July 4, 2026
As 5G FR2, Wi-Fi 7, 77/79 GHz automotive radar, low-Earth-orbit satellites, and K/Ka-band satellite communications all enter active product development, microwave PCBs (1–30 GHz) and mmWave (above 30 GHz) are no longer a niche concern for RF engineers — they're an engineering reality facing most hardware and communications R&D teams in Taiwan. Yet at the spec stage, engineers often get stuck on two questions: which Rogers grade exactly? And is microstrip enough, or do we need CPW? This article walks through 6 microwave material selection criteria, the logic behind major material families, and CPW coplanar waveguide design and manufacturing considerations.
Signal integrity in microwave PCBs is determined by two coupled decisions: the substrate material's electrical stability, and the transmission line's ability to contain fields.
Material decides how fast the signal propagates, how much loss it accumulates, and how much phase drifts – dielectric constant (Dk) controls propagation velocity and impedance, dissipation factor (Df) controls transmission loss, and CTE controls phase stability. Transmission line structure decides whether the field is effectively confined near the trace – microstrip, stripline, and coplanar waveguide (CPW) each have their own frequency ranges and application sweet spots.
The two must be designed together. A premium low‑Dk/Df material paired with the wrong transmission line structure can still introduce more than 3 dB of unnecessary loss in mmWave bands. Conversely, the right CPW choice paired with a Dk‑unstable substrate will see impedance drift with frequency, degrading return loss and beam‑forming accuracy.
The six key material parameters are:
| Parameter | Recommended Value | Physical Meaning | | :--- | :--- | :--- | | Dielectric constant (Dk) | 2–4 (@10 GHz) | Sets propagation velocity and trace width; lower Dk variance means less impedance drift with frequency | | Dissipation factor (Df) | < 0.005 | Main source of dielectric loss; mmWave applications require < 0.002 | | CTE | 10–20 ppm/°C (z‑axis) | Must match copper's CTE to avoid multilayer thermal stress and warpage | | Substrate thickness | 10–20 mil above 10 GHz | Thicker substrates can excite spatial harmonic modes at high frequency | | Loss tangent (tan δ) | 0.0022–0.0095 (10–30 GHz) | Frequency‑dependent dielectric loss characteristic | | Moisture absorption | < 0.1% | Water raises Dk and Df, destabilising high‑frequency performance |
Two engineering details that are frequently overlooked:
Rogers material families dominate microwave PCB selection:
| Series | Dk Range | Structure | Best Suited For | | :--- | :--- | :--- | :--- | | RO3000™ | 3.0–10.2 | Ceramic‑filled PTFE | mmWave, phased array, frequency‑consistent applications | | RO4000™ | 2.55–6.15 | Hydrocarbon ceramic | Mid‑to‑high frequency, process‑friendly, moderate cost | | Rogers TMM® | 3–13 | Advanced thermal microwave laminate | High mechanical strength, chemical stability | | Pure PTFE | 2.6–2.7 | Pure Teflon | Lowest Dk/Df, but difficult to metallise |
The most common selection tradeoff is RO3003 (Dk 3, Df 0.0013) vs RO4350B (Dk 3.66, Df 0.0031): RO3003 has lower loss and is preferred for 24/28 GHz 5G and 77 GHz radar; RO4350B is more process‑friendly and cost‑effective, well‑suited for mainstream 6–15 GHz applications. RO3006/RO3010 (higher Dk) target small‑form‑factor antenna designs.
Cost‑sensitive projects often use a hybrid stackup strategy – Rogers on RF signal layers, FR‑4 on the rest, balancing performance against cost. Taiwan‑based fabricators have mature experience with such hybrid laminations, but designers must account for the CTE mismatch between Rogers and FR‑4, match lamination temperatures carefully, and select appropriate bonding films (FEP, ceramic‑filled PTFE, or LCP) at the stackup stage.
The right transmission line structure can shave 1–2 dB of loss without changing materials. CPW is increasingly replacing microstrip in mmWave applications for four reasons:
CPW is not a universal solution. It requires symmetric grounding, precise gap control (gap width affects impedance more than trace width), and ground‑tie via planning to suppress surface wave modes. The extra layout effort is the main reason many RF engineers use CPW only occasionally rather than as their default.
CPW offers several manufacturing benefits as well:
But PTFE substrates (including RO3003 and other high‑end Rogers materials) carry a hidden process challenge: PTFE's chemical inertness makes copper adhesion difficult. The industry‑standard approach is to apply plasma treatment or sodium etch before plating or lamination – plasma activates the PTFE surface, while sodium etch chemically roughens it. Thicker PTFE boards typically require sodium etch to achieve sufficient peel strength. This is why PTFE microwave boards must be reviewed for process capability at the DFM stage rather than discovered through failed prototypes.
Microwave PCBs are never a “pick one Rogers grade and you're done” problem. Dk/Df variation with frequency, CTE mismatch across layers, PTFE metallisation treatment, CPW vs microstrip selection, CPW ground‑tie via planning – all of these variables are coupled and must be decided together at the spec and layout stage.
In our DFM reviews for microwave and RF boards, three issues come up most often:
eCloud has accumulated extensive prototyping experience with hybrid stackups, PTFE boards, and CPW structures across RF/microwave, mmWave, satellite communications, and automotive radar applications. Planning a mmWave or RF project? Talk to our engineering team during layout – get integrated material, stackup, impedance, and DFM feedback early to avoid unexpected loss or reliability problems between prototype and production.
