New Fast-Cure, High Strength Phenolic Resin Systems for Pultrusion

J.G. Vaughan, University of Mississippi E. Lackey, University of Mississippi D. Green, ANGUS Chemical Company R. Swedo, ANGUS Chemical Company T. Johnson, ANGUS Chemical Company D. Trauth, ANGUS Chemical Company

Presented at the American Composites Manufacturers Association Conference October 6-8, 2004, Tampa, Florida USA

ANGUS Chemical Company A Subsidiary of The Dow Chemical Company

COMPOSITES 2004 Convention and Trade Show American Composites Manufacturers Association October 6-8, 2004 Tampa, Florida USA

New Fast-Cure, High Strength Phenolic Resin Systems for Pultrusion

by J.G. Vaughan, University of Mississippi E. Lackey, University of Mississippi D. Green, Angus Chemical Company R. Swedo, Angus Chemical Company T. Johnson, Angus Chemical Company D. Trauth, Angus Chemical Company

Abstract
Phenolic resins are well known for their excellent fire, smoke, and smoke toxicity properties (FST); retention of properties after long-term exposure to high temperatures; excellent electrical resistance; and excellent chemical resistance, especially in chlorinated solvents. However, they are also known to be quite difficult to pultrude due to their high viscosities, slow cure rates leading to slow pull speeds, and high porosity content. This paper discusses the development of a new series of low viscosity phenolic novolac resin systems processable at high speeds with excellent mechanical properties and outstanding FST characteristics. These systems feature a unique formaldehyde-free curing technology. Pultrusion conditions and their effect on composite properties are discussed. A comparison of process and mechanical properties for the new phenolic novolac systems are contrasted with alternative thermoset resins.

greater safety than common FRP materials with the phenol-resorcinol resins having slower flame spread, less smoke generated, and reduced toxicity as measured by standard ASTM test methods [4]. Driven by government regulations on smoke and fire performance for materials used in the construction and building industries, one of the first papers to address the pultrusion processing of phenolic resins appeared in 1993 [5]. This paper highlighted the processing problems of the phenolic resins available at that time and commented on the problem of residual water in the phenolic resins, how this lead to entrapped water and voids in the final product, and also how phenolic resins tended to give weak, slow forming gels leading to poor surface appearance with 95 percent of the molecular weight achieved within the last five percent of cure. Other papers from 1993 through 1996 highlight the problems experienced by early pultrusion resin systems such as slow cure rates, short pot life, low pH leading to die corrosion, high viscosity, high moisture content, and poor handling characteristics, and also discuss the development process manufacturers and pultruders were going through to develop workable phenolic pultrusion systems [6-9]. Research focused toward meeting the United States Coast Guard Policy File Memorandum 2-98, Level II led to more advanced pultrusion systems, but processing problems still required constant attention during manufacture [10-12]. Phenolic resins for pultrusion still are not considered to have the ease and speed of processing of polyester resin systems nor the mechanical properties. Their main application is in areas of FST that other resins can not meet. Thus, the present research was undertaken to develop a new phenolic resin system that overcomes many of the previous processing problems commonly associated with phenolics, while maintaining the excellent FST properties for which phenolics are so well known.

Resin System
Previous attempts to pultrude phenolics utilized resoles which are 1-part phenolic resin systems prepared with a phenol / formaldehyde ratio less than 1. This leads to a self reactive system which cures upon heating. This characteristic, while attractive to minimize formulation components, also leads to slow cure even at ambient temperature resulting in poor shelf-life. Thus resoles are often refrigerated during storage to minimize viscosity drift. In contrast, phenolic novolacs are prepared with a phenol / formaldehyde ratio greater than 1. As a consequence, few, if any reactive sites remain in the resin. Without the potential for further reaction, storage of novolacs under ambient conditions without viscosity drift is typical. Curing of novolacs is typically effected by adding a source of formaldehyde. The physical handling of formaldehyde or its derivatives (e. g. hexamine,
1

Introduction
The 1940s to 1960s brought about much development in the chemistry of phenolic resole and novolac resins [1-2]. However, it was not until the 1970s to 1980s that much effort was made in the reinforced composites industry to use these resins [3]. Early resin systems were acid cure, high viscosity, and high free formaldehyde. As of 1988, pultrusion had been attempted only on a trial basis with the pultruded rods failing in compression rather than tension in a flexural test, generally yielding only 80 85% of the mechanical properties of equivalent polyester rods [3]. By the late 1980s and early 1990s, safety issues were becoming more important, and phenol-resorcinol resins were shown to provide
Presented by ANGUS Chemical Company at Composites 2004

paraformaldehyde) however, poses significant worker exposure issues. The pultrusion resin systems studied during this work are two-part formulations consisting of phenolic novolac resins and non-formaldehyde hardeners. Stored individually, the components have a shelf life greater than 6 months at room temperature. The two components are mixed just prior to use. A mold release is added to facilitate processing. After mixing, the formulation has a pot life greater than 24 hours at room temperature. Because the resin component is a novolac rather than a resole, the pultrusion formulations do not need acid catalysis for rapid cure. These formulations, typically near pH 9, are also much less aggressive to the pultrusion die surface; hence, chrome plated dies are not required with this technology. Solvents such as ethanol are useful for adjusting formulation viscosity. A typical formulation contains 10 15 wt. % solvent. The non-formaldehyde hardener component reacts thermally with the novolac resin to generate a crosslinked phenolic polymer. No significant amount of water or formaldehyde is generated in these reactions. The ethanol, being an inert solvent, is discharged as a vapor at the end of the die. The ratio of resin to hardener may be varied to achieve the desired cured resin mechanical properties. Good mechanical properties are routinely generated with formulations based on 2.75 parts of resin to 1 part of hardener. These formulations result in crosslink densities of approximately 75%, but useful properties have been demonstrated over the wider range of 60% to 90% crosslink density.

Results and Discussion

Shelf Life: Due to the physical separation of the resin and hardener in a 2-part system, shelf life considerations are greatly simplified. Research has shown, that unlike resole-based systems, a novolac derived 2-part phenolic pultrusion resin system has excellent shelf life under ambient storage conditions. Refrigerated storage of the resin and / or rapid product turnover is not a requirement. The enhanced shelf life is not unexpected, as phenolic novolacs themselves have essentially infinite shelf life. The hardener is also extremely stable; its activity and viscosity have been demonstrated to be unchanged for over 1 year. The price to be paid for the excellent stability is in the additional formulation / mixing that must be performed on site prior to use. Pot Life: In addition to the excellent shelf life of the individual components, research testing has shown the pot life of the fully formulated phenolic novolac pultrusion resin system to be extremely long. Tests in the laboratory have shown that a fresh, 1500cps formulation requires 24 hours to double in viscosity at room temperature. While it is advantageous to minimize the solvent content of the pultrusion formulation, viscosity drift may be further managed (lowered) by adding small increments of ethanol solvent. A graph showing formulation viscosity versus ethanol content is seen in Figure 1. Pultrusion Line Speed: A critical parameter for any continuous manufacturing operation such as pultrusion is line speed. As the research literature has discussed, phenolic resole systems for pultrusion are notoriously slow. This is due in part to the curing chemistry itself and also the need to flash off large quantities of solvent / water generated in the cured part. The addition of blocked-acid catalysts to accelerate cure rates is sometimes used, but this causes detrimental corrosion to expensive dies. As a result, typical pultrusion rates of resole-based phenolic systems are in the 15 36 in/min range, depending on the part geometry. In contrast, the present novolac-based pultrusion systems have a much lower solvent content and may be pultruded at much higher temperatures than a conventional resole. The combination of these two factors results in a fast curing system that may be pultruded at very high rates. A study relating cure speed, cure temperature and part physical properties was performed. Using a uniform die temperature down the full length of the die, pultrusion was initiated at the lowest speed of interest. The pultrusion rate was then incrementally increased, with parts being collected at each rate until further increases in line speed caused an under-cured part to result. The parts were then subjected to 3-point flexural testing to determine their integrity and the effect of pultrusion conditions on properties. The parameters measured were peak flexural stress, modulus, and peak strain to failure. As seen in Figure 2, increased die temperature allowed processing at higher speeds, albeit with a sacrifice in the
2

Experimental Procedure
Pultrusion Processing: A nominally 1 x 0.125 (25.4 mm x 3.2 mm) rectangular profile reinforced with 68% unidirectional E-glass by volume was produced for all of the pultrusion experiments. All of the pultrusion processing experiments were conducted at the University of Mississippi using commercial pultrusion equipment with an open resin bath. Die zone temperatures from 400F to 550F were used for the three-zone, platen heating system with pull speeds ranging from 12in/min to 72 in/min. Processing data including measured line speed, pull load, and die temperatures were acquired using a computer during the pultrusion experiments. Property Evaluation: Mechanical and physical property testing of the resin and the pultruded composites was conducted to characterize the materials and pultruded product. ASTM D790 3-point flexural testing (16:1 span to depth) was conducted.

Presented by ANGUS Chemical Company at Composites 2004

ultimate physical properties obtainable. Pultrusion above 60 in/min at 525F was not attempted. The parts generated at 475F demonstrated excellent mechanical strength up to a pull rate of 52 in/min after which under-curing resulted in a rapid drop off in mechanical performance. Cross-link Density: One of the most important parameters in determining the performance of a thermoset system is the crosslink density. In a typical resole system, this is dictated by the phenol / formaldehyde ratio and cook cycle used in the resin manufacture. While it is always possible to under-cure a system, the maximum crosslink density attainable in a resole is pre-determined by the resin manufacturer. In contrast, the two part novolac system provides the freedom to incorporate as much, or as little, cross-linking agent in the formulation as desired. A study to examine the effects of cross-link density on the mechanical properties of a pultruded part was performed. Formulations containing resin / hardener ratios of 3.2:1, 2.7:1, and 2.3:1 were prepared. These ratios correspond to approximately 60%, 75% and 90% of the maximum theoretical crosslink density. The formulations were pultruded under identical conditions of die temperature (525F) and pull rate (52 in/min). As can be seen in Table 1, there was a small but predictable effect on the mechanical properties. As the cross-link density increased, the parts became more brittle, failing at lower peak strains. While the variation in mechanical properties of the different formulations was predictable, their pultrusion behavior was not. The two highest cross-link formulations pultruded very smoothly, with well behaved pull force profiles. The pull speeds of these formulations could be increased up to 60 in/min (75% cross-link) and 72 in/min (90% cross-link) although the mechanical properties of the parts decreased proportionately to the increased pull speed. In contrast, the low cross-link density material steadily built pull force during the trial. The pull speed was successfully raised to 60 in/min with a similar, proportional decline in mechanical properties, but the run was terminated in anticipation of locking up the die. The reason for this effect is still unclear.

when processed in the 475 500F range. The resulting optimized resin / hardener system is as follows: Resin / Hardener ratio 2.75:1 Formulation Viscosity 1200cps @ 25C Working Pot Life 24 hours Suggested Cure Temperature 475F The properties attainable with the above formulation / processing conditions are shown in Table 2. What Table 2 does not show is the dynamic operating range of the resin system. Start up conditions for initial trials can be extremely conservative, e.g. 12 in/min, without fear of seizing the die. Additionally, process upsets causing prolonged pull stoppages (1 3 min) have not caused operational problems. The surface finish attainable on all roving parts was exceptional, with high gloss, no surface dusting, blemishes or pitting. Environmental Health and Safety Considerations: As with any pultrusion resin system, environmental health and safety considerations are very important. Tests of the present resin system do not show any reportable quantities of free formaldehyde. Additionally, there is essentially no formaldehyde release during cure as demonstrated by the total lack of formaldehyde odor during processing. The absence of formaldehyde generation was also quantified by badge testing. The ethanol solvent is, of course, released and adequate ventilation to remove the vapors from the die exit is needed. FST: Testing has shown this novolac system to have enhanced processing and mechanical properties relative to typical resole-based pultrusion systems. To complete the resin characterization, it was important to show that these improvements did not come at a sacrifice to phenolics excellent FST properties. To this end, ASTM E662 fire testing was performed on pultruded parts. The results are shown in Table 3. These FST results indicate that the novolac-based resin system retains the key FST properties typical of phenolic resins. The high Tg (>200C) typical of phenolics is also retained. Comparison to Non-Phenolic Pultrusion Resins: While this paper has focused entirely on phenolicsystems, it is instructional to compare the processing and mechanical properties of the 2-part novolac pultrusion resins with more conventional polyester, vinyl ester and epoxy-based systems. A comparison between the phenolic novolac system and published data from other commercial pultrusion resins is seen in Table 4. The data was generated using the same pultrusion equipment, die and testing equipment. Pultrusion speeds exceeding that typical for vinyl ester, and approaching that of polyester, are easily obtainable. Strength and flexibility of the new system were found to be superior to all commercial resins tested. The excellent processing and mechanical properties of the two part novolac resin system coupled with its very high use temperature and FST characteristics provide a unique
3

Ultimate Performance: While each of the above studies demonstrates quantifiable effects of formulation or pultrusion conditions on performance, these initial pultrusion experiments were not designed to determine optimized process conditions. Utilizing the effects of formulation viscosity and stability, cure temperature, crosslink density and pull speeds, an optimization of the formulation for a given set of process conditions was performed. It was determined that pultrusion temperatures exceeding 500F were undesirable from a personal safety and operational viewpoint. Based on this, and the desire to maintain high processing rates, conditions were focused on creating a system that would give optimal pultrusion performance and excellent physical properties
Presented by ANGUS Chemical Company at Composites 2004

combination that should incite renewed interest in phenolic pultrusion resins.

Conclusions
Formaldehyde-free, 2-part phenolic novolac pultrusion resin systems have been shown to have excellent shelf and pot life. The processing conditions and mechanical properties of the new system have also been demonstrated to be far superior to typical pultrusion phenolics, in many cases rivaling vinyl ester and polyester resins. These improvements in properties are achieved without sacrificing the outstanding FST characteristics inherent with phenolic resins.

References
G.C. Little and K.W. Pepper, Proc XIth Int. Cong. Pure Apppl. Chem, London, 1947, 5-603. 2. K. Hultzch, Kunststoffe, 1947, 37-205. 3. K. L. Forsdyke, Phenolic Matrix Resins The Way to Safer Reinforced Plastics, 43rd Annual Conference, Composites Institute, SPI, February 1-5, 1988, 18C:1-9. 4. T.H. Dailey, Jr. and J. Shuff, Phenolic Resins Enhance Public Safety by Reducing Smoke, Fire, and Toxicity in RP, 46rd Annual Conference, Composites Institute, SPI, February 18-21, 1991, 15D:1-5. 5. James V. Gauchel, Richard N. Lehman, and Jay J. Beckman, Pultrusion Processing of Phenolic Resins, 48rd Annual Conference, Composites Institute, SPI, February 8-11, 1993, 2B:1-4. 6. T. H. Dailey, Jr., Ronald W. Allison, and Michael W. Klett, An Investigation of Flame, Smoke, Toxicity and Mechanical Properties of Pultruded Phenolic Composites, 48rd Annual Conference, Composites Institute, SPI, February 8-11, 1993, 21B:1-6. 7. T. H. Dailey and Michael W. Klett, Phenolic Pultrusion: Innovations for an Emerging Technology, 49rd Annual Conference, Composites Institute, SPI, February 7-9, 1994, 9D:1-4. 8. Shahid P. Qureshi, W. Hayes Ingram, and Clint Smith, High Performance Phenolic Pultrusion Resin, 51st Annual Conference, Composites Institute, SPI, February 5-7, 1996, 23D:1-6. 9. Michael W. Klett, A Cost-Effective Approach to High-Strength Phenolic Composites for Infrastructure Applications, 51st Annual Conference, Composites Institute, SPI, February 5-7, 1996, 23E:1-9. 10. J. G. Vaughan and E. Lackey, Tips on Phenolic Pultrusion, Composite Fabricators Association, San Antonio, TX, October 21-24, 1998. 11. J. Frank Mayfield and John G. Taylor, Advanced Phenolic Pultruded Grating for Fire Retardant Applications, 31st International SAMPE Technical Conference, 1999, 142-149. 1.

12. Shahid P. Qureshi, Recent Developments in Phenolic Resins Technology and Composites Applications, 31st International SAMPE Technical Conference, 1999, 150-162. 13. Lackey, E., J.G. Vaughan, F. Washabaugh, D. Usifer, and P. Ubrich, Effects of Fillers on Pultrusion Processing and Pultruded Properties, Interantional Composities EXPO 99 Proceedings - Composites Institute, 21E:, 1-10, 1999. 14. Lackey, E., J.G. Vaughan, R. Shanku, and T. Pettijohn, Experimental Examiniation of Promoted vs. Traditional Initiator Cure Systems for Pultrusion, Journal of Advanced Materials, 20(1), 10-15, 1998. 15. Chaitali Meda, Moisture Durability Study of Pultruded AOC P920/Glass Polyester Composite Materials, MS Thesis, Universisty of MS, 2003.

Biographies James G. Vaughan; Barnard Professor and Associate Dean, University of Mississippi. Ellen Lackey; Associate Professor, University of Mississippi, CCT. G. David Green; Product Research Leader, ANGUS Chemical Co. Raymond J. Swedo; Sr. Product Research Specialist, ANGUS Chemical Co. Thomas L. Johnson; Sr. Manager, Commercial Development, ANGUS Chemical Co. Daniel M. Trauth; Sr. Process Research Specialist, ANGUS Chemical Co.

Presented by ANGUS Chemical Company at Composites 2004

Tables

Table 1 Effect of Cross-Link Density on Physical Properties Resin / % of Peak Stress Modulus Peak Hardener Max.Theoretical (ksi) (Mpsi) Strain ratio Crosslink (%) 3.20 : 1 60 192 7.20 2.87 2.73 : 1 75 178 7.30 2.57 2.31 : 1 90 165 7.07 2.43 Table 2 Optimized Flexural Properties of 2-Part Phenolic Novolac Resin System Pultrusion 3-point Flexural Flexural Modulus Peak Flexural Speed (in/min) Stress (ksi) (Mpsi) Strain (%) 40 208 7.30 3.13 44 199 7.17 3.07 48 205 7.23 3.07 52 191 7.13 2.93

Table 3 ASTM E662 Smoke Generation Test Resin System Test Mode Ds 1.5 Ds 4.0 Ds Max Novolac 2-Part Flaming 0 1.4 13.4 Novolac 2-Part Non-Flaming 0 1.9 12.4

Table 4 Comparison of Phenolic Novolac Resin System with Typical Pultrusion Systems Resin System Pull Speed 3-point Flexural- Flexural Strain (in/min) Strength (ksi) to Failure (%) Typical epoxy / E glass [13] Typical vinylester / E glass [14] Typical polyester / E glass [15] Typical Phenolic Resole / E glass Phenolic Novolac 2-Part Resin System / E glass 18 48 60 12 52 169 177 184 152 191 2.5 NA 2.6 2.3 2.9

Presented by ANGUS Chemical Company at Composites 2004

Figure 1
Effect of Ethanol Content on Accelacure Formulation Viscosity Viscosity measured at ~ 230C 3000
Formulation Viscosity, cps

2500 2000 1500 1000 500 0 10 11 12 13 14 15 Ethanol Content, wt.%

Figure 2
Effect of Pultrusion Conditions on Flex Properties - Peak Stress
230 220 210 200 190 180 170 160 150 140 130 30 40 50 60 Pultrusion Speed (in/min)

Peak Stress (ksi)

Die Temp 475F Die Temp 525F Die Temp 450F Die Temp 500F

Presented by ANGUS Chemical Company at Composites 2004

Form No. 319-00611-11/01/04