IGBTs
Combining BJT and MOSFET advantages, latch-up.
The Insulated Gate Bipolar Transistor (IGBT) is a power semiconductor device that combines the high-input-impedance voltage-controlled gate of a MOSFET with the low on-state voltage drop characteristic of a bipolar junction transistor. This combination makes it the preferred switching device for medium to high power applications such as inverters for industrial motor drives, traction systems, and renewable energy converters where both switching speed and conduction efficiency are critical.
Core Concept Explanation
The IGBT structure looks like a power MOSFET with one crucial difference: the N+ drain substrate of the MOSFET is replaced by a P+ substrate. This P+ collector layer, when forward biased, injects holes into the N-drift region. These holes attract electrons from the emitter side, flooding the N-drift region with both carriers. This minority carrier injection, exactly like in a BJT, dramatically reduces the resistivity of the drift region. The result is a very low on-state collector-emitter voltage VCE(sat), typically 1.5 V to 3 V even at high currents, far lower than a comparable power MOSFET at high voltage ratings.
The equivalent circuit of an IGBT is a MOSFET whose output drives the base of a PNP BJT. The gate voltage controls the MOSFET channel, which provides base current to the BJT. The BJT then carries the bulk of the collector current. This is why the IGBT has MOSFET-like gate control (voltage driven, no steady-state gate current) but BJT-like output characteristics with a VCE(sat) rather than a simple resistive drop.
The critical weakness of the IGBT compared to the MOSFET is its current tail during turn-off. When the gate is pulled low, the MOSFET channel turns off quickly. However, the minority carriers stored in the N-drift region by the PNP BJT cannot be instantly removed. They recombine slowly, causing the collector current to decay with a tail waveform. This tail current flows while the collector-emitter voltage has already risen to VDC, causing a significant turn-off switching loss. The duration of this tail is the fundamental limitation on IGBT switching frequency, typically restricting them to below 50 kHz for high-voltage devices.
Mathematical Expression
The on-state conduction loss per device in a half-bridge leg is calculated using VCE(sat). For an IGBT carrying average current Iavg:
P_cond = VCE(sat) x Iavg. This is fundamentally different from MOSFET conduction loss which is I2 x RDS(on). For high current devices, VCE(sat) remains roughly constant with current (like a diode), so IGBT conduction loss scales linearly with current whereas MOSFET loss scales with the square. This means at high currents, IGBTs are more efficient for conduction. Turn-off loss: E_off = (1/2) x VDC x IC x t_tail, where t_tail is the current tail time constant, typically hundreds of nanoseconds to microseconds.
Practical Understanding
IGBTs dominate in applications above approximately 600 V and moderate switching frequencies. In a 3-phase inverter for a 415 V motor drive, six IGBTs operate as switches. Their VCE(sat) of about 2 V at rated current gives acceptable conduction loss, and switching at 10 kHz to 20 kHz is well within their capability. At the same voltage rating, a power MOSFET would have very high RDS(on) due to the silicon limit, making IGBT the better choice.
A dangerous failure mode specific to IGBTs is latch-up. In the IGBT structure, there is a parasitic NPN transistor formed by the N+ emitter, P-body, and N-drift regions. Combined with the PNP BJT, these form a parasitic PNPN thyristor (SCR). If the collector current exceeds a critical value, or if dv/dt is too high, the parasitic SCR can latch on. Once latched, gate control is completely lost and the device conducts full current regardless of gate signal, typically destroying it. Modern IGBTs use design techniques like a short-circuit P+ region under the N+ emitter to suppress the parasitic NPN and prevent latch-up under normal and fault conditions.
Given:
Collector current IC = 50 A, VCE(sat) = 2 V
VDC = 600 V, current tail time t_tail = 600 ns, switching frequency f = 10 kHz
Why this formula applies:
For IGBT, conduction loss is linear in current (not quadratic like MOSFET).
Turn-off loss dominates switching losses due to current tail.
Formula:
P_cond = VCE(sat) x IC x duty_cycle (assume 0.5)
E_off = 0.5 x VDC x IC x t_tail
P_sw = E_off x f
Substitution:
P_cond = 2 x 50 x 0.5
E_off = 0.5 x 600 x 50 x 600e-9
P_sw = E_off x 10000
Calculation:
P_cond = 50 W
E_off = 0.5 x 600 x 50 x 6e-7 = 9 x 10^-3 J = 9 mJ
P_sw = 9e-3 x 10000 = 90 W
Final Answer:
Conduction loss = 50 W, Switching loss = 90 W, Total = 140 W per IGBT.Exam Tip: IGBT conduction loss = VCE(sat) x IC (linear), whereas MOSFET conduction loss = I2 x RDS(on) (quadratic). GATE questions often ask to compare these at a given current. Latch-up occurs when the parasitic SCR inside the IGBT turns on, causing loss of gate control.
- MOSFET channel turns off quickly when gate is pulled low, cutting off electron injection into drift region.
- Holes stored in N-drift by PNP BJT continue recombining, creating the current tail that causes switching loss.
- Latch-up occurs when NPN (N+ emitter, P-body, N-drift) and PNP (P+ collector, N-drift, P-body) form a triggered SCR.
- Once the parasitic SCR latches, collector current cannot be turned off by gate, resulting in device destruction.
- P+ shorting region under N+ emitter bypasses the NPN base-emitter junction, preventing latch-up in modern designs.
Quick Revision
- IGBT = MOSFET gate control + BJT current carrying. P+ collector injects holes into N-drift for conductivity modulation.
- VCE(sat) is typically 1.5 V to 3 V, constant with current. Conduction loss = VCE(sat) x IC.
- Current tail at turn-off: minority carriers in N-drift recombine slowly, causing energy loss E_off = 0.5 x VDC x IC x t_tail.
- Latch-up: parasitic SCR (NPN + PNP) turns on at high current, causing loss of gate control.
- IGBTs preferred above 600 V and below 50 kHz. MOSFETs preferred below 600 V and at high frequency.
- GATE trap: IGBT does NOT have RDS(on). It has VCE(sat). Conduction loss formula is different from MOSFET.
IGBT Fundamentals
Examine BJT-MOSFET integration and latch-up.
Q1.How does the structural design of an IGBT combine the advantages of MOSFETs and BJTs?
Related Articles
IGBT
Insulated Gate Bipolar Transistor structure, tail current.
8 min read
Power Diodes
P-I-N structure, switching characteristics.
9 min read
Power MOSFET
VDMOS structure, safe operating area.
12 min read
Power Diode
Reverse recovery characteristics, Schottky diodes.
12 min read
GTO
Gate Turn-Off thyristor.
12 min read