lable at ScienceDirect

Polymer 55 (2014) 5073e5079

Contents lists avai

Polymer

journal homepage: www.elsevier .com/locate/polymer

Synthesis and characterization of a renewable cyanateester/polycarbonate network derived from eugenol

Benjamin G. Harvey a, *, Andrew J. Guenthner b, Gregory R. Yandek b, Lee R. Cambrea a,Heather A. Meylemans a, Lawrence C. Baldwin a, Josiah T. Reams c

a US NAVY, NAWCWD, Research Department, Chemistry Division, China Lake, CA 93555, USAb Air Force Research Laboratory, Rocket Propulsion Division, Edwards AFB, CA 93524, USAc ERC, Inc., Air Force Research Laboratory, Rocket Propulsion Division, Edwards AFB, CA 93524, USA

a r t i c l e i n f o

Article history:Received 6 April 2014Received in revised form7 August 2014Accepted 16 August 2014Available online 26 August 2014

Keywords:Renewable polymerCyanate esterHomogenous network

* Corresponding author. Tel.: þ1 760 939 0247; faxE-mail address: benjamin.g.harvey@navy.mil (B.G.

http://dx.doi.org/10.1016/j.polymer.2014.08.0340032-3861/Published by Elsevier Ltd.

a b s t r a c t

A homogenous polycarbonate/cyanate ester network has been prepared from a renewable, eugenol-derived bisphenol. The pure polycarbonate exhibited a Tg of 71 �C, Mn ¼ 8360, and polydispersity of1.88. An 80:20 blend of cyanate ester: polycarbonate was prepared and thermally cured. The presence ofthe polycarbonate had no significant effect on the cure behavior of the cyanate ester. Small Angle LaserLight Scattering (SALLS) and DSC were used to analyze the blend and no phase separation was observedeither during or after cure, suggesting that a homogenous network was generated. TMA of the resultingcomposite material revealed a single Tg of 132 �C (tan d), roughly 55 �C lower than the Tg of the purepolycyanurate and 60 �C higher than the polycarbonate. A solvent extraction study showed that thepolycarbonate could be quantitatively separated from the thermoset matrix after cure. This result provedthat no chemical grafting occured under the cure conditions employed. The excellent miscibility of thepolycarbonate and cyanate ester coupled with the efficient cure of the blend to a homogenous networksuggests that these types of blends may have applications for fabrication of toughened compositestructures.

Published by Elsevier Ltd.

1. Introduction

Sustainable thermoplastics and thermosets derived fromrenewable phenols have a number of potential advantages overthose derived from petroleum sources. These materials can beconsidered carbon sinks that are generated from CO2 by a combi-nation of photosynthesis (by plants) and chemical manipulation.Petroleum-based polymers can also be considered carbon sinks, butthere is no net benefit to the overall carbon balance as one carbonsink (crude oil) is merely replaced with another. In contrast,increased utilization of sustainable polymers has the potential toactually reduce the amount of atmospheric carbon dioxide in theshort term and to be carbon neutral in the long term. Anotherunique aspect of renewable polymers lies in their structural di-versity. For example, the presence of various functional groups inrenewable phenols may provide opportunities to reduce thenumber of synthetic steps required to generate a final polymer [1]

: þ1 760 939 1617.Harvey).

and has an impact on the properties of resins without subjectingthe monomers or polymers to additional chemical processes. Thesestructural features can affect important parameters including wateruptake, hydrophobicity, thermal stability, and glass transitiontemperature.

Over the last few years our group has studied a variety ofcyanate esters derived from sustainable phenols [2e5]. Mostrecently, we described the synthesis and characterization of acyanate ester resin derived from the renewable phenol eugenol [6].To expand the potential applications of this resin it became of in-terest to investigate blends of the cyanate ester with a thermo-plastic that was also prepared from eugenol. In this manner a 100%bio-based composite material would be realized. Although cyanateesters have a number of physical properties that make themdesirable in high performance applications, one of their drawbacksand thermoset materials in general, lies in their modest fractureresistance. Toughening approaches for thermosets date back to thelate 1960s, where rubber was shown to improve the fracturestrength of epoxy resins [7]. More recent studies have explored theuse of reactive rubbers to improve the toughness of cyanate esters,however, the use of these materials led to significant reductions in

B.G. Harvey et al. / Polymer 55 (2014) 5073e50795074

both the Tg and mechanical strength [8]. In general, elastomericmodification has been shown to decrease elastic modulus, yieldstrength, and creep resistance. Furthermore, elastomer modifica-tion of highly crosslinked thermosets, such as cyanate esters, is notan effective approach since matrix yielding is the dominanttoughening mechanism. To mitigate these issues, the effects ofblending cyanate esters with various thermoplastic resins,including poly(ethylene phthalate) [9], poly (ether imide) [10],polysulfones [8,11e13], and polycarbonates [14,15] have beenstudied. In most cases the thermoplastic phase separates duringcure of the cyanate ester with toughening of the bulk materialdependent on the resulting morphology. The morphology of thematerial is affected by the composition andmolecular weight of thethermoplastic, cure temperature, kinetics, and pressure, amongother factors.

The formation of a micro-sized particulate phase having a ma-jority composition of the modifying agent is widely considered tobe the most effective morphology for imparting toughness tothermosets. However, this approach to toughening relies onreaction-induced phase separation which is sensitive to cure tem-perature. During the production of composite parts from highlyexothermic cure, significant temperature gradients cannot beavoided and practical control of morphology is difficult to achieve.In contrast, morphologies that allow alternative tougheningmechanisms and are not dependent on cure temperature may beuseful for composite part fabrication. For example, sub-micronphase separation that results in domains smaller than opticalwavelengths of light or interpenetrating networks have also beenshown to provide augmented fracture strength. In one study, pol-ycarbonates blended with epoxy resins were shown to becompletely miscible upon cure at loadings of up to 12% [16]. Ho-mogenous formulations of this type had improved flexuralmodulus compared to related heterogeneous blends. Another studyon phenoxy/epoxy blends showed that at high epoxy cure rates,homogeneous networks were produced that demonstrated higherfracture toughness compared to heterogeneous morphologies, atsome expense to blend Tg [17]. This effect has also been observedwith polycarbonate/cyanate ester blends. Loadings of up to 50%polycarbonate with cyanate esters have been shown to form singlephase interpenetrating networks with improved toughness [15].From a design standpoint, there are two keys to achieving thismorphology. First, the modifying polymer's structure shouldclosely match that of its intended thermoset host. Second, themolecular weight of the thermoplastic should be kept relativelylow [18], both to generate a homogenous network and to limitviscosity increases, thus having minimal effects on fiber reinforcedcomposite processing. With the goal of developing a compatiblethermoplastic toughener for the eugenol derived cyanate ester, thecurrent work explores the synthesis and characterization of a low-molecular weight polycarbonate and cyanate ester/polycarbonatenetwork derived from eugenol.

2. Experimental

2.1. General

4,40-(butane-1,4-diyl)bis(2-methoxyphenol) (1) and 1,4-bis(4-cyanato-3-methoxyphenyl)butane (2) were prepared as previ-ously described [6]. Triphosgene and pyridine were obtained fromAldrich and used as received. NMR spectra were collected on aBruker Avance II 300 MHz NMR spectrometer. Samples were pre-pared in CDCl3 and spectra were referenced to the solvent peaks(d ¼ 7.26 and 77.16 ppm for 1H and 13C spectra, respectively).Fourier Transform Infrared Spectroscopy (FTIR) was carried outusing a Thermo Nicolet Nexuus 6700 FTIR equipped with the Smart

iTr attenuated total internal reflection (ATR) accessory, singlebounce diamond crystal. The detector type was a liquid nitrogencooled MCTA. FTIR spectra are an average of 32 scans, at 4 cm�1

resolution, and have been baseline and background corrected.

2.2. Polycarbonate synthesis

In a typical synthesis 1 (1.006 g, 3.3mmol) was dissolved in 8mLpyridine and the solution cooled to �20 �C with stirring. Tri-phosgene (0.384 g, 1.3 mmol) was added and the solution wasallowed to warm to room temperature. An additional 4 mL ofpyridine was then added to dissolve residual solids. Stirring over-night yielded a thick graymixturewhichwas poured into 200mL ofwater to give a white precipitate. The aqueous supernatant wasdecanted and replaced with methanol. The mixture was heated to~50 �C and the solid was dispersedwith vigorous stirring for 30minin themethanol solution. The supernatant was decanted off and thesolid was dried first under a stream of nitrogen and then in a vac-uum oven at 50 �C overnight to yield 0.622 g (57%) of an off-whitepowder. The polymer was further purified by dissolving in a min-imum amount of methylene chloride followed by reprecipitation inmethanol. 1H NMR (CDCl3) d 7.12 (d, J ¼ 8.0 Hz, 2H, Ph), 6.80e6.70(m, 4H, Ph), 3.87 (s, 6H, OMe), 2.63 (bs, 4H, CH2), 1.67 (bs, 4H, CH2).

2.3. Polycarbonate endcapping

This procedurewas conducted as described in the literature [19].100 mg of polycarbonate was dissolved in 6 mL of pyridine. t-Butyldimethylsilyl chloride (0.468 g, 3.1 mmol) was dissolved intriethylamine (1 mL) and this solution was then added to the pyr-idine solution. The reaction was stirred under nitrogen at 50 �C for24 h. The solvent was removed under reduced pressure to yield aresinous solid. The solid was then dissolved in a minimum ofdichloromethane and precipitated in methanol. After a secondprecipitation, the resulting solid was collected, washed withmethanol, and dried in a vacuum oven. The molecular weight of theendcapped polymer was determined by comparing the 1H NMRintegrals for the tert-butyl group from the endcap to the methoxygroup from the repeat unit. The Mn calculated by this method was9575 Da.

2.4. Solvent extraction of cyanate ester/polycarbonate network

An intimate mixture of 160 mg of 2 and 38 mg of the poly-carbonate was added to a circular silicone mold with a diameter of20 mm and a depth of 2 mm. The mold was heated to 150 �Cunder a slow flow of nitrogen and then held at that temperaturefor 30 min. All of the solids melted to form a clear, homogeneousmixture. The temperature was then increased 10 �C every 10 minuntil it reached 210 �C and was then held at that temperature for24 h. After cooling to room temperature, the puck was removedfrom the mold, broken into several large pieces, and then trans-ferred to a glass vial. 5 mL of dichloromethane was added to thevial and the sample was gently heated to 35 �C for 15 min. Thedichloromethane solution was then decanted and any loose piecesof material were collected on a frit. The dichloromethaneextraction was repeated 5 times and then the glass vial and fritwere placed in a vacuum oven to dry. The dichloromethane ali-quots were combined and the solvent removed under reducedpressure to leave a yellow residue. After drying, the insolublefragments of the puck weighed a total of 153 mg. The solubleresidue from the dichloromethane fractions weighed 42 mg. Atotal of 195 mg of sample (98.5%) was accounted for in theexperiment.

B.G. Harvey et al. / Polymer 55 (2014) 5073e5079 5075

2.5. Thermal analysis

DSC measurements of the polycarbonate were performed on aTA Instruments Q100 calorimeter under 50 mL/min of flowingnitrogen or air. The sample was subjected to a heat-cool-heatcycle from 50 �C to 200 �C with a ramp rate of 10 �C/min. DSCof the cyanate ester/polycarbonate blend was performed on a TAInstruments Q200 calorimeter under 50 mL/min of flowing ni-trogen. The sample was subjected to a heat-cool-heat cycle from50 �C to 350 �C at 50 �C/min. A sample for TMA analysis wasprepared by combining 26.8 mg of 2 and 7.1 mg of the poly-carbonate in a silicone mold measuring 3 � 20 mm and 3 mm indepth. The blend was melted and cured as described above.Oscillatory TMA of the resulting bar was conducted with a TAInstruments Q400 series analyzer under 50 mL/min of nitrogenflow. The bar was held in place via a 0.05 N initial compressiveforce with a macroexpansion probe while the probe force wasmodulated at 0.05 Hz over an amplitude of 0.04 N (with a meancompressive force of 0.1 N). Thermogravimetric analysis (TGA)was performed on a TA Instruments Q5000 analyzer with eithernitrogen or air flow of 50 mL/min. The samples were heated fromambient temperature to 600 �C at 5 �C/min. TGA/FTIR data wascollected with a Thermo Nicolet Nexuus 6700 FTIR interfaced via aheated gas cell and transfer line (held at 150 �C) to a TA in-struments Q50 TGA. The TGA was set to ramp from ambienttemperature to 600 �C at a rate of 10 �C/min. Gas phase FTIRspectra were taken every 30 s.

Fig. 1. Synthesis of a cyanate ester resin

2.6. Small Angle Laser Light Scattering (SALLS)

Amixture of 20% (by mass) polycarbonate and 80% cyanate esterwas prepared by melt mixing at 80 �C. Upon heating and melting ofthe cyanate ester a clear solution was readily obtained. A droplet ofthe mixture was pressed between two glass slides for Small AngleLaser Light Scattering (SALLS) measurements. The in-situ thermalcure of the cyanate ester was induced through the use of a pro-grammable Instec hot stage during the SALLS experiment. Atunable power 5 mW laser diode of wavelength 635 nm was uti-lized as the light source. During the experiment, the unpolarizedlaser light was transmitted through the film sample while thescattered radiation was detected using an Apogee Instruments, Inc.Alta U6 CCD camera. Scattered radiation patterns were detectedand analyzed at prescribed times while the temperature of thesample was increased from ambient up to 250 �C at 10 �C/min.

3. Results and discussion

The polycarbonate was prepared by reaction of the bisphenolwith triphosgene in pyridine at room temperature (Fig. 1). Theproduct was isolated as a white solid by precipitation in waterfollowed by a methanol wash and reprecipitation. Two poly-carbonate samples were initially prepared using different batchesof the bisphenol. The purified polymers were characterized by GPC,1H NMR, IR, DSC, and TGA. The first polymer sample (EPC1) hadMn ¼ 4318 and a polydispersity (Mw/Mn) of 2.25. A second sample

and polycarbonate from eugenol.

Fig. 3. FTIR spectrum of the polycarbonate.

B.G. Harvey et al. / Polymer 55 (2014) 5073e50795076

prepared from a different batch of bisphenol yielded a secondpolycarbonate (EPC2) with Mn ¼ 8360, and a polydispersity (Mw/Mn) of 1.88. No significant differences in purity between the twostarting batches of bisphenol were evident, suggesting that evenminor impurities may affect the molecular weight. The peaks in the1H NMR spectrum of the polycarbonate were relatively sharp andshowed a distinct downfield shift of the aromatic protons, while theresonances due to the methoxy groups and bridging methyleneshad nearly identical chemical shifts to the starting bisphenol(Fig. 2).

The IR spectrum of the polycarbonate (Fig. 3) exhibited adistinctive carbonyl stretch at 1774 cm�1, while the phenolicendgroups were observed as a subtle, broad peak at ~3500 cm�1.DSC analysis of EPC1 showed a Tg of 51 �C (Fig. 4), while TGA undernitrogen showed a 5% weight loss at 310 �C, a 10% weight loss at355 �C, and a char yield (600 �C) of 8%. TGA in air showed signifi-cantly lower stability with a 5% weight loss at 289 �C, a 10% weightloss at 310 �C, and a char yield of 0% (Fig. 5). DSC analysis of EPC2showed a Tg of 71 �C consistent with its higher molecular weight.Both the glass transition temperatures and thermal stabilities of thepolycarbonates are similar to other recently reported aliphatic/ar-omatic polycarbonates [20].

Although o-methoxy groups on the aromatic rings of cyanateesters have been shown to have a detrimental effect on thedegradation temperature of the resulting thermosets [4], thethermal stability of the eugenol-derived polycarbonate appears tobe dictated by the overall polymer structure. In the case of thecyanate esters, the cyanurate ring system is more thermally stablethan the subunits, while in the case of the polycarbonate, thepolymer structure has lower or comparable stability to theoxygenated phenolic subunits. To further investigate this issue anddetermine if the o-methoxy groups play a significant role in thethermal degradation of the polycarbonate, a TGA/IR study of thematerial was conducted. A gas phase spectrum collected at asample temperature of 450 �C showed evolution of phenoliccompounds with similar spectra to eugenol, along with carbondioxide, methane, and putative cyclocarbonates (Fig. 6).

Cyclic carbonyls have been identified in the thermal decompo-sition products of conventional polycarbonates [21], but for theeugenol-derived polycarbonate, a broad absorption centered at

Fig. 2. 1H NMR spectra of 1 (top) and the polycarbonate (bottom). The

1834 cm�1 suggests formation of benzodioxolones (Fig. 7). Thisunique decomposition mode is facilitated by the ortho-methoxygroups and may represent an important degradation pathway forpolymers derived from oxygenated renewable phenols.

After characterizing the polycarbonate and evaluating its ther-mal stability and decomposition pathways, the miscibility of EPC1and the cyanate ester alongwith the cure behavior andmorphologyof the blend was evaluated by SALLS. The overall scattering in-tensity as a function of film temperature was integrated and isplotted in Fig. 8. 20% polycarbonate does not inhibit cyanate estercrystallization as evidenced by the relatively high scattering in-tensity at room temperature. Cross-polarized microscopyconfirmed the crystallinity of the film. A significant decrease in theintensity of the scattered light was observed at the melting point ofthe cyanate ester resin which was depressed by nearly 20 �Ccompared to the pure resin. After this initial drop in intensity, nofurther change in scattering intensity was observed either during orafter the cure of the cyanate ester. The lack of a significant increasein the intensity of scattered light indicates that no phase separationoccurred at length scales comparable to the wavelength of theincident light [22]. To provide further evidence, a DSC experimentwas conducted with the blended material. The DSC scan (Fig. 9)clearly shows the initial glass transition temperature of EPC1 and

phenolic proton is clearly observed in the bisphenol at 5.6 ppm.

Fig. 6. FTIR data of gas phase decomposition products collected at 450 �C.

Fig. 7. Proposed decomposition pathway for the polycarbonate resulting in formationof benzodioxolones.

Fig. 4. DSC data for EPC2 and EPC1. The heat flow has been normalized for clarity.

B.G. Harvey et al. / Polymer 55 (2014) 5073e5079 5077

the melting transition of the cyanate ester, as well as complete cureon heating to 350 �C. On completion of cure and a second scan toestablish a baseline, the DSC sample was removed from itscontainer, examined, and found to be optically clear. Although adistinct glass transition temperature was not observed after com-plete cure, as is sometimes the case for thermosetting networkpolymers, the combination of DSC and SALLS data clearly indicatethat no phase separation took place during cure.

To further characterize the cure behavior of the blend, an 80:20(cyanate ester:EPC2) blend was cured and then evaluated by TMA(Fig. 10). As expected based on the SALLS and DSC data, only one Tgwas observed at 132 �C (tan d) compared to the Tg of the purecyanate ester (186 �Cdtan d) and that of EPC2 (71 �CdDSC). Thisvalue is similar to that predicted from the Fox relation (140.5 �C) inwhich the K- parameter is estimated as Tg1/Tg2 in the Gordon-Taylorequation [23,25]. The presence of only one distinct Tg providesadditional support for the existence of a homogenous network.

To investigate the lack of phase separation during cure, an end-group analysis of the polycarbonate was conducted. The synthesisof the polycarbonate should yield phenolic end-groups which mayreact with cyanate esters to give imidocarbonate linkages [24]. Toquantify the amount of residual phenolic end-groups, the poly-carbonate was allowed to react with TBDMSCl in pyridine/trie-thylamine and was then worked up by precipitation in methanol.1H NMR spectroscopy of the product revealed incorporation ofTBDMS in the product (See Supporting Information). By comparingthe integrals from the methoxy groups to those of the tert-butyl

Fig. 5. TGA data for the polycarbonate in both a nitrogen and air environment.

group on the endcaps, a molecular weight of 9575 Da was calcu-lated. This agrees reasonably well with the GPC results (Mn ¼ 8360)and may be higher due to the reprecipitation steps which would beexpected to reduce the amount of lighter oligomers. To determine ifany chemical grafting was taking place during the cure reaction, amixture of EPC2 and the cyanate ester were cured to form a ho-mogenous puck. The puck was then repeatedly extracted withwarm CH2Cl2 and after evaporation, the extract was weighed. In theextraction experiment, 160 mg of the cyanate ester and 38 mg ofthe polycarbonate were used. After extraction, the total massaccounted for in both fractions was 195 mg or 98.5% recovery. Themass of the polycarbonate extract was 42 mg, while the residualcured cyanate ester weighed 153 mg. The ease of extraction andessentially quantitative recovery of the polycarbonate proves that

Fig. 8. Average scattering intensity as a function of temperature measured from SmallAngle Laser Light Scattering (SALLS). Snapshots of scattered intensity accompanied bycorresponding optical micrographs are also shown.

Fig. 9. DSC data for the cyanate ester polycarbonate blend (prepared with EPC1).

B.G. Harvey et al. / Polymer 55 (2014) 5073e50795078

no significant reaction between polycarbonate end-groups and thecyanate ester resin occur. The relatively low molecular weight ofthe polymer and the flexible 4-carbon chain between aromaticrings are likely important structural characteristics for the forma-tion of the homogenous network in the absence of chemicalgrafting.

Although this work represents an important start, the tough-ening effect resulting from this type of homogenous cyanate ester/polycarbonate network still needs to be quantified. DMA studies, aswell as impact and flexural testing of composite panels preparedfrom the blended/cured networks will need to be conducted toevaluate this toughening approach.

4. Conclusions

A new polycarbonate with both rigid (aromatic) and flexible(aliphatic) components was synthesized from the renewablephenol eugenol. In contrast to polycarbonates derived from con-ventional phenols, two key structural differences; the presence ofan aliphatic bridging group between the aromatic rings, andmethoxy-groups ortho to the carbonates, resulted in some uniqueproperties. The aliphatic bridging group imparted a modest Tg tothe polymer, while the presence of the methoxy-groups appearedto promote a unique thermal decomposition pathway that may be

Fig. 10. TMA data for an 80:20 (cyanate ester: polycarbonate) blend. E0 is the apparentstorage modulus and E00 is the apparent loss modulus.

important for other renewable polycarbonates. To explore theutility of the new polycarbonate, a 100% bio-based thermoset/thermoplastic blend was prepared from a eugenol-derived cyanateester resin and the polycarbonate. SALLS, DSC, and TMA data allshowed formation of a homogenous network upon cure, whileNMR spectroscopy and an extraction experiment confirmed that nochemical grafting was taking place. The excellent miscibility of thecomponents upon cure suggests that formation of a homogenousnetwork from a polycarbonate and cyanate ester derived from thesame parent phenol may be a viable method for generating atoughened thermoset composite. This general method is likely notdependent on cure temperature and may therefore be morecompatible with fiber reinforced composite processing than con-ventional toughening approaches.

Acknowledgments

The authors would like to thank the Strategic EnvironmentalResearch and Development Program (SERDP–ProjectWP-2214), theOffice of Naval Research (ONR), and the Air Force Office of ScientificResearch (AFOSR) for funding this work.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.polymer.2014.08.034.

References

[1] Meylemans H, Groshens T, Harvey B. Synthesis of renewable bisphenols fromcreosol. ChemSusChem 2012;5:206e10.

[2] Davis M, Guenthner J, Sahagun C, Lamison K, Reams J, Mabry J. Polycyanuratenetworks from dehydroanethole cyclotrimers: synthesis and characterization.Polymer 2013;54:6902e9.

[3] Cash J, Davis M, Ford M, Groshens J, Guenthner A, Harvey B, et al. High Tgthermosetting resins from resveratrol. Polym Chem 2013;4:3859e65.

[4] Meylemans H, Harvey B, Reams J, Guenthner A, Cambrea L, Groshens T, et al.Synthesis, characterization, and cure chemistry of renewable bis(cyanate)esters derived from 2-methoxy-4-methylphenol. Biomacromolecules2013;14:771e80.

[5] Davis M, Guenthner A, Groshens T, Reams J, Mabry J. Polycyanurate networksfrom anethole dimers: synthesis and characterization. J Polym Sci Part A:Polym Chem 2012;50:4127e36.

[6] Harvey B, Sahagun C, Guenthner A, Groshens T, Cambrea L, Reams J, et al.A high-performance renewable thermosetting resin derived from eugenol.ChemSusChem 2014;7:1964e9.

[7] McGarry F, Proe R. Building design with fiber reinforced materials. Proc R SocLond Ser A 1970;319:59e68.

[8] Hwang J, Cho K, Yoon T, Park C. Effects of molecular weight of polysulfone onphase separation behavior for cyanate ester/polysulfone blends. J Appl PolymSci 2000;77:921e7.

[9] Iijima T, Katsurayama S, Fukuda W, Tomoi M. Modification of cyanate esterresin by poly(ethylene phthalate) and related copolyesters. J Appl Polym Sci2000;76:208e19.

[10] Harismendy I, Del Rio M, Marieta C, Gavalda J, Mondragon I. Dicyanate ester-polyetherimide semi-interpenetrating polymer networks II. Effects ofmorphology on the fracture toughness and mechanical properties. J ApplPolym Sci 2001;80:2759e67.

[11] Suman J, Kathi J, Tammishetti S. Thermoplastic modification of monomericand partially polymerized bisphenol a dicyanate ester. Eur Polym J 2005;41:2963e72.

[12] Hwang J, Cho K, Park W, Huh W. Phase separation behavior of cyanate esterresin/polysulfone blends. J Appl Polym Sci 1999;74:33e45.

[13] Bucknall C, Gomez C, Quintard I. Phase separation from solutions of poly(ethersulfone) in epoxy resins. Polymer 1994;35:353e9.

[14] Wu T, Prevorsek D, Wertz D. Low color compositions comprising a cross-linked polycyanurate polymer and a thermoplastic polymer U S Patent4334045 1982.

[15] Hay N. In: Hamerton I, editor. Chemistry and technology of cyanate esterresins. Dordrecht: Springer Science þBusiness Media; 1994. p. 176e7. andreferences contained therein.

[16] Lu M-S, Chen J-L, Li Y-S, Chang F-C, Li M-S, Ma C-C. The kinetics, thermal andmechanical properties of epoxy-polycarbonate blends cured with aromaticamine. J Polym Res 1998;5:115e24.

B.G. Harvey et al. / Polymer 55 (2014) 5073e5079 5079

[17] Teng K-C, Chang F-C. Single-phase and multiple-phase thermoplastic/ther-moset polyblends: 2. Morphologies and mechanical properties of phenoxy/epoxy blends. Polymer 1996;37:2385e94.

[18] Iijima T, Tochimoto T, Tomoi M. Modification of epoxy resins with poly(arylether ketone)s. Appl Polym Sci 1991;43:1685e92.

[19] Bolton D, Wooley K. Synthesis and characterization of hyperbranched poly-carbonates. Macromolecules 1997;30:1890e6.

[20] Swileh B, Al-Hiari Y, Kailani M, Mohammad H. Synthesis and characterizationof polycarbonates by melt phase interchange reactions of alkylene and ary-lene diacetates with alkylene and arylene diphenyl dicarbonates. Molecules2010;15:3661e82.

[21] Puglisi C, Sturiale L, Montaudo G. Thermal decomposition processes in aro-matic polycarbonates investigated by mass spectrometry. Macromolecules1999;32:2194e203.

[22] Jyotishkumar P, Moldenaers P, George S, Thomas S. Viscoelastic effects inthermoplastic poly(styrene-acrylonitrile)-modified epoxy-DDM system dur-ing reaction induced phase separation. Soft Matter 2012;8:7452e62.

[23] Schneider H. In: Kulshreshtha A, Vasile C, editors. Handbook of polymerblends and composites volume 3B. Shropshire UK: Rapra Technology Limited;2003. p. 390e3.

[24] Shieh J-Y, Yang S-P, Wu M-F, Wang C-S. Synthesis and characterization ofnovel low-dielectric cyanate esters. Polym Sci Part A: Polym Chem 2004;42:2589e600.

[25] The simplified Gordon-Taylor equation is expressed as 1/Tg(blend) ¼w1/Tg1 þw2/Tg2. In this equation w1 and w2 are the weight fractions of the respectiveblend components. See reference 23 for more information.