Advanced MOS Device Physics by Norman G. Einspruch
By Norman G. Einspruch
Contains contributions from a dozen pros from the inner most quarter and academia. Discusses a number of machine physics themes of specific curiosity to and college researchers in electric engineering, computing device technological know-how, and digital fabrics. Emphasizes actual description, mode
Read or Download Advanced MOS Device Physics PDF
Best engineering & transportation books
Extra info for Advanced MOS Device Physics
First, the implant also increases the doping level under the p s o u r c e / d r a i n diffusions, increasing the junction capacitance; this is dis cussed further in the following section. The second drawback is that addi tional phosphorus makes the /7-channel transistors more strongly enhance ment mode. For example, the devices used for the measurements shown in Fig. 8 had conventional n-doped polysilicon gates, and this resulted in thresholds of nearly - 2 V with the highest doses. This is too negative for high-performance C M O S applications, since the high absolute values of threshold would result in low drive current and hence low circuit speed.
S U M M A R Y The benefits and the drawbacks of device scaling cannot be properly visualized without a coherent accurate physics-based M O S transistor model. In this chapter, the essential features of one such model are presented. In drain current modeling, the formulation is greatly simplified by incor porating only the dominant scaling effects, namely mobility degradation due to the oxide field and velocity saturation, without sacrificing accuracy. Based on the modeling and experimental results, a device is more "longchannel like" if it has a long channel length or thinner gate oxide.
One method to make /7-channel thresholds less strongly enhancement mode is to use a gate material with a higher work function, such as p-doped polysilicon . However, the incorporation of p polysilicon into a C M O S process presents several difficulties in terms of process integration. First, boron from a heavily d o p e d polysilicon gate can more easily penetrate a thin gate oxides than can p h o s p h o r u s , leading to potentially u n c o n t r o l l e d threshold shifts. Second, if n-doped polysilicon is still to be used with the n-channel devices, then a means of achieving good ohmic contacts between the two gate materials must be found.