1. Structure of IGBT
An IGBT is a three-terminal device with a gate (G), a collector (c), and an emitter (E). The structure, simplified equivalent circuit, and electrical symbol of an IGBT are shown in the figure.
The figure shows a cross-sectional view of the internal structure of an N-channel IGBT (N-IGBT) combining an N-channel VDMOS FFT and a GTR. The IGBT has an additional P+ injection region compared to a VDMOSFET, forming a large-area PN junction J1. Because minority carriers are emitted from the P+ injection region to the N-base region when the IGBT is turned on, the conductivity of the drift region is modulated, giving the IGBT a strong current-carrying capacity. The N+ layer between the P+ injection region and the N-drift region is called a buffer. The presence or absence of a buffer determines the different characteristics of the IGBT. An IGBT with an N* buffer is called an asymmetric IGBT, also known as a punch-through IGBT. It has advantages such as low forward voltage drop, short turn-off time, and low tail current during turn-off, but its reverse blocking capability is relatively weak. An IGBT without an N-buffer is called a symmetric IGBT, also known as a non-punch-through IGBT. It has strong forward and reverse blocking capabilities, but its other characteristics are inferior to those of an asymmetric IGBT.
The simplified equivalent circuit shown in Figure 2-42(b) indicates that the IGBT is a Darlington structure composed of a GTR and a MOSFET. Part of this structure is driven by the MOSFET, and the other part is a thick-base PNP transistor.
2. Working principle of IBGT
In simple terms, an IGBT is equivalent to a thick-base PNP transistor driven by a MOSFET. Its simplified equivalent circuit is shown in Figure 2-42(b), where RN is the modulation resistor in the base region of the PNP transistor. From this equivalent circuit, it is clear that an IGBT is a composite device with a Darlington structure, composed of a transistor and a MOSFET. Because the transistor in the figure is a PNP transistor and the MOSFET is an N-channel field-effect transistor, this type of IGBT is called an N-channel IGBT, with the symbol N-IGBT. Similarly, there is a P-channel IGBT, or P-IGBT.
The electrical symbol for an IGBT is shown in Figure 2-42(c). An IGBT is a field-controlled device whose turn-on and turn-off are determined by the gate-emitter voltage UGE. When the gate-emitter voltage UGE is positive and greater than the turn-on voltage UCE(th), a channel is formed within the MOSFET, providing base current to the PNP transistor and thus turning the IGBT on. At this time, holes (minority carriers) injected from the P+ region into the N- region modulate the conductivity of the N- region, reducing the resistance RN of the N- region, allowing even a high-voltage IGBT to have a very small on-state voltage drop. When no signal is applied between the gate and emitter or a reverse voltage is applied, the channel within the MOSFET disappears, the base current of the PNP transistor is cut off, and the IGBT turns off. Therefore, the driving principle of an IGBT is basically the same as that of a MOSFET.
① When UCE is negative: J3 junction is in reverse bias and the device is in reverse blocking state.
② When uCE is positive: UC
1) Conductivity
The structure of an IGBT silicon wafer is very similar to that of a power MOSFET. The main difference is that the JGBT adds a P+ substrate and an N+ buffer layer (NPT - non-punch-through IGBT technology does not add this part). One MOSFET drives two bipolar devices (devices with two polarities). The substrate application creates a J1 junction between the P and N+ regions of the transistor body. When a positive gate bias causes the P-base region under the gate to invert, an N-channel is formed, and an electron flow occurs, generating a current exactly like a power MOSFET. If the voltage generated by this electron flow is in the range of 0.7V, J1 will be forward biased, and some holes will be injected into the N- region, adjusting the resistivity between N- and N+. This reduces the total power conduction loss and initiates a second charge flow. The final result is that two different current topologies temporarily exist within the semiconductor layer: an electron flow (MOSFET current); and a hole current (bipolar). When UCE is greater than the turn-on voltage UCE(th), a channel is formed in the MOSFET, providing base current to the transistor, and the IGBT turns on.
2) On-state voltage drop
The conductivity modulation effect reduces the resistance RN, resulting in a smaller on-state voltage drop. The on-state voltage drop refers to the voltage drop UDS of the IGBT when it enters the conducting state, and this voltage decreases as UCS increases.
3) Turn off
When a negative bias is applied to the gate or the gate voltage is below the threshold, the channel is disabled, and no holes are injected into the N-region. In any case, if the MOSFET current drops rapidly during the switching phase, the collector current gradually decreases because a minority of carriers (less than 10) remain in the N-layer after commutation begins. This reduction in residual current (wake) depends entirely on the charge density at turn-off, which is related to several factors such as the amount and topology of dopants, layer thickness, and temperature. The decay of minority carriers gives the collector current a characteristic wake waveform. The collector current will cause increased power dissipation and cross-conduction problems, especially in devices using freewheeling diodes.
Given that the wake current is related to minority carrier recombination, its value is closely related to the chip's Tc, IC2, and uCE, and also closely related to hole mobility. Therefore, depending on the temperature reached, it is feasible to reduce this non-ideal current effect on the end-device design. When a reverse voltage is applied between the gate and emitter or no signal is applied, the channel within the MOSFET disappears, the transistor's base current is cut off, and the IGBT is turned off.
4) Reverse blocking
When a reverse voltage is applied to the collector, J is reverse biased, and the depletion layer extends into the N-region. This mechanism is crucial because excessively reducing the thickness of this layer will prevent effective blocking. Furthermore, excessively increasing the size of this region will continuously increase the voltage drop.
5) Positive blocking
When the gate and emitter are shorted and a positive voltage is applied to the collector terminal, the J junction is controlled by a reverse voltage. At this time, the externally applied voltage is still absorbed by the depletion layer within the N-drift region.
6) Latch
An IGBT has a parasitic PNPN thyristor between its collector and emitter. Under certain conditions, this parasitic device can turn on. This phenomenon increases the current between the collector and emitter, reduces the control capability of the equivalent MOSFET, and often causes device breakdown. This thyristor turn-on phenomenon is called IGBT latch-up. Specifically, the causes of this defect vary, but are closely related to the state of the device.
