Reactive processing of textile fiber-reinforced thermoplastic composites – An overview

Composites: Part A 38 (2007) 666–681 www.elsevier.com/locate/compositesa

Review

Reactive processing of textile fiber-reinforced thermoplastic composites – An overview K. van Rijswijk *, H.E.N. Bersee Delft University of Technology, Faculty of Aerospace Engineering, Design and Production of Composite Structures, Kluyverweg 3, 2629 HS, Delft, The Netherlands Received 20 December 2005; received in revised form 14 May 2006; accepted 14 May 2006

Abstract Thermoplastic composites offer some interesting advantages over their thermoset counterparts like a higher toughness, faster manufacturing and their recyclable nature. Traditional melt processing, however, limits thermoplastic composite parts in size and thickness. As an alternative, reactive processing of textile fiber-reinforced thermoplastics is discussed in this paper: a low viscosity mono- or oligomeric precursor is used to impregnate the fibers, followed by in situ polymerization. Processes that are currently associated to manufacturing of thermoset composites like resin transfer molding, vacuum infusion and resin film infusion, might be used for manufacturing of thermoplastic composite parts in near future. This paper gives an overview of engineering and high-performance plastic materials that are suitable for reactive processing and discusses fundamental differences between reactive processing of thermoplastic and thermoset resins. Ó 2006 Elsevier Ltd. All rights reserved. Keywords: A. Thermoplastic resin; A. Fabrics/textiles; E. Resin transfer moulding (RTM); Vacuum infusion

Contents 1. 2.

*

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reactive thermoplastic material systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Vinyl polymerization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Ring-opening polymerization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Reactive processing of engineering plastics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.1. Thermoplastic polyurethanes (TPU) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.2. Polymethylmethacrylate (PMMA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.3. Polyamides. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.4. Polyesters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.5. Polycarbonate (PC). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4. Reactive processing of high-performance plastics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.1. Polyetheretherketone (PEEK) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.2. Polyetherketone (PEK) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.3. Polyethersulfone (PES) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.4. Polyphenylenesulfide (PPS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Corresponding author. Tel.: +31 15 2788981; fax: +31 15 2781151. E-mail address: [email protected] (K. van Rijswijk).

1359-835X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.compositesa.2006.05.007

667 668 668 668 669 669 669 669 671 671 672 673 673 674 674

K. van Rijswijk, H.E.N. Bersee / Composites: Part A 38 (2007) 666–681

3.

4.

2.4.5. Polyethylenenaphthalate (PEN) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.6. Polybutylenenaphthalate (PBN) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reactive processing of textile reinforced thermoplastic composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Reactive processes for thermoplastic composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.1. Structural reaction injection molding (SRIM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.2. Vacuum infusion (VI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.3. Resin film infusion (RFI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Differences between reactive processing of thermoplastic and thermoset composites . . . . . . . . . . . . . . . . . . . . . . . Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction Thermoplastic composites (TPCs) offer some substantial advantages over thermoset ones [1]. Due to the higher toughness of the matrix, they offer a higher impact resistance. Manufacturing cycle times consisting of melting the matrix, shaping and consolidation by cooling are significantly shorter than for their thermoset counterparts, which require a time consuming curing step. In addition, TPCs can be welded [2] and recycled [3]. In structural composite applications, textiles are commonly applied as reinforcement because of the high fiber volume fractions that can be obtained and because of the possibility to tailor the load bearing capacity through the fiber lay-up. Traditionally, textile fiber-reinforced TPCs are melt processed by stacking alternating layers of fiber textiles and polymer sheets in a hot-press [4]. After heating the package above the polymer melting point, the press is closed to obtain the required product shape. In a subsequent cooling step the product solidification takes place followed by demolding. The main disadvantage of TPCs is the need for high processing temperatures and pressures, caused by the high melt viscosity of the matrix. In addition, proper impregnation of the fiber at a microlevel might prove difficult and often results in products with a locally high void content.

667

674 675 676 676 676 676 677 677 678 678

A solution to improve the fiber impregnation is to bring the matrix and the fibers in more intimate contact before the final molding step, or in other words, to reduce the required flow length of the polymer matrix. Various concepts of these intermediates have been developed [1] such as co-mingled textiles that consist of both reinforcing and polymer fibers, textiles made of powder coated fibers and partially or fully consolidated panels (semi-pregs and pre-pregs [4]), see Fig. 1. Besides the additional costs of these intermediate products, additional disadvantages are encountered such as the de-bulking of commingled textiles and the occurring fiber waviness, the adherence of the powder coatings and the occurrence of electrostatic discharges during processing of powder coated textiles. Semi- and pre-pregs have poor drapability and often contain some solvent residue when they are made by solution impregnation: when PEI (polyetherimide) pre-pregs are heated for instance evaporation of the NMP (N-methylpyrrolidinone) solvent can be easily seen. The additional remark has to be made that although most plastics can be remolded numerous times without losing properties, some high-performance plastics slightly degrade during processing due to their high melting points: cross-linking occurs for instance in molten PPS (polyphenylenesulfide) [5]. Any intermediate processing step that requires melting therefore increases the number of crosslinks, hence resulting in a more brittle matrix.

Fig. 1. Processing steps for manufacturing thermoplastic composite parts through melt- and reactive processing.

668

K. van Rijswijk, H.E.N. Bersee / Composites: Part A 38 (2007) 666–681

An alternative solution to melt processing presented in this article is reactive processing of TPCs: after impregnating the fibers with a low viscosity mono- or oligomeric precursor, polymerization of the thermoplastic matrix is conducted in situ. Polymerization can be initiated by heat or UV radiation and might require the addition of a catalyst system, which can be added to the precursor prior to impregnation. Due to their low molecular weight, precursors have extremely low melt viscosities and proper fiber impregnation is therefore achieved without the need for high processing pressures. Moreover, through reactive processing, textile fiberreinforced TPCs can be even manufactured through lowpressure infusion processes such as resin transfer molding (RTM) [6], structural reaction injection molding (SRIM) [6] and vacuum infusion [7]. These are all classic manufacturing processes for manufacturing of thermoset composites. In these processes, a dry fiber pre-form is placed into a closed mold after which the precursor is infused by applying vacuum or pressure, typically less than 10 bar. Following polymerization, the composite product is demolded. Additional advantages of this type of processing are:

2. Reactive thermoplastic material systems For reactive processing of thermoplastic composites, the in situ polymerization of the matrix basically has to meet the following requirements: a high molecular weight linear polymer has to be formed at sufficiently high conversions without the generation of unwanted by-products. Suitable polymerization types are consequently narrowed down to addition polymerizations of mono- and difunctional species, of which vinyl polymerization and ring-opening polymerization are most common. 2.1. Vinyl polymerization Vinyl polymers are polymers made from vinyl monomers; small molecules containing carbon–carbon double bonds. During polymerization the double bonds are broken into single bonds, resulting in two free electrons. The free electrons are used to join monomer units to form a long chain of many thousands of carbon atoms containing only single bonds between atoms, see Fig. 2. 2.2. Ring-opening polymerization

 Larger, thicker and more integrated products can be obtained than what is currently achievable with melt processing.  A thermoplastic composite with a chemical fiber-tomatrix bond can be obtained, due to the fact that polymerization takes place around the fibers. This can for instance be accomplished by using silane coupling agents [8].  In addition to the textile reinforcement, nano-particles can be added to the unreacted monomer in order to obtain a fiber-reinforced polymer nano-composite through for instance exfoliation [9–13].

As the name already suggests, ring-opening polymerization (ROP) is based on a polymerization mechanism in which ring-shaped molecules (cyclics) are opened into linear monomers or oligomers and subsequently connected into high molecular weight polymers without generating by-products, see Fig. 4. ROP initially received attention as clean alternative for polymerization routes that result in the generation of nasty by-products or require the use of large amounts of hazardous solvents. Production of PEEK and PPS for instance makes use of high-boiling solvents such as diphenyl sulfone and dichlorobenzene [14,15], whereas toxic phosgene gas is used for interfacial phosgenation polymerization of polycarbonates [16]. In addition, molecular weights obtained through ROP are generally higher than for the same polymer obtained through polycondensation reactions. The first reason for this is that polymer chains are broken in a subsequent purification step, during which unreacted monomer and catalyst residues are removed by applying heat, solvents and high shear forces. A second reason is that molecular weights are kept

In this article, an overview is given on thermoplastic material systems that can be reactively processed; a distinction is made between engineering and high-performance plastics. Various reactive processes are discussed for the manufacturing of fiber-reinforced thermoplastic products and the differences between reactive processing of thermoplastic and thermoset resins are highlighted.

O O

C

N

R1

N

C

O

+

HO

R2

OH

C

O N H

R1

N H

C

O

R2

O

N H

R2

N H

n

a) Polyurethane O O

C

N

R1

N

C

O

+

H2N

R2

NH2

C

O N H

R1

N H

b) Polyurea Fig. 2. Free radical vinyl polymerization of polyurethanes and polyureas.

C

n

K. van Rijswijk, H.E.N. Bersee / Composites: Part A 38 (2007) 666–681

669

low deliberately in order to keep the melt viscosity reasonable low for further processing through the melt. The fact that a so-called ring-chain equilibrium exists, which can be shifted by temperature or the addition of cleverly selected catalysts not only offers the possibility for in situ polymerization of polymer stock shapes and composites but also for recycling through cyclo-depolymerization (CDP) [17]. In the following sections, suitable thermoplastic material systems for reactive processing are introduced. A distinction is made in reactive processing of engineering and high-performance plastics.

ization–repolymerization (DPRP) mechanism. Upon heating, depolymerization into lower weight fractions takes place, which results in a significant viscosity reduction. Subsequent cooling induces repolymerization and molecular weight rises again up to its starting value. A minimum viscosity of a few Pa s is obtained at a processing temperature of 270 °C. Although further heating reduces the viscosity into the mPa s-range, the material loses its ability to repolymerize [23]. Recyclability of the Fulcrum resin through re-grinding and injection molding without the loss of mechanical properties has been demonstrated [24].

2.3. Reactive processing of engineering plastics

2.3.2. Polymethylmethacrylate (PMMA) Free radical vinyl polymerization of the methylmethacrylate monomer (MMA) into its polymer PMMA is usually conducted at temperatures in excess of 40 °C using peroxide initiators [25], see Fig. 3. At lower temperatures the reaction time will soon exceed 16 h, whereas at higher temperatures the danger exists that due to the exotherm of 462.2 J/g the monomer will soon start to boil, resulting in voids in the final product. During polymerization the density increases from 0.9 g/cm3 (monomer) to 1.2 g/cm3 (polymer), and in order to reduce shrinkage, usually a pre-polymer (solution of PMMA in its monomer) is used. The melt viscosity of the pre-polymer is higher than that of the monomer (0.10 Pa s at 50 °C), but is still low enough to cast (PlexiglasÒ or LuciteÒ) windows, which is the main application of PMMA due to its transparent nature. As far as composite processing is concerned, a monomer impregnation method of natural fibers [26] and a reactive injection pultrusion process are mentioned in literature [27,28]. In order to increase the reaction rate, which is necessary in a continuous process, the reaction temperature of the pre-polymer had to be increased to 160 °C, which is well above the glass transition temperature of the amorphous polymer.

2.3.1. Thermoplastic polyurethanes (TPU) Polyurethanes are among the most widely applied resin materials specifically developed for reactive processing [18]. Di-isocyanates react with di-alcohols through free radical vinyl polymerization in a matter of seconds when processed at around 60–80 °C, as is shown in Fig. 2a. Sometimes, diamines are used instead of di-alcohols to polymerize socalled polyureas, see Fig. 2b. Most polyurethanes have a thermoset nature, although thermoplastic polyurethanes exist. These are however commonly sold as fully reacted granules or powder for melt processing. Dube´ et al. [19], however, demonstrated the feasibility of reactive injection pultrusion (RIP) of TPU composites based on the abovementioned chemistry. They pointed out that the high reactivity of the resin requires a fast responding process control system, since any deviation from ideal conditions can easily lead to a significant reduction in the final polymer properties. Reactive processing of TPU based on a different type of chemistry was recently developed by Dow Chemicals (USA) and is currently applied by the Fulcrum Composites Company for the manufacturing of continuous fiberreinforced TPU pultrusion profiles [20–22]. Whereas most polyurethanes are reactively processed from their monomers, Fulcrum TPU uses high molecular weight linear polymer as starting point and makes use of a depolymer-

O R

R

2 R

O

O R

CH3

O

+

O

CH3

CH3

O

H2C R

2.3.3. Polyamides Anionic polymerization of lactams is the oldest and up to now the most developed way for reactive processing of

CH2

H2C

C

n O

O

O CH3

O

O CH3

Fig. 3. Free radical vinyl polymerization of PMMA.

CH3

670

K. van Rijswijk, H.E.N. Bersee / Composites: Part A 38 (2007) 666–681

thermoplastics through ROP. In the early 1940s, ROP of caprolactam into PA-6 was discovered and has been commercially exploited by for instance BASF, Bayern, DSM and Monsanto. Anionic polymerization of laurolactam into PA-12 has recently gained interest due to the work conducted at EMS Chemie A.G. and the Ecole Polytechnique Fe´de´rale de Lausanne, both in Switzerland.

polymerization [40]. The complex kinetics of the anionic polymerization of caprolactam have been studied extensively, mainly focusing on the autocatalytic effect of the exothermic temperature rise [40–48]. Numerous reactive processes have already been developed for unreinforced and reinforced PA-6 such as (rotational) casting [49–53], reaction injection molding (RIM) [54–56], reinforced reaction injection molding (RRIM) [57,58], structural reaction injection molding (SRIM) [59–61] and reactive injection pultrusion (RIP) [62,63]. Technology developed by DSM allows PA-6 to be fully depolymerized in a cost-effective way into caprolactam of virgin quality [64]. In order to reduce mold shrinkage and to increase the toughness, a rubber-modified block-copolymer called NyrimÒ was developed by DSM and is currently traded under the name AP NylonÒ by Bru¨ggemann Chemical, Germany [65–69]. Due to the pre-polymer activator of AP NylonÒ, the viscosity is slightly higher (g = 60–90 mPa s [66]) than of the unmodified resin (g = 10 mPa s [70]). In addition to these rubber block copolymers also varieties with branches and cross-links have been investigated [71,72]. A vacuum infusion process for manufacturing of polyamide-6 composite wind turbine blades is currently being developed at The Delft University of Technology [70,73–77].

2.3.3.1. Polyamide-6 (PA-6). The anionic ROP of e-caprolactam (Tm = 69 °C) into high molecular weight polyamide-6 (PA-6), see Fig. 4, is a catalyzed reaction performed at 130–170 °C [29]. Final conversions of up to 99.3% can be obtained in 3–60 min, depending on the type and amount of activator and catalyst added. Typical activators used are N-acyllactams, whereas metal caprolactamates are commonly used as initiator [30–36]. Commonly two material batches are prepared, which after mixing start to polymerize: a monomer–activator batch and a monomer– initiator batch. Due to the anionic nature, the reaction is easily terminated by proton donating species, such as for instance moisture. Therefore, storage and processing have to be conducted in an absolutely moisture free environment. Since processing takes place below the polymer melting and crystallization point, polymerization and crystallization take place simultaneously, resulting in solid highly crystalline PA-6 [37–39]. The reaction is exothermic (DHpolymerization = 166 J/g, DHcrystallization = 144 J/g), which leads to a significant temperature increase during

2.3.3.2. Polyamide-12 (PA-12). Polyamide-12 is anionically polymerized from x-laurolactam (Tm = 154 °C) using similar activators and initiators as discussed in the previous

R O

O

M

H N

m.

+

O

-

N

N

+

+

(CH2)

(CH2)

n

(CH2)

n

n

Lactam monomer:

Initiator:

Activator:

Caprolactam (n=5)

Metal caprolactamate

N-Acyllactam

Laurolactam (n=11)

Anionic ring opening polymerization

O (CH2) O

n

NH

R

-

N

(CH2)

Polyamide:

n

M

+

O

m

PA-6 (n=5) PA-12 (n=11)

Fig. 4. Anionic ring-opening polymerization of polyamides.

O

K. van Rijswijk, H.E.N. Bersee / Composites: Part A 38 (2007) 666–681

paragraph on PA-6, see Fig. 4 [78]. In contrast to anionic PA-6, polymerization has to be conducted above the final polymer melting point (175 °C) to increase the polymerization rate and to avoid premature crystallization [79]. As a consequence, when processing at 180–240 °C, an additional cooling step is required prior to demolding [80]. The molten monomer has an initial viscosity of 23 mPa s [81] and has to be kept in nitrogen protective environment to prevent initiator deactivation. The reaction exotherm is about 53 J/g [79] and total mold shrinkage is 8.3–9.6% [82]. Reactive PA-12 is currently marketed by EMS Chemie A.G., Switzerland, who also developed a one part activator–initiator solution called GrilonitÒ that can be stored indefinitely in inert atmosphere. This solution, which can also be used for anionic polymerization of PA-6, no longer requires pre-mixing of two separate material batches that slowly polymerize over time. Pultrusion [83] and SRIMlike processes [84–87] for PA-12 composites are currently being developed at the Ecole Polytechnique Fe´de´rale de Lausanne (EPFL), Switzerland, the National University of Ireland, Ireland and the Institut fu¨r Verbundwerkstoffe, Germany. Caprolactam and laurolactam can be anionically copolymerized to tailor polymer properties. Material properties such as the strength and stiffness, glass transition and the melting point are in direct relation with the monomer ratios [88–90]. 2.3.4. Polyesters Synthesis of macrocyclic polyesters and the ring-opening metathesis polymerization (ROMP) thereof was initially developed by D.J. Brunelle and his research group at the General Electrics Corporation, USA, in the late 1980s and early 1990s. Initially aiming at polycarbonates, see next paragraph, reactive processing of both polyetherteraphthalate (PET) and polybutyleneteraphthalate (PBT) was developed. The latter is currently being marketed specifically for the production of composites under the name CyclicsÒ by the Cyclics Corporation, USA [91–93]. 2.3.4.1. Polyethyleneteraphthalate (PET). Macrocyclic oligomers can be obtained through cyclodepolymerization (CDP) of linear PET and subsequently repolymerized through ROMP into high MW PET [94,95]. Fig. 5 shows H

polymerization of PET using a cyclic dimer (Tm = 225 °C). Up to 100% conversion at 250–325 °C are obtained in several hours without a catalyst, whereas addition of a catalyst reduces the reaction time to 3–15 min at 225 °C. The initial melt viscosity of the cyclic precursors is 30 mPa s [96,97]. 2.3.4.2. Polybutyleneteraphthalate (PBT). Depolymerization of linear PBT yields a macrocyclic oligomer mixture, which can be repolymerized directly into solid high molecular weight (Mw = 445,000) semi-crystalline PBT at 180– 200 °C by addition of a titanium initiator [98,99], see Fig. 6. The oligomer mixture has a melt viscosity of 150 mPa s at 150 °C, which drops to 30 mPa s at 190 °C. When processed at 190 °C in protective atmosphere, the viscosity reaches 1 Pa s after approximately 5 min and final conversions of 95–99% are obtained within 30 min [100]. Although polymerization itself is not exothermic, approximately 67 J/g of heat is generated in the subsequent crystallization phase. Final polymer properties strongly depend on the polymerization temperature. When isothermally polymerized below its melting point (Tm = 220–267 °C), the PBT obtained is highly crystalline and tends to become brittle (i.e. elongation at break = 1.8%), due to a phenomena called cold-crystallization [101] A subsequent melting and cooling cycle brings back the more ductile behavior. Properties of initially reactively processed PBT are largely unaffected after mechanical–thermal recycling (re-grinding followed by injection molding) [3]. In addition, PBT can be recycled chemically by depolymerization into the cyclic oligomers or all the way into its monomers dimethylterephthalate and butanediol [17]. RTM-like processes for manufacturing fiber reinforced PBT are developed at Delaware University (USA) and KU Leuven (Belgium) [102–105], whereas the National University of Ireland is developing Resin Film Infusion technology, making use of a one-component monomer–catalyst system [106,107]. Through ring-opening polymerization, PBT copolymers with PET can be easily obtained [91]. 2.3.5. Polycarbonate (PC) Macrocyclic Bisphenol-A (Tm = 200–210 °C) can be polymerized into polycarbonate through ROMP conducted

H O

O O

O

H

H

H

H

H H

H

O H

H

Ring opening

H H

H

H O

671

H

O

O (C2H4)

O

O

metathesis polymerization

O

O

Fig. 5. Ring-opening metathesis polymerization of PET.

n

672

K. van Rijswijk, H.E.N. Bersee / Composites: Part A 38 (2007) 666–681 O

Cl

O

O

O

O

Cyclic oligimer

+

HO(CH2)4OH

formation

(CH2)4

m

Cl

m = 2-7

O

Ring opening metathesis polymerization

Initiation by Ti(OR)4

O

O

O

O

Propagation

O

O

O

O

Ti(OR) 3 R

(CH2)4

O

Ti(OR) 3 R

(CH2)4

O

m

n

Fig. 6. Ring-opening metathesis polymerization of PBT.

H3C

H3C

CH3

CH3

Cyclic oligomer formation from Bisphenol-A OH

O

O

O

O

O

Cl

Cl

H3C

O

m

CH3

O

Ring opening metathesis polymerization O

using anionic initiators

O

n Fig. 7. Ring-opening metathesis polymerization of polycarbonates.

at 240–280 °C using anionic activators [108], see Fig. 7. When polymerized at 250 °C in protective atmosphere, the initial viscosity of the macrocyclic melt is 1 Pa s and in 2–5 min high molecular weight PC is obtained (Mw = 300,000) with conversions of over 99% [109]. The reaction is entropy driven, which means that no exothermic heat is generated during polymerization. ROMP of macrocyclics consisting of Bisphenol-A and hydroquinone at 300 °C results in a solvent resistant PC, which can be obtained in either amorphous or semi-crystalline form [16,110]. Other versions reported in literature are crosslinked [111] and copolymerized PCs [112]. Salem et al. successfully produced glass fiber reinforced PC composites through reactive processing, although an additional consol-

idation step in a hot-press was necessary to reduce the void content [113]. 2.4. Reactive processing of high-performance plastics After the successes of reactive processing of engineering plastics such as PU, PMMA and PA-6, several attempts were made to develop similar technology for processing of high-performance plastics, mainly focusing on ROP of polyarylethers. Initial results, however, brought complications to light, which were directly related to the inherent properties of high-performance plastics, which make them so interesting in the first place: an extreme stiff polymer backbone combined with outstanding chemical resistance

K. van Rijswijk, H.E.N. Bersee / Composites: Part A 38 (2007) 666–681

and thermal properties. Whereas the relatively flexible engineering plastics are easily converted into cyclic precursors of only a single or a few monomer units, a much larger number of monomer units is required to form so-called macrocyclic precursors of the more rigid high-performance plastics. The higher molecular weight of the cyclic precursors brings up the following problems:  Synthesis of the cyclic precursors has to be conducted in high-dilution or pseudo-high-dilution conditions, which is explained by the fact that just before ring closure the ends of the relatively long polymer chains are rather far apart from each other. In case the solid concentration during synthesis is too high, it is more likely that a chain will react with a neighboring chain rather than having its two ends react together to form a cycle. Synthesis therefore requires a lot of solvent and leads to relatively low yields [17].  The macrocyclic precursors obtained are often an oligomer mixture, rather than a substance consisting of single sized rings, which is related to matters discussed in the previous point. Various oligomers might differ in properties such as melting points, solubility in the other oligomers within the same mixture or even in being amorphous of crystalline [114,115]. These differences complicate synthesis and further processing.  In order to obtain high conversions, polymerization has to be conducted at high temperatures for two reasons: (i) the processing temperature has to exceed the melting point of all oligomers, and (ii) the viscosity of the growing polymer chains has to be kept low enough to induce sufficient chain mobility. Unfortunately, at temperatures ranging from 300 to 400 °C side reactions like cross-linking are unavoidable, which strongly reduce the polymer performance [115]. Recent advances in cyclics technology are (i) the use of monomer units containing meta and ortho rather than para

HO DMAc

673

linkages in order to produce macrocyclics with a lower molecular weight (at a slight reduction of thermal stability of their equivalent polymer) and (ii) isolation of specific oligomers with a low melting point or an amorphous character in order to reduce the required polymerization temperature [114]. In the next paragraphs an overview of reactive processing equivalents of common high-performance plastics is given. 2.4.1. Polyetheretherketone (PEEK) Literature mentions synthesis of 45–60–90 membered macrocyclics from 4,4-difluorobenzophenone and hydroquinone in pseudo-high-dilution conditions at a yield of 60%, which polymerized at 350 °C in 5 min using caesium fluoride as initiator [116], see Fig. 8. Another source reports synthesis of cyclic 2-mers, 3-mers and 4-mers, but fails to discuss ring-opening polymerization [15]. 2.4.2. Polyetherketone (PEK) According to Jiang et al. [117–119], macrocyclic oligomers for polyetherketones (PEK) were produced at a yield of 54%. These cyclics (Tg of 127 °C) formed a clear melt at 280 °C and could subsequently be polymerized for 1 h at the same temperature up till a conversion of 93.5%. The resulting linear polymer is slightly branched and has a Tg of 216 °C, which is slightly lower than commercially produced PEK (Tg = 228 °C) due to the presence of oligomers that failed to polymerize. Ring-opening polymerization at 390 °C for 30 min of a cyclic PEK di-mer into an amorphous polymer (Tg = 162 °C) was reported by the same authors [120]. The potential of reactive processing of high performance thermoplastics is clearly demonstrated by the work conducted at McGill University, Canada [121,122]. It was shown that macrocyclic PEK containing a 1,2-dibenzoylbenzene moiety, see Fig. 9, has a stable melt viscosity (80 mPa s at 330 °C) and could be polymerized at 340 °C in 30 min after addition of a nucleophilic initiator.

OH

Toluene

+

K 2CO3

O

O

Macrocyclic oligomer O

formation

O

C

m

F

m = 45, 60 or 90

F

O

Ring opening polymerization O

O

C

Initiation by CsF, 350 oC n

Fig. 8. Reactive processing route of PEEK.

674

K. van Rijswijk, H.E.N. Bersee / Composites: Part A 38 (2007) 666–681

generated during production of commercial PEK-390 and polymerize these in 25 min at 385 °C [123].

CH3 O CH3

2.4.3. Polyethersulfone (PES) ROP of PES cyclic precursors, conducted at 300 °C for 2 h, is shown in Fig. 10 [124]. Different types of poly arylene sulfone cyclics are discussed in [125,126].

O

O

n Fig. 9. Macrocyclic precursor for PEK.

Baxter et al. managed to polymerize cyclic PEK in 30 min at 300 °C using a Caesium fluoride initiator into a semi-crystalline polymer (Tg = 168 °C, Tm = 302 °C) [114]. In addition, they showed that instead of using macrocyclics specifically synthesized for ring-opening polymerization, one could also use the cyclic oligomer byproduct that is

2.4.4. Polyphenylenesulfide (PPS) Melt polymerization of cyclic PPS pentamer and hexamer at 300 °C under Nitrogen environment is discussed in [127], see Fig. 11. The resulting high molecular weight polymer is highly crystalline and has a melting point of 277 °C, which is comparable to that of commercial grades PPS (Tm = 285 °C). An alternative method for preparing the same cyclic PPS precursor is discussed in [128]. 2.4.5. Polyethylenenaphthalate (PEN) According to [129], a PEN (polyethylenenaphthalate) macrocyclic oligomer mixture with a melting point of 250–285 °C was prepared in a 57% yield. In 25 min ROP

O F

DMSO

O

F

S O

Macrocyclic oligomer

+

formation O

O HO

O

S

m

OH

S O

O

Ring opening polymerization

O

S

O

Initiation by CsF, 300 oC n

Fig. 10. Reactive processing route of PES.

S

Macrocyclic oligomer formation S

S

from diphenyl disulfide m

Free-radical ring opening polymerization

S

Initiation by DTB, 300 oC n

Fig. 11. Reactive processing route of PPS.

K. van Rijswijk, H.E.N. Bersee / Composites: Part A 38 (2007) 666–681

675

O Cl

O

Macrocyclic oligomer Cl

formation

O

O

(CH2)

x

O DABCO

O

+

CH2Cl2

m

x = 2: PEN HO

(CH 2)

x

OH

x = 4: PBN

O

O

O

Ring opening polymerization

(CH2)x

Initiation by Bu2 SnO, 275 oC

O

n

Fig. 12. Reactive processing route for PEN and PBN.

weight impurities in the cyclic oligomer mixture strongly affected the final polymer properties [129]. To summarize this chapter, Fig. 13 shows the melt viscosities and processing temperatures of various monoand oligomeric thermoplastic precursors. For comparison, the same data is also given for common thermoset resins and thermoplastic polymers suitable for melt processing. Table 1 compares the processing temperatures for some common thermoplastic matrices for both melt- and reactive processing. It can be seen that whereas the reduction in processing temperatures for engineering plastics is

was conducted in the presence of a peroxide initiator at 295–300 °C, see Fig. 12. The final polymer (Tm = 261 °C) suffered from impurities present in the cyclic oligomer mixture and initiator residue. 2.4.6. Polybutylenenaphthalate (PBN) PBN (polybutylenenaphthalete) macrcocylic precursors containing various size oligomers (Tm = 150–220 °C) were prepared at a 75% yield. ROP was subsequently conducted in the presence of a tinoxide catalyst at 275 °C and was completed in 15 min, see Fig. 12. The linear low molecular

100000

10000

PA-12

1000

Melt viscosity [Pa·s]

PEEK

PES

PMMA Melt processing of engineering plastics

PEKK

PA-6

PEI

PBT

100

PPS Melt processing of highperformance plastics

10

Reactive processing of thermoset resins

PC

1

vinylester

Reactive processing of thermoplastic oligomers

epoxy

ETPU

polyester

0.1

PMMA PA-6

0.01

PBT

PEK

PA-12 Reactive processing of thermoplastic monomers

0.001 0

50

100

150

200

250

300

350

400

450

Processing temperature [°C] Fig. 13. Melt viscosities and processing temperatures of various matrix materials for both reactive and melt processing.

676

K. van Rijswijk, H.E.N. Bersee / Composites: Part A 38 (2007) 666–681

Table 1 Comparison of processing temperatures for melt- and reactive processing for some common thermoplastic matrix materials Thermoplastic matrix

PA-6 PA-12 PBT PMMA PC PET PES PPS PEEK

Processing temperature (°C) Melt processing

Reactive processing

Reduction

230–290 230–270 250–270 220–260 265–360 265–325 330–390 330 380–390

140–160 180–245 180–200 120–160 250 250–325 300 300 350

70–150 0–90 50–90 60–140 15–110 0–15 30–90 30 30–40

significant, the difference for high-performance plastics is relatively small. The reason for this is the high melting point of the oligomeric precursors for high-performance plastics, which are almost as high as their high molecular weight linear equivalents. The next section discusses three potential manufacturing processes to produce textile fiber-reinforced thermoplastic composites from the resin materials introduced previously. 3. Reactive processing of textile reinforced thermoplastic composites In this section, three reactive processes for manufacturing of textile fiber-reinforced TPCs are introduced: structural reaction injection molding, vacuum infusion and resin film infusion, see Fig. 14. Reactive processes for neat polymer parts like casting [49,51,53,130,131], reaction injection molding (RIM) [54–56,66,132,133] and rotational molding [50,67–69,134], and for manufacturing of short fiber reinforced composite parts like reinforced reaction injection molding (RRIM) [57,58,135,136] are discussed in literature, as well as reactive injection pultrusion (RIP) of continuous fiber-reinforced profiles [19,28,63, 80,83,137].

3.1. Reactive processes for thermoplastic composites 3.1.1. Structural reaction injection molding (SRIM) SRIM uses high pressure (8–10 bar) to impregnate a dry fiber pre-form, which is placed between two solid mold halves, see Fig. 14a [80–82,84–87,101–104,138–140]. The process is closely related to Resin Transfer Molding (RTM) of thermoset composites [6]. In order to obtain a proper fiber impregnation, 1 Pa s is generally regarded as the maximum viscosity limit of the reactive mixture [141]. Upon completion of polymerization, the composite part can be demolded. As the size of the product increases so does the clamping force required to keep the mold closed during injection of the resin. As a consequence, the size of parts manufactured through SRIM is limited. The reactive material systems are separated into two material feeds, in order to prevent premature polymerization. Just before entering the mold, the material feeds from two tanks are mixed and polymerization commences. The big advantage of SRIM is the fast cycle time, whereas the high tooling costs are the main disadvantage. 3.1.2. Vacuum infusion (VI) As the name suggests, vacuum is used to compact and impregnate a dry fiber pre-form, which is placed between a solid and a flexible mold half, see Fig. 14b [60,61,70,73–77]. Although the impregnation speed is lower compared to SRIM, the fact that atmospheric pressure is sufficient for mold clamping causes that the maximum achievable part size is only limited by the pot-life of the reactive system. After mixing the contents of the two material tanks, the reactive mixture (maximum viscosity: 1 Pa s) is dispensed into a buffer vessel, which is required to separate the pressure required for dispensing and the vacuum necessary to promote infusion. The big advantages of VI are the virtually unlimited size of the parts that can be produced and the low cost tooling due to the low pressures involved. The disadvantages are related to the flexible mold half, which often can be used only once and leads to a poor surface quality on one side of the product.

Fig. 14. Schematic representation of three reactive processes for manufacturing of thermoplastic composites: (a) structural reaction injection moulding (SRIM), (b) vacuum infusion (VI) and (c) resin film infusion (RFI).

K. van Rijswijk, H.E.N. Bersee / Composites: Part A 38 (2007) 666–681

677

Reactive processing of thermoplastic composites seems to have a lot in common with manufacturing of thermoset ones. There are however a few significant differences, which are briefly summarized below:

Heated equipment is therefore necessary to keep the unreacted mixture in the melt state. Demolding techniques that make use of the brittle nature of thermoset resins, such as peel plies in vacuum infusion and break lines through resin rich areas, are not applicable due to the tough nature of the thermoplastic material. In contrast to thermoset resin waste during processing, reacted and unreacted resin in tubes (vacuum infusion) can be fully recycled, as well as waste from mechanical trimming such as flashes. The performance of composites in not solely determined by the fibers and the matrix, but also by the fiber-tomatrix interphase. In order to improve this bond, glass fibers for instance are usually coated with silane coupling agents: di-functional compounds with the ability to bond with both the fibers and a polymer matrix of choice. A non-compatible coupling agent results in a weak interphase or even prevent the polymerization of reactive resins. Coupling agents have been developed for numerous thermoset composites resins and for thermoplastic composites manufactured through melt processing. Coupling agents specifically for reactive processing of thermoplastic composites have not been developed yet, but have recently become a topic of interest [142,143]. In addition to fibers with suitable coating, processing aids for, for instance, vacuum infusion of thermoplastic resins such as transport tubes, vacuum bagging films and sealant tape are not commercially available as such. Suitable high-temperature consumables, which are chemically resistant against the unreacted thermoplastic materials, therefore still need to be selected.

 The polymerization rate increases with increasing temperature, which is true for both thermoplastic and thermoset resins. When processing reactive thermoplastic materials with a semi-crystalline nature, however, one has to keep in mind that crystallization is adversely affected by temperature. The processing temperature has to be chosen such that polymerization and crystallization are well balanced. When the temperature is too low, crystallization will be too fast and reactive chainends and monomer can get trapped inside crystals before they can polymerize. On the other hand, when the temperature is too high, the final degree of crystallinity is reduced, which reduces the strength, stiffness and chemical resistance of the polymer [73].  Some reactive thermoplastic materials, like PA-6, PA-12 and PBT, have a melt viscosity, which is an order of magnitude lower than of common thermoset resins. As a consequence, the occurring capillary forces during impregnation of the fiber pre-form are significant and form a potential source for voids and runner formation [81,138].  Whereas most thermoset resins are liquid at room temperature, most thermoplastic precursors are still solid.

Perhaps the most striking difference is the fact that whereas reactive processing of thermoset composites forms the mainstay of the composite industry worldwide, the potentially great thermoplastic resins presented in this study have not found any significant composite application so far. This cannot be related to the expected properties of reactively processed thermoplastic composites, given the fact that their melt-processed equivalents have already found extensive application. Careful examination of the references of this study demonstrates the lack of cooperation between polymer chemists and composite processing engineers. The presented material systems are mainly developed for processing of neat polymer products and are yet to be adapted to meet the specific requirements for composite processing. Being unable to address critical processing issues with the currently available resin systems and not being persistent enough in having the resin materials adjusted to its specific needs, the composites industry has not been able yet to bring reactive processing of thermoplastic composites to the next level. More recent publications give evidence of cooperation initiatives between chemists and engineers, which is paramount for putting an end to the current standstill and for bringing reactive

3.1.3. Resin film infusion (RFI) RFI is similar to pre-pregging of thermoset composites: hand lay-up is used to stack alternating layers of fiber textiles and sheets of unreacted reactive mixture in a mold, see Fig. 14c [106,107]. Upon applying heat and pressure, the sheets melt and impregnate fibers, after which polymerization initiates. In case the sheets do not have sufficient mechanical properties for handling, they can be directly cast on top of a supporting fiber textile. Also powderimpregnated textiles can be used. To maintain reactivity, no premature solid-state polymerization or deactivation during storage or handling of the unreacted sheets or pre-pregs should take place. During melting and impregnation a significant volume reduction of the lay-up takes place and often multiple de-bulking steps are necessary to compress the lay-up. The atmospheric pressure is usually sufficient for de-bulking of thin-walled parts and allows the use of a flexible mold half. In this case, the viscosity of the reactive mixture is again limited to 1 Pa s. De-bulking of thicker parts is more difficult and often required a second solid mold half and significant pressure, which simultaneously allows the use of oligomeric precursors with a higher melt viscosity. The ease of preparing the lay-up and the fast impregnation are regarded as the main advantages of RFI, whereas de-bulking is the main disadvantage. 3.2. Differences between reactive processing of thermoplastic and thermoset composites









678

K. van Rijswijk, H.E.N. Bersee / Composites: Part A 38 (2007) 666–681

thermoplastic processing into commercial practice. Only then a comparison between the presented material systems can be made to determine which is best suitable for manufacturing of composites, in terms of storage, handling, processing and of course composite properties. It will be interesting to know which of these materials will ultimately become the ‘epoxies and polyesters of reactive thermoplastic composite processing’. 4. Conclusions Reactive processing of textile fiber-reinforced thermoplastic composites has been discussed as an alternative to melt processing for manufacturing of larger, thicker and more integrated parts. An overview of suitable thermoplastic material systems for reactive processing was given ranging from commercially available engineering plastics to high-performance plastics that are still in the pioneering stage. Three manufacturing processes, commonly used for reactive processing of thermosets, were introduced and differences between reactive processing of thermoset and thermoplastic composites were pointed out. The recent focus of both polymer chemists and composite engineers on reactive processing of thermoplastic composites not only demonstrates the interest in expanding the application field of thermoplastic composites but also provides the necessary means to accomplish this. References [1] Talreja R, Ma˚nson JA. Polymer matrix composites. Oxford: Elsevier Science Ltd.; 2001. [2] Stavrov D, Bersee HEN. Resistance welding of thermoplastic composites – an overview. Compos Part A: Appl Sci Manufact 2005;36:39–54. [3] Steenkamer DA, Sullivan JL. On the recyclability of a cyclic thermoplastic composite material. Compos Part B: Eng 1998;29B:745–52. [4] Weteringe BJ, van Dongen M, Bersee HEN. Rubber forming of thermoplastic composite parts for the eaglet rudder. In: Proceedings of 15th international conference on composite materials (ICCM-15). Durban, South Africa, 27 June–1 July, 2005. [5] Bulakh N, Jog JP, Nadkarni VM. Structure development in curing of poly(phenylene sulfide). J Macromol Sci—Phys B 1993;32:275–93. [6] Rudd D. Resin transfer molding and structural reaction injection molding. In: ASM Handbook, 2001. p. 492–500. [7] Hoebergen A, Holmberg JA. Vacuum infusion. In: ASM Handbook, 2001. p. 501–15. [8] Plueddemann EP. Silane coupling agents. New York: Plenum Press; 1982. [9] Chen G, Weng W, Wu D, Wu C. Nonlinear conduction in nylon-6/ foliated graphite nanocomposites above the percolation threshold. J Polym Sci Part B: Polym Phys 2004;42:155–67. [10] Weng W, Chen G, Wu D. Crystallization kinetics and melting behaviors of nylon6/foliated graphite nanocomposites. Polymer 2003;44:8119–32. [11] Weng W, Chen G, Wu D, Chen X, Lu J, Wang P. Fabrication and characterization of nylon 6/foliated graphite electrically conducting nanocomposites. J Polym Sci Part B: Polym Phys 2004;42:2844–56. [12] Du X, Xiao M, Meng Y, Hay AS. Synthesis of poly(arylene disulfide) – vermiculite nanocomposites via in situ ring-opening polymerization of macrocyclic oligomers. Polym Int 2004;53:789–93.

[13] Vlasveld DPN. Fibre reinforced polymer nanocomposites. Applied Sciences, Aerospace Engineering, Delft University of Technology, Delft, 2005. [14] Wang Y-F, Hay AS. Macrocyclic arylene ether ether sulfide oligomers: new intermediates for the synthesis of high-performance poly(arylene ether ether sulfide)s. Macromolecules 1997;30(2): 182–93. [15] Qi Y, Chen T, Xu J. Synthesis of cyclic precursor of poly(ether ether ketone). Polym Bull 1999;42:245–9. [16] Brunelle DJ. Solvent-resistant polycarbonates. Trends Polym Sci 1995;3(5):154–8. [17] Hodge P, Colquhoun HM, Williams DJ. From macrocycles to macromolecules – and back. Chem Ind 1998:162–7. [18] Randall D, Lee S. The polyurethanes book. John Wiley & Sons Ltd.; 2002. [19] Dube´ MG, Batch GL, Vogel JH, Macosko CW. Reaction injection pultrusion of thermoplastic and thermoset composites. Polym Compos 1995;16(5):378–85. [20] d’Hooghe EL, Edwards CM. Thermoplastic composite technology; tougher than you think. Adv Mater 2000;12(23):1865–8. [21] DowFulcrum. Thermoplastic pultrusion promises new synergies. In: Reinforced plastics, 2001. [22] Edwards CM. Thermoplastic pultrusion promises new synergies. In: Reinforced plastics, 2001. p. 34–8. [23] Edwards CM, personal communications, 25 August 2004. [24] d’Hooghe EL, Hoek B, Edwards CM. Applications for fulcrum thermoplastic composite technology. Available from: www. DowFulcrum.com. [25] FME N. Materiaalomschrijving van Thermoplastische Kunststoffen: PVC, PMMA, PE en PP. Culemborg: Stam Technische Boeken B.V.; 1976. [26] Brouwer WD. Composite applications using coir fibres in Sri Lanka. Delft: Delft University of Technology; 2003, p. 19. [27] Ma C-CM, Chen C-H. Pultruded fiber reinforced thermoplastic poly(methyl methacrylate) composites. Part I. Correlation of processing parameters for optimizing the process. Polym Eng Sci 1991;31(15):1086–93. [28] Ma C-CM, Chen C-H. Pultruded fiber reinforced thermoplastic poly(methyl methacrylate) composites. Part II: Mechanical and thermal properties. Polym Eng Sci 1991;31(15):1094–100. [29] Hedrick RM, Richard Jr WR. Reinforced polyamides and process of preparation thereof. USA: Monsanto Company; 1968. [30] Bolgov SA, Begishev VP, Malkin AY, Frolov VG. Role of the functionality of activators during isothermal crystallization accompanying the activated anionic polymerization of e-caprolactam. Polym Sci USSR 1981;23(6):1485–92. [31] Udipi K, Dave´ RS, Kruse RL, Stebbins LR. Polyamides from lactams via anionic ring-opening polymerization: 1. Chemistry and some recent findings. Polymer 1997;38(4):927–38. [32] Ueda K, Yamada K, Nakai M, Matsuda T, Hosoda M, Tai K. Synthesis of high molecular weight nylon 6 by anionic polymerization of e-caprolactam. Polym J 1996;28(5):446–51. [33] Russo S, Imperato A, Mariani A, Parodi F. The fast activation of ecaprolactam polymerization in quasi-adiabatic conditions. Macromol Chem Phys 1995;196:3297–303. [34] Ricco L, Russo S, Orefice G, Riva F. Anionic poly(e-caprolactam): relationships among condition of synthesis, chain regularity, reticular order, and polymorphism. Macromolecules 1999;32: 7726–31. [35] Mateva R, Petrov P, Rousseva S, Dimitrov R, Zolova G. On the structure of poly-e-caprolactams, obtained with bifunctional N-carbamyl derivatives of lactams. Eur Polym J 2000;36:813–21. [36] Scelia RP, Schonfeld SE, Donaruma LG. Some effects of cocatalyst structure on the anionic polymerization of e-caprolactam. II. J Appl Polym Sci 1967;11:1299–313. [37] Hornsby PR, Tung JF. Characterization of polyamide 6 made by reactive extrusion. II. Analysis of microstructure. J Appl Polym Sci 1994;54:899–907.

K. van Rijswijk, H.E.N. Bersee / Composites: Part A 38 (2007) 666–681 [38] Mateva R, Delev O, Kaschcieva E. Structure of poly(e-caprolactam) obtained in anionic bulk polymerization. J Appl Polym Sci 1995;58:2333–43. [39] Frunze TM, Shleifman RB, Zaitsev VI, Kurashev VV, Babchinitser TM, Korshak VV. Effect of the cooling conditions upon the structure and mechanical properties of poly-e-caproamide prepared by anionic polymerization of e-caprolactam. Polym Sci USSR 1972;14(4):1071–9. [40] Rigo A, Fabbri G, Talamini G. Kinetic study of anionic polymerization of 6-caprolactam by differential calorimetry. Polym Lett Ed 1975;13:469–77. [41] Malkin AY, Frolov VG, Ivanova AN, Andrianova ZS. The nonisothermal anionic polymerization of caprolactam. Polym Sci USSR 1979;21:691–700. [42] Malkin AY, Ivanova SL, Frolov VG, Ivanova AN, Andrianova ZS. Kinetics of anionic polymerization of lactams. (Solution of nonisothermal kinetic problems by the inverse method). Polymer 1982;23:1791–800. [43] Greenley RZ, Stauffer JC, Kurz JE. The kinetic equation for the initiated, anionic polymerization of e-caprolactam. Macromolecules 1969;2(6):561–7. [44] Stratula-Vahnoveanu B, Vasiliu-Oprea C. Anionic polymerization of lactams in microdispersion. I. Kinetic aspects of the process of caprolactam polymerization. Polym-Plast Technol Eng 2002;41(5): 981–95. [45] Dave´ RS, Kruse RL, Stebbins LR, Udipi K. Polyamides from lactams via anionic ring-opening polymerization: 2. Kinetics. Polymer 1997;38(4):939–47. [46] Dave´ RS, Kruse RL, Udipi K, Williams DE. Polyamides from lactams via anionic ring-opening polymerization: 3. rheology. Polymer 1997;38(4):949–54. [47] Ueda K, Nakai M, Hosoda M, Tai K. Synthesis of high molecular weight nylon 6 by anionic polymerization of e-caprolactam. Mechanism and kinetics. Polym J 1997;29(7):568–73. [48] Qu X, Ding H, Lu J, Wang Y, Zhang L. Isothermal and nonisothermal crystallization kinetics of MC nylon and polyazomethine/MC nylon composites. J Appl Polym Sci 2004;93: 2844–55. [49] Miscevic M, Catic I, Sercer M. Production of nylon-6 castings by anionic polymerisation of e-caprolactam. Int Polym Sci Technol 1993;20(11):90–8. [50] Rusu M, Ibanescu C, Murariu M, Bordeianu A, Balint S, Andrei E. Centrifugal casting of polyamide 6. I. Influence of thermal treatment. Polym Polym Compos 1998;6(3):143–6. [51] Schaaf S. Monomergussverfahren und nylon-RIM-verfahren. Kunststoffe-Plast 1986(3):24–6. [52] Vass JA. Taking advantage of cast nylon. Mach Des 1994:70–4. [53] Worthington M. The many wonders of cast nylon. IAPD Mag 1996:32–3. [54] Hedrick RM, Gabbert JD, Wohl MH. Nylon 6 RIM. American Chemical Society; 1985, p. 135–162. [55] Long SD, Dawood I, Coates PD, Johnson AF. Influence of reaction injection moulding (RIM) processing variables on the mechanical properties of a commercial linear segmented polyamide. Plast, Rubber Compos Process Appl 1995;23(3):161–74. [56] Lee KH, Kim SC. Engineering analysis of reaction injection moulding process of nylon 6. Polym Eng Sci 1988;28(7):477–84. [57] Mateva R, Ishtinakova O, Nikolov RN, Djambova C. Kinetics of polymerization of e-caprolactam in the presence of inorganic dispersed additives. Eur Polym J 1998;34(8):1061–7. [58] Te Nijenhuis K, Addink R, Van der Vegt AK. A study on composites of nylon-6 with hollow glass microspheres. Polym Bull 1989;21:467–74. [59] Karger-Kocsis J. Fracture and failure behavior of glass fiber mat-reinforced structural nylon rim composites. J Polym Eng 1992;11(1–2):153–73. [60] Pillay S, Ning H, Vaidya UK, Janowski GM. Liquid molding of carbon fabric reinforced nylon matrix composite laminates. In:

[61]

[62]

[63] [64]

[65]

[66]

[67]

[68]

[69] [70]

[71]

[72]

[73]

[74]

[75]

[76]

[77]

[78]

[79]

[80]

679

Proceedings of 7th international conference on flow processes in composite materials, Newark (USA), July 7–9, 2004. Pillay S, Vaidya UK, Janowski GM. Liquid molding of carbon fabric-reinforced nylon matrix composite laminates. J Thermoplast Compos Mater 2005;18:509–27. Cho BG, McCarthy SP, Fanucci JP, Nolet SC. Fiber reinforced nylon-6 composites produced by the reaction injection pultrusion process. Polym Compos 1996;17(5):673–81. Ning X, Ishida H. RIM – pultrusion of nylon-6 and rubbertoughened nylon-6 composites. Polym Eng Sci 1991;31(9):632–7. Booij M. Polyamide-6 recycling to caprolactam – DSM closes the loop. In: Proceedings of polyamide 2000 world congress, Zurich, Switzerland, March 14–16, 2000. Gardlund ZG, Bator MA. In situ polymerization of nylon–polyurea block copolymers. I. Synthesis and characterization. J Appl Polym Sci 1990;40:2027–35. Gabbert JD, Hedrick RM. Advances in systems utilizing nyrim nylon block copolymer for reaction injection molding. Polym Process Eng 1986;4(2-4):359–73. Harkin-Jones E, Crawford RJ. Rotational moulding of liquid nyrim in biaxially rotating heated moulds. Plast, Rubber Compos Process Appl 1995;24(1):1–6. Harkin-Jones E, Crawford RJ. An investigation of the flow behaviour of rotationally moulded NYRIM. Plast, Rubber Compos Process Appl 1995;23(4):211–20. Harkin-Jones E, Crawford RJ. Mechanical properties of rotationally molded Nyrim. Polym Eng Sci 1996;36(5):615–25. Van Rijswijk K, Bersee HEN, Jager WF, Picken SJ. Optimisation of anionic polyamide-6 for vacuum infusion of thermoplastic composites: choice of activator and initiator. Compos Part A: Appl Sci Manufact 2005;37:949–56. Kotel’nikov VA, Surin NN, Cherepanov SG, Persits IE, Danilevskaya LB, Konova IO, et al. Synthesis of cross-linked nylon RIM block copolyamides. Polym Sci, Ser A 1995;37(12):1198–205. Mathias LJ, Sikes AM. Star-branched nylon 6: the effects of branching and crosslinking on polymer properties. In: Proceedings of cross-linked polymers: chemistry, properties, and applications, Denver, CO (USA), April 5–10, 1988. Van Rijswijk K, Bersee HEN, Beukers A, Picken SJ, Van Geenen AA. Optimisation of anionic polyamide-6 for vacuum infusion of thermoplastic composites: influence of polymerisation temperature on matrix properties. Polym Test 2005;25:392–404. Van Rijswijk K, Joncas S, Bersee HEN, Bergsma OK, Beukers A. Sustainable vacuum-infused thermoplastic composites for mw-size wind turbine blades-preliminary design and manufacturing issues. J Solar Energy Eng 2005;127:570–80. Van Rijswijk K, Koppes K, Bersee HEN, Beukers A. Processing window for vacuum infusion of fiber-reinforced anionic polyamide6. In: Proceedings of 7th international conference on flow processes in composite materials (FPCM-7), Newark (USA), July 7–9, 2004. Van Rijswijk K, Vlasveld DPN, Bersee HEN, Picken SJ. Vacuum injection of anionic polyamide 6. In: Proceedings of 4th international conference on composite structures and technology (ICCST4), Durban, South Africa, January 21–23, 2003. Van Rijswijk K, Vlasveld DPN, Van Rhijn P, Bersee HEN, Beukers A, Picken SJ. Process considerations for liquid molding of composites based on anionic polyamide 6. In: Proceedings of 14th International Conference on Composite Materials (ICCM-14). San Diego (USA), July 14–8, 2003. Luisier A, Bourban P-E, Ma˚nson J-AE. Initiation mechanisms of an anionic ring-opening polymerization of lactam-12. J Polym Sci: Part A: Polym Chem 2002;40:3406–15. Luisier A, Bourban P-E, Ma˚nson J-AE. In-situ polymerisation of polyamide 12 for thermoplastic composites. In: Proceedings of 12th international conference on composite materials (ICCM-12), Paris, France, 1999. Zingraff L, Bourban P-E, Wakeman MD, Kohler M, Ma˚nson J-AE. Reactive processing and forming of polyamide 12 thermoplastic

680

[81]

[82]

[83]

[84]

[85]

[86]

[87]

[88]

[89] [90]

[91]

[92]

[93] [94]

[95]

[96]

[97]

[98]

[99]

K. van Rijswijk, H.E.N. Bersee / Composites: Part A 38 (2007) 666–681 composites. In: Proceedings of 23rd SAMPE Europe international conference, Paris, France, April 9–11, 2002. p. 237–48. Michaud V, Zingraff L, Verrey J, Bourban P-E, Ma˚nson J-AE. Resin transfer molding of anionically polymerized polyamide 12. In: Proceedings of 7th international conference on flow processes in composite materials (FPCM-7), San Diego (USA), July 7–9, 2003. Zingraff L, Bourban P-E, Michaud V, Ma˚nson J-AE. Liquid composite moulding of anionically polymerised polyamide 12. In: Proceedings of 14th international conference on composite materials (ICCM-14), San Diego (USA), July 14–18, 2003. Luisier A, Bourban P-E, Ma˚nson J-AE. Reaction injection pultrusion of PA12 composites: process and modelling. Compos Part A: Appl Sci Manufact 2003;34A(7):583–95. Verrey J, Michaud V, Ma˚nson J-AE. Processing of complex parts with thermoplastic RTM techniques. In: Proceedings of 24th international SAMPE Europe conference, Paris, France, April 1– 3, 2003. p. 553–61. ´ , McDonnel P, Connor MT, Eder R, O ´ Bra´daigh CM. Ma´irtı´n PO Process investigation of a liquid PA-12/carbon fibre moulding system. Compos Part A: Appl Sci Manufact 2001;32(7):915–23. Zingraff L, Michaud V, Bourban P-E, Manson JA. Resin transfer moulding of anionically polymerised polyamide 12. Compos Part A: Appl Sci Manufact 2005;36:1675–86. Rosso P, Friedrich K, Wollny A, Mulhaupt R. A novel polyamide 12 polymerization system and its use for a LCM-process to produce CFRP. Journal of Thermoplastic Composite Materials 2005;18:77–90. ´ Bra´daigh CM. Development of a polyamide Greaney M, O copolymer resin transfer molding system for thermoplastic composites. In: Proceedings of 7th international conference on flow processes in composite materials (FPCM-7), Newark (DE, USA), July 7–9, 2004. Kubota H, Nowell JB. Changes in morphology of cast nylon 6 through copolymerization. J Appl Polym Sci 1975;19:1521–38. Ricco L, Russo S, Orefice G, Riva F. Caprolactam–laurolactam copolymers: fast activated anionic synthesis, thermal properties and structural investigations. Macromol Chem Phys 2001;202: 2114–21. Brunelle DJ, Bradt JE, Serth-Guzzo J, Takekoshi T, Evans TL, Pearce EJ, et al. Semicrystalline polymers via ring-opening polymerization: preparation and polymerization of alkylene phthalate cyclic oligomers. Macromolecules 1998;31:4782–90. Ciovacco J, Winckler SJ. Cyclic thermoplastic properties and processing. In: Proceedings of 45th international SAMPE symposium, Long Beach (CA, USA), 2000. Winckler SJ. CBT resin wordt thermoharder met thermoplastische eigenschappen. Kunststof Rubber 2002(10):22–3. Youk JH, Kambour RP, MacKnight WJ. Polymerization of ethylene terephthalate cyclic oligomers with antimony trioxide. Macromolecules 2000;33(10):35–94-3599. Youk JH, Boulares A, Kambour RP, MacKnight WJ. Polymerization of ethylene terephthalate cyclic oligomers with a cyclic dibutyltin initiator. Macromolecules 2000;33(10):3600–5. Nagahata R, Sugiyama J-I, Goyal M, Goto M, Honda K, Asai M, et al. Thermal polymerization of uniform macrocyclic ethylene terephthalate dimer. Polymer 2001;42:1275–9. Nagahata R, Sugiyama J-I, Goyal M, Asai M, Ueda M, Takeuchi K. Solid-phase thermal polymerization of macrocyclic ethylene terephthalate dimer using various transesterification catalysts. J Polym Sci: Part A: Polym Chem 2000;38:3360–8. Brunelle DJ, Serth-Guzzo J. Titanate-catalyzed ring-opening polymerization of cyclic phthalate ester oligomers. Polym Preprints, Am Chem Soc, Div Polym Chem 1999;40(1):566–7. Brunelle DJ, Bradt JE, Serth-Guzzo J, Takekoshi T, Evans RA, Pearce EJ, et al. Semicrystalline polymers via ring-opening polymerization: preparation and polymerization of alkylene phthalate cyclic oligomers. Macromolecules 1998;31(15):4782–90.

[100] Eder RHJ, Winckler SJ. Processing of advanced thermoplastic composites using cyclic thermoplastic polyesters. In: Proceedings of 22nd SAMPE Europe conference, Paris, France, 2001. p. 661–72. [101] Parton H, Baets J, Lipnik P, Devaux J, Verpoest I. Liquid moulding of textile reinforced thermoplastics. In: Proceedings of 11th European conference on composite materials (ECCM-11), Rhodos, Greece, 2004. [102] Parton H, Verpoest I. Reactive processing of textile reinforced thermoplastics. In: Proceedings of 14th international conference on composite materials (ICCM-14), San Diego (USA), July 14–18, 2003. [103] Parton H, Verpoest I. Thermoplastic liquid composite molding: production and characterization of composites based on cyclic oligomers. In: Proceedings of 7th international conference on flow processes in composite materials, Newark (DE, USA), July 7–9, 2004. [104] Parton H, Verpoest I. In situ polymerization of thermoplastic composites based on cyclic oligomers. Polym Compos 2005:60–5. [105] Chatterjee A, Mallon PJ, Dweib MA, Saied S, Gillespie Jr JW, Heider D. Development of an elevated temperature vacuum assisted resin transfer molding system for thermoplastic composites. In: Proceedings of 34th International SAMPE Technical Conference, Baltimore (MD, USA) 2002. ´ Bra´daigh CM. Resin film infusion of [106] Coll SM, Murtagh AM, O cyclic PBT composites: a fundamental study. In: Proceedings of 25th SAMPE Europe international jubilee conference, Paris, France, 2004. ´ Bra´daigh CM. Resin film infusion of [107] Coll SM, Murtagh AM, O cyclic PBT composites: consolidation analysis. In: Proceedings of 7th international conference on flow processes in composite materials (FPCM-7), Newark (DE, USA), February 7–9, 2004. [108] Stewart KR. Melt polymerization of BPA cyclic polycarbonate oligomers: rheokinetics of polymerization relevant to reactive processing. Polym Preprints, Am Chem Soc, Div Polym Chem 1989;30(2):575–6. [109] Brunelle DJ, Evans TL, Shannon TG, Boden EP, Stewart KR, Fontana LP, et al. Preparation and polymerization of cyclic oligomeric carbonates: new route to super-high molecular weight polycarbonate: an overview. Polym Preprints, Am Chem Soc, Div Polym Chem 1989;30(2):569–70. [110] Brunelle DJ, Krabbenhoft HO, Bonauto DK. Preparation of crystalline and solvent resistant polycarbonates via ring-opening polymerization of cyclic oligimers. Polymer Preprints USA 1993;34:73–4. [111] Rosenquist NR, Fontana LP. Formation of crosslinked networks via polymerization of oligomeric cyclic carbonates in the presence of polyfunctional co-reactants. Polym Preprints, Am Chem Soc, Div Polym Chem 1989;30(2):577–8. [112] Evans TL, Carpenter JC. Use of cyclic carbonate oligomers in the synthesis of bisphenol-A polycarbonate/polydimethylsiloxane copolymers. Polym Preprints, Am Chem Soc, Div Polym Chem 1990;31(1):18–9. [113] Salem AJ, Stewart KR, Gifford SK, Berenbaum AM. Fabrication of thermoplastic matrix structural composites by resin transfer molding of cyclic bisphenol-A polycarbonate oligomers. SAMPE J 1991;27(1):17–22. [114] Baxter I, Colquhoun HM, Hodge P, Kohnke FH, Lewis DF, Williams DJ. Macrocyclic oligomers of the aromatic polyetherketone ‘‘PK99’’: synthesis, fractionation, structural characterisation and ring-opening polymerisation. J Mater Chem 2000;10:309–14. [115] Hall AJ, Hodge P. Recent research on the synthesis and application of cyclic oligomers. React Funct Polym 1999;41(1-3):133–9. [116] Chen M, Gibson HW. Large-sized macrocyclic monomeric precursors of poly(ether ether ketone): synthesis and polymerization. Macromolecules 1996;29(16):5502–4. [117] Jiang H, Chen T, Bo S, Xu J. Novel macrocyclic precursors of phenolphthalein poly(arylene ether ketone) and poly(arylene ether sulfone): synthesis and polymerization. Macromolecules 1997;30(23):7345–7.

K. van Rijswijk, H.E.N. Bersee / Composites: Part A 38 (2007) 666–681 [118] Chen C, Zhou H, Wu Z, Zhang WJ. Macrocyclic oligomers of the phenolphthalein poly aryl ether ketones: synthesis and characterization. Polymer Preprints, Am Chem Soc, Div Polym Chem 1998;39(2):531. [119] Jiang H, Chen T, Qi Y, Xu J. Ring-opening polymerization of polyarylates. Indian J Chem Technol 1998;5:146–50. [120] Jiang H-Y, Qi Y-H, Chen T-L, Xing Y, Lin Y-H, Xu J-P. Isolation, spectra, structure, and ring-opening polymerization of strained macrocyclic aryl(ether ketone) dimer. J Polym Sci: Part A: Polym Chem 1997;35(9):1753–61. [121] Wang Y-F, Chan KP, Hay AS. Rheological and chemorheological studies of cyclic aryl ether ketone and aryl ether thioether ketone oligomers containing the 1,2-dibenzoylbenzene moiety. J Appl Polym Sci 1996;59:831–43. [122] Chan KP, Wang Y-F, Hay AS, Hronowski XL, Cotter RJ. Synthesis and characterization of novel cyclic (aryl ether ketone)s, cyclic (aryl ether phthalazine)s, and cyclic (aryl ether isoquinoline)s. Macromolecules 1995;28(20):6705–17. [123] Ben-Haida A, Colquhoun HM, Hodge P, Williams DJ. Cyclic oligomers of poly(ether ketone) (PEK): synthesis, extraction from polymer, fractionation, and characterisation of the cyclic trimer, tetramer and pentamer. J Mater Chem 2000;10:2011–6. [124] Mullins MJ, Woo EP, Chen CC, Murray DJ, Bishop MT, Balon KE. Synthesis and polymerization of aryl ether cyclooligomers. Polym Preprints, Am Chem Soc, Div Polym Chem 1991;32(2):174–5. [125] Ganguly S, Gibson HW. Synthesis of a novel macrocyclicarylene ether sulfone. Macromolecules 1993;26(10):2408–12. [126] Gibson HW, Ganguly S, Yamaguchi N, Xie D, Chen M, Bheda M, et al. Macrocyclic monomers for high performance polymers. Polym Preprints, Am Chem Soc, Div Polym Chem 1993;34(1):576–7. [127] Miyatake K, Yokoi Y, Yamamoto K, Tsuchida E. Selective and efficient synthesis of cyclic hexakis(thio-1,4-phenylene) through oxidative polymerization of diphenyl disulfide. Macromolecules 1997;30(15):4502–3. [128] Wang Y-F, Chan KP, Hay AS. Synthesis and novel free-radical ring-opening polymerization of macrocyclic oligomers containing sulfide linkage. Macromolecules 1995;28(18):6371–4. [129] Hubbard P, Brittain WJ, Simonsick Jr WJ, Ross III CW. Synthesis and ring-opening polymerization of poly(alkylene-2,6-naphthalenedicarboxylate) cyclic oligomers. Macromolecules 1996;29(26): 8304–7.

681

[130] Dreger DR. Cast nylon – for rugged, long-wearing parts too large to mould. Mach Des 1974:75–7. [131] Tao Z, Chen S, Zhang H, Yu G. Monomer casting nylon – a satisfactory material for wind turbine blade. In: Proceedings of Asian and Pacific area wind energy conference, Shanghai, August 1–4, 1988. p. 251–2. [132] Boscoletto AB, Trezza G, Andreis B, Milan L, Tavan M, Furlan P. Anionic polyamides modified with poly(oxypropylene) by ‘‘oneshot’’ RIM technology: structural and morphological characterization. Macromolecules 1992;25(21):5752–8. [133] Ning X, Ishida H. Dynamic mechanical analysis of RIM nylon-6. J Polym Sci: Part B: Polym Phys 1991;29:1479–92. [134] Rusu G, Ueda K, Rusu E, Rusu M. Polyamides from lactams by centrifugal molding via anionic ring-opening polymerization. Polymer 2001;42:5669–78. [135] Gabbert JD, Hedrick RM, Tierney PA. Reinforced polymer compositions. USA: Monsanto Company; 1971. [136] Williams GT. Reinforced and foamed nyrim – an overview. Eng Plast 1992;5(1):48–52. [137] Chen C-H, Wang W-S. The development of glass fiber reinforced polystyrene for pultrusion. I: Process feasibility, dynamic mechanical properties, and postformability. Polym Compos 1998;19(4): 415–22. [138] Verrey J, Michaud V, Ma˚nson J-AE. Capillary effects in liquid composite moulding with non-crimp fabrics. In: Proceedings of 14th international conference on composite materials (ICCM-14), San Diego (USA), 2003. [139] Weyrauch F, Stadtfeld HC, Mitschang P. Simulation and control of the LCM-process with future resin systems. In: Proceedings of 7th international conference on flow processes in composite materials (FPCM-7), Newark (DE, USA), February 7–9, 2004. [140] Otaigbe JU, Harland WG. Studies in the properties of nylon 6-glass fiber composites. J Appl Polym Sci 1988;36:165–75. [141] Potter K. Resin transfer molding. London: Chapman & Hall; 1997. [142] Otaigbe JU. Effect of coupling agent and absorbed moisture on the tensile properties of a thermoplastic RRIM composite. J Appl Polym Sci 1992;45:1213–21. [143] Laura DM, Keskkula H, Barlow JW, Paul DR. Effect of glass fiber surface chemistry on the mechanical properties of glass fiber reinforced, rubber-toughened nylon 6. Polymer 2002;43: 4673–87.