DARPA
How DARPA Reshaped RF and Microwave Engineering
DARPA's operating model is unusual: a flat organization with no permanent laboratories, built around six technical offices and a rotating cadre of program managers who each own a portfolio for three to five years. A manager publishes a set of hard, falsifiable metrics, often a 5 to 10 times leap over the current best demonstrated number, then funds whoever can credibly chase them. For RF, the offices that matter most are the Microsystems Technology Office (MTO) and the Strategic Technology Office (STO). MTO has historically owned the semiconductor and component programs that produce the transistors, data converters, and integrated circuits everything else depends on.
The clearest example is gallium nitride. In the early 2000s GaAs dominated microwave power amplifiers, but its power density capped what a radar transmit module could deliver per square millimeter of die. The Wide Bandgap Semiconductors RF program (WBGS-RF) drove GaN-on-SiC HEMT processes through three contractual phases, ending with a qualified 0.25 micron and 0.14 micron foundry technology hitting 5 to 8 W per millimeter of gate periphery at X-band. That density, paired with a breakdown field roughly ten times silicon's, is why a single GaN MMIC can now replace a stack of GaAs parts in a phased-array transmit/receive module.
Beyond devices, DARPA has pushed system architectures: digital and hybrid beamforming arrays, wideband software-defined receivers fed by multi-gigasample data converters, and heterogeneous integration that bonds GaN, GaAs, InP, and CMOS on a common substrate (the DAHI and T-MUSIC programs). Each of these is now visible in commercial 5G and satellite hardware, a transition that typically lags the original DARPA demonstration by 8 to 12 years.
From TRL Gate to Production Transition
DARPA programs are scored against Technology Readiness Levels rather than schedule alone. The agency deliberately lives in the valley between basic research and a fielded product, usually TRL 2 through TRL 5, and the explicit goal of a successful program is a transition partner (a service lab or a prime contractor) willing to carry the technology from TRL 6 to deployment. For an RF supplier, that means deliverables are measured data against the program metric, for example output power, power-added efficiency (PAE), noise figure, and reliability from accelerated high-temperature operating-life testing, not narrative reports.
Governing Metrics for an RF Program
Pdensity = Pout / Wgate (W/mm); GaN-on-SiC ≈ 5 to 8 W/mm at X-band vs. GaAs ≈ 1 to 1.5 W/mm
Power-added efficiency (a common DARPA metric):
PAE = (Pout − Pin) / PDC × 100%
Output power in dBm:
Pout(dBm) = 10 log10(PmW), so 10 W ≈ 40 dBm
Example WBGS-RF-class target: ≥ 10 W output at 30 GHz with PAE > 40% from a single GaN MMIC. With high stage gain (Pin small, so PAE ≈ drain efficiency), 10 W out at 40% implies DC draw PDC ≈ 10/0.40 ≈ 25 W, dissipating ≈ 15 W as heat (PDC − Pout), which sets the thermal and packaging challenge.
DARPA Versus Adjacent Defense Research Paths
| Vehicle | Typical horizon | TRL band | Funding scale | RF relevance |
|---|---|---|---|---|
| DARPA BAA program | 3 to 5 years | 2 to 5 | $10M to $100M+ | GaN, beamforming, wideband converters |
| SBIR / STTR Phase 1 | ~6 months | 2 to 3 | $150K to $250K | Feasibility of a single device or block |
| SBIR Phase 2 | ~2 years | 3 to 5 | $1M to $2M | Prototype MMIC or module |
| Service S&T (AFRL, ARL, NRL) | 3 to 7 years | 3 to 6 | Varies | Maturation and transition partner |
| Program of Record | 5 to 15 years | 6 to 9 | $100M to $billions | Production radar / EW hardware |
Frequently Asked Questions
Which DARPA programs created today's GaN RF technology?
The Wide Bandgap Semiconductors program (WBGS-RF), roughly 2003 to 2010, is the direct ancestor of defense-grade GaN. It matured GaN-on-SiC HEMT processes through three phases to a manufacturable 0.25 and 0.14 micron foundry technology delivering 5 to 8 W/mm at X-band, about 5 times GaAs density. Follow-ons such as NEXT, DAHI, T-MUSIC, and T3 pushed heterogeneous integration and W-band devices. That foundry maturity is why GaN MMIC power amplifiers are now standard in radar T/R modules and 5G mmWave front ends.
What is a DARPA Technology Readiness Level requirement on RF contracts?
DARPA scores programs on Technology Readiness Level (TRL), a 1 to 9 maturity scale, and typically works in the TRL 2 to 5 band. A Phase 1 deliverable might be a TRL 3 device demonstration; a Phase 3 transition to a program of record expects TRL 6 or higher. Hitting a gate means measured S-parameters, output power, PAE, and reliability data (for example MTTF from accelerated GaN HTOL testing) against the published program metrics.
How does a small RF company win DARPA funding?
Most RF work flows through Broad Agency Announcements (BAAs) and the SBIR/STTR programs. A company answers a program manager's quantitative metrics with a white paper, then a full proposal if invited. SBIR Phase 1 (about $150K to $250K over 6 months) funds feasibility, Phase 2 (about $1M to $2M over 2 years) funds a prototype, and Phase 3 is the unfunded production transition. Winning proposals tie a hard RF number, such as 10 W at 30 GHz with PAE above 40 percent, to a genuine leap past the state of the art.