its role in precursor dissociation, C dissolution, M-/FLG nucleation and domain growth/merging. Although the structure of as-formed graphitic layers on crystalline transition metal surfaces under ultra-high vacuum conditions has been extensively studied in surface science, a central question remains: what M-/FLG quality can be achieved with CVD, in particular, if for cost effectiveness sacrificial polycrystalline metal films/foils and less stringent vacuum/CVD process conditions are used. There have been numerous recent reports of large area M-/FLG CVD on for instance poly-crystalline Ni and Cu, including integrated roll-to-roll processing. However, there is currently very limited understanding of the detailed growth mechanisms, and the mostly empirical process calibrations provide little fundamental insight in to how the process and M-/FLG quality/domain size can be optimised. Herein, we study M-/FLG CVD by complementary in situ probing under realistic process conditions with the aim of revealing the key growth mechanisms. We focus on poly-crystalline Ni films and simple one-step hydrocarbon exposure conditions. However, as highlighted by Figure 1, even for such seemingly simple CVD conditions, the parameter space is manifold which leads to ambiguity in the interpretation of post-growth process characterisation and motivates our in situ approach. For catalyst metals with a high C solubility, such as Ni, current literature typically assumes C precipitation upon cooling as the main growth process. 8] M-/FLG precipitation has been studied in detail for slow, near thermodynamic equilibrium thermal cycling of C doped crystals. 9] For CVD, however, the conditions are distinctly different (Figure 1): an isothermal C precursor exposure phase, which represents a variation in composition rather than temperature, is followed by a typically fast cooling or thermal quenching. Hence kinetic aspects are important. Additionally, competing processes might influence the growth outcome such as etching of M-/FLG in a reactive atmosphere, for example, hydrogen or water, during the CVD process. By combining in situ, timeand depth-resolved X-ray photoelectron spectroscopy (XPS) and in situ X-ray diffraction (XRD), we can clearly show that M-/FLG growth occurs during isothermal hydrocarbon exposure and is not limited to a precipitation process upon cooling. While the fraction of M-/FLG due to isothermal growth and precipitation upon cooling strongly depends on process conditions, we show that the former is dominant for the low-temperature CVD conditions used. We find that M-/FLG nucleation is preceded by an increase in (subsurface) dissolved C with the formation of a solid solution of C in the Ni film, which indicates that graphene CVD is not a purely surface process. We discuss our data here in the context of simple considerations of C solubility and diffusivity as well as rate equations of the basic contributing processes, in order to establish a framework to guide future improvements in graphene CVD by a more fundamental understanding. We perform in situ XPS during low-pressure CVD of M-/FLG from hydrocarbon precursors on Ni(550 nm) films. Figure 2A Figure 1. Illustrative processing profile for a simple one-step hydrocarbon exposure consisting of four major phases: catalyst pretreatment, C dissolution into the catalyst during initial precursor exposure, isothermal M-/FLG growth with continued precursor exposure, M-/FLG growth by precipitation upon cooling. The key catalyst and M-/FLG properties that may be defined at each phase of growth are also listed.
Read full abstract