Answers are grounded in this paper and cite their sources.
Low-temperature Cu-Cu direct bonding (<150°C) for 3D integration faces a severe bottleneck from interfacial voids caused by limited atomic diffusion. While nanocrystalline Cu (NC-Cu) has emerged as a promising bonding material owing to its high stored energy, the thermodynamic influence of initial crystallographic orientation remains unclear. Here, we systematically investigate the orientation-dependent bonding thermodynamics of NC-Cu thin films with (100), (111), and (101) orientations. Despite high internal energy, (100) and (111) Cu consume significant energy for internal defect curing (dislocation glide and twin annihilation), inducing void aggregation. Conversely, (101) Cu, despite not possessing the largest absolute total driving force, directs the largest fraction of its available positive driving force toward macroscopic grain-boundary migration, thereby achieving extensive interfacial-crossing growth and enhanced overall mechanical reliability. We demonstrate that reliable bonding requires targeted allocation of grain boundary energy, establishing a novel crystallographic design principle for advanced heterogeneous integration and solid-state joining technologies.
No takes yet. Share an insight, caveat, or question.
Han et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: