4.2.1 Fabrication of the Simulated Devices

All four HBT structures are MOCVD epitaxially grown on GaAs substrates. The first device, later on referred to as Dev. 1, has the following layer sequence: The cap is formed by an n$^+$-InGaAs/n-GaAs layer. The emitter consists of 80 nm Si-doped n-Al$_{0.28}$Ga$_{0.72}$As with 20 nm graded layers on its top and bottom, respectively. The 120 nm GaAs base is carbon doped (p$^+$ = 3.10$^{19}$ cm$^{-3}$), followed by a 700 nm GaAs collector (n$^-$ = 2.10$^{16}$ cm$^{-3}$) and a 700 nm GaAs subcollector (n$^+$ = 5.10$^{18}$ cm$^{-3}$). The simulated device structure is shown in Fig. 4.13.

Figure 4.13: Simulated device structure of AlGaAs/GaAs HBT

In two devices (Dev. 2 and Dev. 3) the graded AlGaAs layer next to the base-to-emitter junction is replaced by a 20 nm InGaP ledge layer (n = 4.10$^{17}$ cm$^{-3}$ and 3.10$^{17}$ cm$^{-3}$, respectively). Both devices differ in base carbon doping (p$^+$ = 3.10$^{19}$ cm$^{-3}$ vs. 4.10$^{19}$ cm$^{-3}$) and layer thickness (80 nm vs. 120 nm). A 40 nm InGaP emitter (n = 4.10$^{19}$ cm$^{-3}$) HBT with 120 nm base thickness is the last considered device type (Dev. 4). All HBTs are processed by etching a double mesa structure and passivated by Si$_3$N$_4$. The simulated device structure is shown in Fig. 4.14.

Figure 4.14: Simulated device structure of InGaP/GaAs HBT