For decades, a power MOSFET designer faced a wall: raise the breakdown voltage and the on-resistance climbs, because the drift region must get thicker and more lightly doped. Super junction (SJ) technology broke through that wall with charge-balanced pillars, and it changed the economics of high-voltage switching supplies. This article explains the physics behind the breakthrough, what an SJ datasheet actually tells you, and how to select one for PFC, flyback, and LLC stages without falling for headline numbers.
Why Conventional MOSFETs Hit a Wall
A planar high-voltage MOSFET supports reverse voltage across a drift region—a thick, lightly doped layer that blocks the field. To withstand more volts, the region must be thicker and less doped, which raises the resistance of the conducting path. The result is a brutal scaling law: specific on-resistance (RDS(on) × die area) grows roughly with breakdown voltage to the power of 2.5. At 600 V and above, a planar device needs a large, expensive die to keep RDS(on) acceptable, and the large die brings large capacitances and slow switching.
The classic trade-off can be summarized in one sentence: in a conventional MOSFET, blocking voltage and conduction resistance are both controlled by the same drift region, so improving one worsens the other.
How Super Junction Breaks the Trade-Off
A super junction MOSFET replaces the uniform drift layer with alternating P and N pillars. Under reverse bias, the P and N pillars deplete each other laterally. That mutual depletion acts like a much lighter effective doping, so the structure can block high voltage without needing a thick, highly resistive region. The pillars are then doped more heavily than a conventional drift layer, which slashes the on-resistance path.
This charge-balance concept is why SJ parts are sometimes called “cool” or “charge-compensated” MOSFETs. The balancing is delicate: if the P and N charges are not matched, breakdown voltage collapses and avalanche ruggedness suffers. Manufacturing a well-balanced pillar structure—often by deep trench filling or repeated epitaxy—is hard, which is why SJ MOSFETs took years to become mainstream and why process control is a real supplier differentiator.
In practical terms, a 600–650 V SJ MOSFET can offer roughly a 3–5× reduction in specific on-resistance compared with a planar device of the same class. That translates to smaller die, lower conduction loss, or both.
What Super Junction Actually Buys You—and What It Costs
The headline benefit is low RDS(on) at high voltage, but the structure changes several other characteristics, and not all of them are free:
| Characteristic | Planar MOSFET | Super Junction MOSFET |
|---|---|---|
| RDS(on) vs die size | Higher specific resistance | Much lower specific resistance |
| Output capacitance Coss/Qoss | Generally higher per die area | Lower Qoss benefits at high voltage |
| Gate-drain charge Qgd | Moderate | Tends to be lower, faster switching |
| Body-diode recovery | Soft, slow | Slower, “harder” recovery with more Qrr |
| dv/dt ruggedness | Proven | Improved in modern generations, but design-dependent |
| Cost at high voltage | Higher die cost | Lower per RDS(on), but more complex process |
Two consequences deserve special attention in switching supplies:
Body-diode recovery. The SJ structure’s charge balance also affects the body diode. In half-bridge and LLC topologies, where current freewheels through the body diode, the recovery charge (Qrr) and its temperature behavior can create losses and ringing. Designers often add an external fast-recovery diode in parallel or choose the operating point to avoid hard body-diode commutation. This is not a reason to avoid SJ parts—it is a reason to check the reverse-recovery data instead of ignoring it.
Gate charge and dv/dt. Lower Qgd is generally good for switching speed, but the fast edge can excite parasitic resonance with PCB inductance. The datasheet’s gate-charge curve and the switching waveforms in the application note matter more than the single Qg number on the front page.
Reading an SJ Datasheet Like an SMPS Designer
When comparing super junction parts, ignore the marketing summary and read these values:
- RDS(on) at the real junction temperature. SJ datasheets list RDS(on) at 25 °C and show a temperature multiplier curve. At 100–125 °C, on-resistance typically rises 60–100%. Compare parts at your operating temperature, not at 25 °C.
- EAS (avalanche energy). The charge-balanced structure makes avalanche behavior process-sensitive. If the circuit can see unclamped inductive spikes—flyback leakage, motor stall, hot-plug events—EAS and the single-pulse avalanche curve decide survivability.
- Qoss and Coss characteristics. In LLC and other resonant stages, the energy to charge/discharge Coss is stored and partly recovered; the Qoss versus voltage curve is more useful than the single capacitance value.
- Body-diode reverse recovery (trr, Qrr). For bridges and synchronous stages, recovery data at the real di/dt and temperature is essential.
- Switching times and di/dt control. Check the test circuit conditions. Different suppliers test with different gate resistances; compare like for like.
Where Super Junction MOSFETs Fit in SMPS
Power factor correction (PFC). The boost stage in active PFC operates from 100–400 V DC and switches at tens to hundreds of kilohertz. The combination of high blocking voltage, low RDS(on), and low Qgd makes SJ MOSFETs the mainstream switch for continuous-conduction-mode PFC in adapters, chargers, and server supplies. The boost diode’s recovery behavior must be matched to the switch, because it feeds directly into the MOSFET’s turn-on loss; the diode decision is the subject of a companion article in this series on PFC rectifiers.
Flyback converters. The 600–650 V rating suits universal-input flyback designs (including QR and active-clamp variants). The low RDS(on) reduces primary conduction loss, and the low Qgd keeps switching loss manageable. Avalanche energy matters here because transformer leakage can push the drain voltage above the steady-state clamp.
LLC resonant converters. SJ MOSFETs are common in the primary side of LLC converters, where ZVS reduces switching loss. The important parameters shift to Coss and the body diode, and designers often review the half-bridge behavior carefully.
For the device families and typical parts used in these stages, the SMPS application section of the Good-Ark site lists super junction MOSFETs alongside the bridge rectifiers and fast-recovery diodes that complete the supply chain.
A Sizing Example: 300 W CCM PFC Boost
To make the selection concrete, consider a 300 W universal-input PFC boost stage with a 400 V DC output, switching at 100 kHz in continuous conduction mode. The MOSFET sees a peak current of roughly 5 A and an average current near 2.5 A in the conduction interval. The figures below are an illustrative calculation for method only; real values depend on the specific device, gate drive, and waveform.
Step 1: pick the voltage class. Universal input with 264 V AC peak gives about 373 V peak bus voltage; with switching spikes and derating, 600–650 V is the standard class. Step 2: estimate conduction loss. A 0.35 Ω SJ MOSFET at 100 °C has an on-resistance near 0.6 Ω, so conduction loss is approximately I² × RDS(on) × duty, on the order of 2–3 W under this duty assumption. Step 3: estimate switching loss. With a low Qgd part and a 10 Ω gate resistor, Eon + Eoff might land near 100–150 µJ per cycle, adding 10–15 W at 100 kHz—which is why PFC designs tune the gate resistor and choose the boost diode for low recovery charge to keep the MOSFET’s turn-on current low. Step 4: check the thermal budget. The total loss of 12–18 W in a TO-220 or DPAK with a heatsink determines the junction temperature; the datasheet’s thermal resistance and the RDS(on) temperature multiplier close the loop.
The point of the exercise is not the specific numbers—they depend on the exact part and waveform—but the method: every parameter in the loop (RDS(on) hot, Qgd, thermal resistance, boost-diode recovery) interacts, and an SJ part chosen on the 25 °C RDS(on) headline alone will overheat before the lab results match the spreadsheet.
Topology-to-Parameter Matrix: PFC, Flyback, LLC
The same SJ MOSFET family is tuned differently for each topology. The table below maps the topology’s dominant stress to the parameter that should lead the selection:
| Topology | Dominant stress | Lead parameter | SJ strength to exploit |
|---|---|---|---|
| CCM PFC | Hard-switched turn-on, high dv/dt | Qgd, boost-diode recovery, EAS for surge | Low Qgd keeps switching loss down; check EAS for line-surge events |
| QR flyback | Valley switching, transformer leakage spikes | EAS, RDS(on) hot | Avalanche energy absorbs leakage; low RDS(on) cuts primary loss |
| LLC (half-bridge) | ZVS transitions, body-diode conduction | Coss/Qoss, body-diode recovery, dead-time fit | Low Qoss eases ZVS; verify body-diode behavior in the bridge |
Use the table as the first filter: it tells you which columns of the datasheet to read for your topology before you compare part numbers. The remaining checks—junction-temperature loss, package thermal path, and supply qualification—are then applied to the shortlist rather than to every device on the market.
Frequently Asked Questions
Why is super junction better than planar for high-voltage MOSFETs? The charge-balanced P/N pillar structure decouples blocking voltage from conduction resistance, so a 600–650 V part can offer much lower RDS(on) per die area than a planar device.
Do super junction MOSFETs have a weak body diode? Modern SJ parts have improved recovery, but the body diode is generally not as fast or soft as a dedicated fast-recovery diode. In bridge and resonant topologies, verify trr/Qrr and consider an external diode if needed.
Can SJ MOSFETs handle avalanche? Yes, but the charge-balance process makes ruggedness design-sensitive. Check the datasheet’s EAS rating and single-pulse avalanche curve for your circuit’s worst-case energy.
Are super junction MOSFETs limited to offline SMPS? They are most common in offline power conversion (PFC, flyback, LLC), but they also appear in lighting drivers, solar micro-inverters, and other high-voltage switching applications where the same voltage-loss trade-off applies.
How do SJ MOSFETs compare with SiC at 650 V? SJ remains the cost-effective silicon answer for most offline supplies; SiC wins where higher frequency, higher temperature, or tighter efficiency targets pay for the premium. The choice is system-level, and the switching-device decision for SiC is covered separately in this series.
Before You Choose an SJ MOSFET
Start from the topology, not the datasheet front page: PFC asks about Qgd and boost-diode recovery, flyback asks about avalanche energy, and LLC asks about Coss and body-diode behavior. Apply the topology-to-parameter matrix first, then size the junction-temperature loss and thermal path for the shortlist. The SMPS application section shows how SJ MOSFETs combine with bridge and recovery diodes in a full supply design, and the contact page connects you to the team for part-level selection support.