Advances in Polyurethane Pultrusion: Cure Modeling and Second Generation Resin Systems

American Composites Manufacturers Association February 21-23, 2011 Las Vegas, Nevada USA Advances in Polyurethane Pultrusion: Cure Modeling and Second Generation Resin Systems by Michael Connolly and Dan Heberer Huntsman Polyurethanes Auburn Hills, MI (DSC) and rheological experiments 4. Both cure rate and viscosity are dependent on time, temperature, conversion, pressure, etc., and they also mutually influence each other. Kinetic and rheological models which express curing rates and viscosity as a function of the conversion and temperature can be applied to experimental DSC and rheology data as has been reported 5. As pultruders have become knowledgeable with injection processing, first generation polyurethane resins have found early, niche success in the marketplace due to their strength and toughness. However, new opportunities such as cargo load floors, electrical cable and various thick section pultrusions demand additional capabilities of resin systems. To enable these and other applications, second generation resins have been developed to have higher Tg or lower viscosity to increase usetemperature or improve wet-out and line speed in thicksection structural profiles. Abstract: Over the past decade, pultrusion with polyurethane (PU) resins has begun to mature and is being adopted by an ever-widening group of custom pultruders and captive users. The strength and toughness of polyurethane resins are now widely recognized and polyurethane process knowledge is becoming more commonplace. Using 'first generation' resin systems, PU profiles are now in production as window lineals, ladder rails and a wide range of structural profiles. However, to enable further expansion of PU resin use in pultrusion, process innovations and 'second generation' formulations are needed. In this paper, advances in thermal modeling of die cure will be discussed, along with new formulations for higher glass transition temperature and low viscosity for improved fiber impregnation and higher line speeds. Introduction With its fast reactivity and the absence of volatile organic compounds, polyurethane (PU) resins have become increasingly adopted in pultrusion applications. These resins offer high strength and durability in the final composites, combined with the potential for high pultrusion line speeds 1-3. In order to exploit these qualities, a good knowledge of resin cure and viscosity build-up during processing is essential. Although a trial-and-error approach can eventually lead to the desired knowledge, a more detailed understanding of the physical and chemical parameters allows a better and faster prediction of the chemo-rheological behavior. The two main parameters to characterize curing systems are the curing rate and the resin viscosity, which can be determined by differential scanning calorimetry Cure Modeling: Polyurethane Resin Chemistry To improve and optimize tooling design and processing, detailed knowledge is crucial regarding the time and temperature at which the effects of the network formation takes place in polyurethane resins. Two suitable techniques to study the process of curing are differential scanning calorimetry and rheology. Mechanism of Cure Kinetics A mechanistic kinetic model can be derived from the main simplified reaction mechanism. The isocyanate forms an activated intermediate with the catalyst (rate determining step), which then reacts further with the polyol to form a polyurethane linkage. NCO- + CAT «C-1 C-1 + OH- U + CAT Assuming all side reactions are unimportant and physical effects (e.g., diffusion) can be neglected, then the reaction mechanism can be simplified further to a single reaction described by a conversion equation, namely: where is an effective curing rate dependent upon temperature, catalyst concentration and initial isocyanate concentration; α is the isocyanate conversion and the initial isocyanate-to-polyol ratio. These assumptions hold true mainly at the beginning of the experiment. Differential scanning calorimetry and 1

rheometry can be used to obtain the kinetics of this mechanistic model. Experimental Techniques to Study Cure Kinetics A typical polyurethane resin formulation for pultrusion processes has been used to study the rheokinetic parameters. It consists of SUPRASEC 9700 polymeric MDI-based isocyanate and RIMLINE SK97007 polyol blend. Differential scanning calorimetry (DSC) experiments were performed on a TA Instruments DSC 2920 using hermetically sealed pans and sample weights of approximately 10 mg. Both isothermal and dynamic scans were run at temperatures between 20 to 60ºC or heating rates between 3 and 15ºC/min, respectively. Because of the high reactivity, the reaction starts as soon as the isocyanate and polyol are mixed. The sample preparation must therefore be done swiftly in order to avoid loss of exotherm data during the first stages of the reaction. FIGURE 1 shows dynamic DSC runs at three different heating rates. The maximum peak shifts towards higher temperatures with increasing heating rates. The relative area under the curves up to a given temperature is a good measure of the conversion. In all curves, a shoulder on the right, high temperature side of the peak can be distinguished indicating inhibition of the reaction. This reaction delay is due to a gelation effect, discussed below, which transforms the reaction rate from being kinetically limited to diffusion controlled. Rheological measurements were performed on a TA Instruments AR2000 stress-controlled rheometer with a 25-mm aluminum parallel plate setup. The time dependent increase in viscosity (Pa s) was monitored during isothermal runs at temperatures between 20ºC and 60ºC at a torque of 10 µn m. FIGURE 2 shows isothermal rheometry data at different temperatures. After an initial stage of a small viscosity build up, the viscosity then increases quite sharply. This process is called gelation; it is the formation of an infinite network at which the viscosity increases to infinity. From an initial liquid-like behavior, the system transitions into solid-like behavior. Gelation time can be defined in several ways. One simple means, used in this study, can be to take the time when the viscosity reaches 10,000 Pa s. As can be seen in FIGURE 2, the gelation process will take place faster and also more abruptly at higher temperatures. Increased temperatures will lead to faster reaction rates and eventually shorter curing times. After gelation, the curing will still continue, but the process will be much slower. This is because the reactive groups are restricted in their movement. Because of this diffusion effect, complete conversion is seldom reached. Experimental Fit For the formulation previously described, the following curing kinetics was obtained from DSC and rheological experiments: with a heat of reaction of 280 J/g. Pultrusion Process Modeling The main processing challenges for pultruders are the speed at which the pultrusion line will operate, the pot life of the resin, the die temperature profile, the die geometry and the reinforcement schedule. Previous work 6 described best practice in injection box design for producing good fiber wet-out. The present study will focus on how polyurethane composites cure in pultrusion dies. To model the die in a pultrusion process, the resin conversion and temperature distributions must be predicted. The die thermal model should be able to predict whether the resin is fully and uniformly cured as well as the location of the maximum exotherm along the length and through the thickness of the profile. The model should also be able to predict the influence of operating conditions, such as line speed, heater configuration, etc., on the conversion and temperature distribution. The present study evaluates pultruded sections of varying thickness using polyurethane resin and glass fiber reinforcement. The conversion, α, and the temperature, T, distributions are governed by the following equations: where U x is the line speed; and λ are the thermal properties of the composite, namely heat capacity and thermal conductivity; H r the heat of reaction and r the curing kinetics; and x the direction in which the part is being pultruded and y the distance from the center of the part. The boundary conditions of the thermal model are: T(0, y) = T 0 α(0, y) = 0 T(x, y wall ) = T wall 2

Using curing kinetics obtained from DSC and rheological experiments and combining them with thermal properties of glass fiber and polyurethane resin, predictions of the curing process can be made. Model Application FIGURE 3 shows the predictions of temperature and conversion in a pultrusion process. Since the composite profile is symmetric along the Z-axis (thickness direction), only the top half of the part is modeled. The core or center of the part is along the bottom axis of FIGURE 3 and the die wall at the top. The first 250 mm in length represent the injection box where resin is impregnated into the fiber bundle and little or no curing takes place. The die wall is heated to a maximum temperature of 210ºC and curing of the resin begins. Heat is conducted towards the core of the part as the reaction starts. Since the reaction is exothermic, the temperature rises in the core to a maximum of about 215ºC. When the part leaves the die, it is fully cured. FIGURE 4 compares the temperature and conversion history along the core and the die wall. Conversion of the resin in the core lags conversion at the wall due to the earlier and higher temperature at the die wall. The peak of the temperature in the core also lags the wall temperature due to the heating profile at the wall and the later conversion of the resin in the core. Influence of Line Speed and Heating Rate on Curing FIGURES 5 and 6 show the influence of line speed on resin curing in the die. As line speed is increased, the residence time of the part in the die is shortened and the peak core temperature moves down the die. Similarly, the completion of cure also occurs later in the die. This effect can be seen more clearly in the core conversion and temperature histories shown in FIGURE 6. Predicted exotherms in the profile at both line speeds show good agreement with experimental data. The influence of different die heating profiles can also be studied using the model. FIGURES 7 and 8 compare temperature and conversion distributions for two different heating rates, one where the die is heated up early, and another where later heating occurs. Again, when compared with experimental data for core temperature history, the trends and agreement of the model and experiment are good. Second Generation PU Resin Systems Early polyurethane pultrusion resins were described as long as thirteen years ago 7-9. These prototype resin systems revealed the potential for line speed improvements and unique toughness in pultruded PU profiles, but these materials did not prove particularly robust in a production environment and did not yield profiles of commercial quality. Continued development of PU pultrusion formulations lead to commercially viable resins by 2005. These initial resins now have a proven track record in production over the past 6-7 years. With new opportunities expanding the market for pultrusion profiles, new PU pultrusion formulations are required to meet performance, productivity and cost targets. Commercial window lineals and other structural profiles with long term outdoor exposure requirements must be painted to maintain surface appearance and properties, including strength and electrical properties. However, many high durability paint and powder coating systems require bake temperatures up to, and sometimes exceeding 150 C to ensure good coating consolidation and cure. Since these temperatures are near or above the glass transition temperature of many standard PU resins, a higher Tg resin is required to minimize the impact of coating bake temperatures on composite properties and profile dimensional stability. Hence, a new resin was developed with a Tg 20 C higher than the standard resin. The unique strength and toughness of PU resins allow the potential of pultruded composites to displace traditional materials of wood, steel and concrete in many structural applications. However, such profiles can be as thick as 20-mm or more and may be fabricated with many layers of rovings, random mat, and/or heavy engineered fabric. Obviously, ensuring complete fiber wetout is critical to end-use performance, but doing so is exceedingly difficult in the dynamic wet-out of pultrusion injection dies. As has been learned in resin infusion and resin transfer molding, decreasing resin viscosity to less than 500 mpa*s greatly improves fiber wet-out. Thus, a reduced viscosity PU resin was developed to reach this critical level with a modified polyol and isocyanate compared to a standard system. Isothermal viscosities of the standard PU, the high Tg resin and the reduced viscosity system are shown in FIGURE 9. The initial viscosity of the high Tg version is similar to the standard resin, but unsurprisingly, the viscosity rises much faster due to the higher cross-link density. Thus, the working time (time in the injection zone and time to wet the fibers) will be reduced. The low viscosity variant has an initial viscosity less than half of the standard resin (490 vs. 1100 mpa*s). Viscosity of this variant rises in parallel to the standard system. Using a criterion of 2000 mpa*s for maximum working viscosity, the high Tg variant could be used for only 12-13 minutes, while the standard and low viscosity variant reach this critical level after 20-25 minutes. Dynamic mechanical analysis (DMA) data for the three resin variants is shown in FIGURE 10. The low viscosity variant has a Tg ~10 C lower than the standard resin, but this Tg remains sufficient for the majority of 3

PU pultrusion application requirements. The high Tg variant has a glass transition 20 C higher than the standard resin. With a Storage Modulus G onset over 150 C, this higher Tg system should prove more stable in high temperature exposure in paint ovens or perhaps even in electrical transmission cables. Physical properties of neat resin plaques of each resin are give in TABLE 1. Results are remarkably similar for each resin with the notable exceptions of Tg, as discussed above, and tensile elongation-to-failure. As might be expected, the high cross-link density of the high Tg system leads to reduced tensile elongation of 5.5%, which is still quite good compared to other non-pu resins of similar Tg. It is also worth noting that all of these PU variants exhibit ductility and none fail in flexural testing. Conclusions: Kinetic and rheological data have been obtained by dynamic and isothermal DSC as well as viscosity buildup measurements. The results could be successfully fitted to Arrhenius-based models. The obtained parameters were subsequently used in a mathematical model of a pultrusion die. The model can predict temperature and conversion distributions in a pultrusion die. It can be used to study the influence of line speed, heating rate and part thickness on curing process and shows good agreement with experimental data. This kinetic model has been embedded into a predictive computational tool to assist technical service personnel for pultrusion die design and process setup. Two new polyurethane resin systems were developed and evaluated on commercial pultrusion dies. The first system was designed with a glass transition of 150 C (by DMA G onset) which is 20 C higher than a standard PU pultrusion resin. A second PU resin was developed for improved wet-out in thick-section profiles and designed for reduced viscosity of ~500 mpa*s, a 55% reduction compared to the standard PU resin. Both resin systems are now commercially available. Disclaimer: While the information and recommendations in this publication are, to the best of our knowledge, information and belief, accurate at the date of publication, NOTHING HEREIN IS TO BE CONSTRUED AS A WARRANTY, EXPRESS OR OTHERWISE. IN ALL CASES, IT IS THE RESPONSIBILITY OF THE USER TO DETERMINE THE APPLICABILITY OF SUCH INFORMATION AND RECOMMENDATIONS AND THE SUITABILITY OF ANY PRODUCT FOR ITS OWN PARTICULAR PURPOSE. NOTHING IN THIS PUBLICATION IS TO BE CONSTRUED AS RECOMMENDING THE INFRINGEMENT OF ANY PATENT OR OTHER INTELLECTUAL PROPERTY RIGHT, AND NO LIABILITY ARISING FROM ANY SUCH INFRINGEMENT IS ASSUMED. NOTHING IN THIS PUBLICATION IS TO BE VIEWED AS A LICENCE UNDER ANY INTELLECTUAL PROPERTY RIGHT. Except where explicitly agreed otherwise, the sale of products referred to in this publication is subject to the general terms and conditions of Huntsman International LLC or of its affiliated companies. Huntsman Polyurethanes is an international business unit of Huntsman International LLC. Huntsman Polyurethanes trades through Huntsman affiliated companies in different countries such as Huntsman International LLC in the USA and Huntsman Holland BV in Western Europe. RIMLINE and SUPRASEC are registered trademarks of Huntsman Corporation or any affiliate thereof, in one or more countries, but not all countries. Acknowledgments: The authors would like to thank Roel van Boxel and Mark Brennan of the Core Science Team at Huntsman Polyurethanes in Everberg, Belgium for the performing thermal analysis and rheology experiments and developing the pultrusion cure model. Author(s): Michael Connolly, Ph.D. is Product Manager Urethane Composites for Huntsman Polyurethanes in Auburn Hills, MI. Michael has been employed for 15 years at Huntsman and has spent the past 10 years in polyurethane composite technical and market development. He received his doctoral degree in Polymer Science and Engineering from the University of Massachusetts-Amherst in 1989. Dan Heberer, Ph.D. is a graduate of the Polymer Science Department at the University of Akron. His career has primarily focused on the development and commercialization of adhesives for automotive applications. Currently, he is a Technical Manager for elastomers and composites in the America s Business Development Team at Huntsman Polyurethanes in Auburn Hills, MI. 4

References: 1. Connolly M., King J.P., Shidaker T.A., Duncan A.C., Pultruding Polyurethane Composite Profiles: Practical Guidelines for Injection Box Design, Component Metering Equipment and Processing, Proceedings of Composites 2005, American Composites Manufacturers Association, Sept. 2005. 2. Connolly M., King J.P., Shidaker T.A., Duncan A.C., Processing and Characterization of Pultruded Polyurethane Composites, Proceedings of the 8th World Pultrusion Conference, European Pultruders Technical Association, March 2006. 3. Connolly M., King J.P., Shidaker T.A., Duncan A.C., Characterization of Pultruded Polyurethane Composites: Environmental Exposure and Component Assembly Testing, Proceedings of Composites 2006, American Composites Manufacturers Association, Oct. 2006. 4. Yousefi A., Lafleur P.G., Kinetic Studies of Thermoset Cure Reactions: A Review, Polymer Composites, Vol.18, pp 157-168, Apr. 1997. 5. Connolly, M., van Boxel, R. and Brennan, M., Rheo-Kinetics and Cure Modeling for the Pultrusion of Polyurethane Composites, Proceedings of the 1 st Global Pultrusion Conference, Baltimore, MD, June 2009. 6. Brennan M., Connolly M., Shidaker T.A., CFD Modeling of the Closed Injection Wet-Out Process For Pultrusion, Proceedings of the 9th World Pultrusion Conference, European Pultruders Technical Association, March 2008. 7. Joshi, R.R., Varas, L.L., Padsalgikar, A.D., Polyurethanes in Pultrusion: Styrene Free Alternative Systems, International Composites Expo 1999-SPI, (Society for the Plastics Industry), April 1999. 8. Coffee, H.D., Perry, M.J., Processing Techniques for Reinforced Thermosetting Urethane Composites, Proceedings of Composites 2001-CFA, (Composites Fabricators Association), October 2001. 9. Vaughn, J.G., Lackey, E., Coffee, H.D., Barksby, N., Lambach, J.L., Pultrusion of Fast-Gel Thermoset Polyurethanes: Processing Considerations and Mechanical Properties, Proceedings of Composites 2003-CFA, (Composites Fabricators Association), October 2003. 5

FIGURE 1: Typical Experimental DSC Data for the Curing of Polyurethane Resins FIGURE 2: Viscosity Build-up of a Polyurethane Resin at Different Temperatures 2

FIGURE 3: Predicted Contours of Temperature (Top) and Resin Conversion (Bottom) in a Pultrusion Process with a Line Speed of 60 mm/min with a Part Thickness of 6.4 mm FIGURE 4: Comparison of Core and Die Wall Temperature and Conversion History for a Pultrusion Process Running at a Line Speed of 60 mm/min with a Part Thickness of 6.4 mm 3

FIGURE 5: Influence of Line Speed on the Temperature and Conversion Distributions in a Pultrusion Die for a Composite Part with a Thickness of 6.4 mm - 30 mm/min [Top] and 60 mm/min [Bottom] FIGURE 6: Influence of Line Speed on the Core Temperature and Conversion History for a Pultrusion Part of Thickness 6.4 mm [S = Simulation, E = Experimental] 4

FIGURE 7: Influence of Different Die Heating Profiles on the Temperature and Conversion Distributions in a Pultrusion Die with a Line Speed of 45 mm/min and Part Thickness of 6.4 mm Early Heating (Top) and Late Heating (Bottom) FIGURE 8: Influence of Different Heating Profiles on the Core Temperature and Conversion History in a Pultrusion Die with a Line Speed of 45 mm/min with a Part Thickness of 6.4 mm [S = Simulation, E = Experimental] 5

10000 9000 s ) * 8000 a P 7000 (m * 6000 ta E 5000 sity c o 4000 is V3000 x le p 2000 m o 1000 C 0 Standard PU Low Viscosity PU High Tg PU 0 5 10 15 20 25 30 35 40 Time (min) FIGURE 9: Viscosity vs. Time of Polyurethane s 1.E+10 1.E+09 ) a ' (P G1.E+08 s lu u d o M1.E+07 e a r h S e g 1.E+06 r a to S Standard PU Low Viscosity PU High Tg PU G' Onset 131 \C G' Onset 121 \C Tan Delta Max 138 \C Tan Delta Max 147 \C G' Onset 151 \C Tan Delta Max 166 \C 1.E+05 0.00 25 50 75 100 125 150 175 200 Temperature ( C) 1.00 0.90 0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10 e lta D t e n g a n T FIGURE 10: DMA of Polyurethane Pultrusion Neat Resin Samples 6

TABLE 1: Physical Properties of Polyurethane Pultrusion Neat Resin Plaques RESIN Standard PU Low Viscosity PU High Tg PU Initial Mixed Viscosity (mpa*s) 1100 490 950 Gel Time (min) 100 g Mass in 125-ml Cup Tensile Modulus (GPa) ASTM D638 18 ± 2 18 ± 2 13 ± 1 2.90 2.85 2.84 Tensile Yield Strength (MPa) No Yield 93 92 Tensile Break Strength (MPa) 83 91 88 Tensile Yield Strain (%) No Yield 6.7 6.4 Tensile Strain-at Break (%) 9.4 8.5 5.5 Flexural Modulus (GPa) ASTM D790 2.96 3.00 2.90 Flexural Yield Strength (MPa) 117 128 119 Flexural Break Strength (MPa) No Break No Break No Break Flexural Yield Strain (%) 7.1 7.6 7.1 Flexural Strain-at Break (%) No Break No Break No Break Tg - DMA tan delta Peak ( C) 147 138 166 Tg - DMA G' Onset ( C) 131 121 151 Heat Deflection Temperature ( C) 4.5 MPa - ASTM D648 132 127 146 Notched Izod ( J/m ) 64 43 43 Hardness - Shore D ASTM D2240 Water Absorption (%) 24H @ 25 C - ASTM D570 Specific Gravity ASTM D792 81 80 79 0.35 0.36 0.27 1.23 1.23 1.24 7