Paper MN-ThP7
Creation of Co-planar Oxide Pillars for Fabricating Overhanging Metal Structures

Thursday, November 12, 2009, 6:00 pm, Room Hall 3

Magnetic resonance force microscopy (MRFM) combines the nanoscale resolution of scanned probe microscopy with the three-dimensional, isotopically specific imaging capabilities of magnetic resonance imaging. The ultimate goal of MRFM is to achieve single-proton imaging resolution and create an atomic-resolution three-dimensional image of a individual molecule. At this level, the technique could achieve such feats as the structural determination of a single copy of a protein or macromolecular complex, making it a fantastic tool for biological study.

The key technology for MRFM is extremely sensitive, magnet-tipped cantilevers. While extensive effort has gone into fabricating such cantilevers, thermally-limited cantilever sensitivity is seldom achieved in practice because of surface-induced dissipation. The design of our cantilever minimizes this noise by extending the magnet past the cantilever tip. In our current cantilever fabrication scheme, we deposit the magnet on the device layer of a silicon-on-insulator wafer, and then create an overhanging magnet by using an isotropic sulfur hexafluoride etch to partially remove the silicon under the magnet. While successful, this process raises concerns over possible damage to the magnet from the etch species, and there is some degree of variability in the length of the overhang because of the very high silicon etch rate. To mitigate these issues, we have developed an alternative process in which the magnet is deposited over a pillar of silicon oxide extending through the device silicon layer and conformal with the top of this layer. This innovation removes the plasma etching step of our previous approach, and the length of overhang can be controlled by changing the lithographic placement of the magnet relative to the edge of this pillar. The pillar is created by localized oxidation of the device silicon, followed by chemical mechanical planarization. In this poster we will present our progress on this work, as well as present ideas for uses of our innovation well beyond cantilevers for MRFM.