Fig. 6: Validation of the MLP model for fracture target space in crystal carbon systems and simulations of fracture toughness of a-C:H. | npj Computational Materials

Fig. 6: Validation of the MLP model for fracture target space in crystal carbon systems and simulations of fracture toughness of a-C:H.

From: Transferable machine learning model for multi-target nanoscale simulations in hydrogen-carbon system from crystal to amorphous

Fig. 6

Bond length change versus normalized stress intensity factor by the value of fracture toughness (\({{K}_{{\rm{{\rm I}}}}/K}_{{\rm{{\rm I}}}{\rm{C}}}\)): (a) for diamond [001] crack front propagation on the (110) plane and (c) for graphene in-plane crack front propagation along the zigzag edge. The ‘L’ in (a) and (c) represented the distance between the atomic centers on the site of the crack-tip. The energy and force change versus separation distance extrapolated from decohesion simulations: (b) for diamond (110) and (d) for graphene zigzag edge. The ‘D’ in (b) and (d) represented the separated distance between the created surfaces or edges. The snapshots for the processes above were inserted with the gray spheres representing the carbon atoms, and the blue spheres representing hydrogen atoms. Fracture toughness (\({K}_{{\rm{{\rm I}}}{\rm{C}}}\)) extrapolated from the MD simulations: (e) for a-C with different density (2.0–3.2 gcm−3) and (f) for a-C:H with different hydrogen content (10–40 at.%H).

Back to article page