Renewable energy systems generate electrical energy in different forms. In order to match the source and load, power conversion circuits are required. In power conversion, efficiency is the key factor. High efficiency results in low power loss, hence the compact systems. In modern power conversion circuit semiconductor devices are used as switches. Power diodes, bipolar junction transistors (BJTs), metal oxide field effect transistors (MOSFETs), insulated gate bipolar transistors (IGBT), silicon-controlled rectifiers (SCRs), gate turn-off thyristors (GTOs), and integrated gate commutated thyristors (IGCTs) are the most widely used power devices in the industry. These devices serve a broad spectrum of power levels and frequencies [1]. In high-power and medium-voltage systems (>2.2 kV) such as high voltage DC (HVDC), electric trains, medium-voltage drives, wind farms, and large solar plants, power levels range from a few megawatts to a few hundred mega watts, and switching frequency ranges from 50 Hz to 1 kHz. SCRs, GTOs, and IGCTs are used at this power level [2]. In high performance drives, electric vehicles, and standalone renewable energy sources, the switching frequency is in the range of a few kHz to 500 kHz, and the power level ranges from 10 kW to 10 MW. IGBTs are the most popular in this power range and MOSFETs are used at low voltage applications [2,3]. In communication systems, audiovisual equipment, biomedical instruments, computers, etc., the power level ranges from a few megawatt to, say, a few kilowatts, and the switching frequency ranges from hundreds of kilohertz to megahertz. In these applications the operating voltage is low and hence MOSFETs are the most suitable [3]. Diodes are required in all power levels and frequency ranges. Schottky diodes and most recently silicon carbide (SiC)-based Schottky diodes are preferred in low voltage and high-switching-frequency circuits [4].

A power semiconductor diode is a two terminal P–N junction device. Power semiconductor diodes are designed to carry higher currents and withstand large voltages. The P–-N junctions are formed by a masked diffusion of impurities [1–3]. The size of diffusion mask, length of diffusion, and magnitude of diffusion temperature has a direct impact on device characteristics. The power diodes have very high breakdown voltages. This high breakdown voltage is achieved by controlling depletion layer boundaries, which is achieved by means of floating field plates or guard rings. Usually a silicon-based P-type semiconductor (anode) is diffused with a silicon-based N-type semiconductor (cathode). Contact terminals are formed using aluminum. The circuit symbol and v–i characteristics of a power diode are as shown in Figure 13.1.

A power transistor is a three terminal device with emitter, collector, and base, and has two types of configuration, namely n–p–n and p–n–p. The circuit symbols of a bipolar junction transistor (BJT) are given in Figure 13.3a.

Power MOSFETs entered power electronic applications in the early 1980s. Power MOSFETs are unipolar devices, in which majority carriers constitute the current. There are two types of MOSFETs, namely enhancement and depletion type. Enhancement MOSFETs are used for power applications. The power MOSFET has a four-layered structure. The n + pn −n + (p + np −p +) structure forms an enhancement type n–channel (p–channel) MOSFET. Gate is insulated from the body using a SiO₂ layer, as shown in Figure 13.6.

The IGBT is based on MOS-bipolar integration. MOSFET structure is used to give the necessary base drive for the bipolar transistor. The basic structure of the n-channel asymmetric blocking (also known as punch through – PT) IGBT is given in Figure 13.8. The N-type drift region is grown over a P⁺ substrate. Initially, high doping is used to create a buffer layer. After this the remaining N-drift region is lightly doped to increase the voltage-blocking capability. In nonpunch through (NPT) or symmetric IGBTs, the P⁺ semiconductor of collector (C) is diffused from the back side of the N-drift region. The N⁺ buffer layer is absent. Then the deep P⁺ body regions are formed to eliminate the effect of parasitic thyristor. Gate oxide is grown after the formation of body regions. The N⁺ dopant is diffused into the body region to form an emitter (E).

Gallium nitride (GaN) and silicon carbide (SiC) are wide band gap materials. GaN and SiC have high critical electric filed strength (e.g., 3.2 eV for SiC and 3.4 eV for GaN) compared to silicon (e.g., 1.12 eV) [4,5,8]. For a given breakdown voltage requirement, width of the layer reduces. This results in smaller sized and reduced on-state resistance, which in turn reduces conduction loss. SiC-based Schottky diodes are now available with a voltage-blocking capacity as high as 3 kV. Other SiC-based power semiconductor devices such as SiC-based thyristors, SiC-based IGBTs, and SiC-based MOSFETs are experimentally proven but have a long way to go before they are made economically viable [4,8,14]. Hence, only a brief discussion is presented in this book. However, SiC has a promising higher-voltage rating for these devices, which was not possible to achieve with silicon. At present SiC-based IGBTs and FETs are available in the voltage range between 600 V/50 A and 1.7 kV/50 A [12,15]. SiC has very high thermal stability compared to silicon. Experimental applications of SiC-based IGBTs in multilevel converters, inverters for drives, and grid-connected inverters are reported in [6,7]. SiC does not melt, instead it gradually sublimes at high temperatures, which makes it impossible to form large monolithic crystals. This requires a modified sublimation process, which is very expensive. However, the performance advantages offset a higher manufacturing cost in certain applications.

Silicon-controlled rectifiers (SCRs), GTOs, and IGCTs are the most popular power semiconductor devices in the thyristor family. SCRs are the oldest semiconductor devices, invented by General Electric in 1957. An SCR is a four-layer device with alternating P- and N-type semiconductors, as shown in Figure 13.12a. The basic structure is constructed by starting with a lightly doped N-type silicon wafer. This forms the drift region and determines the breakdown strength of the thyristors. The anode P⁺ region is formed by diffusing the dopants on one side of the substrate. The p-gate and N⁺ region are diffused on the other side of the substrate.

The gate turn-off thyristor (GTO) is similar to the SCR but has the ability to turn off using gate current. The GTO is also a four-layered device like the SCR. The doping is modified and the width of the cathode region is reduced to achieve turn-off through gate current [1,14]. Since the GTO has gate turn-off capability, the asymmetrical structure with the N-buffer region is preferred. The asymmetric GTOs with a shorted anode structure have a fast turn-off but their reverse voltage-blocking capacity is very low, say from 20 V to 30 V [1,14]. The GTOs without a shorted anode are slow in turn-off but have higher reverse voltage-blocking capacity. The v–i characteristic of a GTO is almost similar to that of an SCR. However, a minor difference is that, for anode currents less than a latching current, the GTO behaves like a BJT. For this operating point, the GTO returns to the turn-off position when gate current is removed. GTOs are not operated in this region.

Integrated gate commutated thyristor (IGCT) is also a four-layer p–n–p–n device. A buffer layer is introduced at the anode to reduce thickness of the anode layer [14,18]. The electron can be extracted efficiently during turn off by the gate current. The above modification results in reduced on-state voltage drop, compared to a GTO. The IGCT has a hard drive feature $d{i}_G/dt>1000\text{A}/\text{μ}\text{s}$, which can be operated without snubbers. The hard drive ensures that the anode current commutates quickly [1,14]. Figure 13.16 shows the circuit symbol and turn-off waveforms of an IGCT. In order to turn off, the gate drive circuit should supply a fast turn-off current pulse. With this current pulse, the cathode side n–p–n transistor structure turns off within 1 μs and leaves the anode side p–n–p transistor structure base open-circuited, forcing the anode current to zero. Due to a very short gate pulse the turn-off energy at the gate is reduced drastically as compared to a GTO. Also the IGCTs have a very short tail current interval, which makes these devices much faster compared to GTOs. Due to the high-pulsed current requirements of the gate circuit, the IGCT’s turn-on/off gate drive unit is supplied as an integral element of the device by the manufacturers [9,10]. IGCTs are meant for high-voltage and high-power applications, where switching frequency is quite low, hence they are optimized for low conduction losses. At present IGCTs with a forward-blocking capacity of 10 kV and a current rating as high as 6 kA are manufactured [10].

Power semiconductor devices span a wide spectrum of power levels from milliwatts to a few hundred megawatts. The switching frequency requirements range from 50 Hz at high power levels to several GHz at low power levels. Efficiency and reliability are the key factors in the power converter circuits, hence proper selection of devices is very important.