Through silicon vias (TSVs) are the conducting pathways that transfer power and signals in 3D integrated technologies. They are made utilizing a variety of semiconductor manufacturing processes including deep reactive ion etching, electrochemical deposition, and chemical mechanical planarization (CMP). Notably, small perturbations in the electrochemical deposition rate leads to thicker overburden and excessive polishing downstream. We demonstrate unique methods for TSV electrochemical deposition endpoint detection of based on the current and voltage transients. Copper is typically used as the conductive fill of the TSV. Void-free, bottom up copper electroplating is required for this process, and is determined by the presence of additives to the copper bath. Small changes in the etch process, or plating additives can greatly impact the TSV filling rate, thus overplating is required to account for process variation. Furthermore, with each design iteration, TSV locations and/or size may change, effectively changing the filling process, which would require yet another set of filling experiments to define a new timing specification. Removal of the over-plating on top of the via, results in extra CMP processing. All these difficulties with TSV filling could be addressed if there as a method for determining when the via filling was complete.
Current and/or voltage transients during the electrochemical filling process provide a unique signature when the copper deposit reaches the surface of the wafer. What is more interesting is that these signals only vary slightly with TSV size, depth, and quantity, making this detection method suitable for automation in a variety of settings. This has allowed us to integrate these various signatures into a software reads the electrochemical out and detects when TSV filling has completed. Having this end-point detection has the potential to reduce the number of experiments needed (or even eliminate) to generate TSV filling protocols as well as reduce the amount of the CMP needed post filling. SNL is managed and operated by NTESS under DOE NNSA contract DE-NA0003525.
