next up previous contents
Next: 6.4 Comparison between Low Noise, Power, and Millimeter Wave HEMTs Up: 6.3 Millimeter Wave HEMTs Previous: 6.3.2.5 Dependence on the Passivation Thickness


6.3.3 Optimization of Millimeter Wave HEMTs

The optimization of DC and RF performance basically requires a maximized gm and minimized CGS and CGD. Based on the results presented in this section it appears unrealistically that the maximum gm of over 800 mS/mm could be increased significantly by a reduction of dGC still providing good reproducibility of the manufacturing process.

The investigations of the dependence on LR revealed that it is favorable to increase LR to about 130 nm in order to obtain fT and fmax between their corresponding maxima. LR should not be chosen larger if the passivation thickness is reduced because a reduction of dP shifts the maxima towards smaller LR. For HEMTs on InP LG of 100 nm are commonly used [71, 81]. With such a gate length CG could be reduced by about 100 fF/mm if a HEMT with dCG = 10 nm is considered. For this case gm is expected to remain constant.

Provided a tight process control the passivation could be reduced to a thickness of about 150 nm which would leave a residual passivation thickness on the cap layer of 50 nm. The improvements of CGS and CGD due to an increase in LR and a reduction in dP can be estimated only roughly from the previously shown characteristics. Simulation of the millimeter wave HEMT with dGC = 10 nm, LG = 100 nm, LR = 130 nm, and a total passivation thickness of dP = 150 nm revealed gm = 773 mS/mm, fT = 136 GHz and fmax = 322 GHz at the bias point VDS = 2.0 V and VGS = 0.8 V.
 



next up previous contents
Next: 6.4 Comparison between Low Noise, Power, and Millimeter Wave HEMTs Up: 6.3 Millimeter Wave HEMTs Previous: 6.3.2.5 Dependence on the Passivation Thickness

Helmut Brech
1998-03-11