Research Interests

Complex 3-D structures have many technologically relevant applications such as increasing transistor densities in microelectronic devices while reducing interconnect RC time delay and power consumption, increasing efficiencies of thin film solar cells, and comprising microelectromechanical systems.  A common way to fabricate 3-D structures is to successively transfer films from a handling wafer to a substrate on which the 3-D stack is constructed.  This transfer of films requires two fundamental steps: bonding and releasing.  All steps along the process require precise control of the fracture properties of the thin films, both to ensure mechanical integrity and reliability of the final fabricated device and to ensure the proper release of the handling substrate at the appropriate time.  Both the bonding and releasing processes in this research are novel mechanisms that exploit surface interactions between neighboring layers in the thin film stack and allow precise control of the fracture properties of the thin films in the stack.

Figure 1. Schematic of “bond and release” process to transfer IC layer from a handling wafer to a stack constructed on a base Si substrate with several previously transferred IC’s.

The bonding process uses the aluminum-induced crystallization (AIC) of amorphous silicon to produce a thin, strong bond at low temperatures using materials already common to the semiconductor industry.  This novel bonding technique is effective at temperatures as low as 270 °C, approximately 100 °C lower than temperature typically used for copper bonding.  Furthermore, the potential exists for this bonding technique to be effective at even lower temperatures since it is known that AIC of amorphous silicon can occur as low as 150 °C.  With a fracture energy of approximately 10 J/m², the mechanical integrity of this bond is superior to those of other common low temperature bonding techniques such as dielectric bonding and direct bonding.  The current focus of this research area is developing strategies to lower the temperature requirements further.

Figure 2. SEM of bonding by Al-MIC of a-Si showing a) a region with Al where crystallization and successful bonding occurred, and b) a region without Al.

The releasing process takes advantage of the attraction of hydrophilic porogen (pore generating molecules) particles that segregate to hydrophilic substrates.  The relatively volatile porogen decomposes around 300 °C and leaves behind porosity when it is removed from an organosilicate matrix.  The segregation of the porogen to the substrate interface is shown clearly through thin-film adhesion studies and secondary ion mass spectroscopy.  Moreover, fracture experiments and thermodynamic modeling of the system both indicate that this process is most effective at low porogen loadings.  The fact that the releasing mechanism is based on porogen burnout suggests that it can be thermally activated when the adhesive fracture energy decreases as porogen is removed.  Continuing work will attempt to control the chemistry of the templating particles to tune the fracture energy of the release layer and the temperature at which the releasing mechanism activates.

Figure 3. Release process showing a) the porogen templates distributed in the dielectric layer, b) controlled template segregation to interface, c) interface release with template burnout, and d) the dielectric layer and interface fracture energies.