Predicting the adhesion and delamination strength of carbon films on metals by high-throughput ab initio calculations

TL;DR

This work tackles the challenge of strong, predictable adhesion between carbon-based coatings and metallic substrates by performing a high-throughput, ab initio study of 60 diamond/metal interfaces across four diamond terminations. Using the TribChem workflow and the Zur algorithm to generate minimal-strain, physically meaningful interfaces, the authors compute adhesion energies, charge transfer, and structural/electronic responses, revealing that adhesion scales with the geometric mean of constituent surface energies and that fracture location can be anticipated from cohesive energies. They further show that surface graphitization weakens adhesion while rehybridization of surface carbon and charge accumulation correlate with stronger bonding, offering a simple yet effective screening descriptor for material design. Overall, the study provides a predictive framework for selecting carbide-forming metals and diamond terminations to optimize adhesion, with clear implications for DLC coating design and interlayer engineering.

Abstract

Diamond and diamond-like carbon (DLC) coatings are widely employed for their exceptional mechanical, thermal and chemical properties, but their industrial application is often limited by weak adhesion to metallic substrates. In this work, we employ a high-throughput ab initio approach to systematically investigate the adhesion of diamond-metal interfaces, combining a set of technologically relevant metals (Al, Ag, Au, Cr, Cu, Fe, Ir, Mg, Mo, Pt, Rh, Ti, V, W, Zn) with the C(111), C(111)-2x1 (Pandey reconstructed), C(110), C(100), that are most common in diamond and are representative of different types of bonds present in DLC. Thanks to our automated and accurate computational protocol for interface construction and characterization, databases are populated and relevant trends are identified on the effect of surface graphitization, ability to form carbides and metal reactivity on carbon film adhesion and delamination strength. Beyond capturing trends, our workflow yields predictive insights. Indeed, we found that adhesion energy scales with the geometric mean of the constituent surface energies, providing a simple descriptor for rapid screening; while comparing the work of separation with the metal's cohesive energy anticipates the fracture location under tensile loading. A novel method based on g(r) analysis is introduced to identify when contact with a metal drives rehybridization of surface carbon from sp2 to sp3, the structural signature of improved resistance to delamination. These structural changes are mirrored by an electronic rearrangement at the interface, quantified by a charge-accumulation descriptor that strongly correlates with adhesion.

Figures (9)

• Figure 1: Schematic representation of the computational workflow implemented in TribChem to calculate solid interface properties.
• Figure 2: Maximum percentage strain applied on the metal constituent in the interface supercell. Negative values correspond to compressive strain (red regions), whereas positive values represent tensile strain (blue regions).
• Figure 3: Adhesion energies of diamond/metal interfaces. Darker colors represents higher $|E_{adh}|$. Metals are ordered from left to right according to decreasing surface energy.
• Figure 4: Atomistic models of interfaces between the four considered low-index diamond terminations and representative metallic substrates.
• Figure 5: Adhesion energies as a function of the geometric mean between diamond and metal surface energies. Metal substrates are differentiated by colors, while symbols represent different diamond terminations. The grey dashed line represent the linear fit. The corresponding fitting parameters and quality indicators are summarized in Tab.\ref{['tab:linear_fit']}. Data are plotted using the same scale.