Thermoplastics, thermosets and elastomers

Heating reveals different architectures

A flexible cable sleeve is stretched, released and later reshaped while warm. Another sleeve springs back just as well but cannot be remoulded by the same route. Which one is an elastomer? Potentially both. Which one is thermoplastic? That asks a different question about its processing history.

Thermoplastics can soften into a processable state when suitable temperatures allow chain motion. Conventional cured thermosets contain permanent networks and cannot simply be remelted like ordinary thermoplastics. Their actual operating and degradation ranges still depend on chemistry and formulation.

Thermoplastic describes a polymer that can repeatedly soften on heating and become firmer on cooling within a suitable range. The required molecular motion must occur on the processing clock. In an ordinary uncrosslinked thermoplastic, chains can eventually move relative to one another without cutting their backbones. Entanglements and other interactions can still make that motion slow; softening is not instantaneous free flow.

Thermosetting describes a material that forms a persistent network through curing; the cured product is a thermoset. Curing changes chemical connectivity. Cooling an already formed thermoplastic changes its physical state without requiring that same network-forming reaction. Heating can accelerate a cure, but the category does not say that every formulation is cured by heating or provide a temperature/time recipe.

The two processing histories are not a stiffness ranking. An uncrosslinked glassy thermoplastic can be quite rigid at the use temperature. A cured network with flexible strands can be rubber-like. A chemical-family name also may not settle the category: a thermoplastic polyurethane and a permanently crosslinked polyurethane are different architectures. Formulation and the actual state matter.

Ask which connection must survive, and which must release

Consider a representative block-copolymer thermoplastic elastomer. A block is a sequence of similar repeat units joined to another sequence within the same molecule. In an A–B–A molecule, relatively rigid A end blocks can aggregate with A blocks of neighbouring molecules into small domains. The more mobile B middle portions connect some of those domains. Such domains act as physical junctions while they hold the chain ends on the observation timescale.

A styrenic example has polystyrene-rich glassy domains with more mobile rubbery middle blocks. Other thermoplastic elastomers use different associations or crystalline junctions. A domain is a small region of material, not a new covalent bridge between every molecule inside it. Some molecules loop back to one domain or have a loose end, so an A–B–A label alone does not prove a particular network or modulus.

At the chosen use temperature, long enough persistence of the junctions prevents unconstrained whole-chain displacement during a short stretch. At a suitable higher shaping condition, the junctions can rearrange or release chain ends quickly enough for the body to flow under the applied forming load. The end blocks remain covalently attached to their own middle blocks: releasing an end from a physical junction is not the same as cutting that backbone.

Cooling in a new shape can make physical junctions persistent again, possibly with different neighbouring ends. The material need not return to the object’s pre-processing shape: that earlier arrangement of junction partners was allowed to change. This differs from releasing a short use-temperature stretch while the relevant constraints persist. The first figure follows that distinction with selected fragments, not a measured network or a molecular length scale.

In a conventional cured thermoset, permanent covalent connections remain when segmental mobility increases, provided bonds neither break nor exchange. The body can become more compliant without becoming a freely separating melt. Heating until bonds degrade is not ordinary reversible remelting. Special networks deliberately designed to exchange bonds can be reconfigured under their own conditions; their existence does not make an unknown conventional cured material melt-processable.

Rubbery describes a response regime

An elastomer can undergo large recoverable deformation because flexible chains rearrange within a constraining network or equivalent physical structure. Thermoplastic elastomers use reversible structural constraints. These categories therefore overlap in useful ways; they do not mean hard, medium and soft grades of one universal scale.

In a rubbery response, mobile portions between persistent constraints change conformation, the spatial shapes available to the same connected chain. A stretch biases those portions toward more extended arrangements. There are generally more accessible coiled arrangements than strongly aligned ones, giving a statistical tendency to retract when the external load is removed. Entropy here describes that multiplicity of arrangements; the explanation does not require every backbone bond to lengthen by the macroscopic percentage.

Constraints are essential to distinguish useful recovery from unconstrained relative chain motion. They can be permanent chemical links in a conventional cured rubber or physical junctions in a thermoplastic elastomer. Chain segments also need adequate mobility on the test clock. Cooling a rubbery material until segmental rearrangement becomes slow can make its response glassier even though no new covalent links were formed.

A single material can therefore be both an elastomer under specified use conditions and a thermoplastic under specified processing conditions. A cured silicone rubber is a representative thermoset elastomer; a styrenic block material can be a thermoplastic elastomer. Neither label assigns hardness, tensile strength, chemical resistance or an allowable service temperature to every member of its family.

Real elastomers can show delayed recovery, hysteresis, drift or damage. Greater extension is not unlimited reversible stretch, and a small residual measured soon after unloading is not proof of a permanent set. Viscoelasticity, temperature and time develops controlled histories and response models. Here we compare supplied observations without fitting another constitutive law.

Measure recovery against the extension that was imposed

Use an original invented observation set, separate from the schematic. Three strips P, Q and R begin with an original marked gauge length L₀ = 40.0 mm at 25 °C. Each is extended to L_max = 60.0 mm in the same 5.0 s ramp, held there for 10.0 s, unloaded over 5.0 s and left at zero external force for 60.0 s. The reported L_after is then measured at the same 25 °C. Assume intact gauges and exact stated values for this arithmetic exercise; this is not a standard test or a named-polymer dataset.

All three have imposed engineering strain ε_max = (L_max − L₀)/L₀ = 20.0/40.0 = 0.500, or 50.0%. Equal imposed strain does not imply equal force or work; those records are not supplied. The after-unloading lengths are P: 55.0 mm, Q: 41.0 mm and R: 40.4 mm.

Define a recovery fraction for this particular comparison: r = (L_max − L_after)/(L_max − L₀). Its numerator is the extension removed by the stated recovery reading; its denominator is the extension imposed relative to the original gauge. Both use matching length units, so r is dimensionless. The remaining engineering strain is instead ε_after = (L_after − L₀)/L₀. These quantities have different denominators.

P has recovered 5.0/20.0 = 0.250, or 25.0%, of the imposed extension, leaving 15.0 mm and ε_after = 37.5%. Q has recovered 19.0/20.0 = 95.0%, leaving 1.0 mm and ε_after = 2.50%. R has recovered 19.6/20.0 = 98.0%, leaving 0.4 mm and ε_after = 1.00%. For Q, 95% extension recovery is not a 95% reduction in the original total length.

The second figure keeps the 20.0 mm imposed-extension basis common. These are finite observations after one specified history, not universal recovery percentages for categories. No threshold in this lesson defines elastomer for every standard or application. Longer recovery, another temperature, repeated cycles or another maximum extension may give different values; a residual cannot be named permanent without the required further evidence.

Combine the history with independent connection evidence

The fictional records supply more than the length test. Independent structure information identifies P as uncrosslinked, Q as a material with reversible physical junctions and no permanent network, and R as an already cured conventional covalent network. A separate 120 s shaping comparison uses uniform 160 °C, the same forming fixture and a common constant forming force. P and Q each flow and retain a new shape on cooling to 25 °C. R becomes more compliant but does not show that flow. No backbone scission, bond exchange, further cure or mass loss occurs in these stipulated shaping records; R’s network remains intact. These are invented qualified conditions, not recommended processing settings.

Together the observations support P as thermoplastic but do not show useful rubber-like recovery for this particular large-extension history. Q combines thermoplastic shaping with substantial rubber-like recovery at the use condition: a thermoplastic-elastomer interpretation is supported. R combines a conventional thermoset network with substantial recovery: it is the thermoset-elastomer comparison. High recovery has not proved that R can be remelted; remoulding Q has not proved that it is rigid during use.

Suppose an invented sleeve screen requires r ≥ 0.950 after this exact 60 s recovery and an independently demonstrated ordinary heat-reshaping route. Q meets both supplied screens; R meets recovery but misses that reshaping route, while P misses recovery. If the recovery requirement changes to r ≥ 0.960, Q no longer meets it, and none meets both. Passing these two screens alone does not establish load capacity, seal performance, fatigue life or safety.

An unknown sample needs the same separation of evidence. No flow in a short heating test could mean insufficient mobility, insufficient stress, persistent physical junctions or a permanent network. Flow after overheating could include degradation. A label such as rubber-like, one indentation hardness, one melting feature or one insoluble residue is insufficient by itself. Crystallinity, crosslinks and network formation explains why order and chemical connectivity need separate evidence.

A processing opportunity is not a performance guarantee

Change the sleeve’s use temperature while keeping its task. If Q’s physical junctions no longer persist for the required hold, it can creep or retain a changed shape even though it was rubber-like in the original 25 °C record. A higher shaping temperature is useful precisely because constraints can release there; it does not automatically provide a wider service range. R’s retained covalent connectivity also does not guarantee an unchanged modulus, chemical stability or service life.

Reheating a thermoplastic can permit reshaping, but useful recycling additionally needs compatible composition, controlled contamination and a qualified processing history. Chain scission, reactions, additives, orientation, porosity and cooling structure can change during reuse. A new shape is evidence of forming, not evidence that all properties returned to their original values. “Why processing history matters” develops that state dependence.

Cutting a cured thermoset into particles does not undo the network inside each particle. Reusing those particles as a filler is a different route from remelting the old network into an equivalent homogeneous polymer. Chemical recovery or deliberately exchangeable networks are also different routes with their own inputs and evidence. Neither thermoplastic nor thermoset is a complete environmental ranking.

Use the categories to ask better questions: Which constraints are present? What can move over the actual temperature and time history? Which recovery or forming result was measured? Record state, formulation, direction, geometry, conditioning, load/ramp/hold/recovery history and uncertainty. Unknown polymers should not be heated or chemically treated to identify them using this lesson: real processing needs the material-specific safety and qualified equipment conditions.

Worked examples

Try the reasoning

1. Can every cured thermoset be reshaped by ordinary melting?

   Show the answer and reasoning

   No.

   A permanent network prevents that conventional route. Increased segmental mobility can soften the response while the conventional chemical connections remain. Bond exchange, chemical recovery or degradation would be a different, specifically evidenced route.

2. Does thermoplastic mean soft at room temperature?

   Show the answer and reasoning

   No.

   Its state depends on transitions and operating conditions. Glassy thermoplastics can be rigid; rubber-like response can occur in thermoplastic or cured-network materials. Use the specified time, temperature and formulation.

3. An unfamiliar Q-type strip starts at L₀ = 50.0 mm, is stretched to 75.0 mm, and is 52.0 mm after the same ramp/hold/recovery timing at 25 °C. Calculate r and remaining strain. Does it meet a separate r ≥ 0.950 screen?

   Show the answer and reasoning

   r = 23.0/25.0 = 0.920 = 92.0%; remaining strain is 2.0/50.0 = 4.00%. It misses that recovery screen.

   Use the new imposed extension in r’s denominator, rather than the previous strip’s 20.0 mm. The original length remains the strain denominator. The calculation does not change the structural category or establish final recovery.

4. In a separate Q repeat with the original 40.0 → 60.0 mm stretch, the length is 43.0 mm after 10 s recovery but 41.0 mm after 60 s. What are the two r values? May the 10 s residual all be called permanent?

   Show the answer and reasoning

   85.0% after 10 s and 95.0% after 60 s; the earlier residual is not all permanent.

   The recovered lengths are 17.0 and 19.0 mm, divided by the same 20.0 mm imposed extension. The measured recovery between readings directly rules out treating the entire earlier 3.0 mm residual as permanent. It still does not establish the eventual limit.

5. Q is formed into a longer shape while its physical junctions can rearrange and is then cooled unloaded. A learner expects it to recover the object’s original pre-processing length simply because it is an elastomer. What has changed?

   Show the answer and reasoning

   The arrangement that constrained the original object was allowed to rearrange during processing; cooling can stabilise a new reference shape.

   Rubber-like recovery from a short use-temperature stretch assumes the relevant constraints persist. It does not require memory of every previous shape after deliberate remoulding. Re-forming physical junctions can involve different partners while the covalent backbones remain intact.

6. A Q-like part is rubber-like in a short 25 °C test. At a new use temperature its physical junctions release on a 20 min hold, without chain scission. Is continued shape recovery guaranteed by the word elastomer, and has loss of persistence necessarily broken the backbone?

   Show the answer and reasoning

   Neither claim follows. The relevant constraint no longer persists on that use clock, and physical release need not cut the backbone.

   The response depends on temperature and time relative to junction persistence and chain motion. The original classification record is not an allowable-temperature or lifetime guarantee. Measure the new load and recovery history under that condition.

7. An unknown soft sample shows no flow during one short heating observation. A report calls it a cured thermoset and proposes a higher temperature until it flows. What conclusion and action are unsupported?

   Show the answer and reasoning

   A permanent network has not been established, and uncontrolled extra heating is not an identification or remelting method.

   Slow mobility, physical constraints, stress and duration can also suppress observed flow. Overheating can cause chemical change, decomposition or other hazards. Use appropriate material-specific structural and qualified processing evidence instead.

8. The only uncertainty in a separate Q reading is L_after = 41.0 ± 0.2 mm, interpreted here as a stipulated possible interval; L₀ = 40.0 and L_max = 60.0 mm are exact. Find the corresponding r interval. Is r ≥ 0.950 established for every value in that interval?

   Show the answer and reasoning

   L_after from 40.8 to 41.2 mm gives r from 0.960 down to 0.940. The whole interval does not meet r ≥ 0.950.

   r decreases as L_after increases. Evaluate both endpoints and reorder them: 94.0% to 96.0%. The central 95.0% result sits on the threshold. This is interval propagation for supplied possible bounds, not a confidence interval or a general uncertainty model.

What to carry forward

• Separate architecture, response and processing category.
• State the relevant temperature regime.
• Mobile strands plus persistent constraints can give rubber-like recovery; junction release during shaping is a different history.
• Use the stated recovery denominator and observation time; finite remaining strain is not automatically permanent.
• A reversible shaping route, recovery performance and a qualified reuse outcome need their own evidence.