PhD - Leuven | More than two weeks ago
Three-dimensional integrated circuits (3D-ICs) offer the possibility to solve delay and power problems in the conventional planar 2D-ICs. The so-called memory wall, the difficulty of accessing data fast enough, could be overcome by the application of different stacking options enabled using the third dimension. One attractive option which allows the device stacking at room temperature and the possibility to reach high alignment accuracy is direct wafer to wafer bonding (i.e., hybrid bonding).
Direct wafer bonding is realized at room temperature without any additional intermediate layer but only exploiting Van der Waals forces establishing between extremely clean and smooth dielectric surfaces which are brought in contact after a so-called activation process consisting of a plasma and clean processes.
It has been proved that this activation process drastically improves bonding performance but still a complete understanding of the reasons behind that are still to be described. Different activation processing can improve bonding strength but at the same time can also impact the surface of the metallic connections. Understanding the effect of dielectric activation combined with the passivation of the exposed Cu surface will be of prime importance. In general, a fundamental insight in the physical mechanisms beyond direct bonding will need to be understood.
Also, an extremely parameter that still needs to be assessed is the level of cleanliness especially when looking at applications as CFET where the interface of the bonding dielectrics will be part of a transistor. Impact of particles at the interface should be studied considering the current limitations of the characterization tools available in the field.
Another challenge of hybrid bonding is the formation of edge voids formed when bringing wafers together. Understanding of the void formation mechanism (wafer morphology, water contribution...) will also be important to optimize the hybrid bonding technology even further.
In this PhD research, you will enhance activation routines to improve hybrid bonding and try to unravel the physical mechanisms behind surface activation processes and exact formation mechanism of edge voids. Furthermore, you will investigate the impact of particles at the interface of the bonding dielectrics, engineer test material and characterization routines to define cleanliness levels enabling the bonding.
You have experience or are open to learn new characterization techniques (e.g. FTIR, water contact angle, bond strength techniques...) to realize an increased fundamental understanding of the mechanism of hybrid bonding. Scripting and programming skills are a plus.
You will interact closely with researchers at imec and KU Leuven.
Required background: Physics, Engineering technology, Chemistry or equivalent
Type of work: 20% literature study, 10% modelling, 70% Experimental work
Supervisor: Stefan De Gendt
Co-supervisor: Serena Iacovo
Daily advisor: Serena Iacovo
The reference code for this position is 2023-002. Mention this reference code on your application form.