Next steps revealed for 32nm mode
The demands of the semiconductor market are constantly evolving as end user devices, for example smart phones, become smaller, thinner and lighter and are expected to offer much greater integrated functionality, yet are required to use less power to extend battery life. This market explosion continues to push semiconductor manufacturers to increase performance and reduce overall footprint, yet remain cost-efficient. The competition is increasingly fierce. The advanced architectures of memory devices is driving development in materials and processes, resulting in greater complexity in device manufacturing, especially at the gate level where molecular chemistry takes place. ‘Traditional’ semiconductor materials no longer offer the performance required. With the industry moving towards the 32nm mode and beyond, chemical innovation and supply chain collaboration are assuming greater importance and are key enablers in the development and implementation of future semiconductor technologies.
At SEMICON West recently, SAFC Hitech announced a new five-year chemical roadmap for Metalorganic Chemical Vapour Deposition (MOCVD) and Atomic Layer Deposition (ALD) processes on silicon semiconductor substrates. This new roadmap for silicon semiconductors adheres to the International Technology Roadmap for Semiconductors (ITRS) guidelines and maps an extensive programme of materials development across a number of layers within the semiconductor. These include materials for high-k dielectrics in logic and memory devices, additional functional memory architectures, electrodes in DRAM or gate stacks, barrier layers, wiring, and low-k dielectrics. The roadmap was introduced to encompass the wide range of chemicals and advanced molecules now being researched, developed, tested and introduced for large-scale production in the semiconductor markets over the next five years.
The expansion of the periodic table
We are experiencing an expansion of use of the elements in the periodic table, that is, a significant increase in the numbers of complex molecules being used in the deposition processes at the wafer level. Silicon oxides and other traditional materials have long been used as pre-metal dielectrics, but as the semiconductor industry moves from the 65nm node through 45nm, 32nm and beyond, the demands placed on the electro physics of the silicon device require the development of new enabling chemistries.
In our experience, the semiconductor market is entering a true ‘age of chemistry’, where continual materials evolution and the development of novel materials utilising previously untried elements will be vital to enable future technology nodes. We are already manufacturing precursors for the deposition of oxides and binary oxides such as aluminium oxide, hafnium oxide, hafnium silicate and zirconium oxide, as well as complex rare earth oxides in our production processes. These metal oxides have a very high-k value and will continue to be widely exploited in the next couple of years if the industry stays on course with Moore’s law.
Over the course of the new product roadmap, we plan to introduce more complex high-k oxides for silicon semiconductor manufacturing, such as hafnium zirconium based layers, which offer greater flexibility as they can be doped with other materials like silicon, nitrogen, aluminium, lanthanum and yttrium to meet individual customer requirements in creating a layer that functions well for a particular device design. Beyond that, the research and development of, for example, lanthanide and strontium chemistries, binary metals and complex metals oxides or iterations of oxides will facilitate the delivery of the 50+ k values needed for future technology nodes.
Collaboration = seamless delivery
In addition to the continuing developments in the type of properties of the materials used in semiconductor manufacturing, greater collaboration between process engineers, tool manufacturers and chemical companies is taking place to ensure a seamless process for the safe delivery of these new molecules to the wafer’s surface. ALD, the current ‘technology of choice’, works via a chemical reaction of the molecule with an oxidant or reductant, where the reaction is more important than the temperature (unlike CVD). However, these new complex molecules can be volatile and seriously affected by thermal issues. Therefore, there is a critical need to maintain chemistries at lower temperature if possible to prevent decomposition of these molecules before they reach the process chamber. The key is that molecules cannot break down too early, and must remain stable in order to react.
Materials issues will continue to be addressed with novel chemistries, which are likely to not be as stable as they were when semiconductor manufacturing used thermally stable gases only. Higher temperatures increase the ability in deposition chemistry to achieve increased uniformity, yet such temperatures mean molecules are more likely to decompose. New vaporisation methods and materials, allied with advances in processes that deliver these materials to the process chamber as quickly as possible, must be created to adopt and maintain the industry’s rate of progress and meet its expectations to ensure minimal degradation, uniformity and high purity across increasingly pronounced trenches and stacks, while reducing contamination and gate leakage.
Looking to the future
The materials that are anticipated to be key in high-k dielectrics and other application areas are those that possess the proper physical chemical characteristics and which meet or exceed the correct requirements in deposition capability across all device architectures where high deposition rate is critical. These materials will possess the appropriate electrical properties and must be offered at an acceptable cost to consumers. The physical challenge facing the industry is to minimise particle contamination and maximise uniformity when earlier process steps introduce uncontrolled decomposition of precursors or poor maintenance of the tool, chamber or tubing. At the higher level, we expect to see increasing attention to shrinking device geometries further and increasing the capability of the surface.
The integration of new materials requires the continued development of process steps such as vaporisation, in line with advances in molecular design. Collaboration between SAFC Hitech, tool manufacturers and specialised vaporiser manufacturers is already underway to introduce advanced precursor materials that enable new thin films at the sub-32-nm technology node to be deposited. Such new materials, allied with the advanced processes that deliver them into the process chamber as quickly as possible, enable industry expectations to continue to be met. We expect to see increasing specialisation in vaporiser technology, for example, ‘flash’ vaporisation, as this aids to prevent the molecules from degrading in the wrong location, and tool design from both bespoke manufacturers and by tool makers ‘in-house’. Expect a mix of both, allied with customer input to win the market. For the time being at least, ALD will remain the best technology for achieving these goals.
We fully expect to witness an increased commitment to this turnkey collaborative approach, with increasing levels of co-operation between chemical, tool and vaporiser companies and end customer IDM’s via multiple joint development programs.
