How do I select between a pHEMT and a mHEMT process for a millimeter wave LNA?

Selecting between a pHEMT (pseudomorphic High Electron Mobility Transistor) and an mHEMT (metamorphic HEMT) process for a millimeter-wave LNA involves evaluating the trade-offs between noise figure, gain, operating frequency, cost, and availability. The pHEMT uses an InGaAs channel grown pseudomorphically (lattice-matched with strain) on a GaAs substrate: the indium content is limited to approximately 15-25% (higher indium increases the lattice mismatch strain beyond what the pseudomorphic layer can tolerate), the channel electron mobility is 5,000-8,000 cm^2/V-s, providing f_T of 80-150 GHz and NF_min of 0.5-1.0 dB at 40 GHz. The mHEMT uses a higher indium content InGaAs channel (40-80% indium) grown on a GaAs substrate through a graded metamorphic buffer layer that gradually transitions the lattice constant: the higher indium content provides higher electron mobility (10,000-15,000 cm^2/V-s) and sheet charge density, achieving f_T of 200-400 GHz and NF_min of 0.3-0.7 dB at 40 GHz (30-50% lower noise than pHEMT at the same frequency). Selection criteria: use pHEMT when the operating frequency is below approximately 40 GHz (where pHEMT noise figure is adequate), cost is important (pHEMT processes are more mature and less expensive), the LNA must be integrated with other circuits (pHEMT is available in multi-function MMIC foundries), and high reliability is required (pHEMT processes are well-characterized for space and military applications). Use mHEMT when the operating frequency is above 40 GHz and the absolute lowest noise figure is required, the application is noise-sensitive (radio astronomy, deep-space receivers, passive mmW imaging), and the extra cost of the mHEMT process is justified by the performance improvement.