High-electron-mobility transistors (HEMTs and HEM FETs) use a heterojunction and a two-dimensional electron gas (2DEG) channel to reach very high speed, gain, and low noise in RF, millimeter-wave, and power circuits. This article explains their layer structure, materials, modes, growth methods, reliability, modeling, and PCB layout in clear steps.
HEMTs and HEM FETs Basics
High-electron-mobility transistors (HEMTs or HEM FETs) are field-effect transistors that use a boundary between two different semiconductor materials instead of a single, uniformly doped channel as in a MOSFET. This boundary, called a heterojunction, lets electrons move very quickly in a thin layer with low resistance. Because of this, HEMTs can switch at very high speeds, provide strong signal gain, and keep noise low in high-frequency circuits. Common material systems such as GaN, GaAs, and InP are chosen to balance speed, voltage strength, and cost, so HEMTs see wide use in modern high-frequency and high-power electronics.
2DEG Channel in HEMTs and HEM FETs
In HEMTs, the high mobility comes from a very thin layer of electrons called a two-dimensional electron gas (2DEG). This layer forms at the boundary between a wide-bandgap layer and a narrower-bandgap channel. The channel is undoped, so electrons move with fewer collisions, giving a fast, low-resistance path for current.
Steps in 2DEG formation:
• Donor atoms in the wide-bandgap layer release electrons.
• Electrons move into the lower-energy narrow-bandgap channel.
• A thin quantum well forms and traps the electrons in a sheet.
• This 2DEG sheet acts as a fast channel controlled by the gate.
Layer Structure in HEMTs and HEM FETs
n⁺ cap layer (low bandgap)
Provides a low-resistance path for the source and drain contacts. The cap is removed under the gate to keep the channel controlled.
n⁺ wide-bandgap donor/barrier layer
Supplies electrons that fill the 2DEG and helps handle high electric fields.
Undoped spacer layer
Separates the donors from the 2DEG so electrons see fewer collisions and can move more easily.
Undoped narrow-bandgap channel/buffer
Holds the 2DEG and lets current flow quickly at high frequencies and high fields.
Substrate (Si, SiC, sapphire, GaAs, or InP)
Supports the whole structure and is chosen for heat handling, cost, and material match; GaN-on-Si and GaN-on-SiC are common in power and RF HEMTs.
Material Options for HEMTs and HEM FETs
| Material system | Main strengths | Typical frequency range |
|---|---|---|
| AlGaAs / GaAs | Low noise, stable, and well-developed | Microwave to low mmWave |
| InAlAs / InGaAs on InP | Very high speed, very low noise | mmWave and higher |
| AlGaN / GaN on SiC or Si | High voltage strength, high power, hot-ready | RF, microwave, power switching |
| Si / SiGe | Works with CMOS, better mobility than silicon | RF and high-speed digital |
pHEMT and mHEMT Structures in HEMTs and HEM FETs
| Type | Lattice approach | Main benefits | Typical limits/tradeoffs |
|---|---|---|---|
| pHEMT | Uses a very thin, strained channel kept below a critical thickness to match the substrate | High electron mobility, low defects, stable performance | Channel thickness is limited; stored strain must be managed |
| mHEMT | Uses a graded “metamorphic” buffer that slowly changes the lattice constant | Allows high indium content and very high speed (high fT) | More complex buffer, higher risk of crystal defects |
Enhancement and Depletion Modes in HEMTs and HEM FETs
Depletion-mode HEMTs (dHEMT, normally on)
• Often found in AlGaN/GaN structures where a 2DEG forms by itself.
• The device conducts at VGS = 0V; a negative gate voltage is needed to shut the channel off.
• Can reach very high power levels and high breakdown voltage but needs extra care to make the system fail-safe.
Enhancement-mode HEMTs (eHEMT, normally off)
• Built so the channel is off at VGS = 0V.
• Methods include gate recess, p-GaN gate, or fluorine treatment to shift the threshold to a positive value.
• Acts more like a MOSFET, which can make power and automotive circuits easier to protect and control.
RF and Millimeter-Wave Roles of HEMTs and HEM FETs
In RF and millimeter-wave circuits, HEMTs and HEM FETs are widely used because they can switch very fast and add only a small amount of noise to the signal. Their structure gives them high gain and lets them work at frequencies where many silicon devices start to struggle.
In these systems, HEMTs often serve as low-noise amplifiers that boost weak signals with minimal added noise, and as power amplifiers that drive stronger signals at high frequency. Advanced HEMT technologies can keep useful gain well into the millimeter-wave range, so they see wide use in very high-frequency communication and sensing circuits.
GaN HEMTs and HEM FETs in Power Conversion
GaN HEMTs and HEM FETs are now used as main switches in high-efficiency, high-frequency power converters in the 100–650 V range. They have much lower switching loss than much silicon MOSFETs, so they can run at hundreds of kilohertz or even into the megahertz range while still staying efficient.
These devices also offer low on-resistance and low charge, which helps cut both conduction and switching losses. Their strong electric field and good temperature handling support smaller magnetics and more compact power stages. To get these benefits safely, the gate drive, PCB layout, and EMI control must be carefully planned so that fast voltage edges and ringing stay under control.
Epitaxial Growth for HEMTs and HEM FETs
MBE (Molecular Beam Epitaxy)
• Uses ultra-high vacuum and very precise control of growth.
• Common in research and low-volume, very high-performance HEMTs.
MOCVD (Metal-Organic CVD)
• Supports high wafer throughput.
• Used for commercial GaN and GaAs HEMTs, balancing performance and production cost.
Reliability and Dynamic Behavior in HEMTs and HEM FETs
GaN-based HEMTs and HEM FETs can run into reliability issues when they switch at high voltage and high power. Traps in the buffer, surface, or interfaces can catch charge during switching, which raises dynamic on-resistance and cuts current, leading to current collapse compared with DC operation.
Strong electric fields and high temperatures near the gate can add extra stress. Over time, repeated switching, heat, humidity, or radiation can slowly change values like threshold voltage and leakage, so good thermal design and protection support long-term stability.
Conclusion
HEMT and HEM FET behavior comes from the 2DEG channel, chosen material system, and pHEMT or mHEMT structure, shaped by enhancement or depletion mode design. Together with MBE or MOCVD growth, traps, dynamic resistance, and thermal limits define real performance. Accurate RF and power models plus careful PCB and packaging choices keep the operation stable.
Frequently Asked Questions [FAQ]
What gate-drive voltage do GaN HEMTs need?
Most enhancement-mode GaN HEMTs use about 0–6 V gate drive.
Do HEMTs need special gate drivers?
Yes. They need fast, low-inductance gate drivers, often dedicated GaN driver ICs.
Which packages are common for HEMTs and HEM FETs?
RF HEMTs use RF ceramic or surface-mount packages. Power GaN HEMTs use QFN/DFN, LGA, low-inductance power packages, or some TO-style packages.
How does temperature affect HEMT performance?
Higher temperature raises the on resistance, reduces current, lowers RF gain, and increases leakage.
How are HEMTs tested in power converters?
They are checked with a double-pulse test to measure switching energy, overshoot, ringing, and RDS(on).
What safety measures are important for high-voltage GaN HEMTs?
Use reinforced isolation, proper fuses or breakers, surge protection, correct creepage and clearance, controlled dv/dt, and protected gate drive.
