Miscibility and Morphologies of Poly(arylene ether phenylphosphine oxide/sulfone) Copolymer/Vinyl Ester ResinMixtures and Their Cured Networks

SHENG WANG, JIANLI WANG, QING JI, A. R. SHULTZ, T. C. WARD, J. E. MCGRATH

Chemistry Department and National Science Foundation Science and Technology Center for High Performance PolymericAdhesives and Composites, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061-0344

Received 14 January 2000; revised 26 June 2000; accepted 28 June 2000

ABSTRACT: Nonreactive bisphenol A-based poly(arylene ether triphenyl phosphine ox-ide/diphenyl sulfone) statistical copolymers and a poly(arylene ether triphenyl phos-phine oxide) homopolymer, each having a number-average molecular weight of about 20kg/mol, were synthesized and solution-blended with a commercial dimethacrylate vinylester resin. Free-radical cured systems produced morphologies that were a function ofboth the amount of phosphonyl groups and the weight percentage of the copolymers.For example, highly hydrogen-bonded poly(arylene ether phenyl phosphine oxide) ho-mopolymer/vinyl ester resin mixtures were homogeneous in all proportions both beforeand after the formation of networks. Copolymers containing low amounts (#30 mol %)of the phosphonyl groups displayed phase separation either before or during cure. Thephase-separated cured materials generally had phase-inverted morphologies, such as acontinuous thermoplastic copolymer phase and a particulate, discontinuous vinyl esternetwork phase, except for systems containing a very low copolymer content. The resinmodified with a copolymer containing 30 mol % phosphine oxide comonomer showedimproved fracture toughness, suggesting the importance of both phase separation andgood adhesion between the thermoplastic polymer and the crosslinked vinyl ester fillerphase. The results suggested that the copolymers with high amounts of phosphine oxideshould be good candidates for interphase sizing materials between a vinyl ester matrixand high-modulus carbon fibers for advanced composite systems. Copolymers with lowamounts of phosphonyl groups can produce tough, vinyl ester-reinforced plastics. Thechar yield increases with the concentration of bisphenol A poly(arylene ether phosphineoxide) content, suggesting enhanced fire resistance. The incorporation of thermoplasticcopolymers sustains a high glass-transition temperature but does not significantlyaffect the thermal degradation onset temperature. © 2000 John Wiley & Sons, Inc. J PolymSci B: Polym Phys 38: 2409–2421, 2000Keywords: vinyl ester; bisphenol A poly(arylene ether phenyl triphosphine oxide/diphenyl sulfone); copolymers; hydrogen bonding; scanning electron microscopy (SEM);transmission electron microscopy (TEM); fracture toughness; interphase material; mor-phologies; miscibility

INTRODUCTION

Thermosetting structural adhesive and compositematrix networks based on epoxy and dimethacry-

late vinyl ester (VE) resin oligomers are impor-tant materials that display a number of excellentmechanical and adhesive properties (Scheme 1).1

These oligomers are widely used as adhesives andthermosetting matrix materials for composites re-inforced with glass, carbon, and aramide fibers.However, unmodified networks exhibit low-im-pact resistance in comparison with most engi-

Correspondence to: J. E. McGrath (E-mail: jmcgrath@vt.edu)Journal of Polymer Science: Part B: Polymer Physics, Vol. 38, 2409–2421 (2000)© 2000 John Wiley & Sons, Inc.

2409

neering thermoplastics. Many approaches havebeen investigated to address these problems. Theincorporation of reactive liquid rubbers, for exam-ple, butadiene–acrylonitrile copolymers, is awidely used method to improve fracture tough-ness, especially of cured epoxy resins.2–9 One dis-advantage of this approach is that polydienerubbers degrade properties such as stiffness,strength, heat distortion temperature, and oxida-tive and chemical resistance. An alternative ap-proach to improve fracture toughness is to incor-porate high-performance thermoplastics, eitherreactive or nonreactive, into epoxy resins.10–15

Relatively few attempts have been made to im-prove the impact resistance of cured VE becauserubber additives such as vinyl-terminated buta-diene–acrylonitrile copolymers are not misciblewith bisphenol A (BPA)-based hydroxy etherdimethacrylate styrene solutions.16–18 Very littlehas been published on using high-performancethermoplastic materials to toughen commercialVE resins.

It is hypothesized that adhesion improvementsbetween a network matrix and fibers might beachieved with polymeric sizings on the fiber thatare miscible with the network matrix and aremore adherent to the fiber surface. This might beachieved via specific interactions.19 In our labora-tory, various phosphine oxide-containing poly-mers, such as poly(arylene ether)s,20 polyim-

ides,21 and cyanate oligomers,22 were prepared,and investigations on a poly(arylene etherphenyl phosphine oxide/sulfone) copolymer/phe-noxy resin model blend system showed that themiscibility of the two components could beachieved; this depended on the molar content ofphosphine oxide in the copolymer.23,24 The spe-cific interaction (hydrogen bonding) between thephosphonyl groups of the copolymer and hydroxylgroups of the phenoxy resin was demonstrated tobe the driving force for miscibility.24 However,studies on the miscibility between phosphine ox-ide-containing polymers and thermosetting oli-gomers such as epoxy and VE resin have not beenreported. Relatively few thermoplastic polymershave been used to improve the interfacial adhe-sion of a VE matrix to the major reinforcing fi-bers.25 This research has systematically studiedthe effect of phosphonyl content in the copolymeron its miscibility with a VE resin and on themorphologies and toughness of the cured VEsystem.

EXPERIMENTAL

Materials

BPA was kindly supplied by Dow Chemical Co.The activated halide monomer 4,49-dichloro di-

Scheme 1. Structures of BPA poly(arylene ether triphenyl phosphine oxide) ho-mopolymer and diphenyl sulfone copolymers, phenoxy resin, and related systems.

2410 WANG ET AL.

phenyl sulfone (DCDPS) was provided by BP–Amoco. The monomer 4,49-bis(fluorophenyl) phe-nyl phosphine oxide (BFPPO) was either synthe-sized in our laboratory or provided by Zeneca.20

Commercial VE resin (Derakane 411-400; 28 wt% styrene) was supplied by Dow Chemical. Anexperimental sample of methacrylate-endcappedVE (;690 g/mol) without styrene was also do-nated by Dow Chemical and was used as received.A liquid initiator, t-butylperoxybenzoate (Ato-chem), was used as received.

Synthesis of BPA Poly(arylene ether phenylphosphine oxide/sulfone) Statistical Copolymers

Phenolic hydroxyl-terminated BPA poly(aryleneether phenyl phosphine oxide/sulfone)s with anumber-average molecular weight of 20 kg/molwere synthesized via aromatic nucleophilic sub-stitution reactions with a controlled excessamount of BPA. The molar ratio of the reactantmonomers required for obtaining the final prod-uct with the desired molecular weight was calcu-lated according to the modified Carothers equa-tion.26 A 1-L, four-necked, round-bottom flaskequipped with an overhead stirrer, a nitrogeninlet, and a Dean–Stark trap with a reflux con-denser was employed for all of the polymeriza-tions. As an example, a copolymer with 20 mol %phosphine oxide was synthesized according to thefollowing schedule: 38.0944 g (0.1669 mol) of BPAwere added along with 10.2528 g (0.03263 mol) ofBFPPO and 37.4743 g (0.1305mol) of DCDPS.The amount of BFPPO and DCDPS can be variedwith various molar ratios, depending on the tar-get composition of the copolymer. A 15 mol %excess of potassium carbonate (26.5 g, 0.192 mol)was also charged as the weak base for thenucleophilic reaction. A dimethylacetamide/tolu-ene mixture (70/30 v/v; 560 mL/240 mL) was usedas the initial reaction medium. The reaction washeated to a 135 °C reflux for 4 h under a nitrogenatmosphere to dehydrate the system. Then, mostof the toluene was removed, and the reaction tem-perature was kept at 160 °C for 16 h; thisachieved a high molecular weight. The viscousliquid was allowed to cool to room temperature,was diluted with dimethylacetamide, and was fil-tered; acetic acid was added to protonate the phe-nolate end groups. The polymer was precipitatedinto methanol, redissolved in chloroform, againprecipitated into methanol, filtered, and dried ina vacuum oven at 150 °C for 24 h.

Resin Preparation and Curing

The VE control sample was prepared by the ad-dition of a 1 wt % liquid initiator, t-butylperoxy-benzoate (t-butylperoxybenzoate/dimethacrylate/styrene molar ratio ' 1/20/52), to the VE resin atapproximately 80 °C. The mixture was stirreduntil homogeneous and then was degassed in avacuum oven for about 1 min. Care was taken toavoid a significant loss of the styrene reactivediluent. Thermoplastic copolymers containing 20mol % or more phosphonyl groups were success-fully dissolved at 5–20 wt % into the commercialVE resin. When the mixture was homogeneous,the 1 wt % initiator was added, and the mixturewas degassed. Both the control and modified mix-tures were added to preheated silicone rubbermolds at 100 °C and covered with a steel plate inan air convection oven. The resin was cured at100 °C for 1 h and postcured at 140 °C for anotherhour and then cooled slowly in the oven.27

Characterization

Intrinsic viscosities were measured in chloroformat 25°C. Gel permeation chromatography (GPC)measurements were performed to characterizethe molecular weights and molecular weight dis-tributions of the thermoplastic copolymers.N-Methyl pyrrolidone containing 0.02 M phospho-rus pentoxide was employed as the solvent at 60°C.28 A Waters 150C instrument, with a differen-tial refractive index detector and a Viscotek dif-ferential viscometer in parallel, allowed the cal-culation of absolute molecular weights with theuniversal calibration technique.

Fourier transform infrared (FTIR) measure-ments utilized a Nicolet Impact 400 instrumentwith a resolution of 2 cm21 for an average of 256scans. The copolymer, the styrene-free VE resin,and their blends with various thermoplastic mod-ifier compositions were dissolved in chloroform.The solutions were cast onto NaCl plates anddried overnight in a vacuum oven at 120 °C toremove the solvent.

Glass-transition temperatures (Tg’s) of thethermoplastic copolymer/VE resin cured blendswere measured with a PerkinElmer DSC-7 differ-ential scanning calorimeter at a heating rate of 10°C/min. All the results reported were obtainedduring a second heat after cooling from 250 °C.The midpoint temperature of the specific heattransition during the second heat was taken asthe value of Tg.

MISCIBILITY AND MORPHOLOGIES 2411

A PerkinElmer DMA-7e instrument was em-ployed for dynamic mechanical analysis (DMA)measurements. The samples were analyzed withthe three-point bend mode at a frequency of 1 Hzand a heating rate of 5 °C/min.

Thermogravimetric analyses (TGAs) were per-formed with a PerkinElmer TGA-7 instrument.Samples of 4–6 mg were heated at 10 °C/minfrom 25 to 800 °C in an air atmosphere.

The fracture surfaces of the modified resinswere examined by scanning electron microscopy(SEM) with a model ISI-SX-40 at an acceleratingvoltage of 20 or 30 kV. Fracture surfaces wereproduced at ambient temperature. All of the frac-ture surfaces were coated with Au to avoid charg-ing problems during the electron beam scans.

Transmission electron microscopy (TEM) wasconducted to determine the morphologies of thecured polymer modified VE networks. The barsamples were trimmed and ultramicrotomed witha Reichert–Jung Ultracut E equipped with a dia-mond knife. The sample thicknesses were about700 Å. The electron micrographs were taken witha Philips 400T TEM instrument with an acceler-ation voltage of 100 kV.

The critical-stress-intensity factor, KIC, a mea-sure of plane-strain fracture toughness, wasmeasured by the single-edge-notched-bendingmethod, following ASTM Standard D 5045 91.The molded samples were ground with emery pa-per to obtain a rectangular shape (3 3 6 3 40mm). Each specimen was sawed to generate anotch; then, a crack was initiated by tapping witha liquid-nitrogen-chilled fresh razor blade. Thesamples were tested in the three-point bendingmode at a test rate of 1 mm/min on an Instronmodel 1123 apparatus. Approximately 10 sam-ples of each blend were tested, and the fracturetoughness values were calculated by the formulaprovided in ASTM D 5045 91.

RESULTS AND DISCUSSION

The intrinsic viscosity and GPC results for thehydroxyl-terminated BPA poly(arylene ether phe-nyl phosphine oxide/sulfone) homopolymer andcopolymers are shown in Table I. The data showthat the target number-average molecularweight, 20 kg/mol, was achieved. The homopoly-mer had a Tg of 200 °C. The objective of thisinvestigation was to focus on the effect of thechemical structure of these copolymers on theproperties of the thermoplastic-modified net-works. The goal was to achieve good mechanicalproperties while a relatively low viscosity wasmaintained in the thermoplastic/VE blends.

Hydrogen Bonding

Further understanding of the specific interactionbetween the hydroxyl group of VE and the phos-phonyl group of BPA poly(arylene ether phenylphosphine oxide) was achieved with FTIR mea-surements.24 Figure 1 displays infrared spectra of

Table I. GPC and Intrinsic Viscosity Characterization of Hydroxyl-Terminated BPA-Based Poly(arylene ethertriphenyl phosphine oxide/diphenyl sulfone) Homopolymer and Copolymers

Sample

BFPPO/DCDPS GPC [h] (dL/g) [h] (dL/g)

Molar Ratio Mn Mw Mw/Mn NMP (60 °C) CHCl3 (25 °C)

BPA–P20 20/80 21.9 42.1 1.9 0.31 0.40BPA–P30 30/70 21.3 40.8 1.9 0.32 0.37BPA–P50 50/50 22.7 42.0 1.8 0.30 0.35BPA–P100 100/0 24.2 57.9 2.4 0.24 0.45

Figure 1. FTIR spectra of BPA–P100/VE (withoutstyrene) blends at various weight ratios recorded atroom temperature in the hydroxyl stretching region:(A) BPA–P100, (B) VE, (C) VE/BPA–P100 (80/20 w/w),and (D) VE/BPA–P100 (50/50 w/w).

2412 WANG ET AL.

the VE, the thermoplastic modifier, and theirblends. For simplicity, only a commercial VEwithout styrene was employed. The hydrogen-bonded hydroxyl-stretching vibration of the VEwas originally at about 3445 cm21. When it wasblended with BPA-based poly(arylene ether phe-nyl phosphine oxide), it shifted to a lower fre-quency, as reported earlier for the phenoxy sys-tem.24 At 50 wt % BPA–P100, the VE hydrogen-bonded hydroxyl-stretching band peak showed asignificant shift down to 3300 cm21. The resultsindicate a strong hydrogen-bonding specific inter-action between the hydroxyl group of the oligo-meric dimethacrylate VE resin and the phospho-nyl group. Although the samples were dried in avacuum oven at 120 °C, a partial polymerizationof VE should not affect our conclusions becauseonly the band shift of the hydrogen-bonded hy-droxyl group needed to be considered.

Miscibility

The phosphine oxide containing the miscibilitiesof the copolymers with a commercial VE resinwere examined. Table II qualitatively shows thatthe miscibility of the VE/copolymer systems de-creased with increasing copolymer concentrationand increased with the amount of phosphine ox-ide (i.e., the hydrogen-bonded site) in the copoly-mer. The commercial VE (Derakane 411-400) con-tained 28 wt % styrene, which is a nonsolvent forthe BPA-based poly(arylene ether sulfone) con-trol. The 10 mol % BFPPO copolymer was onlyslightly soluble in this commercial resin. Increas-ing the amount of phosphine oxide in the copoly-mer significantly increased solubility in the VEoligomer/styrene solution because of the hydrogen

bonding between hydroxyl groups of VE resin andthe phosphonyl groups. When the resin blendswere cured, phase separation could be observed tooccur with copolymers that had low phosphineoxide content. In contrast, the resin modified bythe homopolymer poly(arylene ether phenyl phos-phine oxide) (BPA–P100) was transparent beforeand after curing. These results showed that theincorporation of phosphine oxide significantly im-proves the miscibility between the copolymer andVE network matrix. The copolymer modifier wasrestricted to 20 wt % or less because of the high-viscosity considerations.

These results can be qualitatively explainedaccording to the thermodynamic equation, DG5 DH 2 TDS, where DG, DH, DS, and T are thechange of the Gibb’s free energy, change of en-thalpy, change of entropy, and temperature, re-spectively. DG is largely dependent on DH in thespecific interaction system. A strong attractiveinteraction (hydrogen bonding) between BPA-based poly(arylene ether sulfone/phosphine ox-ide) yields DH ! 0, so DG can become negative.With DG , 0, a homogeneous mixture of the co-polymer and the VE resin can be generated. Dur-ing network formation of the VE resin, DS be-comes smaller, and phase separation can occur, ifthe attractive interaction between the copolymerand the cured VE resin is not strong.

Morphology

One goal was to study the effect of triaryl phos-phine oxide concentration in the copolymers onthe miscibility of and/or the adhesion to the curedVE resin network. The fracture surface morphol-ogies of the free-radical cured composites were

Table II. Influence of Composition on the Miscibility of BPA-Based Poly(arylene ether triphenyl phosphineoxide/diphenyl sulfone) Copolymer and VE Resin before Curing and after Curing

Blend

BPA–Px in the Blends (wt %)

5 10 15 20

Sol’n Poly Sol’n Poly Sol’n Poly Sol’n Poly

VE/BPA–P20 E 2 E 2 E 2 E 2VE/BPA–P30 1 2 1 2 E 2 E 2VE/BPA–P50 1 1 1 1 1 E 1 E

VE/BPA–P100 1 1 1 1 1 1 1 1

Sol’n 5 solution, polymer dissolved in commercial VE; poly 5 polymer network after curing; 1 5 transparent; E 5 translucent;2 5 opaque. BPA–P10 is only partially soluble in commercial VE.

MISCIBILITY AND MORPHOLOGIES 2413

characterized by SEM. Figure 2 shows the effectof phosphine oxide content on the morphologies at20 wt % copolymer loadings. Lower amounts ofphosphine oxide in the copolymers (BPA–P20,BPA–P30) produced mixtures that phase-sepa-rated with the cured VE network to produce thedispersed phase and the thermoplastic copolymeras the continuous phase. Increasing phosphonylconcentration (BPA–P30) produced a less sharpphase boundary between the copolymer and theVE. A further increase to 50 mol % resulted in afuzzy boundary, no doubt because of high misci-bility with even the cured VE. The homopolymer(BPA–P100) generated a homogeneous networkbecause of the high miscibility of the two compo-nents. The morphology also depends on the con-centration of the thermoplastic copolymer. As anexample, Figure 3 shows the morphology changewith the increase of the weight percentage of

BPA–P30 in the BPA–P30/VE mixtures. The co-polymer formed a dispersed phase in the VE net-work at 5 and 10 wt % copolymer loadings, butcopolymer BPA–P30 showed continuous morphol-ogy with deformed VE spheres at the two higherloadings. This latter condition has also beentermed phase inverted morphology.15 In these sys-tems, phase inversion occurred at 15 wt % BPA–P30.

TEM was also used to investigate the morphol-ogies of the cured networks. Figure 4 shows agradual change of morphology with increasingamounts of phosphine oxide in the copolymer at20 wt % copolymer loadings. The darker phaserepresents the BPA–Px phase because of itshigher electron density relative to the VE phase.Both BPA–P20- and BPA–P30-modified VE net-works exhibited phase separation; the copolymersformed the continuous phase, and the VE net-

Figure 2. SEM micrographs of the fracture surfaces of cured VE resin modified withBPA–Px with 20 wt % BPA–Px: (A) VE/BAP–P20, (B) VE/BAP–P30, (C) VE/BAP–P50,and (D) VE/BAP–P100.

2414 WANG ET AL.

works formed the dispersed phase. The BPA-30modified VE had a relatively narrow size distri-bution of VE network particles in comparisonwith the BPA–P20 system. Phase separation wasbarely discernable in the cured BPA-50/VE mix-ture, and the cured PBA–P100/VE mixture wasclearly homogeneous. These TEM results wereconsistent with the SEM results on the fracturesurfaces. Apparently, weak hydrogen bonding be-tween phosphonyl groups of the BPA–Px copoly-mers and the hydroxyl groups of the VE is suffi-cient to solubilize the precursor. However, highconcentration afforded strong hydrogen bonding,leading to homogeneity even in the cured net-work. Figure 5 shows the morphologies of variousweight percentages of BPA–P30 in cured BPA–P30/VE mixtures. At 5 wt % BPA–P30, the copol-ymer appears to be the dispersed phase in a con-tinuous VE matrix. At 10 wt % BPA–P30, cocon-

tinuous phases of the copolymer and VE networkappear to exist. At 15 and 20 wt % loadings of theBPA–P30, the thermoplastic copolymer forms thecontinuous phase and the VE network forms thedispersed phase in the cured mixtures. Thisphase-inversion behavior is consistent with theSEM results.

The chemical resistances of the cured VE resincontrol and cured VE/copolymer mixtures wereexamined by the immersion of sample bars inchloroform. The cured VE resin control bar re-mained clear and transparent even after a monthof immersion. However, the cured bars containing20 wt % BPA–P20 or BPA–P30 copolymer becamewhite and began to disintegrate. After about amonth, these samples disintegrated into smallpieces. The cured bars with higher weight per-centages of copolymer disintegrated faster. Curedbars containing 5 or 10 wt % BPA–P30 were

Figure 3. SEM micrographs of the fracture surfaces of cured VE resin modified withBPA–P30 at various weight-percentage loadings: (A) 5% BPA–P30, (B) 10% BPA–P30,(C) 15% BPA–P30, and (D) 20% BPA–P30.

MISCIBILITY AND MORPHOLOGIES 2415

Figure 4. TEM micrographs of cured VE resin modified with BPA–Px at 20 wt %loadings: (A) VE/BAP–P20, (B) VE/BAP–P30, (C) VE/BAP–P50, and (D) VE/BAP–P100.

2416 WANG ET AL.

Figure 5. TEM micrographs of cured VE resin modified with BPA–P30 at variousweight-percentage loadings: (A) 5% BPA–P30, (B) 10% BPA–P30, (C) 15% BPA–P30,and (D) 20% BPA–P30.

MISCIBILITY AND MORPHOLOGIES 2417

highly resistant to chloroform and did not disin-tegrate even after a month of immersion. A curedbar containing 15 wt % BPA–P30 disintegrated inchloroform. SEM and TEM measurements bothshowed that the cured mixtures that disinte-grated in chloroform had VE microgel particlesdispersed in a continuous thermoplastic copoly-mer phase. Their lack of solvent resistance isreadily understood. Nondisintegrating cured mix-tures were observed when the copolymer existedas the discontinuous (or cocontinuous?) phasewithin the VE network. Cured mixtures of VE/BPA–P100 were homogeneous and resistant toBPA–P100 leaching from the VE network.

These observations indicate that it will be nec-essary to have reactive end groups12–15 or reac-tive groups pendent on the copolymer chains29 toassure that the thermoplastic copolymer will ex-ist as dispersed, phase-separated toughener do-mains. This need for the toughener to partiallyreact with the matrix material to prevent inver-sion during the network formation has been notedin step-growth reaction matrix systems. In theVE resin free-radical reaction, gelation occurs atvery low conversions. Rapid autoacceleration ofthe polymerization in regions where gel hasformed tends to yield discrete microgel particles,with the thermoplastic toughener expelled as acontinuous surrounding phase. This phase inver-sion occurs at lower toughener concentrationsand at less unfavorable network/toughener inter-actions in free-radical chain-growth reactionsthan in step-growth reactions. The need for areactive toughener in the former instance is,therefore, more stringent.

Effect of the Modifier on the Thermal Stability andMechanical Properties

Thermal Stability

TGA results showed that the incorporation of 20wt % of the copolymers or BPA–P100 homopoly-mer did not significantly affect the 5% weight-losstemperature, probably because the initial degra-dative process begins with the thermoset fraction(Fig. 6). In the high-temperature regime, the charyield increased with the phosphine oxide contentof the copolymer. As known, poly(arylene etherphenyl phosphine oxide/sulfone)s are fire-resis-tant copolymers.20(b) Therefore, the incorporationof the copolymers as tougheners in VE resinshould and did increase the char yield.

Thermal Transition Behavior

Figure 7 presents the tan d (mechanical loss tan-gent) versus temperature curves of cured VEresin modified with BPA–P30 at various compo-sitions. Similarly, Figure 8 lists the tan d versustemperature curves of VE resin modified withBPA–Px at 20 wt % loadings. The DMA resultsshowed a slight increase of Tg of the modifiedresins because the modifiers had higher Tg’s (Fig.9). In all the resins, the tan d versus temperaturecurves showed only a single damping peak nomatter what the loading and composition in thecopolymers were. Although the copolymer and VEnetworks were phase-separated, the tan d peak ofthe higher Tg could not be reached before samplefailure. Phase separation was expected in the net-works modified with copolymers containing lowamounts of phosphine oxide. Although cocontinu-ous phases seemed to be observed on the fracturesurfaces at high loadings of copolymers by SEM,TEM showed only crosslinked VE particles sur-rounded by a continuous thermoplastic phase.The Tg’s of the thermoplastic polymers24 aremuch higher than that of the VE network,27 andthe amount of thermoplastics polymer is rela-tively low. Therefore, because of experimentallimitations, differential scanning calorimetry(DSC) could not detect the Tg of the thermoplasticpolymers in the presence of the rubbery VE net-work. Only a single Tg that was nearly the sameas that of the cured VE control was observed(Fig. 10).

Fracture Toughness

Table III shows the stress concentration factors,KIC, of various cured VE/copolymer mixtures at

Figure 6. TGA of cured BPA–Px/VE with 20 wt %BPA–Px at a heating rate of 10 °C/min in air. Thenumbers in parentheses are the temperatures at 5%weight loss.

2418 WANG ET AL.

four copolymer contents. The cured control VEwas very brittle. For all the compositions exam-ined, the fracture toughness was significantly im-proved with the incorporation of only 5 wt % ofthermoplastic polymer. High amounts of thermo-plastic copolymers led to significantly improvedfracture toughness. However, these materialsconsisted of thermoplastic copolymer reinforcedby VE microgel filler particles. The fracturetoughness increased to more than two times thatof the control with a 20 wt % loading of the copoly-mer BPA–P30. These results showed that bothphase separation and interaction between the VEand the thermoplastic polymer were importantfor the improvement of the fracture toughness. As

is well known, phase separation between the twocomponents is often a requirement to increaseimpact strength. Also, good adhesion helps totransfer stress from the continuous phase to thereinforcing filler particles. Small amounts ofphosphine oxide in the copolymer may not enablemaintenance of good adhesion, whereas largeamounts can lead to insufficient phase separationor even homogeneous systems. Therefore, aproper choice of the phosphine oxide content al-lows for both phase separation and good adhesion.Good phase separation with still strong adhesionbetween a thermoset network and a modifier cansignificantly increase the fracture toughness inreactive modifier systems.15 In a previous inves-tigation, bis(4-fluorophenyl)-3-aminophenylphos-phine oxide (amino DFTPPO) was incorporated asa comonomer into a BPA-based poly(sulfone). The

Figure 7. Mechanical loss tan d of cured BPA–P30/VE at various BPA–P30 weight percentages at1 Hz and a heating rate of 2 °C/min.

Figure 8. Loss tan d of cured BPA–Px/VE mixtureswith 20 wt % BPA–Px at 1 Hz and a heating rate of2 °C/min.

Figure 9. Tg (tan d peak temperature) of cured BPA–Px/VE mixtures at various loadings.

Figure 10. DSC thermograms of cured VE/BPA–Pxmixtures at an 80/20 (w/w) composition. Arrows indi-cate the location of Tg.

MISCIBILITY AND MORPHOLOGIES 2419

resultant copolymers were used to toughen anepoxy resin. Only 2.5 mol % of the reactive phos-phine oxide comonomer in the copolymer wasneeded to optimize the toughness,29 even thoughno detailed study of the morphology was at-tempted. In a separate study, the BPA–Px copoly-mers were used to toughen an epoxy resin: onlylow concentrations of phosphonyl groups in thecopolymers were needed to significantly increasethe fracture toughness.30

CONCLUSIONS

Mixtures of nonreactive BPA-based poly(aryleneether triphenyl phosphine oxide/diphenyl sulfone)copolymers with a number-average molecularweight of 20,000 with a commercial VE resin ex-hibited increasing miscibility as a function ofphosphine oxide content. FTIR spectra of BPA-based poly(arylene ether triphenyl phosphine ox-ide) homopolymer/VE (no styrene) mixturesshowed that strong hydrogen bonding betweenthe hydroxyl groups of the VE and the phosphonylgroups of the copolymer was the driving force forthe miscibility of the two components. Althoughthe polysulfone control was insoluble, the copoly-mer mixed homogeneously in all proportions ifthe phosphonyl groups exceeded 50 mol %. Thehomopolymer also formed homogeneous mixtureswith a commercial VE resin, and the networksappeared to be transparent after curing. This sug-gests the possible use of the homopolymer orphosphine oxide-containing copolymers as inter-phase materials between a VE resin matrix andglass, carbon, or aramide fibers.

The morphologies of cured poly(arylene ethertriphenyl phosphine oxide/diphenyl sulfone) co-polymer/VE resin mixtures depended both on thephosphine oxide content of the thermoplastic co-polymer and on the weight fraction composition of

the mixtures. Only at 10 wt % or less of thecopolymer with a sufficient molar fraction of phe-nyl phosphine oxide units did the thermoplasticform a discrete internal phase in a crosslinked VEresin matrix. In general, phase inversion oc-curred, producing a continuous thermoplasticphase reinforced by discrete microgel VE resinparticles. Although a similar phase inversion oc-curs during cures in systems with nonreactivethermoplastics mixed with step-growth network-forming resins (e.g., epoxies), the phase inversionis greatly encouraged by the rapid, low-conver-sion attainment of gel in the free-radical curesystem. This further points out the need to havereactive groups on the thermoplastic componentof the latter systems to assure that it is forced toform the internal toughening phase.

The fracture toughness of all the cured thermo-plastic-modified VE resin materials exceeded thatof the cured VE resin control. However, in mostinstances the materials were composed of thethermoplastic reinforced by high loadings of par-ticulate VE resin. The improved fracture tough-ness does indicate good adhesion of the thermo-plastic to the cured VE resin.

The incorporation of copolymer in the VE resinsystem did not significantly influence the 5%weight-loss temperature. Also, the observed Tg ofthe cured VE resin remained about the same asthat of the unmodified cured resin. Both DMAand DSC could detect the Tg of the thermoplasticpolymer, but only the Tg of the cured VE resinwas detected. However, because of the phase-in-verted morphologies of most of the networks,their solvent (chemical) resistances were poor.The amorphous continuous thermoplastic phasewas extracted from the samples, and the materi-als disintegrated to yield the discrete microgelparticles of the cured VE resin. As discussed pre-viously, reactive thermoplastic modifiers are ex-pected to provide the desired toughening and in-

Table III. Critical-Stress-Intensity Factor KIC (MPa z m0.5) of Polymer-Modified Vinyl Ester

Sample

Weight-Percentage Loading of the Polymer in the Network

5 10 15 20

VE/BPA–P20 0.55 6 0.05 0.54 6 0.08 0.50 6 0.08 0.64 6 0.08VE/BPA–P30 0.64 6 0.10 0.66 6 0.08 0.68 6 0.22 0.80 6 0.11VE/BPA–P50 0.56 6 0.10 0.66 6 0.09 0.66 6 0.13 0.63 6 0.09VE/BPA–P100 0.52 6 0.15 0.47 6 0.07 0.47 6 0.09 0.63 6 0.13

KIC of the control VE was 0.36 6 0.09 MPa z m0.5.

2420 WANG ET AL.

creasing chemical resistance of the cured VEresin matrices, as was observed in analogous ep-oxy systems.15

The authors appreciate support from the National Sci-ence Foundation Science and Technology Center forHigh Performance Polymeric Adhesives and Compos-ites and the Office of Naval Research. Many thanks aredue to Mr. Steve McCartney for the TEM measure-ments and to Mr. Frank Cromer for the SEM measure-ments. Dr. H. Li is also acknowledged for her experi-mental help.

REFERENCES AND NOTES

1. International Encyclopedia of Composites; Lee,S. M., Ed.; VCH: Weinheim, 1990; Vols. 1–5.

2. Sultan, J. N.; McGarry, F. J. Polym Eng Sci 1973,13, 29.

3. Bucknall, C. B.; Yoshii, T. Br Polym J 1978, 10, 53.4. Rubber Modified Thermoset Resins; Riew, C. K.;

Gillham, J. K., Eds.; Advances in Chemistry Series208; American Chemical Society: Washington, DC,1984.

5. Butta, E.; Levita, G.; Marchetti, A.; Lazzeri, A.Polym Eng Sci 1986, 26, 63.

6. Jakusik, R.; Jamaani, F.; Kinloch, A. J. J Adhes1990, 32, 245.

7. Kasemura, T.; Kawamoto, K.; Kashima, Y. J Adhes1990, 33, 19.

8. Guild, F. J.; Kinloch, A. J. J Mater Sci 1995, 30,1689.

9. Verchere, D.; Pascault, J. P.; Sautereau, H.; Mos-chiar, S. M.; Riccardi, C. C.; Williams, R. J. J ApplPolym Sci 1991, 42, 701.

10. Bucknall, C. B.; Partridge, I. K. Polymer 1983, 24,639.

11. Bucknall, C. B.; Gilbert, A. H. Polymer 1989, 30,213.

12. Hedrick, J. L.; Yilgor, I.; Wilkes, G. L.; McGrath,J. E. Polym Bull 1985, 13, 201.

13. Hedrick, J. L.; Yilgor, I.; Jurek, M.; Hedrick, J. C.;Wilkes, G. L.; McGrath, J. E. Polymer 1991, 21,2020.

14. Wilkinson, S. P.; Ward, T. C.; McGrath, J. E. Poly-mer 1993, 34, 870.

15. Yoon, T. H.; Priddy, D. B., Jr.; Lyle, G. D.;McGrath, J. E. Macromol Symp 1995, 98, 673.

16. Ullett, J. S.; Chartoff, R. P. Polym Eng Sci 1995, 35,1086.

17. Dreerman, E.; Narkis, M.; Siegamann, J. R.; Do-diuk, H.; Dibenedetto, A. T. J Appl Polym Sci 1999,72, 647.

18. Burchill, P. J.; Pearce, P. J. In CRC Encyclopedia ofPolymer Materials; Salamone, J. C., Ed.; CRC:Boca Raton, FL, 1994; pp 104–112.

19. Lesko, J. J.; Swain, R. E.; Cartwright, J. M.; Chin,J. W; Reifsnider, K. L.; Dillard, D. A.; Wightman,J. P. J Adhes 1994, 45, 43.

20. (a) Smith, C. D.; Grubbs, H. J.; Webster, H. F.;Gungor, A; Wightman, J. P.; McGrath, J. E. HighPerform Polym 1991, 4, 211; (b) Riley, D. J.; Gun-gor, S. A.; Srinivasan, S. A.; Sankarapandian, M.;Tchatchoua, C.; Muggli, M. W.; Ward, T. C.;McGrath, J. E.; Kashiwagi, T. Polym Eng Sci 1997,37, 1501.

21. Tan, B.; Tchatchoua, C. N.; Dong, L.; McGrath,J. E. Polym Adv Technol 1998, 9, 84.

22. Abed, J. C.; Mercier, R.; McGrath, J. E. J Polym SciPart A: Polym Chem 1997, 35, 977.

23. Srinivasan, S.; Kagumba, L.; Riley, D. J.; McGrath,J. E. Macromol Symp 1997, 122, 95.

24. Wang, S.; Ji, Q.; Tchatchoua, C. N.; Shultz, A. R.;McGrath, J. E. J Polym Sci Part B: Polym Phys1999, 37, 1849.

25. Broyles, N. S.; Verghese, K. N. E.; Davis, S. V.; Li,H.; Davis, R. M.; Lesko, J. J.; Riffle, J. S. Polymer1998, 39, 3419.

26. Odian, G. Principles of Polymerization, 2nd ed.;Wiley Interscience: New York, 1981; p 113.

27. Li, H.; Rosario, A. C.; Davis, S. V.; Glass, T.; Hol-land, T. V.; Davis, R. M.; Lesko, J. J.; Riffle, J. S. JAdv Mater 1997, 28, 55.

28. (a) Konas, M.; Moy, T. M.; Rogers, M. E.; Shultz,A. R.; Ward, T. C.; McGrath, J. E. J Polym Sci PartB: Polym Phys 1995, 33, 1429; (b) Konas, M.; Moy,T. M.; Rogers, M. E.; Shultz, A. R.; Ward, T. C.;McGrath, J. E. J Polym Sci Part B: Polym Phys1995, 33, 1441.

29. Pak, S. J.; Lyle, G. D.; Mercier, R.; McGrath, J. E.Polymer 1993, 34, 885.

30. Wang, J.; Wang, S.; Ji, Q.; McGrath, J. E.; Ward,T. C. Manuscript in preparation, 2000.

MISCIBILITY AND MORPHOLOGIES 2421