Release control in anhydrous lip balm molding is determined primarily by the thermal path imposed on the molten wax-oil mass after the dosing nozzle closes. On continuous lines using 12- to 36-cavity silicone or coated aluminum molds, a 4.5 g stick in a 15 mm diameter cavity cools through the congealing range of the wax phase while the mold is transported through a divided air-impingement tunnel. The release event at the end of the tunnel is not simply a surface-lubricity problem; it is the mechanical result of radial shrinkage, crystal network formation, and interfacial adhesion between the cooled stick wall and the mold surface. In-mold thermocouples placed at the fill centerline and 2 mm from the wall, logged at 0.5 s intervals with 0.1°C resolution, reveal that the cooling segment between 60°C and 25°C produces the largest differentiation in release force. Below this band, the material has developed a solid-like shell; above it, the material remains sufficiently fluid to reflow after partial demolding. Published data for this specific configuration is limited because mold release force is not covered by a single harmonized test standard. However, cooling tunnel manufacturers' operating manuals, instrument calibration records under ISO 17025:2017, and process control logs under ISO 9001:2015 clause 8.5.1 support the use of cooling curve parameters as release-control variables.

Which Cooling Curve Variables Exert the Greatest Leverage on Mold Release Force?

The variables with the greatest leverage are the cooling gradient from activation of the cooling tunnel to the end of the crystallization plateau, the duration of dwell at the congealing onset temperature, the temperature differential between mold wall and stick core at demolding, and the discharge temperature after final chill. The cooling gradient is best measured with a thermocouple cast into a sacrificial stick at the mold sidewall. For candelilla-rich bases, a gradient steeper than 5 K/min between 60°C and 25°C produces a fine, poorly interlocked wall layer that tends to smear against the mold sidewall during ejection. Data from pilot-scale mold trials, conducted with a reciprocating 24-cavity aluminum mold and a force-deflection texture analyzer operating at a pull speed of 5 mm/s, indicate that release force drift at startup correlates more strongly with mold wall temperature than with air setpoint. The absence of a standardized release-force method means that interlaboratory comparison requires that the fixture geometry, pull angle, pull speed, and mold conditioning history be reported alongside the value. Methods aligned with ASTM D1321-20 for needle penetration and ASTM D3418-21 for crystallization and melting transitions are acceptable for raw material qualification, but neither standard directly predicts mold release force because the mold constraint and surface energy contribution are absent.

When mold release is framed as a shrink-fit problem, the controlling material property is the volumetric contraction developed between the congealing point of the dominant wax phase and the demolding temperature. In candelilla wax, the congealing point typically falls between 68°C and 72°C by ASTM D938-20, but the crystal network continues to densify below this range. Because the mold wall constrains radial contraction, the stick maintains interfacial contact until the cumulative shrinkage exceeds the surface roughness contact area and the mold-coating adhesion threshold. If the product is ejected before this point, the top edge of the stick tears because the central core remains warmer and more ductile than the peripheral shell. Temperature mapping in a mold cavity with thermocouples placed at 2 mm, 6 mm, and 12 mm from the top surface has recorded a vertical gradient of 4°C to 7°C across the fill height during unidirectional bottom cooling. This gradient is often neglected in single-point mold temperature checks but directly controls local shrinkage and the release path. The practical consequence is that releasing the stick from the warmer top side produces higher shear at the surface, often observed as circumferential drag lines under low-angle illumination.

Three-Zone Air Impingement and Recalescence Control

Divided cooling tunnels on production lines typically separate the cooling path into a first mild zone, a second recalescence-control zone, and a third final-chill zone. In the first zone, air at 15°C to 20°C and a velocity of 1.5 m/s to 3.0 m/s reduces surface temperature without shock-nucleating a thick amorphous skin. In the second zone, the air temperature is often lower, but the product remains near a latent-heat plateau because the crystallization exotherm balances heat removal. The recalescence event is visible in the cooling curve as a temporary flattening or slight temperature rise of 2°C to 4°C in the core; failing to account for this event causes zone timers to be set too short, so the product reaches final-chill before the crystal network has formed. When the third zone then drops the stick below 10°C, the outer shell contracts onto the mold while the still-mobile central material is pulled away from the bottom, producing stress cracks that later appear as axial splits after ejection. This is a critical threshold zone because a process window narrower than ±5°C in zone-2 return air temperature is typical for wax-rich bases with low oil content. Condensation on mold surfaces becomes an operator-independent release disruptor when the packaging room dew point exceeds the mold surface temperature; pre-drying of tunnel supply air to a pressure dew point below 0°C is therefore specified when ambient relative humidity is above 60%.

At the formulation level, the release behavior is governed by the crystallization tendencies of the wax phase. Candelilla wax forms a hard, brittle network with a narrow congealing interval, while carnauba wax shows a higher drop melting point near 82°C to 86°C by ASTM D127-19 and contributes gloss but can raise mold adhesion if it segregates to the wall during initial cooling. Beeswax-rich bases present free fatty acid groups that can complex with unanodized aluminum mold surfaces, causing organic aluminum carboxylate residue and pitting after repeated batches; this failure mode is minimized by using anodized or PTFE-coated cavities or by replacing beeswax with high-triglyceride candelilla. Microcrystalline wax grades with congealing points between 60°C and 90°C by ASTM D938-20 have higher oil-binding capacity but lower volumetric shrinkage, so their release force remains high unless the mold is allowed to dwell at the demolding temperature for an extended period. Polyethylene wax increases drop melting point to 100°C to 140°C and shifts demolding into a brittle regime; discharge below 25°C can cause longitudinal fracture because the part behaves as a rigid body with minimal strain-to-break. Published data for this specific configuration is limited; the indicated ranges come from supplier technical bulletins and process development records rather than a universal specification.

When Mold Release Force Exceeds the Yield Strength of the Solidified Shell, Ejection Damage Becomes Statistically Detectable

The threshold at which ejector-pin force exceeds the compressive yield strength of a partially solidified stick is line-specific, but the damage visible at the top annular rim can be quantified with a constant-rate texture analyzer. An ejection pin with a diameter of 3 mm bearing on the bottom face of a 4.5 g stick can generate localized stress above 0.2 MPa when release force exceeds 2 N. At that point, the top rim tears because the peripheral shell is only partially load-bearing. In a 24-cavity mold, cavity-to-cavity variation in release force is often non-normal and bimodal; the upper tail consists of cavities that show delayed crystallization or wall adhesion. Gauge repeatability and reproducibility studies aligned with ISO 17025:2017 show that release-force measurement variation is minimized if the mold is conditioned for a minimum of 20 cycles before data collection. Needle penetration values by ASTM D1321-20 at 25°C are used as an indirect mechanical consistency check, but a direct release fixture is required because penetration does not capture mold-wall adhesion. For candelilla-based sticks, penetration below 20 dmm at 25°C is often associated with brittle ejection when combined with a short dwell time below the congealing interval. That combination can be managed by lowering the first-zone cooling gradient and increasing the recalescence dwell, not by raising the air temperature alone.

Cooling Curve Logging Must Capture Recalescence Events

Data-logging protocols that record only air temperature fail to capture the mold-side recalescence that determines release force. A usable cooling curve for release control includes the fill temperature, the time to reach the wax congealing onset, the duration of the latent-heat plateau, the minimum core temperature before ejection, and the discharge temperature of the stick after ejection. The recalescence plateau in a candelilla-rich base may be visible at 32°C to 38°C and may last 10 s to 25 s in a mold cavity; missing this plateau in the logged trace causes incorrect assignment of the crystallization endpoint. For process traceability, thermocouple calibration is performed against a certified dry-block calibrator with an uncertainty of ±0.2°C. Infrared spot pyrometry is used only for non-contact mold surface checks under ASTM E2758-10; emissivity adjustments for silicone and anodized aluminum mold surfaces require separate calibration curves. A cooling rate calculated from the slope between 60°C and 25°C is not equivalent to the air setpoint reduction because the product load and latent heat interact with tunnel air velocity. Production records should therefore use product-side thermocouples, not inlet-air thermocouples, to establish release control limits.

Beyond the instrumented mold, the release behavior of the finished stick is sensitive to the post-ejection thermal history. Isothermal holding of the demolded stick at 18°C to 22°C for 4 h to 12 h before packaging allows polymorphic densification and reduces subsequent surface oiling. This holding step does not influence mold release force directly because release has already occurred, but it reduces dimensional recovery that can be misinterpreted as release-induced deformation. In-line checks of stick diameter after 24 h in a controlled cabinet should be compared against the mold cavity diameter to separate immediate mold friction from delayed shrinkage. Mold surface preparation interacts with the cooling curve: a surface roughness of 0.4 µm to 0.8 µm Ra typically provides sufficient microtexture to reduce flat contact without creating mechanical interlocking. Polished cavities with roughness below 0.1 µm Ra show longer release force transients after cleaning because the stick sidewall conforms to the smooth surface. Surface roughness parameters are measured according to ISO 4287:1997. Silicone rubber cavities release more easily than unanodized aluminum but have lower heat transfer coefficients; therefore the cooling curve must be extended or the mold temperature lowered. Release agents such as medium-chain triglyceride sprays lower release force but migrate into the anhydrous base; migration is measurable by gas chromatography after extraction, and residual levels must remain within the formulation's cosmetic safety dossier under EU Cosmetics Regulation 1223/2009. The use of volatile silicone release sprays is limited by health and environmental restrictions under REACH Regulation 1907/2006, Annex XVII. Avoid combining free fatty acid-rich beeswax with unanodized aluminum molds because aluminum carboxylate residues form at the interface; if beeswax is required, use anodized or coated cavities and monitor surface pH after cleaning.

Because wax composition controls the position and width of the congealing plateau, a single cooling curve is not universally transferable. The following table compiles the operational boundaries recognized in supplier technical bulletins and production records; it is a line-specific comparison rather than a release-force specification.

On scale-up from a bench mold to a continuous tunnel, the cooling curve changes shape because the mold wall thermal mass and conveyor speed alter the effective cooling rate. A bench mold chilled in a refrigerator may show release force below 1 N, while the same formula in a 24-cavity line with air impingement exhibits release force above 2 N when the mold has not reached steady state. This discrepancy is caused by differences in heat flux at the bottom plate and by the changing surface temperature of the mold as it cycles. In the production line, mold temperature at the point of fill is influenced by the preceding mold cycle; therefore the first 20 to 50 cycles after startup usually produce release outliers. This startup effect is more pronounced on aluminum molds than on silicone rubber molds because aluminum cools faster and reaches a lower surface temperature before the first shot. A defined startup protocol, including mold preheating and rejection of early shots, is a necessary part of release control. Process windows narrower than ±5°C in zone-2 air temperature require that the tunnel controllers be tuned with less than 1°C overshoot; older on-off controllers may allow temperature swing outside that window and create intermittent release failure on the next cycle. Proportional-integral-derivative control with a 2 s scan time is typical for maintaining this narrow band.

The melt viscosity of the anhydrous base at the dosing nozzle affects the initial wetting of the mold and therefore the adhesion footprint. Viscosity can be measured by a rotational rheometer with a cone-plate geometry at 75°C and a shear rate of 10 s⁻¹; values between 0.1 Pa·s and 2.0 Pa·s are common for pour-fill lip balm bases. The measurement procedure is aligned with ISO 3219:1993. Lower-viscosity fills wet the mold more completely and can increase release force despite easier filling. High-viscosity fills may trap air at the mold bottom and produce release lines on the lower face. The fill temperature should be selected to place the melt viscosity in a range that gives a complete cavity fill without excessive wall wetting; this balance is line-specific but can be checked by measuring the contact angle of the molten base on the mold material using a goniometer under ASTM D7334-08. Contact angle below 30° on the mold material generally corresponds to higher wetting and greater adhesion; contact angle above 60° corresponds to poor filling and void defects. The mold surface therefore must be specified together with the cooling curve, not separately.

Amide-containing wax interactions also create measurable release features. Erucamide, when added as a surface slip modifier, blooms to the stick surface during the final cooling stage and can reduce mold release force, but if the cooling gradient exceeds 5 K/min the bloom is non-uniform and appears as white streaks. Attenuated total reflectance FTIR with a germanium crystal identifies the amide carbonyl near 1640 cm⁻¹ and 1648 cm⁻¹; quantification by calibration to a known erucamide film provides a release-film uniformity check. Because erucamide is a primary amide, combining it with free fatty acid-rich bases at elevated fill temperatures can promote surface reactions that alter mold wetting. Published data for this specific interaction in anhydrous lip balm bases is limited; the observation is derived from general amide-fatty acid chemistry and production-scale mold-release records. In addition, formulations that contain transition-metal stearates for pigment dispersion should not be paired with unanodized aluminum molds under high-humidity cooling air, because galvanic differences accelerate mold surface oxidation and release-force drift.

High-oil-content bases create a specific processing conflict. The cooling curve plateau is shorter and less defined because the liquid oil fraction suppresses bulk crystallization, yet the mold release force often rises because the solidified wax shell is thinner and the liquid oil fills the interfacial contact area. In a base containing more than 30% by mass of medium-chain triglycerides, the stick may not shrink enough for release until it is cooled below 15°C, at which point the exterior becomes brittle and the interior remains fluid. This conflict is managed by lowering the cooling gradient to 2 K/min and by discharging at a surface temperature near 18°C after an in-mold dwell of 5 min. The exact percentages and temperatures are line-specific; processing records from a twin-lane seeder line show that release force outliers increase when the oil phase is not fully incorporated before filling. Oil pockets are created if the wax is not fully melted and dispersed at a temperature above the highest-melting wax's drop melting point. A jacketed mixing kettle with a bottom anchor blade at 25 rpm to 50 rpm is typical for batch consistency. The batch should be held at 85°C to 90°C for candelilla-carnauba bases for at least 30 min before pouring, then cooled to the fill temperature.

Validation of a changed cooling curve must be performed with a blocked design that holds wax batch, fill temperature, and mold type constant. At minimum, three cooling gradients are tested: a slow profile near 1 K/min, a moderate profile near 3 K/min, and an aggressive profile near 6 K/min from 60°C to 25°C. Release force is measured on at least 50 consecutive sticks per profile, with cavity identification recorded. The upper control limit is set at the 95th percentile of the release-force distribution unless the observed distribution is bimodal. A bimodal distribution indicates that two cooling paths exist in the mold, usually caused by uneven air impingement across the tunnel. The test records the maximum pull force, the displacement at maximum force, and the residual surface damage graded under a 3× microscope. Damage scoring should be compared against a physical set of retained sticks from an approved master batch, stored in sealed glass jars at 20°C to avoid surface condensation.

Differential scanning calorimetry under ASTM D3418-21 at a cooling rate of 10 K/min is used for material characterization, but the cooling rate in a mold is often below 1 K/min to 5 K/min; therefore the DSC onset temperature is not identical to the in-mold congealing temperature. Process engineers should overlay the in-mold cooling curve with the DSC-derived crystallization exotherm to identify whether the tunnel cooling capacity is sufficient to maintain the plateau without excessive air-chill. If the in-mold trace shows a sharper recalescence spike than the DSC trace, that indicates nucleation was delayed at the wall and release variance may follow. This type of overlay is more predictive of release behavior than single-point mold temperature because it captures the time spent in the transition region where crystallite size and interfacial adhesion are set.