What Temperature Does to a Material's Behaviour

Why this matters

A material does not have one set of properties. It has a set of properties at the temperature you tested it, and a different set at the temperature it actually lives at. The part that held on the bench at shop temperature and split on the third hot afternoon was not defective, it was operating outside the window its published numbers were measured in. Every rating you look up - pressure, tensile strength, hardness, chemical resistance - was measured at a stated temperature, and most of them fall off a cliff somewhere above or below it. If you do not know where that cliff is for the class of material in your hand, you are guessing on every substitution.

Temperature does four separate things

Keep them separate in your head, because they fail differently and they show different symptoms.

1. It changes stiffness and strength continuously. Metals lose yield strength as they warm and gain it as they cool. Plastics soften and lose pressure capacity long before they visibly deform. This is gradual and predictable, and it is what a derating table describes.
2. It crosses transitions where behaviour changes character. A ductile steel goes brittle below its transition temperature. An elastomer goes from rubbery to glassy below its glass transition and stops sealing. A thermoplastic passes its heat-deflection point and starts to creep under a load it carried fine an hour earlier. These are not gradual; the part behaves like a different material on the other side.
3. It drives time-dependent effects. Creep, stress relaxation, and compression set all run faster hot. A bolted joint at elevated temperature loses clamp load with no one touching it. A gasket takes a permanent set and no longer springs back when the joint cools.
4. It accelerates chemistry. Oxidation, chemical attack on an elastomer, adhesive cure, lubricant breakdown, and corrosion all speed up with heat. A fluid that a seal tolerates at room temperature can attack the same seal at operating temperature.

A part can pass on all four counts and still fail because the fifth thing - dimensional change - pinched it. That belongs to the joint, not to the material, and it is covered separately.

The transitions worth knowing by class

The lesson in the table is that the hot-side and cold-side failures are not mirror images. For rigid plastics the hot side is a pressure problem and the cold side is an impact problem. For steel it is the reverse. Do not carry one material's habits onto another class.

The gate: extremes, not averages

Here is the rule this article is built on, and both cases below are run through it exactly as written.

• Unit of analysis: one part, at the worst temperature that specific part sees, not the loop average, not the room temperature, not the setpoint. A circulator seal 10 inches downstream of a heat exchanger sees the exchanger's outlet, not the return.
• Gate, both halves required (AND, not OR): the material's published continuous-service limit sits above the highest temperature the part actually reaches, with headroom; and the lowest temperature the part reaches sits above its brittle or glass transition.
• Headroom starting point: 20 degrees F on the hot side for a static part in a controlled space, more where the temperature is not tightly controlled or where a limit control could fail high. That is a practice default to tune to your own service, not a published spec. Where the manufacturer publishes a derating curve, read the rating at your actual temperature instead of applying a flat margin.
• Step when a half fails: move up one material class. Do not shade the number, do not average two readings to get under the limit, and do not treat a peak as an outlier because it is brief. A peak is where the part fails.

Case one: the seal that passes

A hydronic circulator seal. Measured supply temperature at the pump body on a full call for heat: 200 degrees F. The mechanical room is unheated and drops to 55 degrees F in January with the boiler off.

The seal is an ethylene propylene class elastomer, the common choice for hot water and glycol. Published continuous-service limits for that class in water service commonly run to about 250 to 300 degrees F, and its low-temperature flexibility extends well below zero.

Run the gate. Hot half: take the conservative end of the published range, 250 degrees F, against a 200 degree F peak. That is 50 degrees F of headroom, which clears the 20 degree F starting point with room to spare. Cold half: 55 degrees F is nowhere near the class's low-temperature limit. Both halves hold, so the gate passes and the material stays.

Now the part that is easy to miss. Passing the temperature gate does not clear the fluid. Put a petroleum-based lubricant or a petroleum-carrying additive in that same loop and this elastomer class swells and softens, because it has poor resistance to petroleum oils. That is the chemistry axis, it is independent of the temperature gate, and it is the single most common way a correctly rated seal gets destroyed.

Case two: the fitting that fails

Same shop, a different job: a plastic fitting on a line that reaches 165 degrees F at peak. The fitting is a standard rigid vinyl class, and the tech reaches for it because it is on the truck and the pipe it is joining is the same material.

Hot half of the gate: the published maximum service temperature for that class is 140 degrees F. The part sees 165 degrees F. The gate does not just fail on headroom, it fails on the ceiling itself, by 25 degrees F. That ends the analysis and the fitting comes off the truck.

Look at what the pressure rating was already doing before the ceiling was reached, because it explains why "it only gets hot for a few minutes" is not a defence. Pressure ratings for this class are published at 73 degrees F and derate steeply with temperature; the published multipliers fall to roughly 22 percent of the 73 degree F rating by the time you reach 140 degrees F. A fitting carrying a 150 psi rating at 73 degrees F is therefore good for roughly 33 psi at 140 degrees F. If the line runs at 60 psi, the fitting was already past its rating at 140 degrees F, which is 25 degrees F below the temperature it actually sees.

Step, as written above: move up one class. The chlorinated vinyl class is commonly rated at 180 degrees F at 100 psi, which is the standard rating point to design against rather than the higher figure sometimes quoted and its own derating table, which is why it exists. Check the derating at 165 degrees F rather than assuming the higher ceiling solves the pressure question, because it may not at high line pressure.

The failure mode if you skip this: the joint does not burst on the bench and it does not burst on the day you install it. It creeps. The fitting deforms slightly under pressure at temperature, relaxes when it cools, and does it again every cycle until the joint weeps or lets go, usually weeks later, usually with nobody there. That is why it comes back as a water-damage claim rather than a callback.

Rate effects: the failure with no event

Creep and stress relaxation are the two temperature effects that produce no symptom you can catch on a single visit, because nothing happens fast.

Creep is a material slowly deforming under a constant load. It runs faster hot and it never reverses. A plastic bracket carrying a steady weight, a gasket under bolt load, a hanger strap on a hot line - all of them slowly move.

Stress relaxation is the same physics seen from the other side: the deformation is held constant and the force decays. This is what empties a bolted joint. Nobody loosened it, the clamp load simply bled off at temperature. Joints that run hot are the ones that need a re-torque check after they have been through their first full heat cycle.

Compression set is the elastomer version. A seal squeezed at temperature stops springing back. Pull a set gasket and it will look flat and take a fingernail impression rather than rebounding. That gasket sealed fine while hot and clamped, and leaked the first time the joint cooled and moved.

What temperature does not explain

Heat is blamed for more faults than it causes, and the confusion costs a diagnosis. A thermal cause tracks a temperature. If the fault tracks a clock, a cycle count, or a load instead, look elsewhere.

• A joint that weeps at every temperature is a preparation or material-compatibility fault, not a thermal one.
• A part that fails at a repeatable run time regardless of ambient is a control, a duty-cycle, or a lubrication problem.
• A part that fails only after a specific fluid change is chemical attack, even if it also happens to be hot.
• A part that fails on first startup is a fit, torque, or handling problem. Materials do not degrade in one cycle.

The test that separates them: does the symptom move with outdoor or process temperature across separate visits? If two visits at different temperatures produce the same symptom at the same severity, temperature is not the variable.

Checking you got this right

• Write down the actual measured peak temperature at the part, with the meter and location noted, rather than the setpoint or the design figure.
• Confirm the published limit you used is for continuous service, not an intermittent or short-term excursion rating. The two numbers are often 30 to 50 degrees F apart and they are not interchangeable for a part that lives there.
• Confirm the pressure rating you used was read at your temperature, not at the room-temperature headline number.
• Confirm the fluid, not just the temperature. A pass on temperature and a fail on chemistry looks identical in the failed part.
• On a hot joint, schedule the re-check after the first full heat cycle rather than before you leave. Clamp load measured cold on a joint that has never run tells you nothing about what it holds hot.

References

• Manufacturer published derating tables for pressure ratings at elevated temperature; ratings for rigid vinyl and chlorinated vinyl piping classes are published at 73 degrees F and derate with temperature
• Manufacturer elastomer compatibility and temperature-range data for the specific compound, not the generic family name
• ASTM standard test methods for heat deflection temperature and compression set, as referenced on material data sheets
• See related: Expansion and Contraction in the Field; Gasket + O-Ring Material Reference; Reading Rust and Corrosion Patterns