Two letters on a datasheet—”trench” or “planar”—tell you more about a power MOSFET than most of its front-page numbers. The structure determines how the device achieves its on-resistance, how fast it switches, how much gate charge it needs, and where it will and will not be cost-effective. This article compares the two technologies at the level a design engineer needs, explains why trench devices dominate low-voltage power conversion, and shows how to read the structure story out of ordinary datasheet parameters.
Where Trench Technology Came From
The first power MOSFETs used a planar DMOS structure: the gate ran across the surface of the die, and current flowed from source to drain through a channel at the surface and then down through the drift region. Increasing current density meant shrinking cell pitch, but the planar layout hit limits—the surface channel consumed area, and the JFET effect between neighboring cells squeezed the current path.
Trench MOSFETs turned the channel vertical. The gate is etched down into the silicon as a trench, so the channel runs along the trench sidewall. That orientation packs far more channel width per unit area, eliminates the JFET pinch region, and lets current flow straight down to the drain. The result was a step change in cell density and a sharp drop in specific on-resistance for low- and mid-voltage parts.
Structure Comparison at a Glance
| Feature | Planar MOSFET | Trench MOSFET |
|---|---|---|
| Channel orientation | Horizontal at the surface | Vertical along the trench sidewall |
| Cell density | Limited by lateral geometry | Much higher |
| Specific RDS(on) at low voltage | Higher | Lower (the main reason trench won) |
| Gate-drain overlap | Larger (lower Qgd per RDS(on)) | Smaller, but design-dependent |
| Switching behavior | Generally slower at low voltage | Faster edges, less stored charge |
| Ruggedness | Well understood, proven | Mature, but needs good termination design |
| Cost below ~150 V | Higher per RDS(on) | Lower per RDS(on) |
| High-voltage blocking | Standard approach historically | Now used with charge balance at 600 V+ too |
The table makes the pattern obvious: below roughly 100–150 V, trench technology is dominant for power switching; planar parts survive in niche roles where their simplicity or specific characteristics still fit.
What Trench Changes in Real Parameters
RDS(on) and current density. The vertical channel removes the lateral JFET constriction, so the specific on-resistance drops dramatically. For a 30–100 V MOSFET, a trench design can roughly halve the RDS(on)×area product compared with planar of the same generation. Designers see the result as smaller packages for the same current, or lower loss in the same footprint.
Gate charge. Trench devices generally reduce the gate-drain charge component because the overlapping area between gate and drain is smaller and better controlled. Lower Qgd means a shorter Miller plateau and faster switching for the same drive strength. This is why trench MOSFETs pair naturally with high-frequency buck converters and synchronous rectification.
Body diode. The body diode of a trench MOSFET is still a PN junction with stored charge; the structure itself does not fix recovery. In synchronous buck stages, the dead-time behavior and any negative-current commutation still need the same care as with planar parts. Some designs add Schottky diodes in parallel or rely on the controller’s timing to minimize body-diode conduction.
Avalanche and SOA. Ruggedness depends on termination and edge design, not just the trench geometry. Modern trench parts have good avalanche ratings, but the safe-operating-area curves and EAS numbers must be checked for the specific application—motor drives and inductive loads are the classic cases where avalanche energy matters.
Where Trench MOSFETs Dominate
The list of applications is essentially the list of modern power conversion:
- DC-DC converters. Buck converters in servers, telecom, and POL modules use trench MOSFETs for both high-side and synchronous low-side switches, where low RDS(on) and low Qg drive efficiency.
- Synchronous rectification. Low-voltage, high-current outputs are the trench MOSFET’s home turf—the low RDS(on) directly converts into efficiency points.
- Load switches and power management. Battery protection, power path switching, and hot-swap circuits value the small footprint and low conduction loss.
- Motor drivers. Brushless DC and brushed motor bridges use trench MOSFETs for their low on-resistance and fast switching.
- Low-voltage industrial and consumer power. Power tools, chargers, and appliance drives all follow the same logic: at 12–100 V, trench is the cost-effective way to get low loss.
For the discrete families and packages that support these stages, the product catalog on the Good-Ark site provides the family-level overview. The same selection discipline applies whether the switch is a trench part at 40 V or a super junction part at 650 V; the application-side decisions—logic-level direct drive and synchronous rectification—are covered by companion articles in this series.
Trench Variants: Shielded, Split-Gate, and Logic-Level Options
The basic trench structure has evolved into several variants, and the datasheet rarely says which one you are looking at. Understanding them helps you interpret parameter differences between suppliers:
- Standard trench. The gate trench runs the full depth of the active region. Simple, proven, and still dominant in low-voltage power switching.
- Shielded-gate (split-gate) trench. A second, source-connected electrode below the gate shields the drain voltage from the gate. The result is a large reduction in gate-drain capacitance (Qgd), faster switching, and lower switching loss—at the cost of a slightly more complex process. These parts shine in high-frequency converters.
- Trench with logic-level option. The channel design can be adjusted to lower the threshold and guarantee RDS(on) at 4.5 V or 2.5 V, letting the part be driven directly from logic-level outputs for moderate loads. The drive-margin checks for that case are covered in the companion article on logic-level MOSFETs.
The shielded-gate family is worth a specific mention: when a datasheet shows an unusually low Qgd relative to Qg, or a “low gate charge” marketing claim at 40–100 V, it is usually a shielded or split-gate trench design. For a 500 kHz buck converter, the difference between a standard trench and a shielded-gate part can be several watts of switching loss and a different EMI profile.
Paralleling trench MOSFETs. When one device cannot carry the current, multiple parts are paralleled. Trench MOSFETs share current reasonably well in the on-state because RDS(on) has a positive temperature coefficient, but the switching transient is shared by the layout: each gate trace must see the same drive impedance, and the drain-source loops must be symmetric. A well-paralleled pair performs better than a single larger die in some thermal layouts because heat spreads across the board—but only if the gate drive is symmetric and the current-sense path is not shared in a way that creates positive feedback.
Reading the Structure Out of a Datasheet
The datasheet rarely says “trench” prominently on older parts, but the numbers tell the story:
- RDS(on) versus package. If a small package like DPAK or SO-8 can carry tens of amps at low voltage with reasonable RDS(on), the die is almost certainly trench.
- Qg versus RDS(on). Plot the figure of merit RDS(on) × Qg. Trench parts cluster at the low corner; planar parts sit higher for the same voltage class.
- Threshold and gate drive. Trench parts are available with standard and logic-level options; the choice depends on the drive voltage, not the structure.
- Reverse recovery data. When the body diode matters, compare trr and Qrr at the operating temperature—structure alone does not predict recovery quality.
- SOA and avalanche curves. For industrial loads, these curves decide whether the part survives fault conditions; read them at the worst-case case temperature.
Voltage-Band Technology Map
The structure decision maps cleanly onto voltage bands. Use this map as the first filter, then compare datasheets within the recommended column:
| Voltage band | Recommended structure | Why | Watch out for |
|---|---|---|---|
| Below 40 V | Trench, including logic-level options | Lowest specific RDS(on) and small footprint | Gate-drive margin in 3.3 V/5 V logic designs |
| 40–100 V | Trench and shielded-gate trench | Best RDS(on) × Qg figure of merit for fast switching | Gate-loop ringing at high dv/dt |
| 100–200 V | Trench or planar, by application | Transition zone; both exist, cost and SOA decide | SOA and thermal behavior under pulse loads |
| Above 200 V | Super junction families | Voltage-loss trade-off favors charge balance | Body diode and Qoss in bridge topologies |
Two practical consequences follow from the map. First, in the low-voltage bands the RDS(on) × Qg figure of merit is the fastest differentiator: a shielded-gate trench part with similar RDS(on) but half the Qgd usually wins in a 500 kHz converter even if its headline current is the same. Second, the map prevents a common mistake—choosing a planar part in the 40–100 V band because its RDS(on) looks acceptable, when a trench device delivers the same conduction loss with less die area and better switching.
The structure is the first filter, not the last. Package, thermal path, and gate-drive compatibility still decide the final part number, but those checks belong to the shortlist that the map produces.
Frequently Asked Questions
Is a trench MOSFET always better than planar? Not universally. Trench technology wins on specific on-resistance and usually on gate charge for low- and mid-voltage parts, but planar devices still exist in niche roles. Compare the actual RDS(on), Qg, and ruggedness data for your voltage class.
What voltage range does trench technology cover? Modern trench MOSFETs span from a few volts to 100–150 V for mainstream power switching, and trench structures also appear in 600 V+ super junction and IGBT devices. The classic trench advantage is strongest below 150 V.
Do trench MOSFETs have a body diode problem? The body diode is a silicon PN junction with recovery charge, regardless of structure. In synchronous and bridge topologies, verify dead time and recovery behavior, and consider an external Schottky where needed.
How do I choose between a trench MOSFET and a super junction MOSFET? Voltage class decides most cases: super junction parts dominate 400–900 V offline conversion, while trench parts dominate low-voltage DC-DC and load switching. Select by application voltage, then by the RDS(on) × Qg figure of merit within the recommended band.
What is the difference between a standard trench and a shielded-gate trench? The shielded-gate design adds a source-connected electrode that shields the gate from the drain voltage, cutting Qgd and switching loss. It is the usual reason one datasheet shows a much lower gate charge than another at the same voltage and RDS(on).
Does trench technology improve avalanche ruggedness? Not by itself. Avalanche capability depends on the termination and edge design, not the trench geometry. Check the EAS rating and the single-pulse avalanche curve for inductive-load applications.
Choosing by Voltage Band
Trench technology won the low-voltage world because the vertical channel packs more current into less die with lower on-resistance and gate charge than planar predecessors. The decision rule is now simple: read your operating voltage, find the band in the technology map, and compare parts within that column using the RDS(on) × Qg figure of merit. Package and thermal checks finish the selection. The product catalog groups the low-voltage families and packages, and the contact page is the direct channel for part-level support.