Abstract
Porous sintered silver is widely used as a die-attach material in high-power electronic packaging because its thermal transport capability strongly influences heat dissipation in SiC and other wide-bandgap power devices. During fabrication, the sintered silver layer develops a porous mesoscale structure in which a continuous silver backbone is interrupted by distributed pores, altering conductive pathways and reducing effective through-thickness thermal conductivity. However, most existing statistically generated porous models are constrained primarily by porosity and lower-order spatial statistics and do not explicitly preserve topological features such as backbone continuity, pore connectivity, and directional conduction pathways. As a result, the discrepancy between two-dimensional (2D) and three-dimensional (3D) representations of porous sintered silver remains insufficiently quantified. To address this limitation, this study develops a statistically controlled comparison framework for 2D and 3D porous sintered silver mesostructures. Structures with matched porosity and spatial correlation characteristics are generated using a Gaussian random field approach to ensure comparable lower-order morphology across dimensionalities. Local morphology metrics are first evaluated to verify statistical consistency, followed by the use of topological descriptors and weighted connectivity diagrams to characterize global connectivity and directional through-thickness transport pathways. Finite element heat-conduction simulations are then performed to evaluate effective thermal conductivity. The results show that, although the 2D and 3D structures are comparable in porosity and lower-order morphology, they differ significantly in topology and connectivity. Three-dimensional networks retain stronger backbone continuity and more alternative conduction paths, whereas 2D structures fragment earlier and show stronger bottleneck behavior as porosity increases. Consequently, thermal conductivity decreases more rapidly in 2D models, highlighting the role of dimensionality in mesostructure-informed thermal transport prediction.
| Original language | English |
|---|---|
| Title of host publication | 2026 27th International Conference on Electronic Packaging Technology, ICEPT 2026 |
| Publisher | Institute of Electrical and Electronics Engineers Inc. |
| Edition | 2026 |
| ISBN (Electronic) | 9798319531117 |
| DOIs | |
| State | Published - 2026 |
| Event | 27th International Conference on Electronic Packaging Technology, ICEPT 2026 - Xi'an, China Duration: 5 Aug 2026 → 7 Aug 2026 |
Conference
| Conference | 27th International Conference on Electronic Packaging Technology, ICEPT 2026 |
|---|---|
| Country/Territory | China |
| City | Xi'an |
| Period | 5/08/26 → 7/08/26 |
Keywords
- Connectivity analysis
- Gaussian random field approach
- Porous sintered silver
- Statistically controlled comparison
- Thermal conductivity
Fingerprint
Dive into the research topics of 'Connectivity and thermal transport in porous sintered silver: A statistically controlled 2D-3D comparison'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver