Porous Organosilicates for On-Chip Applications: Dielectric Generational Extendibility by the Introduction of Porosity
Dense alkyl-functionalized organosilicates have dielectric constants that are 25-30% lower than silica, which allows ultralow dielectric targets(κ< 2.2) to be achieved at reduced porosity levels relative to those required for silica. Partially condense
- PDF / 639,394 Bytes
- 36 Pages / 439.37 x 666.142 pts Page_size
- 105 Downloads / 147 Views
e unrelenting drive toward decreasing device dimensions and increasing on-chip densities will result in increasing signal delays due to capacitive coupling and crosstalk in the back-end-of-line (BEOL) interconnect wiring [1]. The RC delays depend on the resistivity of the wiring metal, the metal dimensions, and the dielectric constant of the insulating media. While the switch P.S. Ho et al. (eds.), Low Dielectric Constant Materials for IC Applications © Springer-Verlag Berlin Heidelberg 2003
168
W. Volksen et al.
from aluminum metallurgy to copper [2] has provided some relief, this effect will be short-lived without a concurrent decrease in the dielectric constant of the insulating media. The latter has fueled a somewhat frantic search for new low-k materials to replace the traditional silicon dioxide insulators [3]. Assuming a dielectric scaling that is commensurate with the generational decrease in wiring dimension and pitch, dielectric constant targets of 2.6–2.9 and 2.0–2.2 for the 0.13 and 0.10-μm technology nodes, respectively, are required. Fortunately, there are a plethora of dielectric materials with dielectric constants in the 2.6–2.9 range including both CVD and spin-on candidates [3]. These materials may be roughly characterized as either inorganic-like or organic polymers. Virtually all of the CVD materials are inorganic-like consisting mainly of carbon-doped oxides of silicon of various composition and structure. The viable spin-on candidates include both inorganic-like and organic polymers. An increasingly important consideration for future technology nodes is the dielectric extendibility to materials with lower dielectric constants. For dielectric constants in the range of 2.0, the only available homogeneous alternatives include highly fluorinated alkane derivatives and fluoroethers such as Teflon [4] and Teflon-AF [5]. For unfluorinated systems the only foreseeable route to materials with dielectric constants of 2.0 or below involves the introduction of porosity. This route is currently a more viable option for spin-on rather than CVD candidates. Porous materials provide a number of porosity-specific integration challenges including chemical, mechanical, and electrical properties as well as thermal management. To mitigate these issues, the porosity should be homogeneously distributed and the pore sizes small with respect to the minimum device dimensions. For future generations, this will require pore dimensions of 200 ˚ A or less. Ideally, the porosity should also be closed-cell (noninterconnecting) since the uptake of water or environmental contaminants during integration must be minimized, although this may be difficult for porous materials with dielectric constants significantly less than 2.0. The superposition of porosity-specific requirements and the usual integration challenges associated with the introduction of new materials provide daunting obstacles for material scientists and integration engineers. The drop in the dielectric constant of porous materials arises from the decrease in film density cau
Data Loading...
