Electron Beam Curing Of Composites.pdf

Felipe Wolff-Fabris Volker Altstädt Ulrich Arnold Manfred Döring

Electron Beam Curing of Composites

Wolff-Fabris, Altstädt, Arnold, Döring Electron Beam Curing of Composites

Felipe Wolff-Fabris Volker Altstädt Ulrich Arnold Manfred Döring

Electron Beam Curing of Composites

Hanser Publishers, Munich

Hanser Publications, Cincinnati

The Authors: Dr.-Ing. Felipe Wolff-Fabris, University of Bayreuth, Bayreuth, Germany Prof. Dr.-Ing. Volker Altstädt, University of Bayreuth, Bayreuth, Germany Dr. Ulrich Arnold, Karlsruhe Institute of Technology, Eggenstein-Leopoldshafen, Germany Prof. Dr. Manfred Döring, Karlsruhe Institute of Technology Eggenstein-Leopoldshafen, Germany Distributed in the USA and in Canada by Hanser Publications 6915 Valley Avenue, Cincinnati, Ohio 45244-3029, USA Fax: (513) 527-8801 Phone: (513) 527-8977 www.hanserpublications.com Distributed in all other countries by Carl Hanser Verlag Postfach 86 04 20, 81631 München, Germany Fax: +49 (89) 98 48 09 www.hanser.de The use of general descriptive names, trademarks, etc., in this publication, even if the former are not especially identified, is not to be taken as a sign that such names, as understood by the Trade Marks and Merchandise Marks Act, may accordingly be used freely by anyone. While the advice and information in this book are believed to be true and accurate at the date of going to press, neither the authors nor the editors nor the publisher can accept any legal responsibility for any errors or omissions that may be made. The publisher makes no warranty, express or implied, with respect to the material contained herein.

Library of Congress Cataloging-in-Publication Data

Electron beam curing of composites / Felipe Wolff-Fabris ... [et al.]. p. cm. Includes bibliographical references. ISBN-13: 978-1-56990-473-2 (hardcover) ISBN-10: 1-56990-473-1 (hardcover) 1. Electron beam curing. 2. Composite materials--Curing. I. Wolff-Fabris, Felipe. TP156.C8E44 2010 620.1’18--dc22 2010035942 Bibliografische Information Der Deutschen Bibliothek Die Deutsche Bibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliografie; detaillierte bibliografische Daten sind im Internet über abrufbar. ISBN 978-3-446-42405-0 All rights reserved. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying or by any information storage and retrieval system, without permission in writing from the publisher. © Carl Hanser Verlag, Munich 2011 Production Management: Steffen Jörg Coverconcept: Marc Müller-Bremer, www.rebranding.de, Munich Coverdesign: Stephan Rönigk Typeset: Manuela Treindl, Fürth Printed and bound by Druckhaus “Thomas Müntzer” GmbH, Bad Langensalza Printed in Germany

Contents 1 Introduction to Electron Beam Curing of Composites.. . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1 Principles of the Technique.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 Chemical Aspects of the Curing Reaction.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.2.1 Free Radical Polymerization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.2.2 Cationic Polymerization.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.2.3 Network Formation.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.3 Parameters Affecting Electron Beam Curing.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 1.3.1 Impurities.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 1.3.2 Irradiation Dose.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 1.3.3 Initiator Content.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 1.3.4 Thermal history. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 1.3.5 Irradiation Energy.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 1.4 Electron Beam Curing Facilities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 1.5 Safety Issues.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 2 Aspects of Electron Beam Curable Materials.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 2.1 Initiators.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 2.1.1 Onium Salt Initiators.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 2.1.1.1 Iodonium Salt Initiators.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 2.1.1.2 Sulfonium Salt Initiators.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 2.1.1.3 Other Onium Salt Initiators and Related Compounds.. . . . . . . . . . . 34 2.1.2 Metal Complex Initiators.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 2.1.2.1 Cyclopentadienyl Iron(II) Arene Complex Initiators.. . . . . . . . . . . . 37 2.1.2.2 Silver Alkene Complex Initiators.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 2.2 Neat Resins.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 2.2.1 Free Radical Polymerizable Resins.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 2.2.2 Cationically Polymerizable Resins.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 2.3 Toughened Resins.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 2.3.1 Liquid Reactive Rubber.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 2.3.2 Core-Shell Rubber Particles.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 2.3.3 Thermoplastics.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 2.3.4 Block Co-polymers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 2.3.5 Inorganic Nanoparticles.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 2.3.6 Hyperbranched Polymers.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 2.4 Interfacial Properties Between Fibers and Matrix.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 2.4.1 Fiber Surface Treatment and Use of Sizings.. . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 2.4.2 Processing Conditions.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 2.5 Residual Stresses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 2.6 Effect of Post-Curing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

VI

Contents

3 Electron Beam Curing Applied to Composite Molding Technologies.. . . . . . . . . . . . . 85 3.1 Layer-by-Layer Assembly. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 3.2 Prepregging.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 3.3 Vacuum Bagging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 3.4 Pultrusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 3.5 Filament Winding.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 3.6 Resin Transfer Molding (RTM) and Vacuum Assisted RTM (VARTM).. . . . . . . . . 90 3.7 Lost Core Molding.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 4 Current Limitations and Potentials for Electron Beam Curing.. . . . . . . . . . . . . . . . . . . 93 4.1 Cost Analysis of Electron Beam Curing Processes.. . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 4.2 Comparison Between Thermal and Electron Beam Cured Materials in Terms of Properties.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 4.3 Summary of Potential Applications for Electron Beam Curing.. . . . . . . . . . . . . . . . . 98 5 Research Trends and Projects in the Field of Electron Beam Curing.. . . . . . . . . . . . . 101 6 Examples of Electron Beam Curing Applications.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 6.1 Automotive.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 6.1.1 Composite Concept Vehicle, Daimler-Chrysler.. . . . . . . . . . . . . . . . . . . . . . . 105 6.1.2 Composite Armored Vehicle, US Army.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 6.2 Aircraft Industry. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 6.2.1 Patch Repairs for Civil and Military Applications.. . . . . . . . . . . . . . . . . . . . . 107 6.2.2 T-38 Talon, US Air Force.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 6.2.3 F/A-18 Hornet, Northrop Grumman. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 6.3 Space Applications.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 6.3.1 Space Shuttle Venture Star, Lockheed and NASA.. . . . . . . . . . . . . . . . . . . . . . 111 6.3.2 Satellite’s Flywheel, AFS Trinity and NASA.. . . . . . . . . . . . . . . . . . . . . . . . . . . 113 6.3.3 Satellite’s Reflector Dish, Acsion and CASA (Spanish space agency). . . . . 114 References.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 Appendix 1: Key Players in Electron Beam Curing.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 Appendix 2: List of Commercially Available Materials for Electron Beam Curing.. . . . 125 Subject Index.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127

1

Introduction to Electron Beam Curing of Composites

1.1

Principles of the Technique

Electron beam curing is based on the principle that fast electrons have the ability to initiate chemical changes in the material. The absorption of accelerated electrons in matter, with typical energies in the keV and MeV range, results on the initiation of secondary electrons as a consequence of the energy degradation process. These electrons enter into a Coulomb interaction with the atoms or molecules of the absorber, which finally leads to the formation of radicals, ions, trapped electrons, and an excited state of molecules or atoms [1]. Polymeric substances, which are predominantly high molecular weight organic compounds, such as plastics and elastomers (rubber), respond to radiation in several ways. Depending on the irradiation energy, these materials can either undergo degradation or chain crosslinking. On the other hand, electron beam irradiation can also be employed to prepare polymers from monomers or from oligomers (essentially very low molecular weight polymers). In almost all of these reactions, a liquid (reactive resin) is converted into a solid more or less instantaneously. This transformation of monomers, oligomers, and polymers by irradiation to a rigid body is referred to as curing. A wide range of high performance fiber reinforced polymers is currently based on thermosetting matrices, which are mainly a liquid reactive resin such as polyester, vinyl ester, epoxy, or bismaleimide resin. Thermosetting composites are cured at either ambient or elevated temperatures, and a hard solid is obtained due to the formation of a three-dimensional chemical cross-linked network. Typically, the use of radiation crosslinking decreases the curing time considerably, and allows processing at ambient temperature. Moreover, the use of high-energy electron beams allows the curing of thick fiber reinforced thermosetting composites. For instance, parts having a thickness up to 4 cm can be cured using a double-sided irradiation process. Studies regarding electron beam curing of polymers have been conducted since the 1960s, mainly by Charlesby and Chapiro [2, 3], and in particular the field of electron beam curing of thermosetting composites has been investigated since the 1970s. Saunders and Singh [4] have been working on EB-curing of numerous resins, such as acrylated epoxies. They have not only studied the properties of the EB-cured composites, but have also investigated the effect of electron beam treatment on resins, fibers, and on the fiber sizing. Concerning the manufacture of such composites, Beziers et al. [5] have published a number of papers on filament wound electron beam cured composites. There are a number of advantages of using electron beam irradiation to cure composites compared to conventional thermal processing. The reduced overall manufacturing costs,

2

1  Introduction to Electron Beam Curing of Composites

discussed in Section 4.1, is one of the major driving forces for the development of this new technology. Moreover, the opportunity to work at room temperature is one of the main advantages of this process and it derives from the fact that the reaction mechanism involved in the polymer curing does not require thermal activation. In particular the conservation of room temperature during the process allows the use of less expensive materials for the molds, and above all, it reduces the formation of internal stresses in the product, caused by the difference in the expansion of the cured system and the mold. Nevertheless, during irradiation the temperature of the materials can increase depending on different parameters. Heat evolution in the material is caused by the occurrence of two phenomena: the absorption of radiation energy and the exothermic polymerization reaction [6]. Curing times of composites are much shorter when using electron beam curing, leading to increased production rates. Furthermore, the resins (which are not designed to cure thermally) are more stable and thus have longer shelf life, making low temperature storage unnecessary. In addition, the electron beam process has environmental advantages because the materials have reduced emission of volatiles in comparison to typical thermally curable resins [7]. Finally, electron beam processing is a continuous operation, and components can be electron beam treated immediately after they have been produced. This continuous operation makes production scheduling and inventory control easier and reduces the number of identical molds needed to economically manufacture products, as compared to using a batch process, such as thermal curing in an autoclave.

1.2

Chemical Aspects of the Curing Reaction

High-energy electrons, when impacting matter, generate ionic species, free radicals, and molecules in excited states capable of initiating and sustaining polymerization. Depending on the chemistry of the resin system being irradiated, polymerization can proceed via a free radical or a cationic mechanism [8]. There are, however, some marked differences between cationic and free radical polymerizations. For example, there is nearly a first-order dependency of the polymerization rate on the irradiation intensity in cationic polymerizations, rather than the half-order relationship that is observed in free radical polymerizations. The rates of cationic polymerizations are also highly temperature dependent, whereas the rates of free radical polymerizations display little temperature dependence. The most striking difference between these two polymerization mechanisms is the ability of cationic polymerizations, once initiated, to proceed further in the absence of radiation. Most free radical polymerizations cease within a millisecond after irradiation shutdown because of rapid termination [9]. In the next sections, these two polymerization mechanisms will be presented in detail.

1.2  Chemical Aspects of the Curing Reaction

1.2.1

3

Free Radical Polymerization

Radicals can have lifetimes ranging from a sub-second timescale up to long-term persistence, depending on their chemical environment. Free radicals are generally very reactive and therefore free radical polymerization [10, 11] reactions exhibit frequently low selectivities. Consequently, several side reactions can occur and thus radical reactions are usually not as well defined as ionic polymerization reactions. The polymerization of monomers containing double bonds is often conducted thermally by free radical polymerization. The reaction cascade comprises three major steps: initiation, propagation, and termination (Fig. 1.1). Initiation is usually triggered by the use of peroxides. These peroxides homolytically decompose upon heating to form free radicals that add to the monomer, thus transferring the active site. Such initiators are not required for electron beam initiation since the irradiation of vinyl-containing monomers, such as acrylates and methacrylates, leads to the cleavage of double bonds to generate free radicals. R1

R2

R4

R3 R⋅

Initiation R4

R1

R

Propagation

⋅ R3

R2

R1

R2

R4

R3

Chain transfer

n

R4 R1 R4 R1

⋅

R 3

R R

2

3

R R2 n

R'⋅ Termination

R'H

Radical recombination R4 R1 R4 R1

R4 R1 R4 R1 R'

R R3 R2 R3R2 n

H

R

+

R'⋅

R3 R2 R3R2 n

Figure 1.1: Dominating reactions in the free radical polymerization of unsaturated monomers

4

1  Introduction to Electron Beam Curing of Composites

Polymerization proceeds via propagation steps that involve successive addition of vinylcontaining monomers to the active radical centre. Termination occurs, similarly to thermally curable systems, by several processes including radical recombination, chain transfer, disproportionation, and radical trapping. However, the polymerization induced by high dose rate electron beam curing can differ significantly from polymerizations initiated via thermal decomposition of initiators, as the high concentration of initiating radicals in electron beam curing affects the balance between propagation and termination [12]. A major concern of free radical polymerization and cross-linking processes regards the fact that this reaction can be inhibited by oxygen, which can react with the initiating radicals as well as the propagating free radical chains. The reaction with oxygen either prevents chain propagation or simply terminates the propagating chain. Accordingly, the residual oxygen concentration in the irradiation zone must be lowered to 100–1000 ppm (depending on the thermosetting system) by inert-gas purging [1].

1.2.2

Cationic Polymerization

A series of monomers such as olefins, vinylethers, or heterocyclic compounds like cyclic esters and ethers can be polymerized via cationic mechanisms [13–15] using directly Brønsted or Lewis acids as initiators or species that generate such acids upon energy impact either by heating or irradiation. In this context, sophisticated studies on the thermal curing of epoxy resins by cationic polymerization have been carried out employing boron trifluoride-amine complexes as initiators and it has been shown that these Lewis acid-base adducts exhibit high efficiencies [16–24]. According to the mechanism outlined in Fig. 1.2, polymerization is initiated by protonation of the epoxide ring. The resulting oxonium species are able to attack further epoxy groups, thus starting propagation reactions. n O R R-NH2⋅BF3

+

F3B-

O R

N H

H

R

O+ R

R H

O

[R-NH⋅BF3]-

O+ n

R

Figure 1.2: Cationic polymerization of epoxides initiated by boron triflouride-amine complexes

Polymerization propagation can proceed via two competing pathways (Fig. 1.3) referred to as the activated chain end (ACE) and the activated monomer (AM) mechanism. The former is characterized by monomer addition to the chain end, whereas the latter comprises alternately reactions of the protonated monomer with hydroxyl containing species or other proton donors and proton transfer reactions to the monomer. The activated center is switching continuously between the growing polymer chain and the monomer, thus maintaining a living polymerization. The mechanisms can proceed simultaneously and their extent mainly depends on the monomer/proton donor ratio as well as the polymerization stage. Termination of cationic polymerization can occur by reaction of cationic species with nucleophiles and therefore the nature of further components in the compositions can play a major

5

1.2  Chemical Aspects of the Curing Reaction

Activated chain end (ACE) mechanism O

n O R

HO

O R

H+

H O+

R

R HO

R

+

R

O+

O

n

R

R

Activated monomer (AM) mechanism

R

O '

HO-R

HO

O+

R'

HO

R

H

R

O R

R' +

H O+

Figure 1.3: Cationic polymerization of epoxides via ACE and AM pathways

role. The basicity and nucleophilicity of the involved species should be kept at a minimum and this applies particularly to anions present in the formulations. Non- or at least weakly basic anions such as SnCl5–, TiCl5–, BF4–, PF6– or SbF6–, which are frequently employed as counter­ions in cationic initiators, are suitable and can shield cationic propagating species from nucleophiles, e.g., water, by ion-pair formation, thus reducing termination reactions. Chain transfer is another prominent termination process (Fig. 1.4) that is promoted by protic species such as water or alcohols. These can react with cationic centers, thus generating protons that can accomplish again monomer activation and polymerization initiation. Chain transfer can also occur by inter- and intramolecular reactions of active centers with oxygen atoms in the already grown polyether chain. In contrast to radical-mediated reactions, these cationic curing reactions do not suffer from inhibition by molecular oxygen. (a)

R1

RXH

O+

R1

O

XR R2

R2

+

H+

R2 3

R (b)

R1

O+

+ 2

R

O

R3

O R4

O

O

R1

O+ R4

Figure 1.4: Chain transfer reactions in the cationic polymerization of epoxides by reactions of propagating species either with protic species (a) or oxygen in the polymer chain (b)

It is assumed that radiation induced cationic polymerizations proceed via similar pathways. However, due to radical and ionization reactions by direct interaction of the monomers and/ or initiators with high energy radiation, these processes can be, from a mechanistic point of view, much more complicated and, naturally, strongly depend on the radiation type. Since the discovery of highly efficient onium salt initiators for UV induced cationic polymerizations (see also Chapter 2), these light induced processes have been investigated in detail and the

6

1  Introduction to Electron Beam Curing of Composites

technology became well-established with numerous applications, e.g., in the fields of photoresists, printing inks, stereolithography, or adhesives [25–28]. Some research groups studied also curing by γ-irradiation, employing monomers such as epoxides [29–35], vinylethers [36], styrenes [37–39], or THF [40] and it was shown by direct comparison that similar results can be obtained using electron beam irradiation [29, 33]. Employing the latter technology, remarkable progress was made during the last years [5, 41–45] and activities in this research field concentrate on the cationic polymerization of commercially important epoxy resins ranging from mechanistic studies on a molecular level [46–50] to applications such as the preparation of high-performance composites [41–45, 51–53]. A significant stimulus was certainly the finding that classical initiators, originally developed for UV curing, exhibit also high efficiencies in electron beam processing. An interesting post-irradiation phenomenon observed for the electron beam curing of cationically polymerizing systems is that curing takes place even after exposure to the electron beam. A plausible explanation for this phenomenon could be related to the ions generated during radiation. The cationic cure process by electron beam irradiation can therefore be divided into two parts, “dark” reactions (reactions that occur after irradiation) and “light” reactions (reactions that occur during irradiation) [54]. The “light” reactions are characterized by a very rapid response to irradiation, with a reaction rate that increases approximately linearly with the applied dose, at least within the limitation of dose rate and total dose per pass. These reactions are also characterized by a rapid rate decay after irradiation, with a decay time constant that appears to follow an Arrhenius relationship. “Dark” reactions show a behavior quite different from that of “light” reactions. “Dark” reactions do not decay on the same time scale as the “light” reactions, although they possibly also have a finite life time. It is also reasonable to assume that “dark” reactions proceed unabated during irradiation while “light” reactions superpose these “background” reactions.

1.2.3

Network Formation

For both afore-mentioned polymerization mechanisms, free radical and cationic, the use of electron beam processing leads to a crosslinking reaction and therefore results in the formation of a compact and rigid network. The crosslinking reactions are controlled by reaction diffusion, instead of the dominance of molecular diffusion, and therefore the crosslinking reaction rate gradually decreases. The reaction rate rapidly slows down when the glass transition temperature (Tg) of the cured resin reaches the curing temperature. This event is known as vitrification. In addition, the increasing gel fraction must reduce the number of curable functional groups in the network due to saturation effects of crosslinking. Therefore, at this stage, the unit radiation dose brings about less curing reaction [55]. Although electron beam curing is a non-thermal activated cure process, it is known that during irradiation the temperature of the material increases mainly due to the exothermic behavior of the polymerization reaction. The effect of polymerization reaction heat can therefore be favorably used to obtain high conversion, as it delays vitrification, where the mobility of molecules

1.3  Parameters Affecting Electron Beam Curing

7

is restricted. Studies show that the increase of the temperature during electron beam curing provides more mobility to monomers and oligomers, resulting in higher final conversion [56]. Furthermore, the importance of the temperature reached during cure was shown also on the network homogeneity and the glass transition temperature of the matrix [6]. In composite materials the reinforcing fibers greatly suppress the resin exothermic reaction, both by displacing the resin and by absorbing the heat evolved. This is especially relevant for carbon fibers, which show high heat conductivity in fiber direction. These fibers added to the resin can be seen as an inert material that acts like a heat sink, as opposed to the exothermic resin. Hence it is possible that the crosslinking reaction in a composite occur to a lower extent as compared to the neat resin [57]. Therefore, a way to avoid this lower degree of cure is to attain a heat overshoot, produced by the radiation or the reaction energy of the resin, which may compensate the lower overall heat generation [58].

1.3

Parameters Affecting Electron Beam Curing

Many of the most attractive features of electron beam curing arise from the potential of unparalleled flexibility this process offers with respect to processing temperatures and rates of reaction. This flexibility means that electron beam irradiation can potentially cure components either quickly or slowly, at high or low temperature, and even selectively cure sections of a component, if required [59]. However, the same features of electron beam curing that lead to this flexibility also lead to the complexity that is potentially one of its biggest drawbacks. The numerous interacting phenomena involved in radiation curing of polymers and composites renders it difficult to predict a priori the appropriate process conditions required to produce high-quality robust components. In this section, the major aspects influencing the curing reaction of thermo­ setting polymers are discussed.

1.3.1

Impurities

Impurities in resins can slow down polymerization by “eating” the active species. Therefore, it is necessary to gain some critical concentration of cations or free radicals to compensate the effect of impurities in order to start propagation and consequently crosslinking of the chains [60]. The induction dose is related to this critical concentration of active species and can be attributed to the delay of the polymerization reactions due to kinetic reasons that cause an increase of the activation energy. Comparing systems having the same chemical nature, the difference in the induction dose is related to the presence of traceable amounts of protons acceptors impurities in the resin mixture.

8

1  Introduction to Electron Beam Curing of Composites

In free radical polymerization, the presence of oxygen can reduce curing efficiency or prevent crosslinking entirely. Oxygen inhibition is caused by deactivation of intermediate states, by formation of oxygenated radicals, and by the creation of oxidizing species including peroxy radicals. As a consequence, the surfaces remain tacky (oxygen-inhibited) and exhibit totally inadequate properties. To avoid this inhibition, high dose rate procedures can cure polymers before the diffusion of oxygen occurs [61]. Commercially available resins absorb a significant amount of water from the air. The water is often found on the surfaces of reinforcing components such as fillers or fibers. Such absorbed water plays an important role in the cationic polymerization reaction. Results suggest that water, at a concentration comparable to that found in resin exposed to typical ambient conditions, reduces the molecular weight of the resulting polymeric product [62]. Further substances that should be avoided in cationic polymerization include [63]: active nitrogens (amine, hydrazines, etc.), anionic surfactants, calcium carbonate and basic clays, alkaline materials, and strong anions (Cl–, Br–, OH–). Some of the impurities may not have a direct impact over the cationic curing reaction; however, these materials may lead to low final properties of the cured polymer. For instance, resin systems containing impurities such as ethylene glycol and chlorobenzene can still be cationically cured, and their curing degrees are even higher than that of the pure resin (without impurity), but the glass transition temperature of these cured resins is lowered by about 13–15 °C. Some types of impurities may serve as a plasticizer in the cured resin and some types of impurities may take part in the chemical reaction of the resin system [64].

1.3.2

Irradiation Dose

An important variable for electron beam processing is the dose delivered. Dose is defined as energy absorbed per unit mass and is expressed in “kiloGray “ (kGy), where one “Gray” is defined as one “Joule” per “kilogram”. The dose must be high enough to cause the desired effect (minimum dose), yet not as high as to cause any other undesired effects (maximum dose) [65]. In respect to free radical polymerization, unlike conventional chemical or thermally initiated curing, the radiation process can create all of the initiating free radicals in a fraction of a second. Although desirable from a process perspective, the resulting exotherm could thermally degrade, foam, or even ignite the part. Furthermore, a very high initiation rate (dose-rate) also enhances radical-radical reactions that cause premature termination of the propagating chain with concomitant reduction in the resulting average chain length of the polymer [66]. Although cationic polymerization proceeds in a different manner than free radical polymerization, similar care should be taken in respect to the irradiation dose. Processing is done in several passes to prevent over-heating of the material and tooling. Directed air cooling can be applied during the time between passes. The temperature rise is rapid during the first pass due to a combination of polymerization exotherm and direct electron beam heating. Additional passes accumulate the dose to maximize composite properties, but produce little additional exothermic effects [67].

1.3  Parameters Affecting Electron Beam Curing

9

The majority of polymer matrix composites produced by electron beam polymerization cure at doses ranging from 70–250 kGy. The specific dose depends on the specific resin used, the initiator used and its concentration, fiber type, and the percentage of fiber content. For the reasons mentioned before, this specific dose is usually subdivided in a series of low-dose irradiation passes.

1.3.3

Initiator Content

Initiators are only required for cationic polymerizations, as used for the cure of epoxy resins. The higher the initiator content in the resin systems, the higher is the concentration of reactive species under the same doses of radiation and as faster is the curing reaction rate [55]. However, at very high concentrations of initiator, plastification of the polymer matrix by photoinitiator degradation products may take place. This is reflected as a direct consequence in a decrease of the glass transition temperature of the resin [56]. Therefore, each initiator has an optimum concentration for a defined curing dose. When the dosage of initiator is below the optimum value, the concentration of radiated-induced active species is rather low in the resin system, leading to incomplete cure. In this case, further irradiation exposure or thermal post-curing is required to increase the conversion and the crosslinking density of the epoxy resin, and consequently to attain a higher glass transition temperature. The optimum content of most commercially available initiators is found to be between 1 wt.% and 3 wt.%, depending on the specific resin formulation.

1.3.4

Thermal history

The maximum temperature rise due to radiation exposure, that means the heating due only to the stoppage of electrons in the matter, can be estimated [68]: T=

D Cp

where: T = temperature rise in °C D = exposure dosage in kGy Cp = material specific heat in J/(g °C). Based on this equation, standard epoxy resins (Cp = 1.84 J/g °C) under usual electron beam irradiation curing conditions (dose of 35 kGy per pass), should exhibit a temperature rise of approx 19 °C per pass.

10

1  Introduction to Electron Beam Curing of Composites

However, the thermal energy causing the temperature rise in the irradiated samples is not only originated from the deceleration of the high-energy electrons, but also from the exothermic polymerization reaction. The temperature rise due to the chemical reaction may reach values up to 150 °C or 200 °C, depending on the specific irradiation conditions. This simultaneous thermal treatment affects not only the degree of cure and consequently the material’s properties, but the resulting temperature gradient in the material also causes residual stresses which affect the long-term performance of the resin system [69]. Therefore, the accurate knowledge of temperature evolution during curing is essential to control and optimize the EB cure process. After an induction dose, a steep temperature increase occurs up to a maximum value. After the pronounced peak, the temperature of the samples increases slowly due to the continuing absorption of radiation energy. A typical profile can be seen in Fig. 1.5, where the temperature of resins with 1 wt.% initiator and without initiator (this means: polymerization is not taking place) is shown as a function of the irradiation dose. The temperature rise during irradiation and curing was not only affected by the type of initiator used, but also by the viscosity and type of epoxy resin employed. The higher the viscosity, the lower the temperature rise [70]. Nevertheless, the major factor influencing the temperature rise on cationically cured resins is the amount of initiator. The increase of the initiator concentration causes a generalized increase of the temperature values reached during irradiation. The temperature profile depends on the balance between the rate of the heat production and the rate of heat released by the system towards the environment. Consequently, if there is no significant heat release to the environment, the temperature in the resin will increase rapidly to a very high value. Therefore, an increase in the initiator concentration or in the applied dose rates speeds up the curing reaction rate and the rate of heat evolution (see Fig. 1.5), without a

Figure 1.5: Temperature as a function of irradiation dose during electron beam irradiation of epoxy resins [30] (Reprinted with permission from Elsevier)

1.3  Parameters Affecting Electron Beam Curing

11

similar increase of the heat transfer rate from the reacting system towards the surroundings [30]. The type and amount of reinforcement fibers also influences the maximum temperature (Tmax) reached during curing, depending on the ability of the fibers to transfer heat to cooler locations in the sample. When comparing the same fiber volume content in composites, higher Tmax’s will be reached with glass fibers as compared to carbon fibers. The relatively low heat conductivity of the glass fiber leads to less cooling of the process zone. The glass transition temperature (Tg) of the crosslinked resin was also influenced by the fiber content in a way similar to how Tmax was influenced. When plotting Tg versus Tmax, a linear relationship could be observed. These results indicate that at higher curing temperatures, the mobility and the reactivity of the specific resin increases, thereby allowing the resin to come closer to its “fully cured” or “most possibly cured” state [71]. In order to increase the mobility of the reactive species during cure, heating of the samples prior to irradiation may have a positive effect on the curing process, as suggested by literature [71]. In this case, the curing reaction starts at around 80 °C. This resulted in Tmax values roughly 60 °C higher than in samples cured starting at room temperature. Consequently, a higher Tg was obtained. The additional heat development causes an increased mobility of the reactive species, and a higher degree of crosslinking was attained. Of course, in most cases it would not be economically interesting to heat up the composites in an oven prior to curing. Using a more powerful electron beam accelerator could compensate for a small lack of thermal energy in the bulk of the material, i.e., more heat would be produced in the material by the slow down of the incoming electrons. However, this would decrease the energy output of the system. A more effective solution would be the addition of highly exothermic reactive diluents to the resin in order to obtain a higher temperature during cure and a fully cured matrix.

1.3.5

Irradiation Energy

The irradiation energy is directly related to the effective beam penetration, and therefore to the maximum composite thickness that can be treated uniformly. The penetration depth can be estimated from the following relationship [72]: Wa = where:

E 2. 2

Wa = material areal weight in g/cm2 E = Electron beam energy in MeV.

Typical electron beam accelerators used for composite curing allow an energy of 10 MeV. Therefore, assuming a density of 1600 kg/m3 for the composite material, the maximum thickness that can be exposed is estimated to be 21 mm. The penetration depth of electrons as a function of energy is shown on the Fig. 1.6.

12

1  Introduction to Electron Beam Curing of Composites

Energy (MeV)

12 10

Carbon fibre composite 3

8

(density = 1.6 kg/dm )

6

Unit density

4 2 0

0

5

10

15

20

25

30

35

40

Penetration (mm)

Figure 1.6: Penetration depth in carbon fiber reinforced composite and in water

Table 1.1 summarizes some examples of penetration depth as well as the temperature rise for a variety of materials, assuming an accelerator energy of 10 MeV. The temperature rise is estimated for typical composite curing conditions at a dose of 35 kGy per pass. The total curing cycle demands usually four passes. Table 1.1: Temperature Heating and Electron Beam Penetration of Materials.

Material

Density

Penetration (E = 10 MeV)*

Specific heat

Temperature rise (D = 35 kGy)**

[g/cm3]

[cm]

[J/(g °C)]

[°C]

Aluminum

2.74

1.63

0.97

36

Steel

7.68

0.58

0.46

76

Epoxy

1.24

3.59

1.84

19

PVC foam

0.24

18.56

0.97

36

Polyurethane foam

0.77

5.79

1.67

21

Phenolic

1.55

2.87

1.30

27

Carbon fiber/epoxy

1.59

2.80

1.13

31

Glass fiber/epoxy

2.17

2.05

0.63

56

* Equation E = 2.2 · Wa ** Assumes adiabatic process. ∆T = Dose/Specific heat

1.4  Electron Beam Curing Facilities

1.4

13

Electron Beam Curing Facilities

Electron beam equipment has been used for the past 50 years for radiation processing of materials, but this included mainly irradiation of wires, tubes and films, or sterilization of medical products [12]. The number of worldwide installed electron beam machines used for industrial applications has been significantly increasing over the last 20 years. As it can be seen in the Fig. 1.7, the number of industrially used accelerators has more than tripled between 1991 and 2010. Electron beam machines play a significant role in the processing of polymeric materials; a number of different accelerator designs and different energy levels are available. Industrial electron beam accelerators with energies in the 150–300 keV range are in use in applications where low penetration depth is needed, such as curing of surface coatings. Accelerators operating in the 1.5 MeV range are used where more penetration is needed (i.e., crosslinking of cable insulation). High-energy commercial electron beam accelerators, operating in the range from 3–10 MeV, are used for applications such as curing of composites for aerospace applications and sterilization of boxes filled with disposable medical devices. Energies above 10 MeV are seldom used because of the imminent danger of generating radioactive high‑Z elements [7]. Theoretically, the amount of composite materials being cured by EB is directly proportional to the electron beam power, measured in kW. In practice, the requirement to stay in the specific temperature limits the rate at which curing is done. Industrially, electron beam processing is performed in high power accelerators (> 50 kW), which are commercially available and are equipped with a variety of material handling systems, capable of significant throughput [73].

Number of industrial EB accelerators

2500

2000

1500

1000

500

0 1990

1995

2000

2005

2010

Year

Figure 1.7: Evolution of the estimated number of EB accelerators over the last 20 years

14

1  Introduction to Electron Beam Curing of Composites

5 4

1 6 3

7

1. Pressure vessel filled with insulating gas 2. Titanium foil 3. Deflecting magnet 4. High voltage generator 5. Hot cathode 6. Concrete shield 7. Product

2

Figure 1.8: Working principle of an industrial electron accelerator for material processing

Many of the electron beam irradiation facilities that are used commercially today are captive; they are owned by a particular company and used exclusively on that company’s specific product line. There are other facilities that service many customers or companies coming to them to purchase beam time, and they process a whole range of different materials. A large variety of composite products can be cured at the existing industrial irradiation facilities. However, if very large composite products need to be manufactured, the radiation source and the target room need to be appropriately designed such that both the source and the product can be moved and rotated for uniform irradiation of the product [43]. A typical electron beam accelerator operates on the principle of a Braun tube, as shown in Fig. 1.8. Inside a volume (the beam tube), where the air has been extracted, electrons are generated by heating a filament (cathode). A strong electrical field causes the emitted electrons to be drawn (extracted) from the area surrounding the filament. The electrons are additionally accelerated and collimated as they travel down the beam tube and through a magnet. This magnet scans the electron beam back and forth in a high frequency alternating magnetic field (ranging from 100 Hz – 400 Hz) so that at the end of the scan horn the electron beam fans out and is projected onto the product, as a “shower” of electrons. A titanium foil separates the vacuum region from the atmosphere. Conveyor-type handling systems are widely used for irradiation of simple parts. The scan horn sweeps a “curtain” of electrons back and forth, while the transport system moves the part under the horn. Either cart-on-track systems or conveyor belts can be used to move the product through the concrete radiation-shielding maze. A composite part moving on a conveyor belt under the scan horn is shown in Fig. 1.9. Other facilities for EB curing and bonding beside the traditional closed maze have been proposed and demonstrated. Because narrow concrete walls and a fixed scan horn place limitations on the size and shape of composite parts to be cured, Aerospatiale and Science Research Laboratory have built large shielded rooms with movable accelerators. These facilities are accessible to large composites parts through sliding doors or removable concrete blocks.

1.5  Safety Issues

15

Figure 1.9: Electron beam irradiation of a composite part moving on a conveyor belt [44] (Reprinted with permission from Elsevier)

1.5

Safety Issues

The stoppage of fast electrons generates X‑rays, which are hazardous to human health. X‑rays cause cell damage that can lead to cancer formation or genetic mutations. Even at low dosage, X‑ray exposure can cause skin burning and general radiation syndrome [74]. The human lethal dose is approximately 4 to 5 Gy [75], which is extremely low compared to the dose required for the cure of composites (around 150 kGy). Thus, electron beam facilities must provide adequate shielding for the staff. Employees working in the area where electron beam equipment is operating must be monitored for exposure to ionizing radiation with film badges that detect and quantify any exposure to stray irradiation. Employers must train workers thoroughly in the operation of the equipment and proper safety and hygiene, and must keep records of exposure of their personnel to ionizing radiation. The majority of industrial electron beam installations rely upon poured concrete (density 2.35) or pre-cast concrete block to provide shielding from the X‑rays emitted when the unidirectional electron beams impinge upon a material [76]. Guidelines limiting personnel exposure have been established, which have led to the construction of concrete beam vaults of typically 1.5–3.5 m thickness for beam energies in the range of 3–10 MeV. Experienced accelerator manufacturers and designers of electron beam processing facilities provide more than adequate worker protection, including adroitly designed labyrinths and safety interlocks. A schematic view of such facility is shown in Fig. 1.10.

16

1  Introduction to Electron Beam Curing of Composites

Figure 1.10: Standard facility of electron beam irradiation site [77] (Reprinted with permission of the publisher)

Electron beam curing has been successfully used as an environmentally friendly, low-VOC (Volatile Organic Compounds) technology. All types of electron beam products can be handled safely as long as the proper industrial hygiene practices and engineering controls are utilized. Materials polymerized by electron beam irradiation are similar to those used in standard thermal curing method, and hazards regarding skin and eye contact are equivalent for both techniques. Thus, most people can work safely with electron beam curing materials by using the proper protective clothing and handling procedures. It is important to remember that electron beam curable materials do not evaporate, so spills and incidental contamination will remain until cleaned up. Because electron beam curing materials do not dry out or cure under normal ambient conditions, they remain liquid and can be cleaned up easily with less aggressive solvents, such as soap and water or citrus and vegetable oil cleaners [78]. Electron beam curable materials are designed to be reactive under controlled conditions, but a few products have the potential to generate a great amount of heat during uncontrolled bulk polymerization, mainly if they are placed in contact with other chemicals. If this polymerization reaction occurs in a sealed container, pressure can build up, which can result in blowing off the lid or violent rupture of the container.

2

Aspects of Electron Beam Curable Materials

2.1

Initiators

During the past decades, free radical electron beam curing, particularly of (meth)acrylates, was developed elaborately based on well-established technological and scientific principles [79–83]. In contrast, ionic polymerization in the field of electron beam curing was not widely investigated and the lack of suitable initiators restricted further developments in this field. The combination of conventional cationic initiators, typically Lewis or Brønsted acids, with multifunctional monomers resulted in nearly instantaneous and irreversible gelation. This high reactivity and lack of latency made it impossible to employ most of these cationically polymerizable systems in commercially important applications. A major breakthrough in ionic curing reactions initiated by electron beam irradiation came along with the key discovery of Crivello et al. that certain onium salts, such as diaryliodonium [Ar2I]+MFn– or triarylsulfonium salts [Ar3S]+MFn– with weakly or non-nucleophilic anions (MFn– = BF4–, PF6–, AsF6– or SbF6–), as seen in Fig. 2.1, can act as highly efficient photoinitiators for the cationic polymerization of virtually all known types of cationically polymerizable monomers, including epoxides, vinyl ethers, oxetanes, oxazolines, styrenes, and many others [9]. Moreover, these salts display excellent latency and, after dissolving of the initiator in a multifunctional monomer, polymerization can be triggered on demand by irradiating the mixture. Initial investigations concentrated on UV curing [84–89], e.g., of coatings, and it could also be shown that these compounds are very well suited as initiators for electron beam processing [90–93], even of deep section substrates and composites [94]. R2 R1

I+ MFn-

R2 S+ R1

MFn-

R3

Figure 2.1: Diaryliodonium and triarylsulfonium initiators (MFn– = BF4–, PF6–, AsF6– or SbF6–)

Various initiation mechanisms resulting in the formation of very strong Brønsted acids HMFn and/or highly reactive carbocations CR3+ capable of monomer activation and polymerization initiation have been discussed [95]. The dominating initiation pathways are summarized in Fig. 2.2 and can be divided into “non-redox pathways” covering HMFn and/or cation formation in a direct or multi-step procedure and “redox initiation pathways” where the onium salt functions as an oxidant converting radicals into cationic species. A cascade of reactions can take place and, to complicate matters further, some of these mechanisms can also proceed simultaneously.

18

2  Aspects of Electron Beam Curable Materials

Non-redox initiation pathways

• Multistep initiation

Irradiation

Initiator decomposition by heterolytic and/or homolytic bond dissociation

Reaction of emerging species with the monomer

Initiator rearrangement

HMFn and/or cation formation

HMFn release

• Direct initiation

Irradiation

Initiator decomposition and direct release of HMFn and/or (carbo)cations

Redox initiation pathways

Irradiation Generation of radicals stemming from

the monomer

initiator-monomer coupling products

external radical generating species added to the composition

Oxidation of radicals by the initiator Formation of HMFn and/or cations

Figure 2.2: Main initiation pathways in iodonium and sulfonium salt-induced cationic polymerization

According to Fig. 2.2, the following monomer activation and polymerization (Fig. 2.3) is initiated by Brønsted acids and/or cations generated upon radiation-induced onium salt decomposition. HMFn

+ Monomer

CR3MFn

+ Monomer

H(Monomer)+MFn-

n Monomer

CR3(Monomer)+MFn-

H(Monomer)n+1MFn

n Monomer

CR3(Monomer)n+1MFn

Figure 2.3: Cationic polymerization initiated by strong Brønsted acids HMFn or carbocations CR3+

2.1  Initiators

19

It was demonstrated that various essential characteristics of iodonium and sulfonium salt initiators can be assigned to the respective cation and anion moieties of these compounds [9, 96]. For the efficient initiation of polymerization by Brønsted acids, super acids derived from the initiator anion with Hammett acidities [97] (H0 values) in the range of –15 down to –30 are required. One of the strongest Brønsted acids known and thus one of the most powerful initiating species for cationic polymerization is HSbF6 [97, 98]. This acid is readily generated in situ by photolysis of onium salt initiators bearing the SbF6– anion. Fluorinated tetraphenylborates, e.g., B(C6F5)4– or B[C6H3(CF3)2]4–, represent another class of highly efficient anions and onium salt initiators bearing these anions exhibit activities similar to the corresponding hexafluoroantimonate compounds [99]. The rate of chain propagation depends on the basicity and nucleophilicity of the anion and polymerization rates increase with decreasing basicity and nucleophilicity of the anion [95]. Thus, within a series of photoinitiators with structurally identical cations but different anions MFn– or B(C6R5)4–, initiator activities increase in the order BF4– < PF6– < AsF6– < SbF6– ≈ B(C6F5)4– ≈ B[C6H3(CF3)2]4– Therefore, the nature of the anion and its stability determines the strength of the acid formed during photolysis and its corresponding initiation efficiency. The nature of the anion also determines the character of the propagating ion pair. This affects directly the polymerization kinetics and whether termination can occur by anion-splitting reactions. On the other hand, several important initiator properties, e.g., radiation absorption characteristics, radiation sensitivity, compatibility with other additives such as photosensitizers, solubility in the monomers, toxicity, or thermal stability strongly depend on the cation type and substituents at the cation framework. The thermal stability, for instance, has a direct impact on the latency of these photoinitiators and determines, ultimately, whether they can be employed in industrial applications. Besides the onium initiators, a radically different class of cationic photoinitiators was developed by the research department of the Ciba-Geigy Corporation. It was found that cyclopentadienyl iron(II) arene complexes generate Lewis acidic iron species upon irradiation by release of the arene ligand [100]. Subsequently, the arene ligand is substituted by monomer species and ringopening polymerization starts in the coordination sphere of the metal. However, the use of these initiators appears restricted to monomers that can effectively bind to the coordinatively unsaturated iron centre, e.g., epoxides. Very recently, another class of metal complex initiators for cationic polymerization was described. It was shown that silver olefin complexes are highly efficient initiators, especially for the electron beam processing of epoxy resins [101]. The most important initiators, both onium salt and metal complex initiators, their preparation, commercial availability, and initiation mechanisms are discussed in details in the following sections.

20

2  Aspects of Electron Beam Curable Materials

2.1.1

Onium Salt Initiators

The discovery that diaryliodonium and triarylsulfonium salts are easily accessible and efficient photoinitiators for cationic polymerization led to the development of various other onium salt initiators [95], mainly based on nitrogen, sulfur, or phosphorus compounds. However, most of the commercially available initiators for electron beam processing are diaryliodonium and triarylsulfonium salts of weak bases, due to their outstanding performance. Many parallels exist between diaryliodonium and triarylsulfonium salts [9]. For example, the initiation mechanisms of diaryliodonium salts follow a course very similar to that observed for triarylsulfonium salts. Triarylsulfonium salts undergo electron-transfer photosensitization like diaryliodonium salts, but they are not as readily photosensitized as diaryliodonium salts, due to their significantly lower (less positive) reduction potentials. However, the relatively low reduction potentials of the triarylsulfonium salts contribute markedly to the excellent thermal stability of these compounds even in the presence of the most nucleophilic and reactive monomers. As mentioned above, some key features of these initiators are governed by the substituents at the cation and can be tuned by appropriate choice. Therefore, a variety of synthesis strategies have been developed facilitating access to a series of onium salts with the desired substituents and properties. 2.1.1.1

Iodonium Salt Initiators

Compared to diaryliodonium salts, dialkyliodonium and alkylaryliodonium salts are much more difficult to prepare and to handle. Hence, alkyl-substituted iodonium salts, in contrast to alkyl-substituted sulfonium salts, did not find practical applications as initiators [102]. Figure 2.4 shows several strategies that have been developed to synthesize diaryliodonium salts in high yields via electrophilic aromatic substitution employing iodyl sulfate (IO)2SO4 (Eq. 2.1), potassium iodate KIO3 (Eq. 2.2), iodine(III)acylates I(OOCR)3 (Eq. 2.3), aromatic iodoso compounds ArIO (Eq. 2.4), iodoso diacetates ArI(OAc) 2 (Eq. 2.5), or aromatic iodoxy compounds ArIO2 (Eq. 2.6) as iodine source [102–105]. The synthesis route depends on the desired substituents at the aryl moieties and diaryliodonium salts bearing electronwithdrawing groups are favorably prepared starting from iodylsulfate, whereas iodonium salts with electron-donating groups are advantageously obtained using potassium iodate. Another efficient method to couple electron rich aromatic substrates is their electrophilic substitution reaction with hydroxy(tosyloxy)iodobenzene ArI(OH)OTs (Eq. 2.7) [106]. The reagent can be obtained in a one pot synthesis (Fig. 2.5) by oxidation of iodobenzene with peracetic acid, yielding the iodoso diacetate and subsequent addition of p-toluenesulfonic acid (Eqs. 2.8 and 2.9) [107]. Furthermore, reactions of hydroxy(tosyloxy)iodoarenes with aryltrimethylsilanes proceed with silicon carbon bond cleavage and offer a very selective pathway to diaryl­iodonium salts with tunable substituent placement in both rings [108]. Symmetrical iodonium salts with the same substituents at both aryl groups are obtained via reactions (2.1) to (2.3) whereas unsymmetrical substituted iodonium salts can be prepared according to equations (2.4) to (2.7).

21

2.1  Initiators

(IO)2SO4

+

H2SO4

Cl

4

I+

2 Cl



HSO4

KIO3

+

H2SO4, Ac2O

2

(2.1)

Cl -

(2.2)

I+ HSO4-

I(O2CR)3

+

2

HX

OMe



I+

MeO

IO

H2SO4

+



I(OAc)2

H2SO4

+

(2.6) R

HSO4-

OMe



I+

R

I+ OH

(2.5) R

HSO4H2SO4

+



I+

R

IO2

R

HSO4-

+

(2.4)

I+

R

(2.3)

OMe

X-

H2SO4

(2.7)

I+ HSO4-

-

TsO

R



Figure 2.4: Synthesis strategies for the preparation of diaryliodonium salts via electrophilic aromatic substitution: Eq. 2.1 to 2.7

I

CH3CO3H, CH3CO2H

I(O2CCH3)2

- H2O p -TsOH ⋅ H2O - 2 CH3CO2H

I(O2CCH3)2

(2.8)



I+ OH TsO-

(2.9)



Figure 2.5: Synthesis of hydroxy(tosyloxy)iodobenzene: Eq. 2.8 and 2.9

According to the procedures outlined in Fig. 2.4, iodonium salts with hydrogen sulfate, halogenides, or p-toluolsulfonate as counterions are obtained in most cases. These are not compatible with a cationic polymerization pathway due to the nucleophilicity of the anions. The anions combine with emerging cationic species thus interfering with the polymerization reaction. Anion exchange (Eq. 2.10) with bulky, non- or extremely low-nucleophilic complex metal fluorides MFn–, such as BF4–, PF6–, AsF6– and SbF6–, or fluorinated tetraphenylborates, such as B(C6F5)4– or B[C6H3(CF3)2]4–, yield highly active polymerization initiators. Some commercially available iodonium salts are listed in Table 2.1. Ar2I+X– + NaMFn → Ar2I+MFn– + NaX

(2.10)

22

2  Aspects of Electron Beam Curable Materials

Table 2.1: Commercially Available Aryliodonium Salts

Iodonium salt

I+

CAS-No.

Company a

Name or order number

178233-72-2

Rhodia Silicones S.A.S. TCI Europe N.V. ABCR GmbH & Co. KG Gelest, Inc.

Rhodorsil 2074 I0591 AB134236 OMBO037

139301-16-9

Sartomer Co., Inc. Aldrich

SarCat CD-1012 445835

121239-75-6

ABCR GmbH & Co. KG Hampford Research, Inc. GE Silicones

AB154263 OPPI OPPI

62613-15-4

TCI Europe N.V. Apollo Scientific Ltd. 3B Scientific Corporation Fluorochem Ltd. ABCR GmbH & Co. KG

D2248 PC0756 3B3-063894 009164 AB142786

58109-40-3

TCI Europe N.V. Aldrich 3B Scientific Corporation Pfaltz & Bauer Chemicals Carbone Scientific Apollo Scientific Ltd. Alfa Aesar Fluorochem Ltd. ABCR GmbH & Co. KG PCI Synthesis

D2238 548014 3B3-017307 D51230 C-22159 PC8214 A17934 009165 AB129475 600013

66003-76-7

TCI Europe N.V. Aldrich 3B Scientific Corporation Apollo Scientific Ltd. Alfa Aesar Fluorochem Ltd. ABCR GmbH & Co. KG

D2253 530972 3B3-063895 PC5029 L17444 009166 AB180350

61358-25-6

Midori Kagaku Co., Ltd. ABCR GmbH & Co. KG TCI Europe N.V.

BBI-102 AB171136 B2380

139301-16-9

Sartomer Co., Inc. Aldrich

SarCat CD-1012 445835

B(C6F5)4-

I+

O

SbF 6-

C12H25 OH

I+

O

C8H17

SbF6-

I+ AsF6-

I+ PF6-

I+ CF3SO3-

I+ PF6-

I+ SbF6-

O

C12H25 OH

2.1  Initiators

23

Table 2.1: (continued) Commercially Available Aryliodonium Salts

Iodonium salt

I+

CAS-No.

Company a

Name or order number

84563-54-2

Midori Kagaku Co., Ltd. Aldrich Acros Organics 3B Scientific Corporation ABCR GmbH & Co. KG TCI Europe N.V.

BBI-105 530999 37503 3B3-073732 AB171137 B2381

62051-09-6

Midori Kagaku Co., Ltd.

BBI-101

61358-23-4

Midori Kagaku Co., Ltd.

BBI-103

71786-70-4

Deuteron GmbH UFC Corporation Nanjing Chemlin Chemical Industry Co., Ltd.

UV 1242 DDHA NP50764

69842-75-7

Deuteron GmbH

UV 2257

153699-26-4

Nanjing Chemlin Chemical Industry Co., Ltd.

NP43110

139301-16-9

Sartomer Co., Inc. Aldrich

SarCat CD-1012 445835

CF3SO3-

I+ BF4-

I+ SbF6-

I+

H25C12

C12H25

SbF6-

I+ PF6

-

I+ B(C6F5)4-

I+ SbF 6-

a



O

C12H25 OH

Some selected suppliers

Several further developments in the field of iodonium salt initiators have been described and promising results were obtained using initiators adapted to the respective requirements by incorporation of irradiation-sensitive substituents, e.g., fluorenones [109]. Other strategies comprised the development of polymers containing iodonium salt moieties in the main chain [110] or the use of tailored sensitizers [111], e.g., dyes [112]. The mechanism of UV-induced cationic polymerization, especially of epoxy resin monomers, employing diaryliodonium salt initiators was intensely investigated by several research groups [84–89, 113–117] and electron beam initiation proceeds via similar reaction pathways [90–94]. The process is complex and comprises a cascade of reactions (Fig. 2.6), starting with photoexcitation of the iodonium salt, followed by heterolytic and homolytic carbon-iodine bond dissociation. The resulting aryl cations, aryliodine cation radicals, and aryl radicals are highly reactive and give rise to the formation of Brønsted acids HMFn in subsequent reactions.

24

2  Aspects of Electron Beam Curable Materials

I+ MFnirradiation

∗ I

+

MFnheterolytic bond dissociation

I

homolytic bond dissociation

I+⋅

+

⋅

MFn-

MFn-

R-H

I+ H

R-H

+

R

MFn-

R

+

HMFn

I

+

HMFn

Figure 2.6: Interaction of diaryliodonium salts with radiation and formation of super acidic species HMFn

Depending on the anion MFn–, superacids, such as HSbF6 with Hammett acidities down to H0 values around –30, can emerge upon irradiation, capable of monomer protonation and thus starting cationic polymerization. Apart from the basic reactions outlined in Fig. 2.6, a series of other initiation reactions [118], e.g., release of protons via radiation-induced decomposition and rearrangement reactions of the initiators (Fig. 2.7) [117], have been discussed. I

+

+

H + +

I

⋅

I I+

+

⋅

Figure 2.7: Generation of Brønsted acids via radiation-induced decomposition and rearrangement reactions of iodonium salts

H+

2.1  Initiators

25

Reactions where the iodonium salts serve as efficient oxidants (Fig. 2.1 and Fig. 2.8), e.g., oxidation of radicals emerging after high-energy electron irradiation and stemming from the monomer (Eq,.2.11) [91, 93, 94], direct reaction of aryliodine cation radicals with the monomer followed by oxidation of the coupling product (Eq. 2.12) [118–199], or generation of reactive carbocations by reaction of iodonium salts with easily oxidizable free radical photoinitiators (Eq. 2.13) [115, 120–122] yield highly reactive cationic species and have been observed particularly in the case of electron beam irradiation. In this context, reduction of the iodonium salt initiator by low-energy electrons followed by formation of protonic acids (Eq. 2.14) was also proposed as another probable mechanism [91, 92, 94]. As mentioned above, some of these mechanisms can also proceed simultaneously and superpose. Thus, electron beam curing of an epoxy acrylate in the presence of diphenyliodonium hexafluorophosphate was investigated and it was shown that acrylate as well as epoxide groups were involved in the polymerization [123]. The results lead to the assumption that acrylate polymerization proceeds via a radical process induced by slow electrons and phenyl radicals, whereas epoxide polymerization was induced by hydrogen fluoride, stemming from the decomposition of preliminarily generated HPF6. H

H

irradiation

O

-e

⋅

O+⋅

-

O

+

H+

(2.11) +

+

Ar2I+

⋅

+

I

⋅ O

O I+⋅

+

O

⋅

I



H + O

(2.12) + I

O

+

O OMe

- ArI, Ar OMe

e-

Ar2I+

⋅

C

- C6H5CO⋅

OMe

⋅

I

⋅

irradiation

C C

I+

Ar2I+

H+

- ArI, Ar⋅

OMe

I

+

⋅

+

MFn-

MFn-

Figure 2.8: Formation of cationic species via reduction of iodonium salts

H

+

O

OMe +

(2.13)

C



OMe H+

(2.14)

HMFn



26

2  Aspects of Electron Beam Curable Materials

A detailed study on the electron beam-induced cationic polymerization of epoxide monomers using iodonium salt initiators showed that polymerization depends mainly on the type of epoxide monomer, the onium salt concentration, the counterion, and the radiation dose [124]. A mechanism according to Eqs. 2.11 and 2.14 in Fig. 2.8 was proposed and it was concluded that the polymerization proceeds mainly via ion pairs, rather than free ionic species. Cationic polymerization of vinylether monomers was also investigated (Fig. 2.9) and the reaction proceeds via α-alkoxy radicals [91, 92, 124], analogous to the mechanisms outlined in Eqs. 2.11 and 2.14 related to cyclohexene oxide as monomer. O

irradiation

R

O

⋅

- e-

R

Ar2I+

O+⋅

R

O

R +

O

+

I

+

⋅

R

+

H+

⋅

Figure 2.9: Generation of cationic species upon electron beam irradiation of vinylether monomers in the presence of diaryliodonium salts

Recent research on the electron beam curing of iodonium salt-containing compositions focused on the use of epoxide monomers. The technology was compared to thermal and UV curing [125, 126] and the influence of various parameters such as resin type [55, 64, 127], initiator type and concentration [55, 64, 127–129], the presence of coinitiators [130], impurities [64], diluents [129] or hydroxyl containing compounds [62], dose and dose rate [55, 64, 127, 128], temperature [64], thermal treatment after irradiation [55, 128], as well as the influence of fiber surfaces [131] in the case of composites were studied. It should be pointed out that iodonium salts are among the most efficient initiators for electron beam processing. The compounds are meanwhile, like in UV curing, well-established and some of the most promising material features could be obtained employing these initiators [132–134]. 2.1.1.2

Sulfonium Salt Initiators

Aryl-substituted sulfonium salts, both symmetrically and unsymmetrically substituted, can be prepared using readily available diarylsulfides Ar2S, thiophenols ArSH, or diarylsulfoxides Ar2SO as starting materials. Among a variety of synthetic procedures [135–140] some strategies turned out to be superior in terms of applicability, versatility, scope, efficiency, yields, and costs (Fig. 2.10). These comprise, with respect to diarylsulfides and thiophenols as substrates, copper-catalyzed arylations using diaryliodonium salts (Eqs. 2.15 and 12.6) [141, 142]. Another approach is the self-condensation of diarylsulfides (Eq. 2.17), resulting in thiophenoxy-substituted triarylsulfonium salts [143, 144]. Considering diarylsulfoxides as substrates, reactions with aryl and alkyl Grignard reagents [145], favorably in the presence of trialkylsilyl triflates RSiMe2OTf [146] (Eq. 2.18) as well as reactions with aromatic hydrocarbons in the presence of aluminum halides [147] (Eq. 2.19), or phosphorus pentoxide-methanesulfonic acid

27

2.1  Initiators

P2O5‑MsOH [148] (Eq. 2.20) afford an extensive series of aryl-substituted sulfonium salts. According to the latter procedure, diarylsulfoxides can also be coupled with diarylsulfides (Eq. 2.21), yielding dicationic sulfonium salts [148].

S

Cu2+

I+

+

SH

+

Cu2+, NR3

S



X-

Ac2O, CH3CO3H, H2SO4

S

(2.16)

S+

X-

2



X-

I+

2

(2.15)

S+

X-

S

(2.17)

S+ X -

S

S

1. RSiMe2OTf 2. HX

O S

+

MgBr

(2.18)

S+



X-

O S

AlX3

+

(2.19)

S+ X-

O S

+

OR



MsOH/P2O5 S+

OR



MsO-

O 2

S

+

S

MsOH/P2O5 S+

S -

MsO

(2.20)

(2.21)

S+ MsO

-



Figure 2.10: Synthesis strategies for the preparation of triarylsulfonium salts: Equations 2.15 to 2.21

28

2  Aspects of Electron Beam Curable Materials

Apart from triarylsulfonium salts, sulfonium salts bearing both aryl and alkyl substituents represent another useful class of initiators (Fig. 2.11) and alkyldiphenyl sulfonium salts, for example, can be obtained via the acid-catalyzed alkylation of diphenylsulfide with alcohols and ethers (Eq. 2.22) [149, 150]. Preferably methanesulfonic acid CH3SO2OH (MsOH) is used in these reactions. Another versatile approach is the reaction of sulfur compounds such as diarylsulfides or tetrahydrothiophene with alkyl or benzyl halides via nucleophilic substitution (Eqs. 2.23 and 2.24) [151, 152]. Trialkyl sulfonium salts are considerably less stable than their aryl-substituted counterparts and are also photochemically inactive. Due to their strong alkylating properties, the compounds can initiate spontaneously cationic polymerizations [153] but did not find broad applications. S

+

Me

MsOH

OMe

(2.22)

S+ MsO-



Br Br-

S+ +

(2.23)

S

MeO

CH2Br

+

CH2 S+

MeO

S

Br-

(2.24)



Figure 2.11: Synthesis strategies for the preparation of sulfonium salts bearing alkyl, aryl and benzyl substituents: Equations 2.22 to 2.24

Two special types of sulfonium initiators bearing alkyl and 4-hydroxyphenyl or phenacyl substituents have been introduced by Crivello et al. (Fig. 2.12) [156]. The former can be obtained by condensation of dialkylsulfoxides with phenols [157] (Eq. 2.25) and the latter were synthesized by reaction of phenacylbromides with dialkylsulfides (Eq. 2.26) [158, 159]. O OH

+

CH3

S

HCl/CH3OH

CH3

- H2O

O C CH2Br

CH3 HO

ClO

+

CH3

S

CH3

(2.25)

S+ CH3

CH3

(2.26)

C CH2 S+ CH3 Br-



Figure 2.12: Synthesis strategies for the preparation of sulfonium salts bearing alkyl and 4‑hydroxyphenyl or phenacyl substituents: Equations 2.25 and 2.26

As described above for iodonium salt initiators, arylsulfonium salts with nucleophilic anions, e.g., halides, methane‑, trifluoromethane- or p-toluenesulfonates, have to be modified in a

29

2.1  Initiators

subsequent reaction step by anion exchange with bulky and minimal nucleophilic and basic anions, such as complex fluorides MFn– (Eq. 2.27). Ar3S+X– + NaMFn → Ar3S+MFn– + NaX

(2.27)

Direct synthesis of hexafluoroantimonate salts without additional ion exchange steps is also possible by reaction of glycidylethers with sulfides in the presence of hexafluoroantimonic acid hexahydrate HSbF6 · 6 H2O (Eq. 2.28) [151]. Some commercially available sulfonium salts are listed in Table 2.2. R1

O

O

+

R2

S

R3

OH

HSbF6x6H2O R1

O

R2 S+

SbF6R3



(2. 28)

Table 2.2: Commercially Available Arylsulfonium Salts

Sulfonium salt

S+

CAS-No.

Company a

Name or order number

75482-18-7

Bepharm Ltd. Nanjing Chemlin Chemical Industry Co., Ltd. AK Scientific, Inc. 3B Scientific Corporation Carbone Scientific Charkit Chemical Corporation AmplaChem

B22500 NP43032 P842 3B3-033856 C-34453 DTS-102 Aa34035

74227-35-3

Union Carbide Sartomer Co., Inc. Aldrich ABCR GmbH & Co. KG Degussa

Cyracure UVI 6990 SarCat CD-1011 531138 AB134315 Degacure KI-85

71449-78-0

Charkit Chemical Corporation AK Scientific, Inc. ABCR GmbH & Co. KG Nanjing Chemlin Chemical Industry Co., Ltd.

DTS-103 R033 AB134440 HC43022

89452-37-9

Gelest, Inc. Nanjing Chemlin Chemical Industry Co., Ltd.

OMAN076 HC42708

S

PF6-

Mixtures of

S+

S+

S

PF6-

PF6-

S+

S

PF6-

S+

S

SbF6-

S+ SbF6-

S

S+ SbF6-

30

2  Aspects of Electron Beam Curable Materials

Table 2.2: (continued) Commercially Available Arylsulfonium Salts

Sulfonium salt Mixtures of

S+

CAS-No.

Company a

Name or order number

159120-95-3

Union Carbide Sartomer Co., Inc.

Cyracure UVI 6974 SarCat CD-1010

145612-66-4

Midori Kagaku Co., Ltd. Aldrich 3B Scientific Corporation ACC Corporation SynQuest Fluorochemicals Apollo Scientific Ltd. Advanced Technology & Industrial Co., Ltd.

BDS-105 527076 3B3-073750 CHEM-16599 6670-3-12 PC3895 5000973

66003-78-9

3B Scientific Corporation Advanced Technology & Industrial Co., Ltd. Carbone Scientific ACC Corporation Charkit Chemical Corporation AmplaChem Aldrich SynQuest Fluorochemicals Apollo Scientific Ltd.

3B3-002775 3587440 C-21931 F-02760 TPS-105 Aa-33468 526940 6670-3-10 PC4475

57835-99-1

Pfaltz & Bauer Chemicals Advanced Technology & Industrial Co., Ltd. Charkit Chemical Corporation

T30958 1619412 TPS-102

57840-38-7

Pfaltz & Bauer Chemicals Nanjing Chemlin Chemical Industry Co., Ltd.

T30957 NP14298

S+

S

SbF6-

SbF6-

S+

S

SbF6-

S+ CF3SO3-

S+ CF3SO3-

S+ PF6-

S+ SbF6-

a



Some selected suppliers

2.1  Initiators

31

The main radiation-induced reactions of sulfonium salts (Fig. 2.13) are very similar to those described previously for iodonium salts (Fig. 2.6) [84–89, 115–117, 160–162]. Homolytic as well as heterolytic carbon-sulfur bond cleavage yields highly reactive cations and radicals resulting in, after further reactions with other substrate components, the formation of strongly acidic Brønsted acids HMFn. Subsequently, cationic polymerization can start by protonation of the monomer.

S+ MFnirradiation

∗

S+ MFnheterolytic bond dissociation

homolytic bond dissociation

⋅

+

S+⋅

S

MFn-

MFn-

R-H H R-H

S+

+

R

MFn-

S

+

R

+

HMFn

S

+

HMFn

Figure 2.13: Interaction of triarylsulfonium salts with radiation and formation of super acidic species HMFn

32

2  Aspects of Electron Beam Curable Materials

As described for the iodonium salt initiators (Figs. 2.7 and 2.8), various additional initiation mechanisms, e.g., radiation-initiated rearrangement reactions of the sulfonium salts [163] yielding Brønsted acids (Fig. 2.14) or dissociation of benzyl-substituted sulfonium salts [164, 165] yielding benzyl carbocations (Fig. 2.15), were proposed.

S

+

+ S+

S+

⋅

+

H

⋅

S

H+

+

Figure 2.14: Generation of Brønsted acids via radiation-induced decomposition and rearrangement reactions of triarylsulfonium salts CH3 S+

HO

irradiation

CH2

HO

S CH3

+

H2C+



(2.29)

S irradiation

S+ H2C

S +

H2C+

(2.30)

CN

S CN



Figure 2.15: Formation of benzyl cations by radiation-induced dissociation of sulfonium salts. Equations 2.29 and 2.30

Especially in the case of electron beam irradiation, reduction of sulfonium cations by slow electrons (Eq. 2.31 in Fig. 2.16) and radical oxidation by sulfonium salts yielding highly reactive carbocations and protons (Eqs. 2.32–2.34) were also discussed [91, 92, 94]. However, triarylsulfonium salts are significantly less reactive in such oxidation reactions than iodonium salts due to their significantly lower (less positive) reduction potential (E0 ≈ –1.2 V for sulfonium salts and E0 ≈ –0.2 V for iodonium salts vs. saturated calomel electrode SCE) [95]. Within these radical oxidation processes, oxidation of radicals derived from the monomer (Eq. 2.32), monomer/initiator coupling products (Eq. 2.33), or radical coinitiators (Eq. 2.34) are possible mechanisms [166].

33

2.1  Initiators

e-

S+

S

⋅

+

H+

MFn-

+

(2.31)

HMFn

MFn-



H

H

irradiation

O

⋅

O+⋅

O

H+

+

(2.32) +

+

Ar3S+

⋅

+

S

⋅

O



O

S+⋅

+

O

H

⋅

S

+ O

(2.33) + S

O

+

O OMe

irradiation

C C

Ar3S+

H+

- Ar2S, Ar⋅ OMe

- C6H5CO

OMe

C

(2.34)

+

C

- Ar2S, Ar⋅

OMe

+

O

OMe

Ar3S+

⋅

⋅

H

⋅

S



OMe

Figure 2.16: Formation of cationic species via reduction of sulfonium salts. Equations 2.31 to 2.34 CH3 +

HO

S

irradiation

CH3

CH3 O

S

CH3

(2.35)

HMFn

+

MFn-



O

CH3

C CH2

S+ CH3

irradiation

O

CH3

C CH

S

CH3

+

-

MFn

(2.36)

HMFn



Figure 2.17: Formation of Brønsted acids upon irradiation of 4-hydroxyphenyl- and phenacylsulfonium salts. Equations 2.35 and 2.36

In the case of 4-hydroxyphenylsulfonium and phenacylsulfonium salts, direct release of Brønsted acids from the initiators is promoted by the formation of the corresponding sulfur ylides (Fig. 2.17) [156–159].

34

2  Aspects of Electron Beam Curable Materials

Several sulfonium salts have been tested successfully as initiators for electron beam processing, predominantly in the curing of epoxy resins [90–94, 132–134, 167, 168] and the influence of various parameters such as epoxy resin and initiator type, initiator concentration, impur­ ities, radiation dose or temperature has been evaluated. However, direct comparison with the performance of analogous iodonium salts revealed that sulfonium salt initiators are frequently somewhat less efficient [64, 169, 170] and this could be basically due to their significantly lower oxidation power. 2.1.1.3

Other Onium Salt Initiators and Related Compounds

In addition to diaryliodonium salts, analogous diarylchloronium and diarylbromonium salts were also described (Fig. 2.18). These halonium salts can be obtained by reaction of the corresponding haloarene with an appropriate benzenediazonium salt in the presence of tri­fluoro­acetic acids [171]. It was shown that the compounds are, like the analogous iodonium salts, efficient cationic photoinitiators [172, 173]. However, preparation is more difficult and applicability is restricted due to their thermal lability [9]. In this context, triarylselenonium salts were also described [174], but did not show major advantages over their triarylsulfonium analogs (Fig. 2.18). Furthermore, these compounds are considerably more costly to prepare.

X+

Se+

MFn-

MFn-

X = Cl, Br

Figure 2.18: Structures of diarylchloronium, diarylbromonium, and triarylselenonium salt photoinitiators

Aryldiazonium salts [175–177] represent another class of initiators and these compounds are among the first onium salts that have been tested for photoinitiated cationic polymerization [178]. The salts can be obtained by reaction of aniline derivatives with sodium nitrite and a Brønsted acid. The latter component provides the non- or at least low-nucleophilic anion, required for cationic polymerization. Like in the case of diarylchloronium and diaryl­ bromonium salts, applicability of aryldiazonium salts is limited due to their thermal lability. Furthermore, nitrogen evolves upon irradiation, thus affecting material properties. However, some diazonium salts, such as N-morpholino-2,5-dibutoxybenzene diazonium hexafluorophosphate MDBZ from Sanbo Chemical Co. (Fig. 2.19), are commercially available. OBu O

N BuO

Figure 2.19: Structure of MDBZ from Sanbo Chemical Co.

N2+PF6-

2.1  Initiators

N+

N+

-

MFn

OR

(a)

N

OR

MFn-

(b)

O

S+

O

2

R

MFn-

(c)

O R1

+ P N P

N+

MFn-

Me

MFn-

(e)

(d)

O

S+

CH2

Me

MFn-

N+⋅ MFn-

C

(f)

(g)

Ar

R

R1 O

O

I+ R

+

O

R -

Ar

35

+

Ar

S

MFn

-

MFn

MFn

(h)

(i)

Ea+ (MFn-)a O

-

(j)

R2

b

E = B; a = 1; b = 2 E = Si, Ge; a = 1; b = 3 E = P; a = 2; b = 3

(k)

Figure 2.20: Onium-type initiators based on (a) alkoxypyridinium, (b) alkoxyisoquinolinium, (c) pyrazinium, (d) bis-(triphenyl phosphoranylidene) ammonium, (e) “non-ylide” and (f ) “ylide generating” sulfoxonium, (g) triarylaminium, (h) pyrylium, (i) thiopyrylium, (j) diaryliodosyl and (k) 1,3-diketonate salts

Numerous other onium-type initiators, such as alkoxypyridinium and alkoxyisoquinolinium [179], pyrazinium [180], bis-(triphenyl phosphoranylidene) ammonium [181], “non-ylide” and “ylide generating” sulfoxonium [182], triarylaminium [183, 184], pyrylium [184], and thiopyrylium [185, 186], diaryliodosyl [187, 188], or 1,3-diketonate salts of boron, silicon, germanium, and phosphorus [189] (Fig. 2.20), as well as sensitizers [190] have been described during the past decades. However, these compounds have been developed with a strong focus on photocuring and many of them have not yet been investigated as initiators in electron-beam induced curing processes. Several phosphonium salts were also successfully employed as photoinitiators [95]. The compounds can easily be prepared by reaction of phosphines with chlorides or bromides via nucleophilic substitution [191–195]. Triarylphosphonium salts of the type Ar3P+–CH2–R can release protons upon photolysis, thus generating the corresponding phosphorus ylides and

36

2  Aspects of Electron Beam Curable Materials

Brønsted acids capable of initiating cationic polymerization (Eq. 2.37) [196–203]. The acidity of the salts is particularly distinctive if a resonance stabilized ylide can arise and this can be controlled by choice of the substituent R [204].

R

irradiation

SbF6

P

R

P

+

(2.37)

HSbF6

However, a series of mono‑, di‑, tetra‑, and oligophosphonium salts (Fig. 2.21) has been tested as initiators for the polymerization of epoxy resins and, although showing good performance in thermal curing, exhibited only moderate activity in electron beam processing so that only partial curing was observed [205]. O P

P

O

O

P

SbF6

SbF6

SbF6

O

P SbF6

SbF6

O

O

P

P

SbF6 P

P

SbF6

P

(SbF6)2n+2 P

P P n

Figure 2.21: Mono‑, di- (n = 0), tetra- (n = 1) and oligophosphonium salt (n > 1) photoinitiators

SbF6

2.1  Initiators

2.1.2

37

Metal Complex Initiators

Several photoinitiator systems for cationic polymerization based on metal compounds [100, 206–209], e.g., manganese carbonyl complexes [210, 211], titanium complexes [212, 213], zirconium complexes [214], aluminum complexes [215], aluminum complexes combined with silylethers [216, 217], silver and copper compounds [218], ferrocene [219, 220], or cyclohexadienyltricarbonyl iron complexes [221] have been described. An outstanding performance was observed using cyclopentadienyl iron(II) arene complexes [222, 223] and these have also been tested, in addition to light-initiated processes, as initiators for electron beam induced cationic polymerization [166]. Very recently, silver olefin complexes were introduced as a new class of highly efficient initiators for the electron beam curing of epoxy resins [101]. 2.1.2.1

Cyclopentadienyl Iron(II) Arene Complex Initiators

Compared to similar organometallic compounds, cyclopentadienyl iron(II) arene complexes exhibit several advantages, e.g., tunable absorption characteristics by variation of the arene ligand, high thermal stability, high efficiency, particularly in the cationic polymerization of epoxides, and good availability due to moderate synthesis expenditure. The complexes can be prepared from ferrocene by substitution of one cyclopentadienyl ligand by an arene and subsequent anion metathesis to introduce a non-nucleophilic counter anion (Fig. 2.22) [224–226].

Fe

+

R

Al/AlCl3 - C5H5-

Fe+ AlCl4R

H2O/MFn- AlCl4-

Fe+ MFnR

Figure 2.22: Synthesis of cyclopentadienyl iron(II) arene complexes

Some of these iron complexes are commercially available, e.g., Irgacure 261 from Ciba (Fig. 2.22, R = isopropyl and MFn– = PF6–). The complexes were employed in the fields of coating [227], bonding [228], or imaging [229] and a mechanism for the light-induced initiation of cationic epoxide polymerization, based on sophisticated investigations on the coordination chemistry of such complexes [230], was proposed (Fig. 2.23) [231–233]. This mechanism implies activation of the complex upon irradiation and substitution of the arene ligand by the monomer followed by epoxide ring-opening in the coordination sphere of iron and thus initiating cationic polymerization. The hexafluorophosphate salt of the cyclopentadienyl iron(II) cumene complex (Fig. 2.22, R = isopropyl and MFn– = PF6–) was employed as initiator for the electron beam-induced polymerization of 3,4-epoxycyclohexylmethyl-3,4-epoxycyclohexanecarboxylate and a mechanism comprising both ionization of the epoxy resin by electron irradiation yielding HPF6 (Fig. 2.24, Eqs. 2.38–2.41) and epoxide activation by coordination to the iron center

38

2  Aspects of Electron Beam Curable Materials

hν +

O

Fe MFn

Fe+

R

-

- C6H5R

O

R

R

O

MFnO

R

O

MFn-

Fe

R

R

R

O

O

+

R

Figure 2.23: Light-induced epoxide polymerization in the presence of cyclopentadienyl iron(II) arene complexes

followed by ring-opening polymerization according to Fig. 2.23 was proposed [166]. It was pointed out that the anion PF6– could play an active role in the initiation process. Decomposition of HPF6 can produce Brønsted-acidic HF and Lewis-acidic PF5 (Fig. 2.24, Eq. 2.41) and the latter is known for its capability to initiate polymerization of cyclic ethers [234, 235]. Activity of the iron initiator in terms of curing degree vs. radiation dose was compared to corresponding sulfonium and iodonium hexafluorophosphates and activity increased in the order iron  sulfonium < iodonium salt. These significant differences in activity are probably due to the different initiation mechanisms. Formation of Brønsted acids by the onium salts, one central motif in multistep onium salt initiation, could be more efficient than direct interaction of the iron complex with the monomer according to Fig. 2.23. The lower activity of the iron complex can also be explained by its much lower (less positive) reduction potential [236] (E0 ≈ –1.8 V vs. SCE) compared to the onium salts (E0 ≈ –1.2 V for sulfonium salts and E0 ≈ –0.2 V for iodonium salts vs. SCE). Thus, the oxidation power of the iron complexes is significantly lower and generation of highly reactive initiating species, e.g., by radical oxidation as outlined above for the onium salts (Fig. 2.8 for iodonium salts and Fig. 2.16 for sulfonium salts) is less efficient. R

electron beam irradiation

O

R O



R O H

O

+

(2.38)

e-

+

+

⋅ O

+



(2.39)

R

R O+ H

HPF6

O+⋅

R

R +⋅

R

+

HF

PF6-

+

O

PF5

+

HPF6



(2.40) (2.41)

Figure 2.24: Epoxide monomer ionization by electron beam irradiation followed by HPF6 generation and decomposition. Equations 2.38 to 2.41.

2.1  Initiators

2.1.2.2

39

Silver Alkene Complex Initiators

Silver alkene complexes [Ag(alkene)a]MFn with weakly basic anions MFn– such as SbF6– represent another class of highly efficient metal complex initiators for cationic polymerizations and very recently, a series of such complexes was successfully employed as initiators for the electron beam induced curing of epoxy resins [101]. The initiators can easily be prepared by reaction of commercially available silver salts AgMFn with various linear and cyclic mono‑, di- and oligo-alkenes (Fig. 2.25), according to methods described in the literature [237–242]. In many cases the complexes precipitate from the reaction mixtures and can be isolated by filtration in yields above 95%.

AgSbF 6

+

a alkene

solvent

[Ag(alkene) a]SbF6

Alkene =

OH

O

O OEt

O

OH

O O

OH

COOH

Figure 2.25: Synthesis of silver alkene complex initiators and overview of the alkenes used

The coordination number of the silver atom in such complexes is usually two, three, or four and depends on the alkene and the anion. The use of alkenes with more than one double bond can result in the formation of coordination polymers and the structural complexity and variety increases with the number of double bonds in the alkene ligand. The reaction of AgSbF6 with 1,7-octadiene, for example, yields a one-dimensional coordination polymer (Fig. 2.26) consisting of silver centers alternately linked by one and two alkene moieties [101]. Important initiator properties such as resin solubility, radiation sensitivity, thermal latency, or reactivity can be controlled by choice of the alkene ligand. Formulations of epoxy resins with these initiators are light sensitive but can be stored in the dark without loss of activity and pot

40

2  Aspects of Electron Beam Curable Materials

SbF6-

SbF6-

Ag+

Ag+ Ag

+

SbF6-

n

Figure 2.26: Structure of the coordination polymer composed of AgSbF6 and 1,7-octadiene

lives at room temperature are in the range of at least several months. Thermal curing of such resin systems is also possible, but temperatures significantly above 100 °C are necessary and thus, the resins can be safely handled at temperatures up to 100 °C during some hours, even when exposed to daylight. The initiators exhibit high efficiencies in the electron beam curing of bisphenol A based epoxy resins or epoxy novolacs and their performance is similar to highly active iodonium salt initiators (see also Section 2.2.1). Electron beam curing of carbon fiber/epoxy composites is also possible and low initiator concentrations around 1% can be sufficient, if necessary combined with other additives, such as toughening agents, to obtain materials with very promising properties, e.g., high crosslinking degree and glass transition temperatures as well as high tensile strength and fracture toughness. An initiation mechanism was proposed that implicates alkene dissociation from the silver ion upon electron beam irradiation, thus generating Lewis acidic and oxophilic Ag+ species. These metal species can attack the epoxide ring (Fig. 2.27, reaction pathway A) and initiate cationic ring-opening polymerization that is promoted by heat evolved from the exothermic reaction (pathway B) and/or electron beam irradiation (pathway D). However, recent investigations [243] suggest that, at least in the case of some weakly coordinating alkene ligands, the alkenes can be substituted by epoxides (pathway C) already during mixing of the components at room temperature and polymerization is stimulated in the following step by electron beam irradiation (pathway D). As mentioned above, thermal curing is also possible and this process could proceed via pathway C followed by B. Naturally, the anion plays a crucial role, like in the case of onium initiators (see also Section 2.2), and analogous silver complexes with the BF4– anion B ∆T / n O

A electron beam / O [Ag(alkene)a]SbF6

R - x alkene - SbF6

R

[(alkene)a-xAg] O+

[(alkene)a-xAg]

O+ R

R

C ∆T / O

O

n

R

D electron beam / n O R

R

- x alkene - SbF6

Figure 2.27: Probable mechanisms of the electron beam induced cationic polymerization of epoxides employing silver alkene complex initiators

2.2  Neat Resins

41

are much less active than complexes with the SbF6– anion, even with higher initiator concentrations up to 5%. In the case of onium and iron arene initiators, it was pointed out that the oxidation power of the initiator and its capability to oxidize radicals generated by interaction of the monomer with high energy electrons is another important parameter. The reduction potential of silver compounds is significantly higher (E0 ≈ 0.8 V vs. SCE) [244] compared to onium and iron arene initiators and further investigations have to clarify the contribution of such redox processes to the entire curing process. The results obtained with silver alkene complexes show that metal complexes can be a powerful alternative to widespread onium salt initiators. There is a great potential for optimization, e.g., by variation of the metal, the ligand, and the anion and a better understanding of the fundamental initiation mechanisms will certainly result in several new and highly efficient polymerization initiators.

2.2

Neat Resins

The electron beam induced polymerization of acrylic/methacrylic systems, maleic and fumaric polyester resins, maleimides, and thiol-ene systems proceeds without addition of initiators via free radical reaction mechanisms. On the other hand, epoxides, vinyl ethers, oxetanes, oxazolines, and styrene can be polymerized cationically in the presence of an appropriate initiator, which dissociates under electron beam irradiation, as already explained in Section 1.2.2. From the above mentioned resin systems, free radical cured resins based on acrylate and methacrylate functionalities, and cationically cured epoxides obtained by using diaryliodonium or triarylsulfonium salt initiators, are the most promising electron beam curable resin systems for composite applications [12, 58]. The choice between free radical polymerized acrylate/ methacrylate or cationically polymerized epoxy resins as polymer matrix for the manufacture of composites relies on the specifications required for the respective application. Nowadays, both classes of resins are commercially mature and further details about each of them are given within the following sections.

2.2.1

Free Radical Polymerizable Resins

During the past decades, a whole series of resin systems has been electron beam cured with differing results. For example, oligomers containing acrylate and methacrylate groups, as well as dienes and vinyl compounds have been reported to be electron beam curable [108]. Many highly reactive monomers, oligomers, or prepolymers suitable for free radical polymerization can be produced by acrylation of precursors, such as epoxy resins, urethanes, polyesters, polyethers, silicones, oligobutadiene, melamine derivatives, cellulose, and starch employing a variety of acrylating agents, e.g., acrylic acid, acrylamide, hydroxyethyl acrylate, or glycidyl acrylate [79].

42

2  Aspects of Electron Beam Curable Materials

In cooperation with resin suppliers, Aerospatiale France developed the first practical applications of electron beam cured structural composites for solid rocket motors using acrylated epoxy resins. These resins yielded composites that were quite satisfactory with respect to the structural rigidity called for in this application [58]. However, extensive testing of composite parts based on these matrix resins is still required to qualify them for use in manned aerospace applications. Comparing the reactivity of acrylate and methacrylate resins curable by electron beam irradiation, it was shown that acrylate functionalized resins can release more heat and polymerize more rapidly. This is due to differences in molecular motion of monomeric species in the partly crosslinked system [71]. Typical structures of acrylate and methacrylate monomers are shown in Fig. 2.28 and major suppliers are Sartomer with its main trademarks SARTOMER (SR) for (meth)acrylate monomers, CRAYNOR (CN) for (meth)acrylate oligomers, and M-CURE for specialty monomers, as well as UCB Chemicals with its trademark EBECRYL. O

O O

O R

O

O

OH (a) CAS: R = H, 4687-94-9 R = CH3, 1565-94-2

OH

O

O OH

n

R

O O O N

O N R

O

R O

N

O

O O

R

Si

R

O

Si O O n

R

O (b) CAS: R = H, 40220-08-4 R = CH3, 35838-12-1

(c) CAS: R = H, 6999-52-6 R = CH3, 2003-87-4

O

O

N O

N R

R

R

O

R

O N

O

R

(d) CAS: R = H, 959-52-4 R = CH3, 27325-67-3

O

O

O

O R

R

(e) CAS: R = H, 15625-89-5 R = CH3, 3290-92-4

R

O

O O

O

O

O O

O R

R

(f) CAS: R = H, 4986-89-4 R = CH3, 3253-41-6

Figure 2.28: Structures and CAS registry numbers of some commercially available acrylate (R = H) and methacrylate (R = methyl) monomers based on (a) diglycidylether of bisphenol A monomer and oligomers, (b) isocyanuric acid, (c) oligodimethylsiloxane, (d) hexahydro-1,3,5-triazine, (e) trimethylolpropane and (f ) pentaerythritol

2.2  Neat Resins

43

Electron beam curable composites with acrylate/methacrylate matrices have been studied extensively during the last years [245, 246]. These systems exhibited high polymerization rates and allowed for a good control of processing viscosities. Furthermore, the resulting materials showed in general good stiffness. Another advantage is the very long shelf life, as initiators are not incorporated in these resins. Considering the shortcomings associated with such systems, the high cure shrinkage, ranging typically from 8 to 20%, the possibility of oxygen inhibition, the comparatively low glass transition temperatures, and the high moisture absorption should be mentioned [12].

2.2.2

Cationically Polymerizable Resins

The disadvantages of (meth)acrylated matrices for use in high performance composites, e.g., for aerospace applications, directed research and development efforts to another area. In the mid-1990s, researchers from the Oak Ridge National Laboratory (ORNL) and Atomic Energy Canada Limited (AECL) found that commercially available resins could be rendered electron beam curable by incorporation of cationic initiators and excellent materials could be obtained. The range of monomers polymerizable by a cationic mechanism encompassed vinyl- and epoxide-containing compounds and these allowed for the preparation of a wide range of high performance polymers. In the meantime it was shown that almost every type of cationically polymerizable monomer can be polymerized by electron beam irradiation. Naturally, the use of multifunctional monomers such as di- and oligo- vinylethers or epoxides yields highly crosslinked polymers with superior properties compared to systems based on monofunctional monomers [9]. Electron-rich vinyl compounds are exceptionally reactive monomers for cationic polymerizations and Crivello et al. demonstrated that multifunctional vinyl ether monomers are ideally suited for radiation induced cationic polymerizations [247]. However, the high costs of these monomers and their lower performance compared to alternative monomers limited their use in many applications [90]. Especially interesting for electron beam induced cationic curing are epoxide monomers, as these resins and the resulting materials are to a large extent already approved for use in many composite applications, even for numerous highly demanding aerospace applications [43]. Moreover, cationically cured epoxy systems exhibit low shrinkage, usually below 6%, and high glass transition temperatures. In contrast to (meth)acrylated resin systems, curing is usually not affected by oxygen. On the other hand, cationic systems based on epoxy resins tend to cure more slowly than (meth)acrylate systems and the resin systems can be easily poisoned by nucleophilic contaminants [12, 63]. A large assortment of epoxy monomers is commercially available but the suitability of such monomers for electron beam processing differs considerably. Studies show that the reactivity is strongly affected by both steric and electronic factors as well as by the presence of other functional groups in the monomer [45]. The structures of some important electron beam curable epoxy resins are summarized in Fig. 2.29. These can be glycidylethers, e.g., the monomeric (n = 0) or oligomeric (n ≥ 1) diglycidylether of bisphenol A and epoxy phenol (R = H) or cresol (R = methyl) novolacs, glycidyl esters, e.g., the diglycidylester of hexahydrophthalic acid,

44

2  Aspects of Electron Beam Curable Materials

or cycloaliphatic epoxy resins bearing usually two cyclohexene oxide moieties. Some major epoxy resin suppliers are Dow Chemical (trademarks: D.E.R. and D.E.N.), Hexion (EPON, EPICOTE), Union Carbide (ERL), Huntsman (ARALDITE), Leuna Harze (EPILOX), and Nan Ya (NPEL). It should be pointed out that most of the nitrogen containing epoxy resins, e.g., glycidylized 4,4‑methylenedianiline or 4‑aminophenol used for high performance aerospace applications, are, due to their basicity, not compatible with a cationic reaction mechanism. Within the glycidylether based epoxy resins, which are predominantly employed for the manufacture of high performance materials by electron beam curing, reactivities can vary significantly. For instance, the density of epoxy groups in epoxy novolacs is higher than that in bisphenol A based resins and this is reflected in the lower epoxy equivalent weights of the former compared to the latter resins. Hence, the curing reaction is stronger exothermic and this gives rise to a higher temperature in a shorter time. Consequently, epoxy novolacs can reach a

O

O

O

O OH

O

O

n

(a) CAS: 1675-54-3

O

O

O

O O

O

O

O

O

O

O

O O

O R

R

R n

O

(b) CAS: 028064-14-4

(c) CAS: 5493-45-8

O O

O

O O

(d) CAS: 3388-03-2

O O

O O

(e) CAS: 2386-87-0

O

O

(f) CAS: 2754-17-8

Figure 2.29: Structures and CAS registry numbers of some commercially available epoxy resins for electron beam curing: (a) diglycidylether of bisphenol A monomer (n = 0) or oligomers (n ≥ 1); (b) epoxy phenol (R = H) or cresol (R = methyl) novolacs; (c) diglycidylester of hexahydrophthalic acid; (d) 2-(3,4’-epoxycyclohexyl)-5,1-spiro-3,4-epoxycyclohexane1,3-dioxane; (e) 3,4-epoxycyclohexylmethyl-3,4-epoxycyclohexanecarboxylate; (f ) bis(3,4-epoxycyclohexylmethyl)adipate

2.3  Toughened Resins

45

higher crosslinking degree than bisphenol A glycidylethers at a lower radiation dose [127]. The molecular weight of the monomer also influences the reactivity of the resins. With increasing molecular weight, the gel fraction of epoxy resins decreases slightly at a given irradiation dose and this is associated with a limited molecular motion and finally a lower reaction rate. On the other hand, glass transition temperatures increase with increasing molecular weight and this indicates that the molecular weight affects both crosslinking degrees and glass transition temperatures [55].

2.3

Toughened Resins

In fiber-reinforced composite materials, cracks may originate from manufacturing defects, such as microvoids, matrix microcracks, and ply overlaps, or at localized damages caused by in-service loadings, such as subsurface delaminations due to low-energy impacts and hole-edge delaminations due to static or fatigue loads. These cracks grow into the resin-rich interlaminar region between crossed plies and can be redirected into the interply, generating delaminations, which is considered the most critical failure mode limiting the long-term fatigue life of composite laminates [278]. Therefore, tough resin systems are critically important for use in many structural composite applications because they can significantly reduce the probability of catastrophic failure caused by impact damage compared with brittle resins. Additionally, toughness impedes the initiation and progression of damage leading to sudden failure [248]. Typically, the interlaminar fracture toughness of composites, measured in terms of the critical strain energy release rate, is determined for Mode I (tensile mode of crack propagation) [279] and for Mode II (in-plane shear mode of crack propagation) [280]. These properties are used for comparing the resistance of various resin systems against the growth of delamination failure. In general, the delamination growth resistance increases with increasing fracture toughness of the matrix, as seen in Fig. 2.30. However, one should be aware that improving the fracture toughness of a matrix may not completely translate into an equivalent increase in the delamination growth of a laminate [281]. This is due to the fact that tough polymers present a relatively large volume of plastic deformation during crack propagation. However, the presence of fibers in the composite material limits the volume available for plastic deformation of the resin and hence, limits the fracture mechanisms responsible for increasing the toughness. Due to the fact that resins cured with electron beam usually present very homogeneous, highly crosslinked networks, these resins tend to be particularly brittle. Resins intended for use in aircraft, aerospace, and many other applications must therefore be modified to increase toughness and impact resistance to endure many years of harsh service [71]. There are different approaches to modify the resins in order to enhance the fracture toughness, such as modification with liquid rubbers, core-shell particles, thermoplastics, block co-polymers, and nanoparticles. In these cases, the resulting heterogeneous structure promotes new deformation mechanisms that may increase the fracture toughness of the material.

2  Aspects of Electron Beam Curable Materials

Interlaminar fracture energy - GIc (J/m2)

46

500 1 : 1 translation

400 300 200 100 0

0

200 400 600 800 2 Neat resin fracture energy - GIc (J/m )

Figure 2.30: Translation of neat resin fracture energy to composite interlaminar fracture energy, both in Mode I Temperature t0,Mw0, visc0 reaction advances

t1,Mw1, visc1 Tg, neat modifier

2-phase region T cure 1-phase region

Tg, t1 Tg, t0

(LCST) TP Thermoplastic volume fraction

Figure 2.31: Modification of the phase diagram of a thermoplastic-thermoset blend (Lower Critical Solution Temperature diagram) due to the progress of the curing reaction along with time (from t0 to t1)

In the case of rubber, thermoplastic, and block co-polymer modification, the formed heterogeneous structure is the result of a phase separation mechanism. In contrast to most binary thermoplastic blends, whose final morphology is mainly controlled by thermodynamics, for any thermoplastic-thermoset blend there is a continuous change of molecular weight of the thermosetting component, viscosity of the system, glass transition temperature of the phases, and crosslinking density as a result of the polymerization process. The situation is schematically

2.3  Toughened Resins

47

represented for a thermoplastic in epoxy with lower critical solubility temperature behavior, see Fig. 2.31. Thus, the final shape, size, or size distribution of the formed phases is not only determined by the thermodynamics of the process, but also by the reaction kinetics of the thermoset-modifier blend. Hence, a careful consideration of these factors must be exercised when attempting to toughen thermosetting systems employing such modifiers. In the next sections, typical methods used for toughening thermally curable epoxy resins, including the incorporation of XTBN reactive rubbers, core-shell rubbers, high-performance thermoplastics, inorganic nanoparticles and hyperbranched polymers, are briefly presented and the suitability of each one of these approaches for electron beam curable resins is discussed. In the literature, there is a lack regarding systematic and fundamental investigation of the toughness modification of electron beam curable resins. Therefore, the authors will present here in-depth results from a large research work focused on the toughness modification of model epoxy resin systems cured with silver salt complex initiators.

2.3.1

Liquid Reactive Rubber

The incorporation of elastomeric modifiers has been a successful way of enhancing the fracture toughness of thermally curable thermosets, such as epoxy resins [282]. In general, 5–15 wt.% of an initially soluble liquid rubber is added to the resin and phase-separates during curing. Particles with diameters between 100 nm and 10 µm are thereby formed. The addition of a rubber modifier improves the fracture toughness but usually at the expense of bulk properties, such as stiffness or thermal resistance. This is not unexpected because the modulus of the toughness modifier is much lower than the modulus of the thermosetting matrix [283]. Typical rubber modifiers for epoxy resins are low molecular weight liquid butadiene-acrylonitrile copolymer bearing carboxyl (CTBN) [284], amine (ATBN) [285] or epoxy (ETBN) [286] reactive end groups. The rubber-rich domains can promote toughening by various mechanisms that are determined by the amount of dissolved rubber, their domain size, and the extent of phase separation. The formed morphology is clearly influenced by the epoxy-rubber compatibility before curing, the curing agent, and cure kinetics [287]. The toughening effect promoted by the rubber particles has been attributed to plastic shearyielding in the matrix. Owing to the modulus mismatch between the modulus of the rubber particle and the matrix, the rubber particles behave as stress concentrators. Thus, in contrast to a hole, which would produce similar stress concentration, the rubber particles can better bear the applied load. The stress concentration around the particle leads to the initiation of two major energy dissipation mechanisms. These are [287]: yy initiation and growth of multiple localized shear-yield deformation in the matrix that causes an enhanced plastic deformation of the surrounding matrix, and yy cavitation of the rubbery particles which can promote void formation; such voids lower the stress required for shear yielding and so promotes extensive plastic shear deformations in the matrix (shear banding).

48

2  Aspects of Electron Beam Curable Materials

Figure 2.32: Model of the process zone around the crack tip of a rubber-modified thermoset illustrating various toughening mechanisms

These mechanisms enable the formation of a quite large plastic zone ahead of the crack tip, where a large volume of material above and below the crack plane has to undergo plastic deformation. This, in turn, helps absorbing a high amount of energy, thus increasing the fracture toughness of the material. A summary of these mechanisms is illustrated in Fig. 2.32. Although successful for thermally curable resins, this approach cannot be used for electron beam curable epoxy resins, as typical rubber modifiers (e.g., CTBN or ATBN) are based on polybutadiene-co-polyacrylonitrile. The basicity of the acrylonitrile group prevents the cationic curing of the epoxy, as they react with the epoxy macro-cations, thus stopping the crosslinking reaction.

2.3.2

Core-Shell Rubber Particles

Besides the use of soluble liquid rubbers, another form of rubber toughening of thermosets is attained by the application of pre-formed elastomeric particles. These particles, e.g., butadiene rubber or acrylic rubber particles can be obtained by an aqueous emulsion polymerization process [288, 289]. Moreover, core-shell microparticles are prepared to enhance the interfacial adhesion of non-compatible rubber particles by grafting a compatible polymeric shell onto an elastomeric core. Versatile core-shell toughening modifiers consist of a polybutadiene core and a thermoplastic shell of either styrene/acrylonitrile or styrene/methyl methacrylate [290, 291]. Depending on the acrylonitrile or methyl methacrylate content, the compatibility and thus the interfacial adhesion between shell and thermosetting matrix can be varied over a wide range. The toughening mechanisms promoted by core-shell particles in thermally curable low crosslinking density resins is similar to that promoted by liquid rubber (already described in the previous paragraphs) [292]. As expected, the Young’s modulus at room temperature decreases when incorporating the core-shell particles into the thermoset. However, the coreshell particles, in contrast to liquid rubbers, show the advantage of not negatively affecting the

2.3  Toughened Resins

49

glass transition temperature of the thermally cured material [293, 294]. This indicates that the epoxy network is not plasticized by the core-shell presence. On the other hand, a decrease of the glass transition temperature was reported for electron beam cured resins modified with core-shell particles [249]. Such systems were also investigated in detail by the authors. The initial step of the modification of electron beam curable epoxy resins with core-shell particles was a screening investigation of different commercially available products. It is known that, depending on the proprietary manufacturing process, impurities may be present, affecting the electron beam curing reaction. For this purpose, core-shell particles from different suppliers were tested. The main characteristics of these particles are summarized in Table 2.3. A content of 5 wt.% of each of these products was added to the epoxy novolac resin DEN431 (Dow Chemicals). These samples were electron beam cured using 1 wt.% of the silver complex initiator Ag(1,7‑octadiene)1.5SbF6. Table 2.3: Characteristics of Core-Shell Particles Used as Tougheners for Electron Beam Curable Resins

Core-shell particle Albidur EP2240A Nanoresins

Rubber particle content (wt.%)

Rubber particle size (µm)

Base resin & EEW

40

0.1–3

DGEBA 300 g/eq

Albidur EP6220 Nanoresins

40

0.1–3

Novolac

Kaneka ACE MX120 (standard particles)

25

0.09

DGEBA 243 g/eq

Kaneka ACE MX156 (low T toughening)

25

0.09

DGEBA 243 g/eq

Wacker Genioperl P22

100

< 0.1

(none)

The effect of such core-shell particles on the electron beam curing process is evaluated by the degree of crosslinking and the glass transition temperature (Tg) of the modified resins. These properties are summarized in Table 2.4. It is observed that Genioperl P22 is the only modifier that has no effect on the degree of crosslinking of the cured resin. This is due to the fact that this material is a particulate modifier in powder form; that means it does not contain a base resin. All other core-shell rubbers are dispersed in an epoxy resin, which has an epoxy equivalent weight higher than the one of the novolac resin DEN431 (EEW = 176). It is known that the electron beam reactivity of an epoxy resin decreases rapidly by increasing the EEW of the resin. This explains the behavior observed, where systems containing Albidur particles (EP2240A and EP6220) have a degree of crosslinking of approx. 80%, while for systems cured in the presence of ACE particles (MX120 and MX156) this value decreases to approx. 60%. This can be explained by the fact that ACE products have a higher content of base resin (75 wt.%) in their formulations than Albidur products (60 wt.%). Typically, a linear dependence of glass transition temperature on the degree of crosslinking of the untoughened electron beam cured samples is expected. Quite clearly, this relationship is not entirely true for the case of toughness-modified systems, as other mechanisms affecting

50

2  Aspects of Electron Beam Curable Materials

the glass transition temperature are also being incorporated in the system. The best example of this behavior is observed for the resin modified with Genioperl P22. Although the degree of crosslinking remains practically as high as for the unmodified system, its glass transition temperature drops dramatically. This is due to the fact that the glass transition temperature is also dependent on the crosslinking density of the epoxy network, which can be roughly correlated to the storage modulus in the rubbery state (above Tg), as measured by DMA. In order to compare similar rubber contents in the system, the material containing 2 wt.% P22 should be compared to the one containing 5 wt.% EP2240A. In this case, again, although the crosslinking degree of the P22-modified system is higher than the one in the EP2240Amodified system, the glass transition temperature of the former is much lower. Figure 2.33 shows the DMA curves for unmodified system as well as for the modified epoxy systems containing 2 wt.% P22 and 5 wt.% EP2240A. Table 2.4: Properties of Core-Shell Particle Modified Epoxy Systems (DEN431 + 1% Ag(1,7-Octadiene)1.5SbF6)

Degree of crosslinking (%)

Tg (°C)

–

92.8 ± 4.0

190

5 wt.% Albidur EP2240A

83.8 ± 3.5

185

5 wt.% Albidur EP6220

75.7 ± 2.7

171

5 wt.% ACE MX120

65.1 ± 3.3

120

5 wt.% ACE MX 156

61.0 ± 0.6

135

5 wt.% Genioperl P22

91.6 ± 0.7

142

2 wt.% Genioperl P22

95.2 ± 0.8

155

Modifier (5 wt.%)

Figure 2.33: Dynamic mechanical analysis of unmodified and core-shell particle modified systems

2.3  Toughened Resins

51

0.9

KIc (MPa. m1/2)

0.8

5% P22 5% EP 2240

0.7 0.6

2% P22

5% MX120 5% MX156

0.5 0.4

No modifier

5% EP6220

0.3 0

0 2000

3000

4000

5000

Young's modulus (MPa)

Figure 2.34: Mechanical properties of the resin system DEN431 cured with 1% Ag(1,7‑octadiene)1.5SbF6 modified with rubber core-shell particles

It can be clearly seen that the addition of P22 has a significant effect on decreasing the crosslinking density of the epoxy network, leading to a lower glass transition temperature. However, this is not the case for the modifier EP2240A. The modification with core-shell particles is known to increase the fracture toughness of the system and the mechanical properties of the various resin systems, where the fracture toughness is plotted against the Young’s modulus, shown in Fig. 2.34. As expected, it can be seen that the incorporation of core-shell particles in the epoxy resin has a positive effect on the fracture toughness of the electron beam cured system. An increase from 2 wt.% to 5 wt.% P22 corresponds to an increase in fracture toughness of 10%. Quite interesting is the comparison of the system modified with 5 wt.% EP2240 to the system modified with 2 wt.% P22. Although the core-shell rubber content is the same in both cases and the degree of crosslinking is higher for the P22-modified system, its fracture toughness is higher than the one of the EP2240A-modified system. This difference is better understood by looking at the fracture surface of these two systems by scanning electron microscopy (SEM). Figure 2.35 shows the unmodified system, followed by the resin modified with 5 wt.% EP2240A, Fig. 2.36, and the system modified with 2 wt.% P22, Fig. 2.37. The fracture surface of the unmodified system, Fig. 2.35, is glassy, smooth and relatively featureless. Only striations and arrest lines, parallel to the direction of crack growth, are visible, typical for brittle fracture in epoxy materials. These structures result from the superposition of reflected stress waves during crack propagation, which create local compressive stresses causing crack branching.

52

2  Aspects of Electron Beam Curable Materials

(a)

(b)

Figure 2.35: Fracture surface of resin system DEN431 cured with 1% Ag(1,7-octadiene)1.5SbF6: (a) 200×, (b) 1000×

(a)

(a)

(b)

(b)

(c)

(c)

Figure 2.36: Fracture surface of resin system DEN431 cured with 1% Ag(1,7‑octadiene)1.5SbF6 modified with 5 wt.% EP2240A: (a) 200×, (b) 1000×, (c) 5000×

Figure 2.37: Fracture surface of resin system DEN431 cured with 1% Ag(1,7‑octadiene)1.5SbF6 modified with 2 wt.% P22: (a) 200×, (b) 1000×, (c) 5000×

2.3  Toughened Resins

53

Samples containing 5 wt.% of Albidur EP2240A exhibit a much rougher fracture surface, which correlates to the increase in fracture toughness. Looking at the higher magnitude of amplification, it is possible to identify that the core-shell particles are not individually distributed in the epoxy matrix; instead, small agglomerations have formed with dimensions of approx. a few microns. Outside these agglomerations, there is practically no plastic deformation of the resin, and the crack propagates in a quite brittle manner. Inside the agglomerations though, there is clear evidence of an interaction among the particles and the crack tip. Locally, there is an increase of fracture energy due to the cavitation of the rubbery particles, which promote void formation as observed on the fracture surface. Such voids lower the stress required for shear yielding and promote plastic deformation, as the direction of the crack front propagation is changed. However, it is also observed that single core-shell particles do not have the ability to generate this toughness mechanism; they only have an effect when working in agglomerates, because the shear deformation in the matrix around the particles cannot be formed due to the relatively high crosslinking density of the epoxy resin used here. A completely different picture is observed on the fracture surface of the epoxy resin modified with 2 wt.% P22. Here, the morphology reveals very large dispersed domains in the epoxy resin. These domains are actually agglomerates of the core-shell rubber particles. The poly(methyl methacrylate) shell of the Genioperl P22, other than EP2240A, does not contain reactive epoxy groups, which explains the better dispersion of EP2240A in the epoxy resin. Moreover, P22 is delivered in powder form, and this powder actually consists of agglomerates of core-shell particles, as shown in Fig. 2.38 [295]. The size of the single particles is below 100 nm and the agglomerates have sizes of approx. 100 µm. This corresponds to the morphology observed in the case presented here (Fig. 2.37). During the resin preparation, the shear forces created by mechanical stirring do not completely dissolve the agglomerates, which are also observed in the cured material. As the crack traverses these domains, evidence of large plastic deformation is observed, which increases the fracture energy of the resin system. A clear border between the rubber-like domains and the resin is observed, as seen by the differences of roughness of the fracture surfaces. Moreover, by comparing the continuous epoxy matrix of the P22-modified with the EP2240A-modified systems, a striking difference is observed, which is better seen in the 1000x amplified SEM pictures. While in the EP2240A-modified system the epoxy domains without particle agglomerates are very brittle, as indicated by the flat and featureless fracture surface, these domains in the P22-modified system show evidence of plastic deformation, with large occurrences of crack deviation. This is a consequence of the reduced crosslinking density, as observed by DMA curves, due to the very small single core-shell particles distri­ buted within the crosslinked network.

Figure 2.38: Ball-like structure of Genioperl P22

54

2  Aspects of Electron Beam Curable Materials

Table 2.5: Toughening Modification of Epoxy Novolac Resin (30% DEN438 + 70% DEN431) by 10% Core-Shell Particles. Initiator: [Ag(1,7-octadiene)1.5]SbF6

Initiator

Unmodified

Modified (10 wt.% EP 2240A)

Crosslinking (%)

KIc (MPa · m1/2)

Crosslinking (%)

KIc (MPa · m1/2)

0.5 wt.%

74.3 ± 1.5

0.48 ± 0.04

62.0 ± 2.9

0.56 ± 0.03

1.0 wt.%

90.8 ± 2.2

0.47 ± 0.04

73.0 ± 4.3

0.50 ± 0.01

1.5 wt.%

94.3 ± 0.2

0.45 ± 0.02

88.3 ± 3.3

0.48 ± 0.01

2.0 wt.%

95.1 ± 0.8

0.45 ± 0.06

89.7 ± 1.8

0.46 ± 0.03

Among all the core-shell particles studied here, the product EP2240A provides the best compromise between fracture toughness and glass transition temperature. This product has therefore been selected for further investigations. EP2240A was incorporated at 10 wt.% to a formulation of epoxy novolac resins (70% DEN431 + 30% DEN438, from Dow Chemicals) cured with the silver salt initiator [Ag(1,7‑octadiene)1.5]SbF6. Different amounts of initiator were employed in order to be able to correlate differently crosslinked networks with the toughenability of the core-shell rubber particles. The results are summarized in Table 2.5. The results show that the incorporation of core-shell rubbers leads to an increase in the fracture toughness of the system, as previously discussed. Moreover, this gain was also followed by a decrease in the degree of crosslinking of the system. This drop in the degree of crosslinking was compensated by increasing the amount of initiator in the resin from 0.5 up to 2 wt.%. However, at the same time, it is observed that this modification in the crosslinked network of the resin is also affecting the fracture toughness. For instance, in the case of the system containing 2 wt.% initiator, the toughness is reduced to a level similar to the unmodified resin system. This behavior can be associated with the fact that the crosslinking density of the material increases with the amount of initiator, and it can be clearly seen that core-shell particles are only effective as toughness modifiers when employed in combination with low crosslinking density resins. The incorporation of these particles in highly crosslinked materials, such as the resins described, has only a very limited effect of toughening. The fracture surfaces of these modified systems are shown in Fig. 2.39 (0.5 wt.% initiator), Fig. 2.40 (1 wt.% initiator), Fig. 2.41 (1.5 wt.% initiator), and Fig. 2.42 (2 wt.% initiator). The glass transition temperature of each system is also presented. The first conclusion drawn from the fracture surfaces of the modified materials concerns the non-occurrence of multiple localized shear-yielding in the matrix, which should cause an enhanced plastic deformation of the epoxy matrix surrounding the rubber particles. This is known to be one of the major energy dissipation mechanisms in rubber toughened epoxies [292]. However, in the case of electron beam cured resins presented here, this mechanism is suppressed due to the very high crosslinking density of the resin. The high crosslinking density is required for high‑Tg resins, and moreover, this density is particularly high for electron beam cured materials, due to the cationic homo-polymerization reaction.

2.3  Toughened Resins

Figure 2.39: Fracture surface of (70% DEN431 + 30% DEN438) + 10% EP2240A + 0.5 wt.% Ag(1,7‑octadiene)1.5SbF6: Tg: 141 °C; KIc: 0.56 ± 0.03 MPa · m1/2

Figure 2.40: Fracture surface of (70% DEN431 + 30% DEN438) + 10% EP2240A + 1.0 wt.% Ag(1,7‑octadiene)1.5SbF6: Tg: 171 °C; KIc: 0.50 ± 0.01 MPa · m1/2

Figure 2.41: Fracture surface of (70% DEN431 + 30% DEN438) + 10% EP2240A + 1.5 wt.% Ag(1,7‑octadiene)1.5SbF6: Tg: 191 °C; KIc: 0.48 ± 0.01 MPa · m1/2

Figure 2.42: Fracture surface of (70% DEN431 + 30% DEN438) + 10% EP2240A + 2.0 wt.% Ag(1,7‑octadiene)1.5SbF6: Tg: 192 °C; KIc: 0.46 ± 0.03 MPa · m1/2

55

56

2  Aspects of Electron Beam Curable Materials

The second major toughening mechanism involved in rubber toughening of epoxies regards the cavitation of the rubbery particles. For all initiator contents presented here, the fracture surfaces of the modified resin systems indicate that this effect is taking place. As the particles cavitate, voids are formed, which lower the stress required for shear yielding and promote local plastic deformation, as the direction of the crack front is changed. This effect also depends on the crosslinking density of the resin. With 0.5 wt.%, a relatively high number of unreacted epoxy groups in the network are still present. Thanks to the higher local mobility of these groups, the changes in the direction of the crack front are favored, promoting larger plastic deformation and is consequently increasing the toughness of the resin system. This is observed by comparing the homogeneously rough fracture surface of the resin containing 0.5 wt.% initiator to the brittle fracture surface of the resin cured with 2 wt.% initiator. These results demonstrate the intimate relationship between glass transition temperature and toughenability of the resin. A higher crosslinking density, and consequently glass transition temperature, decreases the toughenability of the core-shell particles. In other words, such modifiers are mainly suitable for toughening low‑Tg materials, where both toughening mechanisms, namely shear-yielding of the matrix and rubber cavitation, can take place.

2.3.3

Thermoplastics

In order to overcome some of the disadvantages of rubber toughening of thermosets, extensive studies on thermally curable epoxy resins have been devoted to high-temperature thermoplastic toughening employing poly(ethersulfone) (PES) [296–299], poly(sulfone) (PSU) [300–304], and poly(etherimide) [305–308]. A homogeneous blend is initially obtained for an epoxyamine system modified with a thermoplastic. With the advancement of the curing reaction, the solubility of the thermoplastic in the mixture decreases and may phase separate, as a result of a decrease in the entropic contribution to the free energy of mixing during polymerization [309, 310]. However, this is a competitive process as phase separation and polymerization reaction are taking place simultaneously. Therefore, the morphology development is determined by the ratio of phase separation rate and curing reaction rate. Moreover, the phase separation can be arrested by vitrification or gelation of either phase; thus, a homogeneous material is formed [311]. It has been found that phase separation in general leads to an improvement of the fracture toughness of the epoxy resin without deterioration of other mechanical properties and, more specifically, a co-continuous or phase inverted morphology is necessary to achieve a definite higher fracture toughness of the material [297, 304, 311–313]. A general overview of the different morphologies formed with the addition of a thermoplastic to a thermosetting system is presented in the Fig. 2.43. The toughening effect by addition of thermoplastics has in some cases been attributed to crack deflection caused by the modifier-rich phase [314] and also to ductile drawing of the thermoplastic and induced plastic deformation of the surrounding matrix [315]. However, Kinloch et al. [316] found that the fracture toughness increased continuously with thermoplastic concentration, even though the microstructure of the resulting material changed

2.3  Toughened Resins

57

Figure 2.43: General view of the different morphologies formed upon phase separation of a thermoplastic modified thermosetting system

from a particulate to a co-continuous and then to a phase-inverted morphology. Here, the toughening enhancement was attributed to crack deflection by the microstructure and the high toughness of the PES-rich phase. Lately, several mechanisms have been proposed to explain the toughness increase when adding a thermoplastic phase [315]. These mechanisms can eventually act simultaneously to produce the overall toughening effect. A summary of the different mechanisms is compiled in the Fig. 2.44. As previously mentioned, in the case of thermoplastic toughening of epoxy resins, these blends are typically prepared starting from a homogeneous solution of the modifier in the thermoset precursors (monomer) and phase separation is induced during network formation, as the Load Particle Plastic zone Crack growth direction

Tail behind the particle

Matrix

a) Crack bridging

b) Crack pinning / bowing

Multiple crack planes Shear banding Crack growth direction

c) Crack deflection / bifurcation

d) Shear banding

Figure 2.44: Major toughening mechanisms in thermoplastic-modified thermosets

58

2  Aspects of Electron Beam Curable Materials

molecular weight of the thermoset increases. This mechanism is designated polymerizationinduced phase separation (PIPS). PIPS depends on the interrelation among thermodynamics, competition between polymerization, and phase separation kinetics. It has been largely studied in conventional thermal curing of thermosets. However, in electron beam curing, where gelation and vitrification states are reached very quickly, the occurrence of the phase separation process is questionable. This aspect was investigated by employing various high-Tg thermoplastics as modifiers. The main characteristics of these modifiers are summarized in Table 2.6. A resin system based on novolac epoxy resins (70 wt.% DEN431 + 30 wt.% DEN438) and 1 wt.% of silver initiator Ag(1,7‑octadiene)1.5SbF6 was used for the investigations. Table 2.6: Commercial Thermoplastics Used as Toughness Modifiers

Thermoplastic

Tg (°C)

Molecular weight Mn (g/mol)

Molecular weight Mw (g/mol)

PSU Ultrason S2010

180

13 400

41 800

PSU Udel P1700 NT-11

185

–

30 000

PES Radel A-704 FP

220

16 000

36 000

PPSU Radel R5000NT

220

–

–

PEI Ultem 1000

215

12 000

30 000

The effect of incorporation of 10 wt.% of these thermoplastics on the glass transition temperature of the cured epoxy resin is shown in Table 2.7. In addition, the storage modulus in the rubbery state is also presented. It is known from the theory of rubber elasticity [317] that an increase of the storage modulus above Tg corresponds to an increase in the crosslinking density of the network. However, one should be aware that in the case of very high Tg thermo­ plastics, where their Tg is above the Tg of the thermoset network, the thermoplastic might also contribute to an increase of the storage modulus of the blend. Table 2.7: Dynamic Mechanical Properties as a Function of the Thermoplastic Modification of the Epoxy Resin (70 wt.% DEN431 + 30 wt.% DEN438) + 1% of Initiator Ag(1,7-Octadiene)1.5SbF6

Modifier (content: 10 wt.%)

Tg (°C)

Crosslinking (%)

Storage modulus in the rubbery state (· 10–7 Pa)

–

190

90.8 ± 2.2

7.9

PSU Ultrason

185

92.0 ± 1.1

6.5

PSU Udel

186

64.6 ± 2.5

7.9

PES Radel

190

90.2 ± 0.6

6.6

PPSU Radel

186

79.4 ± 2.7

6.7

PEI Ultem

173

63.2 ± 4.6

9.9

2.3  Toughened Resins

59

From the results in Table 2.7 it can be seen that some of the thermoplastics (PSU Udel, PPSU Radel, and PEI Ultem) significantly decrease the degree of crosslinking of the epoxy resin. This aspect cannot be related to the chemical structure of the thermoplastic, as for instance, the PSU Ultrason has led to higher crosslinking than PSU Udel, although both have the same chemical structure. Therefore, this negative effect on the electron beam curing process can only be related to the presence of impurities from the proprietary manufacturing process of the modifiers, which inhibit the cationic curing. Nevertheless, it is remarkable to observe that, although in some cases the degree of crosslinking of the cured material is much lower than the reference system, the glass transition temperature is practically not affected, due to the high glass transition temperature of the thermoplastics employed. This result indicates that the systems investigated here have formed a homogeneous morphology. In respect to the effect of the thermoplastic modification on the mechanical properties of the epoxy resins, the results regarding fracture toughness and Young’s modulus are shown in Fig. 2.45. 0.9

0.5 K Ic (MPa.m )

0.8 0.7

(70% DEN431 + 30% DEN438) 1% Ag(1,7-Octadiene)1.5SbF6

10% PSU Ultrason

0.6 10% PPSU Radel 0.5 10% PSU Udel

10% PEI Ultem 10% PES Radel

No modifier

0.4 0.3 0

0 2000

3000

4000

5000

Young's modulus (MPa)

Figure 2.45: Effect of thermoplastic modification on mechanical properties

Figure 2.45 indicates that the thermoplastic modification generally leads to an improvement of the fracture toughness of the electron beam cured resins, with particular enhanced properties for the resins modified with PSU Ultrason, PEI Ultem, and PES Radel. The mechanical properties of thermoplastic modified epoxy resins are typically related to the phase separation process and the final morphology of the system [297, 311]. In order to evaluate the presence of phase separation, dynamic mechanical analysis (DMA) is a powerful tool. A two-component heterogeneous blend system, which in our case means phase separated morphology, presents two distinct glass transition temperatures, one for each phase. As a consequence, the two Tg’s are seen as two peaks on the tan  curve of the DMA. In the case of a homogeneous system, only one glass transition temperature is observed, which has an intermediary value between those from the single components. The DMA curves for the different thermoplastic/thermoset blends are shown in Fig. 2.46 (a) to (f).

60

2  Aspects of Electron Beam Curable Materials

10% PES Radel

9

10

8

10

7

0.2 0.1

50

(a)

100 150 200 Temperature (°C)

1010 10

9

0.2

108 10

7

0.1

50

250

(b)

10

10

9

108

0.2

107

0.1

50

(c)

100 150 200 Temperature (°C)

10

10

10

9

10

8

0.2

107

250

0.1

50

(d)

10 9

10

8

10

7

0.2 0.1

50

(e)

100 150 200 Temperature (°C)

10

10

10

9

10

8

10

7

0.2 0.1

50

250

(f )

100 150 200 Temperature (°C)

Figure 2.46: 70% DEN431 + 30% DEN438 + 1% Ag(1,7-Octadiene)1.5SbF6 + 10% Modifier: (a) No modifier, (b) PES Radel, (c) PSU Ultrason, (d) PSU Udel, (e) PPSU Radel and (f ) PEI Ultem

250

Tan delta

10

Storage modulus G' (Pa)

10

250

10% PEI Ultem

Tan delta

Storage modulus G' (Pa)

10% PPSU Radel

100 150 200 Temperature (°C)

Tan delta

Storage modulus G' (Pa)

10

250

10% PSU Udel

Tan delta

Storage modulus G' (Pa)

10% PSU Ultrason

100 150 200 Temperature (°C)

Tan delta

10

Storage modulus G' (Pa)

1010

Tan delta

Storage modulus G' (Pa)

Unmodified system

2.3  Toughened Resins

61

In Fig. 2.46 (c) and (f), which correspond to the systems containing PSU Ultrason and PEI Ultem, respectively, it is possible to observe “shoulders” on the tan  curves. This indicates the presence of two different glass transition temperatures in the systems, and the formation of phase separated morphologies at some level. In order to further investigate these phase separated morphologies, the fracture surface of the epoxy resin modified with 10 wt.% PSU Ultrason was analyzed by scanning electron microscopy, presented in Fig. 2.47 for three different magnifications. First, we focus to the fracture mechanics of these systems. The pictures show a rougher surface than the typical featureless fracture surface of brittle unmodified epoxies (e.g., Fig. 2.35). Furthermore, the plastic deformation of the PSU modified system during crack propagation is taking place at a very small size scale. The main toughening mechanisms observed for this system are crack deflection and bifurcation, indicating that the thermoplastic modifier act on the crack front, deviating it from its principal plane of propagation and causing a crack to split into several secondary cracks. However, in terms of the morphology, even at the maximum amplification obtained by SEM pictures (× 5000), it is not possible to observe thermoplastic phase separated domains, as indicated by the DMA curves. Transmission electron microscopy was therefore employed, as shown in Fig. 2.48.

(a)

(b)

(c)

Figure 2.47: SEM images from fracture surface of epoxy resin containing 10 wt.% PSU Ultrason: (a) × 200, (b) × 1000, (c) × 5000

Figure 2.48: TEM picture of an epoxy resin modified with 10% PSU, stained with RuO4.

62

2  Aspects of Electron Beam Curable Materials

In Fig. 2.48, the dark grey areas correspond to the stained component, which in this case corresponds to PSU, and the light grey correspond to the epoxy matrix. It can therefore be seen that for this electron beam curable resin, there is also phase separation of the modifier. Typically, in thermally curable resins, thermoplastic phase separated domains have a particlelike geometry, with a diameter in the range of a few microns [318]. However, due to the very fast reaction in electron beam curing, although phase separation is also observed, the domains do not have enough time to agglomerate and form particles. Micelle-like domains are then formed, with a size in the nano-scale range, confirming the results regarding phase separation suggested by the DMA curves. Subsequently, after this first screening of engineering thermoplastics suitable as toughness modifiers for electron beam curing, PES Radel and PSU Ultrason were used for further investigations. These modifiers have been employed at different amounts in the novolac epoxy resin system (70 wt.% DEN431 + 30 wt.% DEN438), cured with 1.5 wt.% of silver initiator Ag(1,7-octadiene)1.5SbF6. The effect of these modifiers on the fracture toughness and glass transition temperature of the electron beam cured epoxies is seen in Fig. 2.49. It is interesting to compare the results from the resins containing 10 wt.% PES Radel cured with 1 wt.% silver initiator (Fig. 2.45) to those cured with 1.5 wt.% silver initiator (Fig. 2.49). The resin system cured with 1 wt.% initiator has a KIc equal to 0.59 MPa · m1/2 and a Tg of 190 °C, while that cured with 1.5 wt.% initiator has a KIc equal to 0.48 MPa · m1/2 and a Tg of 205 °C. This lower fracture toughness of the resin system cured with a higher amount of initiator is explained by the fact that the crosslinking density of the network is higher, which leads to a less toughenable network due to a lower local mobility of the segments of the crosslinked network. A similar behavior has already been observed and discussed in the section regarding the toughness modification with rubber core-shell particles. Regarding the comparison among the systems cured with 1.5 wt.% (Fig. 2.49), it is observed that a high amount of PES, namely 15 wt.%, is necessary in order to obtain a significant toughening 0.9 PSU Ultrason PES Radel

0.5 K Ic (MPa. m )

0.8

181 °C

0.7 0.6

175 °C

182 °C

0.5

210 °C 200 °C

0.4

192 °C

205 °C

5

10

0.3 00

0

15

Thermoplastic content (%)

Figure 2.49: Fracture toughness and Tg as a function of modifier content

20

2.3  Toughened Resins

63

effect of the electron beam cured resin. Moreover, it is important to note that this gain in fracture toughness takes place together with an increase of the glass transition temperature. This remarkable behavior is observed due to the higher glass transition temperature and fracture toughness of the PES in comparison to the crosslinked epoxy network. Concerning the modification of the epoxy system with PSU, a positive effect in toughness is already observed for a modifier content of 10 wt.%, with a significant increase at 15 wt.% PSU. Compared to PES, the modification with PSU leads to a higher improvement in fracture toughness, due to its higher toughness. However, PSU has also a lower glass transition temperature than PES, leading therefore to lower values of glass transition temperatures for the modified epoxy resins. As previously mentioned, the effect of toughness modification of epoxy resins by thermoplastics is strongly related to the morphology of these blends, and phase separated domains are known to significantly improve the toughness of thermosets [312, 313]. Regarding electron beam curable resins, although phase separation has been observed, the thermoplastic-rich domains were extremely small, and had only a limited effect on toughening the system. A different approach was therefore proposed, where preformed thermoplastic particles were dispersed in the resin and the system was electron beam cured without the previous dissolution step of these particles (it is important to mention here that the dissolution step of thermoplastics in epoxy is typically based on slowly adding the thermoplastic to the resin and mixing with a mechanical stirrer at approx. 110–130 °C for at least 45 min at 600–1000 rpm). As electron beam curing is an extremely fast process, no dissolution of the particles during curing can be expected when the dissolution step was omitted. PES particles with an average diameter of 200 µm were therefore used in order to compare these two toughening methods, namely dispersion of preformed thermoplastic particles in the resin at room temperature versus dissolution of the thermoplastic in the thermoset precursors prior to curing. The epoxy resin system (30% DEN438 + 70% DER331) was cured with 1.5 wt.% silver initiator Ag(1,7‑octadiene)1.5SbF6. Table 2.8 shows the fracture toughness and glass transition temperatures of these systems as a function of the amount of PES dispersed or dissolved. It is important to note that all modified systems have a constant total amount of 15 wt.% PES. Table 2.8: Effect on the Properties of Electron Beam Cured Epoxy Modified by Dispersed PES Radel Particles and/or Dissolved PES Radel. Resin System: (30% DEN438 + 70% DER331) + 1.5 wt.% Ag(1,7-octadiene)1.5SbF6

Dissolved PES*

Dispersed PES**

Tg (°C)

Crosslinking (%)

KIc (MPa · m1/2)

Modifier –

–

198

97.0 ± 0.3

0.48 ± 0.02

15 wt.%

–

186

86.7 ± 2.5

0.53 ± 0.02

7.5 wt.%

7.5 wt.%

182

90.3 ± 1.5

0.59 ± 0.05

–

15 wt.%

195

92.1 ± 0.4

0.70 ± 0.02

* PES was dissolved in the resin at 130 °C. After cooling down the modified resin, initiator was added ** Dipersion of particles (average diameter 200 µm) at room temperature

64

2  Aspects of Electron Beam Curable Materials

The first conclusion from the results summarized in Table 2.8 is that the dissolution of PES in the resin leads to a slightly lower degree of crosslinking of the cured system, due to dilution effects. For the case of the modification with PES particles, this effect is not as important. As a consequence, the glass transition temperatures of the resins modified with particles tend to be higher than those modified with dissolved PES. Yet, the most striking aspect regarding the results from Table 2.8 concerns the fracture toughness of the resins. As expected, and as already shown and discussed, the incorporation of 15 wt.% of dissolved PES leads to an increase in the toughness of the resins. In addition, there is clearly a significant enhancement for the resins modified with PES particles in comparison to dissolved PES, leading to a higher fracture toughness combined with a higher glass transition temperature. In order to understand the fracture mechanisms that are leading to the higher fracture toughness, the fracture surfaces of these materials were investigated by scanning electron microscopy, presented in Fig. 2.50 (15 wt.% dissolved PES), Fig. 2.51 (7.5 wt.% dissolved PES and 7.5 wt.% dispersed PES), and Fig. 2.52 (15 wt.% dispersed PES). The fracture surface of the resin system modified with 15 wt.% dissolved PES, Fig. 2.50, does not indicate the occurrence of phase separated PES-rich domains in the micrometer range. On the other hand, the PES particles that were simply dispersed in the resin, Figs. 2.51 and 2.52, are still present, indicating that during electron beam curing, although high temperatures may be reached, the period of time at high temperatures is not long enough to dissolve the particles during the curing process. In terms of toughening mechanisms, in the resin system containing uniquely dissolved PES, Fig. 2.50, the roughness of the surface indicates the occurrence of plastic deformation out of the plane of the crack propagation, explaining the higher fracture toughness in comparison to the unmodified system. Moreover, the main toughening mechanisms observed for this system are crack deflection and bifurcation, indicating that the thermoplastic modifier acts on the crack front, deviating it from its principal plane of propagation and causing a crack to split into several secondary cracks. In the case of the epoxy resin modified uniquely with dispersed PES particles, Fig. 2.52, the features of the fracture surface are completely different. There are clearly two distinct domains, a continuous epoxy matrix and dispersed PES particles. Regarding the continuous epoxy matrix, a very brittle fracture is observed, with a featureless surface, suggesting that thermoplastic chains are not present in the epoxy matrix. On the other hand, during crack propagation, the thermoplastic particles show very large plastic deformation, with evidences of crazing as well. Moreover, crack pinning is also observed. These various fracture mechanisms observed on the PES particles are taking place due to the strong adhesion between the particles and the epoxy matrix. Finally, regarding the epoxy system modified with 7.5 wt.% dissolved PES and 7.5 wt.% dispersed PES particles, Fig. 2.51, a combination of the above mentioned mechanisms is observed. The continuous epoxy phase, which is modified with the dissolved PES, shows evidence of crack deviation, crack deflection, and crack pinning, whereas the dispersed PES particles show a strong adhesion to epoxy and undergo large plastic deformation and crazing during crack propagation.

2.3  Toughened Resins

(a)

65

(b)

Figure 2.50: Fracture surface of the epoxy system (30% DEN438 + 70% DER331) cured with 1.5 wt.% Ag(1,7-octadiene)1.5SbF6, and modified with 15 wt.% dissolved PES: (a) 200×, (b) 1000×

(a)

(b)

Figure 2.51: Fracture surface of the epoxy system (30% DEN438 + 70% DER331) cured with 1.5 wt.% Ag(1,7-octadiene)1.5SbF6, and modified with 7.5 wt.% dissolved PES + 7.5 wt.% dispersed PES: (a) 200×, (b) 1000×

(a)

(b)

Figure 2.52: Fracture surface of the epoxy system (30% DEN438 + 70% DER331) cured with 1.5 wt.% Ag(1,7-octadiene)1.5SbF6, and modified with 15 wt.% dispersed PES: (a) 200×, (b) 1000×

In conclusion, comparing the fracture toughness among the three modified epoxy resin systems, it is observed that the fracture mechanisms related to large plastic deformation of the thermoplastic particles require more energy than energy dissipation mechanisms related to the epoxy matrix, such as crack deviation and crack deflection. A further aspect of major interest to be investigated regards the toughening effect of the dispersed PES as a function of the diameter of the particles. Figure 2.53 shows the glass transition temperature and the fracture toughness of systems modified with 15 wt.% PES with different particle sizes. It is important to note that particle sizes equal zero correspond to the PES-modified system in which the modifier was completely dissolved prior to curing.

66

2  Aspects of Electron Beam Curable Materials

0.5 K Ic (MPa.m )

0.8

Tg = 193 °C

0.6

0.4

Tg = 195 °C

Tg = 186 °C

0.2

0

0

50

100

150

200

250

Average size of PES particles (Microns)

Figure 2.53: Effect of particle size on toughening of epoxy resin (30% DEN438 + 70% DER331) cured with 1.5 wt.% Ag(1,7-octadiene)1.5SbF6. All systems containing 15 wt.% PES Radel. (* Particle size equal zero means that the PES was completely dissolved in resin prior to curing)

The trend observed in the last years regarding nano-composites has been rapidly growing due to the fact that smaller particles increase the surface area in contact between the components, which has the ability to boost properties. However, the results observed in Fig. 2.53 clearly show an opposite trend, where increasing the particle size of the modifier causes a further enhancement in fracture toughness of the resin system. The mechanisms involved in toughening by PES dissolved in epoxy (particle size equal zero), namely crack deviation and crack deflection, have already been presented and discussed (Fig. 2.50). The same was done for the resins modified with a particle size of 200 µm (Fig. 2.51), where large plastic deformation of the particles could be observed, causing enhanced fracture toughness. Finally, regarding the system modified with particles with an average size of 50 µm, the fracture surface is presented in Fig. 2.54. Clearly, the morphology shows a two-phase system, where PES particles are homogeneously dispersed in a continuous epoxy matrix. This epoxy phase, similarly to the resin system modified with 200 µm particles, is very brittle and shows a featureless surface, except for the crack pinning mechanism due to the presence of particles. Regarding the PES particles themselves, there is also evidence of strong adhesion to the epoxy matrix, as pull-out of particles is not observed. Moreover, as the crack propagates through the particles, plastic deformation takes place. However, comparing the 200 µm-modified system (Fig. 2.52) to the 50 µm-modified system (Fig. 2.54) there are some clear differences on the fracture mechanics of the particles.

2.3  Toughened Resins

(a)

67

(b)

Figure 2.54: Fracture surface of epoxy system (30% DEN438 + 70% DER331) cured with 1.5 wt.% Ag(1,7-octadiene)1.5SbF6, and modified with 15 wt.% dispersed PES (50 µm): (a) 200×, (b) 1000×

The fracture surface of the larger particles (200 µm, in Fig. 2.52) reveals the presence of voids, which were generated during propagation of the crack. During the formation of these microvoids, new surfaces are generated, therefore increasing the energy necessary to fracture these particles. Quite interestingly, the occurrence of microvoids is only observed for particles having a diameter larger than 100 µm, explaining the higher fracture toughness values for the epoxy resins modified with the 200 µm particles in comparison to those modified with 50 µm particles.

2.3.4

Block Co-polymers

Lately, the modification of brittle polymers with block co-polymers has received considerable attention [319], due to their ability to self-assemble into a variety of ordered nanoscale morphologies. These microphase-separated structures result from the competition between chain connectivity and block immiscibility. Early work [320] on block co-polymers dissolved in reactive solvents such as styrene has been revalorized recently using thermoset precursors such epoxy [321, 323] and phenolic [324] systems. In contrast to classical linear thermoplastics, macroscopic phase separation can be avoided with block co-polymers. In thermally curable resin systems, it was demonstrated that amphiphilic di-block co-polymers, where one block is miscible with the thermoset, are not only dispersible in a reactive system, but also able to order themselves on the nanometer scale in both the unreacted and reacted states [321–324]. Blends of the tri-block co-polymer polystyrene-block-polybutadiene-block-polymethylmethacrylate (SBM) with epoxy resins have been reported in the literature [325–327]. SBM co-polymer was selected, knowing that yy the homopolybutadiene (PB) is immiscible with the epoxy precursor, leading to an initial macrophase separation between PB rubber and epoxy monomers, yy the homopolystyrene (PS) is initially partially miscible with the epoxy precursors (a homogeneous solution is obtained at temperature higher than 90 °C, whereas decreasing temperature involves phase separation between PS and epoxy monomers), and yy the homopoly(methyl methacrylate) (PMMA) is completely miscible with epoxy precursors.

68

2  Aspects of Electron Beam Curable Materials

Depending on the content of the SBM in thermally curable epoxy, distinct structures were obtained. For low concentrations, a micellar structure, formed by both PS and PB blocks, was generated as consequence of the immiscibility of those blocks with the epoxy precursors. The favorable interaction of PMMA with epoxy ensures the structuration of the blend at a nanometer range. Increasing the block co-polymer concentration from 10 to 50 wt.% induced a change to either a more complex “spheres on spheres” (raspberry) structure, with PB nodules localized around PS spheres or a core-shell (onion) morphology composed of PS spheres surrounded by PB shells. The addition of SBM was found to be a powerful way to increase the toughness of epoxy networks, which linearly increased with the PB content in the tri-block, although leading to a small decrease of the Young’s modulus of the epoxy resin system. Similarly, the tri-block copolymer polymethylmethacrylate-block-polybutylacrylate-blockpolymethylmethacrylate (MAM), which has two completely miscible groups with the epoxy precursors, has been investigated as toughness modifier for low crosslinking density thermally curable epoxy resins. Thanks to the miscibility of the external blocks on epoxy, the soft central block has shown an indirect high adhesion to the crosslinked epoxy network. The morphology obtained for such systems consisted of flocculated particles that lead to worm-like micelles by increasing the content of the block co-polymer. MAM was shown to effectively toughen low crosslink density epoxy resins (DGEBA) [328]. The toughening mechanisms identified included rubbery phase cavitation as well as matrix deformation. Results concerning high crosslinking density thermally curable resins (tetra­ functional resin) modified with MAM were also reported [318]. Besides the impressive increase of fracture toughness of the system, the glass transition temperature of the material as well as Young’s modulus were not affected. The morphologies observed were similar to those described for the low crosslinking density resins. Rubbery cavitation as well as crack pinning and crack deviation were identified as fracture mechanisms, as seen in Fig. 2.55 for an epoxy system containing 5 wt.% MAM. Electron beam curable systems modified with the block copolymer MAM M22, from Arkema have also been investigated in the literature [329]. This modifier was used at different contents, namely 7.5 wt.% and 10 wt.%, in an epoxy novolac resin system (70% DEN431 + 30% DEN438) and electron beam cured using 1.5 wt.% of a silver initiator Ag(1,7-octadiene)1.5SbF6. Prior to curing, the modifier was completely dissolved in the epoxy resin. The results regarding the glass transition temperature and degree of crosslinking of the unmodified and modified epoxy resins are shown in Table 2.9.

Figure 2.55: SEM picture of the fracture surface of the epoxy system RTM6® modified with 5 wt.% MAM

2.3  Toughened Resins

69

Table 2.9: Effect of MAM Modification of the Epoxy Resin (70 wt.% DEN431 + 30 wt.% DEN438) + 1.5 wt.% of Initiator Ag(1,7-octadiene)1.5SbF6 on the Electron Beam Reactivity of the System

Modifier

Tg (°C)

Degree of crosslinking (%)

–

199

95.1 ± 0.5

7.5 wt.% MAM

83 and 201

95.1 ± 0.4

10 wt.% MAM

87 and 202

95.3 ± 0.8

The results indicate that the modifier MAM does not have a negative effect on the electron beam cationic curing reaction, as the degree of crosslinking of the epoxy network remains practically identical. Regarding the glass transition temperature of the resin system, an interesting behavior is observed. Instead of a single peak on the tan  curve, two peaks can be observed, indicating that due to the modification with MAM, the resin system has two distinct glass transition temperatures. Figure 2.56 shows the DMA curve for the resin system modified with 10 wt.% MAM. Starting at very low temperatures, one can also observe a peak around –80 °C, which is related to the crankshaft movement of the epoxy chains present in all epoxy networks, independently of the modification of the system. Moving to higher temperatures, the DMA curve clearly indicates that a two-phases morphology has been formed in these MAM-modified systems. Starting with the peak at higher temperatures, at approx. 200 °C, this transition corresponds to the glass transition temperature of the crosslinked epoxy network. This Tg is approximately the same as the one for the non-modified epoxy resin, indicating that there should be very little or no MAM chains dispersed in the epoxy network, which would decrease the crosslinking density, acting as a plasticizer, and therefore decreasing the glass transition temperature.

G' (Pa)

10

10

9

10

8

10

7

G' (Pa)

0.2

0.1

Tan delta

Tan delta

10

0 -100

-50

0

50

100

150

200

250

300

Temperature (°C) Figure 2.56: Dynamic mechanical behavior of epoxy novolac resin (70% DEN431 + 30% DEN438) + 1.5 wt.% of initiator Ag(1,7-octadiene)1.5SbF6 modified with 10 wt.% MAM

70

2  Aspects of Electron Beam Curable Materials

0.9

K Ic (MPa.m 1/2)

0.8 0.7

10 wt% MAM

0.6

7.5 wt% MAM

0.5 0.4

Unmodified

0.3 0

0 2000

3000

4000

5000

Young's modulus (MPa)

Figure 2.57: Mechanical properties of epoxy novolac resin (70% DEN431 + 30% DEN438) + 1.5 wt.% of initiator Ag(1,7-octadiene)1.5SbF6 modified with MAM

Regarding the transition at approximately 90 °C, this peak indicates the presence of MAM-rich domains in the system. It is known that MAM has the ability to self-assemble in a variety of structures, and thereby also leading to phase separation among the PMMA and PBA blocks. When the blocks exhibit phase separation, two distinct glass transition temperatures would be expected, approx. 110 °C for PMMA and –60 °C for PBA, and consequently, two distinct peaks are registered in the tan  curve. The behavior observed here however, with a single peak for the MAM-rich phase, indicates the formation of a homogeneous domain, where the PMMA and PBA blocks are not completely separated from each other. The effect of MAM modification of the epoxy resin with respect to the mechanical properties is shown in Fig. 2.57, where fracture toughness vs. Young’s modulus is presented. Figure 2.57 shows that the incorporation of the block co-polymer MAM causes a loss of approx. 10% in Young’s modulus. At the same time, it is also observed that the incorporation of MAM brings a significant increase to the fracture toughness of the epoxy resin. In order to evaluate the influence of MAM incorporation on the fracture mechanisms of the epoxy resin, the fracture surfaces of the epoxy resin modified with 10 wt.% MAM were investigated by scanning electron microscopy. The micrographs are shown in Fig. 2.58. The surface of the modified material clearly shows a morphology based on two phases, confirming the results from the DMA curves, where the MAM-rich domains (light grey) are dispersed in the continuous epoxy matrix (dark grey). Regarding the fracture mechanics of this material, it is observed that the epoxy matrix undergoes a brittle fracture. This suggests that practically none of the MAM is dispersed outside the phase-separated domains. Quite interestingly, around the MAM-rich domains, for a distance of approximately 10 to 20 µm, the epoxy matrix exhibits a different fracture mechanism, where “fracture waves” are observed, in comparison to the featureless fracture surface at the bulk epoxy. These striations, parallel to the direction of crack growth, result from the superposition of reflected stress waves, which

2.3  Toughened Resins

71

Second phase (low Tg phase) (a)

(b)

(c)

Figure 2.58: Fracture surface of the epoxy novolac resin (70% DEN431 + 30% DEN438) + 1.5 wt.% of initiator Ag(1,7-octadiene)1.5SbF6 modified with 10 wt.% MAM: (a) 200×, (b) 1000× (c) 5000×

create local compressive stresses. The presence of these structures indicates that the stress state of the resin is modified due to the MAM-rich phase-separated domains. The rough surface of the MAM-rich phase confirms the large plastic deformation of these domains, which contribute to an increase of fracture energy of the system. A well-defined border between the particles and the epoxy matrix is not observed. Moreover, by combining this information with the fact that the domains are not pulled out from the epoxy matrix during crack propagation, a very strong adhesion between the two phases is confirmed, as expected, due to the compatibility of the PMMA blocks with epoxy. In order to observe the structure of these MAM-rich domains in more detail, a sample containing 10 wt.% MAM was investigated by transmission electron microscopy, shown in Fig. 2.59. The TEM picture shows a MAM-rich domain dispersed in the epoxy matrix. A closer look reveals a complex structure based on a phase-inverted morphology, in which the continuous phase is MAM with some islands of dispersed epoxy. In general, the MAM domain does not show an organized morphology, as observed by the many different color contrasts. A darker contrast corresponds to the regions richer in PBA and a lighter contrast to PMMA. Only in some punctual regions there is an indication of nano-ordered sub-morphologies, where the

72

2  Aspects of Electron Beam Curable Materials

Figure 2.59: TEM image of epoxy novolac resin (70% DEN431 + 30% DEN438) + 1.5 wt.% of initiator Ag(1,7-octadiene)1.5SbF6 modified with 10 wt.% MAM. Staining agent: RuO4

PBA formed sphere-like structures, observed as black spots. Possibly due to the fast curing reaction by electron beam, these structures do not have enough time to further assemble themselves.

2.3.5

Inorganic Nanoparticles

Some inorganic particulate materials have been considered as toughening agents for thermally curable epoxy materials. In the literature, glass beads [330, 331], alumina trihydrate [332] and silica [333, 334] (diameter in the range of 4–100 µm) were used to improve the toughness of the filled materials. Moreover, such particles also have the ability to improve Young’s modulus without reducing the glass transition temperature of the material. More recently, a new technology has emerged offering promise to increase the mechanical performance of such thermosetting polymers via the addition of a nanophase structure in the polymer, where the nanophase consists of small rigid silica particles [335–338]. The modification with such nanoparticles has shown to increase the fracture toughness of the thermally curable epoxy system. Moreover, due to the very small size of the silica particles, the viscosity of the epoxy resin is practically not affected, assuring ease of the processing. The crack-pinning mechanism, first described by Lange [339], is very often used to explain particle toughening. As shown in Fig. 2.60, the propagating crack front is slowed down and

2.3  Toughened Resins

73

Crack front propagation

Pinning

Bowing

Break-away

Figure 2.60: Scheme of the crack pinning mechanism

has to bow when meeting a line of particles. Thereby, the fracture surface is increased and the fracture energy is distributed over a wider area. Upon break-away of the crack front, the separated crack fronts unite again. Because of slight differences in the heights of the single crack fronts, small ridges are often formed behind the particles. These ridges, often referred to as tails, can be easily detected by observation of the crack surface. More recently, different toughening mechanisms for these materials could be identified, depending on the size of the particle (schematically represented in Fig. 2.61) [340, 341]. For most of the silica particles, the crack propagates by: (a) deviation in matrix; (b) broken silica particles; (c) debonding at the matrix/silica interface; and (d) combination of the above mentioned mechanisms. Therefore, with increasing silica particle size, the ratio of (b) increases and (a) becomes larger. Fracture toughness can be improved with this mechanism. Meanwhile, fine silica particles cannot be broken by cracking, and the crack propagates in the (d) mode mixed with (a) and (c). Large SiO2 particles (b)

(a)

Crack propagation path (c)

(a)

(d)

(c)

Small SiO2 particles

Figure 2.61: Schematic illustration of crack propagation model: (a) Crack deflection at large SiO2 particles; (b) Crack passing through large SiO2 particles; (c) Crack propagation with debonding at the large SiO2 particles/matrix interface; (d) Crack deflection or debonding at the small SiO2 particles/matrix interface

74

2  Aspects of Electron Beam Curable Materials

In nanocomposites, as the silica particles are so much smaller than the crack-opening displacement, it is unlikely that the on-plane mechanisms (crack pinning and crack deflection) are responsible for the complete toughening effect observed for epoxy resins. Scanning electron and atomic force microscopy tests of the fracture surfaces revealed nanoparticles surrounded by voids, providing evidence of debonding of the nanoparticles and subsequent plastic void growth, which also significantly contribute to the increase in toughness [342, 343]. Regarding electron beam curable resins, studies on the influence of nano-silica particles on the curing reaction and mechanical properties of novolac epoxy resins have been carried out. Two grades of nano-silica particle products provided by the company Nanoresins have been investigated. The first one, Nanopox A610, contained 40 wt.% silica in a cyclo-aliphatic epoxy resin (EEC), while the second product, Nanopox A660, contained 50 wt.% silica in a reactive epoxy diluent (TMPO). In addition, the average size of the silica particles in both cases was 20 nm. These products were incorporated into two novolac epoxy resins, DEN431 and DEN438, and cured with 1 wt.% of silver salt initiator Ag(1,7-octadiene)1.5SbF6. The glass transition temperature (Tg) of these materials is summarized in Table 2.10. Table 2.10: Glass transition Temperature Of Nano-Silica Modified Systems

Resin DEN431

DEN438

Modifier

Crosslinking degree (%)

Tg (°C)

–

92.8 ± 4.0

190

5 wt.% Nanopox A610

68.4 ± 5.7

174

5 wt.% Nanopox A660

73.4 ± 0.9

176

–

95.6 ± 0.8

204

5 wt.% Nanopox A610

72.9 ± 0.8

181

5 wt.% Nanopox A660

72.5 ± 1.5

186

The novolac epoxy monomer of the resin DEN438 has a higher functionality than the one of DEN431, which explains the higher glass transition temperature of the former, due to a higher crosslinking density. Nevertheless, in both resin systems, the same trend is observed when Nanopox is incorporated: the degree of crosslinking and the glass transition temperature significantly decrease. This result cannot be justified by the presence of the base resin of A610 (EEC) and A660 (TMPO) alone. Due to the very small amount in the formulation, approximately 3 wt.% EEC for A610 and 2.5 wt.% TMPO for A660, it is improbable that such a negative effect was caused by the lower reactivity of these components alone. Moreover, these two base resins are quite different from each other; nonetheless, the observed effect is similar. Therefore, these aspects suggest that the negative effect on reactivity is possibly related to the presence of impurities from the proprietary manufacturing of these nano-silica based products, which inhibit the electron beam cationic curing reaction. As already mentioned, comparing the two nano-silica products, it was observed that the incorporation of 5 wt.% resulted in similar values in the degree of crosslinking. On the other hand, one can identify a trend showing higher glass transition temperatures for the systems

2.3  Toughened Resins

75

modified with A660 in comparison to A610. In the literature, it is known that nano-silica particles have the ability to increase the glass transition temperature of epoxy resins [335], acting as rigid anchoring points, and in this case, due to the higher amount of nano-silica in A660, this effect becomes obvious. To illustrate the effect of incorporated nano-silica particles on the mechanical properties of the epoxy resins, the fracture toughness and Young’s modulus of the systems are shown in Fig. 2.62 (DEN431) and Fig. 2.63 (DEN438). In both cases, the addition of nano-particles causes a positive increase in Young’s modulus of the material. This was expected, due to the fact that silica particles have a much higher Young’s modulus than epoxy and this property typically follows a rule of mixture. 0.9

DEN431 + 1% (Ag(1,7-Octadiene) SbF ) 1.5

K Ic (MPa. m

0.5

)

0.8

6

0.7

5 wt% A660

0.6 0.5 0.4

No modifier

5 wt% A610

0.3 0

0 2000

3000

4000

5000

Young's modulus (MPa)

Figure 2.62: Effect of nano-particles on mechanical properties of resin DEN431 0.9 DEN438 + 1% (Ag(1,7-Octadiene) SbF ) 1.5

K Ic (MPa.m

0.5

)

0.8

6

0.7 0.6 0.5

5 wt% A610 0.4

5 wt% A660

No modifier

0.3 0 0 2000

3000

4000

Young's modulus (MPa)

5000

Figure 2.63: Effect of nano-particles on mechanical properties of resin DEN438

76

2  Aspects of Electron Beam Curable Materials

Figure 2.64: Fracture surface of the resin formulation DEN438 + 5 wt.% A610 cured with 1 wt.% Ag(1,7-octadiene)1.5]SbF6

Incorporation of nano-silica particles in the resin results in an improvement in fracture toughness. Figure 2.64 shows the fracture surface of the resin DEN438 modified with 5 wt.% A610. Agglomerates of silica particles are not observed, showing a very good dispersion of the nanoparticles in the resin. On the fracture surface, it is possible to identify some voids in the matrix caused by the debonding and pulling-out of the particles during crack propagation. Besides debonding, the main toughening mechanism seen here relates to crack deviation and crack pinning. Unfortunately, due to the high crosslinking density, there is no evidence for plastic void formation and growth around the particles, which would lead to a significant increase in toughness. This justifies the minor improvement in fracture toughness observed here.

2.3.6

Hyperbranched Polymers

Hyperbranched polymers are highly branched macromolecules based on ABx (x ≥ 2) mono­ mers, which have lately obtained much attention. They are not perfectly branched, as dendri­ mers are, but they are more easily produced due to simpler synthetic routes. A study of Glauser et al. [71] showed that, in order to obtain a fine blending of the electron beam curable resin and the hyperbranched additive, it was necessary to add a solvent in order to lower the viscosity. Both dichloromethane and ether were investigated. The phase separation worked well when using dichloromethane, but complete removal of the solvent required high vacuum that strongly increased the risk for gelling. Dichloromethane also turned out to be disadvantageous for electron beam curing of the resin, because the cure was highly affected by small amounts of residual solvent. Ether was preferred because of its lower boiling point and its lesser affinity for the hyperbranched additive. This also increased the driving force for phase separation. Unfortunately, reaction-induced phase separation, as performed by Boogh et al. [250], was not possible by electron beam curing because the cure process is too fast. The phase separation was governed by several factors. First, the viscosity had to be lowered by adding a solvent, so diffusion could occur on a reasonable time scale. But on the other hand, the solvent had to be evaporated in order to obtain phase separation. A compromise was found by evaporating the solvent until the resin had a syrup-like consistence. The phase separation first provided

2.4  Interfacial Properties Between Fibers and Matrix

77

very small particles, which were too small to provide an increase in toughness. Time was the second factor. Sufficient time was not only necessary for phase separation to occur, but also for the particles to mature, or grow to a size in the micrometer range. Toughness increased slightly for these hyperbranched particles modified electron beam curable resins, as cracks simply propagated through the hyperbranched particles or as secondary cracking was initiated from the hyperbranched polymers.

2.4

Interfacial Properties Between Fibers and Matrix

All reinforcing fibers that are traditionally used for advanced composites can be employed for electron beam cured composites. Carbon fibers have shown to be particularly suited for electron beam processing, since they exhibit excellent radiation stability, tolerating about 1000 times the typical curing dose, and as a good electrical conductors they ensure that no charge is built up in thicker composite structures. Aramid fibers are also unaffected by electron beam treatment up to the typical curing doses, a consequence of its highly aromatic structure. The effect of electron beam treatment on the mechanical properties, density, and chemical durability of glass fibers was shown to be negligible up to a dose of 3 MGy. The only visible effect was discoloration [66]. The use of organic fibers, such as polyethylene or nylon fibers, for reinforcing electron beam cured composites must be considered with caution, since the properties of the fibers are affected by the electron beam irradiation. Designers must be aware that special fibers might affect the electron beam curing process, if at all, and consider these changes in the design. It is known that the fiber properties, such as thermal conductivity and heat capacity, may directly influence the curing of the matrix and consequently the final properties of the composite. As already discussed in Section 1.3.4, the type and amount of reinforcing fibers also influences the maximal temperature (Tmax) reached during curing, due to the capacity of the fiber to transfer heat to cooler zones. When comparing the same loading of glass fibers with carbon fibers, a higher Tmax was reached. The relatively low conductivity of the glass fibers leads to less cooling of the curing zone. Historically, much work was conducted to improve the interfacial strength between carbon fibers and thermally cured epoxy systems. Proper surface treatment and sizing materials were developed to optimize this critical aspect of composite performance. The solutions relied initially on trial and error and more recently have been based on a fundamental understanding of the chemical and physical processes that occur in the interphase region. It is well known that fibers have the potential to perturb the polymer network structure in the vicinity of the surface. Several studies have shown that the interfacial shear strength, as measured by debonding fibers from the matrix, is typically 30–60% lower than that of thermally cured analogs [42, 59]. These results associated the interfacial properties as a consequence of the new mechanisms of curing related to electron beam irradiation. These range from chemical incompatibility

78

2  Aspects of Electron Beam Curable Materials

to physical phenomena linked only to irradiation. In the following section, chemical aspects (fiber surface treatment and use of sizings) as well as processing conditions of electron beam curable composites are discussed in regard to the resulting interfacial properties.

2.4.1

Fiber Surface Treatment and Use of Sizings

Concerning the chemical aspects, it is well known that the functional groups on the surface of the fibers play an important role in improving the surface free energy or the work of adhesion between the fiber and the matrix, which affect the mechanical properties of the composites [131]. In thermally cured epoxy-amine resins the reaction of the amine with the carbon fiber surface is as little as 3% of chemical bonding, but this can account for a 25% increase of the interfacial shear strength [252, 253]. The French company Aerospatiale was the first to realize the need for new fiber sizings to improve the adhesion of EB-curable resins to carbon fibers during the development of EB-curable acrylate filament wound rocket motors. They developed specialized coupling agents that improved the fiber-matrix interface by creating covalent bonds between the fiber and the resin. The hydroxyl groups on the carbon fibers reacted with an isocyanate group on the sizing, and the hydrogen bonds on the sizing’s acrylate group are free to crosslink with the resin during electron beam curing. (6-Isocyanato n-hexyl) carbamoyloxyethyl methacrylate was used as a sizing agent [254] in an acrylic matrix reinforced with carbon fibers. The level of interlaminar shear strength (ILSS) was higher when using fibers sized with the acrylic-isocyanate formulation, compared to unsized fibers or epoxy sized fibers. Zhang et al. [131] evaluated the sizing of carbon fiber in reinforced epoxy resins cured by electron beam irradiation. Unsized carbon fibers were oxidized in acid, alkaline, and neutral electrolytes. From ILSS test results it was concluded that H2SO4 and NaHCO3 as electrolytes increased the adhesion; whereas NaOH reduced ILSS of electron beam cured composites. It had been shown that chemical bonding, also in electron beam curing, takes great importance in the increase of adhesion between the carbon fibers and the matrix resin. H2SO4 electrolyte seems to improve the adhesion much better than an alkaline one. The reason for this result is assumed as the following: it is the proton acid that really initiates the cationic polymerization of the electron beam curable resin. When H2SO4 was taken as the electrolyte, many oxygen-containing groups were added to the carbon fiber surface. In addition, the carbon fiber surface chemisorberd some acidic functional groups. These acidic functional groups and oxygen-containing groups benefit the electron beam curing through strengthening the action of proton acid and the reaction with matrix resin during electron beam process. With NaOH as the electrolyte, the active carbon atoms on the carbon fiber surface chemisorbed the OH- groups. It is known that the OH- can be neutralized with the proton. Studies concerning glass fibers can also be found in the literature. In the work of Goodman et al [61], glass fibers from Owens Corning, have been sized with coatings designed for use with specific epoxy resin systems. These fibers show high interfacial properties with thermally cured epoxy resin. However, poor properties were obtained when these composites were cured

2.4  Interfacial Properties Between Fibers and Matrix

79

by electron beam irradiation. The authors associated these results with the fact that the rapid EB cure may not allow the unbound components of the fibers to diffuse into the bulk resin and react with coupling agents. An alternative explanation was that reactivity in the vicinity of the fiber surface might be decreased due to particular compounds present in the sizing, which may locally inhibit cationically curing reactions. Surface treatments on carbon fibers and a sizing suitable for cationically cured epoxy resins have been developed during a CRADA project, developed by a number of companies and research institutes in North America. In chapter 5 this project is described in greater details. This approach was based on the modification of the chemistry at the fiber-matrix interface during radiation curing, and resulted in proprietary formulations. The major results showed that [59]: yy Sol-gel, dialdehyde and isocyanates sizings can improve adhesion; yy Plasma treatment can be very efficient to increase adhesion in carbon fiber reinforced epoxy composites; yy Epoxy-novolac sizing has a limited efficiency to increase adhesion; yy Epoxy sizing with high photoinitiator concentration can significantly increase adhesion. Employing the above-mentioned methods, it was shown, in the frame of this project, that the interfacial shear strength of electron beam curable composites could be increased in the range of 30 to 50%.

2.4.2

Processing Conditions

The low adhesion of electron beam curable resins to reinforcement fibers has also been investigated and evaluated in terms of the irradiation conditions during processing. The effects of dose-time and temperature-time profiles during the curing were investigated. A large number of experiments showed that fiber-adhesion is relatively insensitive to the irradiation procedure [255]. However, within this same study, the thermal post-curing of the irradiated samples showed to increase appreciably the adhesion between fibers and matrix and consequently the properties of the composite. Therefore, the absence of high temperature treatment was identified as a negative aspect influencing the adhesion between the matrix and the fiber surface. Resin cure shrinkage was also evoked as a responsible effect for the low interfacial adhesion in electron beam curable composites. Increased resin cure shrinkage, although a source of residual stresses and dimensional instabilities in laminates, is also identified as a viable method of achieving incremental improvements in composite transverse strength because interfacial debonding is delayed [256]. It is clear that differences exist in interfacial behavior of electron beam versus thermally cured systems. These differences may stem from altered cure mechanisms as well as the novel chemistry of these new resin systems. Although comparable properties between thermal and electron beam curable composites are found in the literature [59], further investigations into interfacial effects of electron beam curable systems are still required. These investigations are

80

2  Aspects of Electron Beam Curable Materials

currently advancing, mainly focusing on the elucidation of the new mechanisms for interface/ interphase formation related to electron beam curing.

2.5

Residual Stresses

Large temperature gradients in the resin during cure can cause the build-up of non-uniform residual stresses that may influence resin properties under service conditions, because temperature is the main parameter to control the conversion rate. Also, nonuniform cure due to severe temperature gradients can lead to nonhomogeneous thermo-mechanical properties within the resin and deteriorate the quality of the material. This can cause severe warping of unbalanced structures. The thermal curing process typically requires a complex heating and pressure cycle that ultimately must reach temperatures ranging from 150–250 °C and pressures as high as 700 kPa. As a result of the large difference between the curing temperature and the operating temperature for most structures, thermally cured composites contain large residual stresses. Considering that one of the most appealing characteristics of electron beam processing is that it can be considered a “cold” process, the products cured by this technique should also show beneficial effects on the final properties, as they should be free of residual thermal stresses [6]. However, it is known that during polymerization reaction, which is an exothermic reaction, high temperatures may be reached. Therefore, it is of great importance to predict and minimize temperature gradients within the resin during electron beam irradiation. A minimization of the temperature gradient inside the sample will also lead to minimized gradients in the degree of cure, establishing a more homogeneous crosslinked structure. During electron beam curing, the temperature gradient may be reduced by controlling processing parameters such as processing temperature (laboratory temperature), temperature and properties of tooling materials, forced convection heat transfer around the curing materials, time between passes, and others. Concerning the irradiation doses, it was shown that a lower dose per pass leads to higher cure levels and lower residual stresses at a given dose level. In addition, cure-shrinkage induces residual stresses that can be appreciable in electron beam cured multidirectional laminates. In summary, there are much less residual stresses induced during electron beam curing as observed with thermally cured composite [257]. A homogeneous dispersion of the initiator for cationically cured materials is essential to assure cure homogeneity, therefore decreasing the temperature gradient of the materials and reducing internal stresses. At the same time, a higher initiator concentration caused a larger temperature gradient within the resin system because of larger heat release due to a higher cure rate. Usually, for electron beam curable resins, the maximum temperature during cure, due to the exothermic polymerization reaction, is found in the core of the material. This is due to conduction and convection boundary effects, as the process is achieved in ambient temperature.

2.6  Effect of Post-Curing

81

Therefore, the material’s properties of the mold used will exert a high effect on temperature gradients. For instance, in a mold made from glass, the maximum temperature appears in the centre of the core; however, an aluminum mold plate caused the maximum temperature to appear close to the open interface subject to surroundings. This is explained by the high value of thermal conductivity of aluminum [69]. Therefore, a balance between irradiation dose, number of passes, amount of initiator, initiator dispersion, and mold properties should be met in order to minimize the effect of internal stresses in electron beam cured composite materials and structures.

2.6

Effect of Post-Curing

Post-curing consists of subsequent steps after samples have been irradiated by an electron beam. The thermal treatment of the samples may improve some final properties of the materials. However, post-curing of electron beam curable resins is not interesting from an economical point of view. In the literature, there is no conformity concerning the efficiency of thermal post-curing of electron beam curable thermosetting polymers. In some cases, improvement of properties was observed; however other studies showed properties to be independent to thermal post-treatment. A literature review presenting these distinct results and main conclusions is presented in this section. Zhang et al. [258] studied DGEBA epoxy resins cured in presence of diphenyliodonium hexafluorophosphate. The gel fraction of the irradiated samples increased after heat treatment, and the increased extent of gel fraction in the samples was inversely proportional to the radiation dose. This is a result of the existence of residual active centers and reactive functional groups in the resin systems after electron beam irradiation. Curing reactions were difficult to sustain because of the glassy and compact crosslinking structures in the resin systems. When the irradiated samples were heated at high temperature, the trapped active centers acquired enough mobility to induce a further curing reaction, involving residual epoxy groups. Only if the temperature of heat treatment was above the glass transition temperature of the irradiated samples, the heat treatment could play a role through the collision and reaction of residual active centers and reactive groups, which can be devitrified by the heating process. Similar observations were made by Sui et al. [55]. Moreover, it was observed that the post curing of DGEBA resins caused not only increased gel fraction and glass transition temperature, but also mechanical properties, such as Young’s modulus. Post-curing was also related to the fact that residual active centers and epoxy groups still exist after radiation and was found to be more significant for low initiator concentrations. This can be explained by higher concentration of unreacted groups and a significantly lower temperature rise during EB irradiation at lower initiator concentration, which limited the mobility of residual monomers. At higher initiator concentration (thus at higher resin temperature) a larger portion of the reactive species was consumed before being trapped in the vitrified system [69].

82

2  Aspects of Electron Beam Curable Materials

Therefore, increasing the polymerization reaction level during the irradiation step decreases the effect of post-curing on the final properties. As mentioned, one way to increase the speed of reaction is to increase the amount of initiator. There are two other methods to improve the polymerization by electron beam irraditation, one is to use more reactive initiators, and another is to increase the irradiation dose. As shown by Janke et al. [63], the glass transition temperature (Tg) of resins incorporating very efficient initiators does not change, even after several thermal post-cures. On the other hand, the Tg’s of resins having inefficient initiators do change after thermal post-cure. However, and more importantly, even after several thermal post-cures, the Tg’s for these resin systems incorporating less efficient initiators never approach the Tg’s of the same resin systems incorporating highly efficient initiators. Defoort et al. [255] observed that a higher dose is required to reach a fully reacted state when electron beam processing is done with low dose increments. This is mostly due to the temperature profile during electron beam processing and to the vitrification of the material after the first passes under the beam. For the materials that were undercured after electron beam curing, the thermal treatment leads to a completion of the polymerization and subsequently to a significant increase of the glass transition temperature. It is also noticeable that the materials that were partially cured by electron beam and by thermal treatment exhibit lower shear modulus than the materials that were fully cured under electron beam. Some negative effects of thermal post-curing of irradiated samples are also reported in the literature [128, 255]. In highly cured samples, the thermal treatment seems to reduce the crosslinking density of the network. The magnitude of Tg’s decrease after a thermal treatment seems to be more correlated to the initiator concentration in the electron beam cured materials than to any other change during reaction. A tentative explanation of this phenomenon is based on the relaxation of the interactions through ionic pairs. The progress of the cationic polymerization in the vitrifying network is believed to be early affected by the difficult segmental diffusion of the bulky counter-anion. Clusters of weak ion pairs could be formed in the glassy material that could then relax to tight but independent ion pairs if some mobility is authorized by a thermal treatment. In respect to fiber/matrix adhesion, thermal post-cure treatment by electron beam curing increased the interfacial shear strength between the fiber and the matrix. The thermal postcure even increased the fiber/matrix adhesion of materials that were cured with an electron beam up to a high level of curing. The increase of the interfacial shear strength after thermal treatment is not related to a change in the properties of the bulk matrix. The increase of adhesion might be due to two phenomena. The thermal treatment may induce some beneficial thermal stresses in the composite material that was not induced during electron beam curing because of the limited temperature increase during processing. This thermal stress is beneficial to fiber matrix adhesion measured at room temperature. Alternatively, thermal treatment at high temperature might induce chemical bonding between the carbon fibers and the matrix that was not present during the high-speed low-temperature electron beam curing. Even a few unreacted epoxy functions remaining in the matrix after electron beam processing could

2.6  Effect of Post-Curing

83

significantly increase the adhesion to the carbon fibers if they react during the thermal posttreatment [255]. To conclude, owing to the pseudo-living character of electron beam curing, the active centers occluded in the glassy samples are ready to resume propagation upon any temperature rise that would provide the network with some mobility [259]. However, thermal post curing of electron beam cured neat resins is only required for low reactive systems. That means that highly crosslinked networks obtained directly by electron beam curing are not positively affected by thermal post-curing. On the other hand, due to a lack of suitable fibers adapted to electron beam curing, thermal post-curing may still improve some of the interphase properties of such composites.

3

Electron Beam Curing Applied to Composite Molding Technologies

Electron beam curing of composites must be combined with a suitable molding technology. This can be an already established method for thermal curing, such as tape or tow placement, hand-layup with hot debulk, vacuum assisted resin transfer molding, pultrusion, and others, or a molding technology specifically used for electron beam curing, such as lost wax molding techniques. Numerous demonstration parts were manufactured in order to illustrate the flexibility of electron beam curing of composites. Some of them are presented in this section together with a critical discussion of the suitability of the molding technology to fabricate parts by electron beam curing. One of the advantages of electron beam curing compared to thermal curing is that the tools do not have to tolerate the typically high curing temperatures. Lightweight inexpensive tooling materials such as injection-molded polyethylene, wood, and cardboard are used for many electron beam cured composite parts. On the other hand, electron beam curing under pressure, although being possible, is considered to be impractical for most products because two of the benefits of electron beam curing, the overall production speed and the simplicity of the process, are lost [68]. Finally, before going into the details of each composite molding technology, one should be aware that some materials, such as releasing agents, may inhibit curing on the surface of the part [72]. This surface inhibition evaluation, primarily due to “proprietary” mold releases, and a judicious analysis of the effect of such materials should be carried out prior to the use.

3.1

Layer-by-Layer Assembly

This technique uses a single-sided female mold, as shown schematically in Fig. 3.1. The method consists of the lay-up of the reinforcement fibers and subsequently the resin is worked into the reinforcement with a brush or a roller. This process is repeated for each layer of reinforcement until the required thickness is built up. After preparing the laminate, the part is irradiated with the electron beam focused on the top surface. However, it is known that interface adhesion has a major influence on mechanical properties of composites, and this technique does not assure a complete coating of the fibers by the matrix. There might be a high fraction of voids. Therefore, only low quality electron beam cured composites are expected to be manufactured by this molding method.

86

3  Electron Beam Curing Applied to Composite Molding Technologies

Prepreg

Mould

Figure 3.1: Schematic composite preparation by layer-by-layer method

Figure 3.2: A portable electron beam gun for in situ curing was fabricated at Science Research Laboratory, Inc. before installation on the ATP head at NASA Marshall Space Flight Center [12] (Reprinted with permission of the publisher)

Main advantages of this technique are the low costs involved with the preparation of the composites, which renders the technique to be effective for low-performance composite applications. An automated processing method derived from the manual layer-by-layer technique, which is already used for electron beam curing applications, is called automated tape placement (ATP) [260]. In this case, the e-beam gun combines the two steps into one process, leading to the added benefit of potentially faster throughput, since ATP lay-up and the curing process are completed at the same time (Fig. 3.2). As curing takes place at each thin layer placement, low electron beam energies are required. These equipments are easier to install and cost less than high-energy systems. To obtain good consolidation, researchers often use a warm debulk and vacuum bag pressure prior to high energy electron beam curing. On the other hand, the auto tape placement (ATP) head applies heat and pressure for only a very short time, leaving little room for possible

3.2  Prepregging

87

errors. Good tape quality with full fiber wetting is critical to produce in situ cured composites without voids. This is because unlike conventional autoclave curing, there is no time for resin flow to fill tape voids, and thus flaws must be avoided during tape fabrication. Voids can also occur if the top layer springs back due to insufficient “memory” before it is cured. The degree of spring-back as a function of EB cure dose, as well as the tack, drape, and memory of the uncured tape are determined by the viscoelastic properties of the EB-curable resin.

3.2

Prepregging

A “prepreg” is a preimpregnated fiber reinforced material where the resin is partially cured or thickened. The fibers are arranged in a unidirectional tape, a woven fabric, or random chopped fiber sheets. Prepregs are made using a hot melt impregnation method. A strict control of viscosity of the resin is a major issue to impregnate the fibers with the desired amount of resin. A major drawback of thermal curable prepregs is the limited life time and the need for storage at sub-freezing temperatures until required for usage. Prepregs made from electron beam curable resins would therefore be advantageous, as these prepregs can be stored for a much longer time at ambient temperature. This method can be employed also for electron beam curable materials. One of the main issues that should be taken into account for cationically cured epoxy resins is the stability of the initiator to visible light. In cases where these initiators react under light, particular care should be taken during the process to avoid exposure and consequently curing of the material. In respect to viscosity, the resins should be prepared to have a viscosity suitable for the processing window, and the maximum temperature should not exceed the temperature where cationic initiators undergo thermal scission (this is usually between 150 and 200 °C). A low energy electron beam accelerator might have to be incorporated to the prepreg line, in order to partially cure the resin to assure the handability of the material for further processing. The preparation of pregregs with electron beam curable resins still requires further investigations. A research team [59] highlighted several problems with the prepregging of electron beam curable resins, many of which related to the use of new fiber surface treatments and new resin systems. In 1996, Oak Ridge National Laboratory successfully produced electron beam curable glass fiber prepregs that were used for the manufacturing of composite fuel tanks and engine inlets for the US Army LONGFOG anti-tank missile. Although complicated, the usage of prepregs is, together with filament winding, one of the mostly used composite molding technologies for the manufacturing of electron beam cured composite parts.

88

3.3

3  Electron Beam Curing Applied to Composite Molding Technologies

Vacuum Bagging

Reinforcement and resin are laid into the one-sided tool. After placing the stack of fiber layers, this is then covered with perforated PTFE release film followed by a porous woven nylon peel ply fabric. This layer provides a barrier between the laminate stack and subsequent stack and offers an escape route through the laminate thickness for air and excess resin. A bleeder cloth is then added which provides a uniform compaction pressure over the laminate surface due to its high bulk factor. Finally, the vacuum bag film, usually nylon, is used to enclose the entire mold and sealed using a butyl mastic tape. The vacuum connection can be made either via a connection fitted to the mold or a fastening in the bag itself. Curing of composites manufactured by vacuum bagging is already used in electron beam curing. An example is shown in Fig. 3.3, where an aircraft part for military applications is passing under the electron beam and is being cured [262]. The accelerated electrons traverse the bag film to reach the composite and to cure the resin. This means the material from which the bag film is made should not undergo degradation or deformation during irradiation. The same statement is valid for the sealing component of the vacuum bag.

Figure 3.3: Composite part manufactured by vacuum bagging under electron beam irradiation [262] (Reprinted with permission of the publisher)

3.5  Filament Winding

3.4

89

Pultrusion

A collection of fibers in the form of rovings, tows, mats, or fabrics is pulled through a resin bath and then through a heated die. Products range from a simple round bar to complex prismatic sections of up to about 1.5 m maximum dimensions (the shape of the section is similar to the pultrusion die). A flying cut-off saw is programmed to cut the product to the desired length. This is an important process to manufacture continuous composite sections, mainly from glass/ vinyl ester or polyester resins. The continuous character of this process permits a high production rate, and can therefore be adequately adapted to electron beam technology [261]. Up to now, there are still no operational electron beam pultrusion lines. In this case, some modifications would have to be taken into account. The replacement of the hot die by a lowdensity die is the first point, in order to assure penetration of the accelerated beams throughout the material. Moreover, the resin bath would have to be completely isolated from the curing area; otherwise, radiation would also cure the material in the bath. Problems associated with the impregnation of the fibers, as described in the prepreg process, would also have to be faced in pultrusion. From an economical point of view, the pultrusion of glass/vinyl ester composites curable by electron beam irradiation would not be suitable, as curing of such resins is relatively fast. Therefore, one of the main advantages of electron beam curing would not be used: the drastic reduction in the curing time. However, for most high performance composites, such as those reinforced by carbon fibers, electron beam curing associated with pultrusion would have a major significance, as this molding technology can assure a high material production rate.

3.5

Filament Winding

Filament wound composites are produced by resin-preimpregnated fiber tows over a rotating male mandrel. Typically, thermal cure takes place after the part is wound. Process variables such as winding pattern, fiber tension, winding speed, and impregnation techniques are directly related to the properties and quality of filament wound composites. This composite mold method is possibly the most frequently used technique for electron beam curing composites. The French company Aerospatiale has manufactured solid propellant rocket motors using this technique for more than 25 years. The casings were traditionally made from filament wound carbon fibers with heat-cured epoxy resins. Researchers found that a combination of electron beam and X‑ray curing could reduce the curing time from four days to less than eight hours, while keeping the structures nearly at room temperature. After filament winding, the casing is taken to the irradiation facility. The product rotates, at a specified speed, with its axis perpendicular to the beam scan direction. The casing is slowly

90

3  Electron Beam Curing Applied to Composite Molding Technologies

moved under the scan horn, in the direction of its axis. It is moved back and forth under the electron beam until a specified dose is applied. To ensure uniform energy deposition, the product must rotate at least once before the conveyor moves the product a distance equal to the radius of the electron beam spot. The speed of rotation must also be different from the speed that the electron beam traverses across the scan horn. The scan width is also set to be less than the minimum diameter of the product, ensuring that the entire beam is directed onto the product surface.

3.6

Resin Transfer Molding (RTM) and Vacuum Assisted RTM (VARTM)

In resin transfer molding (RTM), a dry fiber preform is placed between the molds and impregnated with the resin, which is injected, either by gravitational or pump pressure. The resin is then cured and the mold opened to release the cured component, which may require subsequent finishing operations. The vacuum assisted resin transfer molding (VARTM) technology is almost identical to RTM. The mold for VARTM is linked to a vacuum pump, and therefore a high sealing level of the mold is required. The vacuum system allows the evacuation of air from the fibers, leading to a composite presenting lower void volume than those manufactured by RTM. Both RTM and VARTM are considered to be the principal viable alternatives to prepregs for the manufacture of high-quality composite moldings. The key processing parameters are the viscosity of the resin, which should be below 500 cps, and the length of the required resin flow path [261]. These methods were already adapted to electron beam curing. A major modification is that the mold should be made from a low-density material that allows the passage of the electron beam without undergoing degradation or deformation. Moreover, this mold should stand some mechanical pressure due to the injection pump and should provide adequate sealing for VARTM. A maritime part [66] was produced by this technique. The composite part was layed up on an inexpensive wood/formica tool. A vacuum-assisted resin transfer molding technique was used for the infusion of vinyl ester resin. Immediately after resin infusion, the part was irradiated to achieve cure. Because electron beam curing can be achieved at room temperature, no special consideration for high temperature operation was included in the tool design and construction. Unfortunately, the elevated temperature, resulting from the exothermic reaction during cure, exceeded the desired operating temperature of the mold. This resulted in a loss of vacuum during the EB cure, exacerbating the tendency of the part to delaminate. Nevertheless, the VARTM method showed to be efficient for electron beam curing and the use of a more appropriate mold material will have an impact on the improvement of the composite properties.

3.7  Lost Core Molding

3.7

91

Lost Core Molding

A difficult problem is to fabricate essentially closed parts that are supported internally during cure and before tooling removal. To overcome this problem, the metal industry developed a method called the “lost core process”, which has been adapted for electron beam curing of composites. Wax is a common, inexpensive material that can be easily cast. Because it melts at relatively low temperatures, it is also an excellent candidate for single use applications where destruction of the mandrel during part removal is not a problem or where an encapsulated tool must be removed. The wax is relatively easy to melt out and can be recast through at least several cycles. Similarly, styrofoam is a common, inexpensive material that can be easily cast or cut to shape. A relatively rapid dissolution by common solvents such as acetone is possible. These characteristics make styrofoam a good candidate for single use applications where destruction of the mandrel during part removal is not a problem. However, one should be aware about the sensitivity of the styrofoam and wax to elevated temperatures in which the curing exotherm from the curing of the part actually may melt the core. With electron beam curing, the cure rate and associated exotherms can be controlled, and these parameters allow the part to be successfully cured to the desired shape.

4

Current Limitations and Potentials for Electron Beam Curing

4.1

Cost Analysis of Electron Beam Curing Processes

If electron beam curing of composite materials becomes established as an industrial process, there will be far reaching benefits to the entire composite industry, primarily because of a reduction in manufacturing costs. Six independent economic studies of the EB curing process in aerospace manufacturing applications have concluded that cost savings ranging from 25 to 65% are possible, depending un the part, including both recurring and nonrecurring costs [45]. The studies were performed by the Oak Ridge Center for Manufacturing Technology, Atomic Energy of Canada Limited, Aerospatiale, Boeing, Lockheed Martin, and Northrop Grumman. These impressive savings result from the following characteristics of the EB process: yy Curing time is reduced. The significant reduction in curing time may allow a manufacturer to utilize an electron beam system as an in-line curing technique in their fiber reinforced polymer composite fabrication processes. yy Tooling costs are reduced. In general, electron beam curing is done at ambient temperatures, allowing the use of easily-fabricated, low temperature tools such as foam or wood. This vastly reduces the tooling cost – a major expense for composite applications. This is especially important for fabricating complex shapes, e.g., engine inlets, closed-end structures, cryogenic tanks, and prototyping of parts leading to final production runs. yy Resin stability at ambient temperatures is improved. Electron beam curable resins do not normally auto-cure at room temperature. This leads to an almost indefinite resin shelf-life, which reduces scrap and inventory control costs. Costs associated with the routine cleaning of resin application equipment are also reduced. Acsion Industries has several hundred electron beam curable epoxy formulations, which have been stored at room temperature for more than 4 years without any loss in properties. yy The production of volatiles is minimized. Electron beam cured resins typically produce 0.1% or less volatiles. Volatile-induced delaminations and out gassing, both of which are significant composite quality control issues, are essentially eliminated. Stack emissions of volatile chemicals from EB-curing can be less than 1% of the emissions caused by thermallycured resins, minimizing the impact of environmental and health concerns. yy The use of chemical crosslinking agents for thermosetting resins is eliminated. Chemical hardeners are typically hazardous (toxic) and carcinogenic. The requirement to add hardeners to composite prepreg materials also limits the shelf-life of these resins to only a few months, even when stored at refrigeration temperatures. Electron beam curable prepregs may be stored for much longer periods at room temperature.

94

4  Current Limitations and Potentials for Electron Beam Curing

yy The curing energy-absorption profile can be controlled to allow greater product design flexibility. Electron beam processing allows the degree of crosslinking to be varied in depth and location, and the curing process can be temporarily halted once started. These extraordinary capabilities enable the development of new products, with design-tailored cross-linked profiles, that current curing methods do not allow. Major costs for implementing electron beam curing are related to the electron accelerator itself and the concrete radiation shielding. Costs of constructing new EB curing facilities capable of meeting industrial curing of composites are estimated to reach up to 4 Million Euro. However, the advantages above-mentioned lead to a pay-back period that will naturally depend on the volume of composite parts processed. An economical study from Northrop found a 2 to 4 year payback period for the assumptions of a single program using EB curing, a seven-year equipment life, and a total of 24.5 to 40.8 ton of composite processed [12]. The processing rate is proportional to the power of the accelerator. As an example, an accelerator rated at 10 MeV and 200 kW (such as the Rhodotron manufactured by IBA) is capable of processing up to 4 ton/hour at a minimum dose of 100 kGy. Although the capital cost of accelerators can range from a few hundred thousand to several million Euro, this high throughput rate can result in polymer processing costs as low as a few cents per kg. It is not necessary for users to own and operate the accelerator facility themselves. Several companies around the world offer toll-processing services [65]. The energy efficiency of an electron beam accelerator has been compared to standard autoclaves. As illustrated in Table 4.1, a commercially available 50 kW accelerator delivering a cure dose of 100 kGy at 70% beam utilization uses roughly one-tenth of the energy of a typical autoclave or oven running a 4 hour thermal cure cycle [263]. This demonstrates that radiation processing requires lower expenditure of energy than the conventional thermochemical processes. An economical successful conversion from thermal curing to electron beam curing of composites can be given by the example of Aerospatiale. Aerospatiale, in France, has been manufacturing the engine casings of ballistic missiles based on electron beam irradiation for several years. These casings range from small (app. 300 mm diameter) casings up to relatively big casings (more than 2 m in diameter and more than 7 m in length, 1 ton of composite). Table 4.1: Relative Energy Efficiency of Electron Beam Curing (Adapted from [263])

Equipment type

Equipment specification

10 MeV Electron beam

50 kW

Autoclave Autoclave

Total energy (kW/hr)

Capacity (kg/hr)

Energy efficiency (kW-hr/kg)

400

1800

0.22

12.2 m length 2.4 m diameter

480

273

1.76

15.2 m length 7.6 m diameter

7660

2730

2.81

4.2  Comparison Between Thermal and Electron Beam Cured Materials

95

This process was implemented to decrease the cost of these products. The industrial facility was set up in 1991 and has been continuously used since this date. The initial investment is the electron generator (typically 2 to 4 million Euro), the building shielding (against electron beam and X‑rays) and cinematic systems to move the parts to be cured and/or the electron generator. The return on investment is obtained from recurring costs decrease (tools for the parts and parts curing costs decrease). For composite engine casings, the cycle time for curing was divided by 6, leading to a reduction of the number of manufacturing tools (20% cost reduction), and savings up to 30% for mandrels (the process being a room temperature process, mandrels can be simplified) were attained. Recurring costs of engine casings were decreased by approximately 15% (this, in fact is rather low as compared to potential savings, and is due to a limited production rate). These operational results are complemented by the theoretical promises of the EB process for high production rates of composite materials: it can be several magnitudes faster to cure parts with this method than with thermal curing or the new routes it opens for production streamlining (such as direct curing on grain for engine casings, curing with pre-integrated metallic inserts, or monolithic curing including glue polymerization) [264].

4.2

Comparison Between Thermal and Electron Beam Cured Materials in Terms of Properties

Conventional fabrication techniques for fiber-reinforced polymer matrix composites incorporate the heat-curing step to crosslink the matrix. There are a number of disadvantages associated with the heat cure process [61]: yy Heat curing can be a fabrication bottleneck, limiting throughput, especially for large composite parts. yy Heat curing liberates low molecular volatiles, causing air pollution and possibly composite voids. yy Thermal expansion mismatch between the fiber reinforcement can cause residual stresses upon cooling. Electron beam curing is a method that should overcome most of the limitations of thermal cured composites. However, the impact of this new process on the final properties of the material is still being analyzed, as this is not yet a mature and well-established technology. A number of studies [42, 62, 65] illustrate that electron beam cured resin systems are equivalent or better, in terms of neat resin tensile modulus and glass transition temperature, than thermal cured materials. Moreover, significant processing advantages can be incorporated. However, it must be stated that the electron beam curable systems are deficient in neat resin toughness, and suffer mainly from low interlaminar shear strength, which can be 30 to 50% lower than thermally cured systems. This is a serious obstacle preventing the widespread use of

96

4  Current Limitations and Potentials for Electron Beam Curing

EB cured polymer matrix composites in many applications. Deficiencies in interlaminar shear strength are believed to be the result of poor fiber-matrix adhesion, which is due to the current lack of suitable fibers developed specifically for the electron beam curing technology. A major work concerning the comparison of electron beam curable systems (epoxy and acrylates) to thermal cured systems was presented by Lopata et al. [45]. The key features of this work for neat resins and composites are summarized in Tables 4.2 and 4.3. Electron beam curable materials, as shown in Table 4.2 and Table 4.3, are equivalent or superior to conventional thermosetting epoxies in several important ways. Main features are: yy Improved mechanical properties compared to autoclave-cured thermoset materials. Polymer matrix composites made from electron beam curable epoxies show mechanical properties that meet or exceed those of the state-of-the-art toughened epoxy composites cured in autoclave (Cytec 977‑2). yy Cryogenic and thermal cycling of EB-curable epoxy laminates showed excellent retention of properties. Mechanical properties of composites made from electron beam curable epoxies, after cryogenic and thermal cycling, were unaffected, and in some cases actually increased in values. Electron beam processing of composite structures for cryogenic applications may offer the only technological means of truly controlling thermally induced stresses in the microstructure of these composite parts. Cryogenic tanks must hold fluids that are at temperatures from −200 to −255 °C. Conventional curing must be performed at temperatures of +150 to +200 °C. This wide disparity between service temperature and curing temperature introduces large stresses in the composite component. The curing temperature is a fully independent variable for electron beam curable epoxies. Curing can be performed at room temperature or lower, minimizing residual stresses in the composite structures. yy Void contents comparable to autoclave cured composites. Many composites fabricated from electron beam curable epoxies using conventional hand lay-up and filament winding processes have had less than 1% void content. These void contents have been achieved with just evacuated pressures (100 kPa pressure) and thermal debulking temperatures less than 70 °C. yy Glass transition temperatures rivaling those of polyimides. Glass transition temperatures (tan δ Tgs) for some EB-curable epoxies and their composites have been achieved with values higher than 390 °C. These are the highest Tgs ever reported for any commercially available epoxy resin. Surprisingly, these Tgs also exceed those of some polyimide resins. yy Superb low water absorption values in the same range as cyanate esters. Some of these EB-curable epoxies have exhibited water absorption values of less than 2% after 48 h water boil. Thermally-cured epoxies normally absorb water in the 3–6% range. yy Lower shrinkage. Shrinkage for EB-curable epoxies has been measured in the range of 2–4%. Shrinkage values for thermally-cured epoxies are generally in the range 4–6%. yy EB-curable epoxies have been processed using several, conventional fabrication methods. Many composite parts manufactured via hand lay-up, tow placement, filament winding, resin transfer molding (RTM), and vacuum assisted resin transfer molding (VARTM) have been produced using these materials, thus demonstrating their fabrication versatility.

4.2  Comparison Between Thermal and Electron Beam Cured Materials

97

Table 4.2: Features of Electron Beam Curable Epoxy and Acrylate Resins Compared to Conventional High-Performance Thermally Curable Epoxies (Adapted from [45])

Feature

Thermal cured epoxy resins

EB cured epoxy resins

EB cured acrylates and esters

Mechanical properties

High-performance

High-performance

Low-performance

Composite prepreg storage

Limited lifetime at refrigerated temp.

Extended lifetime at ambient temp.

Limited lifetime at ambient temp.

Environmental and health concern

Low (resin), high (hardener)

Low

Moderate

Material shrinkage (%)

4–6

2–3

8–20

Glass transition temp. (°C)

Up to 300 °C

Up to 400 °C

Up to 150 °C

Residual stresses in cured composites

Moderate to high (thermal mismatch)

Very low

High (matrix shrinkage)

Water absorption (%)

<6

<2

<6

Material cost – ready to use for composite fabrication (Euro/kg)

3–9 (commercial) 14–35 (high performance)

3–7 (commercial) 14–35 (high performance)

2–7 (acrylates) 17–28 (vinyl esters)

Table 4.3: Property Comparison of Electron Beam vs. Thermally Cured Unidirectional Laminates (Adapted from [45])

a



Property

Thermally cured Cytec 977-2

EB cured 8H a

EB cured 3K a

Cure conditions

Autoclave (6 h, 177 °C)

250 kGy (25 °C)

150 kGy (25 °C)

Void content

Not reported

1.77

0.64

Tg (°C) (tan )

200

396

212

0° tensile strength (MPa)

2537

1869

2200

0° tensile modulus (GPa)

166

168

157

0° flexural strength (MPa)

1641

1986

1765

0° flexural modulus (GPa)

147

196

154

0° compressive strength (MPa)

1580

1683

1428

0° compressive modulus (GPa)

152

149

Not reported

0° interlaminar shear strength (MPa)

110

77

89

H2 permeability, Kp s mm/atm s cm2 uncycled at 25 °C

Not reported

9.22 · 10–9

3.43 · 10–8

H2 permeability, Kp s mm/atm s cm2 uncycled at –196 °C

Not reported

< 2.97 · 10–9

1.62 · 10–8

Proprietary resin formulation

98

4  Current Limitations and Potentials for Electron Beam Curing

yy Controlled B-staging during manufacture. Electron beam curable epoxies can be B-staged precisely because of the excellent control of dose delivered to the part. B-staging can also be accomplished by UV-curing, for example, in filament wound parts as they are being manufactured. This allows parts to be made with extremely low viscosity resins, but with B-staging the viscosity is increased to stop resin flow for shipping. Electron beam curing of the B-stage gives the part the same properties as if it was electron beam cured without the B-staging. yy Single cure cycle. EB-curable epoxies require a single cure cycle with curing doses of 100–200 kGy. Exposing the resin to a higher dose does not impair the properties of the composite. This capability allows manufacturers to place resins with different properties in various regions on a single lay-up structure. An example would be a single composite aircraft structure that is made with extremely high service temperature resins in one area of the lay-up, low-modulus, toughened resins in other areas, and high-strength resins in still other areas. The conventional thermal curing process does not allow this approach because of the different curing cycles required for each resin. Electron beam curable resins show equivalent or better properties than state-of-the-art thermally cured resins in terms of mechanical and thermal behavior, and much effort has been placed on the further development of these radiation curable materials. However, electron beam curable composites are still behind thermal cured materials in terms of interlaminar shear strength, due mainly to low adhesion between matrix and fibers. The development of more suitable sizings or fiber treatments is a major issue for the widespread of this new technology.

4.3

Summary of Potential Applications for Electron Beam Curing

Rapid processing by means of electron beam curing offers advantages for both high-volume and low-volume production for the aircraft, aerospace, automotive, marine, and sporting goods industries. High-volume part production is accelerated by the short cure times. Low-volume production and prototyping are made easier by the use of inexpensive, easily constructed fabrication tools and molds. The combination of high-performance electron beam curable epoxies and processing also allows the manufacturing of advanced composite structures that cannot be built using conventional manufacturing technologies. Thick cross-sectional structures, large surfaces with unbalanced fiber reinforcement, structures having embedded metal fasteners, dis­similar fiber, and/or dissimilar resin formulations, can all be processed using electron beam curing. The aerospace industry is already pursuing implementation of electron beam technology in some of its most important market sectors:

4.3  Summary of Potential Applications for Electron Beam Curing

yy yy yy yy

99

Military aviation: advanced military aircraft (several vehicle programs); Commercial aviation: high speed civilian transport; Space exploration: lightweight spacecraft structures; Manned spacecraft: reusable launch vehicles and cryogenic fuel tanks.

Composite suppliers to other industries are also beginning to evaluate the unique capabilities of electron beam curable epoxy resins. Potential applications are already being studied in these diverse markets: yy Transportation, such as low-cost, lightweight composite vehicle structures for automobiles, vans, trucks, railcars, ships, and pleasure boats; yy Infrastructure, such as lightweight, corrosion-resistant building materials, telephone poles, bridges, electrical transmission towers, pier pilings, oil pipelines, and offshore drilling platforms; yy Sporting goods, such as golf club shafts, skis and ski poles, tennis racquets; yy Electronics components, such as printed circuit boards; yy Military (ground and marine) hardware, such as ballistic protection structures, lightweight shelters, composite armored vehicles, missile components, and submarines. Further advances in electron beam accelerator technology and resin chemistry, based on the work in progress, would make the following composite applications feasible in the near future: yy Continuous manufacturing by electron beam curing, using pultrusion and extrusion technologies; yy In-line electron beam curing for filament wound and tow/tape wrapped products; and yy Composite repair technology Small, portable electron beam accelerators are currently available for use in some of these applications. Composite fabrication machinery must be redesigned in order to safely combine them with electron beam processing for optimum results. The Air Canada Winnipeg Maintenance Base has developed a program with Acsion Industries, and is currently using electron beam technology for repair of composites on their Airbus aircraft. The first repair of a composite component using this technology was placed in one of their commercial aircraft as early as in 1998.

5

Research Trends and Projects in the Field of Electron Beam Curing

The present book already provides a large literature review about the main aspects involved in the field of electron beam curable composites, such as curing reaction, toughening of materials, interfacial properties, molding technologies, among others. Therefore, within this chapter, the authors would like to give a general overview about the trends followed by research efforts in the last few years in the field of electron beam curing of composites. The majority of studies and programs related to electron beam curable composites focus on the development of resin systems, optimization formula of curable matrix, and improvement of some properties of the system [61, 71, 73, 249]. These studies were undoubtedly the basis for the development of a range of systems curable by electron beam irradiation, today commercially available. More recently, some works were dedicated to the fundamental study of EB curing. The understanding of how EB curing takes place and its dependence on materials parameters is important to offer statistically design data according to desired goal and accelerate the development and maturity of EB process technology [30, 129]. Radiation dose, for instance, was correlated to the final properties of the EB cured material, such as glass transition temperature and percent of conversion [30, 169]. In situ measurements permitted the establishment of the relation between the progress of polymerization and the glass transition temperature, which was quite useful for chemically defining the characteristics of the material to be obtained. Moreover, the knowledge of the continuous progress of vitrification upon cross-linking polymerization allows exerting a better control over the curing process [128, 265]. In a different study [62], the reaction mechanisms of electron beam induced cationic polymerization of epoxides and the influence of hydroxyl group containing species (e.g., absorbed water in resin system, intermediates formed during reactions, monomers and co-monomers) on the reaction kinetics and polymer network formation was evaluated. Moreover, the inevitable exothermic effect due to the crosslinking reaction and the absorption of high-energy electrons was also investigated, and a valuable understanding of the temperature change and its influence on the properties of EB cured samples could be obtained [58, 127]. Thermal effects induced by the electron beam were also numerically simulated on the basis of the general heat equation applied to a one-dimensional system [266]. The number of research works on electron beam curing has been increasing year by year and it became clear that properties of electron beam cured materials and the degree of cure depend not only on the total dose received but also on the thermal history during cure. Therefore, the reaction mechanism and the kinetics of electron beam initiated polymerization had to be characterized.

102

5  Research Trends and Projects in the Field of Electron Beam Curing

The development and application of a simple process model for electron beam processing of polymer and composites was presented in 2002 [57]. This model incorporated a onedimensional heat transfer analysis coupled with a developed model for the kinetics of radiation curing and could be used to predict internal temperatures and progression of cured during processing. Various enhanced models to describe more accurately the characteristics of the observed behavior were subsequently developed [267]. A hygrothermal model [69] was developed for a thick thermoset resin during EB curing, which could predict not only the temperature evolution, but also the glass transition temperature rise characteristics, dynamic change in the resin yield stress, and water vapor pressure. It has been pointed out that the rates of the photoinitiated cationic polymerizations of commercially available epoxies were lower than those of the corresponding photoinduced radical polymerizations of acrylated monomers. To overcome this drawback, it has become a common practice to employ photosensitizers in combination with onium salt photoinitiators to enhance the reactivity of these cationic systems. In conclusion, there is an apparent acceleration of the rate of polymerization of the monomer in the presence of a photosensitizer as compared to the case where it is absent. A study showed that anthracene can be efficiently employed as photosensitizer for both commercially available triarylsulfonium and diaryliodonium salt cationic photoinitiators [268]. In 1994, a cooperative research and development agreement (CRADA), called “Improved epoxy resins cured by electron beam irradiation” was established in the USA. The purpose of this CRADA was to conduct research and development activities to better understand and utilize the electron beam polymer matrix composites curing technology. This three-year effort involved two Department of Energy (DoE), national laboratories, and 10 industrial partners, including four major aircraft and aerospace companies. CRADA researchers achieved a breakthrough for the composites industry by successfully developing cationic epoxy resin formulations and electron beam curing cycles that reduce processing time and cost while meeting the demanding requirements of high-performance composite structures. Several electron beam curable resin families, including acrylates, methacrylates, vinyl esters, and epoxies, were studied. This developmental effort yielded a broad family of high-performance, EB-curable epoxies possessing a wide range of processing and property profiles [45, 269]; in addition, four patents were filled. The achievements of the first CRADA led to the development of electron beam curable resins possessing engineering properties (strength, service temperature) equal or superior to conventional thermally cured epoxy composites. However, electron beam cured composites had shown deficiencies in shear strength [270], and a second CRADA was initiated in 1999. The essential purpose of the project “Interfacial properties of electron beam cured composites” was to improve the mechanical properties of electron beam cured, carbon fiber reinforced epoxy composites, with a specific focus on composite shear properties for high performance aerospace applications. This project focused [59] on developing and evaluating new electron beam compatible fiber surface treatments, fiber sizings, reactive finishes, chemical coupling agents, wetting agents and modified electron beam curable epoxy resin systems specially for electron beam processing.

5  Research Trends and Projects in the Field of Electron Beam Curing

103

The project team made several discoveries. A number of fiber coatings or treatments were developed that improved fiber/matrix adhesion by 40% or more, according to microdebond testing. Electron beam curable resin properties were improved substantially, with 80% increase in resin toughness, and 25% and 50% improvement, respectively, in ultimate tensile strength and ultimate tensile strain in comparison to the earlier generation of electron beam curable resins. Additionally, a new resin was developed with generally good properties, and a very notable 120% improvement in transverse composite tensile strength versus earlier generations of electron beam cured carbon fiber reinforced epoxies. Despite the aforementioned impressive accomplishments, the project team did not fully realize the project objectives. The best methods for improving adhesion were combined with the improved electron beam 3K resin to make prepreg and uni-direction test laminates from which composite properties could be determined. Nevertheless, only minor improvements in the composite shear strength, and moderated improvements in the transverse tensile strength, were achieved. The project team was not satisfied with the laminate quality achieved, and the low quality (specifically, high void content) laminates compromised the composite properties. There were several problems with the prepregging and fabrication, many of them related to the use of new fiber treatments. Funded by the German Ministry of Education and Research (BMBF), a research project called “Non-thermally Curable Resin Systems” was initiated in 2005. This 3-year project involved three research centers, namely IZM Fraunhofer, Forschungszentrum Karlsruhe, and the department of Polymer Engineering of the University of Bayreuth, as well as the industrial partner Henkel AG & Co KGaA. In the frame of this project, new highly efficient initiators have been developed, based on complex silver salts [101], see also Section 2.1. Moreover, fundamental structure/properties relationships of electron beam curable resins and composites were also investigated in detail, with the focus laying on the toughness modification of such materials (see Section 2.3), which resulted in a number of publications and patents [251, 271–273, 329].

6

Examples of Electron Beam Curing Applications

6.1

Automotive

6.1.1

Composite Concept Vehicle, Daimler-Chrysler

Chrysler Corporation’s advanced technology group investigated methods to produce an allcomposite car body and designed an automotive production facility that included injection molding and electron beam curing. In theory, an entire car can be electron beam cured, even including the adhesive and the paint. However, Chrysler’s advanced technology group determined that there are too many technical risks to this approach, and instead decided to produce an all-composite thermoplastic car frame using injection molding [274]. The result is the composite concept vehicle (CCV), which was presented to the public in 1997. The CCV body is composed of six large injection-molded pieces made of glass-reinforced polyethylene terephthalate (PET). The CCV inner- and outer-frames are bonded with adhesive to themselves and to a steel chassis. These adhesives were prepared from electron beam curable resins. Conventional heat-curing adhesives require several minutes to bond. Total bonding time includes the time to heat the material at the bond-line, plus the time to cure the adhesive. Some of the bonds in the CCV join the plastic to a steel frame chassis. After joining with heat curable adhesives, some sections of these bonds and nearby composite can crack when cooled to room temperature. The cracking is due to the thermal expansion mismatch between the two materials. Therefore, for automotive application, electron beam bonding offers high throughput and “command-cure” bonding near room temperature. Electron beam curable adhesives can be used to bond composites to a metal substructure, avoiding the afore-mentioned de-bonding during the cool-down cycle that can occur due to differences in thermal expansion of the metal and composite.

6.1.2

Composite Armored Vehicle, US Army

The Composite Armored Vehicle Advanced Technology Demonstrator (CAV-ATD) is an armored combat vehicle that demonstrated the first use of a fully integrated armor and structure composite hull. The side skirt panels are composed of three layers including base composite, ceramic tile armor, and signature protection (Fig. 6.1) The S‑2 glass/epoxy base composite and armor tiles are each approximately 1.25 cm thick. Each side skirt is approximately 1 square meter and contains approximately 60 hexagonal tiles [67].

106

6  Examples of Electron Beam Curing Applications

Figure 6.1: Representation of the composite side skirt panels in the composite armored vehicle

A Vacuum Assisted Resin Transfer Molding (VARTM) process was developed as a low cost alternative to Automated Fiber Placement (AFP) for fabrication of the armor panels. Electron beam curing can be used in combination with VARTM for additional processing advantages. Because electron beam curing is a low temperature process, materials with disparate thermal expansion coefficients can be easily cured together. Two spare side skirt panels were fabricated using low cost, low temperature wood tooling with an electron beam cure. The electron beam curing time for the thick-section laminates was less than one hour. Processing steps for these spare skirts include lay-up of the layers on the low-cost tooling, bonding of the ceramic tiles using conventional or electron beam curable adhesive, and injection of electron beam curable resin in a VARTM process. Novel two-component electron beam curable resins have been developed which react up to a suitable B-stage for the handling at room temperature. The composite is then removed from the mold. The electron beam curing of such demonstration was demonstrated to be successful.

6.2  Aircraft Industry

6.2

Aircraft Industry

6.2.1

Patch Repairs for Civil and Military Applications

107

The use of composite patches on metallic or composite structures has gained wide acceptance as an appropriate repair technology for military applications. Potential application opportunities on civil aircraft are very significant, particularly within the framework of the current Aging Aircraft programs. The range of applications includes fatigue life enhancement, corrosion repair, and patching of actual structural cracks. Currently, the technology itself is complex to implement with consistency, due to the very high levels of process control and technician skills required. The principal benefit of this technology is economic, since it offers more structurally efficient repairs, which can be performed in a shorter timeframe than existing metallic repair technology. Instead of replacing the composite parts, a repair patch is used to save money. The repair patch is placed over the damaged area and then thermally glued, using an epoxy resin. A schematic representation of a composite patch repair is presented in Fig. 6.2. An example of this technology is the repair of the vertical stabilizer of the US Air Force F‑18 Hornet [275]. The main challenges of the repair program for the vertical stabilizer using composite patch repair is to avoid skin separation and microcracking, which allows fluid penetration and is consequently responsible for a reduction in properties. Microcrackings can significantly be reduced by repair at room temperature, for which electron beam becomes a fundamental tool. Moreover, the use of electron beam curing makes it possible to accelerate the curing process. As this composite patch repairing technology is already established for military applications, this know-how is now being transferred to civil aircraft applications. An electron beam curable path has been developed and evaluated by the Canadian company Acetek Composites Inc. Acetek Composites Inc. was established in 2001 by a joint venture of Acsion Industries and Air Canada. Acetek’s mandate is to commercialize electron beam technology for aerospace applications through development and certification activities. To support this mandate Acetek Composites Inc. became a Transport Canada qualified Authorized Maintenance Organization (AMO) in 2002. An Airbus A320 aircraft contain composite parts in more than 20 sections. Traditional thermal-cure repair techniques are costly because of the time the aircraft is out of service Carbon fibre prepreg patch repair

Vaccum bag

Prepared surface

Damaged composite structure

Figure 6.2: Typical cross section of a composite patch repair

108

6  Examples of Electron Beam Curing Applications

Lower rear fairing Wing-to-body fairing Figure 6.3: Lower rear and wing-to-body fairings repaired by electron beam curing

due to hand lay-up and curing process flow times. Acetek has collaborated with Air Canada to repair composite secondary structures in their Airbus fleet (A319, A320, and A340) using electron beam techniques in order to validate the numerous electron beam process advantages, particularly with respect to rapid cure cycles, ambient temperature curing, and the possibility of on-aircraft curing to repair parts without loss of material properties. The first demonstrator was a fairing, located at the lower rear of the aircraft fuselage, which was repaired using electron beam irradiation. The location of the rear fairing can be seen in Fig. 6.3. This fairing was in service on the aircraft for 900 landing/take-off cycles over 16 months. This was the first commercial airline application using this technology in the world. The type of damage normally encountered on this component is primarily from baggage loading equipment and objects being thrown up by the aircraft’s wheels. In addition, the component is also exposed to considerable amounts of various fluids, such as oils, hydraulic fluid, and water. The patches used for electron beam repair were made using glass fibers and a proprietary epoxy resin developed at Acsion, called 1L0. The repairs were produced from preimpregnated materials, using similar techniques as in conventional thermal processes. The plies were laid down and compacted under vacuum. Some heat was additionally applied to consolidate the uncured laminate and to remove entrapped air. The parts were prepared for cure with a vacuum bag system that incorporated an opaque covering to prevent premature cure from UV light sources. Subsequently, the parts were irradiated with a high-energy electron beam to fully cure the patch repair. Preliminary testing showed a marked difference in properties between the thermal and electron beam repaired panels. Most notable was the retention of properties at higher temperatures for the electron beam cured panels. Better hot/wet properties were also observed because electron beam cured resin systems tend to have higher crosslink densities compared to thermal systems. The higher crosslink densities prevent water and other liquids from diffusion into the resin system thus causing a loss of physical properties. Subsequently, as a first step in certifying this process for aerospace applications, an electron beam repair on Airbus aircraft wing-to-body fairings has been approved by Transport Canada through release of a Repair Design Certificate (RDC No. C‑RA03‑059). This is the first approved electron beam repair application in aviation history. In parallel with this approval, flight trials have been conducted on fairing and engine cowl components with repairs accumulating thousands of flight hours.

6.2  Aircraft Industry

109

Air Canada is now routinely using electron beam curing for the wing-to-body fairings, a first for any aircraft, military or civilian. Eventually electron beam curing could be used to repair all composite components on the aircraft, minimizing downtime and saving money. It is estimated that every 8 hours of downtime for a plane costs half a million Euro. In the long term, it is possible that repairs could be carried out on the runway by a remotely controlled curing system, with new materials being rapidly laid up to repair the damage and the panel could be cured within minutes, ready for take-off again. Clearly, this system will require the use of low-energy electron accelerators, in order to limit the shielding required. As consequence, curing would be achieved on a layer-by-layer basis, due to the low penetration of such lowenergy systems.

6.2.2

T-38 Talon, US Air Force

The U.S. Air Force Advanced Composites Program Office sponsored a project, which studied the feasibility of electron beam curing for the T‑38 composite windshield frame [276]. The T‑38 is the U.S. Air Force’s primary supersonic trainer and the composite windshield frame is an integral structural part of the windshield assembly system, rated to withstand a 1.8 kg bird impact at 740 km/h. The windshield structure is seen in Fig. 6.4.

Figure 6.4: T-38 composite windshield structure (Courtesy of nasaimages.org)

110

6  Examples of Electron Beam Curing Applications

There were two major goals for this program. The first was to determine whether the electron beam curable prepregs were capable of meeting the requirements of the US Air Force T‑38 supersonic jet trainer composite windshield frame. The T‑38 windshield frame is currently manufactured using an aerospace-grade prepreg composed of 6781 S‑2 woven glass fibers and various 120 °C thermally cured epoxies. The second goal was to develop the lowest cost hand lay-up and debulk process that could be used to produce laminates with acceptable properties. Four resins and three fiber sizings were experimentally examined to evaluate mechanical and analytical properties of laminates at room and elevated temperatures, prepreg tackiness, prepreg out-time capability, and the debulk requirements needed to achieve the required properties. Initially, ten different lay-up and debulk procedures were screened for their suitability towards meeting the program’s goals. It was determined that a hand pressure lay-up and room temperature vacuum bag debulk every four plies produced laminates with properties as good as, if not better than, the other more elaborate lay-up and debulk processes and met the goals of the program. Using this process, laminates were produced presenting void fractions in the range from 1 to 4%. Three of the prepregs studied generated laminates with tensile strengths at average equivalent to the thermally/autoclave-cured prepreg used as a reference system. The compressive and flexural strengths of these EB prepregs were measured to be in the range of 20 to 30% lower than those of the reference system. On the other hand, the moduli for the best electron beamcured prepregs were equivalent or better. The glass transition temperatures of all electron beam cured prepregs were significantly higher than the thermally cured prepreg systems. The high temperature properties of these EB resins strongly suggest two aspects. Firstly, their properties should be compared to other high-temperature, aerospace-grade prepregs to determine if they may have the very important high-temperature capabilities that many aerospace systems require. Secondly, the data suggest that there is sufficient high-temperature capability within the EB curable resins that additional toughening modifiers could be added to improve the lower-temperature (below 177 °C) properties, which are important to a large community within the aerospace industry.

6.2.3

F/A-18 Hornet, Northrop Grumman

The Affordable Polymer Composite Structures (APCS) program examined many of the key issues facing the aerospace industry as it begins to adopt electron beam curing and bonding. The overall goal of the APCS program was to see if electron beam curing could significantly reduce the recurring and non-recurring cost of aerospace composites. Northrop Grumman personnel evaluated electron beam curable materials, developed new processes, accessed electron beam cost benefits, and produced demonstration assemblies, including a full-scale article. The electron beam curable resins evaluated by Northrop included acrylated and methacrylated epoxies, cationic-curing epoxies, a modified bismaleimide, a modified epoxy novolac, and a modified epoxy.

6.3  Space Applications

111

The electron beam cured composite properties approached those of thermal cured composite, except in compression and shear strength. Northrop attributed this to poor resin-fiber adhesion, as indicated by dry fibers in the failed region of the specimens. The research group believed that a fiber sizing or a coupling agent compatible with the EB-curable resin can improve the interface [277]. The APCS process development concentrated on ways to reduce porosity in prepreg composite laminates, which were initially high (in excess of 5%). Hot debulk prior to electron beam cure significantly reduced porosity, and application of external pressure during hot debulk increased the fiber volume in the laminates. For resin transfer molding (RTM) and vacuum-assisted RTM, they found that laminate porosity could be controlled by proper degassing of the resin before injection and by maintaining vacuum integrity throughout the infusion and cure cycle. The APCS program fabricated many parts including skins, covers, doors, and various substructures, as well as a full-scale demonstration article for the F/A‑8 aircraft. Cost studies constituted a significant portion of the APCS program. Northrop conducted cost trade studies for composite monolithic details, sandwich and integrally stiffened structures, and various assemblies. They found that the cost savings potential of electron beam processing varied with the structure, complexity, and manufacturing process, and ranged from approximately 10% in case of simple details with direct process substitution up to 60% for a complex assembly. The cost estimation included the cost of electron accelerator systems, buildings with the required features, part handling systems, operating, maintenance, and repair. Capacity analysis for each facility and return on investment analysis were also performed. The pay-back period naturally depends on the volume of composite parts processed. Northrop found a 2 to 4 year payback period for the assumptions of a single program using EB, based on a seven-year equipment life.

6.3

Space Applications

6.3.1

Space Shuttle Venture Star, Lockheed and NASA

Lockheed and NASA investigated the use of electron beam processing for curing whole wings and fuselage sections, due to the already known advantages of such one-part structures. Major programs included the high-speed civil transport aircraft and the Venture Star, the planned Space Shuttle replacement vehicle. The artistic representation of the Venture Star is seen in Fig. 6.5. In respect to the Venture Star project, electron beam curable 1/10th scale liquid hydrogen tanks were produced. The method used of curing the tank was the following: the tank was laid up in sections using inexpensive foam and wood tooling materials and was first partially cured using Acsion’s 1 kW, 10 MeV accelerator. The components were then assembled, and cured to produce a tank approximately 90 cm × 180 cm × 180 cm. In reality, these tanks would measure

112

6  Examples of Electron Beam Curing Applications

Figure 6.5: The X‑33 or Venture Star, proposed replacement for the space shuttle (Courtesy of nasaimages.org)

30 m × 16 m, and the current autoclave technology would be far too expensive, taking into account both the autoclave itself and energy costs for producing such parts. Electron beam irradiation is therefore the suitable technology for curing such large hydrogen tanks. The total cost using electron beam curing was estimated to be about 51 million Euro, compared to 78 million Euro or more for conventional techniques [263]. Electron beam curing of such parts would also reduce the internal stress in the composite tanks, give lower microcracking, and offer much better hydrogen permeability than thermal cured parts. Pressure tests indicated that the hydrogen barrier properties of the tank were within specifications. However, the X‑33 program was cancelled by the American Government before cryogenic testing could be performed.

6.3  Space Applications

6.3.2

113

Satellite’s Flywheel, AFS Trinity and NASA

Modern satellite’s electromechanical flywheels couple an engine generator with additional rotating mass. These are used for stabilization purposes. Depending on whether the application requires a long or a short discharge time, one may have a small engine generator coupled to a large rotating mass, or the reverse. The term Flywheel Motor Generator (FMG) is used to denote the rotating mass and the engine/generator itself, with its vacuum enclosure, coolant system, and sensors. The rotating mass is a high strength carbon composite flywheel. The position of a flywheel on a satellite is seen in Fig. 6.6. In 1997, AFS Trinity started along with NASA a project for the development of electron beam curing methods for composite flywheel rotors. The flywheel was manufactured by filament winding and cured by electron beam. AFS Trinity delivered aerospace flywheel systems in prototype form in 2005 for flight-testing in 2006. Additional aerospace applications are expected to emerge as the technology advances. So far, the main conclusions obtained from this project are the following: yy Successfully electron beam cured composite rotor, 15 cm outside diameter, 15 cm long, and 2.5 cm thick. yy Developed and successfully demonstrated electron beam process parameters. yy Electron beam curing process costs were 30% lower than thermal cure. yy Fabrication costs (which includes filament winding and electron beam curing) were less than 14 Euro per kilo of fabricated composite.

Figure 6.6: Composite flywheel used in satellite’s stabilization system (Courtesy of nasaimages.org)

114

6.3.3

6  Examples of Electron Beam Curing Applications

Satellite’s Reflector Dish, Acsion and CASA (Spanish space agency)

Reflector dishes are satellite antennas that make up an essential part of all satellite communications systems, and make sure that satellites stay in perfect contact with the ground in the most cost-effective way. Antennas are often large, because radiation pattern control requires a certain physical size, and this most often rules out redundant systems. That is why satellite antennas have to be very reliable, stable and, at the same time, light weight. Composite materials are therefore very suitable for these applications. The advantage of near ambient temperature curing by electron beam irradiation allows the development of composite materials with extremely low coefficients of thermal distortion, essential for composite reflector dishes. Acsion has produced a composite satellite reflector dish by electron beam curing in a partner­ ship with CASA, the Spanish Space Agency. The major requirement was dimensional stability at an operating temperature of 100 °C, which was easily achieved by electron beam curable composites. Moreover, the manufacturing process showed to be advantageous from an economical point of view, as the complete process took less than 6 hours.

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40]

Mehnert, R., Nucl. Instr. Met. Phys. Res. B (1995) 105, p. 348 Charlesby, A., Wycherley, V., Intl. J. of Appl. Rad. Isot. (1957) 2, p. 26 Chapiro, A., Radiat. Phys. Chem. (1979) 14, p. 101 Singh, A., Saunders, et al., Radiat. Phys. Chem. (1996) 48(2), p. 153 Beziers, D., Perilleux, P., et al., Radiat. Phys. Chem. (1996) 48(2), p. 171 Alessi, S., Parlato, A., et al., Proceedings of the 8th Convegno Nazionale AIMAT (2006) Palermo, Italy Clough, R. L., Nucl. Instr. Met. in Phys. Res. B (2001) 185, p. 8 Glauser, T., Kungl Tekniska Högskolan (1999) Crivello, J. V., J. Polym. Sci., A: Polym. Chem. (1999) 37, p. 4241 Buback, M., Van Herk, A. M., Radical Polymerization, Kinetics and Mechanism, Macromolecular Symposia (2007) 248, Wiley-VCH, Weinheim Moad, G., Solomon, D. H., The Chemistry of Radical Polymerization (2005), Elsevier, Amster­dam Goodman, D. L., Palmese, G. R., Chapter in the Handbook of polymer blends and composites (2001) A. Kulshreshtha and C. Vasile, Eds., Rapra Technology Ltd. Goethals, E. J., Cationic Polymerization and Related Processes (1984) Academic Press, London Kennedy, J. P., Maréchal, E., Carbocationic polymerization (1982) John Wiley & Sons, New York Plesch, P. H., The Chemistry of Cationic Polymerization (1963) Pergamon Press, Oxford Matĕjka, L., Dušek, K., et al., J. Polym. Sci., A: Polym. Chem. (1997) 35, p. 665 Matĕjka, L., Dušek, K., et al., J. Polym. Sci., A: Polym. Chem. (1997) 35, p. 651 Matĕjka, L., Chabanne, P., et al., J. Polym. Sci., A: Polym. Chem. (1994) 32, p. 1447 Chabanne, P., Tighzert, L., et al., J. Appl. Polym. Sci. (1994) 53, p. 787 Chabanne, P., Tighzert, L., et al., J. Appl. Polym. Sci. (1994) 53, p. 769 Chabanne, P., Tighzert, L., et al., J. Appl. Polym. Sci. (1993) 49, p. 685 Bouillon, N., Pascault, J. P., et al., Makromol. Chem. (1990) 191, p. 1435 Bouillon, N., Pascault, J. P., et al., Makromol. Chem. (1990) 191, p. 1417 Bouillon, N., Pascault, J. P., et al., Makromol. Chem. (1990) 191, p. 1403 Roffey, C. G., Photogeneration of Reactive Species for UV Curing (1997) John Wiley & Sons, New York Randall, D. R., Radiation Curing of Polymers (1987) Springer, New York Allen, N. S., Edge, M., J. Oil Colour Chem. Assoc. (1990), 73, p. 438 Holman, R., Oldring, P., UV and EB Curing Formulation for Printing Inks, Coatings and Paints (1988) SITA Technology, London Alessi, S., Parlato, A., et al., Radiat. Phys. Chem. (2007) 76, p. 1347 Alessi, S., Calderaro, E., et al., Nucl. Instr. and Meth. in Phys. Res. B (2005) 236, p. 55 Spadaro, G., Calderaro, E., et al., Radiat. Phys. Chem. (2005) 72, p. 465 Dispenza, C., Scró, F., et al., Radiat. Phys. Chem. (2002) 63, p. 69 Kang, P. H., Park, J. S., et al., Macromol. Res.(2002), 10, p. 332 Kozlov, A. A., Doroshenko, V. N., et al., Vysokomol. Soyed. (1983) A25, p. 1505 Penczek, S., Wieteszka, J., et al., Makromol. Chem. (1966) 97, p. 225 Sundell, P.-E., Jönsson, S., et al., in Radiation Curing of Polymeric Materials, ACS Symposium Series (1990) 417, p. 459 Mah, S., Yamamoto, Y., et al., Macromolecules (1983) 16, p. 681 Mah, S., Yamamoto, Y., et al., J. Polym. Sci., Polym. Chem. (1982) 20, p. 2151 Mah, S., Yamamoto, Y., et al., J. Polym. Sci., Polym. Chem. (1982) 20, p. 1709 Ma, X.-H., Yamamoto, Y., et al., Macromolecules (1987) 20, p. 2703

116

[41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] [65] [66] [67] [68] [69] [70] [71] [72] [73] [74] [75] [76] [77] [78] [79] [80] [81] [82] [83] [84]

References

Lopata, V. J., Sidwell, D. R., Radtech Report September/October (2003) p. 32 Berejka, A. J., Eberle, C., Radiat. Phys. Chem. (2002) 63, p. 551 Singh, A., Nucl. Instr. and Meth. in Phys. Res. B (2001) 185, p. 50 Saunders, C., Lopata, V., et al., Radiat. Phys. Chem. (2000) 57, p. 441 Lopata, V. J., Saunders, C. B., et al., Radiat. Phys. Chem. (1999) 56, p. 405 Emmi, S. S., Corda, U., et al., Radiat. Phys. Chem. (2007) 76, p. 1251 Chen, J. H., Johnston, A., et al., Radiat. Phys. Chem. (2006) 75, p. 336 Moon, S.-W., Morgan, R. J., et al., J. Compos. Mater. (2004) 38, p. 357 Moon, S.-W., Morgan, R. J., et al., J. Compos. Mater. (2004) 38, p. 375 Park, J. S., Kang, P. H., et al., J. Macromol. Sci., A Pure Appl. Chem. (2003) 40, p. 641 Alessi, S., Dispenza, C., et al., Radiat. Phys. Chem. (2007) 76, p. 1308 Defoort, B., Drzal, L. T., J. Adhes. (2003) 79, p. 361 Raghavan, J., Baillie, M. R., Polym. Compos. (2000) 21, p. 619 Lee, J., Palmese, G. R., Abstracts of Papers of the American Chemical Society (2004) 228, p. 387 Sui, G., Zhang, Z. G., et al., Mat. Sci. Eng. A (2003) 342, p. 28 Kumar, R. N., Mehnert, R., Macromol. Mater. Eng. (2001) 286, p. 449 Johnston, A., Cole, K. C., et al., 47th International SAMPE Symposium (2002) Glauser, T., Johansson, M., et al., Macromol. Mater. Eng. (2000) 274, p. 25 Eberle, C. C., Janke, C. J., et al., Cooperative research and development agreement – final report (2005) Vorobyev, L., Korenev, S., 32nd International SAMPE Technical Conference (2000) Goodman, D. L., Birx, D. L., et al., Proceedings of the 41st International SAMPE Symposium and Exhibition (1996) Ghosh, N. N., Palmese. G. R., Bull. Mater. Sci. (2005) 28(6), p. 603 Janke, C. J., Norris, R. E., et al., 42nd International SAMPE Symposium (1997) Fengmei, L., Jianwen, B., et al., Rad. Phys. Chem. (2002) 63, p. 557 Kerluke, D. R., Cheng, S., Proceedings of the 2nd annual automotive composites conference and exposition of the society of plastics engineers (2002) Chappas W. J., Devney, B. G., et al., Rad. Phys. Chem. (1999) 56, p. 417 Goodman, D. L., Byrne, C. A., 42nd International SAMPE Symposium and Exhibition (1997) Saunders, C. B., Lopata, V. J., et al., Radiat. Phys. Chem. (1995) 46(4–6), p. 991 Moon, S. W., Morgan, R. J., et al., J. Comp. Mat. (2004) 38 (5) Lopata, V. J., Chung, M., et al., 28th International SAMPE Technical Conference (1996) Glauser, T., Johansson, M., et al., Macromol. Mat. Eng. (2000) 274, p. 20 Norris, R. E., EB Curing of Composites Workshop (1996) E-Beam Services The international composites Expo (1999) Drobny, J. L., Rad. Tech. Polym. CRC Press (2003) Bradley, R., Radiation Technology Handbook (1984) Marcel Dekker, Inc., New York and Basel Berejka, A. J., Emerging applications of radiation processing – Proceedings of technical meeting held in Vienna (2003) International Atomic Energy Agency (1992) Safety Guidelines: Radiation Safety of Gamma and Electron Irradiation Facilities, VIC Library, Vienna, p. 9. Cantrell, R., September/October Radtech report (1999) p. 23 Odian, G., Principles of Polymerization (2004) 4th ed., John Wiley & Sons Inc., New York, Chap. 3, p. 198 Glauser, T., Johansson, M., et al., Polymer (1999) 40, p. 5297 Hayashi, K., Takezaki, J., et al., J. Appl. Polym. Sci. (1988) 36, p. 295 Allen, C. C., Oraby, W., et al., J. Appl. Polym. Sci. (1974) 18, p. 709 Squire, D. R., Cleaveland, J. A., et al., J. Appl. Polym. Sci. (1972) 16, p. 645 Fouassier, J. P., Photoinitiation, Photopolymerization, and Photocuring: Fundamentals and Applications (1995) Hanser, Munich

References

117

[85] Crivello, J. V., in Ring Opening Polymerization (1993) Ed.: Brunelle, D. J., Hanser, Munich, p. 157 [86] Fouassier, J. P., Rabek J. F., Radiation Curing in Polymer Science and Technology (1993) Chapman & Hall, London [87] Fouassier J. P., J. Photopolym. Sci. Technol. (1990) 3, p. 1 [88] Crivello, J. V., in Advances in Polymer Science (1984) Ed.: T. Saegusa, Springer, Berlin, Vol. 62, p. 1 [89] Pappas, S. P., UV-Curing: Science and Technology (1978) Technology Marketing Corporation, Stamford [90] Crivello, J. V., Radiat. Phys. Chem. (2002) 63, p. 21 [91] Crivello, J. V., Nucl. Instr. and Meth. in Phys. Res. B (1999) 151, p. 8 [92] Crivello, J. V., Fan, M., et al., J. Appl. Polym. Sci. (1992) 44, p. 9 [93] Crivello, J. V., US 5260349 (1992) to Polyset Corporation [94] Crivello, J. V., Walton, T. C., et al.,Chem. Mater. (1997) 9, p. 1273 [95] Yağci, Y., Reetz, I., Prog. Polym. Sci. (1998) 23, p. 1485 [96] Park, C. H., Takahara, S., et al., Polym. Adv. Technol. (2006) 17, p. 156 [97] Olah, G. A., Prakash, G. K. S., et al., Superacids (1985) John Wiley & Sons, Chichester [98] Olah, G. A., Prakash, G. K. S., et al., in Encyclopedia of Reagents for Organic Synthesis (2004), Ed.: L. Paquette, John Wiley & Sons, New York [99] Castellanos, F., Fouassier, J. P., et al., J. Appl. Polym. Sci. (1996) 60, p. 705 [100] Meier, K., Coord. Chem. Rev. (1991) 111, p. 97 [101] Barriau, E., Schmidt-Freytag, U., et al., Macromolecules (2008) 41, p. 3779 [102] Crivello, J. V., Lam, J. H. W., Macromolecules (1977) 10, p. 1307 [103] Banks, D. F., Chem. Rev. (1966) 66, p. 243 [104] Beringer, F. M., Falk, R. A., et al., J. Am. Chem. Soc. (1959) 81, p. 342 [105] Beringer, F. M., Drexler, M., et al., J. Am. Chem. Soc. (1953) 75, p. 2705 [106] Crivello, J. V., Lee, J. L., J. Polym. Sci., A: Polym. Chem. (1989) 27, p. 3951 [107] Neiland, O., Karele, B., J. Org. Chem. USSR (Engl. Transl.) (1970) 6, p. 889 [108] Koser, G. F., Wettach, R. H., et al., J. Org. Chem. (1980) 45, p. 1543 [109] Hartwig, A., Harder, A., et al., Eur. Polym. J. (2001) 37, p. 1449 [110] Crivello, J. V., Lam, J. H. W., J. Polym. Sci., Polym. Chem. Ed. (1979) 17, p. 3845 [111] Ichimura, K., J. Mater. Chem. (2007) 17, p. 632 [112] Crivello, J. V., Lam J. H. W., J. Polym. Sci., Polym. Chem. Ed. (1978) 16, p. 2441 [113] Crivello, J. V., J. Polym. Sci., A: Polym. Chem. (2006) 44, p. 3036 [114] Sipani, V., Kirsch, A., et al., J. Polym. Sci., A: Polym. Chem. (2004) 42, p. 4409 [115] Bulut, U., Crivello, J. V., Macromolecules (2005) 38, p. 3584 [116] Boey, F., Rath, S. K., et al., J. Appl. Polym. Sci. (2002) 86, p. 518 [117] Dektar, J. L., Hacker, N. P., J. Org. Chem. (1990) 55, p. 639 [118] Pappas, S. P., in Comprehensive Polymer Science Lastmond, G., Ledwith, A., Russo, S., Sigwalt, P., (Ed.) (1989) Pergamon Press, New York, 4, p. 337 [119] Pappas, S. P., Pappas, B. C., et al., J. Polym. Sci., Polym. Chem. Ed. (1984) 22, p. 69 [120] Ledwith, A., Polymer (1978) 19, p. 1217 [121] Bachofner, H. E., Beringer, F. M., et al., J. Am. Chem. Soc. (1958) 80, p. 4269 [122] Baumann, H., Timpe, H.-J., J. Prakt. Chem. (1984) 326, p. 529 [123] Batten, R. J., Davidson, R. S., et al., J. Photochem. Photobiol. A: Chem. (1991) 58, p. 115 [124] Buijsen, P. F. A., Research Thesis, Delft University of Technology (1996) February 5. [125] Nishitsuji, D. A., Marinucci, G., et al., Nucl. Instr. and Meth. in Phys. Res. B (2007) 265, p. 135 [126] Mascioni, M., Sands, J. M., et al., Nucl. Instr. and Meth. in Phys. Res. B (2003) 208, p. 353 [127] Zhang, Z.-G., Li, Y.-B., et al., J. Appl. Polym. Sci. (2004) 94, p. 2217 [128] Degrand, H., Cazaux, F., et al., Radiat. Phys. Chem. (2003) 68, p. 885 [129] Sui, G., Zhang, Z.-G., et al., Mater. Chem. Phys. (2002) 78, p. 349

118

References

[130] Zhang, Z., Li, Y., Acta Polym. Sin. (2006) 2, p. 284 [131] Zhang, Z., Liu, Y., et al., Compos. Sci. Technol. (2002) 62, p. 331 [132] Janke, C. J., Lopata, V. J., et al., U.S. Patent 5877229 (1999) [133] Janke, C. J., Dorsey, V. J., et al., U.S. Patent 5726216 (1998) [134] Janke, C. J., Lopata, V. J., et al., WO Patent 1997005172 (1997) [135] Crivello, J. V., in Advances in Polymer Science. Saegusa, T. (Ed.) (1984) Springer, Berlin, 62, p. 9 [136] Watt, W. R., Hoffman Jr., H. T., et al., J. Polym. Sci., Polym. Chem. (1984) 22, p. 1789 [137] Lowe, P. A., in The Chemistry of the Sulfonium Group. Stirling, C. J. M., Patai, S. (Ed.) (1981) John Wiley & Sons Ltd., Chichester-New York, p. 267 [138] Crivello, J. V., Lam, J. H. W., J. Polym. Sci., Polym. Chem. (1979) 17, p. 977 [139] Marino, J. P., in Topics in Sulfur Chemistry. Senning, A. (Ed.) (1976) Thieme, Stuttgart, p. 32 [140] Trost, B. M., Melvin Jr., L. S., Sulfur Ylides, Emerging Synthetic Intermediates (1975) Academic Press, London [141] Crivello, J. V., Lam, J. H. W., J. Org. Chem. (1978) 43, p. 3055 [142] Crivello, J. V., Lam, J. H. W., Synth. Commun. (1979) 9, p. 151 [143] Crivello, J. V., Lam, J. H. W., J. Polym. Sci., Polym. Chem. (1980) 18, p. 2677 [144] Crivello, J. V., Lam, J. H. W., J. Polym. Sci., Polym. Chem. (1980) 18, p. 2697 [145] Wildi, B. S., Taylor, S. W., et al., J. Am. Chem. Soc. (1951) 73, p. 1965 [146] Miller, R. D., Renaldo, A. F., et al., J. Org. Chem. (1988) 53, p. 5571 [147] Wiegand, G. H., McEwen, W. E., J. Org. Chem. (1968) 33, p. 2671 [148] Akhtar, S. R., Crivello, J. V., et al., J. Org. Chem. (1990) 55, p. 4222 [149] Badet, B., Julia, M., Tetrahedron Lett. (1979) 13, p. 1101 [150] Badet, B., Julia, M., et al., Bull. Soc. Chim. Fr. (1984) p. 431 [151] Pappas, S. P., Tilley, M. G., et al., J. Photochem. Photobiol. A (2003) 159, p. 161 [152] Kikkawa, A., Takata, T., et al., Makromol. Chem. (1991) 192, p. 655 [153] Hogsed, M. J., U.S. Patent 3418289 (1968) [156] Crivello, J. V., Lee, J. L., Macromolecules (1981) 14, p. 1141 [157] Crivello, J. V., Lam, J. H. W., J. Polym. Sci., Polym. Chem. (1980) 18, p. 1021 [158] Crivello, J. V., Lam, J. H. W., J. Polym. Sci., Polym. Chem. (1979) 17, p. 2877 [159] Crivello, J. V., Lee, J. L., Macromolecules (1983) 16, p. 864 [160] Chen, Z., Webster, D. C., J. Polym. Sci., A: Polym. Chem. (2006) 44, p. 4435 [161] Hartwig, A., Schneider, B., et al., Polymer (2002) 43, p. 4243 [162] Decker, C., Moussa, K., J. Polym. Sci., A: Polym. Chem. (1990) 28, p. 3429 [163] Dektar, J. L., Hacker, N. P., J. Chem. Soc., Chem. Commun. (1987) 46, 2, p. 1591 [164] Lin, M.-P., Kim, H.-B., et al., J. Polym. Sci., A: Polym. Chem. (1992) 30, p. 2365 [165] Saeva, F. D., J. Chem. Soc., Chem. Commun. (1987), p. 37 [166] Davidson, R. S., Wilkinson, S. A., J. Photochem. Photobiol. A: Chem. (1991) 58, p. 123 [167] Nho, Y. C., Kang, P. H., et al., Radiat. Phys. Chem. (2004) 71, p. 241 [168] Kumar, R. N., Mehnert, R., Macromol. Mater. Eng. (2001) 286, p. 449 [169] Boey, F. Y. C., Chia, N. K., et al., J. Appl. Polym. Sci. (2001) 82, p. 3099 [170] Crivello, J. V., Lam, J. H. W., Polym. Mat. Sci. Eng. (1991) 65, p. 319 [171] Olah, G. A., Sakakibara, T., et al., J. Org. Chem. (1978) 43, p. 463 [172] Dektar, J. L., Hacker, N. P., J. Org. Chem. (1991) 56, p. 1838 [173] Crivello, J. V., Lam, J. H. W., J. Polym. Sci. Polym. Lett. (1978) 16, p. 563 [174] Crivello, J. V., Lam, J. H. W., J. Polym. Sci., Polym. Chem. (1979) 17, p. 1047 [175] Müller, U., Utterodt, A., et al., J. Photochem. Photobiol., A (2001) 140, p. 53 [176] Itoh, H., Kameyama, A., et al., Macromolecules (1995) 28, p. 883 [177] Smets, G., Aerts, A., et al., Polym. J. (1980) 12, p. 539 [178] Schlesinger, S. I., Photogr. Sci. Eng. (1974) 18, p. 387 [179] Yağci, Y., Kornowski, A., et al., J. Polym. Sci., A: Polym. Chem. (1992) 30, p. 1987 [180] Park, S.-J., Heo, G.-Y., et al., J. Polym. Sci., A: Polym. Chem. (2003) 41, p. 2393

References

[181] Everett, J. P., Schmidt, D. L., et al., Polymer (1997) 38, p. 1719 [182] Stark, B. P., Irving, E., Ind. Chem. Bull. (1982) 1, p. 164 [183] Ledwith, A., Accounts Chem. Res. (1972) 5, p. 133 [184] Ledwith, A., Polym. Prepr. ACS (1982) 23, p. 323 [185] Ketley, A. D., Tsao, J.-H., J. Rad. Curing (1979) 6, p. 22 [186] Ketley, A. D., Tsao, J.-H., Polym. Prepr. ACS (1978) 19, p. 656 [187] Irving, E., European Patent 0104143 (1983) [188] Irving, E., European Patent 0106797 (1983) [189] Cella, J. A., U.S. Patent 4086091 (1977) [190] Aydogan, B., Gundogan, A. S., et al., Macromolecules (2008) 41, p. 3468 [191] Takeshi, O., Japanese Patent 2004143252 (2002) [192] Kenji, O., Tadashi, B., et al., Japanese Patent 07010893 (1993) [193] Abu-Abdoun, I. I., Aale-Ali, Eur. Polym. J. (1992) 28, p. 73 [194] Gupta, S. N., Thijs, L., et al., Macromolecules (1980) 13, p. 1037 [195] Campbell, T. W., McDonald, R. N., J. Org. Chem. (1959) 24, p. 1246 [196] Park, S. J., Heo, G. Y., et al., J. Polym. Sci. A: Polym. Chem. (2003) 41, p. 2393 [197] Toneri, T., Sanda, F., et al., Macromolecules (2001) 34, p. 1518 [198] Takuma, K., Takata, T., et al., Macromolecules (1993) 26, p. 862 [199] Takata, T., Takuma, K., et al., Makromol. Chem., Rapid Commun. (1993) 14, p. 203 [200] Abu-Abdoun, I. I., Aale-Ali, Eur. Polym. J. (1993) 29, p. 1439 [201] Abu-Abdoun, I. I., Aale-Ali, Eur. Polym. J. (1993) 29, p. 1445 [202] Neckers, D. C., Abu-Abdoun, I. I., Macromolecules (1984) 17, p. 2468 [203] Crivello, J. V., U.S. Patent 4250311 (1980) [204] Zhang, X.-M., Bordwell, F. G., J. Am. Chem. Soc. (1994) 116, p. 968 [205] Arnold, U., unpublished results. [206] Klingert, B., Riediker, M., et al., Comments Inorg. Chem. (1988) 7, p. 109 [207] Lohse, F., Zweifel, H., Adv. Polym. Sci. (1986) 78, p. 61 [208] Palazzotto, M. C., Hendrickson, W. A., European Patent 0109851 (1983) [209] Palazzotto, M. C., U.S. Patent 5089536 (1982) [210] Strohmeier, W., Barbeau, C., Makromol. Chem. (1965) 81, p. 86 [211] Anderson, W. S., U.S. Patent 3709861 (1970) [212] Degirmenci, M., Onen, A., et al., Polym. Bull. (2001) 46, p. 443 [213] Riediker, M., Roth, M., et al., European Patent 0122223 (1984) [214] Kaeriyama, K., Makromol. Chem. (1972) 153, p. 229 [215] Asano, S., Aida, T., et al., J. Chem. Soc, Chem. Commun. (1985) 17, p. 1148 [216] Hayase, S., Onishi, Y., et al., Macromolecules (1985) 18, p. 1799 [217] Onishi, Y., Hayase, S., Japanese Patent 59043018 (1982) [218] Woodhouse, M. E., Lewis, F. D., et al., J. Am. Chem. Soc. (1982) 104, p. 5586 [219] Meleshevich, A. P., Kozlov, A. A., Teor. Eksp. Khim. (1984) 20, p. 492 [220] Kaeriyama, K., J. Polym. Sci., Polym. Chem. (1976) 14, p. 1547 [221] Irving, E., Johnson, B. F. G., et al., European Patent 0094914 (1983) [222] Meier, K., Bühler, N., et al., European Patent 0094915 (1983) [223] Meier, K., Eugster, G., et al., European Patent 0126712 (1984) [224] Doggweiler, H. O., Desobry, V., European Patent 0270490 (1987) [225] Schumann, H., Chem. Ztg. (1984) 108, p. 239 [226] Nesmeyanov, A. N., Vol’kenau, N. A., et al., Tetrahedron Lett. (1963) 4, p. 1725 [227] Mahoney, W. S., Willett, P. S., et al., U.S. Patent 6133335 (1998) [228] Artiga-Gonzàlez, R.-A., Nicolaisen, H.-C., et al., European Patent 0340591, (1989) [229] Bühler, N., Giller, G., European Patent 0182744 (1985) [230] Chrisope, D. R., Park, K. M., et al., J. Am. Chem. Soc. (1989) 111, p. 6195 [231] Roloff, A., Meier, K., et al., Pure Appl. Chem. (1986) 58, p. 1267

119

120

References

[232] Meier, K., Zweifel, H., J. Imaging Sci. (1986) 30, p. 174 [233] Meier, K., Zweifel, H., J. Radiat. Curing (1986) 13, p. 26 [234] Hoene, R., Reichert, K.-H. W., Makromol. Chem. (1976) 177, p. 3545 [235] Andruzzi, F., Pescia, A., et al., Makromol. Chem. (1975) 176, p. 977 [236] Astruc, D., Tetrahedron (1983) 39, p. 4027 [237] Canty, A. J., Colton, R., Inorg. Chim. Acta (1994) 220, p. 99 [238] Masuda, H., Munakata, M., et al., J. Organomet. Chem. (1990) 391, p. 131 [239] Albinati, A., Meille, S. V., et al., J. Organomet. Chem. (1979) 182, p. 269 [240] Lewandos, G. S., Gregston, D. K., et al., J. Organomet. Chem. (1976) 118, p. 363 [241] Quinn, H. W., VanGilder, R. L., Can. J. Chem. (1970) 48, p. 2435 [242] Bressan, G., Broggi, R., et al., J. Organomet. Chem. (1967) 9, p. 355 [243] Arnold, U., unpublished results. [244] Handbook of Chemistry and Physics (2006) Ed.: Lide, D. R., CRC, 87th Ed. [245] Hoyle, C. E., Kinstle, J. F., Radiation Curing of Polymeric Materials, (1992) ACS Symposium Series, American Chemical Society, NY [246] Pappas, S. P., Radiation Curing: Science and Technology (1992) Ed.: Pappas, S. P., Plenum Press, NY [247] Crivello, J. V., Lee, J. L., et al., Journal of Radiation Curing (1983) 10 (1), p. 6 [248] Janke, C. J., Lomax, R. D., et al., Conference SAMPE 2001 (2001) Long Beach, CA, May 6-10 [249] Janke, C. J., Dorsey, G. F., et al., Proceedings of the 28th international SAMPE Technical Conference (1996) Seattle, USA [250] Boogh, L., Pettersson, B., et al., Polymer (1999) 40, p. 2249 [251] Wolff Fabris, F., Toughness Modification of Electron Beam Curable Epoxy Matrices for Carbon Fibre Composites (2010) ISBN 978-3-941492-11-0 [252] Defoort, B., Drzal. L. T., 46th International SAMPE symposium (2001), Long Beach, USA [253] Drzal, L. T., Sugiura, N., et al., Composites Interfaces (1999) 4, p. 337 [254] Thomas, Y., Parisi, J. P., et al., Polymer Bulletin (1992) 29, p. 259 [255] Defoort, B., Drzal L. T., 33rd International SAMPE technical conference (2001), Seattle, USA [256] Bechel, V., Janke, C. J., et al., 46th International SAMPE Symposium (2001), Long Beach, USA [257] Raghavan, J., 46th International SAMPE Symposium (2001), Long Beach, USA [258] Zhang, Z. G., Sui, G., et al., Polymer and Polymer Composites (2002) 10, p. 6 [259] Defoort, B., Boursereau, F., et al., RADTECH Europe Conference 2003 (2003) Berlin, Germany [260] Goodman, D. L., Byrne, C. A., et al., 31st International Technical SAMPE Conference (1999) Long Beach, USA [261] Bader, M. G., Composites: A (2002) 33, p. 913 [262] Farmer, J. D., Warnock, R. B., et al., 43rd International SAMPE Symposium (1998), Anaheim, USA [263] Ivanov, V. S., Radiation Chemistry of Polymers (1992) Utrecht, The Netherlands: VSP BV [264] Parrot, P., His, S., et al., 44th International SAMPE Symposium (1999) Long Beach, USA [265] Defoort, B., Lopitaux, G., et al., Macromol. Chem. Phys. (2001) 202, p. 3149 [266] Defoort, B., Defoort, D., et al., Macrol. Theory Simul. (2000) 9, p. 725 [267] Johnston, A., Cole, K. C., et al., 49th International SAMPE Symposium (2004) Ottawa, Canada [268] Crivello, J. V., Jang, M., J. Photochem. Photobio. A: Chem. (2003), 159, p. 173 [269] Final Project report – Oak Ridge Y-12 plant managed by Lockheed Martin Energy Systems, Inc. for the US Department of Energy CRADA (1998) [270] Johnson, M. A., Radtech Report, (2001) November/December, p. 41 [271] Wolff-Fabris, F., Klophaus, K., et al., German registered patent (2009) [272] Döring, M., Arnold, U., et al., German Patent DE 10 2007 041 988 (2009) [273] Döring, M., Arnold, U., et al., World Patent WO 2008/064959 (2008) [274] Krebs, M., Modern Plastics (1997) October, p. 17 [275] Saunders, C., IBA workshop (2002)

References

121

[276] Farmer, J. D., Janke, C. J., et al., The electron beam cure of fibreglass/epoxy prepregs – final report USAF Advanced composites program office, Oak Ridge National Laboratory and Atomic Energy of Canada Ltd [277] Taggart, D., Sidwell, D., Integrated Airframe Technology for Affordability (IATA) – Final Report (1996) [278] Wilkins, D., Eisenmann, J., et al., Damage in Composite Materials (1980) ASTM STP 775:168. [279] DIN-EN-6033 Carbon fibre reinforced plastics – Test method – Determination of interlaminar fracture toughness energy – Mode I – GIc, DIN – Deutsches Institut für Normung e.V. (1996) [280] DIN-EN-6034 Carbon fibre reinforced plastics – Test method – Determination of interlaminar fracture toughness energy – Mode II – GIIc, Deutsches Institut für Normung e.V. (1996) [281] Altstädt, V., Gerth, D., et al., Polymer (1993) 34(4), p. 907 [282] Drake, R., J. Polym. Mat. Sci. Eng. (1990) 63, p. 802 [283] Pascault, J.-P., Macromol. Symp. (1995) 93, p. 43 [284] Sultan, J., FJ, M., Polym. Eng. Sci. (1973) 13, p. 29 [285] Kunz, S., Sayre, J., et al., Polymer (1982) 23, p. 1897 [286] Chen, D., Pascault, J.-P., et al., Polym. Int. (1994) 33, p. 253 [287] Kinloch, A., Rubber-toughened plastics: Advances in Chemistry Series 222 (1989) Washington, USA [288] Shvetsov, O. Acta Polymerica (1986) 37, p. 573 [289] Liu, S., Nauman, E., J. Mat. Sci. (1991) 26, p. 6581 [290] Nakamura, Y., Tabata, H., et al., J. Appl. Polym. Sci. (1986) 32, p. 4865 [291] Qian, J., Pearson, R., et al., Polymer (1997) 38, p. 21 [292] Ashida, T., Katoh, A., et al., J. Appl. Polym. Sci. (1999) 74, p. 2955 [293] Maazouz, A., Sautereau, H., et al., Polym. Bull. (1994) 33, p. 67 [294] Bécu-Longuet, L., Bonnet, A., et al., J. Appl. Polym. Sci. (1999) 72, p. 849 [295] Schaefer, O., Nordic Rubber Conference (2007) Sandefjord, Norway [296] Bucknall, C., Polym. Eng. Sci. (1986) 25, p. 54 [297] Di Pasquale, G., Motta, O., et al., Polymer (1997) 38, p. 4345 [298] Mimura, K., Ito, H., et al., Polymer (2000) 41, p. 4451 [299] Andres, M., Garmenia, J., et al., J. Appl. Polym. Sci. (1998) 69, p. 183 [300] Varley, R., Hondgkin, J., et al., J. Appl. Polym. Sci. (2000) 77, p. 237 [301] Varley, R., Hodgkin, J., et al., Polymer (2001) 42, p. 3847 [302] Oyanguren, P., Galante, M., et al., Polymer (1999) 40, p. 5249 [303] Huang, P., Zheng, S., et al., Polymer (1997) 38, p. 5565 [304] Min, B., Stachurski, Z., et al., J. Appl. Polym. Sci. (1993) 50, p. 1511 [305] Su, C., Woo, E., Polymer (1995) 36, p. 2883 [306] Recker, H., Altstädt, V., et al., SAMPE J. (1990) 26, p. 73 [307] Girard-Reydet, E., Vicard, V., et al., J. Appl. Polym. Sci. (1997) 65, p. 2433 [308] Bonnet, A., Camberlin, Y., et al., Macromol. Symp. (2000) 149, p. 145 [309] Williams, R. J. J., Rozemberg, B., et al., Adv. Polym. Sci. (1997) 128, p. 95 [310] Riccardi, C., Borrajo, J., et al., J. Polym. Sci., B: Polym. Phys. (1996) 34, p. 349 [311] Pascault, J.-P., Williams, R. J. J., Formulation and Characterization of Thermoset-Thermoplastic blends (2000) John Wiley & Sons, New York, USA [312] MacKinnon, A., Jenkins, S., et al., Macromol. (1992) 25, p. 3492 [313] Cho, J., Hwang, J., et al., Polymer (1993) 34, p. 4832 [314] Bucknall, C., Partridge, I., Polymer (1983) 24, p. 639 [315] Yee, A., Du, J., et al., Toughening of Epoxies (2000) John Wiley & Sons, New York, USA [316] Kinloch, A., Yuen, M., et al., J. Mat. Sci. (1994) 29, p. 3781 [317] Treolar, L. The physics of rubber elasticity (1958), Oxford, England [318] Wolff-Fabris, F., Beier, U., et al., 28th SAMPE Europe International Conference (2007) Paris, France

122

References

[319] Ruckdäschel, H., Sandler, J., et al., Polymer (2007) 48(9), p. 2700 [320] Finaz, G., Skoulios, A., et al., Real Academy of Sciences of Paris (1967) 253, p. 565 [321] Hillmyer, M., Lipic, O., et al., Journal of the American Chemical Society (1997) 119, p. 2749 [322] Grubbs, R., Dean, J., et al., Macromol. (2000) 33, p. 3046 [323] Mijovic, J., Shen, M., et al., Macromol. (2000) 33, p. 5235 [324] Kosonen, H., Ruokolainen, J., et al., Macromol. (2001) 34, p. 3046 [325] Ritzenthaler, S., Court, F., et al., Macromol. (2002) 35, p. 6245 [326] Ritzenthaler, S., Macromol. (2003) 36, p. 118 [327] Zheng, W., Wane, Z., Macromol. (1996) 29, p. 18 [328] Hydro, R., Pearson, R., J. Polym. Sci., B: Polym. Phys. (2007) 45, p. 1470 [329] Wolff-Fabris, F., Toughness modification of electron beam curable epoxy matrices for carbon fibre composites (2009) ISBN 978-3-941492-11-0 [330] Spanoudakis, J., Young, R., J. Mater. Sci. (1984) 19, p. 473 [331] Amdouni, N., Sautereau, H., et al., J. Appl. Polym. Sci. (1992), 46, p. 1723 [332] Moloney, A., Kausch, H., et al., J. Mater. Sci. (1983) 18, p. 208 [333] Nakamura, Y., Yamaguchi, M., et al., Polymer (1991) 32, p. 2976 [334] Nakamura, Y., J. Appl. Polym. Sci. (1992), 45, p. 1281 [335] Kinloch, A., Taylor, A., et al., J. Adhes. (2003) 79(8-9), p. 867 [336] Kinloch, A., Mohammed, R., et al., J. Mater. Sci. (2005) 40(18), p. 5083 [337] Ragosta, G., Abbate, M., et al., Polymer (2005) 46(23), p. 10506 [338] Zhang, H., Zhang, Z., et al., Acta Materialia (2006) 54(7), p. 1833 [339] Lange, F., Philosophical Magazine (1970) 22, p. 983 [340] Kobayashi, T., Strength and Toughness of Materials (2004) Springer Verlag, Tokyo, Japan [341] Igakura, S., Yamamoto, I., et al., J. Mass Spect. Soc. Japan (1992) 29, p. 307 [342] Lee, J., Yee, A., Polymer (2001) 42, p. 589 [343] Johnsen, B., Kinloch, A., et al., Polymer (2007) 48, p. 530

Appendix 1: Key Players in Electron Beam Curing Industrial Groups Aerospace Industry yy Lockheed Martin Corporation, USA yy Boeing Company, USA yy Northrop Grumman Corporation, USA yy Acsion Industries, Canada yy Air Canada, Canada yy Acetek Composites Incorporated, Canada yy EADS Launch Vehicles, France Electron Beam Irradiation yy STERIS Corporation, USA yy E-Beam Services, Inc., USA yy YLA, Inc, USA yy IR-RADEX Co., Hungary yy BGS Beta-Gamma-Service GmbH & Co., Germany yy Laben-Proel Tecnologie Division, Italy Electron Beam Curable Materials yy Applied Poleramic Inc, USA yy Adherent Technologies Inc., USA yy Cytec Industries Inc., USA yy Hexcel Corporation, USA yy Toho Tenax Europe GmbH, Germany yy Henkel AG & Co. KGaA, Germany Research Institutes yy Oak Ridge National Laboratory, Tennessee, USA yy Oak Ridge Centre for Composite Manufacturing Technology, Oak Ridge, USA yy Laboratory for Research on Electron Beam Curing of Composites, Center for Basic and Applied Polymer Research, University of Dayton, Dayton, USA yy Kent State University, Kent, USA

124

Appendix 1: Key Players in Electron Beam Curing

yy Department of chemistry, New York Center for Polymer Synthesis, Rensselaer Polytechnic Institute, Troy, USA yy National Institute of Standards and Technology, Gaithersburg, USA yy AECL (Atomic Energy of Canada Limited), Canada yy Department of Polymer Technology, Royal Institute of Technology, Stockholm, Sweden yy Laboratoire de chimie organique et macromoléculaire, Université des sciences et technologies de Lille, France yy ISOF-CNR – Institute for the Organic Synthesis and Photoreactivity, Bologne, Italy yy Department of Chemical Engineering Processing and Materials, Interdepartmental Research Centre for Composite Materials, University of Palermo, Italy yy Fraunhofer IZM (Institut Zuverlässigkeit und Mikrointegration), Teltow, Germany yy Forschungszentrum Karlsruhe, Karlsruhe, Germany yy Department of Polymer Engineering, University of Bayreuth, Germany yy Institut für Oberflächenmodifizierung e. V. (IOM), Leipzig, Germany yy School of Materials Science and Engineering, Beijing University of Aeronautics and Astronautics, Beijing, China

Appendix 2: List of Commercially Available Materials for Electron Beam Curing Initiators for Epoxy Resins yy Rhodorsil 2074, diaryl-iodonium salt, from Rhodia yy Cyracure UVI 6990, triaryl-sulfonium hexafluorophosphate, from Union Carbide Chemicals and Plastics Co, UCAR Coatings Resins. yy Cyracure UVI 6974, triaryl-sulfonium hexafluoroantimonate, from Union Carbide Chemicals and Plastics Co, UCAR Coatings Resins yy DEGACURE KI-85, triaryl-sulfonium hexafluorophosphate, from Degussa Corporation yy FX-512, triaryl-sulfonium hexafluorophosphate, from 3M Corporation yy CD1012, diaryl-iodonium hexafluoroantimonate from Sartomer Co, Inc. yy OPPI, phenyl-iodonium hexafluoroantimonate, from General Electric Silicones Resins As mentioned in Section 2.2, the resin monomer suitable for electron beam curing may be virtually any of the commercially available materials. In this list, uniquely the resins already successfully cured by electron beam irradiation in research projects are reported: yy yy yy yy yy yy yy yy yy yy

DERAKANE D 411, epoxy-acrylate (vinylester), from Dow Chemicals Tactix 123, diglycidyl ether of bisphenol A, from Dow Chemicals DER332, bisphenol A diglycidyl ether, from Dow Chemicals DEN431, novolac epoxy, from Dow Chemicals DEN438, novolac epoxy, from Dow Chemicals ERL 4221E, difunctional cycloaliphatic epoxide monomer, from Union Carbide Asia Pacific, Inc. Epon 828, diglycidyl ether of bisphenol A, from Shell Chemical Epon 862, diglycidyl ether of bisphenol F, from Shell Chemicals Ebecryl 600 EB600, acrylate functionalized bisphenol-A, from UCB Belgium Ebecryl 610 EB610, methacrylate functionalized bisphenol-A, from UCB Belgium

Subject Index A accelerator –– electron 94, 111 –– electron beam 13, 94, 99 acid –– Brønsted 17, 19, 31, 32, 33, 34, 36, 38 –– Lewis 4, 17 acrylate 3, 41, 42, 43, 96, 102 –– epoxy 25, 110 activated chain end 4 activated monomer mechanism 4 adhesion 64, 68, 71, 78, 79, 82, 85, 96, 98, 103, 111 –– interfacial 48 agent, coupling 78, 102, 111 aramid fiber 77 automated fiber placement 106 automated tape placement 86 B bag, vacuum 88, 108, 110 banding, shear 47 block co-polymers 67 Brønsted 4, 38 Brønsted acid 17, 19, 31, 32, 33, 34, 36, 38 C carbon 113 carbon fiber 11, 77, 83, 89, 102 cationic 43 –– initiator 17, 43 –– mechanism 2 –– polymerization 2, 4, 5, 8, 9, 17, 21, 23, 24, 36, 37 cavitation 47, 56, 68 chain end, activated 4 coating, surface 13 complex –– metal 37 –– silver 39, 40 –– silver olefin 37 –– silver salt 47, 103 content, void 96

co-polymer, block 67 core-shell particle 49 core-shell rubber 48, 62 costs, manufacturing 93 coupling agent 78, 102, 111 crack deflection 57, 61, 64, 66, 74 crack deviation 64, 66, 68, 76 crack pinning 64, 68, 72, 74, 76 crazing 64 crosslinking degree 40, 45, 49, 54 crosslinking density 46, 62, 82 curing, UV 17 D dark reaction 6 debonding 76 deflection, crack 57, 61, 64, 66, 74 degree of crosslinking 40, 45, 49, 54 delamination 45 density, crosslinking 46, 62, 82 depth, penetration 12 deviation, crack 64, 66, 68, 76 dose 6, 9, 26, 34, 45, 79, 80, 81, 82, 87, 90, 101 –– irradiation 8 E electron accelerator 94, 111 electron beam accelerator 11, 13, 94, 99 energy, irradiation 11 epoxy 4, 9, 34, 37, 40, 44, 47, 89, 96, 102 –– acrylate 25, 110 –– methacrylated 110 ester, vinyl 90, 102 exothermic 44, 80, 101 –– reaction 40 F fiber 11, 87 –– aramid 77 –– carbon 11, 77, 83, 89, 102 –– glass 11, 77, 87 –– organic 77 –– placement, automated 106

128

Subject Index

fiber –– surface 26 –– surface treatment 102 –– treatment 98 filament winding 89, 99 fracture toughness 40, 45, 54, 62, 68, 70, 72, 75 free radical 2, 17, 41 –– polymerization 2, 3, 8 G gelation 56, 58 glass 89, 105 –– fiber 11, 77, 87 –– transition temperature 6, 9, 43, 45, 46, 49, 56, 62, 68, 81, 95, 101 –– – high 43 H horn, scan 90 hyperbranched polymer 76 I impurity 7, 26, 34, 49, 59, 74 initiator 9, 17, 81, 82, 87, 103 –– cationic 17, 43 –– silver complex 49 –– silver salt 54, 74 interfacial 101, 102 –– adhesion 48 –– property 77 internal stress 112 iodonium 19, 20, 21, 25, 26 ionic polymerization 17 irradiation dose 8 irradiation energy 11 L latency 19 layer-by-layer 85, 109 lay-up 110 Lewis acid 4, 17 light reaction 6 liquid reactive rubber 47 lost core molding 91 M manufacturing costs 93 mechanism, cationic 2

metal complex 37 methacrylate 3, 41, 42, 43, 102 methacrylated epoxy 110 mold 85, 90, 98 molding 101 –– lost core 91 monomer mechanism, activated 4 N nanoparticle 72 network 6, 101 O onium 20, 34, 35, 38, 40, 102 –– salt 17 organic fiber 77 P particle, core-shell 49 penetration depth 12 phase separation 56, 57, 58, 61, 62, 63, 67, 70, 76 photoinitiator 17, 19 photolysis 35 photosensitizer 102 pinning, crack 64, 68, 74, 76 polymer, hyperbranched 76 polymerization –– cationic 2, 4, 5, 8, 9, 17, 21, 23, 24, 36, 37 –– free radical 2, 3, 8 –– ionic 17 post-curing 9, 79, 81 prepreg 87, 90, 93, 103, 110 processing, radiation 13 property, interfacial 77 pultrusion 89, 99 R radiation processing 13 radiation-shielding 14 radical, free 2, 17, 41 radioactive 13 reaction –– dark 6 –– exothermic 40 –– light 6 repair 99, 107, 108 residual stress 80, 95, 96

Subject Index

rise, temperature 9, 10, 12 RTM 90, 111 rubber –– core-shell 48, 62 –– elasticity, theory 58 –– liquid reactive 47 S safety 15 salt –– complex silver 103 –– onium 17 scan horn 14, 90 separation, phase 56, 57, 58, 62, 63, 67, 70, 76 shear banding 47 shear yielding 47, 53, 54, 56 shelf life 43, 93 shielding 15, 94, 95 shrinkage 43, 79, 80, 96 silver complex 39, 40 silver complex initiator 49 silver olefin complex 37 silver salt complex 47, 103 silver salt initiator 54, 74 sizing 77, 78, 98, 102, 111 solvent 76 stability, thermal 19 stress –– internal 112 –– residual 80, 95, 96 sulfonium 19 surface coating 13 surface, fiber 26 T tape placement, automated 86

129

temperature –– glass transition 6, 9, 43, 45, 46, 49, 56, 62, 68, 81, 95, 101 –– high glass transition 43 –– rise 9, 10, 12 theory of rubber elasticity 58 thermal stability 19 thermoplastics 56 toughening 40, 48, 57, 64, 72, 101 toughness 45, 77, 95, 103 –– fracture 40, 45, 54, 62, 68, 70, 72, 75 treatment –– fiber 98 –– fiber surface 102 U UV 5, 23, 26, 108 UV-curing 17, 98 V vacuum bag 88, 108, 110 VARTM 90, 106 vinyl ester 90, 102 viscosity 10, 46, 76, 87, 90 vitrification 6, 56, 58, 101 void 67, 76 –– content 96 W winding, filament 89, 99 X X-ray 15, 89 Y yielding, shear 53, 56

Wolff-Fabris · Altstädt · Arnold · Döring

Electron Beam Curing of Composites Electron beam curing technology for advanced composites has emerged as a credible and attractive alternative to thermal curing for a number of high performance composite products. Technical advantages, e.g., for aerospace structures, include fast curing at room temperature, using low-cost tooling, and the ability to fabricate large integrated structures. Here, both theoretical and practical aspects of electron beam curing of composites are comprehensively covered, with the intention to bridge the gap between academic knowledge and industrial applications. The reader is introduced to fundamental information regarding curing reactions, material requirements, toughness modification and parameters affecting the curing process, and the analysis of current molding technologies combined with EB curing. A description of research projects and the main topics they address as well as examples of EB curing applications within the composite industry are also presented. The Content:  Introduction to Electron Beam Curing of Composites  Aspects of Electron Beam Curable Materials  Electron Beam Curing Applied to Composite Molding Technologies  Current Limitations and Potentials for Electron Beam Curing  Research Trends and Projects in the Field of Electron Beam Curing  Examples of Electron Beam Curing Applications

www.hanserpublications.com Hanser Publications ISBN 978-1-56990-473-2 www.hanser.de Carl Hanser Verlag ISBN 978-3-446-42405-0

9

781569 904732