Update : November 22, 2023
An Insulated-Gate Bipolar Transistor (IGBT) combines the advantages of a power transistor (Giant Transistor—GTR) and a power MOSFET (Metal-Oxide-Semiconductor Field-Effect Transistor). It exhibits excellent characteristics and has a wide range of applications. The IGBT is a three-terminal device consisting of a gate, collector, and emitter.
The IGBT integrates the benefits of both the power transistor (GTR) and the power MOSFET, offering superior characteristics and finding extensive application. An IGBT (Insulated Gate Bipolar Transistor) is a MOS-structured bipolar device and a power device that combines the high-speed performance of a power MOSFET with the low on-resistance of bipolar technology.
Typically, IGBTs are used in applications with voltages above 600V, currents above 10A, and frequencies over 1kHz. They are commonly found in industrial motors, small-capacity household motors, converters (inverters), camera flash observers, induction heating rice cookers, and more.
IGBTs are generally categorized into two types based on packaging: the molded resin-sealed three-terminal single package, ranging from TO-3P to small surface mount devices, and the module type, where IGBTs are paired with FWD (Free Wheel Diode) in sets (2 or 6 sets) for industrial use. Module types vary in shape and packaging method based on the application and are available in various series.
IGBT is a natural evolution of vertical power MOSFETs for high-current, high-voltage applications, and fast-switching devices. Power MOSFETs, due to their need for a source-drain channel to achieve a high breakdown voltage (BVDSS), exhibit high on-resistance (RDS(on)).
IGBTs overcome these major drawbacks of existing power MOSFETs. Although the latest generation of power MOSFETs has significantly improved RDS(on) characteristics, their power conduction losses at high voltages are still much higher than those of IGBTs. IGBTs have a lower voltage drop, converting into a lower VCE(sat) capability, and their structure allows for higher current density compared to standard bipolar devices, simplifying the IGBT driver schematic.
An N-channel enhancement-type IGBT structure includes an N+ region called the source, with an electrode known as the source electrode. The control region of the device is the gate area, with the gate electrode attached. The channel forms adjacent to the gate region boundary. The P-type region (including P+ and P- regions) between the drain and source, known as the subchannel region, and the P+ region on the other side of the drain area, called the drain injection region, is unique to IGBTs.
This region, together with the drain and subchannel areas, forms a PNP bipolar transistor, acting as an emitter, injecting holes into the drain region for conductive modulation, reducing the device's on-state voltage. The electrode attached to the drain injection region is the drain electrode.
The switching action of an IGBT is achieved by applying a forward gate voltage to form a channel, providing base current to the PNP transistor, thereby turning on the IGBT. Conversely, applying a reverse gate voltage eliminates the channel, stopping the reverse base current, and turning off the IGBT. The driving method of IGBTs is similar to that of MOSFETs, requiring only control of the input pole N-channel MOSFET, hence featuring high input impedance.
When the channel in the MOSFET forms, holes injected into the N- layer from the P+ base modulate the electrical conductivity of the N- layer, reducing its resistance and allowing the IGBT to maintain a low on-state voltage even at high voltages.
The static characteristics of IGBTs primarily include their voltage-current characteristics, transfer characteristics, and switching characteristics.
The voltage-current characteristic of an IGBT is the relationship curve between the drain current and gate voltage, with the gate-source voltage Ugs as the parameter variable. The output drain current is controlled by the gate-source voltage Ugs; the higher the Ugs, the greater the Id.
This characteristic is similar to that of GTR and can be divided into three parts: the saturation region, amplification region, and breakdown characteristics. In the off-state, the forward voltage is borne by the J2 junction, and the reverse voltage by the J1 junction. If there is no N+ buffer region, the forward and reverse blocking voltages can be the same. However, with an N+ buffer, the reverse blocking voltage can only reach tens of volts, thus limiting some IGBT applications.
The transfer characteristic of an IGBT is the relationship curve between the output drain current Id and the gate-source voltage Ugs. It is similar to that of a MOSFET. When the gate-source voltage is below the threshold voltage Ugs(th), the IGBT is off. Once the IGBT is on, Id shows a linear relationship with Ugs over most of the drain current range. The maximum gate-source voltage is limited by the maximum drain current, typically set around 15V.
The switching characteristic of an IGBT is the relationship between the drain current and the drain-source voltage. When the IGBT is on, due to its PNP transistor being a wide-base transistor, its B value is extremely low.
Although the equivalent circuit is a Darlington structure, the current through the MOSFET becomes the main part of the total current of the IGBT. Due to the presence of the N+ region and the conductive modulation effect, the IGBT has a low on-state voltage drop, typically 2-3V for a 1000V rated IGBT. In the off-state, only a small leakage current exists.
During the turn-on process, an IGBT mainly functions as a MOSFET. However, in the latter stages of the drain-source voltage (Uds) reduction, the transition of the PNP transistor from the amplification to saturation region adds a delay. The turn-on delay time is denoted as td(on), and the current rise time is tri. The total turn-on time for the drain current, commonly provided in practical applications, is the sum of td(on) and tri. The fall time of the drain-source voltage comprises tfe1 and tfe2.
Triggering and turning off the IGBT require applying both positive and negative voltages between its gate and base, which can be produced by various driving circuits. When selecting these circuits, factors such as the device's turn-off bias requirements, gate charge requirements, robustness, and power supply considerations must be taken into account. Due to the large gate-emitter impedance of the IGBT, MOSFET driving techniques are applicable for its triggering. However, since the IGBT's input capacitance is larger than that of most MOSFETs, its turn-off bias should be higher than the bias provided by many MOSFET driving circuits.
The switching speed of the IGBT is lower than that of MOSFETs but significantly higher than that of GTRs. The IGBT does not require a negative gate voltage to reduce turn-off time, but the turn-off time increases with the parallel resistance between the gate and emitter. The turn-on voltage of the IGBT is approximately 3-4V, similar to MOSFETs. The saturation voltage drop of the IGBT during conduction is lower than that of MOSFETs but close to that of GTRs, decreasing with the increase in gate voltage.
To date, commercially used high-voltage, high-current IGBT devices have not been widely available, and their voltage and current capacities are still limited, not fully meeting the needs of the development of power electronics technology, especially in high-voltage applications requiring device voltage ratings of over 10KV. Currently, high-voltage applications can only be achieved through techniques such as high-voltage series connection of IGBTs.
International manufacturers like Switzerland's ABB have developed 8KV IGBT devices using soft punch-through principles, and Germany's EUPEC has produced 6500V/600A high-voltage, high-power IGBT devices already in practical use. Toshiba in Japan has also entered this field. Meanwhile, semiconductor manufacturers continue to develop IGBTs with high voltage endurance, large current, high speed, low saturation voltage drop, high reliability, and low cost, mainly using sub-1um fabrication processes, achieving some new progress.
The operation of an N-channel IGBT involves applying a voltage greater than the threshold voltage (V_TH) between the gate and emitter. This action forms an inversion layer (channel) in the P-layer just below the gate electrode, initiating the injection of electrons from the n-layer beneath the emitter electrode. These electrons act as minority carriers for the p+n-p transistor, flowing into the holes from the collector substrate p+ layer, thereby modulating the conductance (bipolar operation) and reducing the saturation voltage between the collector and emitter. On the emitter electrode side, an n+pn- parasitic transistor is formed. If this parasitic transistor operates, it turns into a p+n- pn+ thyristor, and the current continues until the output side stops supplying current. This uncontrollable state by the output signal is typically referred to as latch-up.
To prevent the operation of the n+pn- parasitic transistor, the IGBT is designed to minimize the current gain factor (α) of the p+n-p transistor, typically below 0.5. The latch-up current (I_L) of the IGBT is more than three times the rated (DC) current. The IGBT's driving principle is similar to that of power MOSFETs, with on-off controlled by the gate-emitter voltage (u_GE).
The structure of the IGBT silicon chip closely resembles that of the power MOSFET. The primary difference lies in the addition of a P+ substrate and an N+ buffer layer (not added in the non-punch-through IGBT technology), where one MOSFET drives two bipolar devices. The substrate creates a J1 junction between the P+ and N+ regions of the tube body.
When positive gate bias inverts the P base region under the gate, an N-channel forms, producing an electron flow and generating a current in the manner of a power MOSFET. If the voltage generated by this electron flow is within 0.7V, J1 becomes forward-biased, injecting some holes into the N- region and modulating the resistance between the anode and cathode. This process reduces the total power conduction loss and initiates a second charge flow. The final result is two different current topologies within the semiconductor layer: an electron flow (MOSFET current) and a hole flow (bipolar). When u_GE exceeds the turn-on voltage UGE(th), a channel forms in the MOSFET, providing base current for the transistor, and the IGBT conducts.
The conductance modulation effect reduces the resistance RN, resulting in a low on-state voltage drop.
When a negative bias is applied to the gate or when the gate voltage falls below the threshold value, the channel is inhibited, preventing the injection of holes into the N- region. In any case, if the MOSFET current drops rapidly during the switching phase, the collector current gradually decreases. This reduction is due to the presence of a few minority carriers (tail current) in the N layer after commutation begins. The decrease in this residual current (tail current) depends entirely on the charge density at turn-off, related to several factors such as the quantity and topology of impurities, layer thickness, and temperature. The decay of these minority carriers gives the collector current a characteristic tail current waveform, causing increased power consumption and cross-conduction issues, particularly in devices using freewheeling diodes.
Given that the tail current is related to the recombination of minority carriers, its value should be closely related to the chip's temperature, IC, and VCE, as well as the mobility of the holes. Therefore, based on the reached temperature, it is feasible to reduce the undesirable effects of this current in terminal device design. The tail current characteristics are related to VCE, IC, and TC.
When a reverse bias is applied between the gate and emitter or when no signal is applied, the channel in the MOSFET disappears, and the base current of the transistor is cut off, turning off the IGBT.
When a reverse voltage is applied to the collector, J1 is controlled by reverse bias, and the depletion layer extends into the N- region. If the thickness of this layer is excessively reduced, an effective blocking capability cannot be achieved, making this mechanism crucial. On the other hand, if the size of this region is increased too much, the voltage drop will continuously increase.
When the gate and emitter are short-circuited and a positive voltage is applied to the collector terminal, the P/NJ3 junction is controlled by reverse voltage. At this time, the depletion layer in the N drift region bears the externally applied voltage.
IGBT has a parasitic PNPN thyristor between the collector and emitter. Under certain conditions, this parasitic device may conduct, increasing the current between the collector and emitter, reducing the control capability of the equivalent MOSFET, and often causing device breakdown. This thyristor conduction phenomenon is known as IGBT latch-up. Specifically, the causes of this defect vary and are closely related to the device's state. Generally, static and dynamic latch-up differ as follows: static latch-up occurs when the thyristor is fully conductive, while dynamic latch-up only occurs during turn-off. This phenomenon severely limits the safe operating area.
To prevent the harmful effects of parasitic NPN and PNP transistors, it is necessary to prevent the conduction of the NPN part by changing the layout and doping levels, and to reduce the total current gain of the NPN and PNP transistors.
Furthermore, the latch-up current affects the current gain of the PNP and NPN devices, hence closely related to the junction temperature. As junction temperature and gain increase, the resistivity of the P-base region increases, compromising overall characteristics. Therefore, device manufacturers must maintain a certain ratio between the maximum collector current and the latch-up current, typically around 1:5.
IGBT gate driver design, how to choose the key components
What is a MOSFET (Metal-Oxide-Semiconductor Field-Effect Transistor)?
