Ball Valve Seat Materials: Why PTFE Creeps and PEEK Will Not Conform

A stainless steel ball valve is usually bought on three lines of a datasheet: body grade, pressure class, seat material. The first two are printed and comparable. The third is written as one word — "PTFE" — and that word covers four polymers with four failure modes and prices that differ by a factor of three.

The seat is also the only part of the valve that has to be soft. That requirement sets the temperature limit, the torque, the leak rate and the service life, and it is decided before the body is cast.

A Seat Is a Spring, Not a Gasket

A gasket is squeezed once and left alone. A ball valve seat is squeezed, released, rotated against, and squeezed again — every time the valve moves, and continuously while it sits closed under pressure.

That means it has three jobs at once: seal against the ball, maintain the preload that keeps the seal closed as parts move, and carry the ball's thrust without deforming past recovery. A material that is excellent at the first job can be poor at the second, which is why a seat that passed a factory shell test can weep after a season in service.

Creep and Extrusion Are Not the Same Failure

Both are called cold flow in catalogue copy, and they are fixed in opposite ways.

Creep is compression set: the seat thins under sustained load, uniformly, and contact stress between seat and ball decays with it. It is a bulk property and accelerates with temperature. For virgin PTFE, compression set after 1,000 hours runs about 2–3 percent at 20 °C, 10–15 percent at 150 °C, 20–30 percent at 180 °C and 40–50 percent at 200 °C.

Extrusion is local: the seat lip is pushed into the clearance between ball and seat pocket, shears, and leaves a path. It depends on pressure, temperature, the size of that clearance, and how much unsupported seat area faces it — not on the polymer alone.

The distinction matters commercially. A buyer who sees leakage at temperature and upgrades to PEEK may pay three times the seat cost to fix a creep problem that a filled PTFE would have solved at 1.1 times — or pay for PEEK when the real fault was a 0.3 mm pocket gap that no polymer survives.

Why the Same PTFE Works at DN15 and Fails at DN80

Two size effects decide whether a seat survives, and neither appears on a datasheet.

The first is differential expansion. PTFE expands at roughly 12 to 13 × 10⁻⁵ per kelvin; 316 stainless at about 1.6 × 10⁻⁵. Over a 160 K rise the seat grows about 1.7 percent more than its pocket: a quarter of a millimetre at 15 mm, 0.9 mm at 50 mm, over 1.4 mm at 80 mm. That growth has to go somewhere — it either closes the running clearance and raises torque, or buckles the seat and loses contact in one sector.

The second is unsupported area: sealing perimeter grows with diameter, but the seat's unsupported width does not shrink to compensate, so bending load on the lip rises faster than the sealing force it produces.

This is why a DN15 PTFE seat at 180 °C can run for years while a DN80 in the same service deforms in months. Any seat recommendation that does not ask for the size is incomplete.

Virgin PTFE: The 150 °C Line Nobody Prints

Virgin PTFE is rated to about 200 °C and is the best sealing material in the catalogue below 150 °C: lowest friction, widest chemical resistance, and enough conformity to seal against a ball that is not perfectly round.

Above 150 °C it becomes a maintenance item. At 180 °C it has already lost a fifth to a third of its thickness to compression set, and its pressure capability has fallen with it — a seat that holds 40 bar cold holds roughly 24 bar at 180 °C, about 60 percent of rating. At 200 °C the loss is close to half.

So the practical limit for critical isolation is 150 °C, not the 200 °C on the certificate. The certificate describes the polymer's thermal ceiling; the valve lives at its mechanical one.

RPTFE: Glass Buys Strength and Takes Away Conformity

Reinforced PTFE is the same polymer with a filler, and the filler is the whole point. Fifteen percent glass lifts compressive strength from roughly 30 to 40 MPa and cuts creep noticeably. Twenty-five percent carbon or graphite reaches about 45 MPa, with lower friction and better performance under high pressure-velocity, which is why carbon grades dominate steam and hot oil.

The costs are real. Friction rises, so breakaway torque rises with it. Glass is abrasive: in dirty service it polishes the ball rather than bedding to it. And the filler is chemically vulnerable — glass-filled PTFE is attacked by hydrofluoric acid and hot concentrated caustics.

What the filler does not do is raise the ceiling much. Both grades still stop around 220 °C, because the limit is the PTFE matrix, not the additive.

TFM: The Grade Between PTFE and PEEK

TFM is modified PTFE — same chemistry, lower crystallinity. The result is less creep, lower permeability and better behaviour under repeated thermal cycling, with a service range to about 260 °C and a cost 1.5 to 1.8 times virgin PTFE.

It is the right answer in three situations that are easy to miss: lines that cycle daily between ambient and operating temperature, oxygen service where permeability and cleanliness matter, and vacuum service where a seat that outgasses is a problem before it is a leak.

TFM is not a stronger PTFE so much as a more dimensionally stable one. If the failure is extrusion through a wide gap, TFM will extrude too.

PEEK: No Creep, No Conformity

PEEK is a different material, not an improved PTFE. Compressive strength around 90 MPa, continuous service to 250 °C and short excursions to 300 °C, and effectively no cold flow. It carries Class 2500 in the right body design and resists rapid gas decompression, which is why it appears in high-pressure gas service.

What it gives up is conformity. PEEK will not bed into a ball that is slightly out of round or a bore that is slightly oval. It seals because the geometry is right, not because the material yields — so ball finish, sphericity and pocket alignment must be better than for PTFE. Breakaway torque rises 30 to 50 percent, and the material is brittle under impact, a poor match for a valve that gets slammed.

PEEK fixes creep and extrusion by converting the problem into a machining problem. That is a good trade at 200 °C and a bad one at 80 °C.

The Comparison, in the Numbers That Decide It

Seat material	Service range	Compressive limit	Creep	Breakaway torque	Relative cost	Characteristic failure
Virgin PTFE	−29 to 200 °C	~30 MPa	High above 150 °C	Lowest	1.0	Compression set, lost preload
RPTFE, 15% glass	−29 to 220 °C	~40 MPa	Moderate	Low	1.1	Ball polishing in abrasive duty
RPTFE, 25% carbon	−29 to 220 °C	~45 MPa	Moderate	Low	1.15	Carbon shedding in clean media
TFM	−29 to 260 °C	~35 MPa	Low	Low	1.5–1.8	Extrusion at wide clearances
PEEK	−60 to 250 °C	~90 MPa	Negligible	30–50% higher	2.5–3.0	Leak from poor ball geometry

Torque Is a Seat Decision, Not an Actuator Decision

Actuators are sized from a torque figure, and that figure is a property of the seat at operating temperature.

Three effects run at once. Increasing contact stress raises torque. Creep lowers it, because the seat is losing the interference that generated the stress. Differential expansion raises it again as the seat closes the clearance. The net curve is not monotonic: a valve that turns noticeably easier over its first season is losing preload, not running in.

Size the actuator at operating temperature and maximum differential pressure, with a margin for the seat grade actually supplied. An actuator selected from a PTFE datasheet, installed on a PEEK seat, will stall.

What to Write on the Datasheet

One line is enough, and it should name the grade, the limit and the mechanism:

"Seat: 25% carbon-filled PTFE, service −29 to 220 °C, rated for 180 °C saturated steam; PEEK anti-extrusion ring; breakaway torque not to exceed the stated value at 20 bar differential and 180 °C."

Then ask three questions the line cannot answer. What is the seat stress at operating temperature, not at ambient? What is the ball surface roughness and sphericity? What is the ball-to-pocket clearance, and is there a backup ring in it?

A supplier who answers those three is discussing a seal. One who answers only "PTFE" is quoting a word.

Is RPTFE always better than virgin PTFE?

No. It resists creep and wear better, but friction is higher, glass can polish the ball in abrasive duty, and it is attacked by hydrofluoric acid and hot caustics. Below 150 °C with clean media, virgin PTFE seals better.

Can a PTFE seat handle 200 °C steam?

Only intermittently. Compression set reaches 40–50 percent after 1,000 hours at 200 °C, and the seat loses the preload that keeps it closed. Sustained steam above 180 °C should use carbon-filled PTFE, TFM or PEEK, with the pressure rating derated to match.

Why does a ball valve start leaking a few months after installation?

Usually creep, not wear. The seat has lost thickness under sustained load at temperature, contact stress has fallen, and the leak appears once the valve cools or pressure drops. A ball or body fault would leak from the first cycle.

Does a PEEK seat need a better ball finish?

Yes. PEEK does not conform, so it seals on geometry. Roughness, sphericity and pocket alignment must be tighter than for PTFE, and torque will be 30 to 50 percent higher — the actuator has to be sized for it.