- Conference date: 23-27 March 1998
- Location: Gaithersburg, Maryland (USA)
In the Front-end (transistor forming) area of silicon CMOS device processing, several NTRS difficult challenges have been identified including; scaled and alternate gate dielectric materials, new DRAM dielectric materials, alternate gate materials, elevated contact structures, engineered channels, and large-area cost-effective silicon substrates. This paper deals with some of the characterization and metrology challenges facing the industry if it is to meet the projected needs identified in the NTRS. In the areas of gate and DRAM dielectric, scaling requires that existing material layers be thinned to maximize capacitance. For the current gate dielectric, and its nitrided derivatives, direct tunneling will limit scaling to approximately 1.5nm for logic applications before power losses become unacceptable. Low power logic and memory applications may limit scaling to the 2.0–2.2nm range. Beyond these limits, dielectric materials having higher dielectric constant, will permit continued capacitance increases while allowing for the use of thicker dielectric layers, where tunneling may be minimized. In the near term silicon nitride is a promising substitute material while in the longer term “high-k” materials such as tantalum pentoxide and barium strontium titanate (BST) will be required. For these latter materials, it is likely that a multilayer dielectric stack will be needed, consisting of an ultra-thin (1–2 atom layer) interfacial layer and a high-k overlayer. Silicon wafer surface preparation control, as well as the control of composition, crystal structure, and thickness for such stacks pose significant characterization and metrology challenges. In addition to the need for new gate dielectric materials, new gate materials will be required to overcome the limitations of the current doped polysilicon gate materials. Such a change has broad ramifications on device electrical performance and manufacturing process robustness which again implies a broad range of new characterization and metrology requirements. Finally, the doped structure of the MOS transistor must scale to very small lateral and depth dimensions, and thermal budgets must be reduced to permit the retention of very abrupt highly doped drain and channel engineered structures. Eventually, the NTRS forecasts the need for an elevated contact structure. Here, there are significant challenges associated with three-dimensional dopant profiling, measurement of dopant activity in ultra-shallow device regions, as well as point defect metrology and characterization.
MOST READ THIS MONTH
MOST CITED THIS MONTH
Article metrics loading...