1394
Process Optimization of CoSi2 Formation on P-Doped Poly-Si by Hot Wall-Based Rapid Thermal Annealing

Wednesday, 27 May 2015
Salon C (Hilton Chicago)
J. Y. Lee, H. J. Kim, E. J. Kim, H. S. Song, S. J. Yeom (SK hynix, Inc.), T. Ishigaki, K. Kang, and W. S. Yoo (WaferMasters, Inc.)
Low resistivity junctions and low resistivity contacts are essential for fabricating high performance devices with advanced device modes. Significant effort has been expended to develop low resistivity junctions, industry wide. However, low resistivity contact formation has not received significant attention, other than changing the silicide of choice, depending on required minimum line width or device node. Once the candidate silicide material is chosen for the device node, relatively small improvement in resistivity has been made in recent years, despite silicidation process optimization which include modifications of process steps and integration schemes.  

In this study, we have investigated the effect of annealing temperature and time on the resulting resistivity of Co2Si contacts with P-doped poly-Si using a single wafer furnace-based (hot wall) rapid thermal annealing (RTA) system. The hot wall RTA resulted in significantly (>20%) lower Rs, down to 19.6 ohm/sq. from equivalent RTA process using conventional tungsten halogen lamp-based (cold wall) RTA systems over a very wide process window (RTA temperature and time). Secondary ion mass spectroscopy (SIMS) dopant (P) depth profiling results revealed that the P atoms in the CoSi2 film and at the CoSi2/P-doped poly-Si interface, redistribute very differently under different RTA conditions. The same Rs value can be realized by different P profiles resulting from different RTA conditions. To enable contact resistance reduction, P pile-up near the CoSi2/P-doped poly-Si interface for and to suppress P depletion from P-doped poly-Si layer for resistivity reduction, lower temperature RTA for longer times in the hot wall RTA system was found to be beneficial. 

Wafers with almost identical P depth profiles after CoSi2 formation RTA in the cold wall RTA and the hot wall RTA systems, still showed an Rs difference of ~20%. The wafers annealed in the hot wall RTA system always showed lower Rs values indicating higher P dopant activation in the P-doped poly-Si layer. The difference in wafer heating mechanisms between the cold wall (lamp-based) RTA and the hot wall (furnace-based) RTA may have played an important role in the dopant activation and Rs differences. Hot wall RTA is very promising for realizing the low resistivity contacts required in advanced high performance devices. The simple change of annealing technique may extend the life cycle of CoSi2/P-doped poly-Si contacts. 

The Rs response surface (as a function of RTA temperature and time) and SIMS depth profiles of P atoms will be presented at the meeting.