Toughening of Epoxy Infusion Resins for Composites

by

Stephen C. Nolet Principal Engineer

TPI Composites, Incorporated

373 Market Street Warren, Rhode Island Tel: (401) 247-4010

Douglas J. Sober

Marketing Specialist - Epoxy

Kaneka Texas Corporation 1910 Almadale Farms Pkwy.

Collierville, Tennessee Tel: (713) 503-1558

Nick Miyatake Senior Chemist

Kaneka Texas Corporation

2 Northpoint Drive Houston, Texas

Tel: (713) 304-5971

Abstract

Many new composite applications for epoxy resins are for large or very large structural parts. Windmill rotor blades, bodies for people movers such as trains and buses, boats and military troop transport are ideally suited for manufacture by vacuum resin infusion (SCRIMP®) techniques. To be successful, these large parts need improved physical and thermal properties over current systems available. Many toughening agents already exist for epoxy resins and novel modifiers have been introduced to the marketplace recently. TPI Composites Inc. and Kaneka Texas Corporation have conducted a joint project to examine physical and thermal properties of several toughened resin systems at various loadings of modifier. The base premise was that the resin system be epoxy based and capable of being infused by the SCRIMP® process. The testing was conducted on neat resin castings as well as epoxy infused glass and carbon reinforced fabrics which include uni-directional and multi-axial laminates. Performance properties evaluated include thermo-mechanical properties tensile and compressive properties, fracture toughness (K1c and G1c), compression after impact (CAI), tension/tension fatigue, compressive strength, and others. This paper will discuss the test design and the test results of the composite matrixes.

I. Introduction

TPI Composites is currently engaged with the United States Army to design and build a composite variant for the DoD’s High Mobility High-Mobility Multipurpose Wheeled Vehicle (HMMWV). This effort, designated as the “All Composite Military Vehicle” (ACMV), will be completed with the delivery of three vehicles in the summer of 2007. Figure 1 shows the configuration of the current multi-part welded steel chassis and bolted up aluminum vehicle body. The ACMV configuration of the redesigned vehicle has passed through critical design review and the definition of OML and tooling is currently active. The approach has consolidated dozens of individual metallic parts into five primary structural components.

The configuration of the ACMV is illustrated in Figure 2. A primary composite frame will be vacuum resin infused capturing carbon fiber frame rails into a single integrally molded unit for attachment of drive components, transmission and suspension. The integrally molded body and roof structure will bolt up to the primary frame to create the passenger compartment.

Figure 1: Current production HMMWV welded steel chassis and bolted aluminum body pan. The purpose of this material conversion is to save weight on the baseline vehicle. The baseline vehicle is then heavily armored for crew protection. Current “up-armored” HMMWV’s suffer from a significant loss in payload capacity and are limited in the type and weight of protection that can be carried. A targeted weight savings of 700 lb in the GVW of the vehicle translates into nearly a 50% increase in payload or a 35% increase in crew protection via additional armor capacity. High performance composite laminates, and particularly those structures that are cyclically loaded at high frequency and subject to repeated impact benefit tremendously from the use of “toughened” matrix systems. The use of carbon fiber selective reinforcement drives the TPI HMMWV design toward epoxy resin system. Thermo-mechanical properties of the resin are very important given that regions of the structure are constantly exposed to elevated temperature from engine heat and exhaust.

2

Bolted Attachment of Composite Hatch Cover

Integrally Molded Roof

3

Figure 2: Consolidation of body and chassis components of ACMV. Five primary structural components will replace dozens of individual welded parts and bolted assemblies.

II. Toughening of Epoxy Resin Matrices

The traditional approach of using CTBN (carboxyl terminated butadiene acrylonitrile) to increase fracture toughness and elongation typically increases resin viscosity while reducing thermo-mechanical properties of a given baseline epoxy resin. This trade-off often drives the use of material forms that reduce fiber volume fraction and requires the use of greater section in a given structure resulting in added weight in the final product. A novel alternative to CTBN resin toughening is the incorporation of “Core Shell Rubber” into the epoxy resin blend. Core Shell Rubber (CSR) particles have been used to toughen various polymeric systems since the early 1960s. A simplified depiction of a typical CSR particle can be found in Figure 3. As the name “Cores Shell Rubber” implies, a spherical core of one composition is covered with a shell of another. The core of the CSR particle is synthesized by an emulsion polymerization process and the chemistry of the core provides the impact modification performance and toughening desired. A shell of a second chemistry is then grafted onto the core with the dual purpose of both protecting the core and generating compatibility with the polymer being modified. The core is typically a butadiene / styrene copolymer similar to a rubber compound and the shell is of an acrylate character. Historically over the last 40 years, literally hundreds of CSR products have been synthesized by various manufacturers differentiated by core composition, shell chemistry, size and geometry. The product is generally furnished as a solid powder

Integrally Molded Lower Body/ Floor pan

Integrally Molded Composite Cross Members

Primary Frame - Composite

Bolted or Integrally Molded Composite Bumpers

Integrally Molded Frame Members

Single Integral Molding

comprised of agglomerated CSR particles that have been dried. Powder CSR particles have been successfully used in thermoplastic systems such as PVC and PET. Figure 4 shows the typical applications of the over 1 million metric tons of CSR used globally each year.

Figure 3: Design of Core Shell Rubber (CSR) Particle.

Figure 4: Applications for CSR as Toughening Agent.

CSR particles toughen a polymeric matrix in several different ways. The first theory is that the core itself acts to dissipate the energy of an impact. The cavitation theory of impact modification is that the core deforms effectively mitigating the force of the impact itself. This theory is supported by the change of refractive index in the cores of the CSR near the impact area itself. The composition of the core as well as the overall size of the CSR can change the toughening performance of the CSR. The CSR itself can be optimized for the particular application. The second theory of toughening by CSR is that the particles themselves act as crack terminators. Cracks or fissures that start by whatever mechanism are stopped when they meet a CSR particle. Therefore having small and evenly dispersed CSR in the polymeric matrix increases the probability of the crack meeting a CSR particle. The third theory of toughening is via stress relief. During the curing process, internal stresses can be generated especially in geometrically complex structures. The rubber core “absorbs” making the part less susceptible to microcracking. Any polymeric matrix exhibiting “cohesion” failure can be improved using CSR nano-particles as supported by T-Peel, Lap Shear, and Fracture toughness data. The importance of an even dispersion of CSR or any toughening agent is to maximum the number of effective sites for cavitation crack termination and stress reduction. In addition, a small, homogenously dispersed CSR provides a reproducible result in terms of fracture toughness and adhesion characteristics lot to lot or batch to batch. Consistence is often as important as the absolute value improvement in toughening. Attempts to mix CSR powder with epoxy and other thermosetting resins have resulted in very poor dispersions characterized by rapid settling and inherently non-homogeneous levels of toughness. A lot of effort has been expended trying to mill,

4

2-Layer CSR

Co-Polymer II designed to be compatible with thermosetting resin. This is the shell. The shell is grafted onto the core.

Co-Polymer I designed for impact resistance. This is the core. Normally the core has a chemical link to “rubber”. MBS

Resin

PVC

Application

packaging (bottle, film)PC auto (bumper, panel)

PBT auto (airbag cover)

ABS SAN auto, appliance

Acrylic PVC

PMMA

window, siding, fence

auto, lighting, signage

Volume (mt/yr.)

250,000

Acrylic

350,000

500,000

10,000

grind or shear the solid CSR particles into epoxy resin. Once the particles are agglomerated however, it is impossible to break the solid particle up into the primary CSR domains first synthesized. Kaneka Texas Corporation has developed a process which delivers the CSR particles in various epoxy resins where the individual particles are not agglomerated but are evenly dispersed throughout the liquid resin. Figure 5 shows a beaker containing a 10% dispersion of solid CSR on the left versus the beaker on the right containing Kane Ace® MX 120 product comprised of a 25% dispersion in Bis A epoxy resin. In the Kane Ace® product, the 100 nm CSR particles are evenly dispersed and stable without settling or agglomeration for at least 1 year. Therefore the toughening agent is easily mixed into existing systems and no further agitation is required to keep the CSR in even suspension as a raw material or in the formulated system. Figure 6 shows a cured composite that has been cross-sectioned and stained to highlight the CSR particles within the matrix. The 100 nanometer particles (0.1 micron) are evenly dispersed providing numerous sites for cavitation and/or crack termination.

Figure 5: Dispersion characteristics of 10% solid CSR on left versus 25% CSR MX 120, both in liquid Bis-A epoxy resin.

Figure 6: Cross section showing 100 nm MX 120 CSR in composite polymer matrix.

III. Resin and Laminate Test Program

TPI Composites, with the technical help and support of both Kaneka Texas Corporation (Houston, Texas) and Epoxical (S. St. Paul, MN) have developed a rigorous test plan to investigate the performance of CTBN toughened and CSR toughened epoxy resin developed specifically for use in advanced vacuum assisted resin infusion processes. The investigation utilizes a base-line Bis-A epoxy resin and cyclo-aliphatic curing agent adducted with varying levels of both CTBN or CSR materials. The Kaneka product used to deliver the CSR particles was Kane Ace® MX 125 while the CTBN used for the test matrix was a commercially available CTBN used in formulating resins for composites.

5

The purpose of this extensive project is to quantify significant performance parameters (both process related and mechanical) as a function of the level of adduction. Completion of this effort will provide significant guidance for TPI design staff and manufacturing engineers to choose the most appropriate material system to integrate into the HMMWV project.

Table 1 and outlines the test plan that is currently in progress. Specimen fabrication and testing includes: neat resin casting, and laminates fabricated from continuous reinforcement forms of either glass or carbon fiber. Composite laminates were resin infused as either unidirectional or quasi-isotropic plates. Table 1 shows the test program accomplished on neat cast resins. A control casting of the baseline epoxy (no toughening material adducted) is included in the project. Table 2 outlines the specimens and testing accomplished on unidirectional laminate and Table 3 contains the test plan for specimens and testing for quasi-isotropic laminates.

Resin System Toughener (pph) Cure Conditions/Schedule Test Description Test Std # of Samples

CTBN x13 (0) Mold Temperautre (°C)Assignment of the Glass Transition

Temperatures by Differential Scanning Calorimetry

E1356-03 3

CTBN x13 (5) 25 °C Test Method for Plastics: Dynamic Mechanical Properties D4440-01 3

CTBN x13 (7.5) Demold Time (hr)Test Method for Assignment of the Glass

Transition Temperature By DYNAMIC MECHANICAL ANALYSIS

E1640-04 3

CTBN x13 (10) 24hrTest Methods for Plane-Strain Fracture

Toughness and Strain Energy Release Rate of Plastic Materials

D5045-99 5

Post Cure (°C)/(hr) Test Method for TENSILE Properties of PLASTICS D638-03 5

Test Method for VISCOSITY of RESIN Solutions D1725-04 3

Test Method for GEL TIME and Peak Exothermic Temperature of Reacting

Thermosetting RESINSD2471-99 3

Mold Temperautre (°C)Assignment of the Glass Transition

Temperatures by Differential Scanning Calorimetry

E1356-03 3

MX 125 (25%CSR) (5) 25 °C Test Method for Plastics: Dynamic Mechanical

Properties D4440-01 3

MX 125 (25%CSR) (7.5) Demold Time (hr)

Test Method for Assignment of the Glass Transition Temperature By DYNAMIC

MECHANICAL ANALYSISE1640-04 3

MX 125 (25%CSR) (10) 24hr

Test Methods for Plane-Strain Fracture Toughness and Strain Energy Release Rate of

Plastic MaterialsD5045-99 5

Post Cure (°C)/(hr) Test Method for TENSILE Properties of PLASTICS D638-03 5

Test Method for VISCOSITY of RESIN Solutions D1725-04 3

Test Method for GEL TIME and Peak Exothermic Temperature of Reacting

Thermosetting RESINSD2471-99 3

Epoxical XCSR-XXXA/B

A-Side

Epoxical XCTBN-XXXA/B

100 PPH A-Side 31/29/28/27 PPH B-Side

Formulation

100 PPH A-Side 29/28/27 PPH B-Side

B-Side

4 hr @ 66°C

4 hr @ 66°C (150 °F)

Resin (phr) Hardener (phr)

A-Side B-Side

Table 1: Neat Resin Casting Test Matrix

It should be noted that all test series include specimens manufactured utilizing the control (no adduction with additional tougheners). In all cases a standard cure cycle was implemented that included a four hour post-cure at 66 °C (150 °F). All testing has been completed in the physical sciences laboratory at TPI Composites. Structural tests were conducted on a computer controlled MTS-810 servo-hydraulic universal load frame. Both 350 ohm strain gages and an extensometer were used to collect strain data.

6

Resin System Toughener (pph) Reinf. Laminate Cure Schedule Test Description Test Std # of Samples

CTBN x13 (0) [010] Infusion Temperautre (°C) Standard Test Method for Tensile Properties of Polymer Matrix Composite Material D3039-00e2 5

CTBN x13 (5) [9010] 32 °C (90 °F) Standard Test Method for Tensile Properties of Polymer Matrix Composite Material D3039-00e2 5

CTBN x13 (7.5) [08] Demold Time (hr)Standard Test Method for Mode I Interlaminar Fracture Toughness of Unidirectional Fiber-

Reinforced Polymer Matrix CompositesD5528-01 5

CTBN x13 (10) [908] 24 hr

Post Cure (°C)/(min) 3

[010] Infusion Temperautre (°C) Standard Test Method for Tensile Properties of Polymer Matrix Composite Material D3039-00e2 5

MX 125 (25%CSR) (5) [9010] Standard Test Method for Tensile Properties of

Polymer Matrix Composite Material D3039-00e2 5

MX 125 (25%CSR) (7.5) [08] Demold Time (min)

Standard Test Method for Mode I Interlaminar Fracture Toughness of Unidirectional Fiber-

Reinforced Polymer Matrix CompositesD5528-01 5

MX 125 (25%CSR) (10) [908] 24 hr

Post Cure (°C)/(min) 3

Formulation

Epoxical XCTBN-XXXA/B

Resin (phr) Hardener (phr)

4 hr @ 66°C (150 °F)

4 hr @ 66°C (150 °F)100 PPH A-Side 31/29/28/27 PPH B-Side

A-Side B-Side

B-SideEpoxical

XCSR-XXXA/B

100 PPH A-Side

A-Side

29/28/27 PPH B-Side

Vectorply E-T 1500 Uni-

Dir E-Glass

Vectorply C-LA 1312 Uni-Dir Carbon

Vectorply E-T 1500 Uni-

Dir E-Glass

Vectorply C-LA 1312 Uni-Dir Carbon

Reinforcement Detail

Standard Test Method for Tension-Tension Fatigue of Polymer Matrix Composite Materials D3479-02e1

Standard Test Method for Tension-Tension Fatigue of Polymer Matrix Composite Materials D3479-02e1

Table 2: Uni-directional Laminate Test Matrix

Resin System Toughener (pph) Reinf. Laminate Cure Schedule Test Description Test Std # of Samples

CTBN x13 (0)Vectorply E-LT 2400 E-

GlassInfusion Temperautre (°C) Standard Test Method for Tensile Properties of

Polymer Matrix Composite Material D3039-00e2 5

CTBN x13 (5)Vectorply E-BX 2400 E-

Glass32 °C (90 °F) Standard Test Method for Compressive

Properties of Rigid Plastics D695-02a 5

CTBN x13 (7.5)Vectorply C-

LT 2200 Carbon

Demold Time (hr)Test Method for Compressive Residual

Strength Properties of Damaged Polymer Matrix Composite Plates

D7137-05e1 3

CTBN x13 (10)Vectorply C-

BX 2400 Carbon

24 hr

Post Cure (°C)/(min) 3

Vectorply E-LT 2400 E-

GlassInfusion Temperautre (°C) Standard Test Method for Tensile Properties of

Polymer Matrix Composite Material D3039-00e2 5

MX 125 (25%CSR) (5)

Vectorply E-BX 2400 E-

Glass

Standard Test Method for Compressive Properties of Rigid Plastics D695-02a 5

MX 125 (25%CSR) (7.5)

Vectorply C-LT 2200 Carbon

Demold Time (min)Test Method for Compressive Residual

Strength Properties of Damaged Polymer Matrix Composite Plates

D7137-05e1 3

MX 125 (25%CSR) (10)

Vectorply C-BX 2400 Carbon

24 hr

Post Cure (°C)/(min) 3

Resin (phr) Comp #4 (phr)

D3479-02e1 (MODIFIED)

Standard Test Method for Tension-Tension Fatigue of Polymer Matrix Composite Materials

D3479-02e1 (MODIFIED)

[0/90/±45]s

[0/90/±45]sEpoxical

XCSR-XXXA/B

4 hr @ 66°C (150 °F)

4 hr @ 66°C (150 °F)

Epoxical XCTBN-XXXA/B

100 PPH A-Side 29/28/27 PPH B-Side

Formulation Reinforcement Detail

[0/90/±45]s

[0/90/±45]s

Standard Test Method for Tension-Tension Fatigue of Polymer Matrix Composite Materials

100 PPH A-Side 31/29/28/27 PPH B-Side

Table 3: Quasi-Isotropic (cross-plied) Laminate Test Matrix

At the time of this writing all neat resin casting tests have been completed and reported. Static uni-directional (both axial and transverse) tests have been completed for E-glass specimens. All specimens for the investigation, except for double cantilevered beam testing (use to obtain mode I strain energy release rate, GIC) have been manufactured. Carbon fiber specimens, quasi-isotropic cross-plied laminates and fatigue testing will be completed during the next six month period and will be reported shortly thereafter.

7

IV. Results

All seven resin systems have completed mixed viscosity testing at room temperature (approximately 25 °C). 1000 gram samples of each formulation were blended in accordance with specified stoichiometry. Each specimen was mixed for 60 seconds and viscosity was measured on a Brookfield DV-II viscometer using an LV3 spindle at 60 RPM. These specimens were then cast into 12” by 12” by ¼” thick plates and cured in accordance with the schedule outline in Table 1. These castings were used to harvest specimens for all neat resin testing. A Perkin-Elmer DSC-7 thermal analysis unit was used for differential scanning calorimetry (DSC). The DSC unit was used to determine the shift in the heat flow baseline curve as samples were heated from room temperature to a maximum temperature of 120 °C. Tg is often associated with the temperature at the measured inflection point of the baseline shift as illustrated in figure 7 below. Figure 7 is the DSC scan of the Control Epoxy (no toughening). A Tg of 75.4 °C has been determined.

Figure 7: Temperature scan via DSC of control resin sample.

Figures 8 and 9 contrast the thermal behavior at 5phr of CTBN versus the CSR adducts. In all the DSC testing completed the CTBN resin appears to exhibit a greater loss in Tg at the 5 phr level than resin adducted with CSR at 5phr. However, at higher loadings the Tg of both resin systems appear to converge (7.5 phr and 10.0 phr) and remain relatively insensitive to the loading. These results will be verified by additional testing by dynamic mechanical analysis (DMA) using the same test coupons produced by the experimental design.

8

Figure 8: Temperature scan via DSC of resin sample containing 5phr of CTBN.

Figure 9: Temperature scan via DSC of resin sample containing 5phr of CSR.

Table 4 summarizes viscosity and Tg (as determined by DSC) data. The data shows a clear trend of increasing viscosity and decreasing Tg as a function of toughener loading. However, it is also useful to note that the viscosity build of the CSR loaded resin is significantly less than the CTBN adducted resins. This is a very important factor in considering the infusion of reinforcements for high fiber volume fraction composite laminates.

9

Figure 10 and 11 graphically illustrates the room temperature viscosity and Tg data. Typically, infusion processes for high performance (fiber volume fraction > 45%) reinforced systems require resin viscosities below 350 cps. The addition of a small amount of heating can rapidly reduce viscosity which is traded off against resin pot life; so in all cases a lower starting viscosity is always desirable.

Measured Viscosity Tg via DSC Resin Sample ID

(cPs) (°C)

XCTRL 292 75.4

XCTBN 5.0 484 67.5

XCTBN 7.5 636 67.9

XCTBN 10.0 866 66.5

XCTRL 292 75.4

XCSR 5.0 378 72.3

XCSR 7.5 446 67.2

XCSR 10.0 580 67.2

Table 4: Neat Resin Viscosity data and Tg of cast and post-cured resin plates used for tensile testing and fracture toughness testing.

Epoxy Viscosity as a Function of Toughener Loading (phr)

200

300

400

500

600

700

800

900

0.0 2.0 4.0 6.0 8.0 10.0

Toughener Loading (phr)

Visc

osity

(cPs

)

CTBN EpoxyCSR Epoxy

Figure 10: Room temperature (T = 25 °C) resin viscosity versus toughener loading (phr) for both CTBN and CSR loaded epoxy resin.

10

Epoxy Tg as a Function of Toughener Loading (phr)

65.0

70.0

75.0

80.0

0.0 2.0 4.0 6.0 8.0 10.0

Toughener Loading (phr)

Tg

(°C

)

CTBN EpoxyCSR Epoxy

Figure 11: Tg as determined by Differential Scanning Calorimetry versus toughener loading (phr) for both CTBN and CSR loaded epoxy resin.

Tensile testing of cured neat resin castings illustrate the trend toward higher elongation and reduced modulus as a function of increased loading of the toughening agents (see figure 12). The practices of ASTM D-638 were employed. The data is summarized in Table 5 below and represents the mean of five test results for each resin type. It should be noted that the baseline resin itself shows extremely high elongation as a starting point (72,000 micro-strain is equivalent to 7.2% strain). It is important to note that all the adducted epoxies indicate a loss in tensile strength and modulus. While the tensile strength of the CTBN system is less sensitive to lower loading of the toughener, the CSR adducted epoxy shows a much higher rate of elongation with increased toughener loading. Loss of modulus is similar in both material systems as a function of increasing rubber content. In all cases, the effect of additional resin toughening is apparent on the strain to failure. It is clear that elongation of the resin can be moderated by the level of toughening agent added and is traded off with the resin modulus and ultimate strength.

11

0

1,000

2,000

3,000

4,000

5,000

6,000

7,000

8,000

9,000

10,000

0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18

Strain (in/in)

Stre

ss (p

si)

XCTRL

XCTBN 5.0XCTBN 7.5

XCTBN 10.0

XCSR 5.0

XCSR 7.5

XCSR 10.0

Figure 12: Tensile testing of neat resin casting tensile bar IAW ASTM D-638. The coupon under test is a 5 phr CSR loaded epoxy casting. The data shows greatly enhanced elongation of material with increased loadings of CSR and less impact utilizing CTBN.

Max Stress Strain Elongation Modulus Resin ID (psi) (min/in) at Failure (in) (psi)

Epoxy CTRL 9,879 72,398 0.197 424,623

5.0 phr CTBN 8,850 81,782 0.205 388,519

7.5 phr CTBN 8,566 82,824 0.206 379,344

10.0 phr CTBN 7,754 126,078 0.279 343,520

5.0 phr CSR 8,258 97,876 0.257 380,317

7.5 phr CSR 8,250 135,580 0.322 361,846

10.0 phr CSR 7,693 184,760 0.367 341,459

Table 5: Static test results of cast tensile bar specimens.

Fracture toughness testing using a single edge notch bending (SENB) specimen was completed for all resin castings at each level of toughening agent adducted to the epoxy resin. This testing was carried out in accordance practices outlined in ASTM D-5045. Figure 13 shows the 7.5phr CSR cast material under three-point bending with a carefully

12

machined edge notch and crack initiated at mid span. Fracture toughness is determined according to the geometry of the specimen (including crack geometry, span, depth of beam and thickness) and maximum load attained at a fixed cross-head stroke rate of 0.3”/min.

Figure 13: Single edge-notch bending specimen for determination of neat resin fracture toughness. Specimen shown is epoxy loaded with 7.5phr CSR.

Figure 14, below, summarizes results from the SENB fracture toughness testing. Again it should be noted that the baseline resin system exhibits a very high initial fracture toughness of 2.33 MPa-m1/2. Both the CTBN and CSR adducts initially result in higher measure values of fracture toughness at a 5.0phr loading. Additional loading results in a downward trend in the fracture toughness measurement, but in all cases the average KQ is greater than the unmodified control resin sample. The data does suggest that the CSR adduct is a more effectual agent for increasing the baseline KQ.

2.595

2.298

2.264

2.935

2.7042.580

2.610

1.4

1.8

2.2

2.6

3.0

3.4

0.0 2.5 5.0 7.5 10.0 12.5

Loading of Toughening Adduct (phr)

XCTBN Epoxy ResinXCSR Epoxy Resin

Frac

ture

Tou

ghne

ss (M

pa-m

^1/2

)

Figure 14: Fracture Toughness, KQ, of neat resin as measured by single edge notched bending specimens (SENB).

13

Tensile testing of laminates fabricated from uni-directional e-glass knitted broad goods infused with the adducted epoxy has been completed. Infusion of the high fiber volume fraction (vf > 52%) laminates was accomplished by using the Seemans Composite Resin Infusion Molding Process (SCRIMP®). This vacuum resin infusion method included the use of a heated aluminum plate as the molding surface to allow for mild heating to reduce resin viscosity as typical in an industrial application. Reinforcements were heated to 32 °C (90 °F) during the resin infusion. After infusion heat was turned off and the plate was allowed to cure for 24 hours. After 24 hours the laminate was demolded and given a 4 hour free-standing post cure at 66 °C (150 °F). Straight edged tensile coupons were cut from the SCRIMP®’ed panels. Five coupons were machined with fibers aligned with the longitudinal axis of the coupon (0)° and five coupons were harvested with fibers aligned transverse to the longitudinal axis of the coupon (90°). Fiberglass bonding tabs with a 30° beveled edge were bonded to each side on either end of the test coupons. Instrumentation for each tensile test specimen included a 350 ohm axial strain gage as well as a clip-on extensometer for strain measurement. Generally, the strain gage was used for determination of modulus and the extensometer provided full range measurement of ultimate elongation (strain to failure). The photograph shown in figure 15 illustrates the typical tensile specimen loaded in the jaws of the MTS 810 universal load frame.

Figure 15: Longitudinal (°0) unidirectional e-glass tensile test specimen in load frame. Typical stress-strain results are shown in figure 17 which is a group of results from the coupons infused with 5.0phr CSR epoxy resin. In all cases the stress-strain behavior was

14

highly linear as would be expected in this fiber dominated loading range. A summary of the longitudinal (°0) tensile test results is shown if Table 6.

0

20,000

40,000

60,000

80,000

100,000

120,000

140,000

0.00E+00 5.00E-03 1.00E-02 1.50E-02 2.00E-02 2.50E-02Strain (in/in)

Stre

ss (p

si)

Sample #1Sample #2Sample #3Sample #4Sample #5

Figure 16: Longitudinal (°0) tensile test of uni-directional e-glass specimens infused with 5.0phr CSR epoxy resin. Results show high linearity and little effect on static tensile strength of fiber dominated laminates loaded with toughening agents.

Max Stress Strain Elongation Modulus Resin ID (psi) ( in/in) at Failure (in) (psi)

Epoxy CTRL 112,778 21,724 0.362 5.14E+06

5.0 phr CTBN 95,246 19,735 0.414 4.87E+06

7.5 phr CTBN 102,640 19,790 0.390 4.99E+06

10.0 phr CTBN 95,311 18,808 0.387 5.01E+06

5.0 phr CSR 111,293 21,900 0.381 5.24E+06

7.5 phr CSR 106,548 20,880 0.356 5.13E+06

10.0 phr CSR 108,349 20,684 0.340 5.17E+06

Table 6: Summary of Longitudinal (°0) uni-directional e-glass resin infused composite test specimens. Tensile strength of CTBN infused specimens was modestly affected by resin system modification. The results shown in Table 6 indicate a modest decrease in tensile strength for composites utilizing the CTBN modification. CSR modified resins do not appear to affect static

15

tensile strength of fiber dominated composite laminate. Modulus data is unaffected (within margins of measurement) by resin adducts. Testing of transverse (90°) uni-directional coupons shows similar trending as the neat resin coupons, which is not surprising for this resin dominated property. Figures 17 and 18 shown the stress-strain behavior of the unmodified control epoxy specimens and the test specimens infused with 5.0phr CSR epoxy. The increased non-linearity of the stress-strain curve is noticeable in the modified epoxy. Table 7 summarizes all the transverse (90°) uni-directional tensile data. In all cases, except for the 5.0phr CSR epoxy resin specimens both the tensile strength and modulus trended downward. However, unlike the neat resin coupon the strain to failure or specimen elongation was not significantly improved. This may suggest some incompatibility between the resin and the fiber sizing (finish) that is reducing the fiber/matrix adhesion strength. It is notable that the 5.0phr CSR test specimen exhibited a modest, but not insignificant increase in all transverse (90°) properties.

0

1,000

2,000

3,000

4,000

5,000

6,000

7,000

8,000

9,000

0.00E+00 1.00E-03 2.00E-03 3.00E-03 4.00E-03 5.00E-03 6.00E-03 7.00E-03Strain (in/in)

Stre

ss (p

si)

Sample #1Sample #2Sample #3Sample #4Sample #5

Figure 17: Transverse (°90) tensile test of uni-directional e-glass specimens infused with unmodified (control) epoxy resin.

16

0

1,000

2,000

3,000

4,000

5,000

6,000

7,000

8,000

9,000

10,000

0.00E+00

1.00E-03

2.00E-03

3.00E-03

4.00E-03

5.00E-03

6.00E-03

7.00E-03

8.00E-03

9.00E-03

1.00E-02

Strain (in/in)

Stre

ss (p

si) Sample #1

Sample #2Sample #3Sample #4Sample #5

Figure 18: Transverse (°90) tensile test of uni-directional e-glass specimens infused with 5.0phr CSR epoxy resin.

Max Stress Strain Elongation Modulus

Resin ID (psi) ( in/in) at Failure (in) (psi)

Epoxy CTRL 7,217 5,528 0.040 1.43E+06

5.0 phr CTBN 6,380 6,982 0.042 1.13E+06

7.5 phr CTBN 6,444 6,498 0.045 1.23E+06 10.0 phr CTBN 5,761 5,653 0.040 1.24E+06

5.0 phr CSR 8,498 7,884 0.055 1.44E+06

7.5 phr CSR 6,766 6,462 0.041 1.21E+06

10.0 phr CSR 6,307 5,653 0.040 1.24E+06

Table 7: Summary of Transverse (°90) uni-directional e-glass resin infused composite test specimens. With the exception of specimens adducted with 5.0phr CSR, transverse properties of the test specimens trend downward.

17

The tensile strength and modulus data from tables 6 and 7 are graphically represented in figures 19 and 20. Not surprisingly he fiber dominated properties of the longitudinal (°0) are not significantly affected, with the notable exception of the 5phr loaded CTBN tensile strength (and to a lesser degree modulus). Given the recover of these properties at the 7.5 phr loading, it seems likely that additional data may indicate simple experimental variance as opposed to a physical interaction. Transverse properties (resin dominated) are impacted more significantly as shown in figure 20. The 5.0phr CSR loaded epoxy actually shows a

small but not insignificant increase in tensile strength. This coupled with the increased strain to failure of the specimen and its retention of modulus is very attractive.

40,000

50,000

60,000

70,000

80,000

90,000

100,000

110,000

120,000

0 2.5 5 7.5 10

Loading (phr)

Long

itudi

nal T

ensi

le S

tren

gth

(psi

)

3.00E+06

3.50E+06

4.00E+06

4.50E+06

5.00E+06

5.50E+06

6.00E+06

6.50E+06

7.00E+06

Long

itudi

nal M

odul

us (p

si)

XCTBN Strength

XCSR Strength

XCTBN Modulus

XCSR Modulus

Figure 19: Graphical test results of longitudinal (°0) unidirectional E-glass laminate tensile testing.

1,000

2,000

3,000

4,000

5,000

6,000

7,000

8,000

9,000

0 2.5 5 7.5 10

Loading (phr)

Tran

sver

se T

ensi

le S

tren

gth

(psi

)

6.00E+05

8.00E+05

1.00E+06

1.20E+06

1.40E+06

1.60E+06

1.80E+06

2.00E+06

2.20E+06

2.40E+06

2.60E+06

Tran

sver

se M

odul

us (p

si)

XCTBN Strength

XCSR Strength

XCTBN Modulus

XCSR Modulus

Figure 20: Graphical test results of transverse (°90) unidirectional E-glass laminate tensile testing.

V. Discussion

The test data show that the use of either CTBN or Core Shell Rubber (CSR) in a Bis-A epoxy resin system can significantly change the behavior of that resin system as well as the properties of the composite laminates fabricated using these resins. It is clear that both CTBN and CSR increases the room temperature viscosity of the base resin, however, the

18

19

CSR modified resins exhibit lower rates of increasing viscosity as a function of loading. In all cases laminates were successfully infused using a modest level of heating that can be easily accomplished in an industrial environment. It is apparent, however, that the CSR technology offers processing advantages for vacuum resin infusion of large complex structures owing to the lower viscosity. An important trend that is emerging at this stage of the project is that levels of rubber modification above 5phr have a significant adverse effect on most static resin and composite properties. Glass-transition temperature is expected to trend downward with increasing rubber modification and this is just the case. At low levels (5phr) the CSR had only modest effects on Tg. All mechanical properties, neat resin tensile strength and modulus, longitudinal and transverse composite strength and modulus trend downward for loading levels above 5phr. Even fracture toughness appears to peak at a loading of 5.0phr for both the epoxy modified with CTBN and CSR materials. It is also important to note that the CSR modified epoxy (at a loading of 5phr) resulted in significant mechanical improvements over the baseline resin and CTBN adducted resin in almost every aspect (except for the slight decline in Tg compared to the unmodified epoxy and the fiber dominated longitudinal tensile strength of the composite). Significant attributes of the 5phr loading of CSR include only a small decline in resin modulus with a significant increase in strain to failure. In addition the CSR modified epoxy demonstrates better fiber/matrix adhesion compared to the CTBN modified epoxy. This test program has been designed to evaluate the effects of two different resin adduction technologies intended to improve composite structures. These resin adducts, CTBN and Core Shell Rubber have demonstrated a significant impact on the properties of epoxy resin. It is conceivable that using both CTBN and CSR in the same formulated systems may provide an overall synergistic effect in terms of these performance parameters. TPI Composites, Inc. and Kaneka Texas Corporation will continue to build test specimens and move into cyclic load studies, Mode I strain energy release rate of laminates, as well as damage tolerance testing (compression after impact) to evaluate the most important aspect of resin modification: long term durability of structures molded from these modified resins.

Acknowledgements

TPI Composites, Inc would like to gratefully acknowledge Luis Hinosa, Project Manager and John Stepowski, Program Engineer for the United States Army’s All Composite Military Vehicle Program located at the U.S. Army Tank Automotive Research, Development and Engineering Center, (TARDEC) in Warren, Michigan for funding this project and his personal interest in our efforts. TPI and Kaneka Texas owe a debt of gratitude to Ann Jackason and Eric Frank of Epoxical of S. St. Paul, Minnesota for their extensive support in the formulating of test resins used in this study and the expertise of their personnel which proved invaluable in developing these materials.