A buyer signs off on a cure system because the sheet reads "CBS, CAS 95-33-0, assay 96%". Six weeks later the extruder starts scorching at the die and nobody can explain it, because nothing on that sheet changed. The document that actually describes the compound is the one nobody files: the cure curve.

A cure curve is torque against time at a fixed temperature. Every point on it belongs to a different ingredient, and the three numbers a compounder quotes from it answer three separate questions: how long this compound can sit in a hot machine, how long it takes to crosslink, and how much of the network survives an overcure. An accelerator bought on assay alone answers none of them.

One curve, three questions

ReadingHow it is takenThe question it answersWhat moves it
ML, minimum torqueThe floor of the curve before the riseViscosity of the uncured compound — how it flows and fillsPolymer Mooney, filler, oil, stearic acid level
ts2, scorch timeTime for torque to rise 2 dN·m above MLProcess window: how long before the compound stops flowingAccelerator family, retarders, mixing temperature
t90, optimum cureTime to reach ML + 0.90 × (MH − ML)Press or autoclave residence timeAccelerator activity, activator, cure temperature
MH, maximum torqueThe top of the curveStiffness and modulus of the fully cured partSulfur level, crosslink density, filler loading
ΔM = MH − MLThe height of the curveRelative crosslink development, within one compound onlySulfur/accelerator ratio, filler network
CRI, cure rate index100 ÷ (t90 − ts2)How fast the network actually forms, with scorch removedAccelerator activity, temperature

The definitions follow ASTM D2084 for the oscillating disc rheometer (ISO 3417) and ASTM D5289 for the moving die rheometer (ISO 6502). The two instruments are not interchangeable: the disc machine swings a rotor through 1° or 3° at 1.7 Hz inside a sealed cavity, while the moving die shears a thin pad at 0.5° arc and 1.67 Hz with a fraction of the sample. Numbers from the two are not comparable, and a supplier who sends a t90 without naming the instrument and the test temperature has sent you a number you cannot use.

Number one: ts2 is a process decision, measured at the wrong temperature

Almost every cure curve in a technical data sheet is run at 160 °C. Almost no rubber is processed at 160 °C. Mixing, extrusion and calendering happen between 90 and 130 °C, and that is exactly where scorch does its damage — a compound that gels in the extruder barrel cannot be saved downstream.

Reaction rate roughly doubles for every 10 °C, so walk a ts2 of 3 minutes at 160 °C down the ladder: 6 minutes at 150 °C, 12 at 140 °C, 24 at 130 °C, 48 at 120 °C, roughly an hour and a half at 110 °C. The compound with "3 minutes of scorch safety" on paper actually has hours of it in a warm pin barrel — and that is why the factory number is not ts2 at all. It is Mooney scorch at 120–125 °C.

Mooney scorch (ASTM D1646 part C, ISO 289-2) reports two times against a minimum viscosity Vm: t5, when torque has risen 5 Mooney units, and t35, when it has risen 35. The gap between them is the same idea as CRI on a different instrument, and t5 is the number a plant manager can defend in a scheduling meeting.

A worked example of what a single substitution costs. Replacing MBTS on an equimolar basis with a methylene-bridged benzothiazole in an 80 phr silica S-SBR/BR tread compound moved Mooney t5 from 12.5 to 19.8 minutes at 125 °C — a 58% wider process window — while CRI from moving die rheometry at 160 °C fell from 10.3 to 7.1 min−1. You bought window and paid for it in cycle time. That trade is a purchasing decision, not a lab decision, and it only becomes visible when both numbers sit on the same sheet.

The accelerator ladder, read as scorch safety improves:

  • Dithiocarbamates — ZDEC (EZ), ZDBC (BZ), PZ (ZDMC), TDEC. The fastest and the least forgiving, typically used well under 1 phr. The right answer for thin sections and latex, the wrong one for anything with a long flow path.
  • Thiazoles — MBT (M) and ZMBT. The general-purpose floor of the system. Cheap, fast, and they will scorch a hot compound if you push the loading.
  • Sulfenamides — CBS (CZ), TBBS (NS), DCBS (DZ). Delayed action, which is the entire point: they release the active thiazole only after the compound has had time to flow. DCBS is the most delayed of the three, which is why it is the standard answer for thick sections and high-temperature press cure.
  • Guanidines — DPG. A booster almost never used alone, and a quiet thief of scorch safety: it raises the pH of the system and speeds up everything, including the part you wanted to stay slow.

Number two: t90 is a clock, CRI is a rate

Two compounds can both read t90 = 8 minutes and behave nothing alike. If one spent 4 minutes scorching and 4 curing, and the other spent 1 minute scorching and 7 curing, they share a clock time and have completely different chemistry. That is what CRI exists for: 100 ÷ (t90 − ts2) strips out the induction period and leaves the slope.

The spread is not academic. In one published NBR series, ts2 ran 1.31 to 2.61 minutes and tc90 ran 3.41 to 5.97 minutes, giving CRI values from 29.8 to 51.5 min−1 across compounds that all "cured in four to six minutes". Comparing t90 across that set would have told you almost nothing.

CRI converts directly into money. Take a 20-cavity press with 3 minutes of load and unload. At t90 = 8 minutes the cycle is 11 minutes, or 5.45 cycles an hour; at t90 = 6 minutes it is 9 minutes, or 6.67 cycles an hour. That is 22% more output from one change to the cure system, with no capital spend. Run the same arithmetic in reverse and you see why a delayed-action accelerator that costs more per kilo can still be the cheaper choice.

The trap is that temperature moves all three numbers at once. Going from 150 to 160 °C roughly halves t90, which looks like free capacity, but it halves ts2 as well and pushes natural rubber past the point where the network starts to come apart. Section thickness sets the temperature floor; chemistry should be chosen to fit, not to fight.

Number three: ΔM is not a crosslink counter

It is tempting to read MH − ML as crosslink density and compare it across recipes. Do not. ΔM contains the filler network as well as the chemical network, so a compound with 80 phr carbon black and one with 40 phr will show different ΔM with identical sulfur chemistry. The Payne effect alone can move the plateau further than a 0.5 phr change in sulfur.

ΔM is useful in exactly one comparison: same recipe, same filler, one variable changed. For anything else, use equilibrium swelling and the Flory–Rehner equation, or a rubber process analyser at low strain. A supplier who claims "higher MH means better crosslinking" across two different compounds is telling you about the filler.

Reversion: the part of the curve that goes down

Natural rubber has a downslope. Overcure at high temperature breaks polysulfidic crosslinks faster than the network can mature, and the torque falls. The numbers are large enough to plan around. In a conventional (CV) sulfur system held 30 minutes at 180 °C, torque typically drops from 42–45 dN·m to 25–30 dN·m, a 35–40% loss. The same compound on an efficient (EV) system loses only 10–15% under identical conditions.

The reason is bond energy. A polysulfidic bridge sits near 252 kJ/mol, a monosulfidic bridge near 285 kJ/mol, and a carbon–carbon bond near 352 kJ/mol. An EV system is not a better cure; it is a more expensive bond, paid for in dynamic fatigue and tear strength.

SystemSulfur, phrAccelerator, phrAcc/S ratioNetworkBehaviour at 180 °C
Conventional (CV)2.0–3.51.2–0.40.1–0.6About 95% poly- and disulfidic, 5% monosulfidicReverts quickly; good fatigue and tear
Semi-efficient (Semi-EV)1.0–1.72.5–1.20.7–2.5About 50 / 50Intermediate; the usual compromise
Efficient (EV)0.4–0.85.0–2.02.5–12.5About 20% poly- and disulfidic, 80% monosulfidicHolds the plateau; lower fatigue

Note what the ratio column really means: "EV" is not a product, it is an accelerator-to-sulfur ratio above 2.5. Two compounds both sold as EV cure can sit at 2.6 and at 12, and their heat ageing will not match. Write the ratio, not the label.

Where the cure system alone is not enough, anti-reversion additives repair the network as it breaks. Sodium hexamethylene-1,6-bis(thiosulfate) dihydrate, normally 1.0–2.0 phr, inserts hybrid crosslinks carrying a hexamethylene segment inside the sulfur bridge; at 1.5 phr it typically pulls the 30-minute torque loss at 180 °C from 35–40% down to 15–20%, and in one NR belt compound the reversion rate fell from 40.3% to 10.5% with t90 essentially unchanged. At 5 phr, polysulfide loss between t90 and 60 minutes dropped from 30.9% to 15.9%. Bis-citraconimides work differently, forming thermally stable carbon–carbon crosslinks by Diels–Alder reaction with the conjugated dienes that reversion produces.

Two things not to confuse. SBR and BR usually show a marching modulus instead — the torque keeps creeping up rather than falling — so a flat curve is a virtue in NR and a symptom in SBR. And a compound that reverts on the rheometer will also lose modulus in service, which is why the curve is a service-life test as much as a process test.

Three variables nobody files under "cure system"

The first is the activator. Zinc oxide at 3–5 phr with 1–2 phr of stearic acid makes sulfur vulcanization commercially possible at all: without it, the same compound cures five to ten times slower. The chemistry runs through a soap. Stearic acid reacts at the ZnO surface to give zinc stearate and water, and that soluble zinc is what complexes the accelerator and carries sulfur into the polymer. The reaction is fast at 154 °C dry and at about 75 °C in the presence of water, and the zinc stearate it forms melts near 114.5 °C.

Here is the purchasing point that most often goes wrong. Dosing zinc stearate (CAS 557-05-1) in place of ZnO plus stearic acid does not reproduce the activation. The soap has to be generated in situ at the particle surface, where it also coats and disperses the oxide; pre-made soap added as a powder does something else — it lubricates. Zinc stearate in a rubber recipe is a lubricant and release aid, and the 800 mesh grade is chosen for dispersion, not for activation. Buying it to replace the activator pair is a false economy that shows up as a slow, undercured compound with a normal-looking COA.

The second is the antioxidant, which is also a cure-curve variable. In one NBR study, adding 1 phr of 6PPD (CAS 793-24-8) moved ts2 from 1.31 to 2.61 minutes, tc90 from 3.83 to 5.97 minutes, CRI from 39.7 to 29.8 min−1, and MH from 23.60 to 20.39 lb·in. A protective additive bought a longer process window and charged crosslink density for it. TMQ (CAS 26780-96-1) is milder, and MB (CAS 583-39-1) is a known retarder — for MB, check T90 and the initial melting point (a good lot sits at or above 295 °C), not only the assay.

The third is stearic acid overdose. Above roughly 2–3 phr it stops being a co-activator and starts being a plasticiser and a bloom source: ML drops, cure slows, the surface whitens, and adhesion to brass or fabric suffers. The cheapest way to ruin a cure curve is one more scoop of the cheapest ingredient in the recipe.

What to put on the purchase order

  • Accelerator family and CAS, not just the trade name. CBS 95-33-0, TBBS 95-31-8, DCBS 4979-32-2 and MBT 149-30-4 are four different process windows. A blended or diluted lot shows it in the initial melting point and the sieve residue long before it shows in the cure curve.
  • Delayed-action status in writing. If you are buying for a thick section or a 170 °C press cure, the sulfenamide class is the specification, and substituting within that class changes ts2 by a factor you will notice.
  • Sulfur and accelerator as phr and as a ratio. "EV" without a number is marketing. 0.5 phr sulfur with 3.0 phr CBS and 0.8 phr sulfur with 2.0 phr CBS are both EV and neither ages the same.
  • A rheometer sheet with instrument and temperature on it. ODR or MDR, 160 or 180 °C, arc angle. A t90 without those three is not a specification.
  • Mooney t5 at 125 °C for anything that will be extruded. This is the number that predicts whether the line runs, and it is the one most often missing from the offer.
  • For ZnO: Pb ≤ 30 ppm and a controlled D50 of 0.3–0.6 µm. Tire and conveyor specs treat lead content as a compliance line, and agglomerates above 5 µm in the cured part are a dispersion failure, not a purity failure.

Selection in one page

If the problem isChange thisYou will pay in
Scorching in the extruderMove within the sulfenamide class (CBS to TBBS to DCBS), add a scorch retarder, lower the mixing temperatureLonger t90, lower output per press cycle
Cycle time too longHigher cure temperature, or a more active acceleratorNarrower scorch window, higher reversion risk
Modulus falls off after overcureMove from CV to Semi-EV or EV, add 1–2 phr of a thiosulfate anti-reversion agentDynamic fatigue and tear strength, plus additive cost
Undercure with a normal-looking recipeCheck ZnO and stearic acid first, then dispersion, then sulfurNothing — this is the cheap fix
White bloom on black rubberCut stearic acid and zinc loading, check zinc stearate additionsSome process safety, some release

Grades we supply

Accelerators: CBS (CZ), CAS 95-33-0; TBBS (NS), CAS 95-31-8; DCBS (DZ), CAS 4979-32-2; MBT (M), CAS 149-30-4; ZMBT, CAS 155-04-4; DPG, CAS 102-06-7; TRA (DPTT), CAS 120-54-7 as a sulfur donor for low-sulfur and EV systems; ZDEC (EZ), CAS 14324-55-1; ZDBC (BZ), CAS 136-23-2; PZ (ZDMC), CAS 137-30-4; TDEC, CAS 20941-65-5.

Activators and process aids: zinc stearate 325 mesh and 800 mesh, CAS 557-05-1.

Antidegradants for the same compound: 6PPD, CAS 793-24-8; TMQ, CAS 26780-96-1; MB, CAS 583-39-1.

For the chemistry behind the accelerator choice itself, see matching thiazoles, sulfenamides, dithiocarbamates and thiurams to your cure system; for the ageing side of the same recipe, selecting stabilisers by failure mode; and for the soap grades themselves, metallic stearates and the mesh-size decision. Send us the compound target — t90 at your cure temperature, Mooney t5 at 125 °C, and the section thickness — and we will tell you which family gets you there, and what it will cost in the other two numbers.