Access Official Technical PDF

Share Summary:LinkedInEmail

Controlling graphene heterostructure growth requires more than selecting carbon and boron nitride (h-BN) precursor gases. The behavior of individual carbon atoms at the growth surface—including where they adsorb, how they move and whether they form clusters—can affect whether a new material grows beside or beneath an existing layer. First-principles calculations of carbon on Cu(111) partially covered by graphene or h-BN help clarify these atomic-scale differences and their potential relevance to chemical vapor deposition (CVD) process control.

The results are theoretical rather than a recipe for a specific production process. They nevertheless identify mechanisms that technical teams can consider when tuning growth conditions to form lateral or vertical two-dimensional (2D) material structures.

Why the graphene–h-BN interface matters

Graphene and h-BN have different electronic properties, making their combination of interest for electronic applications. A lateral heterostructure joins the materials along an edge within the same plane; a vertical heterostructure places one material above another. CVD can produce these structures, and the order in which source gases are supplied can influence the growth sequence.

For process development, the key challenge is understanding how the growing material interacts with the substrate and with a layer already present. The study compared carbon adsorption on bare Cu(111) with adsorption at graphene/Cu and h-BN/Cu interfaces. It also examined carbon movement across the surface and the initial formation of carbon dimers, an early step toward larger carbon structures.

Carbon adsorption varies with the underlying layer

On each of the three surfaces, the most stable calculated position for a single carbon atom was associated with an H3 site on Cu(111). At the graphene/Cu and h-BN/Cu interfaces, the carbon atom can bond with an atom in the overlying sheet. The calculations also identified a nearby, less stable configuration in which the carbon atom does not form that bond.

Comparisons among surfaces showed that a carbon monomer is less readily accommodated beneath h-BN than beneath graphene or on bare copper. This distinction matters because the presence of an overlayer does not affect carbon incorporation in the same way for both materials. In the model, the preferred interfacial adsorption position is substantially more favorable than adsorption on the outer surface of either graphene or h-BN.

The study also found that a carbon dimer is less favorable at the graphene/Cu interface than at the h-BN/Cu interface or on bare Cu. Taken together, the monomer and dimer results suggest that carbon may accumulate and begin forming a new structure differently depending on which layer covers the copper.

Diffusion barriers offer clues for CVD process control

Adsorbed carbon must move before it can contribute to growth at another location. The calculated lateral diffusion barrier was 0.11 eV on bare Cu, 0.14 eV at h-BN/Cu and 0.26 eV at graphene/Cu. In this model, carbon therefore diffuses less readily beneath graphene than beneath h-BN or across bare copper.

The pathways are not identical. At both covered interfaces, the atom moves through a nonbonding configuration before reaching neighboring positions. On graphene/Cu, the calculated transition out of the bonding position is the rate-limiting step along the reported lateral path. This more complicated route differs from the path calculated on bare copper, where the atom moves through intermediate surface sites.

The calculations also compared lateral motion with movement into copper’s subsurface. Direct vertical diffusion barriers were higher than the corresponding lateral barriers for all three surfaces. Even when an alternative path through a surface valley was considered, the calculated barrier to the subsurface remained above the lateral barrier. The results therefore favor movement along the surface or interface over penetration into the copper subsurface under the modeled conditions.

For engineers and process teams, these findings highlight why a change in the covering layer may alter carbon transport and nucleation behavior even when the substrate remains Cu(111). Growth conditions that affect the supply and movement of carbon could consequently influence where new material forms. The calculations provide comparative guidance, not a validated operating window: they do not specify a complete CVD recipe or establish how every reactor condition will affect growth.

Edge termination may influence growth direction

The researchers also modeled graphene and h-BN nanoribbons on copper. Ribbons with bare edges curled toward the surface, with their edges terminated by copper. When the edges were fully hydrogen-terminated, the ribbons remained parallel to the copper surface in the modeled structures.

Hydrogen binding differed between graphene and the two types of h-BN edge in the calculations. The authors propose that edge termination could therefore help guide whether growth proceeds laterally or vertically. In their interpretation, copper-terminated bare edges may favor lateral growth, while hydrogen-terminated edges may make vertical growth more feasible by allowing carbon to enter the interface, form dimers and develop into larger flakes.

This is a proposed growth-control mechanism, not a guarantee that a particular hydrogen condition will produce a specific structure. The calculations indicate that edge chemistry is a factor worth considering alongside precursor supply and carbon diffusion when developing a process for lateral and vertical heterostructures.

Practical implications and modeling limits

The study used density functional theory with a van der Waals correction to represent the interaction between the 2D sheet and copper. Its conclusions apply to the modeled structures and conditions. The authors note that confirming the preferred diffusion path rigorously would require further calculations, including larger cells and a different van der Waals correction method.

For technical decision-makers, the practical takeaway is to treat the overlayer and edge state as active parts of the growth system—not simply passive films on a copper substrate. Differences in carbon adsorption, lateral mobility and dimer stability may help explain why graphene and h-BN interfaces respond differently during bottom-up growth.

In short, graphene heterostructure growth may be steered by managing the conditions that determine where carbon adsorbs, how it travels and how flake edges are terminated. These theoretical results offer a framework for evaluating CVD process variables, while experimental validation remains necessary before translating the proposed mechanisms into production controls.

Access Official Technical PDF

About The Author