Predictive Simulations and Experimental Study of AlN Bonding for Better Thermal Dissipation in BSPDN

As the backside power delivery network (BSPDN) technology evolves, optimizing bonding materials for both thermal management and integration remains critical. In this study, we explore the optimal thickness conditions and integration strategies for aluminum nitride (AlN) bonding through simulations and experimental validation. Simulations were conducted with realistic power densities, chip layout, and package structures to evaluate multiple combinations of oxide and AlN thickness options. While oxide bonding is limited in thermal dissipation by its thermal conductivity, reducing its thickness clearly improves performance. However, driving the bond oxide thickness down has a lot of integration challenges. Since AlN thermal performance is less sensitive to thickness, it can be a promising alternative. To assess AlN bonding integration a BSPDN test vehicle was used to study the characteristics of AlN crystallinity, its interfacial composition, its bond strength, interface quality, stresses from thermal mismatch, overlay (OVL) residuals, and overall integrability. These results establish a path to integrating AlN as a viable bonding material for BSPDN.