MRS Spring Meeting location:San Francisco date:24-28 March
Spin-on pure-silica-zeolite MFI films are promising low-k candidates to replace materials currently used in microelectronics manufacturing. Prepared from a nanocrystal suspension containing residual amorphous silica, these films hold mechanical properties superior to other organosilicate low-k candidates. In this study, we have systematically synthesized these zeolite films with different crystallization times in order to alter the proportion of zeolite nanocrystals to residual amorphous silica. An extended characterization of these films shows a trade-off, which is crystallization-time dependant, between the pore size and the dielectric constant. We demonstrate that ultraviolet-assisted curing moderates this trade-off by enhancing the cross-linking of the matrix and, most importantly, by minimizing the hydrophilicity of the films. However, this synthesis approach provides no optimal solution to obtain the necessary pore size and dielectric constant. In addition, we present further evidences that this approach is intrinsically limited to fulfill all the requirements for low-k materials. For example, synthesized films lack homogeneity at a length scale of few tens of nanometers and the achieved roughness is comparable to the interconnect feature sizes. We propose a different strategy that exploits advantages of these zeolite films such as high porosity and stiffness but overcomes their intrinsic limitations. It consists of a sol-gel process with tetraalkoxysilanes and alkyltrialkoxysilanes in presence of organic templates commonly used in zeolite synthesis. Variation of synthesis media (acidic or basic) and flexibility in choice of zeolite organic templates allows a fine tuning of the desired properties. The role of the zeolite organic templates is essentially influencing the final properties of the films. Up to now synthesized low-k films can have dielectric constants as low as 2.0 in combination with small pores, hydrophobicity, and excellent mechanical properties. Further improvement of these results is expected. We are convinced this work will open new avenues for the synthesis of porous films for low-k applications, as well as for other emerging technologies.