Paper TF-ThP34
Combinatorial ALD to Accelerate Integration of New Materials for Nanomanufacturing

Thursday, October 31, 2013, 6:00 pm, Room Hall B

New materials and new device architectures have created an increasing need for Atomic Layer Deposition (ALD) of nano-scale conformal films to be incorporated into the manufacturing process flow. The conformality of ALD films is required to form some of the functional layers in today’s leading-edge planar IC devices, and in three-dimensional (3D) multi-gate CMOS devices such as “finFETs.” However, especially at the beginning of an R&D evaluation, ALD research involves expensive and exotic precursors that may only be available in small quantities. Therefore, there is a need to be as efficient as possible when exploring the potential use of a new ALD precursor for commercial manufacturing.

This drive towards efficiency often leads research in the direction of small volume, “coupon”-based reactors to attempt initial learning. Unfortunately, often results don’t translate from small reactors to full 300mm diameter silicon wafers, and this problem is only expected to become worse as the industry transitions to 450mm diameter wafers. In addition, performing experimentation at 300mm allows for integration of ALD films with other state-of-the-art processes to generate the stacks of materials necessary to extract the key electrical data necessary to evaluate device performance. Therefore, there is a pressing need for a more efficient approach that allows ALD R&D to occur on full 300mm diameter wafers.

One such approach is to use a high productivity combinatorial (HPC™) platform—such as that developed by Intermolecular, Inc.—to dramatically accelerate R&D of integrated ALD films by 10-100x relative to traditional methods. Combinatorial ALD can be integrated with combinatorial processes for wet-etching/cleaning and physical vapor deposition/chemical vapor deposition (PVD/CVD) to engineer interfaces and multilayer stacks for optimal device performance.

In one Collaborative Development Program with an Intermolecular customer, the initial design-of-experiments to integrate an ALD film with novel substrates, overlayers, and annealing resulted in over 3500 required tests to completely explore the potential solutions-space. It was estimated that a conventional R&D approach would have taken 2 years while using the HPC platform allowed the team to identify a combination of materials with the desired properties in just 7 months with a 75% reduction in raw materials expenses for ALD precursors and wafers.