Microwave PCB Materials: Choosing a Substrate Above 1 GHz

Where the Material Becomes the Circuit

Microwave boards cover roughly 1 to 30 GHz, and in that range the laminate is not a passive carrier. Its dielectric constant sets the impedance and the propagation velocity, its dissipation factor converts signal into heat, and its thickness tolerance shifts the impedance across a panel. A substrate that behaved acceptably at 500 MHz can be unusable at 10 GHz, and the difference has nothing to do with the layout.

Material selection is therefore the first design decision, and it is the one that most often determines whether a project succeeds on budget or becomes an expensive exercise in chasing loss.

What a Microwave Substrate Has to Deliver

• Low dissipation factor (Df), which controls insertion loss and therefore the noise figure and the link margin.
• Stable dielectric constant (Dk), across frequency, temperature and production lots, because impedance and delay both depend on it.
• Tight thickness tolerance, since a microstrip or stripline impedance is a direct function of the dielectric height.
• Low copper roughness, because at these frequencies the current travels near the surface and a rough foil lengthens its path.
• Thermal and mechanical stability, for power amplifiers that heat up and for assemblies that have to survive temperature cycling.

Frequency amplifies the sensitivity of every one of those items. A Dk variation that shifts the impedance by two percent at 2 GHz may shift it by five percent at 10 GHz, which is the difference between passing and failing a return loss specification.

The Material Families

• PTFE based laminates. The lowest loss family, with Df values an order of magnitude below FR-4 and excellent dielectric stability. They are also the most difficult to process: PTFE has low surface energy and requires plasma activation before plating, it behaves differently under a drill, and it needs lower lamination pressure.
• Ceramic filled PTFE. Higher dielectric constant by design, which allows smaller circuits for a given electrical length. Useful where the wavelength has to fit a compact package, at the cost of some loss and a tolerance on the Dk.
• Ceramic filled hydrocarbon laminates. A middle ground with better loss than FR-4, easier processing than PTFE and good dimensional stability. Frequently the best cost-to-performance choice for moderate frequencies.
• Standard FR-4. Usable on the digital and power sections of the same board, and on RF paths only at the low end of the range.

The practical decision usually lands on a hybrid stack: the low loss material only on the layers that carry the microwave signal, and FR-4 or a high Tg laminate for the rest. That delivers most of the performance at a fraction of the material cost, and it is the arrangement most commercial products use.

Specifications Beyond Dk and Df

Three secondary parameters are frequently overlooked and frequently decisive.

Copper foil type. Standard electrodeposited foil has a rough surface treatment that improves adhesion but adds loss. Low profile and rolled foils reduce the loss at microwave frequencies, and on a long run they can account for a significant share of the total attenuation.

Thickness tolerance and its distribution across the panel. A tight tolerance on the datasheet and a wide variation across the panel are not the same thing. The coupon measures one location; the board has to work everywhere.

Coefficient of thermal expansion and Tg. Relevant where the board carries power devices that heat, and where the assembly sees thermal cycling. A low loss material with poor dimensional stability can pass an RF test at room temperature and drift in service.

Design Implications

• Compute the geometry from the material. The trace width follows from the Dk and the dielectric height at the operating frequency, not from a table borrowed from another project.
• Use the material datasheet at the right frequency. Dielectric constant varies with frequency, and the difference between the 1 MHz value and the 10 GHz value is significant for some materials.
• Minimise transitions. Every connector, filter pad and via in the microwave path contributes to return loss. Where a layer change is unavoidable, back drill the stub or use a blind via.
• Keep the reference plane continuous and provide a return via next to every signal via.
• Fence the transmission lines with vias at a spacing of a tenth of a wavelength or less, so the line behaves as a shielded structure rather than a radiator.
• Plan the thermal path for any power amplifier on the board, since the loss that the material absorbs becomes heat that has to leave. The approach is the same as in general thermal management work, with the added constraint that the heat path must not disturb the RF geometry.
• Design the connector transition deliberately, because it is the most common source of return loss on a microwave board.

Fabrication Requirements

• Plasma treatment before plating on PTFE materials. Without it, copper adhesion is unreliable and the failure appears later as a lifted pad or a delaminated trace.
• Dedicated drilling parameters, because these materials smear and burr differently from FR-4.
• Low pressure lamination, with a press profile developed for the specific laminate rather than borrowed from a standard stack.
• Registration control to keep the RF traces aligned with their reference planes across the panel.
• Low profile copper where the loss budget demands it, which is a specification on the drawing rather than a shop preference.
• Impedance tolerance typically of plus or minus five percent, rather than the ten percent that ordinary controlled impedance work accepts.

Verification

Two measurements confirm that the delivered board matches the design intent. An impedance coupon measured on the finished panel verifies the dielectric height and the line width against the assumptions used in the simulation, and the method described under TDR impedance testing is the standard one. A coupon or test structure measured for insertion loss establishes whether the material performed as the datasheet suggested at the operating frequency. Both sets of data should ship with the boards.

The other verification that matters is on the assembled product: gain, noise figure and return loss measured on the finished unit. A board can meet its impedance target and still underperform if the assembly introduced a solder volume change at a filter or a contaminated surface. That is why the RF measurement belongs in the test plan rather than being replaced by a functional check, and the considerations overlap with those in any RF oriented PCB assembly program.

Cost

• Standard FR-4: the baseline, usable only at the low end of the frequency range.
• Ceramic filled hydrocarbon: a moderate premium, often the best balance for commercial RF work.
• PTFE based laminates: typically two to five times the cost of FR-4, with the lowest loss grades at the top of the range.
• Hybrid stacks: the cost of the low loss material applied only to the layers that need it, plus the process complexity of bonding dissimilar materials.

Three other items add cost on a microwave board: the tighter impedance tolerance and the coupon measurement it requires, any blind or buried via construction, and a hybrid lamination. The largest lever is the stackup itself, since choosing a low loss material for layers that carry no microwave signal is pure cost with no benefit. Comparing two stackup options against the same performance target is the practical way to make that decision, and a formal custom PCB pricing request covering both is the cheapest way to obtain the comparison. The wider process requirements for this class of board are described under PCB manufacturing.

Typical Applications

5G radio units and small cells, radar and automotive sensors, satellite communication terminals, microwave links for backhaul, aerospace systems, and RF test instrumentation. The shared requirement is a low loss path with a predictable impedance over a defined temperature range.

Questions to Ask a Supplier

Five questions separate a capable microwave fabricator from a general one: which low loss laminates are held in stock and in which thicknesses, whether the press and drill profiles for those materials are already developed, whether plasma treatment is done in house, whether impedance and insertion loss are measured rather than assumed, and what the experience is with the specific construction requested. The material is only half of the outcome; the process data behind it determines whether the board actually performs at the frequency the design targets.

FAQ

What frequency range does a microwave PCB cover? Roughly 1 to 30 GHz, with the material requirement becoming critical above about 6 GHz.

Which material has the lowest loss? PTFE based laminates, with dissipation factors an order of magnitude below FR-4.

Can FR-4 be used at microwave frequencies? Only at the lower end of the range or for non critical paths. Above that the insertion loss becomes the limiting factor.

Why does copper roughness matter so much? Because the skin effect confines the current near the surface, so a rough foil lengthens the current path and adds attenuation.

Is a hybrid stack acceptable? Yes, and it is the most common commercial arrangement, provided the two materials are compatible and the press profile is developed for both.

Summary

Above roughly 1 GHz the laminate stops being a passive carrier and starts determining the electrical result. The controlling parameters are the dissipation factor, which sets insertion loss, the dielectric constant and its stability, which set impedance and delay, the thickness tolerance, which limits how tightly impedance can be held, and copper roughness, which adds loss through the skin effect. PTFE laminates provide the lowest loss and the hardest processing, ceramic filled hydrocarbons the best balance, and FR-4 only covers the low end. Choose a hybrid stack where the performance allows it, compute the geometry from the material data at the operating frequency, verify with measured impedance and insertion loss coupons, and confirm the finished product with an RF measurement rather than a functional check.