next up previous contents
Next: 6.3.2.3 Dependence of fT on dGC Up: 6.3.2 RF Performance Previous: 6.3.2.1 Dependence on the Gate Length


6.3.2.2 Contributions to the Gate Capacitance

To investigate the dependence of CG on dGC and LG the different contributions to CG are separated by simulation according to (62).
 

 
Figure 6.45 Simulated CG versus LG for dGC = 10 nm and dGC = 13 nm. To deduce the parameters A1, A2, and A3 for both dGC the passivated and unpassivated case is shown.
 

In Section 6.2 a possible reduction of A2 by optimization of the T­gate shape was discussed. A change of dGC is expected to have only an impact on the parameters A1 and A2 which are attributed to the epitaxial structure. In  Figure 6.45 the simulated CG versus LG is shown for dGC = 10 nm and for dGC = 13 nm both for the passivated and unpassivated case. As demonstrated in Section 6.1.2.2 and 6.2.3.2 the parameters A1, 10, A2, 10, and A3, 10 for the device with dGC = 10 nm and A1, 13, A2, 13, and A3, 13 for the device with dGC = 13 nm can be deduced. The parameters for both dGC are given in Table 6.5. To investigate the dependence of CG on dGC and LG the different contributions to CG are separated by simulation according to (62).
 

 
Table 6.5 Parameters of the contributions to CG.
A1 [fF/mm]
A2 [fF/mm]
A3 [nF/mm2]
dGC = 10 nm
114
68
5.1
dGC = 13 nm
123
69
4.3
 

The fringe capacitances purely attributed to the epitaxial layers A1, 10 is slightly lower than A1, 13. As expected the values of A2 which are related to the coupling of the contacts are very similar. But for A3 substantial differences occur. Due to the smaller dGC A3, 10 is greatly increased to 5.1 nF/mm2 compared to A3, 13 = 4.3 nF/mm2. Thus to improve fT by a reduction of dGC a large part of the improvements of gm is compensated by the significant increase in CG.
 

next up previous contents
Next: 6.3.2.3 Dependence of fT on dGC Up: 6.3.2 RF Performance Previous: 6.3.2.1 Dependence on the Gate Length

Helmut Brech
1998-03-11