Electronics Guide

Operating Principles

Power diodes conduct current in one direction while blocking voltage in the reverse direction, providing the essential rectification function in power electronics. Unlike small-signal diodes, power diodes must handle currents from amperes to thousands of amperes and block voltages from tens to thousands of volts. The semiconductor junction characteristics that enable this behavior also determine forward voltage drop, reverse-recovery time, and safe operating limits.

Standard Recovery Diodes

Standard recovery diodes, also called line-frequency diodes, are optimized for 50/60 Hz rectification where switching losses are not a primary concern. These devices achieve low forward voltage drop through heavily doped structures that store significant charge during conduction. The stored charge must be removed during turn-off, resulting in reverse-recovery times of several microseconds that limit high-frequency performance. Forward drop is typically near 1 V at rated current, and the corresponding conduction loss dominates the loss budget in a mains rectifier. Large stud-mount and press-pack rectifier diodes reach thousands of amperes and blocking voltages of several kilovolts for electrochemical, traction, and HVDC-adjacent service.

Fast Recovery Diodes

Fast recovery diodes reduce reverse-recovery time to tens or hundreds of nanoseconds through optimized doping profiles and minority-carrier lifetime control, often using techniques such as platinum or gold doping or electron irradiation. These devices suit switching frequencies up to several hundred kilohertz and are commonly used as freewheeling diodes in conjunction with transistor switches. The faster recovery comes at the cost of somewhat higher forward voltage drop, because lifetime killing also raises the on-state voltage, and, in soft-recovery variants, careful tailoring of the recovery waveform to limit voltage overshoot. The figures of merit are the reverse-recovery charge, which sets the turn-on loss in the companion transistor, and the softness factor, the ratio of the falling to the rising portion of the recovery current, which governs the ringing and radiated noise the recovery produces.

Schottky Diodes

Schottky diodes use a metal-semiconductor junction rather than a p-n junction, eliminating minority-carrier storage and providing essentially zero reverse-recovery charge. The forward voltage drop is lower than in junction diodes, typically 0.3 to 0.5 V for silicon Schottky devices against roughly 0.7 to 1 V for a p-n rectifier, which reduces conduction losses in low-voltage outputs. However, silicon Schottky diodes are limited in reverse-voltage capability by rapidly rising leakage current, which is practical to roughly 100 to 200 V; they also exhibit a strong increase in leakage with temperature. Silicon carbide Schottky diodes overcome this limit, extending zero-recovery rectification to 650 V, 1200 V, and 1700 V ratings, with 3300 V parts now commercially available.

Diode Selection Criteria

Diode selection balances forward voltage drop affecting conduction losses, reverse-recovery characteristics determining switching losses, voltage and current ratings matching application requirements, and thermal performance enabling heat dissipation. High-frequency applications demand fast, ultrafast, or silicon carbide diodes, while line-frequency rectifiers can use slower devices with lower forward drop. Surge-current rating and the reverse-recovery softness factor also matter where transient stress or electromagnetic interference is a concern.

Operating Principles

Power MOSFETs are voltage-controlled switches that conduct current between drain and source when a voltage applied to the gate exceeds the threshold voltage. The insulated gate requires negligible steady-state current, simplifying drive circuitry compared with bipolar devices. Majority-carrier conduction eliminates minority-carrier storage, enabling fast switching with minimal charge-related losses.

Device Structure

Modern power MOSFETs use vertical structures that conduct current through the die thickness rather than along the surface. The trench MOSFET structure places the gate in etched trenches, reducing cell pitch and on-resistance. Superjunction MOSFETs add columns of alternating doping to break the conventional relationship between breakdown voltage and on-resistance, which for a conventional structure rises with roughly the 2.4 to 2.5 power of the voltage rating. The superjunction charge-balance principle changes that scaling to something closer to linear, giving on-resistance several times lower than a planar device of the same rating and die area at voltages above roughly 500 V. Commercial superjunction parts cluster at 600 to 950 V, which places them squarely in offline power-supply and photovoltaic-inverter service; below about 200 V the trench structure remains the better choice.

Key Parameters

On-resistance (R_DS(on)) determines conduction losses and is specified at a particular gate voltage and temperature. On-resistance increases markedly with temperature in silicon devices, roughly doubling between 25 °C and 125 °C in a typical high-voltage part; the positive temperature coefficient aids current sharing among paralleled parts but raises losses when hot, so thermal design must use the hot-state value rather than the headline datasheet number. Gate charge characterizes the charge required to switch the device and influences switching losses and driver requirements. Its components matter individually: the gate-drain or Miller charge determines the duration of the plateau during which drain voltage transitions, and the ratio of gate-drain to gate-source charge predicts susceptibility to induced turn-on. Breakdown voltage defines the maximum drain-source voltage the device can block, while the safe operating area describes the voltage-current combinations the device can withstand without damage.

Gate Drive Requirements

Silicon power MOSFETs require gate voltages of typically 10 to 15 V for full enhancement, with current pulses during switching to charge and discharge the gate capacitance. Logic-level devices reach full enhancement near 4.5 V. Gate driver circuits must provide low-impedance paths for rapid gate-voltage transitions; a device with 100 nC of gate charge switched in 50 ns demands a peak drive current of about 2 A. Proper gate-drive design prevents spurious turn-on from dV/dt coupling through the gate-drain capacitance and keeps the gate within its absolute maximum voltage rating, commonly ±20 V for standard silicon parts and considerably narrower for wide-bandgap devices.

Body Diode

The MOSFET structure inherently includes a body diode that conducts when the drain-source voltage reverses. This intrinsic diode has relatively slow, charge-heavy recovery in silicon MOSFETs, which can cause additional losses and reverse-recovery stress in hard-switched bridge configurations. Synchronous rectification, in which the MOSFET channel rather than the body diode carries the reverse current, improves efficiency when timing is properly controlled, with dead time kept short enough to limit body-diode conduction but long enough to prevent shoot-through. Manufacturers offer fast-body-diode superjunction variants with lifetime-controlled recovery for hard-switched bridge topologies, and in resonant converters the body diode is deliberately used under zero-voltage switching, where recovery stress is largely absent.

Operating Principles

The insulated-gate bipolar transistor (IGBT) combines the voltage-controlled input of a MOSFET with the bipolar output characteristics of a power transistor. A MOSFET input stage controls a wide-base bipolar output stage, achieving lower conduction losses than silicon MOSFETs at high voltages while retaining easy gate drive. The bipolar output stage provides conductivity modulation that reduces on-state voltage drop, which is why IGBTs dominate above roughly 600 V where silicon MOSFET on-resistance would be prohibitive.

Device Structure and Trade-offs

IGBTs use vertical structures with an added p+ collector layer that injects minority carriers for conductivity modulation. This injection improves on-state performance but creates stored charge that must be removed at turn-off, producing a tail current that limits switching speed compared with MOSFETs. Device optimization balances on-state voltage against switching speed, with different designs targeting different operating frequencies. IGBTs lack a usable intrinsic reverse diode, so an external or co-packaged freewheeling diode is supplied for inductive-load switching. Reverse-conducting IGBTs integrate that diode onto the same die, improving thermal utilization in the module because the transistor and diode functions share the full silicon area rather than each occupying part of it.

Key Parameters

IGBTs are sold in a small set of standard blocking classes: 600 and 650 V for appliance and single-phase drives, 1200 V for the great majority of three-phase industrial drives on 400 to 480 V mains, 1700 V for larger drives and traction auxiliaries, and 3300, 4500, and 6500 V for rail traction and medium-voltage converters. Collector-emitter saturation voltage (V_CE(sat)) determines conduction losses and is typically 1.5 to 3 V depending on current, class, and temperature. Unlike a MOSFET, the IGBT has a floor in its on-state voltage set by the collector junction, so conduction loss scales less than quadratically with current but never approaches zero at light load. Turn-off tail current from recombination of stored charge contributes significant switching losses at high frequency, and datasheets specify turn-on and turn-off energies (E_on, E_off) in millijoules per pulse at a stated bus voltage, current, gate resistance, and junction temperature, from which the designer computes switching loss directly. Short-circuit withstand capability, commonly specified at 10 microseconds, allows protection systems time to respond to fault conditions. Thermal characteristics, including junction-to-case resistance, determine achievable power dissipation.

IGBT Generations and Types

IGBT technology has advanced through multiple generations with improvements in switching speed, conduction losses, and robustness. Punch-through, non-punch-through, and field-stop designs offer different trade-offs between parameters, with field-stop (also called soft-punch-through) structures now dominant for their balance of low losses and stable behavior over temperature. Trench-gate IGBTs reduce on-state losses compared with planar-gate structures, and carrier-stored trench-gate designs achieve very low on-state voltage at the cost of increased switching losses.

IGBT Modules

High-power applications typically use IGBT modules that package multiple dies with associated freewheeling diodes in thermally efficient configurations. Half-bridge, full-bridge, and three-phase (six-pack) modules provide complete power-stage building blocks. Advanced packaging using sintered die attach, copper wire bonding, and improved thermal interface materials enhances power density, thermal cycling endurance, and reliability. Module selection, gate-drive hardware, protection schemes, and cooling are treated in detail in IGBT Modules and Drivers.

Silicon Controlled Rectifiers

Silicon controlled rectifiers (SCRs) are four-layer PNPN devices that can be triggered from a blocking state to a conducting state by a gate-current pulse but cannot be turned off through the gate. Once triggered, the SCR latches into conduction until the anode current falls below the holding-current level, which in line-commutated circuits occurs naturally at each zero crossing. This latching behavior limits control flexibility but enables handling of very high currents and voltages with minimal gate-drive power. Large press-pack SCRs on wafers of 100 to 150 mm block several kilovolts and carry thousands of amperes in a single device, and their surge ratings tolerate fault currents far beyond anything a transistor of comparable rating survives. Two further parameters govern their use: the critical rate of rise of off-state voltage, above which a device turns on without a gate signal, and the critical rate of rise of on-state current, which limits how fast conduction may spread from the gate region without local overheating.

TRIACs and Related AC Switches

A TRIAC behaves as two antiparallel thyristors sharing one gate, conducting in both directions and therefore switching alternating current directly. Phase control of a TRIAC provides inexpensive regulation for incandescent lighting, universal-motor speed control, and resistive heating, usually triggered by a DIAC or an optically isolated driver. Ratings extend to a few tens of amperes and blocking voltages of 600 to 800 V, sufficient for mains-connected loads.

The main limitations are commutation and noise. A TRIAC controlling an inductive load must regain its blocking capability while current and voltage are out of phase, which snubber networks or dedicated high-commutation device families address. Phase control also chops the mains waveform, generating harmonics and conducted emissions that require filtering, so electronic converters have replaced TRIACs in applications where power quality or dimming of electronic loads matters.

Gate Turn-Off Thyristors

Gate turn-off thyristors (GTOs) add the capability to interrupt current through a negative gate-current pulse. Turn-off requires gate current equal to a substantial fraction of the anode current, demanding a powerful gate driver and typically a bulky snubber to manage the rate of voltage rise. GTOs enabled controlled turn-off for medium-voltage applications but have been largely displaced by IGBTs and IGCTs in new designs.

Integrated Gate-Commutated Thyristors

Integrated gate-commutated thyristors (IGCTs) improve on GTO performance through unity-gain turn-off that diverts the entire anode current into the gate circuit before the anode begins to block voltage. This hard-driven commutation, achieved with a very low-inductance gate unit, eliminates the current crowding that limits GTO performance and largely removes the need for a turn-off snubber. IGCTs achieve among the lowest conduction losses of any fully controllable switch and are used in medium-voltage drives and flexible AC transmission system (FACTS) equipment.

Thyristor Applications

Thyristors remain important for very high power applications including HVDC transmission, static VAR compensators, and industrial heating and electrochemical processes. Their robustness under fault conditions and ability to handle extreme surge overcurrents provide advantages over transistor-based alternatives in specific applications. Line-commutated converters built from long series strings of SCRs still carry most of the world’s installed HVDC capacity and remain the technology of choice for the largest point-to-point corridors, including China’s ultra-high-voltage links at ±800 kV and above, which transmit on the order of ten gigawatts over thousands of kilometers. Voltage-source converters built from IGBTs are taking a growing share of new projects because they control real and reactive power independently, energize passive networks, and support the multiterminal and offshore-wind connections that line-commutated converters handle poorly. Thyristor Systems covers phase control, commutation, snubber design, and HVDC practice at that scale.

Silicon Carbide Properties

The 4H polytype of silicon carbide (SiC) has a bandgap near 3.2 electron volts against 1.1 for silicon, enabling higher breakdown voltage, lower leakage at temperature, and operation at elevated junction temperatures. The critical electric field is roughly ten times that of silicon, so a drift region blocking a given voltage can be about a tenth as thick and far more heavily doped, which is the source of the order-of-magnitude reduction in specific on-resistance. Thermal conductivity roughly three times that of silicon facilitates heat removal from compact die.

Practical limits temper these material advantages. Substrate and epitaxy costs remain well above silicon, though they continue to fall as wafer diameters move to 200 mm. Basal-plane dislocations in the crystal can nucleate stacking faults under bipolar conduction, which is why body-diode conduction in SiC MOSFETs receives screening attention. The material also makes minority-carrier devices unattractive at moderate voltages, so SiC practice is dominated by unipolar MOSFETs and Schottky diodes rather than IGBT-like structures.

SiC MOSFETs

Silicon carbide MOSFETs combine the voltage-controlled input of power MOSFETs with the superior material properties of SiC. Devices rated from 650 V through 1700 V are commercially mature, with 3300 V and higher-voltage parts available for traction and grid applications. SiC MOSFETs achieve dramatically lower on-resistance than silicon at equivalent voltage ratings, enabling smaller, more efficient power converters. The weaker temperature dependence of on-resistance compared with silicon eases thermal design, though gate-oxide reliability and recommended gate-drive windows demand careful attention. Channel mobility at the silicon carbide-to-oxide interface is far lower than in silicon, so the channel contributes a large share of the total on-resistance and manufacturers specify a comparatively high on-state gate voltage, commonly 15 to 20 V, together with a negative or zero off-state bias to suppress dV/dt-induced turn-on. Threshold voltage is lower and drifts more than in silicon, which makes gate-voltage margin and driver clamping a first-order reliability question rather than a detail.

SiC Schottky Diodes

Silicon carbide Schottky diodes extend the benefits of Schottky rectification to voltage ratings impractical in silicon. Their negligible reverse-recovery charge eliminates a significant loss mechanism in high-frequency converters and reduces switching stress on the companion transistor. Commercial parts span 650 V through 1700 V for mainstream converters, with 3300 V discrete devices available for higher-voltage work. Nearly all use a junction-barrier Schottky structure, in which implanted p regions shield the Schottky contact from the reverse field, cutting leakage and adding a bipolar surge path that greatly improves surge-current withstand. SiC Schottky diodes are widely used as power-factor-correction boost diodes and as freewheeling diodes alongside both SiC and silicon switches, often improving efficiency immediately when retrofitted into existing converter designs.

Gallium Nitride Devices

Gallium nitride (GaN) offers an even higher critical electric field than SiC and exceptional electron mobility through its two-dimensional electron gas, enabling very fast switching with low gate and output charge. Lateral GaN high-electron-mobility transistors (HEMTs) fabricated on silicon substrates provide cost-effective, high-performance devices, and the 650 V class dominates commercial production. Devices rated at 900 V are available, notably inside integrated switcher circuits, and 1200 V lateral parts are in early commercialization; the limit is set less by the material than by the thickness of GaN that can be grown on a silicon substrate without cracking. Enhancement-mode devices with a p-GaN gate and cascode arrangements pairing a depletion-mode HEMT with a low-voltage silicon MOSFET both give the normally-off behavior that systems require, in contrast to the native depletion-mode HEMT.

GaN imposes its own drive and application rules. The p-GaN gate of an enhancement-mode device tolerates only about 6 V and behaves as a diode rather than a capacitor, so drivers regulate gate current instead of simply applying a voltage. The lateral structure has no body diode; reverse current instead flows through the channel at a third-quadrant voltage drop that rises with negative gate bias, which puts a premium on short dead times. Most commercial GaN parts carry no avalanche rating, so bus voltage and overshoot must stay inside the rating by design, and dynamic on-resistance, the temporary increase after a hard-switching event caused by trapped charge, is characterized under realistic switching conditions rather than from static measurements alone.

Wide-Bandgap Device Applications

Wide-bandgap devices are transforming power electronics across applications. Electric-vehicle traction inverters use SiC to achieve higher efficiency and power density, supporting the move to 800 V battery architectures. Data-center and consumer power supplies use GaN for high-frequency operation with small magnetics. Photovoltaic inverters and other renewable-energy systems benefit from improved efficiency at high power levels. The higher device cost is increasingly offset by system-level savings in thermal management, passive components, and energy consumption, and continues to fall as production volumes grow. Device-level treatments at higher power appear in Silicon Carbide Devices and Gallium Nitride Devices.

Comparing the Device Families

Below roughly 200 V, trench silicon MOSFETs are difficult to beat on cost and on-resistance, and they serve point-of-load converters, battery systems, and automotive twelve-volt loads. From 400 to 900 V the choice widens: superjunction silicon MOSFETs suit offline supplies at tens to hundreds of kilohertz, GaN HEMTs win where size and frequency dominate, and SiC MOSFETs take over where efficiency at high current justifies their cost.

At 1200 V and above the contest is between silicon IGBTs and SiC MOSFETs. The IGBT holds a cost advantage and tolerates fault current better; the SiC MOSFET cuts switching loss enough to raise frequency, shrink magnetics, and improve part-load efficiency, which is why traction inverters have adopted it. Above a few kilovolts and at line frequency, thyristors and IGCTs remain unmatched for conduction loss and surge robustness. A useful rule is that conduction loss decides the winner in low-frequency, high-current designs, while switching loss decides it wherever frequency is high enough to shrink the passive components.

Voltage and Current Ratings

Device voltage rating must exceed the maximum expected blocking voltage, including switching transients and a safety margin. Current rating must accommodate peak currents while respecting duty cycle and thermal limitations. Derating for temperature, particularly the rise in on-resistance for silicon MOSFETs, must be considered. Safe-operating-area limitations constrain the voltage-current trajectory during switching, especially under hard-switched or short-circuit conditions.

Switching Speed Considerations

The required switching frequency determines acceptable switching losses and thus the speed class of device needed. GaN devices reach the megahertz range, silicon and SiC MOSFETs operate from tens of kilohertz to the low megahertz, IGBTs are practical to tens of kilohertz, and thyristors are limited to line frequency in phase-control applications. Faster devices and higher slew rates reduce switching loss but increase electromagnetic interference and the risk of voltage overshoot, requiring optimization for each application.

Thermal Management

Device power dissipation from conduction and switching losses must be removed to keep the junction temperature within ratings. Thermal resistance from junction to case, case to heat sink, and heat sink to ambient determines achievable dissipation. Package selection affects thermal performance, with larger packages and double-sided cooling enabling lower thermal resistance. Paralleling devices shares the thermal load but requires attention to current sharing, layout symmetry, and matched parameters.

Drive Circuit Requirements

Each device type has specific drive requirements for proper operation. MOSFETs and IGBTs need voltage drive with adequate peak current for fast switching, and gate resistors set the trade-off between switching speed and ringing. Negative gate bias during the off-state improves noise immunity for IGBTs and many SiC MOSFETs. High-side switches require isolated or bootstrap supplies and level-shifted control. Proper gate-drive design prevents spurious operation and ensures reliable switching.

Overcurrent Protection

Power semiconductors require protection from overcurrent conditions that exceed their thermal capability or safe operating area. Current sensing enables fast detection and shutdown. Desaturation detection identifies when an IGBT or SiC MOSFET leaves the saturated, low-loss region during a fault. Fuses and circuit breakers provide backup protection but generally operate too slowly to protect the semiconductor itself, so gate-level shutdown remains the primary defense.

Overvoltage Protection

Inductive loads and circuit parasitics generate voltage spikes during switch turn-off that can exceed device ratings. Snubber circuits and clamping devices absorb energy and limit voltage, while careful layout minimizes the loop inductance that drives overshoot. Active-clamp circuits use the switch itself to limit overvoltage by briefly returning it to conduction. Device selection must account for expected transient voltages with appropriate margin.

Short-Circuit Withstand

IGBTs are commonly specified with a short-circuit withstand time of about 10 microseconds, allowing protection systems to detect and clear faults. MOSFETs, including SiC MOSFETs, generally tolerate shorter durations because of weaker current saturation and rapid self-heating. Short-circuit protection must therefore operate within the device withstand time to prevent destructive failure, making fast desaturation detection and prompt gate shutdown essential.

Thermal Considerations

Junction temperature must remain below maximum ratings under all operating conditions, including worst-case ambient temperature and load. Transient thermal-impedance models, usually published as a Foster or Cauer network, predict temperature rise during pulsed operation. Overtemperature protection using on-die or in-module temperature sensors prevents thermal runaway, and power and thermal cycling are managed to limit fatigue of die-attach and bond-wire interfaces. Adequate thermal-design margin ensures reliable long-term operation. High-voltage devices held at high DC bus voltage are additionally subject to single-event burnout from atmospheric neutrons, so manufacturers publish failure-in-time rates against blocking voltage and altitude, and designers derate the bus voltage accordingly on aircraft and high-altitude installations.