No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Surface Porosity in Out-of-Autoclave Prepreg Processing: Causes and Reduction Strategies
Abstract
Composite laminates cured out-of-autoclave (OOA) using vacuum-bag-only (VBO) prepregs can sometimes exhibit significant surface porosity on the tool-side surface when using only a liquid release agent. This porosity may be eliminated by using a release film or a resin-rich surfacing ply between the tool and part during cure, or by post-cure operations such as gel-coating and painting. Because such solutions can add significant time and cost to the overall manufacturing process, new strategies to inhibit surface porosity need to be developed. However, the mechanisms of void formation and tool-part interactions that lead to such surface porosity are not fully understood. The following paper presents a parametric study that clarifies the underlying causes of surface porosity in OOA-VBO carbon fiber-epoxy laminates and determines the relative importance of key manufacturing parameters, including tool plate surface topology, release methodology, prepreg material, and room-temperature vacuum hold time. Laminates are manufactured under various conditions using traditional OOA-VBO protocols. Then, optical microscopy and image analysis are used to characterize the surface porosity in terms of pattern and area. Thus, these parameters are related to expected part quality.
... Recent studies have suggested that surface porosity results from air that becomes trapped between the prepreg and tool plate during layup [6] [7] [8]. In a previous study, we reported that composite laminates made from woven prepreg exhibit more surface porosity than laminates made from unidirectional (UD) prepreg, an observation attributed to the initial morphology of the prepreg [6]. ...
... Recent studies have suggested that surface porosity results from air that becomes trapped between the prepreg and tool plate during layup [6] [7] [8]. In a previous study, we reported that composite laminates made from woven prepreg exhibit more surface porosity than laminates made from unidirectional (UD) prepreg, an observation attributed to the initial morphology of the prepreg [6]. Air entrapment is less likely with UD prepregs because of the more uniform surface topography. ...
In this study, we employ a parametric approach coupled with surface analysis to identify the source(s) of surface porosity and to develop effective mitigation strategies. Results confirmed that surface porosity was primarily associated with air that was trapped at the tool-prepreg interface during layup. The magnitude and distribution of surface porosity was affected by multiple parameters, including vacuum hold time, freezer and out time, and material and process modifications that affect air evacuation. The results indicated that prepreg out time (and thus tack) and vacuum quality were the primary drivers of surface porosity; for example, surface porosity decreased by 83% after just four days of out time and by 99% after 14 days of out time. These factors were used to formulate guidelines to mitigate surface porosity by (a) increasing the driving force for air evacuation and/or (b) increasing the permeability of the tool-prepreg interface.
... In contrast, UD prepregs typically exhibit a smooth surface (12.7 μm). This is attributed to different fiber architectures of the woven and UD prepregs [316]. The weaving patterns of woven fabrics are such that there are regions (overlaps and underlaps) for easy air entrapment; this phenomenon has been studied using XCT [240]. ...
Optimization of process parameters of major aerospace composites manufacturing techniques is essential for manufacturing high-quality aerospace components. However, studying many parameters through experimental approaches is time-consuming and labor-intensive. Modern X-ray Computed Tomography (XCT) technology has made it possible to monitor the complete manufacturing process “in-situ,” due to its high spatial resolution and very short acquisition time. The scanned data can then be used to generate numerical models for predicting different process parameters. This article presents a comprehensive review of the design and modeling aspects of key aerospace composites manufacturing techniques using the XCT procedure. Key functioning of XCT is reviewed with a view to adopt several modeling approaches ranging from mesoscale to macro-scale models for composites processing. The XCT machine parameters during image acquisition and post-processing determine the accuracy and computation time, which are also discussed in detail. The applications of XCT-based material models are reviewed for process characterization of autoclave and out-of-autoclave manufacturing. Current aerospace manufacturing characterization techniques, combined with advanced modeling approaches using XCT and CFD tools, show great potential for designing modern material models that will help in the design of better manufacturing processes for next generation aerospace parts.
... Furthermore, it has been proposed that air entrapped between the first ply and the tool during layup is a major contributor to the issue of surface porosity in cured laminates. 7 Surface porosity or pitting may not have a large structural effect, but acceptance criteria tend to be strict regarding this issue, as they compromise the aesthetic appearance of components. As a result, rework and finishing operations are often required to improve appearance. ...
Air entrapped during the material deposition stage of prepreg layup can contribute to voidage in the cured laminate, which is a major factor in determining the quality of the resulting laminates and components. Removal of this entrapped air is therefore a priority to achieve maximum part performance. To date, there does not appear to have been any consistent study of the processes by which voids can be entrapped. This paper seeks to address this by presenting direct observations of the contact between a glass tool surface and the first ply of carbon fiber prepreg laid down onto it. The entrapment of air between the tool and ply is investigated, in terms of its distribution and evolution during debulk under vacuum at room temperature. A counterintuitive inverse relationship between initial area of contact and the final level of entrapped air was observed, with higher levels of initial contact leading to higher final levels of entrapped air. An initial examination was also carried out to review the effects of ply terminations and bagging practices on air entrapment.
Prepreg format plays a key role in part quality for composites produced using vacuum bag only (VBO) techniques. To date, however, VBO prepregs have been produced by modifying existing autoclave formats. In this work, we introduce USCpreg, a prepreg format designed specifically for out-of-autoclave cure, featuring through-thickness permeability. We describe the fabrication and analysis of laminates processed with USCpreg, as well as laminates fabricated from traditional VBO prepreg formats. The through-thickness pathways for air transport in USCpreg result in near-zero internal porosity and defect-free surfaces in parts cured under VBO conditions, even under challenging processing conditions. Results highlight the fact that surface and internal porosity depend on prepreg format, and that through-thickness permeability is critical to achieving high quality parts in non-ideal manufacturing scenarios.
The increasing use of advanced composite materials within the aerospace industry and in other emerging fields motivates the development of cost-effective and environmentally friendly manufacturing methods. Out-of-autoclave (OoA) prepregs and vacuum bag-only (VBO) cure can improve on traditional autoclave processing and enable further non-autoclave process avenues such as heated tool cure. This paper describes a multi-national academic research program focused on sustainable manufacturing using OoA prepregs. Four major areas are addressed: defect reduction, process modeling, materials efficiency and advanced tooling. For each, major activities are surveyed, and their potential contribution to sustainability is discussed.
The problem of porosity has been addressed in relation to process modelling of composite materials. The material and processing inter-relationships, critical to the production of a void-free laminate, have been investigated and it is concluded that entrapped bubbles can be collapsed and growth processes can be suppressed, or totally inactivated, by control of pressure and temperature, the principal process variables. As these processes are time-dependent, models based on mass diffusion theory have been developed which predict the rate of growth or collapse of gas bubbles in a resin precursor. Simple experimental methods have been developed to measure the required input parameters for the model and the predictions have been compared with experimental observations.
Mass diffusion theory has been used to predict the rate of growth or collapse of gas bubbles in a resin precursor in order to aid the production of void-free composite materials. Simple models have been successfully compared with experimental data which show that bubble behaviour is influenced by the concentration of mobile species in the bulk resin, the concentration at the bubble/resin interface and the diffusion coefficient of these species in the resin precursor. These parameters are affected by temperature, pressure and the exposure history of the resin. Gas bubbles have been monitored and modelled for different temperatures, hydrostatic pressures and gas concentrations. The mass diffusion theory and the relationships between the processing variables and model input parameters can be used for predicting bubble behaviour during a cure schedule.
Various curing cycles were designed to evaluate the effects of different pressure induced voids on some mechanical properties of carbon/epoxy laminates and optimum cure pressure time. A series of [0/90]3s laminates with varying void contents were fabricated and a characterization of void distribution, size, and shape within the laminates was obtained using ultrasonic c-scan and optical metallography techniques. Short beam shear, three-point flexure and tensile testing were used for mechanical evaluation and the results correlated to void volume fraction and ultrasonic absorption coefficient. The experimental results have shown that there exists an optimum time for applying pressure. The proposed grading evaluation method of ultrasonic inspection is helpful for the nondestructive examination of composites. Also, changes in mechanical performance due to porosity are consistent with previously published studies.
An innovative approach to address surface pitting of out-of-autoclave composites using atmospheric pressure plasma is presented herein. Atmospheric pressure plasma surface preparation, which has shown promise in the surface preparation of composites and metals, were used in two separate approaches to better engineer the releasing interface between the out-ofautoclave prepreg and tooling. Treatment of mold-released tooling with variable chemistry plasma was the first approach, which decreased estimated surface porosity from 2.9 ± 0.6 % to 2.2 ± 0.6 % utilizing helium-hydrogen plasma. The second approach was to coat the Aluminum tool surfaces with thin, mold-releasing films by atmospheric-pressure plasma-enhanced chemical vapor deposition to eliminate the need for release plies and solve additional contoured surface issues release plies cannot tackle. The coating of a plain aluminum and an anodized tool reduced composite surface porosity to 2.3 ± 0.6 % and 1.0 ± 0.8 % respectively. Both approaches and their variations are summarized in this paper.
The present study investigated the effects of three pressure-related process deviations (reduced ambient pressure, reduced vacuum and restricted air evacuation) on the consolidation and quality of flat laminates manufactured by out-of-autoclave prepreg processing. The evolution of laminate thickness, measured in-situ, was correlated with thickness and porosity data for woven and unidirectional fibre bed composites. The process deficiencies are shown to have had distinctive detrimental effects on the rate of thickness change and on the amount, distribution and morphology of voids, and to have been more pronounced for the woven fabric prepreg format due to its higher initial bulk factor.
Fabrication of composite parts from prepregs often requires layup and preparation times of days and even weeks, during which prepregs undergo room-temperature aging. The aging process can compromise compaction, tack, and overall quality of composite parts, and thus a need exists for an accurate and convenient method to monitor the extent of prepreg aging as a function of out-time. Here, we report a method to monitor prepreg age, which involves measurement of changes in glass transition temperature as a function of room-temperature aging time. Samples from three out-of-autoclave prepreg systems were aged in ambient conditions and tested periodically using modulated differential scanning calorimetry. A linear increase in glass transition temperature with prepreg age was noted. Results are discussed in the context of monitoring the chemical aging of epoxy resins that occurs at ambient temperature.
Air evacuation is crucial to achieve low porosity for vacuum-bag-only manufacturing of out-of-autoclave prepregs. In this paper, the air permeability of composite skins was evaluated for in-plane and transverse air evacuation. The air permeability and microstructure were evaluated for three common prepreg fabric architectures: unidirectional, plain weave, and 5 harness satin. Since prepreg permeability is anisotropic, a 3-D pressure gradient can arise in honeycomb skins. In-order to calculate the effective air permeability coefficients of honeycomb skins, the computational fluid dynamics software FLUENT was used to determine the 1-D pressure gradient and corresponding area normal to flow. The air permeability coefficients were correlated to the prepreg microstructure using micro-CT imaging. The results showed that the in-plane air permeability was higher for fabrics with larger visible dry tow areas. The transverse air permeability was higher for prepregs that formed connected macro-porosity networks after lay-up, compared to prepregs with isolated macro-pores.
Tow impregnation as a function of material out-time was investigated for an out-of-autoclave carbon fiber–epoxy prepreg. Prepreg was aged at ambient temperature for 56 days. Every 7 days, laminates were laid up and cured using vacuum bag only processing. Void content was calculated through image analysis of polished sections. Experimental results were used to validate an analytical model for tow impregnation. Model predictions were based on flow kinetics during processing conditions, taking into account increasing degree of cure and evolution of resin viscosity as a function of ambient aging time. The study found that no significant tow porosity occurred within the material’s stated out-life, that tow porosity increased once this out-life was exceeded and eventually stabilized due to the room-temperature vitrification of the resin. The model’s predicted trends were consistent with experimental results, suggesting that an increase in resin viscosity is indeed the main cause of out-time induced tow porosity and providing a means of predicting laminate quality as a function of room temperature aging time.
Out-of-autoclave prepregs based on woven fabrics initially consist of dry tows and resin-rich areas. The tows allow air evacuation in the initial stages of processing and are subsequently infiltrated by surrounding matrix. The following study analyzes the relationship between material properties, process parameters and tow impregnation for three OOA prepregs. First, a representative model for tow impregnation is developed. Then, the model parameters are determined and the model predictions are correlated to impregnation data measured by X-ray microtomography. Finally, the model is used in a parametric study to investigate the effect of fibre architecture, cure cycle temperature and resin initial degree of cure on tow impregnation rate and to predict the possible occurrence of flow-induced micro-voids.
Industrial production of thermoset composite components usually involves the application of a vacuum bagging and autoclave pressure to minimize void percentage, usually to less than 5%. The application of vacuum bagging can be time consuming, with the actual vacuum application limited to a short interval during the curing process to avoid resin starvation. At the maximum vacuum achievable void reduction is only to about 10%, with a higher applied vacuum resulting in larger void diameters. This paper reports on the use of high pressure of up to 7000 kPa by means of an isostatic press, without vacuum application, to effectively reduce the void levels to below 3%. Void reduction occurred only above an initial pressurization level, at about 400 kPa, after which significant reduction was obtained until about 5000 kPa. Further pressurization reduced the void content only marginally.
The effects of various curing pressures (resulting in different void contents) on some mechanical properties of carbon/epoxy laminates have been examined. Thermogravimetry and mechanical spectrometry were used to determine the initial water content of prepregs and the dependence of viscosity on temperature. The pressure conditions were defined with respect to these preliminary test results. A set of curing pressure routes was then designed to produce unidirectional laminates with void contents between 0.3 and 10 vol%. The mechanical properties of unidirectional laminates depend on the void content and this results in the determination of an optimal curing pressure route which involves minimizing the void content, while the mechanical properties are kept at their maximum level. To allow generalization of the experimental results, this study has been performed on two different composite materials.
The ability to predict the viscosity of thermoset resin is important to understand the manufacturing process of composites and optimize the processing parameters. During resin or prepreg storage course, the cure reaction may happen and the degree of cure increases gradually. The storage aging effect reduces the fluidity of resin, and hence alters the processability of resin. In this article, the rheological properties of an epoxy resin and a bismaleimide resin used in composite autoclave process were measured and a viscosity model was established, which can predict the viscosity progression during cure for different aging degree of resin. Moreover, a computer simulation method was used to study the effects of aging degree on the composite consolidation and the processing operations. It is found that the viscosity model of aged resin can be obtained by modified dual Arrhenius model of fresh resin with the dynamic rheological measurement. The resin aging strongly alters the flowability, so influences composite consolidation. According to the simulated results, the processing parameters need to be adjusted to achieve cured composites with appropriate fiber content. POLYM. COMPOS., 2009. © 2008 Society of Plastics Engineers
The effects of cure pressure on resin flow, compaction, void content, and mechanical properties of fiber reinforced composites were investigated. Tests were performed on lami nates made of Fiberite T300/976 and T300/934 unidirectional prepreg tape. Data are pre sented on compaction, resin flow, and void content which lend support to the validity of the Springer-Loos model. Data are also given showing the effects of cure pressure on void content and on the compressive strength and short-beam shear strength and modulus.
The reactivity and processability of prepregs for high performance composites have been investigated as a function of time and storing conditions. The study has been focused on the stability of epoxy matrix carbon fiber prepregs, affected by exposure to controlled environmental conditions before their use in composite manufacturing. Effects of the aging on glass transition temperature, reactivity and processability have been investigated by calorimetry. Dynamic, isothermal and cure simulating tests have been performed to this aim. Results on toughened TGDDMDDS epoxy matrix prepregs are reported. A theoretical kinetic model proposed for the unaged system has been adapted for aged prepregs, by properly evaluating the variations of kinetic parameters with the aging time.
Thermoanalytical measurements and tack tests were both performed using a commercially available carbon fiber/epoxy prepreg system (Hercules 3501–6) to examine changes caused by aging as they affect handling and processability of thermosetting matrix-based composites. Combining these techniques, a relationship between prepreg bulk and surface characteristics in relation to aging was investigated. Isothermal kinetic studies at low temperatures showed maximum conversions (αm) that increased with increasing cure temperatures. In addition, a linear relationship between glass transition temperatures (Tg) and conversions (α) was observed regardless of aging (or cure) temperatures. Energy of separation of prepreg stacks, which may be viewed as a measure of prepreg tack, showed a maximum value at a specific temperature. The maximum energy of separation was observed in the temperature range of 20–25°C above the glass transition temperature for a given sample. However, the maximum energy of separation values decreased with increasing aging times (or conversions), implying that prepreg tack was a viscoelastic property rather than a viscous property of the resin matrix in the prepreg.
The times to gelation and to vitrification for the isothermal cure of an amine-cured epoxy (Epon 828/PACM-20) have been measured on macroscopic and molecular levels by dynamic mechanical spectrometry (torsional braid analysis and Rheometrics dynamic spectrometer), infrared spectroscopy, and gel fraction experiments. The relationships between the extents of conversion at gelation and at vitrification and the isothermal cure temperature form the basis of a theoretical model of the time–temperature–transformation (TTT) cure diagram, in which the times to gelation and to vitrification during isothermal cure versus temperature are predicted. The model demonstrates that the “S” shape of the vitrification curve depends on the reaction kinetics, as well as on the physical parameters of the system, i.e., the glass transition temperatures of the uncured resin (Tg0), the fully cured resin (Tg∞), and the gel (gelTg). The bulk viscosity of a reactive system prior to gelation and/or vitrification is also described.
A generalized theory for the glass transition temperature of crosslinked and uncrosslinked polymers has been developed, which takes into account the influences of end groups, branching, and crosslinking, and their functionality distribution. DiBenedetto's theory was found to correctly characterize the influence of crosslinks on the glass temperature. Normalized to constant crosslink functionality, the crosslink constant is a universal parameter suggesting that the entropic theory of glasses is applicable to crosslinked systems. Data on linear polymers and networks from the crosslinking of polymer chains, vinyl/divinyl-copolymers and step-growth polymers, such as polyurethanes, amine-cured epoxies, or inorganic glasses, are presented.
The modulus, density, glass transition temperature (Tg), and water absorption characteristics of an amine-cured resin [diglycidyl ether of bisphenol A (Epon 828)/diaminodiphenyl sulfone (DDS)] were studied as a function of extent of cure. The glass transition is a function of the extent of cure and reaches a maximum temperature, T, when it is completely cured; specimens with different extents of cure were formed by isothermal cure below T, for different times. After slowly cooling, the density at each extent of cure was obtained at room temperature. Moisture absorption was monitored gravimetrically at 25°C for 2 months at several humidity levels. The room temperature density and modulus decreased with increasing extent of conversion whereas the glass transition temperature and equilibrium water absorption increased. The equilibrium water absorption increased linearly with relative humidity, and the absorptivity increased linearly with specific volume. An interpretation of these anomalous results is made in terms of the nonequilibrium nature of the glassy state. The glass transition temperature increases as the extent of cure increases resulting in a material that is further from equilibrium at room temperature and therefore has more free volume and a greater propensity to absorb water.
Void formation as a function of resin moisture content was investigated to better understand and control process defects in composite parts made from prepreg. In this study, uncured prepreg was conditioned at 70%, 80% and 90% relative humidity and at 35 °C. Conditioned prepreg was laid up into quasi-isotropic laminates and cured using vacuum bag only (VBO) processing (low-pressure), and autoclave processing. Moisture uptake in the resin was measured using coulometric Fischer titration. Void content was measured by image analysis of polished sections of cured laminates. Void fractions increased substantially with increasing moisture content in VBO processed laminates, while autoclave-processed parts remained void-free. Experimental results were consistent with trends predicted using a diffusion-based analytical model. The findings are discussed in the context of causes of voids in prepreg composites.
The resin system affects the shape and size of voids in composite laminates and the associated fracture parameters. Most work on the effects of voids in the literature deals with carbon/epoxy laminates; little experimental data are available on other resin systems of practical importance. This work aims at comparing the effect of voids on carbon/epoxy and carbon/bismaleimide fabric laminates. Specimens with void contents in the range 0 to 6% were produced from both materials. The void content is correlated to ultrasonic attenuation and microscopic inspection was carried out to analyze the sizes and shapes of the voids. The interlaminar shear strength was measured by the short-beam method described in ASTM D2344 and the adequacy of a fracture criterion to represent the experimental data for both materials is assessed. The influence of voids on the interlaminar shear strength of both materials is compared in terms of the corresponding fracture parameters and the shape and the location of the voids.
Ultrasonic imaging in the C-scan mode was used to measure the flow rate of an epoxy resin film penetrating through the thickness of a single layer of woven carbon fabric. Assemblies, comprised of a single layer of fabric and film, were vacuum-bagged and ultrasonically scanned in a water tank during impregnation at 70 °C. The permeability of the fabric was calculated using Darcy’s law. The results demonstrated that ultrasonic imaging in the C-scan mode is an effective method of measuring z-direction resin flow through a single layer of fabric. Comparison of ultrasonic and microscopy images yielded consistent results and demonstrated the effectiveness of ultrasonic imaging as an in situ process diagnostic for monitoring through-thickness impregnation and flow rates.
An experimental method based on a falling pressure measurement is proposed to evaluate the through thickness air permeability of prepregs during cure, with two set-up variants. One concerns the measurement of the inherent air permeability and the other applies to processing conditions, implying the use of consumables and vacuum-bag. The methods were successfully tested and fulfil the criterions used in soil science for the evaluation of the air permeability. Prepreg permeability to air during cure varies between 10−17 m2 and 10−19 m2 and is found to depend on the evolution of the resin viscosity, which shifts from an initially immobile phase to a mobile one and back to an immobile phase.