Brand names are the least useful index to a high frequency PCB laminate guide, because the same brand sells four materials that behave nothing alike. Rogers alone ships a thermoset hydrocarbon, a ceramic-filled PTFE without woven glass, a woven PTFE and a high-Dk ceramic-filled PTFE, and their press temperatures, desmear chemistry and CTE differ by factors of two to ten. The useful organisation is by polymer system, with the brands mapped onto it: four systems cover essentially every laminate in use, and once a design is placed in one of the four, the brand question becomes a question about grades rather than about manufacturers.
Why Brand Is the Wrong Index
A laminate is defined by its polymer system, its reinforcement and its filler, in that order. Everything a fabricator cares about — press temperature, desmear chemistry, drill behaviour, CTE, copper adhesion — follows from those three, and none of them follows from the logo on the datasheet.
Rogers is the clearest example. The RO4000 series is a thermoset hydrocarbon with ceramic filler and woven glass, pressed at 190–200 °C and desmeared with permanganate. The RO3000 series is a ceramic-filled PTFE with no woven glass, pressed at 260–300 °C with an FEP bondply, and immune to permanganate. Both are sold by the same company under names that suggest they are siblings, and a fabricator who treats them the same way will scrap the second one. Organising by brand makes that error easy; organising by system makes it impossible.
The four systems below cover every laminate in current use for RF and high-speed work, including the Chinese grades that now appear on the same shortlists.
The Four Material Systems
| System | What it is | Typical Dk / Df | Process signature | Where it wins |
|---|---|---|---|---|
| Ceramic-filled PTFE, no woven glass | PTFE with ceramic powder, microfibreglass, no fabric | 2.20–10.2 / 0.0009–0.0027 | 260–300 °C, sodium-naphthalene or plasma, soft and cold-flowing | mmWave, phase-critical, lowest loss |
| PTFE with woven glass | PTFE plus glass fabric | 2.20–3.50 / 0.0009–0.0018 | Same press range, better dimensional stability, weave effect present | Large panels, multilayer RF, cost-sensitive PTFE |
| Hydrocarbon / ceramic thermoset | Thermoset resin with ceramic filler, woven glass | 3.38–3.66 / 0.0027–0.0037 | 190–200 °C, permanganate desmear, standard multilayer | 5G infrastructure, automotive radar, mixed FR-4 stacks |
| Modified epoxy / PPO high-speed | Digital laminates with engineered resin, not RF materials | 3.3–3.9 / 0.002–0.012 | 170–190 °C, standard FR-4 line, high Tg | Backplanes, optical modules, high-layer digital |
Two boundaries are worth drawing explicitly. The first is between systems one and two: the absence of woven glass removes the weave effect, which is why phase-critical radar work uses non-woven grades, and it also removes most of the mechanical stiffness, which is why large thin panels are harder. The second is between systems three and four: a hydrocarbon thermoset is an RF material that presses like FR-4, while a high-speed epoxy is a digital material that happens to have low loss. Confusing the two produces designs that specify a 0.002 digital grade for an antenna and then discover that its Dk tolerance is three times wider than the RF grade it replaced.
Rogers: Four Families That Behave Differently
Rogers is where the four-system map is easiest to see, because one company sells into three of the four.
| Family | System | Representative grades | Dk | Df | Note |
|---|---|---|---|---|---|
| RO4000 | Hydrocarbon thermoset | RO4003C, RO4350B, RO4835 | 3.38 / 3.48 / 3.48 | 0.0027 / 0.0037 / 0.0037 | 4835 matches 4350B electrically, adds oxidation resistance |
| RO3000 | Ceramic-filled PTFE, non-woven | RO3003, RO3035, RO3006, RO3010, RO3210 | 3.00 / 3.50 / 6.15 / 10.2 / 10.2 | 0.0010 / 0.0015 / 0.0020 / 0.0022 / 0.0027 | 3210 is the one with woven glass |
| RT/duroid | PTFE, microfibreglass | 5880, 5880LZ | 2.20 / 1.96 | 0.0009 / 0.0019 | 5880 is the mmWave reference; LZ is the low-Dk foam variant |
| 6000 / TC | Ceramic-filled PTFE, specialised | 6010.2LM, 6035HTC, TC350 | 10.2 / 3.50 / 3.50 | 0.0023 / 0.0013 / 0.0019 | 6035HTC trades loss for 1.44 W/m·K thermal conductivity |
| 5000 | Hydrocarbon, FR-4 process compatible | 5000-series grades | varies | low-loss class | Designed to run on an FR-4 line — not a PTFE |
Three facts in that table are misread constantly. RO4003C and RO4350B are hydrocarbon thermosets, not PTFE: they press at 190–200 °C, they tolerate permanganate, and their z-axis CTE is 32–46 ppm/°C rather than 237. RT/duroid 5880 is PTFE with microfibreglass — no woven fabric, which is why it has no weave effect and why its z-axis CTE is 237 ppm/°C. And 6035HTC is not a general-purpose laminate: it exists because 1.44 W/m·K is worth a higher Df in a power amplifier.
Two naming traps are worth recording here because they appear in circulating articles. RO3002 and RO3015 are not products; the 3000 series runs 3003, 3035, 3006, 3010 and 3210. “3001” is a bonding film, not a laminate. The series detail is set out in the RO3000 model guide and the RO4000 series overview.
AGC and Taconic: Woven PTFE as the Mainstream
Since AGC integrated Taconic, the Taconic grades have become the middle of the market: PTFE-based, mostly with woven glass, priced below the Rogers PTFE lines and available in large panels. The brand remains in use commercially, which is why datasheets and articles still say Taconic rather than AGC.
| Grade | Dk | Df | Character |
|---|---|---|---|
| TLY-5 | 2.20 ±0.02 | 0.0009 | Non-woven PTFE, the direct alternative to 5880; z CTE 217 ppm/°C |
| TLX-0 / 9 / 8 / 7 / 6 | 2.45 / 2.50 / 2.55 / 2.60 / 2.65 | about 0.0019 | Woven PTFE, tight Dk steps for antenna work |
| RF-30 | 3.00 ±0.04 | 0.0014 at 1.9 GHz | Workhorse PTFE for base-station antennas |
| RF-35 | 3.50 | 0.0018 | Tg above 315 °C, multilayer friendly |
| RF-10 | 10.2 ±0.30 | 0.0025 | High-Dk for size-constrained antennas |
| NF-30 | 3.00 class | low loss | Non-woven, aimed at 77–79 GHz automotive radar |
The Df figure for RF-30 is worth pausing on: it is quoted at 1.9 GHz rather than at 10 GHz, which is a lower frequency than most comparisons assume. Comparing it with a grade measured at 10 GHz without noting the frequency will make it look better than it is — one of the common traps in this market, and the reason the Taconic performance summary quotes frequencies explicitly.
Panasonic and Isola: High-Speed Digital, Not High Frequency
The Japanese and American high-speed lines are frequently listed in “high frequency laminate” guides and they are not high-frequency materials in the RF sense. They are digital laminates with engineered resin, and they are used in RF products for the digital half of the board.
Panasonic’s MEGTRON series covers approximately 0.004 at 10 GHz for MEGTRON 6 and the 0.002 class and below for MEGTRON 7 and 8, with high Tg and CAF resistance as the accompanying claims. Isola’s Astra MT77 and I-Tera lines occupy the same territory, with MT77 aimed specifically at microwave-adjacent work. The distinction that matters is what each is qualified for: a digital grade is qualified on loss, CAF, thermal reliability and drillability at high layer counts, while an RF grade is qualified on Dk tolerance, Df and phase stability. A digital grade can be excellent in a driver circuit and a poor choice for an antenna feed, because nobody specified its Dk tolerance to ±0.02.
Where they are the right answer is the case described in the optical module model summary: 56 GHz Nyquist channels at centimetre length, where 0.55 dB/in is enough and the mechanical requirements — stability, HDI, CAF — dominate.
Chinese Suppliers: What Is Verifiable
The domestic grades are now a genuine part of the map rather than a cost-only option, and the honest way to describe them is by system and by published parameter, not by claims of equivalence to a specific foreign grade.
Shengyi covers the widest range: hydrocarbon-based RF grades of the S7136H type for base-station work, ultra-low-loss digital grades of the S7439 and S8 class, and PTFE-based grades for millimetre-wave. ZYTech supplies PTFE-based and hydrocarbon-based grades (ZYF and ZYC lines) aimed at base-station antennas and filters. Wingline covers PTFE, hydrocarbon and PPO-based composites, with thermal conductivity as a stated differentiator for radar and power work. Wazam and Huazheng compete in the same segments, and the Taiwanese suppliers ITEQ and TUC serve the data-centre and 5G equipment markets with high-speed lines of the IT and TU families.
Upstream, the constraint has moved from resin to reinforcement: low-Dk glass fabrics, quartz fabric and HVLP copper foil are now the materials that gate the next performance step, and domestic suppliers in that chain are the reason the finished grades have improved. None of this makes a grade equivalent to another; it means the comparison can be done on published Dk, Df, CTE and press window rather than on origin.
The Comparison That Actually Matters: Loss Across the Four Systems
Ranking by Df alone understates the differences because loss also depends on Dk, and the practical figure is dielectric loss per unit length at the frequency of use. The table applies αd ≈ 2.6·f·√Dk·Df to representative grades from each system.
| Grade | System | Dk | Df | dB/in at 10 GHz | dB/in at 77 GHz |
|---|---|---|---|---|---|
| RO4350B | Hydrocarbon thermoset | 3.48 | 0.0037 | 0.18 | 1.39 |
| RO4003C | Hydrocarbon thermoset | 3.38 | 0.0027 | 0.13 | 1.01 |
| TC350 | Ceramic-filled PTFE | 3.50 | 0.0019 | 0.093 | 0.72 |
| RF-30 | Woven PTFE | 3.00 | 0.0014 | 0.063 | 0.49 |
| RO3010 | Ceramic-filled PTFE | 10.2 | 0.0022 | 0.19 | 1.48 |
| RO3003 | Ceramic-filled PTFE | 3.00 | 0.0010 | 0.045 | 0.35 |
| RT/duroid 5880 | PTFE, microfibreglass | 2.20 | 0.0009 | 0.035 | 0.27 |
The RO3010 row is the one that surprises people: Df 0.0022 looks better than 4350B’s 0.0037, yet at 10 GHz it is lossier, because Dk 10.2 carries the square-root term to 3.2 instead of 1.87. High Dk buys area, not loss. Conversely RF-30 at 0.0014 and Dk 3.00 reaches 0.063 dB/in, better than TC350 despite a nominally similar Dk, which is what a lower Df does.
Process Compatibility: The Table That Decides a Factory
A material is only usable if the fabricator can press it, drill it, clean it and plate it. The four systems divide sharply here, and the division is the practical content of this guide.
| Process step | Ceramic PTFE | Woven PTFE | Hydrocarbon | High-speed epoxy |
|---|---|---|---|---|
| Press temperature | 260–300 °C | 260–300 °C | 190–200 °C | 170–190 °C |
| Bonding medium | FEP film or low-flow prepreg | FEP or prepreg | Its own prepreg | Its own prepreg |
| Desmear | Sodium-naphthalene or plasma | Sodium-naphthalene or plasma | Permanganate OK | Permanganate OK |
| Laser microvia | UV, narrow window | UV preferred | CO2 with conformal mask | CO2 standard |
| Mechanical drilling | Smear risk, sharp bits, low aspect | Better than non-woven | Standard | Standard |
| Multilayer | Possible, cold-flow control needed | The usual PTFE multilayer route | Straightforward | Designed for it |
The cold-flow issue is specific to the first two columns: PTFE thins about 4 % under press, which takes a 50 Ω line to about 48.8 Ω with no change to the artwork, and it is invisible in incoming inspection. Any PTFE quote that does not mention a pressed-thickness coupon is a quote for a board that will measure differently from the drawing.
Selection by Application
| Application | System that usually wins | Grade examples | Reason |
|---|---|---|---|
| 77 GHz automotive radar | Ceramic PTFE, non-woven | 5880, RO3003, TLY-5, NF-30 | No weave effect, low Df, stable Dk |
| 5G massive-MIMO antenna | Hydrocarbon or woven PTFE | RO4350B, RO4003C, RF-30, TLX | Cost per area, multilayer compatibility |
| Power amplifier, high duty | Ceramic PTFE with high k | RO6035HTC, high-k PTFE grades | 1.44 W/m·K cuts the junction rise in half |
| Phased array, large panel | Woven PTFE | TLX series, RF-35 | Dimensional stability at panel scale |
| High-Dk antenna element | Ceramic PTFE, high Dk | RO3010, RO3210, RF-10 | Size reduction of the radiating element |
| Satellite payload electronics | PTFE, either | 5880, TLY-5, RO3003 | Outgassing, thermal vacuum stability |
| Backplane, optical module | High-speed epoxy/PPO | MEGTRON 7/8 class, S7439/S8 class | Loss at 56 GHz plus CAF and stability |
The cross-brand comparisons behind these rows are covered in two further articles: the five-brand comparison for market positioning, and the PTFE-versus-Rogers selection guide for the specific confusion between 4350B, 4003C, 3003 and 5880. The grade-level figures are consolidated in the Rogers specification guide.
What a Second Source Has to Match
Substitution between brands is where most of the avoidable cost in this market sits, in both directions: a domestic grade rejected for lacking a logo, and a foreign grade substituted by one with a different Dk window and no requalification.
Six parameters have to match before a substitute is a substitute: Dk within ±0.05 at the frequency of use, and preferably within ±0.02 for phase work; Df in the same class measured at the same frequency; the same thickness and thickness tolerance, since 50 Ω is a ratio; the same in-plane and z CTE, because the press cycle and the via plating were designed around them; the same press window, or the stack has to be re-qualified; and the same peel strength and UL94 status, since both are pass-or-fail rather than graded.
Everything else — brand, country of origin, price — is a commercial question that follows the six. The reasoning is the same as for any selection on measurable indicators, set out in the key indicators for material selection.
Four Misreadings That Keep Circulating
“RO4350B is PTFE.” It is a hydrocarbon thermoset with ceramic filler and woven glass. It presses at 190–200 °C, desmears with permanganate and has a z-axis CTE of 32 ppm/°C. The PTFE grades with a comparable Dk are the 3000 series and Taconic’s RF-30.
“RO3003 and RO3010 are the same material at different prices.” They are 3.00 and 10.2 in Dk, with 0.0010 and 0.0022 in Df, and their CTE differs in the x/y axes: 17/16 for 3003 against 13/11 for 3010. The high-Dk grade exists to shrink an antenna, and it costs loss to do so.
“5880B.” There is no such grade. The variants are 5880 and 5880LZ, the latter at Dk 1.96 and Df 0.0019. Similarly, RT/duroid 6202, 6302 and 6500 do not exist as current products, and “3001” is a bonding film rather than a laminate.
“4835 is a different dielectric from 4350B.” Electrically they are the same grade at 3.48 and 0.0037, with the same 10/12/32 ppm CTE. The 4835 differs in oxidation resistance — about ten times the 4350B figure in the supplier’s accelerated test — and in thermal conductivity at 0.66 W/m·K. Those are the reasons to choose it, and neither of them is loss.
Conclusion
A high frequency PCB laminate guide is useful only if it organises the market the way a fabricator experiences it, which is by polymer system rather than by brand. Four systems carry the field: ceramic-filled PTFE without woven glass at Df 0.0009–0.0027 and 260–300 °C pressing, woven PTFE at 0.0009–0.0018 with better panel stability, hydrocarbon thermosets at 0.0027–0.0037 that press like FR-4 at 190–200 °C, and high-speed epoxies at 0.002–0.012 that are digital materials with low loss rather than RF materials.
Within that map, brand selection is a question of grades and of process fit. The same company sells a thermoset and a PTFE under adjacent names; the same Df can be reached at two different loss figures because Dk enters the square root; and the property that decides a factory is often the press window rather than the electrical table. Placing a design on the map first, and choosing a logo second, is the whole discipline.
XCE PCB stocks and builds across all four systems — Rogers, Taconic, PTFE and F4B equivalents, hydrocarbon and FR-4 — with 2 to 72 layers, panels to 600×1200 mm, impedance to ±10 %, samples in as fast as 24 hours, inspected to IPC-A-600G and IPC-6015. Mixed-system stacks are pressed in separate cycles with pressed-thickness coupons, because on these materials the drawing is a proposal and the coupon is the record.
Frequently Asked Questions
Is RO4350B a PTFE laminate?
What is the difference between RO3003 and RO3010?
Which Taconic grade replaces RT/duroid 5880?
Why do PTFE boards come out at 48.8 Ω instead of 50 Ω?
Can a high-speed digital laminate be used for an RF line?
Do RO3002, RO3015 or 5880B exist?
Need a Stack Built From Two Different Material Systems?
Send us your Gerber files and requirements for a manufacturability review and a tailored quotation.
sales1@xcepcb.com
+86 180 3818 8537