The huge voltage range to be covered by semiconductor devices already suggests that it is not possible to satisfy all requirements with a singe device type. In fact, a wide range of different devices is used. Today, for switching applications it is commonplace to use the two major device types MOSFET (Metal Oxide Semiconductor FET) and IGBT (Insulated-Gate Bipolar Transistor) . The latter one is especially used for voltages above 300V. Currently IGBT modules are available for blocking voltages of up to 6.5kV [20,21] and with a maximum total current of up to a few thousand ampere . For the sake of completeness, both devices will be briefly described, despite this work being focused on MOSFET devices in automotive applications where IGBT devices are rarely used.
The principle of field effect transistors (FET) is to control a current in a solid semiconductor by an electric field. This concept was first proposed by Lilienfeld  in two patents [24,25] granted in 1930 and 1933. The functionality was confirmed in 1948, but due to interface and surface problems and due to the invention and success of the bipolar junction transistor (BJT) the field effect transistor was not further pursued. The introduction of thermally-grown silicon dioxides in 1959  paved the way for MOSFET devices. Compared to the BJT, the advantages of MOS devices are the simpler processing and the better scalability along with reduced power consumption . However, the real breakthrough came with the increasing demands on integrated circuits and the CMOS (Complementary MOS) fabrication methods .
In the power and high-voltage domain the BJT was widely used until the 1980's . The major drawback of the BJT in this regime is the low current gain which required complex and expensive control circuits to generate the base current. These circuits required additional power and the increased heat dissipation was a big issue. Also the introduction of mobile devices increased the demand for higher effectiveness. The attempt of increasing the current gain in BJTs leads to lower breakdown voltages and is therefore also no universal solution. On the other hand, field effect transistors are voltage controlled and no static control current is required. This helped to overcome the control circuit problems. An additional advantage of MOSFET devices is that there is no second breakdown. Higher temperatures lead to a decrease of the carrier mobility, and consequently, the drain current is reduced. This results in a reduced power loss and heat generation. Therefore, in contrast to BJTs, MOSFET devices can be simply connected in parallel , which is the basis for the design of power MOSFET structures.
These power MOSFET devices consist of numerous single MOSFET structures connected in parallel on chip level. At first, the structures were built with a V-shaped gate etched into the silicon giving the device the name VMOS (see Fig. 2.1(a)). In this fabrication method it was problematic to produce stable threshold voltages. Another problem were the high electric fields near the bottom peak of the oxide which degraded the breakdown voltage. An evolution of the VMOS is the UMOS (see Fig. 2.1(b)) where the gate is etched in a U-shape, resulting in a channel current flow vertical to the chip surface. This structure avoided the peak electric field that appeared at the bottom of the V-shaped structure. The UMOS is fabricated using trench etching which was originally introduced for the fabrication of highly integrated dynamic memory cells.
A different fabrication method for the channel structure is used in the double-diffused MOSFET or simply DMOS (see Fig. 2.2). Here, the channel area is built by lateral diffusion of the n- and p-dopants, both masked by the gate. The channel length is determined by the diffusivity of the dopants and by the diffusion process temperature and time. Lithographic limitations do not influence the minimum channel length, since it can be adjusted by changing temperature and time of the diffusion process. Since the gate can be used as mask for the channel diffusion this process is self aligning and a high precision can be obtained.
The insulated gate bipolar transistor (IGBT)  combines the high input impedance of a MOS transistor and the high current densities which are possible due to the bipolar current transport. A very low sheet resistance can be achieved which leads to a low voltage drop and therefore reduces the thermal power generated in the device. In Fig. 2.3 two different forms of IGBT structures are shown. The first structures proposed were fabricated using the double-diffusion techniques as it is used for DMOSFET. The other structure introduces the trench gate of the UMOS to the IGBT and helps to further decrease the on-state voltage drop. The main advantage of the IGBT compared to the thyristor and the gate turn off transistor (GTO) is the capacitive gate control. For low frequencies, this results in a nearly power-less voltage control. Therefore, complex driving circuits which are especially required to turn off a GTO can be avoided.
Comparing the IGBT and the MOSFET in high-voltage applications shows that lower specific sheet resistances in the on-state can be achieved using the IGBT. However, the switching behavior of the bipolar device is much slower compared to MOSFET devices. This is caused by the high density of minority carriers in the drift region coming from the collector while the transistor is switched on. The time needed to switch of the IGBT is therefore mainly determined by the recombination process and therefore the carrier lifetime. The main application field of IGBT devices is for a voltage range above 300V . Since this work focuses on smart power devices in automotive environments the operating voltages are mostly far below 100V. IGBT structures therefore have only little importance for this work.