PEEK property data, and what each number actually changes
Every datasheet gives you the value. This one gives you the consequence: what the row changes on a part you have to specify, quote or inspect.
How to read this
Values are typical published figures from manufacturer datasheets for machining grade stock at 23 °C unless the row says otherwise. They are ranges because grades differ between makers and because a machined part is not a moulded test bar. Use them to shortlist a grade and to sanity check a quote. For a value you will design against, take it from the datasheet of the exact grade and lot you will buy. Where the numbers come from and how mistakes get fixed is set out in our editorial standards.
Unfilled
The tough one. Steam, autoclave cycles, chemicals, impact. Moves the most with temperature.
GF30
30% glass fibre. Stiff and creep resistant while hot, still insulating, abrasive to a mating shaft.
CF30
30% carbon fibre. Stiffest, most dimensionally stable, conducts heat, dissipates static, cannot insulate.
Bearing
Carbon fibre with PTFE and graphite. Lowest friction, best dry wear, weakest of the four.
Mechanical
| Property | Unit | Unfilled | GF30 | CF30 | Bearing |
|---|---|---|---|---|---|
| DensitySets part weight and the price of the blank, because stock is sold by the kilo. | g/cm³ | 1.30 - 1.32 | 1.49 - 1.51 | 1.40 - 1.44 | 1.44 - 1.46 |
| Tensile strengthLoad before yield. Filled grades win here, and lose on toughness. | MPa | 90 - 100 | 150 - 170 | 130 - 140 | 90 - 110 |
| Tensile modulusHow much the part deflects. Unfilled PEEK is about one fiftieth as stiff as steel, which is the number most metal designs get wrong. | GPa | 3.5 - 4.0 | 9 - 11 | 12 - 14 | 7 - 9 |
| Elongation at breakWarning for snap fits and impact. A filled grade that reads 2% will crack where unfilled bends. | % | 20 - 45 | 2 - 3 | 1.5 - 2.5 | 2 - 4 |
| Izod impact, notchedSurvival of a drop or a knock. Bearing blends are the most brittle of the family. | kJ/m² | 5 - 8 | 7 - 9 | 4 - 6 | 3 - 5 |
| Coefficient of friction, dry against steelThe one row that decides a bushing. Glass filled is the worst choice against a metal shaft. | µ | 0.30 - 0.40 | 0.35 - 0.45 | 0.20 - 0.30 | 0.10 - 0.20 |
Thermal
| Property | Unit | Unfilled | GF30 | CF30 | Bearing |
|---|---|---|---|---|---|
| Glass transition, TgThe number to design around. Above it the material softens, expansion jumps and creep becomes the limit rather than strength. | °C | 143 | 143 | 143 | 143 |
| Melting pointFar above any annealing soak, which is why stress relief does not distort the part. | °C | 343 | 343 | 343 | 343 |
| Continuous service temperatureSustained, unloaded. It is not the temperature at which the datasheet strength still applies. | °C | 250 | 250 - 260 | 250 - 260 | 250 |
| Heat deflection temperatureThe honest gap between unfilled and filled. A loaded part hot is a filled grade job. | °C at 1.8 MPa | 150 - 160 | 310 - 320 | 310 - 320 | 260 - 290 |
| Thermal expansion, below TgSteel is about 12 and aluminium about 23. This single row explains most fits that close up hot. | µm/m·K | 45 - 50 | 20 - 24 | 14 - 18 | 18 - 24 |
| Thermal expansion, above TgRoughly triples for unfilled PEEK. Any fit that has to hold above 143 °C needs a filled grade or a deliberate cold clearance. | µm/m·K | 110 - 140 | 30 - 50 | 25 - 40 | 35 - 55 |
| Thermal conductivityAbout one five hundredth of aluminium for unfilled. It is why PEEK traps machining heat and why carbon filled holds tighter tolerances. | W/m·K | 0.25 | 0.40 - 0.45 | 0.85 - 0.95 | 0.6 - 0.9 |
Electrical and environmental
| Property | Unit | Unfilled | GF30 | CF30 | Bearing |
|---|---|---|---|---|---|
| Volume resistivityThe grade decision for insulators and for static dissipation, in one row. Carbon filled cannot insulate. | Ω·cm | > 10¹⁵ | > 10¹⁵ | 10³ - 10⁶ | 10³ - 10⁶ |
| Dielectric strengthAsk for the value at your wall thickness, it falls as the section grows. | kV/mm | 19 - 23 | 16 - 20 | - | - |
| FlammabilityTypical at common wall thickness. The rating is thickness dependent, so quote it with the section. | rating | V-0 | V-0 | V-0 | V-0 |
| Limiting oxygen indexWhy PEEK turns up in rail, aerospace and enclosed equipment. | % | 35 | 38 - 45 | 38 - 45 | 38 - 45 |
| Water absorptionLow enough that parts hold size in humid air, unlike nylon. Saturation is about 0.5% for unfilled. | % at 24 h | 0.1 | 0.06 - 0.1 | 0.06 - 0.1 | 0.06 - 0.1 |
What it means in the shop
| Property | Unit | Unfilled | GF30 | CF30 | Bearing |
|---|---|---|---|---|---|
| Tolerance a shop quotesOn small annealed features. Long features and thin walls move to the looser end. | mm | ±0.05 | ±0.05 | ±0.025 | ±0.05 |
| Surface finishFilled grades show fibre ends at the surface, so a sealing face usually wants unfilled. | Ra µm | 0.4 - 0.8 | 0.8 - 1.6 | 0.8 - 1.6 | 0.8 - 1.6 |
| Tool wear, relative to unfilledPriced into the quote. Fibre filled work carries fresh tooling and often a separate tool set. | - | 1x | 3x | 3x | 2x |
| Stress relief on thin walls and tight boresRough, anneal, finish. Under about 3 mm wall or tighter than ±0.05 mm, assume it is in the quote. | - | yes | yes | yes | yes |
Picking a grade from the rows above
Four questions settle almost every case, in this order.
What a shop can hold as a feature gets deeper
Five rules that come out of this table
The rows above are worth more as decisions than as data.
- Stiffness, not strengthUnfilled PEEK is strong for a plastic and about one fiftieth as stiff as steel. A metal section carried straight across will deflect.
- Tg 143 °C is the real ceilingContinuous service of 250 °C is unloaded. Above Tg, expansion triples and creep decides the part.
- Expansion breaks fits, not strengthPEEK moves about four times more than steel per degree. Calculate the fit at the service temperature, both directions.
- Carbon conductsCF30 is the answer for static dissipation and the wrong answer for an insulator. Check the resistivity row before the strength rows.
- Filled grades cost twiceOnce in stock price, once in tool wear. If the filler does nothing for the part, unfilled saves in both places.
Put the numbers to work
Two calculators use the rows above: expansion and fit at temperature, and the annealing cycle for a wall thickness. Both run in your browser and send nothing anywhere.