This Research award in the Inorganic, Bioinorganic and Organometallic Chemistry program supports work by Professors Lisa McElwee-White and Tim Anderson at the University of Florida to develop processes for the growth of thin films of barrier material for Cu metallization used in the manufacture of advanced integrated circuits. The initial target material is ruthenium, which can serve in interfacial layers for seedless Cu metallization. Mechanism-based design accompanied by computational chemistry allows evaluation of Ru-containing film growth precursors before synthesis. An important aspect is the use of alternative precursor delivery systems designed for all molecules independent of volatility, thus expanding the range of possible precursors. Following preparation of potential precursors, rapid screening selects the most promising candidates for film deposition and characterization. Collaboration with industry facilitates process development and optimization including reactor modeling. Graduate students are trained in interdisciplinary team-based problem solving in chemistry, chemical engineering and materials science. Industrial internships, mentoring of undergraduate researchers and interactions with industrial research personnel are used to broaden the students' perspectives. A workshop that one of the PIs (Anderson) has presented on career development for new and prospective faculty for the last 6 years at national meetings is being adapted for presentation to chemistry faculty.
The deposition of thin films is a key technological capability. Innovation often relies on the ability to deposit films with very specific composition, surface morphology, microstructure, thickness, and interfacial chemistry. Developing a process for deposition of a barrier material for Cu metallization will address a current challenge in the manufacture of advanced integrated circuits.
As feature sizes on computer chips become smaller and smaller, new materials and manufacturing techniques become necessary. Several interesting methods such as chemical vapor deposition (CVD), involve synthesizing materials in place by running chemical reactions on the wafer during chip processing. In order to prevent damage to the chip during processing, minimizing the reaction temperatures is critical. In this NSF-funded project, we developed new reactions for depositing diffusion barrier materials on copper-metallized integrated circuits. Diffusion barriers are necessary to prevent shorting out of the chip by migration of the metal. Tungsten nitride (WNx) and tungsten carbonitride (WNxCy) are candidate diffusion barrier materials. Previously reported co-reactant systems for deposition of WNx employed volatile tungsten compounds such as WF6 or W(CO)6 with ammonia as the nitrogen source. While cost-effective, these systems generally require high deposition temperature (> 400 °C), result in the production of corrosive byproducts, and lead to the incorporation of undesirable impurities in the films. Single source precursors for WNxCy all had the disadvantage of high deposition temperatures (400 – 600 °C). We have been developing chemical vapor deposition methods to deposit WNxCy films from tungsten imido, hydrazido and nitrido complexes at low temperatures as an alternative. Nitrido complexes proved to be the best precursor compounds. We reported the CVD of WNx nanospheres from WN(NMe2)3 at temperatures as low as 75 °C. Continuous WNxCy thin films deposited from WN(NMe2)3 at temperatures as low as 125 °C passed the diffusion barrier test (no passage of copper after annealing at 500 °C for 30 minutes). This result represents a lowering of WNxCy deposition temperature by almost 300 °C, when compared to precursors reported before this project.