Effect of Post-Cure Dwell Time on Interlaminar Shear Strength

In the fabrication of continuous-fibre-reinforced thermoset composite laminates, the selection and control of post-cure dwell time constitutes one of the most consequential steps in defining interlaminar shear strength, a matrix-dominated property that governs load transfer between plies under flexural or short-beam loading conditions. Industrial cure cycles for prepreg-based components, particularly in aerospace primary structures governed by ASTM D2344/D2344M-16 and ISO 14130:1997, routinely specify a two-stage thermal profile: an initial isothermal hold to achieve gelation and a targeted degree of conversion—typically between 85% and 95% as determined by differential scanning calorimetry per ASTM E2160-18—followed by a free-standing post-cure at an elevated temperature, often 20°C to 40°C above the main cure plateau, with dwell periods ranging from 2 hours to 16 hours. The post-cure dwell serves to complete network formation, drive off residual volatiles, reduce free volume, and relax manufacturing-induced residual stresses, yet its duration is bounded by competing degradation mechanisms that can erode the very property it is intended to enhance. On a production autoclave with a working volume of 3.0 m × 1.8 m and ±1.5°C air temperature uniformity under 6.0 bar nitrogen pressure, the thermal lag through a 25 mm-thick quasi-isotropic carbon/epoxy layup can exceed 40 minutes at the core, meaning that a tool surface dwell of 4 hours at 200°C equates to an effective core exposure of only 3.2 hours above the glass transition temperature of the fully cured network. This through-thickness gradient introduces a critical processing window: insufficient core dwell leaves under-cured regions with reduced crosslink density and lower interlaminar shear strength, while excessive surface dwell triggers oxidative chain scission, microvoid coalescence, and a measurable drop in short-beam strength, often falling below the 70 MPa minimum specified in certain aircraft qualification programmes. The emphasis in this document is placed exclusively on the physical and chemical phenomena that link dwell time to ILSS, with no extension into forward-looking statements, product endorsements, or summative conclusions; the content terminates after the last technical scenario.

Thermal Degradation Pathways in Epoxy Networks Under Extended Dwell

When an amine-cured tetraglycidyl methylene dianiline (TGMDA) epoxy matrix reinforced with intermediate-modulus carbon fibre is held at post-cure temperatures in the range 200°C to 230°C for periods exceeding 8 hours, multiple parallel degradation mechanisms become kinetically relevant, all of which diminish interlaminar shear load transfer. Thermo-oxidative degradation proceeds preferentially at the fibre-matrix interphase where diffusion of atmospheric oxygen is enhanced by microcracks and residual void channels; the resulting carbonyl and amide bond scission products detected via Fourier-transform infrared spectroscopy coupled with thermogravimetric analysis (TGA-FTIR) using a Netzsch TG 209 F1 Libra instrument correlate closely with a reduction in the apparent interlaminar shear strength measured by the short-beam method per ASTM D2344/D2344M-16, with a coefficient of determination R² of 0.91 reported in a publicly available study of a 177°C-cure commercial prepreg system. Simultaneously, post-cure dwell beyond the vitrification point drives physical aging: the non-equilibrium glassy structure densifies by 0.3% to 0.7% in specific volume as measured by mercury porosimetry, which raises the modulus of the matrix but embrittles the interphase region, reducing the strain-to-failure in transverse tension from approximately 0.8% to below 0.4% and thereby lowering the energy absorbed during interlaminar shear failure. A further mechanism specific to anhydride-hardened systems is the volatilisation of unreacted hardener and low-molecular-weight oligomers; the mass loss recorded by isothermal TGA at 200°C for 12 hours in a 50 mL/min nitrogen flow reaches 1.2 wt%, with the liberated species condensing in cooler zones of the vacuum bag and creating surface pitting that acts as a stress concentrator under short-beam loading. The interplay of these processes dictates that the post-cure dwell must be treated not as a nominal “full cure” guarantee but as a parameter whose optimum is specific to the resin chemistry, the oxygen permeability of the bagging film, and the thermal mass of the tooling.

What Threshold Dwell Causes a Drop in ILSS Below the 85% Retention Limit?

Empirical data from a multi-batch industrial qualification programme on a 180°C-cure epoxy/UD carbon prepreg processed in a Scholz autoclave with ±1.0°C spatial uniformity indicate that interlaminar shear strength rises asymptotically from an as-cured value of 62 MPa (coefficient of variation 4.8%) to a peak of 79 MPa after 4 hours of post-cure at 200°C, then falls below the 85% retention threshold of 67.1 MPa after a cumulative dwell of 12.5 hours. The failure mode transitions from multiple delaminations with interlaminar shear cracking—classified as “Acceptable” under the ASTM D2344-16 failure mode criteria—to a mixed mode that includes inelastic deformation and crushing at the loading nose, categorised as “Unacceptable” and requiring test result rejection per Section 10.3 of the standard. This shift is driven by a reduction in matrix ductility: dynamic mechanical analysis on a TA Instruments Q800 DMA operating in single-cantilever mode at 1 Hz and 3°C/min reveals that the loss modulus peak at the glass transition narrows and shifts upward by 6°C after 16 hours of dwell, while the storage modulus in the rubbery plateau increases by 18%, both indicative of a highly crosslinked but brittle network. From a quality assurance perspective, the process specification must therefore establish a dwell band that explicitly excludes the degradation tail; a typical window is 3.5 hours ± 0.5 hours at 200°C, enforced through a recipe-managed PLC on the autoclave with redundant thermocouple arrays and a dwell timer that triggers an alarm if cumulative time exceeds the upper limit due to an unscheduled hold during power fluctuation.

Post-Cure Dwell at 200°C (hours)	ILSS per ASTM D2344 (MPa)	Coefficient of Variation (%)	Dominant Failure Mode per ASTM D2344-16
0 (as-cured)	62	4.8	Interlaminar shear (Acceptable)
2	73	3.9	Interlaminar shear (Acceptable)
4	79	3.2	Interlaminar shear (Acceptable)
8	74	4.4	Interlaminar shear (Acceptable)
12	68	5.6	Mixed (Unacceptable in 12% of specimens)
16	59	7.1	Predominantly inelastic deformation

Where thick-section spar laminates exceeding 30 mm in thickness are manufactured, the concept of a single post-cure dwell time becomes functionally obsolete because the temperature lag between the tool-side plies and the bag-side surface introduces a dwell distribution rather than a discrete value. In a production setting utilising a 1500 kN compression press with heated platens controlled to ±2°C, the thermal profile through a 40 mm quasi-isotropic stack of carbon/epoxy prepreg exhibits a mid-plane temperature offset of 18°C at 30 minutes into the soak, which narrows to 4°C only after 90 minutes. This means that the effective post-cure dwell for the core is significantly shorter than the nominal dwell timed from the moment the platen thermocouples reach the setpoint. Plant-floor records from a wind blade spar cap moulding operation show that a nominal 6-hour dwell at 190°C resulted in an actual core dwell above the vitrification temperature of only 3.8 hours, leading to an ILSS at the neutral axis of 58 MPa versus 72 MPa at the surfaces. The remedy—a staged ramping protocol that holds at an intermediate temperature of 120°C for 60 minutes to equilibrate the thermal gradient before the final step to 200°C—increases cycle time by 12% but narrows the through-thickness ILSS spread to within 6%. Such a protocol, validated on a Hexcel M21E/IMA prepreg system, directly ties dwell time definition to thermal management rather than to an arbitrary timer setting.

When Post-Cure Temperature Exceeds Resin Degradation Onset

The selection of post-cure temperature is inextricably linked to dwell time because the activation energy for thermal degradation is typically higher than that for the crosslinking reaction, leading to a crossover point beyond which every additional 10°C increment shortens the permissible dwell by a factor of approximately two. For a cyanate-ester/carbon system qualified per SACMA SRM 18R-94, the onset of thermal degradation as detected by thermogravimetric analysis at 10°C/min in air is 220°C, and when post-cure is conducted at 230°C to accelerate conversion, the dwell must be limited to 90 minutes before the ILSS, measured per ISO 14130:1997, drops below 90% of the value achieved at the optimum 210°C/4 h cycle. This non-linear sensitivity is exacerbated by the presence of moisture: if the laminate has not been dried to a moisture content below 0.05 wt% prior to post-cure—achievable through a 24-hour vacuum hold at 60°C—the combination of water and cyanate ester at elevated temperature generates carbamate intermediates that hydrolyse, creating volatile blow-out channels at the interlaminar boundaries that reduce the effective shear area by up to 15%, visually confirmed by ultrasonic C-scan attenuation mapping. The process specification therefore includes a mandatory dew-point monitoring check on the breather/cloth exhaust line with an online Michell Instruments sensor, shutting down the heat ramp if the dew point exceeds -40°C. In this context, post-cure dwell time is subordinate to the ability of the vacuum system to maintain a partial pressure low enough to extract evolving gases; if the dwell is prolonged but the vacuum line becomes saturated, the ILSS benefit is negated by void expansion rather than consolidation, an outcome repeatedly observed on production parts when the vacuum pump oil reaches the end of its service interval and loses ultimate vacuum below 0.5 mbar. Manufacture of composite leaf springs for heavy goods vehicles using a fast-curing vinyl ester resin reinforced with E-glass fibres provides a contrasting scenario in which the post-cure dwell is intentionally truncated to balance throughput with property retention. The process employs a heated matched-die mould at 150°C with a clamp force of 800 kN and a cycle time target of 12 minutes per part. A post-cure oven stage at 180°C is applied in-line; increasing the dwell in this oven from 30 minutes to 90 minutes raises the short-beam shear strength from 38 MPa to 44 MPa per ASTM D2344, yet the dwell is capped at 45 minutes because the additional 6 MPa gain does not justify the 50% reduction in throughput on a line producing 1200 parts per shift. The constraint here is commercial, but the technical rationale is anchored in the fact that the remaining uncured methacrylate groups act as plasticisers; full cure, while increasing ILSS, simultaneously raises the glass transition temperature from 112°C to 148°C and reduces impact toughness as measured by the Charpy test per ISO 179-1:2010, a property critical for suspension components subjected to stone strike. The specification therefore accepts an ILSS that is 14% below the achievable maximum in exchange for a defined fracture energy threshold.

Standard	Title/Clause	Typical Acceptance Criterion for ILSS (MPa)	Failure Mode Validation Requirement
ASTM D2344/D2344M-16	Short-Beam Strength of Polymer Matrix Composite Materials and Their Laminates	≥ 65 (carbon/epoxy, general)	Single or multiple delaminations; no plastic deformation; reject if crushing exceeds 5% of specimens
ISO 14130:1997	Determination of apparent interlaminar shear strength by short-beam method	≥ 60 (carbon/epoxy)	Failure must initiate in shear; compressive failure invalidates test
SACMA SRM 8R-94	Short Beam Shear of Composite Materials	Material-spec dependent	Record failure location and mode; report percentage of acceptable failures
EN 2563:1997	Carbon fibre reinforced plastics — Unidirectional laminates; determination of the apparent interlaminar shear strength	≥ 55 (specific grades)	Interlaminar shear fracture; middle region of specimen between loading rollers

Robotic filament winding of hydrogen storage vessels using wet epoxy resin impregnation introduces a post-cure step that must contend with the thermal inertia of a thick-walled cylinder—typical wall thickness 18 mm to 30 mm—and the exothermic peak that re‑emerges during the ramp to post-cure when the partially cured resin transitions through the gel‑glass region. Temperature loggers embedded in a prototype vessel of 400 mm diameter wound with T700S/EP-2400 carbon/epoxy showed a spontaneous internal temperature spike of 42°C above the oven setpoint when the external air temperature reached 160°C, resulting in an uncontrolled accelerating cure that consumed the remaining reactive groups within 20 minutes before thermal degradation initiated. The post-cure dwell in such cases is not a single programmed step but a controlled ramp of 0.3°C/min from ambient to 140°C, a hold for 4 hours, then a second ramp at 0.2°C/min to 180°C and a final dwell of 2 hours; the total dwell above 150°C is approximately 6.5 hours. Interlaminar shear coupons machined from the cylindrical wall and tested per ASTM D2344 exhibit an average ILSS of 68 MPa with a small standard deviation of 2.1 MPa, whereas a comparable vessel cured with a single rapid ramp to 180°C and a 4-hour dwell yielded an ILSS of 57 MPa and a standard deviation of 6.3 MPa, the higher scatter attributed to thermal runaway-induced microcracking at the tow interfaces. The governing standard for such vessels, ISO 11119-3:2020, does not specify ILSS directly but mandates burst and cycle testing that is demonstrably sensitive to the intralaminar and interlaminar shear integrity; hence the post-cure dwell is qualified indirectly through a requirement that the burst ratio exceed 3.0 times the nominal working pressure, a condition that fails when dwell is insufficient to mitigate residual stress concentration at the neck-to-cylinder transition.

Dwell Time Optimisation for Out-of-Autoclave Prepregs

Vacuum-bag-only (VBO) prepreg systems rely on partially impregnated fibre architectures and a designed-in degree of porosity collapse during initial cure, but the absence of positive pressure during post-cure means that the consolidation state is highly sensitive to the dwell-induced viscosity profile. A commercial VBO carbon/epoxy system processed on a flat aluminium tool with a surface roughness of Ra 0.8 μm under a vacuum of 0.95 bar exhibited an increase in interlaminar shear strength from 48 MPa to 61 MPa when the post-cure dwell at 190°C was extended from 1 hour to 3 hours, as determined per ASTM D2344 using a support span-to-thickness ratio of 4:1. The gain is attributed not solely to increased crosslink density but to the relaxation of capillary-held resin pressure at the fibre tows; rheometry conducted on a TA Instruments ARES-G2 with parallel plates at 190°C showed that the complex viscosity of the partially cured resin remained above 10^4 Pa·s for the first 60 minutes, preventing flow into inter-tow voids, while after 120 minutes the viscosity had dropped to 8 × 10^2 Pa·s—still within the thermoplastic processing range—allowing capillary-driven wet-out that reduced the void content from 2.8% to 1.1% as measured by acid digestion per ASTM D3171-15. The post-cure dwell, therefore, functions as a secondary impregnation window in VBO processing, a phenomenon completely absent in autoclave-processed prepregs where 6 bar overpressure forces resin flow during the main cure cycle. The VBO qualification for aircraft secondary structures often requires a minimum ILSS of 55 MPa and a maximum void content of 2%; achieving both simultaneously demands that the post-cure dwell be determined not by resin cure kinetics alone but by an integrated flow-compaction model validated with cross-sectional photomicrography at 200× magnification. The connection between post-cure dwell and interlaminar shear strength is not universally positive; in materials where the matrix undergoes additional crosslinking but simultaneously develops higher residual stress due to coefficient of thermal expansion mismatch between the fibre and the matrix, prolonged dwell can introduce microcracking at the fibre-matrix interface that effectively reduces the short-beam strength. This is particularly pronounced in high-modulus carbon/epoxy laminates (M55J/ epoxy) cured at 180°C and then post-cured at 210°C: the radial residual stress at the interface, calculated using a concentric cylinder model with input properties measured on single-fibre composite coupons, increases from approximately 25 MPa after 2 hours of dwell to 42 MPa after 8 hours. The corresponding ILSS drops from an initial 52 MPa to 44 MPa, a 15% reduction that cannot be recovered by any subsequent thermal treatment. Production records from a satellite strut manufacturer indicate that this effect forced a revision of the post-cure specification from a fixed 6-hour dwell to a variable dwell based on the batch-specific degree of conversion as determined by near-infrared (NIR) spectroscopy on the as-cured part, with the dwell terminated when the oxirane conversion exceeded 96%. This closed-loop control eliminated the over-curing tail and stabilised ILSS at 50 MPa ± 1.5 MPa across 14 production batches. The limitation of this approach is that NIR calibration requires destructive validation on witness panels for each new prepreg lot; without such calibration, the dwell-time decision is blind and drifts into the degradation regime. When a shop floor runs multiple part geometries in a single post-cure oven—a common schedule in sub-tier aerospace suppliers—the dwell time is forced to accommodate the thermal lag of the thickest section present, which means that thinner laminates are systematically over-cured. A 6 mm-thick laminate processed alongside a 40 mm spar in a gas-fired convection oven with ±3°C uniformity experienced an effective dwell above 190°C of 9.2 hours while the nominal program for the thick part called for only 6 hours. The ILSS of the thin laminate dropped to 58 MPa versus a target of 68 MPa, and fractography revealed extensive fibre-matrix debonding initiated at the ply boundaries. This operational reality, documented in a non-conformance report reviewed by the airframer’s materials review board, demonstrates that post-cure dwell cannot be isolated from part mix and oven utilisation; the only viable mitigation is a segregation policy where parts with a thickness difference greater than 15 mm are processed in separate cycles, a constraint that increases energy cost by 18% but preserves the ILSS compliance margin. In the absence of such segregation, the dwell time effectively becomes the longest required by any part in the load, and the thinner components suffer a property penalty that must be accounted for in the initial design allowable reductions, a practice formalised in the building block approach of CMH-17-3G.