A molded rubber part with a thick section or an uneven wall thickness can come out of the same cure cycle with genuinely different properties depending on where in the part you test it — fully cured, or even over-cured, near the mold surface, while the core is still under-cured. The part passes as “one batch, one cure recipe” on paper while behaving as several different materials in practice.
Root Causes
Rubber compounds conduct heat far more slowly than the metal molds that cure them, so heat entering from the mold surface takes measurably longer to reach the geometric center of a thick section than it does the surface itself. At any given moment during the cure cycle, the surface and the core of a thick part are genuinely at different temperatures — and because vulcanization is a thermally activated chemical reaction, that temperature difference translates directly into a difference in cure state. Cure meters built around this exact relationship — ASTM D2084 (oscillating disk cure meter) and ASTM D5289 (rotorless/moving-die cure meter) — characterize how strongly a compound’s crosslinking rate depends on temperature and time, which is precisely the relationship that turns a temperature gradient into a cure-state gradient.
The practical result inside one molded part is a surface that reaches cure temperature first, completes its own cure, and — particularly in sulfur-cured natural rubber — can continue reacting into reversion (a documented loss of crosslink density with continued heat exposure past the cure optimum) well before the core has even crossed its own scorch and cure thresholds. The core, lagging in temperature, ends up under-cured relative to the surface: lower crosslink density, lower modulus, and reduced compression-set and heat-aging resistance compared with a properly cured section. This gradient scales with thickness and geometry — thick-to-thin transitions, ribs, and heavy corners are all locations where a single nominal cure time and temperature, correct for the part’s thinnest section, is by definition wrong for its thickest one.
What This Means for Raw Material Selection
Accelerator system choice determines how forgiving a compound is to this reality. A fast, non-delayed accelerator system — one whose rheometer curve rises steeply immediately after scorch onset — pushes the surface deep into cure, and potentially reversion, in the same window it takes the core to merely reach cure-onset temperature. Delayed-action accelerator classes such as sulfenamides are formulated specifically to give the compound an induction period before the cure reaction accelerates; that built-in delay gives a slow-heating core more time to reach temperature and start curing before the surface finishes, narrowing the surface-to-core cure-state gap on the same cure cycle rather than requiring a different one.
Sulfur dosing relative to accelerator dosing compounds this, because that ratio does not just set overall crosslink density — a peer-reviewed NMR study on natural rubber and polyisoprene networks found that raising the accelerator-to-sulfur ratio shifts crosslink formation toward shorter (mono- and disulfidic) bridges rather than longer polysulfidic ones, and shorter sulfidic crosslinks are the more thermally stable, less reversion-prone type. For a thick or geometrically complex part, that makes the sulfur/accelerator ratio a lever for how resilient the cure is across the temperature range the part will actually experience — mold-contact surface down to slow-to-heat core — not only a lever for total crosslink density.
Zinc oxide belongs in the same picture because it is the activator, not a passive filler, in this system: zinc oxide reacts with a fatty acid such as stearic acid to form zinc stearate, which is what activates the accelerator in the first place. Consistency in zinc oxide dosing and dispersion — not just the nominal phr charged — affects how sharply or gradually the cure reaction switches on, which is directly relevant to how well a given cure package tolerates a core that is still heating up while the surface has already reached temperature.
The practical takeaway for a buyer: for thick or complex parts, accelerator system and sulfur/accelerator ratio should be verified against an actual D2084 or D5289 rheometer cure curve run at the part’s real cure temperature, not assumed from a datasheet cure time generated on a thin standard test slab — a system that looks well matched under standard slab-test conditions can still open into a wide surface-to-core cure-state gradient once it is actually run in the part’s thickest section.
Sources: ASTM D2084-19, Standard Test Method for Rubber Property—Vulcanization Using Oscillating Disk Cure Meter; ASTM D5289-19, Standard Test Method for Rubber Property—Vulcanization Using Rotorless Cure Meters; “Influence of Sulfur-Curing Conditions on the Dynamics and Crosslinking of Rubber Networks: A Time-Domain NMR Study,” Polymers (MDPI), 2022, Vol. 14, Issue 4, Article 767; general rubber-vulcanization-science literature (e.g., Mark, Erman & Roland, eds., Science and Technology of Rubber, Academic Press) on cure kinetics, reversion, and accelerator delayed-action behavior.
