Standard FKM (copolymer of VDF and HFP) typically shows a TR-10 near −15°C, while peroxide-curable terpolymers with PAVE sit around −20°C and the GLT family (perfluoroether polymer with a peroxide cure site) reaches roughly −35°C [S2].
The same data also points to FKM (B-type, peroxide cure) near −10°C and ETP-S (Aflas-type) at around 0°C, putting a roughly 35°C spread across the commercial fluorocarbon family for cold-rating decisions [S2].
What TR-10 Actually Measures in an FKM Sample
TR-10 is the temperature at which a frozen, elongated FKM specimen has retracted 10% of its stretch while warming under the procedure in ASTM D1329, the same method used for ISO 2921 [S2]. A die-cut strip, typically 2.0 mm ± 0.2 mm thick, is stretched 50% (the most common TR10/50 notation), cooled below recovery, released, and then warmed at a uniform rate while length is recorded [S2].
The number that comes out of that curve is where the compound starts behaving like rubber again rather than a stiff plastic, which is why TR-10 is the figure most engineers treat as the practical cold floor for an elastomer in a static seal [S1]. A common rule of thumb in the rubber industry is to take the TR-10 in °F and subtract 10°F to set a service limit, which is roughly a 5–6°C safety margin below the test point [S1].
FKM Family Comparison on TR-10 and Other Decision Criteria
Four FKM sub-families are usually enough to cover most seal specs, and the cold side of the comparison is what TR-10 is built to show: [S2]
Standard FKM (A-type, VDF/HFP copolymer, bisphenol cure): TR-10 around −15°C, broadest chemical resistance, lowest cost, the default for general hydrocarbon service. FKM B-type (VDF/HFP/TFE terpolymer, peroxide cure): TR-10 near −10°C, better acid and amine resistance, slightly worse cold. FKM GFL-S or GLT family (perfluoroether backbone, peroxide cure): TR-10 around −35°C, best cold performance, higher cost, narrower media list [S2].
ETP-S (Aflas, TFE/propylene copolymer): TR-10 around 0°C, used for amines, steam, and some acids where standard FKM is not enough; cold rating is the worst in the family [S2]. A wider gap between TR-10 and TR70 on the same warm-up curve often signals stronger crystallization tendency, which matters more for dynamic lips than for static glands [S2]. For a broader material reference, see the fluororubber entry covering cure systems, monomer families, and standard compound limits.
Static vs Dynamic Cold Limits: Why TR-10 Is Not the Whole Story
Catalog minimum-temperature ratings for FKM are usually derived from TR-10, with or without an extra margin, and that is the first place spec mistakes happen [S2]. A static O-ring in a compression gland can usually operate at or slightly below TR-10 because it does not need to flex repeatedly; the squeeze force stays adequate as long as the material is past its glassy transition [S2].
A dynamic seal, by contrast, needs TR-10 well below the operating temperature plus a kinetic margin, because the lip has to keep bending at every cycle and any crystallization during cold soak delays recovery [S2]. For dynamic FKM service below −20°C, the GLT family is essentially the only credible option, and even then service is usually limited to static or slow reciprocating duty [S2].
Adjacent Elastomer Reference Points for the Same Cold Test
TR-10 is not unique to FKM, and reading the FKM number against a few neighbours makes the trade-off visible: [S2]
NBR (nitrile) typically lands around −20°C to −25°C, which is why NBR is often the cheaper pick when cold duty is the constraint and oil resistance is moderate [S2]. HNBR runs near −25°C with better heat and ozone resistance than NBR. EPDM sits around −45°C, cheaper again, but disqualified wherever the service fluid is petroleum. Silicone (VMQ) is near −55°C, and low-temperature FKM (GLT) is sometimes specified not because it is the coldest elastomer available but because it is the cold-compatible one with full hydrocarbon resistance [S2].
For lab context, the temperature-retraction method is also used in DSC and DMA workflows where the same heating ramp and retraction response cross-check transitions seen in thermal analysis; the same recording chain shows up in temperature recorder hardware used in TR test stands. Compression set at low temperature, a related but distinct property, is typically measured on a parallel rig and reported separately from TR-10 [S4].
Standards, Test Geometry, and What the Report Should Show
ASTM D1329 is the US method and ISO 2921 is the equivalent international method, and either one is acceptable for a published TR-10 number on a fluorocarbon compound [S2]. The stretch ratio used during the test must be on the report, since TR10/25 and TR10/50 on the same compound will not produce identical numbers; 50% elongation is the most common reporting condition in industrial seal work [S2].
Specimen geometry is also part of the spec: die-cut strip from a cured slab, about 2.0 mm thick, working length typically 1.5–2.0 inches, and a uniform warm-up rate on the order of 1°C per minute is standard practice across the methods cited [S2]. For testing workflows, a temperature controller driving the chamber bath and a temperature monitor logging the specimen environment are the usual instrumentation; a calibrated temperature measurement chain is what makes the −35°C vs −15°C comparison defensible.
Selection Rules for Specifying TR-10 on a Fluorocarbon Seal
Three rules cover most FKM cold-side spec errors seen in industrial purchasing and seal design. First, never spec a fluorocarbon below its TR-10 in dynamic service without confirming the curve in TR70, because a wide TR-10–TR70 gap signals crystallization that will punish a moving lip well before TR-10 is reached [S2]. Second, pick GLT only when both the cold limit and the media list demand it; for static hydrocarbon service above −20°C, a standard FKM is usually a stronger economic and chemical fit [S2].
Third, when the seal is in a refrigeration, LNG, or low-temperature chemical stream, size the gland and the squeeze for the actual service temperature plus 5–6°C of margin, and verify with a lab TR-10 at the production cure condition rather than trusting catalog generic numbers [S1]. A final trackable signal: a parallel standard deviation across at least three FKM lots is the simplest way to flag compound drift that would otherwise be hidden inside a single TR-10 value [S2].
Related analysis: Rebar Unit Weight: #4, #5 and #6 Bars in lb/ft.