TH 16 INTERNATIONAL CONFERENCE ON COMPOSITE MATERIALS THE DURABILITY OF LIQUID MOLDED CARBON NYLON 6 COMPOSITE LAMINATES, EXPOSED TO AN AGGRESSIVE MOISTURE ENVIRONMENT. Selvum Pillay, Uday K. Vaidya and Gregg M. Janowski [Selvum Pillay]: [email protected] Uday Vaidya: [email protected] The University of Alabama at Birmingham Keywords: hygrothermal effects, durability, VARTM, composite laminates stiffness, and corrosion resistance. However, their susceptibility to environmental degradation, Abstract especially hygrothermal aging and ultraviolet (UV) Liquid molding of thermoplastics has been limited irradiation, depending on the application and by high resin viscosity, high temperature processing formulation of the matrix, has been of major requirements, and a short processing window [1]. concern. The processing parameters for vacuum assisted resin When exposed to humid environments, transfer molding (VARTM) developed by the polymer-based composite structures can absorb authors and previously reported [2] have been moisture, which affects the long-term structural adopted to process carbon/nylon 6 composite panels. durability and properties of the composite [4,5]. The material used in this study was casting grade Since high performance carbon fibers absorb little or polyamide 6, ε-Caprolactam with a sodium-based no moisture compared to the matrix, the absorption catalyst and an initiator. is largely matrix dominated. Moisture penetration is The present work addresses the effects of largely dominated by diffusion. Other mechanisms moisture exposure on the static and dynamic that can contribute are capillarity and transport by mechanical properties of carbon fabric-reinforced, micro-cracks [6]. The rate of moisture absorption thermoplastic polyamide 6 (also referred to as nylon depends largely on the type of matrix, orientation of 6 and/or PA6) matrix panels processed using the fibers with respect to the direction of diffusion, VARTM. The Bao and Yee dual diffusivity model temperature of the water, and relative humidity [7]. [3] was applied to evaluate the moisture uptake for Moisture absorption leads to changes in the the C/PA6, fully immersed in distilled water at thermophysical, mechanical, and chemical 100°C. The model shows good correlation with the characteristics of the matrix by plasticization and experimental data in terms of the percentage of hydrolysis [8-10]. Also, moisture wicking along the moisture absorption with respect to the square root fiber-matrix interface degrades the fiber-matrix of time. Scanning electron microscopy (SEM) bond, which results in the loss of microstructural results show that surface defects induced during integrity. Matrix dominated properties are most processing result in matrix micro-cracks after strongly affected. moisture exposure and the fiber matrix interface is 2.0 Kinetics of Water Absorption compromised. The impact resistance of the material is also lowered due to water ingress. However, the A basic study of the existing models of crystallinity and melting peaks are not affected. water absorption in composite materials was undertaken, to compare results of the experimental 1.0 Introduction moisture uptake/diffusion data for the condition of the specimens fully immersed in distilled water at Polymer matrix carbon fiber-reinforced 100°C. A number of studies have been conducted on composites are widely used in a variety of moisture diffusion in composites, and most agree applications, such as aerospace, mass transit, that Fickian behavior adequately predicts the water automotive, and sporting goods. Their use has ingress in these composites [7,11]. These studies become popular due to their attractive properties, have also shown that the most aggressive condition including high strength to weight ratio, high specific for the composite was full immersion in water at 100° C. The equilibrium absorption was also reached in the minimum time, under 100 h for nylon Pillay1, Vaidya2, Janowski3 6/6 [7]. Most of these studies have been conducted composites and proposed a residual stress-dependent for nonwoven (unidirectional and random chopped model. The basis for the residual stress was the fiber) composite materials. Ishak and Berry [7] consideration that matrix shrinkage is restricted in showed that Fickian absorption adequately modeled both directions, hence large residual tensile stresses the moisture absorption of nylon 6/6 reinforced with may develop after cure. As the diffusion of moisture short carbon fiber. progresses, the swelling of the matrix relieves the Equations 2 and 3 outline the model residual stress within the material and slows down developed by Crank [12] for diffusion of a sub- the absorption. The linear dependency of diffusion stance in one direction only. For an infinitely large on time proposed improved correlation with the plate of thickness h with single phase diffusion, Mt experimental data; however, the fit still showed (mass of water absorbed at time t) by M∞ (mass of some discrepancy. water absorbed at saturation) is given by: Bao and Yee [3] conducted an extensive study on the diffusion kinetics of moisture in both M t 81 ⎡ ⎛ Dt ⎞ 2 2 ⎤ =−1 22∑ ×−EXP⎢ ⎜ 2 ⎟ π ()21n + ⎥ unidirectional and woven composites. They also M ∞ π n=0 ()21n + ⎣ ⎝ h ⎠ ⎦ found that the diffusion was non-Fickian for the (2) woven composite. They conducted an extensive evaluation of the different models and fits with This equation is a series solution to Fick’s experimental data, and found that there were second law, which can then be resolved into two discrepancies with the experimental and predicted conditions of Dt/h2 (Equations 3 and 4, results. They had also evaluated the stress-dependent respectively): model proposed Whitney and Browning [14] and Dt found that the correlation was limited at best. A For > 005. Equation 2 reduces to h2 closer study of the experimental absorption curves showed that there were two distinct linear regions before the equilibrium moisture was reached. They Mt 8 ⎡ ⎛ Dt ⎞ ⎤ found that the moisture uptake was composed of two =−1 EXP −⎜ ⎟ π 2 M π 22⎢ ⎝ h ⎠ ⎥ processes with different rates: a fast initial rate and ∞ ⎣ ⎦ then a slower rate. They proposed the following dual (3) diffusivity model, which is the sum of two Fickian processes with different diffusivities. Dt 075..075 For 2 << 005. Equation (2) reduces to ⎪⎧ ⎡ ⎛ Dt⎞ ⎤ ⎪⎫ ⎪⎧ ⎡ ⎛ Dt⎞ ⎤ ⎪⎫ H MM=−−⎨173exp⎢ .⎜ 1 ⎟ ⎥ ⎬ +−−M ⎨ 173 exp⎢ . ⎜ 2 ⎟ ⎥ ⎬ t ∞∞1 ⎝ 2 ⎠ 2 ⎝ 2 ⎠ ⎩⎪ ⎣ h ⎦ ⎭⎪ ⎩⎪ ⎣ h ⎦ ⎭⎪ (5) 12 M 4 ⎛ Dt ⎞ t = ⎜ ⎟ M π 12⎝ h 2 ⎠ ∞ In this model, D1 , M ∞1 and D2 , M ∞ 2 are (4) the diffusivity and equilibrium uptakes of the fast and slow processes, respectively. The diffusivity D can be calculated from the initial linear portion of the absorption curve from Equation (4). 3.0 Experimetal The moisture absorption does not follow the Carbon Fabric previously described model for woven fabric The carbon fabric used in this study was composites. There have been limited studies on the supplied by US Composites and was a 4 harness mathematical modeling of moisture ingress on satin weave, with a 3K tow size, and an areal density woven composites [3,13,14]. The transport of and thickness of 0.02 g/cm2 and 0.254 mm, moisture in woven composite materials is very respectively. The fabric was washed in acetone to complicated, as it is influenced by the matrix remove lubricants and sizing and then placed in a material, weave architecture, residual stresses, and recirculating oven at 100°C for 1 h to re-move any the fiber/matrix interface. residual moisture and/or acetone. Whitney and Browning [14] studied the moisture diffusion of bidirectional graphite The durability of liquid molded carbon nylon 6 composite laminates, exposed to an aggressive moisture environment . Monomer, Activator, and Catalyst (PA6) particular composite material is independent of AP grade caprolactam monomer was used in temperature and relative. At higher temperatures and this study. The monomer has low levels of moisture relative humidity, the maximum amount of moisture (<0.0015% by mass), a low melting point (Tm = absorbed by the material will be reached in a 69°C), and low viscosity (4.87 mPa•s at 100°C). significantly shorter time than exposure to a lower Bruggolen C20, hexamethylene dicarbamoyl relative humidity and lower temperature. dicaprolactam, and Bruggolen C10, sodium Since the focus of this study was to evaluate caprolactamate, both especially developed for AP of the deterioration in mechanical properties of the caprolactam, were used as the activator and catalyst, carbon fabric-nylon 6 matrix composite, it was respectively. The melt temperature of both C10 and decided to expose the material to the most C20 is 60°C. The monomer, catalyst, and activator aggressive moisture condition, i.e., immersed in were supplied by Bruggemann Chemical U.S., Inc., water at 100° C. The specimens were cut from the and were used without any further processing or panels using a diamond saw blade and then polished purification. Due to the sensitivity to moisture of the with 600 grit silicon carbide paper to minimize catalyst and activator, they were stored and handled cracking or debonding of the edges. in a dry box with the relative humidity maintained at The specimens were then dried in a 8%. recirculating air oven at 70°C and periodically weighed on a high precision balance with a 0.1 mg Processing resolution, until the weight had stabilized. The dry Twenty 300 x 300 mm layers of the carbon weight (Wd) of all the specimens was then recorded. fabric were laid onto the tool surface and bagged and The samples were then placed on a rack that sealed using high temperature film and tacky tape. A prevented them from any contact with each other thermocouple was positioned at the surface of the and fully submerged into a reaction vessel top layer of the carbon to record and monitor the containing boiling water. The temperature of the tempera-ture. The tool temperature was then raised water was maintained at 100°C using a hot plate and to 100°C, and a continuous vacuum of 1 atmosphere oil bath to maintain even heating.