Semiconductor and Device Technology III-V Semiconductors Informational

What is a metamorphic HEMT and when would I use it instead of a standard pseudomorphic HEMT?

A metamorphic HEMT (mHEMT) uses a graded composition buffer layer to accommodate a large lattice mismatch between the substrate (typically GaAs) and the channel material (InGaAs with high indium content). This allows using a higher-performance channel on a lower-cost substrate: (1) Pseudomorphic HEMT (pHEMT): the channel layer (InGaAs) is grown directly on GaAs with constrained (pseudomorphic) strain. The indium content is limited to approximately 20-25% InxGa1-xAs (beyond this: the strain exceeds the critical thickness and the layer relaxes with defects). With 20% In on GaAs: electron mobility ≈ 6,000-8,000 cm²/V·s. fT ≈ 60-120 GHz (for 100-250 nm gate). NF at 28 GHz ≈ 0.7-1.5 dB. This is the standard technology for RF amplifiers from DC to 40 GHz. (2) Metamorphic HEMT (mHEMT): a graded InAlAs buffer is grown on GaAs, slowly increasing the In content from 0% (lattice-matched to GaAs) to 40-80%. The dislocations from the lattice mismatch are confined to the buffer layer. On top of the buffer: a high-In channel (In0.6-0.8Ga0.2-0.4As) is grown. The high-In channel has: much higher electron mobility: 10,000-12,000 cm²/V·s (50-70% higher than pHEMT). Higher electron velocity (especially important at mmWave). The result: higher fT (100-300 GHz for 100-50 nm gate), higher fmax, lower NF (0.3-0.8 dB at 28 GHz), and higher gain per stage. The mHEMT on GaAs achieves performance approaching InP HEMT, but on a less expensive GaAs substrate. (3) InP HEMT: the channel (In0.53Ga0.47As or In0.7Ga0.3As) is grown lattice-matched on an InP substrate. No buffer transition needed (the InP substrate naturally accommodates high-In channels). Mobility: 10,000-15,000 cm²/V·s. fT > 300 GHz. NF is the lowest of any technology. But: InP substrates are smaller (3-4 inch), more expensive, and more fragile than GaAs. When to use mHEMT instead of pHEMT: (a) When the frequency is above 40 GHz and pHEMT gain is insufficient. (b) When the NF requirement is below 0.7 dB at 28 GHz (pHEMT cannot achieve this; mHEMT can). (c) When InP performance is needed but InP cost/availability is unacceptable.
Category: Semiconductor and Device Technology
Updated: April 2026
Product Tie-In: Transistors, MMICs, Evaluation Boards

mHEMT vs pHEMT Technology

The mHEMT represents a clever engineering compromise: achieving InP-like device performance on the mature, cost-effective GaAs platform.

Common Questions

Frequently Asked Questions

Which foundries offer mHEMT?

Commercial mHEMT foundry processes: (1) Fraunhofer IAF (Germany): mHEMT processes at 100 nm and 50 nm gate length. fmax > 500 GHz for 50 nm. Available through European space agency and commercial contracts. (2) UMS (United Monolithic Semiconductors, France): mHEMT at 70 nm and 100 nm. Used for European defense and space applications. (3) WIN Semiconductors (Taiwan): mHEMT under development (as of 2025). (4) OMMIC (France, now part of TDK): mHEMT at 70 nm for D-band (110-170 GHz) applications. The mHEMT foundry landscape is smaller than pHEMT (which is offered by many foundries worldwide). Most mHEMT users are in the defense, space, and high-frequency communications sectors.

Does the buffer layer degrade reliability?

The buffer dislocations can be a reliability concern: (1) DARPAs mHEMT reliability studies show that well-designed buffers achieve reliability comparable to pHEMT (MTTF > 10⁶ hours at 150°C). (2) Poorly designed buffers (high dislocation density > 10⁷ cm⁻²) show degraded reliability: the dislocations can propagate under thermal and electrical stress, increasing leakage and degrading gain over time. (3) For space applications: mHEMT has been qualified for satellite payloads (ESA has flown mHEMT LNAs on several missions). The reliability is adequate for 15+ year mission life. (4) Best practice: source mHEMT from a foundry with demonstrated reliability data (MTTF curves at multiple temperatures, activation energy analysis). Verify the buffer quality with TEM (transmission electron microscopy) cross-sections showing the dislocation confinement.

When should I choose InP HEMT over mHEMT?

Choose InP HEMT when: (1) The operating frequency is above 100 GHz (only InP provides adequate gain at D-band and beyond). (2) The NF requirement is < 0.4 dB at 28 GHz or < 1 dB at 100 GHz (achieved by InP but not by mHEMT). (3) The application is radio astronomy or scientific instrumentation (where the absolute lowest noise is required, regardless of cost). (4) The volume is small (< 1000 units), so the InP cost premium is not the dominant factor. Choose mHEMT when: (1) 40-100 GHz operation with performance between pHEMT and InP. (2) NF of 0.4-0.8 dB at 28 GHz is acceptable. (3) The volume is medium (1,000-100,000 units) and the GaAs cost advantage matters. (4) Supply chain: GaAs foundries are more widely available than InP.

Need expert RF components?

Request a Quote

RF Essentials supplies precision components for noise-critical, high-linearity, and impedance-matched systems.

Get in Touch