How to Choose PCB Materials: FR-4, Rogers, Polyimide, Metal Core

Values are typical ranges for each family, not datasheet limits for a specific product. Dk and Df shift with frequency and resin content, so use the exact laminate datasheet once you have a shortlist.

Start from the hardest requirement, because it usually removes most options. Thermal and reflow stress narrow FR-4 grades; frequency and loss budget decide whether you leave FR-4 at all; motion decides flex construction; heat flux decides whether you need a metal base.

Standard FR-4 suits most digital and industrial boards below roughly 1 GHz on short routes. High-Tg FR-4 improves thermal stability through lead-free reflow and dense assemblies.

The numbers behind the names: Tg (glass transition temperature) marks where the resin softens and expansion accelerates—standard FR-4 sits around 130–140 °C, high-Tg grades at 170–180 °C. Td (decomposition temperature) and Z-axis CTE matter just as much for boards that see multiple lead-free reflow cycles: each pass above 245 °C stresses via barrels, and high-layer-count boards with deep holes are where cheap laminate fails first. For six layers and up, or any board facing double-sided reflow plus rework, high-Tg is inexpensive insurance.

FR-4 Dk varies by lot and frequency—fine for many designs, inadequate when phase or filter tuning is sensitive. Do not assume FR-4 for long high-speed channels without stackup validation. CAF (conductive anodic filament) resistance is the other quiet differentiator between laminate grades—relevant for high-voltage spacing and humid environments, and worth a line in the RFQ when either applies.

Use controlled-Dk laminates when loss budgets, filter tuning, or long high-speed links rule out standard FR-4. Rogers hydrocarbon-ceramic materials (the RO4000 family) process much like FR-4 and hold Dk to tight tolerance—RO4350B vs. RO4003C compares the two most common grades; PTFE-based laminates go lower in loss still but need specialized hole preparation (plasma or sodium etch) before plating, which shows up in cost and lead time.

Hybrid builds place RF layers on Rogers and digital sections on FR-4 to control cost—a two-layer Rogers cap over an FR-4 core is the workhorse construction for radar sensors and RF front ends. Hybrid lamination needs explicit press profiles and impedance verification on both material sets—call out hybrid intent in the RFQ.

Dynamic flex favors adhesiveless polyimide and thin copper. Static fold-to-install can tolerate adhesive-based constructions at lower cost. Material choice follows bend radius and cycle count—not only layer count.

Copper type matters as much as the film: rolled-annealed (RA) copper’s grain structure survives repeated flexing where standard electrodeposited (ED) foil cracks early. Specify RA copper for any dynamic application; ED is acceptable—and cheaper—for static bends. Coverlay, stiffener materials, and adhesive systems complete the construction and are covered in depth in our rigid-flex design guide.

When FR-4 with via farms cannot keep junction temperatures in spec, metal-core PCBs move heat through a conductive dielectric into an aluminum or copper base. Dielectric thermal conductivity and thickness are quoting inputs, not afterthoughts.

Typical MCPCB dielectrics conduct 1–3 W/m·K versus roughly 0.3 for FR-4—an order of magnitude that turns LED junction temperatures around. Aluminum base (usually 5052 at about 138 W/m·K or 6061 at about 167 W/m·K, 1.0–1.6 mm thick) covers most lighting and power work; copper base, at roughly 390–400 W/m·K, more than doubles base conductivity for concentrated hot spots at a higher material cost. Standard metal core is single-layer circuitry; double-sided and multilayer MCPCB constructions exist but add insulated through-hole complexity that deserves a PCB capabilities conversation before you design around it.

State temperature rise targets, bend requirements, frequency range, and compliance needs (UL, RoHS evidence). Material surprises discovered in CAM delay every program.

If you are unsure, describe the environment and let review propose options—still faster than specifying the wrong laminate and re-spinning.