6. Power Semiconductors

Learning Objectives

• Explain what distinguishes a power semiconductor device from a small-signal device
• Describe the four-layer structure and latching behavior of a thyristor (SCR)
• Explain why power MOSFETs are preferred for high-frequency switching applications
• Describe how an IGBT combines MOSFET gate control with BJT-like conduction
• Compare thyristors, power MOSFETs, and IGBTs on switching speed, voltage/current handling, and typical use
• Recognize real-world power electronics applications: motor drives, inverters, and power supplies

Quick Answer

Power semiconductors are diodes, transistors, and thyristors specifically engineered to switch or control large currents and voltages efficiently, rather than to amplify small signals. The three workhorse devices are the thyristor (SCR), which latches into full conduction after a brief gate trigger and is used for high-power AC control; the power MOSFET, which uses voltage-controlled switching for fast, efficient operation at moderate voltage/current levels; and the IGBT, which combines a MOSFET's easy gate control with a bipolar transistor's ability to carry very high current at low conduction loss. Choosing among them is a trade-off between switching speed, voltage/current capability, and conduction losses, and this choice shapes the design of motor drives, renewable energy inverters, and switching power supplies.

What Makes a Semiconductor a "Power" Device

Ordinary small-signal diodes and transistors are optimized for speed, low noise, and precision at currents of milliamps and voltages of a few volts to tens of volts. Power semiconductors are built differently: larger die areas to spread out current and heat, vertical current-flow structures (rather than lateral) to handle high current density, and drift regions specifically engineered to withstand hundreds or thousands of volts in the "off" state without breaking down. The central design challenge in every power device is minimizing two competing loss mechanisms simultaneously — conduction loss (voltage drop across the device while current flows) and switching loss (energy wasted during the brief transitions between on and off) — while safely dissipating whatever heat results.

Visual Learning

Thyristors (SCRs)

A Silicon-Controlled Rectifier is a four-layer PNPN device with three terminals: anode, cathode, and gate. Structurally, it behaves like two interconnected BJTs (one NPN, one PNP) arranged so that each one's collector current feeds the other's base current — a positive feedback loop. In the blocking state, the thyristor holds off voltage in both directions until a brief gate current pulse triggers the internal feedback loop, which then regeneratively latches the device into full conduction. Crucially, once latched on, the SCR stays on even if the gate signal is removed — the only way to turn it off is to reduce the anode current below a minimum "holding current," typically by the natural zero-crossing of an AC waveform.

This behavior — needing only a brief pulse to trigger a device that then conducts by itself until current naturally drops — is exactly what makes thyristors so effective for controlling AC power. In an AC phase-control circuit, delaying the gate trigger pulse later into each half-cycle reduces the average power delivered to a load (as used in light dimmers and motor speed controllers), because the thyristor only conducts for the remaining portion of that half-cycle before naturally turning off at the zero crossing.

Power MOSFETs

A power MOSFET is structurally similar to a small-signal enhancement MOSFET but built with a vertical current path (current flows from the top surface down through the die to the bottom, rather than sideways along the surface), which lets a much larger cross-sectional area carry current and dramatically improves current-handling and heat spreading compared to a lateral structure.

Because the gate is voltage-controlled and fully insulated (drawing essentially no steady-state current), power MOSFETs can be switched on and off extremely quickly — often in tens of nanoseconds — making them the device of choice for high-frequency switching applications like DC-DC converters and switch-mode power supplies, where fast switching directly reduces the size of the transformers, inductors, and capacitors needed. Their main limitation is that on-resistance rises sharply with rated voltage, which makes power MOSFETs efficient up to a few hundred volts but increasingly lossy — and therefore less competitive against IGBTs — at very high voltages.

IGBTs (Insulated Gate Bipolar Transistors)

An IGBT is a hybrid device: it uses a MOSFET-like insulated gate for control (easy to drive, negligible gate current, fast turn-on) but its internal current path includes a bipolar PNP transistor structure, giving it much lower conduction voltage drop at high currents than an equivalent power MOSFET would have — behaving more like a BJT once conducting.

This combination is exactly why IGBTs dominate medium-to-high power applications (from a few kW up into the megawatt range) such as motor drives, industrial inverters, electric train propulsion, and renewable energy inverters: they get the simple voltage-controlled gate drive of a MOSFET, plus current-handling and low conduction loss closer to a bipolar device. The trade-off is switching speed — the bipolar conduction mechanism causes a "tail current" during turn-off that power MOSFETs do not have, so IGBTs switch somewhat slower and have higher switching losses at very high frequencies, making them less suitable than MOSFETs for the highest-frequency converters.

Key Terms

Term	Definition	Related Concept
Thyristor (SCR)	Four-layer PNPN device that latches into conduction after a gate trigger pulse	Regenerative feedback, AC phase control
Holding Current	Minimum anode current required to keep a thyristor conducting	Thyristor turn-off
Latching	Self-sustaining "on" state maintained by internal positive feedback, independent of the trigger signal	Thyristor behavior
Power MOSFET	Vertical-structure MOSFET optimized for high current and fast switching	Switch-mode power supplies
On-Resistance (R_DS(on))	Resistance of a MOSFET's channel while fully on, determining conduction loss	Power MOSFET efficiency
IGBT	Hybrid device combining a MOSFET gate with bipolar-transistor-like conduction	Motor drives, industrial inverters
Tail Current	Residual current during IGBT turn-off caused by stored charge in the bipolar region	IGBT switching loss
Conduction Loss	Power dissipated as heat due to voltage drop across a device while it carries current	Device efficiency trade-off
Switching Loss	Energy dissipated during the brief on/off transition of a switching device	Frequency limitations

Common Mistakes

Misconception: A thyristor can be turned off simply by removing the gate signal, just like a transistor. Why it's wrong: Once a thyristor latches on via its internal positive feedback, the gate loses control entirely — removing the gate current does nothing to stop conduction. The only way to turn it off is to reduce the anode current below the holding current (commonly relying on an AC waveform's natural zero crossing) or use special forced-commutation circuitry. Correct understanding: A thyristor's gate only initiates conduction; it does not maintain or terminate it. This one-shot triggering behavior is fundamentally different from a MOSFET or BJT, whose conduction directly tracks a continuous control signal.

---

Misconception: IGBTs are simply "better" power MOSFETs and should always be preferred for power switching. Why it's wrong: IGBTs have a tail current during turn-off caused by stored minority-carrier charge in their bipolar section, which power MOSFETs do not have. This makes IGBTs switch slower and less efficiently at very high frequencies compared to power MOSFETs. Correct understanding: The choice depends on the application: power MOSFETs are usually better for high-frequency, lower-voltage switching (like DC-DC converters), while IGBTs are usually better for high-voltage, high-current, lower-frequency switching (like industrial motor drives).

---

Misconception: A device's voltage and current ratings alone determine which power semiconductor to use. Why it's wrong: Two devices with identical voltage/current ratings can have very different switching speeds, conduction losses, gate drive requirements, and thermal behavior, all of which strongly affect suitability for a given application. Correct understanding: Selecting a power device requires balancing switching frequency, conduction loss, gate drive complexity, and cost together — not just checking that the voltage/current numbers are "big enough."

Comparison and Connections

Device	Control Mechanism	Typical Switching Speed	Typical Power Range	Best Suited For
Thyristor (SCR)	Brief gate pulse, then self-latching	Slow (line-frequency, tens of Hz to a few kHz)	Very high (up to MW)	AC phase control, high-power rectification
Power MOSFET	Continuous gate voltage	Very fast (tens of kHz to MHz)	Low to medium (up to a few hundred V)	High-frequency switch-mode power supplies
IGBT	Continuous gate voltage	Moderate (a few kHz to tens of kHz)	Medium to very high	Motor drives, industrial inverters, traction

Practice Questions

Recall

1. What is the layer structure of a thyristor, and what are its three terminals? Guidance: Four-layer PNPN structure; terminals are anode, cathode, and gate.

2. Why does a power MOSFET use a vertical current-flow structure instead of a lateral one? Guidance: A vertical structure allows a much larger effective cross-sectional area to carry current, improving current-handling capability and heat spreading compared to a lateral (surface) current path.

Understanding

3. Explain why a thyristor, once triggered, continues conducting even after the gate signal is removed. Guidance: The thyristor's internal two-transistor structure forms a positive feedback loop once triggered — each transistor's collector current supplies the other's base current — so conduction becomes self-sustaining and independent of the gate, until anode current drops below the holding current.

4. Explain why IGBTs have higher switching losses than power MOSFETs at high frequencies. Guidance: IGBTs have a bipolar conduction path with stored minority-carrier charge that must recombine or be swept out during turn-off, producing a "tail current" that power MOSFETs (a purely majority-carrier device) do not exhibit, increasing turn-off energy loss especially as switching frequency rises.

Application

5. A designer needs to control the brightness of an incandescent lamp connected to AC mains using phase-angle control. Which power device is most appropriate, and how does delaying the trigger point in each half-cycle achieve dimming? Guidance: A thyristor (SCR, or a TRIAC for AC). Delaying the gate trigger later into each half-cycle means the device conducts for a shorter portion of that half-cycle before the AC current naturally reaches zero, reducing the average power delivered to the lamp.

6. A power supply needs to switch at 500 kHz for a compact DC-DC converter. Would a thyristor, power MOSFET, or IGBT be most appropriate, and why? Guidance: A power MOSFET, because its voltage-controlled, majority-carrier operation allows very fast switching (hundreds of kHz to MHz) with low switching loss, which thyristors (far too slow) and IGBTs (tail current limits high-frequency efficiency) cannot match at this frequency.

Analysis

7. An electric train uses IGBTs rather than power MOSFETs for its traction motor drive, even though the drive operates at only a few kHz switching frequency. Explain why this choice makes sense given the power levels involved. Guidance: Traction drives require very high current and voltage handling (hundreds of kW to MW range) with moderate switching frequency. IGBTs provide much lower conduction loss at these high currents than an equivalent MOSFET would (which would need many parallel devices and would have high on-resistance losses), and since the switching frequency is modest, the IGBT's tail-current disadvantage matters less.

8. Compare the failure consequences of removing gate drive unexpectedly from a conducting thyristor versus from a conducting power MOSFET. Guidance: Removing gate drive from a conducting thyristor has no effect — it keeps conducting because it has already latched via internal feedback; only reducing current below the holding current turns it off. Removing gate drive from a conducting power MOSFET immediately turns it off (assuming no continued forward voltage sufficient for body-diode conduction), since MOSFET conduction directly tracks a continuous gate voltage rather than being self-sustaining.

FAQ

Why can't ordinary small-signal transistors just be scaled up to handle high power? Scaling up die size alone does not solve everything — power devices need vertical current structures and drift regions specifically engineered to block high voltages, wide-enough current paths to avoid localized hot spots (which can create a destructive positive-feedback thermal effect called "thermal runaway" or, in bipolar-type devices, "second breakdown"), and packaging designed to conduct heat away efficiently. Small-signal device geometries are optimized for speed and precision at low power, not for these thermal and voltage-blocking requirements.

What is a TRIAC, and how does it relate to a thyristor? A TRIAC is essentially two thyristors connected back-to-back in a single package, allowing it to conduct and be triggered in both directions of an AC cycle, unlike a standard SCR which only conducts in one direction. This makes TRIACs especially convenient for simple AC phase-control applications like light dimmers, where control over both positive and negative half-cycles is needed from a single component.

Why do power MOSFETs have a resistance rating (R_DS(on)) instead of just a voltage drop like a diode? Unlike a diode's roughly constant forward voltage drop, a MOSFET's fully-on channel behaves almost like a simple resistor, so the voltage drop across it scales directly with current (V = I × R_DS(on)). This resistance rises with the device's voltage rating because higher-voltage MOSFETs need a longer, lower-doped drift region to block that voltage, and that same region adds resistance during conduction — a fundamental trade-off in power MOSFET design.

How do engineers decide between air cooling, heat sinks, and liquid cooling for power semiconductors? The decision depends on how much power the device dissipates (conduction loss plus switching loss) and how quickly that heat needs to be removed to keep the junction temperature within its safe rated limit. Low-power circuits may rely on natural convection alone; moderate power needs a heat sink (often with a fan for forced air cooling); very high-power industrial or traction systems frequently use liquid cooling to manage the much larger heat loads that MOSFETs and IGBTs handling hundreds of amps can generate.

Are power semiconductors and power ICs the same thing? Not quite. Power semiconductors (thyristors, power MOSFETs, IGBTs) are typically discrete high-power switching devices. Power ICs integrate multiple functions — such as a gate driver, protection circuitry, and sometimes a power MOSFET itself — onto a single chip, simplifying circuit design. Many practical power electronics systems use a mix of standalone power semiconductors for the main high-power switching path and power ICs for control, sensing, and protection.

Quick Revision

• Power semiconductors are engineered for high current/voltage switching, balancing conduction loss against switching loss
• A thyristor (SCR) is a four-layer PNPN device that latches on via internal positive feedback after a brief gate trigger
• A thyristor turns off only when anode current drops below the holding current, not by removing the gate signal
• Thyristors suit high-power, lower-frequency AC control (dimmers, motor soft-starters, high-power rectification)
• Power MOSFETs use a vertical structure and voltage-controlled gate for very fast, efficient switching
• Power MOSFET on-resistance (R_DS(on)) rises sharply with voltage rating, limiting efficiency at very high voltages
• IGBTs combine a MOSFET's easy gate control with a bipolar transistor's low conduction loss at high current
• IGBTs have a turn-off "tail current" from stored bipolar charge, limiting their high-frequency efficiency compared to MOSFETs
• IGBTs dominate medium-to-high power applications: motor drives, industrial inverters, traction systems
• TRIACs are essentially two back-to-back thyristors for bidirectional AC control

Related Topics

Prerequisites: Bipolar Junction Transistors, Field Effect Transistors, PN Junction Diodes

Related Topics: Advanced Semiconductor Devices, Semiconductor Device Applications, Semiconductor Materials

Next Topics: Semiconductor Manufacturing, Advanced Semiconductor Devices, Semiconductor Device Applications