Epoxy Resin Chemistry Volume I [1] 0841205256, 9780841205253

Citation preview

Digitized by the Internet Archive in 2022 with funding from Kahle/Austin Foundation

https://archive.org/details/epoxyresinchemisOO0Ounse_i5u9

Epoxy Resin Chemistry

: _ razed

rps

|

—

ae

: cae -

255 ae ——

a

ee

=

a gine at i

a

oF

“y

2

7

= BS

ome

Epoxy Resin Chemistry Ronald S. Bauer, Eprror Shell Development Company

Based on a symposium sponsored

by the Division of Organic Coatings and Plastics

at the 176th Meeting of the American Chemical Society, Miami Beach, Florida,

September 11—15, 1978.

Am Ces meow aMePnOusals USM

AMERICAN

CHEMICAL

WASHINGTON,

D.C.

Seiad

SOCIETY 1979

WK!

Library of Congress CIP Data Epoxy resin chemistry. (ACS symposium series; 114 ISSN 0097-6156) Includes bibliographies and index.

1. Epoxy resins—Congresses. I. Bauer, Ronald S.,1932— _. -I. American Chemical Society. Division of Organic Coatings and Plastics Chemistry. III. Series: American Chemical Society. ACS symposium series; 114.

TP1180.EGES9 ISBN 0-8412-0525-6

668.374 ASCMC8

79-21858 114 1-271 1979

Copyright © 1979 American Chemical Society

All Rights Reserved. The appearance of the code at the bottom of the first page of each article in this volume indicates the copyright owner’s consent that reprographic copies of the article may be made for personal or internal use or for the personal or internal use of specific clients. This consent is given on the condition, however, that the copier pay the stated per copy fee through the Copyright Clearance Center, Inc. for copying beyond that permitted by Sections 107 or 108 of the U.S. Copyright Law. This consent does not extend to copying or transmission by any means—graphic or electronic—for any other purpose, such as for general distribution, for advertising or promotional purposes, for creating new collective works, for resale, or for information storage and retrieval systems.

The citation of trade names and/or names of manufacturers construed as an endorsement or as approval by ACS of the referenced herein; nor should the mere reference herein chemical process, or other data be regarded as a license or

in this publication is not to be commercial products or services to any drawing, specification, as a conveyance of any right or

permission, to the holder, reader, or any other person or corporation, to manufacture, repro-

duce, use, or sell any patented invention or copyrighted work that may in any way be related thereto. PRINTED

IN THE UNITED

“alle

STATES OF AMERICA

Bins

ACS Symposium Series M. Joan Comstock,

Series Editor

Advisory Board Kenneth B. Bischoff

James P. Lodge

Donald G. Crosby

John L. Margrave

Robert E. Feeney

Leon Petrakis

Jeremiah P. Freeman

F, Sherwood Rowland

E. Desmond

Alan C. Sartorelli

Goddard

Jack Halpern

Raymond B. Seymour

Robert A. Hofstader

Aaron Wold

Jamese De Idole) tr:

Gunter Zweig

FOREWORD The ACS Symposium Series was founded in 1974 to provide a medium for publishing symposia quickly in book form. The format of the Series parallels that of the continuing ADVANCES IN CHEMISTRY SERIES except that in order to save time the papers are not typeset but are reproduced as they are submitted by the authors in camera-ready form. Papers are reviewed under the supervision of the Editors with the assistance of the Series Advisory Board and are selected to maintain the integrity of the symposia; however, verbatim reproductions of previously published papers are not accepted. Both reviews and reports of research are acceptable since symposia may embrace both types of presentation.

CONTENTS UATE

Sic decors EM

Sd

AO

ie

ix

1. The Photoinitiated Cationic Polymerization of Epoxy Resins J. V. Crivello and J. H. W. Lam

....

2.

Photosensitized Epoxides as a Basis for Light-Curable Coatings W. R. Watt

3.

Quaternary Phosphonium Compound Latent Accelerators for Anhydride-Curedi H.poxvenesins mee enh ne coctn ed eet oy J. D. B. Smith

47

4, Development of Epoxy Resin-Based Binders for Electrodeposition Coatings with High Corrosion Resistance ................008W. Raudenbusch

57

5. Water-Borne Coatings Prepared from High Molecular Weight Mine ae coe Moke eG EPOXY ARINGSIS me wee iota ee eles Chen eet E. G. Bozzi and C. Y. Zendig

71

.

6.

Aqueous Epoxy Resins for Electrical Reinforced Plastics Industry . . T. L. Anderson, J. H. Melloan, K. L. Powell, R. R. McClain, J. B. Simons, R. L. Conway, and D. A. Shimp

7.

Epoxide Equivalent Weight Determination by Carbon-13 INuglearsMacnetic: Resonances cat. r a0 oe: cer se) yoeneee 24 2 we sn W. B. Moniz and C. F. Poranski

77

83

8. An Instrumental Method for Measuring the Sterilization Resistances

(Blushin:)

softs @any Coatingsmrr

aera

ocieieietercis

W. Raudenbusch, H. C. J. Venselaar, and D. M. Paul 9.

Phenalkamines—A New Class of Epoxy Curing Agents ......... R. A. Gardiner and A. P. Manzara

10.

The Effect of Alkyl Substituents on the Properties of Cured Hydantom r poxy Resins, so a ets sw nen es hae as weet. E. H. Catsiff, R. E. Coulehan, J. F. DiPrima, D. A. Gordon, and R. Seltzer

11.

Effect of Cross-Link Density Distribution on the Engineering Beh AVIOLI OLA EeONICS ote ier ote te oveais recy we cone a ele iy aie.vires S. C. Misra, J. A. Manson, and L. H. Sperling

12.

Network Morphology and the Mechanical Behavior of Epoxies . . S. C. Misra, J. A. Manson, and L. H. Sperling

13.

Creep Behavior of Amine-Cured Epoxy Networks: Effect of le eortakerania mus ocilea acs 6)8 pqs oy etter sone GUDICHIONDCLE VE S. L. Kim, J. A. Manson, and S. C. Misra Vil

91

14,

Self-Cross-Linkable Polyepoxides Y. Tanaka

15.

Some Studies on the Preparation of Glycidy] 2-Ethylhexanoate D. A. Cornforth, B. G. Cowburn, K. M. Smith, C. W. Stephens, and C. G. Tilley

....

211

16.

Reactions in a Typical Epoxy—Aliphatic Diamine System ......... J. J. King and J. P. Bell

225

17.

Epoxy Resins Containing a Specific Vulnerability J. R. Griffith

259

Tril@x O98 oss

..............ececeseecvecee

..............

sors. wensdue,poet we elesUreong esente eerie 6 (eaereia cera aero

Vili

Gita e eee

197

265

PREFACE See their commercial introduction into the United States in the late 1940's, epoxy resins have established themselves as unique building blocks for high performance coatings, adhesives, and reinforced plastics. Over the past five years epoxy resin sales have experienced an annual growth rate of about 8%. In 1978, production in this country exceeded 310 million pounds, and domestic sales were greater than 280 million pounds. This sustained growth of epoxy resins is a result of the wide range of properties that can be achieved with these versatile materials; uses range from coatings for the ubiquitous beverage can to high performance adhesions in spacecraft as well as for encapsulating their sophisticated electronic components. Work still continues as pressures for more environmentally acceptable and energy efficient systems, and demands for even higher performance, challenge the polymer chemist. This volume covers some of the new resins, curing agents, and application techniques along with new ideas about the chemistry and properties of existing epoxy resin systems. Although comprehensive coverage of this field is not possible in our symposium, it is hoped the papers presented here will help stimulate polymer chemists to answer the challenges facing them. Shell Development Company Houston, Texas 77001

May 21, 1979

RONALD S. BAUER

f

=

(24)

Sertins

ed

ae

«9

Frade

oem

Vy -« froerd)

See

APE

aa

odes ‘about urd sos “mepr-t

age

Lanes Worend Plast Preprinitse. LOG 5) oye Geen RECEIVED May 21, 1979.

Resin York,

cen N

wemclomicds

mange stang

yma

9.

WVolantes

Cea Nes

Ctgs.

and

Photosensitized Epoxides as a Basis for Light-Curable Coatings WILLIAM

R. WATT

Princeton Research Center, American Can Company, Princeton, NJ 08540

Coatings which harden or "cure" by photoinduced polymerization offer a way to reduce air pollution and to conserve energy. With such coatings, the change from liquid to solid is accomplished by an increase in molecular weight rather than by removal of solvent. Although there has been increasing interest in light-curable coatings over the past decade, most published investigations deal with coatings based on free-radical polymerization. Much less has been written about light-curable coatings based on ionic polymerization; and what has been written is of relatively recent origin. One of the earliest descriptions of this subject was a paper presented by the author in 1974 describing photosensitized epoxide coatings which cure by cationic polymerization on exposure to ultraviolet radiation (1). Since that time, several technical papers concerned with coatings based on photoinduced cationic

polymerization Photoinitiators

have appeared for

Cationic

(2-11). Polymerization

of Epoxides

The development of light-curable epoxide coatings depends on the availablility of a suitable photoinitiator. Several compounds

are

known

(see Table

I) which will

initiate

cationic

polymeriza-

tion on exposure to light, but not all of these are suitable for use in commercial coating operations. Some are costly, and their use in coatings applications would be uneconomical. With others, the rates of polymerization are too slow to be of interest in high-speed coating operations. These photoinitiators have generally not been available as off-the-shelf items from chemical suppliers. The necessity to synthesize them, together with the fact that their use as photoinitiators is protected by patents, has been a hindrance to their widespread use. From such information as is available, it appears that among the compounds listed in Table I, the onium salts show the greatest promise of meeting 0-8412-0525-6/79/47-114-017$07.00/0

© 1979 American Chemical Society

EPOXY

18

Table I.

Photoinitiators

for Cross-Linking

Reference

(54)

nitrosamines

(activated

by light and heat)

Diazonium

Fluoroborates

Diazonium

Perchlorates;

Perfluorocarboxylates

Fluorinated alkane sulfonic acid salts of Silver and thallium; or a metal salt of a polyboron acid and a silver halide or aromatic halide. Cyclopentadienylmanganese

tricarbonyl

Bis(perfluoroalkylsulfonyl) Diazonium

CHEMISTRY

of Epoxides

Type Unsaturated

RESIN

salts

other

Aryliodonium

salts

Aryliodonium

salts

than

methane

salt

12) Group Va Aromatic 13) Thiopyryllium

onium

salts

—ay

~—

on[o>)

a

dye

Diazonium difluorophosphate, phosphotungstate, phosphomolybdate, tungstogermanate, silicotungstate and molybdosilicate onium

~~

tetrafluoroborates

plus a cationic

11) Group VIa Aromatic

eRe

—olon —

compounds

silver

(4

on GO GW] GO} onS|] — ioe) 0O j=) —~S~— — Toa — “rH

salts salts

I —

— aos, Bb ~S

2.

WATT

Photosensitized Epoxides

19

the requirements of commercial coatings. On photolysis, they will initiate polymerization of some epoxides, and they have the capacity to promote the rapid cure required in high-speed coating Operations. Three types of onium salts most often mentioned in the literature as photoinitiators for epoxide polymerization are diazonium salts, aryliodonium salts and onium salts of the elements of Groups Va and VIla. Representative examples of each type are shown in Table II.

. Diazonium Salts. The first use of a diazonium salt as an initiator for cationic polymerization of cyclic ethers was reported in 1965 by Dreyfuss and Dreyfuss (12), who announced the polymerization of tetrahydrofuran initiated by thermal decomposition of a "new catalyst" which they identified as benzene-

diazonium material

hexafluorophosphate. was

correctly

(In a later publication

identified

(13) this

as p-chlorobenzenediazonium

hexafluorophosphate). In the same year, a patent was issued to Licari and Crepeau (14) for the photoinduced polymerization of epoxide resins by diazonium tetrafluoroborates for use in the encapsulation of electronic components and the preparation of circuit boards.

In 1966,

E. Fischer was

granted a U.S.

patent

(15) claiming

a method for polymerizing cyclic ethers by means of diazonium salts of perchloric acid and perfluorocarboxylic acids decomposed by heat and/or ultraviolet radiation. The storage stability of mixtures of these salts with cyclic ethers was poor, and the handling of diazonium perchlorates would require special consideration before introduction into large-scale coating operations. Dreyfuss and Dreyfuss (13) reported that p-chlorobenzenediazonium hexafluorophosphate was not a very effective initiator for polymerization of epoxides, based on an observation that thermal decomposition of the diazonium salt in the presence of propylene oxide did not yield a high molecular weight polymer. However, S. I. Schlesinger found that epoxides could be polymerized by a wide variety of diazonium salts, including p-chlorobenzenediazonium hexafluorophosphate, on exposure to ultraviolet radiation (16, 17). He delineated differences in behavior between various salts and identified those which are most effective in promoting rapid polymerization to high molecular weight. In particular, he showed that the hexafluorophosphates are much more effective as photoinitiators than the tetrafluoroborates (18). More than one mechanism has been proposed to explain the catalytic activity of diazonium salts in initiating polymerization

of cyclic

ethers.

Dreyfuss

and Dreyfuss

(13) postulated

that

initiation involves hydrogen abstraction from the cyclic ether by a carbenium ion formed via decomposition of the diazonium salt, followed by polymerization via tertiary oxonium ions associated The polymerization reactions studied by with PF,~ counterions. initiated by thermal decomposition of were Dreyfuss and Dreyfuss

SHTELOMaeS

Wd

‘SA7TNDIIONOYOVW ‘Ol ‘9# Pp

‘SOV (6961) ‘16 Stl

91G‘b6l‘°S

| 9(a 8‘AT'OTISAINO (8 ‘WV1TMWHT

MC'MAZOdVN

SONNSAIY Ar 13 ‘Vv NOILVIGVYONINND G (1)

SM

“aM LN3LVd

EPOXY

SN

AVS s1OJVIIUIOJOYT s2Anviuasaiday unio

SC LIVM

Sdl M(V1SWVT

OW

SH X 1457 SLVHdSOHdOYON WNINOZVIGSANSZNS8AXOHLSW-4d

(9

1dQ

“"1°S MH? ‘YSONISSIHOS 8 ATOTTISAINOD (VY f ‘A10d “19S iy WNISOdWAS LN3LVd 902‘°802°€ 9S% S6E-E8E (9261)

‘NSM3

69¢°628'°€ LNJ1Vd SN "Hf SONSZENIZDS (G

SN

WNINOGOIIANSHdIG

£H90 AJLVHdSOHdONON14VX4SH

19%.

N=N ©® 94g 143VX3SH

@jd @l

J1VHdSOHdONON

9

INS TANSHd IY L WNINOS

e443 sO

IT

IV

20 RESIN

SS9ON3Y355y

CHEMISTRY

SI€I-LOE] (2261) Nvf 8261

3¢

2.

WATT

Photosensitized Epoxides

21

the salt rather than photolysis. Their mechanism would account for a "dark reaction", explaining the limited storage stability some photosensitized- epoxides in the absence of light.

The Schieman

reaction

(19) was cited

by Schlesinger

of

(16) and

by Licari and Crepeau (14) as the probable mechanism by which initiators for cationic polymerization are released by diazonium salts. Rutherford et al (20) showed side reactions to be minimized when the diazonium salt is a hexafluorophosphate. Although heat is generally used to promote decomposition of diazonium salts by the Schieman reaction, the reaction has been shown to occur also under the influence of ultraviolet radiation

(21, 22): ®

Naa

R Phosphorous

©

ie

pentafluoride

Uae

was

(O)-+ + No + PFs R shown

by Muetterties

(23, 24)

to initiate polymerization of cyclic ethers to high-molecularweight polymers. The release of a Lewis acid such as BF3 or PFs in the presence of a material capable of undergoing cationic polymerization would be expected to yield polymeric products, and the photoinduced polymerization of epoxides can be explained according to the mechanism for cationic polymerization proposed by Rose (25)

and others

(26, 27, 28). O

H

So

PF. OHO

i ORE

) 5 Phosphonium

> Ammonium

> Stibonium

Since all these elements are in the same group of the Periodic Table (i.e., Group VA), one might expect the reactivity trend to cantinue going from arsonium compounds to stibonium ones. The fact that stibonium compounds are the least effective in this group suggests that steric hindrance effects, caused by the larger

a role.

be playing

may

cation,

(antimony)

stibonium

CHEMISTRY

RESIN

EPOXY

56

Summary

Quaternary phosphonium compounds have been found to be extremely effective latent accelerators for anhydride-cured epoxy Used at resins employed in electrical insulation applications. very

low

(0.01

concentrations

0.25%

to

level)

resins,

these

in

they have been found to give very fast gel times at elevated temperatures combined with very good storage stability properties Results show that the quaternary phosphonium at room temperature. compounds having non-halide anions give the best compromise of gel Comparireactivity, storage stability and electrical properties. son with other types of latent accelerators reveal that the quaternary phosphonium compounds are among the most effective additives yet discovered for epoxy-anhydride resin systems.

Literature

1.

Cited

Kohn,

L.

Paul

S.,

F.

in "Epoxy

Bruins,

Resin

Chapter 11, 219-228.

2.

3. 4.

5. 6.

7. 8. 9. 10.

ii’.

Technology,"

Interscience

Edited

Publishers,

New

by

York,

1968,

Sillars, R. W., in “Electrical Insulating Materials and Their Application," Institution of Electrical Engineers Monograph Series 14, Peter Peregrinus Ltd., Stevenage, England, 1973, 192-198. German Patent 1,162,439 to Westinghouse (Nov. 1958). British

Patent 869,969 to Shell (June 1961). British Patent 966,917 to Union Carbide (Feb. 1965). Launder, A. J., A.C.S. Div. Org. Coatings Plast. Chem., Preprints, 24, No. 2) (Sept. 1964). Feuchtbaum, R. B., Insulation, 47 (Sept. 1962).

Lee,

H.

and Neville,

K.,

"Handbook

of Epoxy

Resins,''

Hill Book Company, 1967, Chapter 5, p. 20. U.S. Patent 2,928,807 to Devoe & Raynolds Co. (March Rogers, D. A. and Kauffman, R. N., Westinghouse R&D Unpublished Work.

McGraw1960).

Vincent, H. L., Frye; C.9L. and Oppliger,9P. E., in! Epoxy Resins,''Advan. Chem. Ser., 92, R. F. Gould, Ed., American

Chemical

12.

Chapter 14, Langer, S. H.

13.

Langer,

(1957).

Thomas;

S=

and

Society,

Elbling,

Washington,

1. N.,

Ind.

D.C.,

Eng.

1970,

Chem.,

4994113

=

Ho, Blbling, #1. N., Finestone, Abe and W. Ro, J. Appl -sPolym. Sci., Vs) NOweEs 63708

14.

Belgian Patent

15. 16. Io 18.

U.S. Patent 3,759,866 to Westinghouse U.S. Patent 3,784,583 to Westinghouse Smetely, de Mo Bog Ty Apply Polly wSei. U.S. Patent 3,979,355 to Westinghouse

RECEIVED May 21, 1979.

633,330 to CIBA (1963).

(Sept. 1973). (Jan. 1974). wan Press. (Sept. 1976).

(lool):

Development of Epoxy Resin-Based Binders for Electrodeposition Coatings with High Corrosion Resistance W. RAUDENBUSCH Koninklijke/Shell-Laboratorium, Amsterdam (Shell Research B.V.), Badhuisweg 3, 1031 CM Amsterdam-Noord, The Netherlands Electrodeposition (ED) has proved to be a most attractive process for industrial coating, in particular for the priming of automobile bodies. Epoxy resins play a leading role in this field, nearly 6 % of all epoxy resins sold for coatings being used in ED systems. Until

the

early

1970s,

all

ED

binders

used

industrially

were

based on maleinised natural oils, acidic alkyd resins or epoxy resin esters. Later, binders based on maleinised polybutadiene oils (LMPBs= low-molecular-weight polybutadienes) also came into industrial use, particularly in Western Europe. All these systems were of the anionic type, l.e. their water solubility and their deposition at the anode were due to the presence of carboxylate groups in the binder molecules. More recently, cathodic ED, with binders carrying cationic groups (e.g. -N®R5E) , has proved tobe an interesting alternative. This paper describes development work directed towards improving the corrosion resistance of anionic, and more briefly, cationic ED binders. The properties, described in broad terms, of the different classes of anionic binders are shown in Table I. The systems initially used (epoxy esters, maleinised oils and acidic alkyds) generally performed quite well and the rapid rise of the LMPBs, notwithstanding their bath stability problems, was therefore rather surprising. In fact, the advance of LMPB-based binders was due to one key advantage: their superior corrosion resistance when applied to nonphosphated steel substrates. This property is important in modern automotive applications since most cars now used in Western Europe are of the integral body type on which spray-phosphating always leaves some poorly phosphated or even bare areas. Primers offering good corrosion protection on poorly phosphated steel are thus quite essential for ensuring a long lifetime of modern cars. The superior

performance permeability however, has

of LMPBs

in this

respect

was

ascribed

to the

of polybutadiene films Chyis Our more recent thrown a different light on this theory.

0-8412-0525-6/79/47-114-057$05.00/0 © 1979 American Chemical Society

low water work,

TeoTUeyooy SULYeOO poos 04 eqenbope Of ATA

——-

qy

uo Uo

e7eq

peyeydsoyd Tee1s Teeqs

S0UeYSTSAI

sotqysedoud

setytzedoud

AGre

UOTSOZIOD

ch aremetUse EC

ATeA poos

poos09 ATA poos

oqenbepe 04 AZSA poos

paesTUTeTey ‘STTO | OTptoe SpAYTe

HONVNHOANE 4O OINOINV GH

wildvib I

poos

poos

AeA poos zood

LIaA

ATOA

Axody uTser | sz94se

SHHCNTE

Area Area

AzaA

poos poos

poos

poos09 AISA poos

TBUOTSed00 SOTUTNOTIJIP

PSeSTUTSTeW SCNT

58 EPOXY RESIN CHEMISTRY

4.

RAUDENBUSCH

Fundamental

Epoxy Resin-Based Binders

Studies

on

the

Factors

59

Governing

Corrosion

Resistance

When comparing two typical ED binders, a commercial LMPB system and a long-oil epoxy ester, in the standard ASTM salt-spray test (25 um clear coatings on bare steel, stoved at 180 °c for 30 min), we first observed that the ratings (good with LMPB, poor with the epoxy ester) did not depend on the method of application used (ED or spraying from solvent solution). Therefore, we subsequently applied all binders simply from organic solvent solutions, which greatly speeded up our further work. Our second observation was that the epoxy ester coatings failed by loss of adhesion rather than by film degradation. We assumed that of the different polar groups present (carboxyl, ester, ether) the carboxyls play the major role in adhesion promotion, and lose this function upon salt formation with the aqueous alkali formed during salt-spray testing. Consequently, we inquired which other, more alkali-resistant, adhesion-promoting groups might be present in LMPB binders. By infrared spectroscopy we discovered that during curing of LMPB films at 180 °c for 30 min about 40% of the carbon double bonds initially present disappear and substantial amounts

of

hydroxyl

and

(ketonic)

carbonyl

groups

are

formed.

The

assumption that these groups are primarily responsible forthe high corrosion resistance on bare steel was supported by the fact that LMPB coatings cured only by air-drying (20 ore ad) contained much less

OH

and less C=O and showed poor salt-spray ratings. Elaborating on the hypothesis that hydroxyl groups may be important alkali-resistant adhesion promotors, we investigated the salt-spray performance of certain epoxy-resin-based coatings which after curing contain reasonably well-defined levels of unreacted hydroxyl groups. One such system consisted of the solidepoxy resin EPIKOTE 1007 (E-1007) cured with various amounts of hexamethoxymethylmelamine (HMMM). When the hydroxyl content of the resin is known, and the equivalent weight of HMMM is taken as 130 (assuming for steric reasons that only three of the six available methoxymethyl

groups

are

reactive),

it

is

possible

to

calculate the amount

of residual OH groups in cured coatings quite accurately. These expériments (see Table II) showed that good salt-spray resistance on bare steel may be expected if the coating contains between 200 and 400 meq hydroxyl groups per 100 g. It is interesting to note that

good

salt-spray

resistance

appeared

to

be

unrelated

degree of crosslinking achieved in this system. The strong beneficial influence of high hydroxyl corrosion

least

resistance

epoxy

where

Some

systems,

than

on

Thus other work, some

their

also

failed

seen

formed

resins

however,

epoxies,

even though

was

most

to

residual

in

the

of which

give

were

similarly

OH level was

the

levels

model

experiments,

coating

ingredients.

other

major

to

based

good

high

factors certainly also play a role. of these factors appear to be:

on

alkyds

salt-spray

on at

rather

ratings

(300to 400 meq/100g¢). From

our

subsequent

*q

*qo00aTp

WLSV

SouUeYSTSOL 1 ep g

step |

qeoueystser jo

keids-77es HO

HO-LLLG

elddLS

SSUnmEd: 00|/beu S

pema

queq4uos

Tenprtssey

azeeTO G2 wn s¥urzeoo poettdde £q SuTfeads Worg OTUBSIO JUSATOS UOTINTOS pus Ppenogs 92 OCl Jo sz07 OF “UTM suleTske PSUTeiuOS G°O My fo -“ThsTOnqoq-d OLUOT PlowSe “ASATeISS :sSuTqyey ‘Spaeyoasgyeun=0| G=6 mm ssoT Jo uoTseype Wory “Yoyveros =g OL um ssoT Jo SuOTSaYype =) G| um ssoTJO SUOTSOYpe ‘*04e=0 [TB109 SSOT jo ‘uoTseype

poos ATOA poos

zood poog

suo

TAUQZE

TANASE

WAWH/OOL SONILVOO NO GYVE

‘qoBdmr y 48

,SLOMIdH,

poos AISA

*e

ssoupszey

ZpTOuyong Gq

SHILYAIOUd JO

Vidi Eta Et

60 EPOXY RESIN CHEMISTRY

4,

RAUDENBUSCH

Epoxy Resin-Based Binders

- Susceptibility - Hydrophilicity to result

of cured of cured

fromthe

coatings to films. Poor

presence

61

alkaline hydrolysis. salt-spray ratings were found

of too many polar groups

in the coatings.

- Steric (?) factors. Generally we observed that the use of even modest amounts of fatty acids or their esters in OH-rich binders substantially decreases the corrosion resistance. A possible explanation would be sterical shielding of the adhesion-promoting groups

by

Development

the

long

fatty

of Anionic

ED

acid

tails.

Binders

with

High

Corrosion

Resistance

The "high-hydroxyl" principle was discovered in experiments in which the coatings were conventionally applied from organic solvent solutions. We then set out to develop ED binders conforming tothis principle. Apart from offering good corrosion resistance, such binders

must

of

course

meet

a number

-

solubility or dispersibility good bath stability, good ED behaviour, crosslinkability on stoving, good mechanical and chemical

-

acceptable

Note:

preparation

of further

requirements, such as

in water,

methods

film

properties,

and

costs.

The formulations described in this section are covered by the German Patent Application OLS 2700537, and other applications in a number of countries. Much

solid

of

epoxy

our

resin

work

was

(EPIKOTE

based

type

on

the

1001

or

following

1004)

approach:

terminal

Into

a

hydroxyl

groups were introduced by reacting the epoxy groups with stoichiometric quantities of hydroxy acids such as lactic or, preferably, dimethylolpropionic acid (DMPA). The resulting hydroxyl-rich "back bones" (8-12 OH groups/molecule) were then reacted with a cyclic

anhydride groups

such

as

necessary

succinic for

anhydride

solubilisation

(SA)

to introduce the carboxyl

in water:

CH,-O# CH.,-000-CH,-CH,-COOH -oco-cCH,.-OCO-C-CH oe eee ee ee CH,-OCO-CH,,-CH,-COOH OH CH,-CO — CH,-OH OH 0ou 2 HOOC-C-CH a 12 So oe OH CHp-CO . OH CHDOH . OH OH OH OH OH OH OH -COOH CH,-OCO-CH,,-CH,, OH ‘84! cHon on (DMPA) 2} 2 { 2) { m4 O SNCS Ce CH,.-OCO-C-CH oe 3 2° i ee ei EPIKOTE 2 CH,.-OCO-CH,,-CH,,peer 2 ae 1004 WATER-SOLUBLE OH-RICH BACKBONE

RESIN

62

EPOXY

It

soon

became

"hard",

i.e.

when

binders

oil-free

these

that

clear

flowed

applied by ED they

RESIN

poorly

CHEMISTRY

too

were

during

the

actual deposition step and consequently formed rough coatings (nos. 1 and 2 in Table III). Better-flowing binders were obtained when the succinic anhydride was replaced by a substituted material, viz. a premaleinised

acid

fatty

3, Table

(no.

III).

Unfortunately,

how-

ever, this approach always led to a deterioration in salt-spray resistance, an effect typical for modifications with natural fatty acids (see above). We then discovered that we could improve the softness and the flow of the binders by incorporating a derivative of a synthetic fatty acid. This derivative is the glycidyl ester of an a-branched monocarboxylic acid with 10 carbon atoms and is commercially avail-

able

under

compound, of the

the trade-name CE

10 could

carboxyl

Ge

be

groups

Oe

CARDURA easily

present

E10

(CE 10).

incorporated in

the

Being

a monoepoxy

by reaction

with

binders:

EN

CH,.-O-CO-C-CH 2 | 3

OH OH o OH

CH,-OCO-CH,,-CH,,-COOH Pax ay C CHa=CH=CHs=0CO= 57 CH CH, Oco CoH ig 4S tS (CEM 10)

OH OH OS

CH-OCO-CH,-CH,-COOH Oe

de { CH,,-OCO-CH

2 ~CH,,-COOH WATER-SOLUBLE RESIN OH “Hy CH,-OCO-CH_=CH “CO0=SCH -CH-cH CoOL Cor 57 CHy—COO-CH,,-CH-CH,,-COO GH,-0-CO-C-CH., OH -OCO-CH_-CH.— os CH,,-OCO-CH,,-CH,,~COOH OH OH OH OH

. CH,-OCO-CH,-CH, -CO00-CH.-CH-CH.-ocoae CH, : 97 000-C

CH,-O-CO-C-CH 2

|

3

CH, 5 - OCO - CH, = CH, = COOH OH-RICH

(No.

ED

BINDER

dinsTable

117)

oH g

part

4.

RAUDENBUSCH

Epoxy Resin-Based Binders

TABLE HYDROXYL-RICH

Binder composition? Molar ratios

63

III

ANIONIC

ED BINDERS®

Hydroxyl ED content©, | behaviour®

stability2

meq/100

(7 a/40 °c)

Bath

g

E-1004/LA/SA

Mechanical coating properties

poor

fair

good

poor

fair

good

good

fair

good

Tpreerhalte'> E-1004/DMPA/SA i)

ete

Wh

2

fs Wes

fe

4 | E-1004/DMPA/SA/CE

Wien

10

fe

E-1001/AA/DMPA/TMA/CE

Af

tee

elas

eae

very

good

good

ees i2a

Commercial LMPB binder (comparison) ee

a.

All

binders

clear

at

(except

coatings

180 °C for

no.

applied

6) combined to

bare

steel

with by

HMMM

ae

(weight

spraying

or

by

ratio

95/5);

ED,

folioued

25 um thick by

stoving

30 min.

b.

E-1001, E-1004: solid EPIKOTE epoxy resins; LA: lactic acid; DMPA: dimethylolpropionic acid; SA: succinic anhydride; TMA: trimellitic anhydride; MAFA: premaleinised fatty acid; CE 10: CARDURA E10; AA: adipic acid.

e. d.

Calculated. 10 % aqueous

e.

binders), and Spray-applied

f.

Both

ED-

and

solutions

containing

triethylamine coatings; spray-applied

for

ethylene

glycol

mono-n-butyl

ether

(30

% on

neutralisation.

coatings;

} for

salt-spray

ratings

see

Table

II.

A binder

(no.

type

of this

in combination

III),

4 in Table

with small amounts of melamine or phenolic resins showed good ED behaviour and good mechanical film salt-spray

resistance

only

was

RESIN CHEMISTRY

EPOXY

;

:

64

slightly

lower

CARDURA-free analogue (no. 2 in Table III). lesser detrimental effect on the salt-spray

than

as crosslinkers, properties. Its of

that

a

CE 10 seems to have a resista than nce natural

acids, possibly because its hydroc chain arbon is much shorter. The hydrolytical stability of binder no. 4 was not very good and our next task was to improve this aspect. Binders rendered water-soluble by the acidic halfester principle contain a hydrolytically weak site, viz. the halfester linkage. fatty

OH-RICH (Saw packBong [~ 2+ =° a ss °

OH-RICH BOE. | Gscos oN COOH BACKBONE

CYCLIC ANHYDRIDE

ACIDIC

HALFESTER

The reason is probably that the saponification of the halfester is assisted by the neighbouring carboxyl function. We screened the hydrolytic stability of binders containing different types of half-

esters

by storing

aqueous

solutions

(10 % solids

content)at ko °c

and periodically checking the pH, the conductivity and the ED behaviour. In this way we found that binders containing halfesters of maleic or phthalic acid are quite unstable. The stability of binders with succinic halfesters is better but probably still insufficient for most industrial applications. Halfesters of tri-

mellitic anhydride (TMA) of succinic acid. To our

are considerably more stable than those surprise we found that modification of the

trimellitic halfesters with the above-mentioned monoepoxide CE 10 enhanced the stability even further. If about equimolar amounts of TMA and CE 10 are used there are sufficient carboxylic functions left to guarantee a good solubility of the binders in water:

OH-RICH AVe tons

OHL

ae

0 i € a ue

COOH

" 0

0 re \ + CH, at CH aneCH, aaa 0200

TMA

Colic

CHO

|

oF

OH-RICH Mt BACKBONE [7 °-©—

COO-CH_-CH=CH_-0-C0_-¢_# e ‘ 7

COOH CARDURA

FF

E10-TMA

MODIFIED

ED BINDER

4.

RAUDENBUSCH

Epoxy Resin-Based Binders

65

From our laboratory experience we believe that such systems are about as stable as the industrially well-known binders which are rendered water-soluble with premaleinised fatty acids. A typical hydroxyl-rich binder which is modified with TMA and and CE 10 is no. 5 in Table III. In this binder the backbone is slightly different from that of the preceding binders, i.e., EPIKOTE 1004 is replaced by a combination of EPIKOTE 1001 (two moles)

and

adipic

acid

(one

mole).

This

was

done to further

improve

the

mechanical properties of the coatings. With melamine or phenolic resins as crosslinkers binder no. 5 shows an attractive combination of good ED behaviour, good mechanical film properties and high bath stability. Its salt-spray resistance is comparable to that of a commercial LMPB system (no. 6 in Table III). Binder no. 5 also shows self-crosslinking behaviour, i.e. it gives coatings with an interesting level of performance

when

In view binder

no.

stoved

of these

5 to

in

the

absence

encouraging

customers

under

of

external

results,

the

code

we

LR-2052.

reactions so far have been quite positive and the high salt-spray resistance of this binder steel

were

fully

crosslinkers.

decided

to

submit

Customer

our views regarding on non-phosphated

confirmed.

With a view to lowering raw material costs (the DMPA used in binder LR-2052 is rather expensive), we later developed another OH-rich binder, designated LR-2053. This systemis based ona linear backbone containing a liquid epoxy resin, EPIKOTE 828, diphenylolpropane, adipic acid and CE 10 (molar ratio 3/2/2/2) and again is rendered

water-soluble

with

a TMA/CE

10

combination

(2.5

moles

TMA

and 2.6 moles CE 10 per mole of backbone). This binder performs as well as LR-2052, with an only marginally lower salt-spray resistance. With

both

binders,

LR-2052

pigmented ED paints. We out active ingredients. Table

IV.

The

general

and

LR-2053,

we

have

prepared

used a simple Ti0o based pigmentation withTypical data for these paints are shown in

performance

of both

pigmented

systems

very satisfactory and the salt-spray ratings were comparable those of unpigmented coatings. In conclusion, we can say that it is possible to design anionic ED binders based on epoxy resins which give coatings a much

better

salt-spray

coatings

based

features

of the

on

resistance

conventional

epoxy

resin

on

anionic

based

bare

steel

binders.

binders

are

than

The (1)

that

main

the

was

to

with of

design

of a presence

sufficiently high level of hydroxyl groups in the crosslinked coatings, and (2) the omission of natural fatty acids. In later work we applied the same principles to cationic ED binders. Cationic

ED

Binders

A chief advantage claimed for coatings applied by the cathodic ED is their superior corrosion resistance, and this has been the main incentive for our work in this area. Initially, we tried to establish whether the cathodic deposition process as such would lead to improved corrosion resistance of the coatings. To this end

66

29TTeI9g ,ZOOL /S8St gc°0 S*0n/00L 0g*0 cay

0Se

tl e°

G6L

gz°0 0/001 $* €g°0

JeTTeIEg |,001

t< SLL/SL 022 SLL-OS1L 2 Of 0921

Qo-ce

/S98 SL

“(L/e2

poos £i9A

Otqer

un

Syoom

A/wo

SLNIVd

pessed £6-6

T


|=

hexahydrophthalic

IPDA

=

isophoronediamine 3,3 -dimethyl-5-aminomethylcyclohexylamine

MCHA

= 4,4'-methylenebis(cyclohexylamine)

MDA

= 4,4'-methylenedianiline

MMCHA

anhydride

= 4,4'-methylenebis(2-methylcyclohexylamine) m-xylylenediamine

ia

PMDA

MDA,

=

= 1,3-bis(aminomethyl) benzene

p-menthanediamine

= 1-methy1-4-(l-amino-l-methylethyl)cyclohexylamine TETA

=

triethylenetetramine

=

TMDA

=

2,2,4-trimethylhexane-1

1,4,7,10-tetraazadecane ,6-diamine

RESIN

EPOXY

126

CHEMISTRY

This resin mixture, which had a low viscosity, typically about 2500 mPa.s, was found to avoid the crystallization tendency Particular utility has been found of pure diglycidyl DMH, Ila.

for the reversible high degree of swelling of a room temperature cured version of a DMH-based resin, as a water-permeable topcoat

The methylethy lhydantoin-based for marine antifouling paints (4). the greatest range of gave VIII, Table Resin IIb, as shown in This is in Peavoonilicicy when the amine curative was varied. keeping with its intermediate degree of alkyl substitution. A few resins were given elevated temperature cures with aromatic amines; here the effect of alkyl substituents on water uptake could be seen, but the overall behavior was intermediate between the elevated-temperature anhydride cures and the room temperature amine cures. Solvent/Chemical

Resistance

The hydrophobic shielding of the hydantoin ring by alkyl substituents affected all the solvent-solute interactions of cured resins. Two of the resins and the DMH-based resin mixture were cured with a commercially available aromatic amine mixture derived from aniline-formaldehyde condensation, identified in Table VII. Weight gain and solvent plasticization were followed in a number of solvents and aqueous media. Some of the exposure was at 60°C as well as at room temperature. For the hydrophobic Resin IIk, the results for exposure to various solvents are shown in Table IX. The solvents are

listed

in increasing

order

of solubility

parameter

&(5).

It

is

evident that (within the limitations of solubility parameter theory) this cured resin was lyophilic in the range of $= 9 9.6; there was also a specific interaction with methanol. Table X shows the generally hydrophobic response of this cured resin in aqueous media (listed in increasing order of pH). Only strong acid had a deleterious effect. Again at the other extreme, the DMH-based hydrophilic resin mixture

This

was

cured

studied,

resin

was

with

results

hydrophilic

shown

(Table

in Tables

XII),

XI

though

and

less

XII.

so

than

the room temperature amine-cured systems of Table VI, but note its general lyophobicity, as shown in Table XI. Of the non-aqueous media, only methanol and hot trichlorethylene showed much swelling or plasticization in 16 weeks. Presumably the poorly shielded DMH rings permitted strong interchain interactions; this virtual crosslinking provided general solvent root cance! Resin IIb, which proved somewhat less hydrophilic than the DMH-based resin mixture (see Table VIII), was studied with certain solvents only, as shown in Table XIII. Its lyophobicity was somewhat less than the DMH-based mixture, and in the two aqueous media it was somewhat less affected. Thus, again, its behavior reflected the intermediate degree of alkyl substitution.

10.

CATSIFF ET AL.

Cured Hydantoin Epoxy Resins

TABLE ROOM

TEMPERATURE

Resin:

IIb

AMINE

127,

VIII

CURING

OF

HYDANTOIN

EPOXY

(1,3-diglycidyl-5-methyl-5-ethylhydantoin) Water

Uptake,

%

% Retention Flex

Curative Formulated

RESINS

1 Day

amine

TMDA (a) 30%eq TMDA ae MCHA

(a)

(a)\ (a).'

15%Zeq TMDAY foes ey

4 Wk

Modulus

4 Wk

ORD

Dy,th

80

Tea

5e3

80

>

ee

2

5

6.0

(b)

MMCHA

(a)

tel

eed

(b)

1,4-BAC

(a)

Zeek

Wik 3th

(b)

1,3-BAC MXDA IPDA TETA

(a) (a) (a) (a)

1.9 2.4 (c) 1334

10.4 9.6 -(c)

(b) (b) (b) 10 (d)

All All

were were

Except water

cured cured

where at

noted,

room

Zeq = percent

(a) (Cb)

at H/R = 1.0 amine hydrogen/epoxide. at room temperature 14-21 days. exposure

was

by immersion

temperature.

of

amine

equivalents

See Table VII for identification. Loos prt ttlesto cut.

(c)

Fragmented

(ye

Atel

day

in deionized

of

RESIN

EPOXY

128

TABLE SOLVENT

RESISTANCE

IIk

Resin:

OF

IX

AMINE-CURED

AROMATIC

CHEMISTRY

HYDANTOIN

RESIN

EPOXY

[1,3-diglycidyl-5-ethyl-5-(2-methylbutyl)hydantoin] IDi.

=

1285¢

Solvent

Uptake,

4

Sokubielnt ya Lerapr

Solvent

Parameter(5)

% Retention of Flex Modulus

°C

1 Wk.

16 Wk.

DS;

0.03

0.54

100

25

0.02

0.10

110

60

Ons

0.9

95

8.9

DS, 60

0.08 O75

0.8 Beis)

100 85

8.91

25

1.9

2958

GG

DS

Ona

eZ.

100

9 0G

25 60

36D (a)

5)5) 5 1 ==

30 ==

Chloroform

9.16

25

(a)

ve

--

Acetone

9.62

D5,

(a)

ae

—

25

-0.08

DS

Ua gil

Heptane

TO

Kerosene

Xylene

Ethyl

Acetate

Benzene Trichlorethylene

Isopropyl

Alcohol

16 Wk.

Cb)

15

OF43

(g)

100

11.44 Methanol

0

ZOe

5

All were cured with aniline-formaldehyde curative. H/R = 1 amine hydrogen/epoxide equivalent

Gelled (a)

16 hr./65°C.

Fragmented

Postcured (b)

See

Table

2 hr./150°C

Cracked

(g)

At

8 weeks

VII.

10.

CATSIFF ET AL.

Cured Hydantoin Epoxy Resins

TABLE CHEMICAL AMINE-CURED

Resin:

IIk

129

X

RESISTANCE

OF

HYDANTOIN

AROMATIC

EPOXY

RESIN

[1,3-diglycidyl-5-ethyl-5-(2-methylbutyl)hydantoin]

LDS

28s C

Liquid

Uptake,

%

% Retention

Aqueous

Temp.

Medium

“se

1 Wk.

25%

DS

Bod

2 60

it DS

DEO 4.8

105 25

Water

25 60

Io I)

Siew) 4.4

mae Q5

10%

NaCl

25 60

I all 2.4

Fall Sige)

os 95

10%

NH,

D5)

Wg

4.3

95

50%

NaOH

25 60

-0.04 0.05

HCl

54 Acetic

Acid

Flex

16 Wk. 38

0.01 0.14

All were cured with aniline-formaldehyde curative. ab vem Valine H/R = 1 amine hydrogen/epoxide equivalent

Gelled

16 hr./65°C.

Postcured

2 hr./150°C

Modulus

16 Wk. 40

105 90 See

of

RESIN

EPOXY

130

TABLE

Resin:

70% 30%

XI

AMINE-CURED AROMATIC OF eee eee

RESISTANCE SOLVENT Reo NC SOLVENT

Ila IV.

CHEMISTRY

HYDANTOIN

EPOXY

RESIN

[1,3-diglycidyl-5,5-dimethylhydantoin] [1-glycidyl-3-(2-glycidoxypropyl)5 ,5-dimethylhydantoin]

IDT = 125°C Solvent Solubility

Solvent

Parameter(5)

Uptake,

Z

Temp.

% Retention of Flex Modulus

°C.

1 Wk.

16 Wk.

16 Wk.

ye)

OW

0.84

100

25) 60

0.09 ORS

O29 hed

100 105

Ong

IES, 60

0.04 On

ike Bp

SS 95

iS). SIL

2S)

0.06

0.6

100

9-16

DS

0.07

10

100

he IMS

25 60

ORL 0.4

isa 6.8

95 80

Chloroform

8) NG

25

@i

I 325)

Acetone

hola

29

Orel

0.96

SNS)

Alcohol 11.44

25

0.00

Onis

105

14750

25

6.9

AMM sO)

15)

Heptane

ToDo

Kerosene

Xylene

Ethyl

Acetate

Benzene Trichlorethylene

Isopropyl

Methanol

All were cured with aniline-formaldehyde curative. H/R = 1 amine hydrogen/epoxide equivalent

Gelled

16 hr. /65°C,

“Postcured

2 hr./150°¢

100

See

Table

VII.

10.

CATSIFF ET AL.

Cured Hydantoin Epoxy Resins

TABLE CHEMICAL

saree

thes Ila 30% IV

XII

RESISTANCE

AMINE-CURED

131

OF

HYDANTOIN

AROMATIC

EPOXY

RESIN

[1,3-diglycidyl-5,5-dimethylhydantoin] [l-glycidyl1-3-(2-glycidoxypropyl)5,5-dimethylhydantoin] IpT-=

125¢

Liquid

Uptake,

Z

% Retention of Modulus

Aqueous

Temp.

Medium

GC.

1 Wk.

16 Wk.

16 Wk.

2a peEG kL

25

(a)

SS

==

25 60

ibste! 8.0

10.1 159

75 50

Water

25) 60

Ie) seh

10.6 (2 7)

ae 60

10%

Nacl

DS: 60

13 4.8

6.5 Ohs5)

oS 70

OK

NH,

25

1.8

is hey)

55

50%

NaOH

25) 60

0.005 OFZ

(0). 745) = (ret

54 Acetic

Acid

Flex

110 85

All were cured with aniline-formaldehyde curative. See Table VII. H/R = 1 amine hydrogen/epoxide equivalent

Gelled (a)

16 hr./65°C.

Fragmented

Postcured

2 hr./150°C

EPOXY

132

TABLE SOLVENT/CHEMICAL AROMATIC Resin:

IIb

AMINE-CURED

RESIN

CHEMISTRY

XIII RESISTANCE

HYDANTOIN

OF

EPOXY

RESIN

[1,3-diglycidyl-5-methyl-5-ethylhydantoin] IDT

149°C Solvent

Solubility Parameter(5)

Temp.

Uptake,

16 Wk.

Z

% Retention Flex

Modulus

16 Wk.

uGe

1 Wk.

De)

0 {Ol

1.0

105

OR EruG

il i 6.5

105 90

Chloroform

oe

Pde

100

Acetone

ono

Bia

85

(a)

--

(ohe5)

90

Solvent

Ethyl

Acetate

8.91

Trichlorethylene

25%

HCL

10%

NH,

25

All were cured with aniline-formaldehyde curative. H/R = 1 amine hydrogen/epoxide equivalent Gelled 16 hr./65°C. Postcured) 2ontre/150.C (a)

Fragmented

See

Table

VIL.

of

10.

CATSIFF ET AL.

Cured Hydantoin Epoxy Resins

133

Conclusions

By varying the substituents at the 5-position on the hydantoin ring, it is possible to control the intermolecular forces of hydantoin-based epoxy resins. This control is manifested to some extent in the monomeric resins, but is far more evident in the glass transition temperature of highly crosslinked resin-curative systems, and particularly in the hydrophilichydrophobic balance of amine-cured systems. Within limits prescribed by the substituents on the hydantoin ring, the hydrophilic-hydrophobic balance is also greatly affected by the organic moieties introduced by the amine curative. Analogous principles may be invoked in considering the lyophilic-lyophobic balance of aromatic amine-cured hydantoin epoxy resins. Experimental Mold

Procedures

Construction.

Casting

molds were prepared from preplates and square-angled The gasket sheets were about 1.6 mm thick and two or more were plied to the casting thickness desired, and spring-clamped between plates. The molds were set vertically with the top edge open for pouring, and cleaned mold-release-coated glass U-shaped silicone rubber gaskets.

usually

were

preheated

in

an

oven

mixture as fluid as possible. was about 20 cm square.

to

The

keep

most

the

common

viscous

plate

Specimen Preparation. The demolded castings using a circular saw with a diamond cutting wheel specimens for the various tests.

casting

size

used

were cut to provide

Casting. Castings of liquid epoxy resins cured with hexahydrophthalic anhydride (HHPA) were prepared by heating the liquid resin in a stirred 3-necked flask to which the calculated amount of

premelted

degas

HHPA

the mixture;

was

added.

then

A vacuum

was

drawn

to

2 phr benzyldimethylamine

partially

(BDMA)

was

added

as accelerator. After further degassing, the mixture was poured into casting molds prepared as described above. The filled molds were put into an oven at the desired gelling temperature; subsequently, either the oven was reset to the final cure temperature or

the

molds

were

transferred

to

a

second

oven,

as

convenient.

Castings of solid epoxy resins cured with HHPA were prepared similarly, except that the resin also had to be premelted. Except where noted in Table II, HHPA-cured castings were gelled at 80°C (at least 3 hours and usually overnight) and then cured

Jae

toO.ce

Castings cured with the commercial from aniline-formaldehyde were prepared The castings, but without accelerator.

was

about

90% methylenedianiline

(MDA)

aromatic like the aromatic

and

tended

amine derived HHPA-cured amine mixture

to partially

EPOXY

134

RESIN

CHEMISTRY

solidify on standing, so it had to be prewarmed and well mixed In the work reported here, the degassed castings use. were gelled 16 hr./65°C and given a final cure Of 2 bre /Lo0GGs For room temperature curing, preheating was used only when absolutely necessary to reduce viscosity and facilitate mixing. For this reason, Resin Ile was not considered suitable for room temperature curing, and Resin Ila was so cured on only a few Such resins could be used in liqoccasions, after premelting. In a similar uid mixtures, as noted in Tables VI, XI, and XII. xylamine) , is(cyclohe methyleneb and MDA as such amines, solid way, MCHA, could only be used in liquid mixtures. With all the amine curatives, the stoichiometry used was one amine hydrogen/epoxide; e.g., TETA, having two primary and two secondary amines/molecule, would have six equivalents/mole. The mixture of liquid resin and liquid amine was stirred and degassed in a 3-necked flask, just as for the thermal castings, but an ice bath was used to control any exotherm in the flask, so that rapid viscosity increase due to reaction would not occur to interfere with degassing. After pouring into the mold, the casting was kept at room temperature for at least 14 days before demolding. When a mild postcure was desired, at least 7 days/room temperature curing preceded it; the postcure was 6 hr./100°C. before

Weight Gain and Flexural Modulus. Weight gain and plasticization by absorbed liquid were measured on small flex bar specimens (7.62 cm x 2.54 cm x 0.32 cm). Most exposure was by immersion at room temperature, but some specimens were immersed at 60°C, and some specimens were exposed to 95% R.H. at 35°C in a controlled-humidity chamber in parallel to their immersion in water.

Deionized

water,

ten

organic

solvents,

and

five

aqueous

solutions were used as immersion media, as noted in the appropriate tables. Exposure times ranged from 24 hours to 4 weeks for the water- or humidity chamber-exposed samples, and from 1 to 16 weeks for the solvent- or aqueous media-exposed samples. At the desired intervals, each specimen was removed from its environment, rinsed if necessary for safe handling, wiped dry, and weighed. Then the flexural modulus was measured, to a maximum outer-layer strain of 0.5%. Thus, the flex measurement was essentially non-destructive and the specimen could be returned to its exposure.

Initial Deformation Temperature (IDT). IDT was determined as the temperature at which a standard test bar (1.27 cm wide x 0.635 cm deep), centrally loaded on a 100 mm span, deflected an additional 0.25 mm under a load that gave a maximum (outerlayer) stress of 1.82 MPa, while being heated at a rate of 2°C/min. The operating procedure followed ASTM Method D648-72 for "Deflection Temperature of Plastics Under Flexural Load" (6), but the maximum strain reached at the IDT was half that at the "deflection temperature under load" (DTUL), so IDT was

10.

CATSIFF ET AL,

several

Cured Hydantoin Epoxy Resins

degrees

lower

than

135

DTUL.

Abstract

Hydantoin epoxy resins having glycidyl groups in the l- and 3-positions and one or two alkyl groups in the 5-position were prepared by the Bucherer reaction, followed by treatment with epichlorohydrin. These resins were crosslinked with hexahydrophthalic anhydride to examine the effect of alkyl substituents on the glass transition temperatures of the cured systems. Higher alkyl substituents shielded the hydantoin rings and gave lower glass temperatures. The same shielding effect was observed in the reduced hydrophilicity of higher alkyl-substituted hydantoin epoxy resins cured with triethylenetetramine. Steric factors - branching at or close to the hydantoin ring - raised the glass transition temperature while maintaining the shielding effect. Amines of different structures were used as room temperature curatives with a few representative resins, to observe the effect on hydrophilic-hydrophobic balance. Solvent effects were examined on aromatic amine-cured resins; the most hydrophilic cured system proved to have the broadest range of lyophobicity. Acknowledgement This

nical

work

could

assistance

thanks go to A. rT. Doyle, Literature

and

not

. Eldin,

S.,

. Weiss,

J.

Lukens, Testing

RecEIvED

carried

out

without

colleagues. L.

the

tech-

Particular

Alter,

H.

B.

Dee,

Cited

. Habermeier,

K.

been

of many

Dr. J. H. Bateman and Messrs. D- Neaditch, and J. Velten-

Catsiftf, E. H.; Dee, 1978, 55 (7), 54.

Hoy,

have

support

L.

J.

Angew.

private

Organic

J.

H.

B.;

Seltzer,

Makromol.

R.

Chem.,

Modern

1977,

Plastics,

63 (921),

63.

communication.

Coatings

Paint

and

Technol.,

Plastic

1970,

Chemistry,

42 (541),

1978,

39,

567.

76.

R.P., Ed. "Annual Book of ASTM Standards;"" Am. Soc. for and Materials: Philadelphia, PA, 1976, Part 35, p. 219.

May 21, 1979.

PE

hams ew)

Cat ale (ere et

;

pee,

ane

We

4

DAY

GLigs ori Ns ea (02 ee is > >

ge

a

aris ea

ay.

4. ;

4

¥ia

ai

an

ator

Te Sete

We

h(t

SVG,

Cet s

AL! *$

| a

tig

a4, (Oe)

tes fire neta

re.

ae”

in

>

ad

ae oh

OTA wis fiat May SOA pe ie Pl gts as +e rr Mm ere jets aoa oe

Z

22S

b

“e's

Pg patie. Fi ej uiaw® qd

gy? “ agiire (Ge

a

mite

aoe

4a

|

©

ee

ing

“edie

lia

Vis

Tee

a

ahr oe gigenlaeer Fay

RAT,

ee

aa Ge hang

beas,

ee

nee we Gs Le

nafi

Tepew “aeacca

miggh baad

Apes

+

iwit ys aoe 8p Tee

Ore,

=.

:

a ls ehaleys

'

Y

2am

-

tn

aa

* sy 4 aaron § har > Wheat tity i! ep es

ee)

a

Sean, oa

:

ean

att,

ys

revt nd Paneba0 a

Sve

t.

Dat

a

7

1 oes

=

:

7

:

civ

en

a

mo

bse

a

pte’ Ts

ina + a)

#4

2

et

ie.

pay At.

+7ops ne

“ es chs

; tes

~ o(

ae

©

tino

os

:

t

i

oe ee

atlary

n4 be

ero)

Celie

+

tz

oS

Vad

thee)

yay tied

~

os

staph

mit ae

clgneuieale

~~ "a4 wsmg

rh ae

a

a

ah.

o ec!

tel

:

ve

eee eh

AP

ol

Shes

Sy |

2 OUTB peTal pee

i

ve :

i

thee

a

oh

ot? er

yeh

a4

7

(en

It Effect of Cross-Link Density Distribution on the Engineering Behavior of Epoxies S. C. MISRA,

J. A. MANSON,

and L. H. SPERLING

Materials Research Center, Lehigh University, Bethlehem, PA 18015

Fundamental knowledge of relationships between characteristics, synthesis or processing, structure, and mechanical and other properties is required of crosslinked networks used for critical and demanding applications. Thus, epoxy resins, which are widely used for engineering purposes, have received great attention in the past decade. Indeed, many papers have covered many aspects of epoxies such as: stoichiometry (1-5), prepolymer structure (6-8), diluents (4,9,12), fillers (13-22), heat treatment (3,22), and effects of cure conditions (23525) However, relatively little attention has been given to properties such as creep and to the effects of the distribution of crosslink density. Certainly typical commercial epoxy resins usually

exhibit a distribution of molecular weights which should result in a distribution of crosslink density in the final network. The broadening of the transition region has been qualitatively attributed to this distribution effect (26). Unfortunately, existing studies were of insufficient scope to permit correlation of the engineering behavior of epoxy networks and the distribution of crosslink density or Mc (the average molecular weight between crosslinks).

Because

of

its

relevance

to

engineering

application,

such

a study was thought to be of considerable interest. With this in mind, a program was begun to examine the effects of crosslinking and crosslink distribution on several aspects of behavior of high-T, epoxies (27-31). It was decided that a thorough characterization of viscoelastic behavior in blends of two epoxy resins, each having a quite different molecular weight, should enable correlations of distributions of Me and other network variables with the engineering behavior. Accordingly, the

crosslink density (M.) was varied by curing a homologous series of bisphenol-A-based epoxy prepolymers with methylene dianiline (MDA). Networks were also prepared at constant crosslink densities by blending low and high-molecular-weight members of the homologous This paper summarizes the following properties of the series. networks:

the

state

of

cure,

M,,

dynamic

mechanical

0-8412-0525-6/79/47-114-137$05.00/0

© 1979 American Chemical Society

spectroscopy

(DMS), creep, and stress-strain detailed results for each topic separately.

CHEMISTRY

RESIN

EPOXY

138

More discussed

and impact behavior. will be presented and

Experimental

The following Epon prepolymers (Shell Chemical Materials. Epons 825, 828, 834 (all liquid to semi— Company) were used: Approximate comand 1004 (all solids). 1002, solid); and 1001,

of polymerization)

(degree

positions

were

by gel

determined

per-

Bimodal blends were made by mixing Epon meation chromatography. Sample preparation and 825 with Epon 1004, as shown in Table 1. the method of curing have been described in the preceding paper

(29) « Table

I.

Compositions

Sample

Mo

E-1 E-2 F-1 F-2 F-3 E-3 F-4 E-5 F-5 E-7 E-4

of

Homopolymers

Epoxy

308 326 326 413 419 430 430 740 740 1004 493

Resin

and

Bimodal

(wt.

825 (100) 828 (100) 8255 (91) e004 8251(62) = 0045 825 (60) + 1004 834 (100) 5) (C57) ae AO: 1001 (100) 825 *(20)-+ 1004 1400 (100) 834 (100) (with

Blends

% in blend)

C9) (33) (40) (83) (80) slight

excess

ee ee eee

Materials Characterization. Measurements of complex moduli, including the storage modulus E', loss modulus E",

loss tangent model DDV-II

made

of

ees Young's and

tan 6, were made using a Rheovibron viscoelastometer, (Toyo Measuring Instrument Co.). Measurements were

at 110 Hz over

a temperature

range

from

-80°C

to about

40°C

above Tg, at a heating rate or 1°C/min. Icing of the specimens at low temperatures was avoided by sweeping dry nitrogen through the chamber. Tensile measurements were made using an Instron tester at room temperature, according to ASTM test D-638-68, type IV

with

a crosshead

speed

of Immfmin

(0.05"/min).

Young's

modulus,

E, ultimate tensile strength, Oqen YLeld strength. Oy» and ultimate elongation, ¢€,, were determined. In all cases, three to four specimens per sample were obtained that did not fail prematurely.

Impact tests were conducted according to the ASTM test D-256-70, method B (Charpy type, notched). On the average, 4 to 5 specimens were tested. Studies of creep as a function of time and temperature were carried

out

using

a

Gehman

Torsional

Stiffness

Tester

(32).

Creep

1].

MISRA ET AL.

Cross-Link Density Distribution

139

experiments covered the entire range of viscoelastic behavior from the glassy to the rubbery region. Modulus readings were taken at various intervals of time between 10 sec and 1000 sec. The inverse time-dependent compliance, 1/J(t), was calculated from the angle of twist of the specimen induced by the torsion wire as indicated by the apparatus, and the time-dependent Young's moduMaster lus, E(t), was assumed to approximately equal 3/J(t). curves were obtained by shifting data obtained at various temperatures with respect to a reference temperature (129°C in this Each curve of reduced modulus was shifted with respect to case). the curve at the reference temperature until all fit together to Master curves for various materials could form a smooth curve. then be easily compared. Temperatures corresponding to a value of 0.2 GPa for E(10 sec) were taken as T,'s. Average swelling and soluble ourer. measurements were conducted at room temperature using acetone as the solvent. Equilibrium swelling was achieved within 15 days, and the swelling ratio, q, was calculated by the formula: swollen volume of S e

extracted,

the

dry volume

Density measurements the density gradient

network

of

the

(1)

network

were made at room temperature, 20°C, using method. Duplicate runs were made in all

cases. A differential scanning calorimeter (DSC Corporation) was used to determine the extent

20-mg

specimens

were

tested

at

a scanning

1-B, Perkin-Elmer of cure; 10-mg to

rate

of 10°C/min.

An

exothermic peak on the thermograph indicates the heat of reaction whereas an endothermic peak in the amorphous polymer indicates the presence of residual stresses or the occurrence of a transition such as the glass transition. The presence of an exothermic peak in the DSC-scan of a pre-cured sample is an indication of incom-

plete

curing. The unreacted epoxy and amine groups present in the samples (degree of cure) were determined by chemical titration, using the procedure outlined by Bell (33) with a slight modification. Cured epoxy samples were filed and about 1 g of the resulting powder was placed in a round-bottomed flask with 50 ml of 0.2N pyridinium and 25 ml of distilled 50 ml of reagent isopropanol, chloride, The mixture was refluxed for 30 min and stirred continuwater. ously

the meter

with

a magnetic

solution and

was

a buret,

After

cooling

a 250-ml

beaker.

stirrer.

poured

into

standardised

0.505N

NaOH

to

room

Using was

added

temperature,

a potentioin

5-ml

From this point on, increments until the pH increased to 7.0. 0.2-ml increments were added until a sharp increase in pH was Since the response in the amine titration noted (pH 7 10.5). region was very slow and the pH did not stablize ommiestil te jee) 10) wanstin after each addition of NaOH, the readings were made at 10-min intervals.

RESIN

EPOXY

140

Results

and

CHEMISTRY

Discussion

Distribution of molecular weights in the commercial resins. Gel permeation chromatography (GPC) indicated that the liquid and semi-solid prepolymers (Epon 825, 828, and 834) were composed of

cH ie CH, -CH-CH, 702) OE: H e 3

/\

0 (M = 340) and m= 0 and m = 1 species

m = m =

weight

of

Table

the

Sample

as

Composition

epee

M

E-1

825

352

E-2 E-3 E-5 E-6 E-7

828 834 1001 1002 1004

380 592 1000 1360 1996

Resin

m

Ore

The 1 (M = 620) components. was dependent on the average

prepolymer,

II.

OH CH \ ie l H-CH,+O-€)6- (2)-0CH,,-CH-CH,

given of

the

Table

in

m=0

(96%) m=0 (86%) m=0 (29%) m=l, m=3, m=l, m=3, m=l, m=3,

present

and

and and and and and

of

II.

Commercial

Components

percentage molecular

Epoxy in

Resins the

prepolymer

m=1

(4%) m=l1 (14%) m=1 (712) m=7 (major component) m=12 (major component) m=l2 (major component)

The solid prepolymers were composed as follows: Epon 1001, m=/ (M=2328), m=3 (M=1192), and m=1; and Epon 1002 and 1004, m=12 (M=3748), m=3, and m=l (the major component being m=12). The exact composition of each component in the solid prepolymers could

not

be completely

resolved

from

the

peaks

of m=7(Epm

1001)

or

m=12

(Epm 1002 and 1004) components, Thus, the bimodal blends (Series F) of Epon 825 and Epon 1004 had the following components: m=0, m=l1, m=3, and m=12. The amount of m=0 component was at a maximum in sample F-l but at a minumum in sample F-5. On the other hand, samples F-2, F-3, and F-4 contained both m=0 and m=12 species in appreciable quantities. Degree of cure. The titration results given in Table III reveal that from 98% to 100% of the functional groups had reacted. Similarly no evidence for incomplete cure was observed by DSC or by dynamic mechanical spectroscopy (DMS). However, it may be pointed out that 2% unreacted functional groups could result in detectable incoherence (see reference 29) in the networks prepared

from high-MW

epoxy

prepolymers.

1].

MISRA ET AL.

Cross-Link Density Distribution

141

Table III. Unreacted Epoxy and Amine Concentration in the Cured Sample Sample

Epoxy

content,

meq/g

Amine

sample

content,

meq/g

sample

E-1 B=2 R=3 E-4 E-5 E-6 F=1 B72 F=3

0.00 0.06 0.00 0.09 0.04 0.03 0.00 0.00 0.03

.00 =OOZ 0.00 -0.04 0.00 0.00 0.00 = O09) 0.03

F-4

0.04

0.04

i=)

0.09

0.06

Dynamic mechanical properties. The parameters characterizing the dynamic mechanical behavior of Series E and Series F are summarized in Table IV. Several of the viscoelastic parameters will now be discussed in detail. Table IV. Series

mM? ©

Specimen

ce en ee E-1 F-2 E=3 E-4 E-5 E-7 F-1 F-2 F-3 F-4 F-5

TP B

SEF)

WEEE

RRRCERT a DLT g sas

max

Jalse

setacc

°C

of

Ca cpedyl

Et

“Based

35) =40 46 -46 550 9-60) “40 -45. -45 ho eS

on M.'s of maximum

E' refers

to

in

the

0/62 6.75 0.90 0.95 10s 91-52 0.75 0.75 0.80 075 el 25 the tan

21 11 175 175 157 138 211 185 175 168 157

components 6

glassy

state

“18¢pa = 1010 dynes/cm f= 3d Clog 7E))/dT oat Ty (+ LEG)

D07 0845 C97 Or 42 Moe: (opal) 165 0.21 145 0.10 175 eOrO65e 1o7 = OL02 17 021 T66o Oe 7 150) 017 eeOrt 125

Sn

Fe

Oe

Whos Fail HO, 00 19) "le. 21 Dh OT ee 19

ee 5.6 waley ony 2975S) SELOTO 5.6 a 5.9 eS TG 9.1

I ET Se ME

ae 326 430. 493 Tae 14007, 326 413 419 aS 70)

From

Dynamic Mechanical Properties E and Series F Epoxy Networks

RESIN

EPOXY

142

CHEMISTRY

The effect of the distribution At M, = 430, the T of the

Glass transition temperature. of M. on T, is shown in Figure 1.

commércial longer chains, chains (see

blend ae ate F=-4 was 13°C lower than that of its This indicates that the counterpart sample E-3. amount as the shorter same the in present though

Table I), dominate the final network and thus control the Ty of Due to the higher reactivity, the longer the blend sample F-4. molecules react first (29) to give microgels swollen with the unreacted smaller molecules which react later to form an interpenetrating network. Thus the secondary microgels result in a On the other hand, at M. = shell and core type structure (29). 326 and M, = 740, the blend samples F-1 and F-5 did not have a T,

from

different

their

E-2

counterparts,

commercial

E-5,

and

respectively. At these M.'s, large differences in distribution between blends and commercial resins could not be obtained because the number of possible permutations of distribution at these Mo's was very restricted; at lowM, (326), Epon 825 was the major component (90%), whereas at high Me (740) Epon 1004 was the major component (80%).

Young's long-range

modulus. segmental

temperature,

E!,

was

As

expected

motions found

to

are be

for

the

glassy

frozen), nearly

state

Young's

(in which

modulus

independent

of

at

Ma, Of,

room Ehe

distribution of M, (Table IV). All values fall within the range of 2 + 0.2 GPa compared to an average of 1.3 GPa obtained from tensile

stress-strain

Rubbery the

present

Graessley

curves

modulus. networks

(34)

is

(to

be

discussed

later).

Figure 2 shows that the rubbery unaffected by the distribution

calculated

the

free

energy

of

modulus of Mes

deformation

and

of the

equilibrium dimensions of network chains for networks of Gaussian random coils and showed that the elastic properties of such networks depend upon the total number of elastically active chains (average M,) and elastically active junctions, but are independent of the chain length distribution (M, distribution), junction functionality distribution, and the detailed pattern of connectivity. Thus, the present experimental results are in agreement with the theoretical predictions of Graessley. B-transition

increased

with

bution

M.

of

temperature.

M, but

(Figure

was 2).

The

§-transition

unaffected

by the

This

because

is

so

temperature

changes Tp,

is

in a

the

result

distriof

the

motion of small segments of the main chain, rather than the larger segments involved in the glass transition. Therefore, Tg should depend on the average M, and not on the distribution of Mo:

Height of the tan 6 peak. The value of the tan 6 increased asymptotically with M.. As expected, the bimodal samples F-2, F-3, and F-4, than their counterparts of

which showed broader the commercial resin

transition samples of

blend regions Series

11.

MISRA ET AL.

Cross-Link Density Distribution

143

160

140

300

Figure 1.

400

600

800

1000

1500

Glass transition temperature as a function of M, for Series E and Series F networks: (O) commercial resins; (()) bimodal blends

EPOXY

144

4,0

UNEXTRACTED

4,O

EXTRACTED NETWORKS

RESIN

CHEMISTRY

NETWORKS

—O.

g/ml p,

"4

O

COMMERCIAL

4

BIMODAL

RESINS

BLENDS

J/M

STRENGTH IMPACT 300

400

700

1000

1500

Me

Figure 2.

Impact strength, E,, Tg, and density as a function of M, for Series E and Series F networks

1]. E,

MISRA ET AL. also

showed

broadening M.-

Cross-Link Density Distribution broader

Slope of the temperature curve Moe

of

In

the

and

is attributed

(mn),

the

tan

6 peaks

broadening

transition region. The in the glass transition

accordance

slope

flatter

to

to

the

creep

obtained

from

results

the

145

of

(Figure

the

3).

This

distribution

of

slope of the E' vs. region also varied with

(discussed

dynamic

later)

mechanical

values

data

(see

Figure 3), of the blend samples F-1, F-2, F-3, and F-4 were close In contrast, the to that of the commercial-resin sample E-l. slope of the blend sample F-5 was steeper than that of its commercial counterpart, The broadening of the sample E-5. region of the blend samples F-1 to F-4 is attributed sence of broad distributions of M, in them.

Tensile behavior. Tensile properties fell in ranges, as shown in Figure 4: (Control Series E),

0.3)

GPa,

o,,

Ey = 150.(+ trends as a

= 71 (+7)

MPa,

transition to the pre-

the following E, = 1.2 (+

c¢ = 11 (+ 2) %; (Bimodal

053) GPa, oy = poe Pa function of M, were observed,

Series F),

eos 9 (eZ). in contrast

2. NO to the case of M, changed by changing stoichiometry (27,28). Thus the tensile properties at room temperature are essentially independent of M. and its distribution. This general relative insensitivity to changes in M, and its distribution is not surprising, because the glassy state properties should be more closely related to the cohesive energy density than to network structure (35). Furthermore, the shift factors (log ap) obtained from creep experiments (see below) indicate that the motion of the main chains of these

networks is very restricted. Therefore, as observed experimentally, it can be expected that the room-temperature tensile properties of these epoxy networks having bulky bisphenol-A groups would be independent of average M. or distribution of M.. Impact strengths varied inversely with M., values ranging from 50 J/m for E-l to 34 J/m for E-6. However, as shown in Figure 2, impact strength was not affected by changes in the distribution of M..

shift glass

Creep behavior. Figure 5 shows the composite curve of the factors and also indicates that the shift factors in the transition region do not follow the WLF equation: log

aT 7

GRC

aleie era

they follow an equation Instead, below (also see Figure 6):

log

a

a

of

-H,/ 2.303

@)

AF eet) the

RT

Arrhenius

type

given

(3)

where R is the gas constant, T is the absolute temperature, and Thus, shift factors for all H, is the apparent activation energy. samples E-5, E-6, and including study, present the of samples the E-7 which have relatively long chains and would be expected to be

EPOXY RESIN CHEMISTRY

146

-n

SLOPE,

6

TAN MAX

300

500

700

1000

1500

ec Figure 3.

Tan 8maz and slope of the transition region as a function of M.: (O) commercial resins; (/\) bimodal blends

ll.

MISRA ET AL.

Cross-Link Density Distribution

147

~w

%ELONGATION

MPa G2

GPa Ey

300

Figure 4.

400

600

800

1000

1500

Tensile properties of Series E and Series F networks as a function of M.: (O) commercial resins; (A) bimodal blends

EPOXY

148

RESIN

CHEMISTRY

10

Figure 5.

Shift factors as a function of (T — T,) for Series E and Series F networks

11.

MISRA ET AL.

Cross-Link Density Distribution

149

LOG ay

40)

Figure 6.

Shift factors as a function of (1/T — 1/T,) for Series E and Series F networks

EPOXY

150

RESIN CHEMISTRY

flexible, surprisingly gave a better fit to an Arrhenius relationship as is generally the case for polymers in the glassy Thus it appears that the motion of the main chains of the state. networks is relatively restricted even beyond the epoxy present

It appears that glassy state. responsible for the restricted even beyond the glassy state.

the

bulky

bisphenol-A

groups

are

motion of the main chains observed It was also noted that the activation energy was independent of M, and distribution of M, in those networks; values of 230 kcal/mole and 75 kcal/mole above and below The values are in good agreefit all samples. T , respectively, mént with values determined here for bisphenol-A-based epoxies cured

with

MDA

(31)

and

elsewhere

for

similar

epoxies

cured

with

amines Cow) To determine the effect of the distribution of M, on creep, master curves for all the samples were drawn at 129°C instead of selecting T,_ as the reference temperature. The temperature corrections were not used to draw the master curves shown in Figure 7 because these corrections, as mentioned earlier, were small and the experimental data did not extend to long ranges of temperature in the rubbery region. The characteristic creep time (to) was taken as the time required to relax to a value of log E(t) = (log Eg + log Er)/2, where E, and E, are the glassy and rubbery moduli, respectively. The slopes (n) of the master curves were determined at the time t,. The characteristic creep time (t,) was found to be inversely related to M, by the following relationship:

log (rhc)

= K,/M,

(4)

where log t = -10.1 = Com at M, = ©, obtained by extrapolation), and) Kj >= 5.4 9x 103 (a constant). The statistical correlation coefficient was 0.99, showing excellent fit. Figure 8 shows that the blend samples F-1 and F-5 appear to have higher t,'s than their counterparts E-2 and E-5, respectively, which were prepared from commercial resins. On the other hand, sample F-4 has a lower t, than its commercial counterpart E-3. These results indicate that the differences in distribution of Me affect to. Samples E-2 and E-3 have network chains composed of epoxy molecules having molecular weights of 340 and 620, respectively, whereas the major portions of samples E-5,-E-6, and E-7 are composed of epoxy molecules having a M.W. of 2328 or 3748. Furthermore, the blend samples F-l and F-5 have a larger proportion of epoxy molecules having a molecular weight of 340 than their

respective

counterparts

E-2

and

E-5

(commercial

resins);

this results in a higher t,. On the other hand, the blend sample F-4 has two epoxy species having M.W.'sof 340 and 3748, respectively, in equal amounts and its commercial counterpart sample E-3 has epoxy species having M.W.'s of 340 and 620. The appreciable amount of large molecules in F-4 results in lower ta Thus, it is concluded that t,, besides being sensitive to M.;, is also very sensitive to the distribution of Mas tT. would be lareerjor

1].

MISRA ET AL.

Cross-Link Density Distribution

151

2

Pa 3/y(t)s DYNES/CM 3/u(t), 10

LoG(t),

Figure 7.

sec

Master curves for Series E and Series F networks (T, = 129°C)

?

-nx10

SLOPE,

300

Figure 8.

500

1000

1500

Characteristic creep time and slope of the transition region as a func-

tion of M, for Series E and Series F networks: (A, A) commercial resins;

bimodal blends

(O, @)

EPOXY

152

chains,

or

a decrease

in M,

whether the network has greater amounts of respectively. transition region was found to broaden with

upon

depending larger The slope of

smaller

smaller

the

(Figure

thatean

results in the broadening from dynamic For example,

and

Shibayama

of polyesters,

indicate

results

These

8).

increase in crosslink density slope, as is often observed.

studies

RESIN CHEMISTRY

(36)

Suzuki

had

of the mechanical

shown

also

density. Furthermore, the bimodal blend samples F-1 to F-4 have slopes close to that of sample E-l1, prepared from the commercial resin whereas sample F-5 (bimodal blend) has a slope close to that of sample E-7 (commercial resin). The lower values of slope observed in the blend samples F-1 to F-4 are a result of a broader distribution of relaxation times which is not present in the respective counterparts of Series E. Similarly, sample F-5 showed a higher slope than its counterpart E-5 (even though the slope measured for E-5 is lower than it should be), because the former had a narrower distribution of molecular weights which resulted in a narrower distribution of relaxation times (see Table I). It is concluded that the broadening of the transition region could take place either by an increase in the average crosslink density or by the broadening of the distribution of crosslink density. that

the

dispersion

broadens

an

with

increase

in

crosslink

Swelling and extraction. Swelling and extraction data are shown in Figure 9. The T, of the blend sample F-4, which had 57Z Epon 825 and 43% Epon 1008, had indicated that Epon 1004 forms the more continuous network. The swell ratio of this sample is the same as that of its commercial counterpart, sample E-3. Similarly the blend samples F-l1 and F-5 also have swell ratios that are the same as their respective commercial counterparts E-2 and E-5. This indicates that the swell ratio is not very sensitive to the distribution of M.On the other hand, samples F-2 and F-3 have swell

ratios

that

are

lower

than

expected

(see

Figure

9).

These

samples have higher proportions of low-molecular-weight prepolymer (i.e. 62 and 60% by weight, respectively). It is highly probable that the composition of these samples is in a critical range in which

a

network another phase.

phase

inversion

exists;

in

other

words,

some

part

of

the

probably has Epon 1004 as the more continuous phase while part of the network has Epon 825 as the more continuous Such a mixed network might well result in anomalous swell-

ing behavior,

with

slightly

lower

values

of

the

swell

ratio

than

expected if the low-M epoxy network dominates swelling. The bulk density of the samples at room temperature decreased with M. and, as expected, was not affected by the distribution of M, at a constant M,, as shown in Figure 2.

ll.

MISRA ET AL.

Cross-Link Density Distribution

153

q

RATIO, SWELL

%EXTRACTION

300

Figure 9.

400

600

800

1000

1500

Swell ratio and % extraction as a function of M, for Series E and Series F networks: (O) commercial resins; (A\) bimodal blends

EPOXY

154

RESIN CHEMISTRY

Conclusions

of conclusions.

of Luts and

The effects in the following Properties

affected

by

its

the

distribution

distribution

of

may

be

summarized

Mo.

The glass transition temperature, as is well known, I. Slight changes in the distribuincreases with a decrease in M.. However, when significant changes in tion of M, do not affect T,. the distribution of M, exist, Ty is governed by the dominant com-— ponent in the final network. 2. The slope of the transition region at Ty increases with M, but decreases (i.e. broadens) with an increase in the breadth of the distribution of M. 3. The height of the tan 6 peak increases with Mc. Furthermore, as expected, an increase in the breadth of the distribution of M, results in a broadening of the tan 6 peak along with a decrease in the height of the peak. 4. ‘The characteristic creep time |(t.) 1s the most)sen— sitive variable to the distribution of M,. Its value is higher or lower depending on whether the network is composed of higher fractions of short or long chains, respectively. Properties nora becteds Dye Lie salSEEl bul Toma teiier I. The rubbery modulus (E,) decreases with an increase in Me but as predicted theoretically @4) is not affected by the

distribution sent Mae main

of Me.

2. The room-temperature tensile properties for the preepoxy networks are independent of M,. or the distribution of This, independence is ascribed to the restricted motion ez sine chains, even above T,, as deduced from the creep experiments.

1S not

3. The soluble content artected by changes in the

increases steadily with distribution of M..

M, but

4. The density increases with a decrease in Ma, buts affected by changes in the distribution of M,. 5. The swell ratio increases with M, but is unaffected by the distribution of M., except in some unusual networks having a phase inversion. 6. The room-temperature impact strength increases with a decrease in M, but is unaffected by the distribution of Moc

not

7.

decrease

The

§-transition

temperature

increases

with

a

in Me but is not affected by the distribution of M.: 8. The experimental shift factors, for DGEBA-epoxy networks, are independent of IMs Ole isle: Ghusjeresjayhjestiow @ae Ii Instead of following the WLF equation, values of the shift factor follow an Arrhenius-type relationship with temperature, indicating: that the chain segment mobility is restricted even in the rubbery Staten

1].

MISRA ET AL.

Cross-Link Density Distribution

155

Acknowledgements

The

authors

Materials

wish

to

DE.

DESCUSSTONSHWLED

acknowledge

through

Laboratory,

Kim

Sool.

support

Contract and

the

help

Products & Chemicals, Inc.) in performing tions are also very much appreciated. Literature

from

the

Air

Force

F33615-75-C-5167.

No.

Mr.

of

the

C.

Worman

(Air

characteriza-

GPC

Cited

ik4

MAGWeehyeineh. WN, ehovel BY Gy Ws ah, Polym, Sci. (1970), A-2(8), ae ae 437. Topand Kiine, De EY eitats J. Appl. Polym. Set. (1972), 16, SUAS).

-) 4.

Hirai, I. aud Kline). &. J. Appl, Polym. sci sulk Whiting, D. A. and Kline, D. E. J. Appl. Polym. hs

(19/3) Sci.

7217,

(1974),

ANOVAs}.

a O:

Selliyys

Ks

aimal Mbuiibeies

ih, We

dis Wee,

Seite

GLO),

1,

i22

8.

HP

Kitoh, M. and Suzuki, K. Kobunshi Ronbunshu (1976), 33, 1, 10% NenyOn AO Em and sINielsenk macmt mie MaciOmollmio Glen mmGlhlem. CEDOFeen S(2) 9 22). aca. No acd yamuchi, Rk. J. Appl.- Polym. Sei. (1973), 17, LT Sse eeiota aNerodd Kumanotant, J... Appl. Polym. Sel. (1977), 21,

i222

PAE WES Moehlenpah,

bata, N. and Kumanotani, J., J. Appl. Polym. Sei. (1971), 15, DEA es (OUST ikem bands hind chen Ge ieee TOCeeGCOlhs Reinte> laste, ConposeavA Ven SOCe last Lidl Oy) na 27s. sein alles

Tee

iO.

13.

Ay.

J. Appl. Polym.

Tee 16.

Manson, J. A, 14(1), 469. Mansons) ..A. CUQTE5 DSc mecks Ace. DeRosset, W.

I7.

Bascom,

ie

W.2D.;

E.;

Sci.

Ishai,

O5;

and

Dibenedetto,

(1969), 13, 1231.

and

Chiu,

E. H.

ACS

and

Chiu,

E.

156 23%

24. re Zor ile Zoe

BY). 30. SUL Be 3\3he 34. BD 36. Tc

EPOXY

Morgan,

and

J.

R.

O'Neil,

E.

J.

in "Chemistry

RESIN

and

CHEMISTRY

Properties

Ed. by S. S. Labana, Academic of Crosslinked Polymers.” see Oo: Clog) PEeGSSyaN«ed Suzuki, K., et.al., Kobunshi Ronbunshw, (19/76) ,893(3)e047 oe Brett, Cx Le oJ. Apple Polym, Sela) (1976) 20 Gar. Nielsen, LE.) J. wacromol. scl. (lv69) Css 80s. eva, wie ANGS Syoyeveilslnys, Ths Isl, Guatel Kalin, So Iho Iuael Mepeisie. AFML Contract No. F33615-75-C-5167, 1977 (AFML-TR-77-109. Kam

ore Lise

OK

bon

Mase

Mansorls

Ji

mars

Hergtez )Cie Cram Roem Wisma

Janiszewski, J. ACS Org. Coatings and Plastics Chemistry (1977), 37, 2445 in press, Polym. Eng. Sei. nici, So 5g Mineo, di, WN. ele Sixeeibtoe. Iho il, Werks yuiilai= Catlone mehoor Misra, S. C. Ph.D. Dissertation, Department of Chemical Engineering, Lehigh University, 1978. KM eo Lise hanson. A aeande MESSE mC. mee id Sm OUD lle tes@rl WOT. ASTM D-1043, American Society of Testing Materials, PhiladelDida

earn

Beil

eds Pee

Graessley,

O72) re W.

Pol yine Sele W.

(Lor 0)ceAe Coal Te

(1975), 8(6), 865. Kaelble, D. A. In "Epoxy Resin Chemistry and Technology." Ed. by C. A. May and Y. Tanaka, M. Dekker, 1973, Chap. 5. Shibayama, K. and Suzuki, Y. J. Polym. Sci. (1965), A-3, OSI « Kito, M. and Suzuki, K. Kobunshi Ronbunshu (1976), 33, 19.

RECEIVED May 21, 1979.

Macromolecules

2 Network Morphology and the Mechanical Behavior of Epoxies S. C. MISRA, J. A. MANSON,

and L. H. SPERLING 627 Lehivh Universit

Materials Research Center, Coxe Laborato

Bethlehem, PA 18015

Does

igh

University,

Since the properties of thermosetting polymers depend on their network structures, the morphology of such polymers has long been a subject of considerable theoretical and practical interest. Thus, it is well known that the tensile strength of thermoset resins is less than predicted theoretically on the basis of the breakage of primary van der Waal's bonds (1,2), and it was proposed long ago (3) that this discrepancy was due to the rupture of weak regions created during network formation. Indeed, it has been shown (4) that high internal stresses can be developed during curing, especially when the curing rate is low. This view of the role of structural defects is also given credence by modern theories of fracture mechanics (5), which emphasize the concept of the concentration of stress at a flaw. Two-Phase

Networks

Other investigators (4,6-33) have emphasized a view of the network essentially as a composite, with a high-crosslink-density (essentially spherical) phase (often considered as a microgel) embedded in a less-crosslinked matrix. In fact, it is probably generally accepted that, regardless of specific details, the curing of thermosets results in an inhomogeneous, two-phase network. Inhomogeneity has been attributed to incompatibility or to non-optimum curing conditions (8) but has also been proposed to be inherently characteristic of all gelling systems (14). It has also been postulated nature, at least for

(7) that these microgels some epoxy systems.

are

colloidal

in

The existence of inhomogeneities has been inferred from results of diverse investigations using techniques such as electhermomechanical tron and optical microscopy (13-26, 29,30,31,32), (34), differential swelling (12,35), and dirtferential measurements

In contrast, the use of microtomed scanning calorimetry (36). thin sections and small-angle x-ray scattering fail to indicate Dispersed phases have been referred to two-phase structures (37). by such terms as "micelles," "globules," "floccules," "nodules," In this study, the term "microgel" will be used. and "microgels."

0-8412-0525-6/79/47-114-157$06.50/0 © 1979 American Chemical Society

EPOXY

158

It-has

been

suggested

by Bobalek

et al.

(27),

RESIN CHEMISTRY

Solomon

C32)"

and Labana et al. (14) that the two-phase system is produced by Some inmicrogelation prior to the formation of the macrogel. vestigators (4,14,16) have postulated that these microgels are loosely connected to the surrounding matrix, and have also suggested (12) that these loose connections are developed during the latter stages of the curing process. In general, two levels of sizes have been reported in the literature--one (type A) ranging from 6 nm to 40 nm, and the It has also been other (type B) ranging from 20 um to 200 pm. shown (4) that slow curing rates result in larger microgels (type Bs which result in a network having a higher Ty, density, and resistance to etching. The surface properties of the network depend on the surface energy of the mold material and on the The size and density of microatmospheric environment (7,30). gels have also been related to the presence of plasticizer (23), radiation damage (26), prolonged exposure to heat (17) and aging OLE hemTe Summ C24)re Studies of interfaces (30,31) have also shown that certain substrates such as Teflon and silicone-coated sheets give featureless surfaces; however, subsequent etching of the surface reveals a two-phase structure. It was further shown that the microgel size decreases with increasing amounts of catalyst. Both low tensile strengths and nodular morphology in thermosets have been related to differences in crosslink density (/,10; 14,27,28,32). At the same time, the fact that the yield strength of some epoxies is fairly independent of stoichiometry has been attributed to the role of microgels as primary flow particles (29). Thus, morphology must play an important role in determining network properties. However, surface morphology should not affect the mechanical behavior of the network as much as bulk morphology. Diffusion phenomena, on the other hand, should depend on the morphology of both the outer surface and of the bulk. Though there have been several postulates of morphological changes during the crosslinking process and of the morphology of the final network, experimental evidence has been scarce. As part of a comprehensive study (38) of the effects of crosslink density on the behavior of epoxies and other polymers, the present morphological investigation was conducted on epoxies with emphasis on the roles of stoichiometry, molecular weight of the epoxy prepolymer, and the distribution of molecular weight bet— ween crosslinks (M,). Bisphenol-A-based prepolymers were used throughout; most were cured with methylene dianiline (MDA), a few were cured with a polyamide. Experimental

Table

Sample Design. 1. To examine

The the

various formulations are effect of stoichiometry,

described in Series A was

12.

MISRA ET AL.

prepared Chemical of M, at

Network Morphology and Mechanical Behavior

159

using different proportions of Epon 828 with MDA (Shell Co.). To provide a baseline for examining the effects constant stoichiometry, Series E, based on the following

prepolymers,

was

evringsagent.

prepared

“Hpon

$25,

using

°828;

stoichiometric

834,

°1001,°1002,

amounts

and

of

1004

MDA

as

(Shell

Chemical Co.) To vary the distribution of Me, series f was pre— pared to yield resins which had the same average M, values as This was done by blending Epon 828, Epon 834, and Epon 1001. Epon 825 (narrow distribution of molecular weight) with various proportions of Epon 1004. The distributions of M. thus obtained Finally, to comare essentially bimodal, with little overlap. pare behavior with that of a polyamide-cured epoxy, Sample G-1

was prepared Versamid 140

by curing Epon (General Mills

Table

I.

Series

Sample

828

with

Chemical

an

Compositions of the Samples Epoxy Resin Series A, E, and A

Series

Amine/epoxy

No.

Ratio

A-7 A-8 A-10

Wg FAsib @ Seal Ihs@eoa

A-11 A-14

easel 4anar

M29?

proportion

equivalent

Prepared F. E and

Sample

epoxy

No.

Gute)

1523} 526 326

E-1 E-2 F-L

825(100) 828(100) 825(91)+1004 (9)

308 326 326

370 592

F-2 F-3

825(62)4+1004 (38) 825(60)+1004 (40)

413 419

430

iLa(@gik

924

E-3

834 (100)

ikateieall

1922

F-4

825(57)+1004

A-20

2 Oe ib

E-5

1001 (100)

Gia

F-5 E-6 | E-7

using

Bell's

F

Mo

A-16

“calculated

in

Series

A-18

ee reasons discussed mens A-16 to A-20 may error will not affect stoichiometry.

of

Co.).

equation

825(20)+1004 1002(100) 1004 (100)

(43)

430

740 (80)

740 980 1400

(39).

by Bell (39), actual M. values for specibe somewhat in error. In any case, the the trends observed as a function of

Se

Resin

Preparation.

For

systems

using

liquid

epoxy

prepoly-

mers (e.g., Epon 825 and 828) with MDA as the curing agent, the After heating curing cycle was similar to that used by Bell (39). to 80°C, the resin and curing agent were mixed together, evacuated 5 to 15 min. to remove air bubbles, and cast and cured as 45 min. in a circulating air oven at 60°C, 30 min. at follows: atel50-C,vand finally slow cooling to room sent. 60°C 92,

for

EPOXY

160

RESIN

CHEMISTRY

The mold assemblies comprised clamped 13~-cm by 13temperature. cm Mylar sheets, separated by 0.5-m, 1.5-mm, or 6-mm Teflon or ethylene-propylene-copolymer spacers backed by glass plates. This cure cycle was reported by Bell to give essentially complete curing, and was also shown (through chemical titration Gir jelne epoxy and amine groups) to be similarly effective in this labora-

tory (38,41,42).

Solid epoxies were first melted and then evacuated to remove the In order to avoid air entrapment, the entrapped air bubbles. curing agent was mixed in, using a magnetic stirrer, under vacuum. Samples were prepared using the following curing cycle: 1.5 hr. tel On Gameollowedsby es lowsecooOliic stom GOom at OOkGeeZmon lina The Versamid-140-cured sample (G-1) was one of the temperature. samples prepared by Manson and Chiu (40), using the following cure cycle: overnight curing at room temperature, 2 hr. at 60°C, Dine, Me NOOO, amc 4 ie. ate WO" Using these techniques, it was possible to prepare reproducible samples suitable for testing, albeit with considerable difficulty in the case of solid epoxies. Electron Microscopy. Various etching techniques were tried to study the micro- and macro-structures of MDA-cured epoxy net-— works ("macro" implying a morphological feature on the scale of 1 um or larger). Etching may be presumed to preferentially attack regions of relatively lower local crosslink density. After the examination of etched samples under the ETEC Autoscan scanning electron microscope (SEM), it was concluded that etching for 7 hr. with 1 M aqueous Cr903 at 80°C was more effective than the following techniques: etching for 30 min with HF; etching for 15 days with acetone; etching for 10 hr. with argon at high voltage under vacuum. The study of etching time indicated that at least one hr. was needed to produce adequate etching with Cr503. Finally, etching times of 4 hr. and 7 hr. were found to be convenient for samples having M.'s above and below 500, respectively. The gross morphology was examined under the SEM while the fine structure was examined using a Phillips 300 transmission electron micro-

scope (TEM), ed surfaces.

using two-stage carbon-platinum replicas of the etchIt should be noted that one must be very careful in

interpreting

statements

nodule on the

about

on a two-stage replica polymer surface.

features

arises

seen

from

in replicas.

a dimple

(not

Thus

a

a nodule)

Isolation of Microgels. The following experiments were made to isolate microgels during the process of curing and from the : etchant solution: 1. The formation of microgels during the curing process was verified experimentally on samples E-2 and E-5 by dissolving them in acetone, before gelation. Sample E-2, which has a longer gelation time than E-5, was dissolved after curing at 60°C for 30 min and at 80°C for another 15 min whereas sample E-5 was

12.

MISRA ET AL.

Network Morphology and Mechanical Behavior

161

dissolved after curing for only 10 min at 100°C. The solution was diluted to a concentration of 100 ppm, and electron microscope grids were made and examined under the TEM. 2. Etched samples were removed, the etching solution was diluted

with

electron

deionized

microscope

Similarly,

a

less

water

grids

dilute

5) was placed on a clean examined under the SEM.

to

were

a very

low

prepared

solution

(from

microscope

concentration,

by platinum the

slide,

etching

allowed

and

shadowing. of

sample

to dry,

E-

and

Results

Figure 1 shows typical macrostructures, with feature sizes from ca. 10 um to 40 um, revealed by progressive etching by the aqueous Cr 903. The effects of different variables on morphology are discussed below.

Effect

under

the

of Stoichiometry. Examination of the etched surfaces SEM (Figures 2 and 3) revealed a two-phase structure.

Dimples were observed on the surface, the dimple size increased with deviation from equivalent stoichiometry. Samples having an excess of epoxy appeared to have fewer dimples than samples having a similar M, but excess amine. The sizes of the dimples varied from ca. 10 um to 70 um, depending on the percent excess of reactants, in general agreement with sizes reported by Cuthrell

(7)

above ences

an amine/epoxy ratio were negligible. The two-stage replicas

the

and

existence

Selby

of

and

two

Miller

(37).

of

It

1.6/1

examined

phaseson

a much

was

also

observed

that

the morphological

differ-

with

revealed

the

finer

TEM

scale

also

(Figure

4).

The

replicas showed nodules on the surfaces, corresponding to dimples on the polymer surfaces. The dimple sizes in these samples were

in the range of 25 nm to 50 nm except for the case of 100% excess amine (A-20) where a large distribution of sizes, ranging from 20 nm to 200 nm, was observed. The distribution of phase size tends to broaden with increasing proportions of excess amine; the average size also tended to increase with the proportion of excess amine. On the other hand, an increase in the amount of excess epoxy did not show significant changes from the morphology observed at equal stoichiometry. These observations differ from those of Racich and Koutsky (31), who found that the domain size decreased with increasing amounts of catalyst and with the preThey argued that sence of saturated vapor of the curing agent. the crosslink density increases with increasing amounts of catalyst,

in

resulting

indicate

that,

as

smaller

reported

domain

tends to increase with a decrease in the domain in the excess-amine case;

ed

by changes

in

crosslink

density

The

sizes.

by Racich

and

present

Koutsky,

the

results

domain

size

but only crosslink density, is not significantly affect-

for

the

excess-epoxy

case.

EPOXY

162

Figure I.

RESIN

CHEMISTRY

Micrographs of Cr,O;3-etched surfaces of Epon 1004-MDA (Specimen E-7). Etching times: (A) 2 hr; (B) 4 hr; (C) 7 hr; (D) 11 hr

12,

MISRA ET AL.

Network Morphology and Mechanical Behavior

Figure 2. Scanning electron micrographs of 7-hr-etched samples having different stoichiometric ratio of MDA to Epon 828: (A) 0.7/1; (B) 0.8/1; (C) 1/1; (D) 1.1/1

163

164

EPOXY

RESIN

CHEMISTRY

Figure 8. Scanning electron micrographs of 7-hr-etched l i j stoichiometric ratio of MDA to Epon 828: (A) 1.4/1; (B) 16/1:( C)an eyoe H

12.

MISRA ET AL.

Network Morphology and Mechanical Behavior

Figure 4. Transmission electron micrographs of replicas of 4-hr-etched samples having different stoichiometric ratio of MDA to Epon 828: (A) 0.7/1; (B) 0.8/1; (C) 1.1/1; (D) 1.4/1; (E) 1.6/1; (F) 2.0/1

165

EPOXY

166

RESIN

CHEMISTRY

When the Effect of Molecular Weight of the Prepolymer. weight of molecular the changing by changed was density crosslink similar to that the epoxy prepolymers, a two-phase structure, found for the case of variable stoichiometry, was observed even A marked difference in at the macro level (Figures 5 and 6). morphology was observed in networks prepared from liquid resins The as compared to those from solid or semi-solid resins. liquid-resin (Epon 825, 828) samples had dimples which had a dis— tribution of size ranging from 5 um to 25 um. In addition to dimples, small ridges were also seen in both agglomerated and isolated forms. In samples from semi-solid and solid resins (Epon 834, 1001, 1002 and 1004), etch-resistance (and, hence, presumably highly crosslinked) shells were observed throughout the cross-section (Figure 7), encapsulating the less crosslinked material. These shell-and-core-type structures were present in these

samples

in a

range

of

sizes

from

2.5

um

to

25

um.

It

was

found that the range narrowed but that the average size increased with an increase in the molecular weight of the prepolymer. This typical morphology was also observed in samples that were solvent cast from acetone solutions and cured after most of the acetone evaporated. Since it has been shown that curing does not take place in the presence of acetone in the Epon-MDA system (39), it can be assumed that the acetone-cast samples were similar to the bulk-cured samples. Indeed no differences in mechanical proper-

ties

in bulk

and

acetone-cast

samples

were

observed

(41,

Ch.

6).

Etched under similar conditions, solvent-cast samples showed the same shell-and-core-type morphology (Figure 7d), indicating that this typical morphology was not due to inadequate mixing. All the samples made from solid or semi-solid prepolymers had the same fine structure, with the size of the discontinuous phase in the range from 15 nm to 25 nm, whereas the samples made from liquid prepolymers had slightly larger discontinuous phases in the range from 25 nm to 50 nm: Effect of the Distribution of Molecular Weight in the Prepolymer. The gross morphologies (both at macro and micro levels) in the bimodal blend samples (Series F) appeared to be approximately similar to those of the counterparts prepared from the Series E commercial resins (see Figure 8 for a typical comparison). Effect of High-Molecular-Weight Curing Agent. Sample G-l with Versamid 140) also showed the shell-—and-core-type morphology (Figure 9). The sizes of the shell-and-core structure (range from 2.5 um to 25 um) and the discontinuous phase were larger than those cured with MDA, probably because this sample (cured

had

a higher

M,.

Figure

9 shows

a shell

(encircled

in black)

which was not etched open by the chromic acid at the time the sample was removed. A semi-open shell can also be seen which looks similar to a dimpled nodule (though of a much larger size)

12.

MISRA ET AL.

Network Morphology and Mechanical Behavior

Figure 5. Scanning electron micrographs of 7-hr-etched networks prepared from equivalent stoichiometric amounts of Epon-MDA resins: (A) Epon 825; (B) Epon 828; (C) Epon 834; (D) Epon 1001; (E) Epon 1002; (F) Epon 1004

167

168

EPOXY

RESIN CHEMISTRY

Figure 6. Transmission electron micrographs of replicas of networks prepared from equivalent stoichiometric ratios of Epon-MDA resins etched for 4 hr: (A) Epon 825; (B) Epon 828; (C) Epon 834; (D) Epon 1001; (E) Epon 1002; (F) Epon 1004

12.

misra ET AL.

Network Morphology and Mechanical Behavior

Figure 7. Scanning electron micrographs of Epon 1001-MDA sample (E-5) etched for 7 hr: (A) and (B) sample cross section; (C) sample surface; and (D) surface of solvent-cast sample

169

170

EPOXY

RESIN CHEMISTRY

Figure 8. Scanning electron micrographs of Series F networks along with their Series E counterparts etched for 7 hr: (A) Epon 828; (B) blend (Epons 825 and 1004) equivalent to Epon 828; (C) Epon 1001; (D) blend (Epons 828 and 1004) equivalent to Epon 1001

12.

MISRA ET AL.

Network Morphology and Mechanical Behavior

geal

Figure 9. Scanning electron micrographs of Epon 828 resin cured with Versamid 140 (network comprising 25% glass beads by volume; etched for 7 hr)

EPOXY

2 shown

by Racich

and

Koutsky

(31).

structure observed in to fast curing, Epon

shell-and-core-type prepolymers was due

To check if the semi-solid and solid

RESIN CHEMISTRY

828 (liquid prepolymer) and MDA (in stoichiometrically equivalent This sample, after being amounts) were cured at 150°C for 4 hr. etched as described earlier, did not show any difference in morphology from the sample cured through the regular curing cycle. It therefore appears that this peculiar morphology depends on the molecular weight of the prepolymer and not on the rate of curing or inhomogeneous mixing. Isolated Microgels. The electron micrographs (Figures 10, lla and 11b) clearly show the individual microgels (ranging in size from 20 nm to 200 nm) and clusters of microgels that were etched out from the network surface. The appearance of these microgels is in good agreement with the observations made on replicas and indicates that sample E-2 (prepared from liquid

prepolymers) as

had

compared

larger

and

sample

can

pack

microgels

E-5

(sizes

(prepared

from

from

50 nm

in a hexagonal

solid

to

150

nm)

sizes from 25 nm to 60 nm). These micrographs prove that the dimples, observed through replicas, on the surfaces of the epoxy networks were due to the etching of the weak connections of the microgels or clusters of microgels and not to the etching of material having a low average crosslink density. The platinum shadows in Figures lla and 11b indicate that the microgels are solid and nearly spherical in shape (a configuration having minimum surface

energy)

to

prepolymer,

array.

Similarly the electron micrographs (Figures lle and 11d) show the existence of microgels even before gelation (darker phase) in the size range from 30 nm to 100 nm. Thus, in both cases discrete microgel particles were observed, indicating that microgel formation takes place before gela-~ tion and that the microgels are dispersed and loosely connected to

each

to

a weaker

other

discontinuous

and

in

some

continuous phase

cases

(non-stoichiometric

phase.

(though

These etched

results

out

first

also as

compositions)

show

clumps)

that is

the com-

posed of microgels that are stronger than the continuous phase. It is the connections between the aggregates that are weak. The size and number of these gels probably continues to increase until the viscosity becomes high enough for physical gelation to take place. Discussion

General

earlier,

Proposed

electron

Model

microscopic

for

Network

evidence

Formation.

shows

that

As

mentioned

microgel

particles form prior to gelation (Figures 10 and 11), both with low and high molecular weight epoxies and with epoxy and amine-rich systems. This observation is in accord with the suggestions of other investigators (14,27,28). Some investigators have observed

12.

MISRA ET AL.

Network Morphology and Mechanical Behavior

173

Figure 10. Scanning electron micrographs of isolated secondary microgels from Epon 828-MDA sample (E-2)

174

EPOXY

RESIN CHEMISTRY

Figure 11. Transmission electron micrographs of isolated primary microgels: (A) and (B) gel particles etched out by Cr,O; solutions at 70°C after 4 hr; (C) and (D) microgels formed before gelation of the epoxy resins. [A and C—E-2; B and D—E-5]

12.

MISRA ET AL.

these microgels reported larger

Network Morphology and Mechanical Behavior

in sizes smaller than 0.5 um while others have than 5 ym. The present results indicate the pEee

sence of microgels and a higher level

that

the

systems

175

at both levels; a (2 wm to 50 um).

studied

by other

lower level (10 nm to 50 nm) It is, therefore, probable

investigators

also

had

micro-

gels at both levels. In order to differentiate between the two size levels, the term primary microgels is used to refer to microgels in the size range from 10 nm to 50 nm, whereas secondary microgels refers to sizes larger than 1 pm. In view of the morphological and property data at hand, a model for network formation is proposed. The principal points to consider include the following questions: the basic morphological units, the phase continuity, and the crosslink density and other properties of the various entities. The proposed mechanism which essentially comprises a synthesis of the mechanisms proposed by Bobalek et al. (27), Solomon (33) and Labana et al. (14), considers that the formation of the macrogel takes place in three sequential steps (Figure 12): formation of the primary microgels, formation of the secondary microgels, and formation of the macrogel. Following Labana et al. (14), it is also proposed that the secondary microgels and the macrogel are not completely coherent (that is, they have some unconnected dangling chains). Formation of the primary and secondary

Primary Microgels. The reactivities of the amine groups in MDA are approximately in the ratio of 1.4/1 (39). The primary amine groups should, therefore, react first to give a linear structure. Due to a combination of an exothermic heat of reaction with poor heat transfer, the temperature should rise in the vicinity of these molecules, resulting in the reaction of the secondary amine groups. Indeed, it has been shown that the primary and secondary amine groups react almost simultaneously even when their reactivities differ by a factor of two (43), The growth of the nuclei continues until the reactive polymer molecules can diffuse to the reactive sites of the nuclei. During this process several new nuclei are also developed. Therefore, a distribution of size of the primary microgels would be expected at all times during the curing procThis phenomenon has been predicted statistically by Labana cess. et al. (14), who termed it a nucleation process resembling the one that occurs in crystallization.

After a certain conFormation of the Secondary Microgels. centration has been reached, the primary microgels and the growing nuclei begin to interact with each other and give rise to new These nuclei grow in part nuclei for the secondary microgels. physical sintering encourage will which forces capillary to due and in part due to the reaction of unreacted functional groups in Thus, the secondary microgels are not as the primary microgels. The size of a particular coherent as the primary microgels. secondary microgel would depend on the concentration of the

EPOXY

176

PRIMARY NUCLEI

GROWING NUCLEI

PRIMARY MICROGEL

PRIMARY MICROGEL

SECONDARY

SECONDARY NUCLEI

SECONDARY MICROGEL

RESIN

MACROGEL

MICROGEL

MACROGEL

Figure 12.

Proposed model for network formation

CHEMISTRY

12.

MISRA ET AL.

Network Morphology and Mechanical Behavior

primary

microgels

tion

the

in

V77

near the secondary nucleus. A size distribusecondary microgels should also be expected,

Formation

of

the

Macrogel

(Final

Network).

At

a critical

solids concentration (~74% for monodispersed spheres arranged in a hexagonal packing), the secondary microgels pack together and the experimental gel point is observed. At this stage, phase inversion takes place during which the secondary microgels become the matrix and the unreacted or partially reacted prepolymers become the dispersed phase. By the time of complete reaction, the secondary microgels become loosely connected to each other with the help of the interstitial prepolymer or due to selfdiffusion of reactive dangling groups. Thus, the interconnections between the secondary microgels would be weaker than those between the primary microgels. The properties of the final netIn network should be affected by the coherence of the network. the secondary works prepared by non-stoichiometric compositions, microgels would be embedded in a matrix of lower crosslink density but still connected to each other through chemical bonds. Based on the above model, the morphological differences and mechanical behavior observed in the samples of different series are interpreted below. Morphological Differences. In epoxy prepolymers, the reactivity increases with the number of hydroxyl groups present in the prepolymer chain (44). Thus, an increase in the molecular weight of the epoxy prepolymer would not only reduce its diffusivity but would also increase its reactivity, resulting in a reduced time of microgel formation. This combined effect would result in a smaller size of the primary microgels, as observed experimentally. In the case of non-stoichiometric compositions, the size of the primary microgels should be expected to be governed both by the value of M, and by the structure, because the diffusivity and the reactivities of the prepolymers should be almost the same

at

all

stoichiometries.

In

the

Epon

828/MDA

system

(Series

A), an excess in amine concentration should initially yield many linear molecules whose length would depend on the amount of excess amine. The premature reaction of the secondary amines of these long molecules would result in larger microgels. Thus, in the case of amine excess, the size of primary microgels, as observed experimentally, should increase with increasing amounts In the case of excess epoxy, a branched struc-— of excess amine. ture having four epoxy molecules attached to one amine molecule Thus, the size of the microgels formed at 100% is formed (39).

epoxy-excess would be much smaller than those formed at 100% The experimental results showed a similar trend. amine excess. For systems having a slower rate of reaction (such as samples of Series A, sample E-1, sample E-2, and sample F-2, the higher diffusivity which were made from liquid prepolymers)

EPOXY

178

RESIN

CHEMISTRY

and lower reactivity result in slightly larger primary However, the loose connections of the adjacent microgels. secondary microgels can be readily attacked, resulting in the etching of the secondary microgels and thus formation of dimples

The semi-solid on the surface (as was observed experimentally). of high-M and low-M prepolymer and solid resins are composed the amount of low-M component decreases with the components; The more reactive highaverage M of the prepolymer (41, Ch. 3). M components form the microgels that get dispersed in the low-M component, which reacts later to give a high crosslink density A decrease in the proportion of the low-M component shell. should increase the size but decrease the number of the shelland-core-type of structure as was found experimentally. Dynamic Mechanical Spectroscopy. As pointed out above, the heterogeneous morphology, observed in the present study, implies network flaws (weak connections between the primary and secondary microgels). However, the low cyclic strains applied in dynamic mechanical tests detect only the effect of the basic network structure and not the network flaws. Therefore, as observed

experimentally not

indicate

(38,41),

dynamic

heterogeneity

Soluble Content. increase inthe dimple

in

mechanical the

spectroscopy

should

samples.

The morphological studies indicated an size of Mo. Ajidimple actually represents

a bit of material extracted from the network. This observation is in good agreement with the extraction results (41,42) that showed an increase in soluble content with Mee

Mechanical Behavior. At present, it is not possible to reconcile all the diverse dependences on stoichiometry. However, clearly there is no a priori reason for trends in all properties to

be

related.

By

a modification Q

i}

of

(S20ES

the

Griffith

ae)

iby/2

equation

(45)

Ci)

where o,, = ultimate tensile strength, E = Young's modulus, S = fracture energy, and a = characteristic flaw size. Thus og at constant E depends on how S varies witha; S may change while

0, does not, if changes in a compensate the Epon 828/MDA system, both Bell (46)

for changes in S. For and Kim et al. (45) have shown that 0, 1s almost independent of stoichiometry. Furthermore, it has also been shown that impact toughness is at a maximum not at 100% stoichiometry but at a 1/1.4 epoxy/amine ratio (37,45,46). The impact strength for the case of excess epoxy was also reported to be higher than that at equal stoichiometry. Under these circumstances, it has been shown that the apparent critical flaw size a is a linear function of the amine/epoxy ratio, changing from 32 um to 141 um (45).

12.

MISRA ET AL.

Network Morphology and Mechanical Behavior

179

Whether the critical flaw size corresponds to the aggregates first formed prior to phase inversion or whether it corresponds to the inclusions is, of course, not known. Indeed, the reality of a flaw corresponding to the calculated value of a is not proven. Nevertheless, the close agreement between microscopic evidence of the present study and the computed flaw size (45) is strongly suggestive of a correlation. Though the modulus, E, does not decrease (in fact increases slightly (45,46) with an increase in amine content, the overall plastic deformation should increase due to the plasticizing character of the amine. Hence, S increases (E itself may

increase time,

that

due

the

tensile

epoxy

to

flaw

excess,

enhanced size

strength the

continuity

increases

a

actually

decreased

of

the

network).

little

but

not

increases

flaw

size

as

At

fast

a little.

(corresponding

the

as

But to

same

S so

with

an

smaller

aggregates) more than balances out the lower value of S so that the strength again increases. However, the toughness otherwise seen at high amine contents is vastly reduced by the high loading rate in impact; while with epoxy excess, the smaller flaw size is able to delay fracture. In the case in which the M, is varied by varying the molecular weight of the epoxy resin at equal stoichiometry (Series E), the ultimate tensile strength is independent of M, but the impact strength decreases with M, (38,41). When the molecular weight of the epoxy is increased, a different behavior occurs. The inherent S of the epoxy components is low compared to that of amine

impact

and,

hence,

strength

the

tensile

strength

does

not

increase

and

decreases.

The Glass Transition Temperature. Samples having a bimodal distribution of molecular weight (Series F) showed the same Tg as their counterparts having a broad distribution in the high (> 700) and the low (< 400) ranges of M,.. At intermediate Manee the T, was found to be lower in samples having a bimodal distribution (42). Though the morphology of the bimodal and broad distribution was the same, the properties of the former will be determined by the dominant component in the final network. If this is a high-molecular-weight resin (which is more reactive), this network may be expected to dominate and T, will be low, perhaps because the initial fast reaction pasuits in greater At intermediate compositions, a judicious balancing incoherence. Thus properties May occur and result in a more coherent network. may be governed by the low-molecular-weight component.

The fact that small-angle X-ray scattering and stained or unstained microtomed thin sections fail to indicate a two-phase structure (37,41) also supports the present model, which explains that heterogeneity in networks is primarily due to incoherence on a local scale and not to major variations in average crosslink density.

EPOXY

180

CHEMISTRY

RESIN

it does capable

Although this discussion is clearly speculative, explain some of the behavior noted and suggests ideas testing in the laboratory.

of

Conclusions

From the present study and the studies of other investigators (discussed earlier), it is clear that crosslinked networks in many epoxy resins and in some alkyd resins are heterogeneous in nature. According to the proposed model, small primary microgels ranging in size from 10 nm to 100 nm are formed much before the inset of physical gelation. These primary microgels agglomerate together through weak connections to produce secondary microgels ranging from 0.5 wm to 50 um. The secondary microgels coalesce together and the experimental gel point is observed. The final crosslinked network is produced after complete reaction of the unreacted prepolymers left in the interstices of the coalesced secondary microgels. As compared to primary microgels, these secondary microgels are generally connected to each other through

"weaker"

or

less

coherent

links.

In networks

prepared

through

non-stoichiometric compositions, the secondary microgels are embedded in a matrix of lower crosslink density but still connected to each other through chemical bonds. The coherence of the final network (macrogel) depends on the density strength of the connections between the secondary microgels which in turn governs its tensile properties. In the case in which the prepolymers have a bimodal distribution of molecular weight, the properties of the network are governed by the component that is more dominant in the microgels. Though sufficient experimental evidence is not presently available, the optimization of curing conditions should also play

an

important

role

in achieving

maximum

coherence

in

the

network.

Acknowledgements

The

of

this

authors

work

wish

through

to

AFMC

acknowledge

Control

support

No.

al support from the Materials Research Fund is also much appreciated. Literature

1. 2.

3.

4. 9.

for

the

F33615-75C-5167. Center

and

the

first

part

Addition-— Ford

Motor

Cited

deBoer, J. Lohse, F.;

H. Trans. Faraday Soc., 1936, 3S Geel Oe Schmid, R.; Batzer, H.:; and Fisch,W.

Br.

Poly.

URE UIC Ee le(Oy Houwink, R. J. Soc. Chem. Ind., London 1936, 225, 24Jg Trans. Havaday SOGs 5693653291225 Cuthreli Sek oh. Je Apol. Polym. Set. 5 1967, 11,7949: Andrews, E. H. "Fracture in Polymers," American Elsevier, 1968.

12.

MISRA ET AL.

Network Morphology and Mechanical Behavior

Solomon, D.'H. and 1966521057 1898s Cuchi: clue

Rem

homed

Hopwood,

J.

Appl.

Polym.

J.

J.

Appl.

Scliy)

Polym.

1068.0

181 Sci.,

12,.01263.

Kaelble, D. H. In "Epoxy Resins, Chemistry and Technology," ed. by C. A. May and Y. Tanaka, Marcel Dekker, NY, 1973, Chapter 5. Blockland, R. and Prins, W. J. Polym. Sci., 1969, A-2(7), IESY)5) a Gallacher, L. and Bettelheim, F. A. J. Polym. Sci., 1962, HR AE ie Balyuzi, H. H. M. and Burge, R. E, Nature, 1970, 227, 489.

ee

14.

16.

ZO Dipl

Zits

Kenyon,

A.

S.

and

1969, A-3(2), 275.

Nielsen,

E.

J.

Macromol.

Sci.—Chem.,

ir eB ee one

IMeWeGIO}S\5 Wig Ig “Sheviis Nis eG MNeeil, esto GNO7/sh, ils SECA 5 Isle: Labana, S. S.; Newman, S.; and Chompff, A. J. In "Polymer Networks,"' ed. by A. J. Chompff and S. Newman, Plenum Press, MOThS. wa see Bratch, b- Be aud Spurr, Rh. AS J.) Polym.) Sei 5 195950 355 391 Erath, E. H. and Robinson, M. J. Polym. Sci., 1960, C-3, 65 BpUrT, Ke ALY Ind. Eng. Chem, 71957, 49° 1839. Miyamoto, T. and Shibayama, K. Kobunshi Kagaku (Eng. Ed.) ason 2 ees. Loskutov, A. I.; Zagrebennikova, M. P. and Arsen'eva, L. A. Vysokomol. Soedin, 1974, B-16, 334. Wohpetedler. Hy PP.) J. Polym. Seis, 1963, C=3,1 JJ Rochow, (4G. Anal .« Chem." 1961,. 33,5 1810. Rochow, T. G. and Rowe, F. G. Anal. Chem., 1949, 21, 461. Kessenikh, R. M.; Korshunova, L. A.3; and Petrove, A. V. Folyscscea. Usskh, 1972, 14,2339. Basins Vo. Y.: Korunskii, LL. M.; Shokal'skaya, QO. Y.; and mleksaudrov, N. V. Poly. Sci. USSR, 1972, 14, 2339. Carswell, TT. S. “Phenoplasts, Their Structure, Properties and Chemical Technology," Interscience: New York, 1947, Chapter 5. Neverov, A. N. Vysokomol. Soedin, 1968, A-10, 463. BobAllciom ion Gesm Moore mr Riss) eviy puSo oes) ands Lees CaG: Japp. seOlyile soci, 4196/7, 1151593. Solomon De Hes Lort..b. Ca; and Swift, J.D. J. Appl.

Polym.

Sci., 1967, 11, 1593.

Morgan, R. J. wa, 7/3 IPS

and

O'Neal,

Racich, J. Ll. and Koutsky,

Be

L.

J.

E.

J. A.

7H), Zillivibe Racich, J. L. and Koutsky, J. of Crosslinked Polymers," ed. Press, Inc.: New York, 1977, Morgan, R. J. and O'Neal, J. of Crosslinked Polymers," ed. Press, Inc.: New York, 1977,

Polym.

Plast.

J. Appl.

Tech.

Polym.

Eng.,

Sci.,

1976,

"Chemistry and Properties A. by S. S. Labana, Academic p. 303. "Chemistry and Properties E. by S. S. Labana, Academic p. 289.

182 33. 34. 35, 36, 3/7, 386

39a HO. 4) Aro LSS

44.

4D.

46.

EPOXY

RESIN

CHEMISTRY

Solomon, D< H. J. Macromol. Sciin (Reve), 07 meek) Lo. Batzer, H.; Lohse, F.; and Schmid, R. Agnew Makromol. Chem., L973, 29/30, 369. Dumke, Wo Je Wollym, Site, I9G/, CHilo, 1497 Kreibick, Ul) T. andy SehmidjeR: (J sPolymey Sell, 97 o,goymp. NOs BS ILI s “Selby, K. and Miller, i. 2. J. .Material (Seijal o75-e2 0, 2. Weisel, do Nos Sereilsine., Iho iio§ ame Kin, §. Ibo Watmall Repoee AFMC Contract No. F33615-75C-5167, April 1977, AFML-TR-//HOO Bell Joe P. ee rolym. sein mo 10g S) nm cis Waiisomn Wo No eine! Clu, > Ws Do Bolym. Set,, L979, Symp NOs Gil, Q5¢ Misra, S:-€.) Ph. b. Dissertation. Lehteh) University 9 io. weer GS, Cos Werisem, do No8 Bune Spercilime., 1, Bi, Wikis volume, 1978. eACLteIiam Mele

blnchm

Reb

as

andm

Sachem

ar moO laymen

LOW, M25 33:5 Bowen, D. O. and Whiteside, R. €. Jn. “Epoxy Resins! in Advances in Chemistry Series, American Chemical Society: Washington, D.C. 1970, 92,48. Kim, 9S. 1.3; Skibo; Mi; Manson, J. A.; Hertzbere, RieWo: and Janiszewski, J. Presented at American Chemical Society meeting, Chicago, August 1977; Poly. Eng. and Sci, in press, 1978. Bell, J.P. Jv Appl. seolym. Sins h970, lace SOL.

RECEIVED May 21, 1979.

13 Creep Behavior of Amine-Cured

Epoxy Networks:

Effect of Stoichiometry S. L. KIM’, J. A. MANSON,

and S. C. MISRA

Materials Research Center, Coxe Laboratory

#32, Bethlehem, PA 18015

Creep studies are not only important in the theory of viscoelasticity but also of great practical importance in any engineering application in which the polymer must sustain loads for long times. With a thermoset resin, creep can vary a great

deal, depending on the degree of crosslinking, the perfection, and the morphology of the network structure. While many studies of the effects of crosslinking on behavior have been made, most have been restricted in scope and many have used relatively uncharacterized materials. For this reason, a systematic examination was conducted to elucidate the

effects of crosslink density and its distribution on a wide range of properties including basic viscoelastic response, stress-— strain and impact behavior, and fatigue crack propagation. The same base prepolymer type and curing agent were used throughout, and a given crosslink density was obtained by changing the stoichiometry and by changing the molecular weight of the prepolymer at constant stoichiometry. Related

Studies

In some epoxy systems (1,2), it has been shown that, as expected, creep and stress relaxation depend on the stoichiometry and degree of cure. The time-temperature superposition principle (3) has been applied successfully to creep and relaxation behavior in some epoxies (4-6)as well as to other mechanical properties

(5-7). More recently, Kitoh and Suzuki (8) showed that the Williams-Landel-Ferry (WLF) equation (3) was applicable to networks (with equivalence of functional groups) based on nineteencarbon aliphatic segments between crosslinks but not to tighter networks such as those based on bisphenol-—A-type prepolymers Relaxation in the latter resin cured with m-phenylene diamine. followed an Arrhenius-type equation. t1Current

address:

Goodyear

Tire

Fiber

& Rubber

and

Co.,

Polymer

Akron,

Product

Ohio

Research

44316.

0-8412-0525-6/79/47-114-183$05.00/0

© 1979 American Chemical Society

Division,

EPOXY

184

CHEMISTRY

RESIN

In this study, the imperfection of the networks was varied by varying the stoichiometry of an Epon 828-methylene dianiline Related studies of morphology, other properties, and system. creep as a function of molecular weight and distribution of the prepolymer are described elsewhere (9-13). Experimental

Epon 828 epoxy resin and methylene the Shell Chemical Co. The resin together, mixed, degassed, and cast

The materials used were dianiline (Tonox), both from and curing agent were melted

between glass plates. The cure cycle was as follows: 45 min at 60°C, 30 min at 80°C, and ®2,5)hreat 150°C) then slow coolingsts room temperature. The amine to epoxy ratio (A/E) was varied from Compositions are given in 0.7 to 2.0 (in terms of equivalents). Table 2. Table

“mM

I.

is

mechanical

tive

Composition

calculated

based

measurements

M. throughout

of

Epoxy

on

(10),

this

the

and

Specimens

for

rubbery

moduli

used

as

a

Creep

relative

Study

from and

dynamic effec-

study.

Creep was measured using a Gehman torsional tester (14). [In fact, the shear compliance J(t) was measured explicitly; the data are presented as moduli by taking E(t) = 3/J(t).] Master curves were obtained in the usual manner (3). Results

and

Discussion

Typical creep-modulus behavior is shown in Figure 1; viscoelastic parameters deduced from the master curve are given in Table II, and discussed below.

Glass-transition

temperature

(T,).

The modulus-temperature

curves: for. al lathes epeciicnsanere: composed by plotting the 10-sec modulus at each temperature, as shown in Figure 2. The glasstransition temperatures (corresponding to a modulus of 0.2 GPa) are given in Table I. Calculations show that T, is inversely proportional to the average molecular weight between crosslinks

13.

KIM ET AL.

1

Amine-Cured

Epoxy Networks

9

a

= ~x
two straight lines for all the specimens were Bpeeiocde as shown Figure 5, consistent with an Arrhenius-type relationship:

1a lane where E, is

R is the gas the apparent

straight mole

(227

lines,

-E ee a

=

the activation

eale/mole)

above

Z

2.303 RT

T is the constant, activation energy.

356

2)

absolute temperature, and From the slopes of the

were

found

kJ/mole

(85

energies

Ty and

in

to be 950 kJ/ kcal/mole)

EPOXY

188

Xi

ne Alle

Saal)

20

40

RESIN

CHEMISTRY

60

(T-Tg), °K Figure 3. Composite curve of experimental shift factors as a function of (T — T,) (symbols are as in Figure 2). The curve predicted by the WLF equation (3) is indicated by the solid line.

13.

KIM ET AL.

Amine-Cured Epoxy Networks

Le,

Pers

oie 7,

Toe

189

Poel

So

iT x 102, (ek)! Figure 4.

Plots of experimental shift factors for epoxy specimens vs. inverse temperature (symbols are as in Figure 2)

-30

S200

10

0

10

20

30

(/T-1/Tq) x102, (°K)! Figure 5.

Composite plot of experimental shift factors for all epoxy specimens as a function of (1/T — 1/T,) (symbols are as in Figure 2)

CHEMISTRY

RESIN

EPOXY

190

As shown in Table III, these values are in excellent below T.. agreement with the values observed by Kitoh and Suzuki (8) in a bisphenol-A-type epoxy cured with phenylene diamine: 962 kJ/mole (230 kcal/mole) above T, and 381 kJ/mole (91 kcal/mole) below Ty. Values also agree reasonably well with the apparent activation energies that were determined in our parallel study in which M, In comwas varied by changing prepolymer molecular weight (13). a assuming Ty, [at activation of energies parison, apparent viscosity of 1012 n.s.m72 CI iy ieee er eniaped to range between 900 kJ/mole (215 kcal/mole) for ve LOOP Grand 51259) kai/ mole (301 kcal/mole) for aa 168°C. Further, the WLF equation predicts a continuous decrease in Eg as T increases above Ty.

Table

III.

Comparison of Apparent for Relaxation in

Energies

Activation

Range

of

T

E>

kJ/mole

of

Epoxies

(kcal/mole)

950° (227) | 960° (230) | 962° (230) 3607 (86)

356 (75)

“This study. Parallel study with M. changed molecular weight (13).

Kitoh with

and

Suzuki

m-phenylene

Thus,

while

our

(8),

381° (91)

by changing

for bisphenol-A-type

prepolymer

resin

cured

those

in

diamine. values

of

Ea

(as

well

as

references

8and 13) agree well with WLF predictions at T, when ee ae LO00SC; they do not agree quantitatively or qualitatively for higher-Ty specimens at T, or for any specimens at T >» Ty. Interestingly, other authors (5,14) have reported validity of the WLF treatment for less-densely-crosslinked epoxy systems. Also, we find that a plot of data by Murayama et.al. (]7) shows that an Arrhenius-type treatment of dynamic mechanical shift factors is preferred for oriented poly(ethylene terephthalate). Characteristic

creep

time.

All

master

curves

were

empiri-

cally shifted to the most convenient common temperature, 150°C, as shown in Figure 6. For convenience, only transition regions are shown. An increase in crosslink density (nearing to equivalent stoichiometry) shifts the curves to longer times as expected. The chracteristic creep time, t., was to a modulus value of log E(t) = (log

taken as the time to creep Eg + log Er)/2, where E and E, are the glassy and rubbery modulus, respectively. [While Tc is analogous to a retardation time, the latter designation is not used, because our te is determined from plots of 3/J(t), not

J/(t),

and

is

hence

not

numerically

equal

to

the

corresponding

13.

KIM ET AL.

Amine-Cured

Epoxy Networks

191

Pa

3/J(t) f(T), x

log t, second

Figure 6.

Creep-modulus master curves for all epoxy specimens as a function of log t at a reference temperature of 150°C

EPOXY

192

retardation

Characteristic

and

also

as

wt, is,

time.

creep

time

a function

strictly shown

is

of 1/M,

speaking,

Voie et

eet .

a relaxation The

again

T/M

the

of

function

7).

(Figure

line relationship between log T,~ and in the behavior of these specimens:

a

a

as

CHEMISTRY

RESIN

good shows

time. ]

A/E

ratio

straight-— consistency

Usth SS 10°

(3)

M

Ges c where log t,,, is found by extrapolation to equal -18, which is the characteristic creep time at about 150°C for a network having an infinite value of M.. Since both te and Ty are inversely related to M, and the activation energies of the shift factors are independent of M., a common segmental motion must be involved in the creep behavior of all these specimens varying the crosslink density merely shifts the curves along the time axis. Distribution of relaxation times, H(t). iuethis discussion, using plots of 3/J(t), we shall discuss the distribution of characteristic response times in terms of relaxation times. The relaxation-time spectrum, H(t), can be determined as

a first

approximation H(t)

= =

(18) Sed

by

the

|CCl

-E(t)

following sda CLogee,)

relationship: tr

(4)

d [log E(t) ] ilo ewe) —

Plots of H(t) vs log t for When data are replotted at bution of relaxation times

all specimens are given in Figure 8. the T, of each specimen, the distrifor tee 5 different specimens nearly coincide with each other until the rubbery region, where some scatter is seen (Figure 9). This confirms that the same mechanism for relaxation is involved for all 5 specimens. The slope of H(t) (i.e., dilog H(7)]/d. (loevejeat the glass transition re¢cion (the right-hand part of Figure 9) is about -0.42, which compares fairly well with the value of -0.33 obtained by Kitoh and Suzuki (8) for a bisphenol-A epoxy resin cured with m-phenylene diamine, considering the differences in the epoxy system and in the testing modes. Clearly the data for our parallel study (13) in which Me is varied by changing prepolymer molecular weight are indis-— tinguishable from those of this study (note triangles in Figure 9). Conclusions

In conclusion, the temperature dependence of shift factors for the networks studied here do not follow the WLF equation, but rather an Arrhenius-type relationship. The apparent activation energies are independent of stoichiometric variation [as they are

when

M, is varied

by changing

prepolymer

molecular

weight

(SD als

13.

KIM ET AL.

Amine-Cured Epoxy Networks

second c’

log 7%

A/E

imc x04

Figure 7. Characteristic creep time for all epoxy specimens as a function of stoichiometry and M,: (@) epoxy-rich; (QO) amine-rich; (A) different series with M, varied by changing prepolymer molecular weight (13)

Pa H(t)

log t, second

Figure 8.

Distribution of relaxation times for all epoxy specimens as a function of log t (symbols as in Figure 7)

193

194

EPOXY

Pa

H(t),

=

0 log t, second

Figure 9.

5

10

RESIN

CHEMISTRY

13.

KIM ET AL.

Amine-Cured

Epoxy Networks

195

Nearing to stoichiometry (decrease in M.) merely shifts the creep curves to a longer time scale. Although the crosslink density was lowered (increase in M.) more than four-fold by varying the A/E ratio, apparently the aromatic rings in the bisphenol-A and diamine group must be controlling (stiffening) the segmental motions in the networks. Thus, considering the results of this work and those of Kitoh and Suzuki (8) and Murayama et. al. (17), we might expect an Arrhenius-type relationship in the temperature dependence of shift factors for relatively rigid networks. Acknowledgement

The authors gratefully acknowledge financial support from the Air Force Materials Laboratory through Contract No. F33615fo-G—5 167 « Literature

1.

Kosuga,

Cited

N.

and

Tsugawa,

S.

Kobunshi

Ronbunshu,

1975,

B25

BVae Z.)sDelmonte, J. Plastics Technol.., 1958, 4, 913. 3. Ferry, J. D. “Viscoelastic Properties of Polymers," 2nd Ed., John Wiley & Sons, Inc., New York, NY, 1970, Chap. 11. Ae negcaris. eb. s RheolsActa.,. 9625.2, 92. Teac ble Dan eI. Appl POlym.so0d.. 1905509, 012137 Dee itO, Ne rand oato, Mi gd. Polym. Sci.-C., 5196/7, 165471069. 7

MeCrum,

1970,

N.

B4(1),

G.

and

Pogany,

G.

A.

J.

Macromol.

Sci.—Phys.,

109.

8.

Kitoh, M. and Suzuki, K. Kobunshi Ronbunshu, 1976, 33, 19. Manson. J. A.> Sperling, I. H. and Kim)-S, L... “Influence of Crosslinking on the Mechanical Properties of High-T Polymers,'' Technical Report AFML-TR-/77-109, July 1977. 10. Kim, S. L. and Manson, J. A. ''Dynamic Mechanical Behavior of Amine-Cured Epoxy," 19th Canadian High Polymer Forum, Ottawa, Canada, August 1977, to be published. Lie kins lat Skibo, Mss Manson, J.0A.; Hertzberg, R. Wo and Janiszewski, J. Polym. Eng.-Sci., 1977. 12. Misra, S. C. Ph.D. Thesis, Lehigh University, 1978. taweiisracesec ©. Manson .aleeA. and Sperling, us. H.,) This Publica tons L979) ASTM D-1043, American Society of Testing Materials, 14. Philadelphia, PA. Sci.-Rev. Macromol. Chem., LOGO J. Macromol. Nielson, L. BE. 15. CSOs WMHs 'Mechanical Properties of Polymers," Van Nielsen, L. E. 16. Nostrand Reinhold, New York, 1962. J. Polym, Murayama, T.; Dumbleton, J. H. and Williams, M. L. 17.

Que

18.

Sci., 1968, part A=-2, 6, 7/87. "Properties and Structure of Polymers," Tobolsky, A. V. John Wiley & Sons, Inc., New York, NY, 1960, Chaps

RECEIVED

May 21, 1979.

Pare

elles

me. |) =

sori ~~ - i. An®

se =

1 ai

)

pore

a

ee a! Le ae: YT| 2?

>

@

:

2)

eer

ealZe

te

afd.

L@

.¢

ae

& Aine

ALG

‘cg

oes

> ene

bet

, vee

tc ISMAD

»* “

ee ag

ao

~

=

Te

raeeeiqe

if

4

|

.oteioue

ae eFtb

a

\mevactsy

e

=

0.8

=

SS

eB

A

=zb5

os i

a eeO Ne aX a!

2S oe = = and

0.2

0.0 0.0

0.2

0.4

VINYLPYRIDINE ( MOLE

0.6 IN

0.8

1.0

MONOMER

FRACTION )

Journal of Polymer Science, Polymer Chemistry Edition

Figure 1. Monomer—copolymer composition curve for 1-ethenyl-4-(2,3-epoxy-1propoxy) benzene (M;) with 4-vinylpyridine (Mz): (——)r1 = 0.467 and r, = 0.638

=

«=

lu

=

= _ on Qo oz

ZN OS) = © = oe ms Ww == = WwW = —

Aa

a

-

©

= as

0.0

0.2

0.4

VINYLPYRIDINE ( MOLE

0.6 IN

0.8

1.0

MONOMER

FRACTION )

Journal of Polymer Science, Polymer Chemistry Edition

Figure 2. Monomer-—copolymer composition curves for 1-ethenyl-4-( 2,3-epoxy-1propoxy) benzene (M,) or 2,3-epoxy-I-propyl methacrylate (M i) with 2-vinylpyridine (M,): (——) 1-ethenyl-4-(2,3-epoxy-I-prpoxy) benzene with 1, = 0.556 and r, = 1.25; (-—-) ga cron! menyl meihacryiaic with r; = 0.51 and r, = 0.

EPOXY

200

RESIN

CHEMISTRY

1,0

td =

0.8

=

= (=>)

—

foe

Ss == =

=

= =

0.6

oO

= a

Le

uw

0,4

=

= ce ata = =S (se

0.2

—

0,0 0.0

0.2

0.4

VINYLPYRIDINE ( MOLE

0.6 IN

0.8

1,0

MONOMER

FRACTION )

Journal of Polymer Science, Polymer Chemistry Edition

Figure 3. Monomer—copolymer composition curves for 1-ethenyl-4-(2,3-epoxy-Ipropoxy) benzene or 2,3-epoxy-l-propyl methacrylate (M,) with 5-ethyl-2-vinylpyridine (M,): (. ) 1-ethenyl-4-(2,3-epoxy-1-propoxy) benzene with r,—= 0.639 and r, = 1,38; (~——) 2,3-epoxy-l-propyl methacrylate with r, = 0.57 and r, = 0.62

‘{‘°py

COLES ‘weH

°*5

‘9D/0V “gg

CCA “0961

eePOGL

I°TOS

“(sq ut)

tose

‘yYATOK

Jo BO

(eqeqe

ejzeq

MON

eqzeq Cava

O

Pp

5 Fz We oeaeq “Te FO eANYeMI

opeytwe, °C) SOLER

jo

TO

a7 961

JO

TW

bunGOzN

Hunox

sobpUeday.

We

pue

39-0 poesn AOF q souSeTejoy ejep FO *SONTPA UOTRETNOTeSO 3e G°*OQ Z-STOW FO ueany YAM SeTTazZTUOAAANQOS "D009 e

WATOd™

UOTReZTASBWATOgG

Apsuuey

* ToworyeNH)

gee © me BY OL ‘°weUud

MiG

:SUOTRTpPUOD

oeutTptTzAdqTAuta-7-TAyAg-G

7Z0°O

LS°Q

sTOW

79°0

JO

‘uOTRZezTAsuATOdOD

‘szewouow

VO ule = so%qc0,0—

tS

£9. k= Lie rs 0 a‘q0?r0 page LO

out ptazAdqAuta-Z

Bayo 0 = Be Neil pa0a 0= ei Oe av Ge (2

ter

1S °0 zo

TAdoad-{[-Axody-¢'Z ASWOUOW ARTATIOLSY

tw om

CoO

8c.0 a‘q0?r0 Oh joene

te mw Od. 60 LOS 2) CV. O9GIe eUAT

On05 CORO Se Bic

olqiV Lis GOO = Com

uotzeztzAswATod

PO°T aque 70 p#qa8 0

UT

490UeTOSIeAUT

-oirpAyeir}e4 DdL ow. CO ob Se ee

OV -1™ mwofgsl70

fe)

oe dol

€6 0 Tal Sa a cle 0 p’q3s8 "0

z0

Ciel

bAG60 wL secre (abet p/qot aL T6°0 al eid

JO JOWA|Og ‘adUaI9g JewAjog Ajsiwayy UOI}IPY jeuINosS

(tw) eqeTAroeyqew seutptazAdqTAuta yAtm p(*W) ut sentTeA 29-6 pue sotzey JO uotzeztzswATodoD Eas OL dpe

201 Self-Cross-Linkable Polyepoxides TANAKA

14,

EPOXY

202

Q=1.30,

e=-0.50

for

2-vinylpyridine,

and

RESIN

Q=1.37,

CHEMISTRY

e=-0.74

for The

Il. in Table given are these pyridine; other the from calculated curve compostion monomer-copolymer

the

monomer

reactivity

r ,=0.514

ratios,

and

r3=0.715,

with points experimental the to fit a poorer gave with obtained data the describe To 4-vinylpyr idine. and Q=1.5 with curve calculated the idine, 2-vinylpyr that with Q=1.4 and e=-l.1l, fit than e=-1.2 was poorer a] form @e=-150) “and @o=1.2 with that than better but calcuThese ethenyl-4-(2,3-epoxy-l1-propoxy )benzene. O=1.3, Q=1.2, with e=—-l.1; lated e==1/2: “and Q=1.5, e=-1.0 for l-ethenyl-4-(2,3-epoxy-l-propoxy)benzene gave the to points corresponding an excellent fit closely with 5-ethyl-2-vinylpyridine. The copolymers with these pyridines have a charac-— teristic band due to an intermolecular hydrogen bond v(OH) of OH group of polymers about at 3500/cm as shown in Figure 4. The bands at 1660 to 1640/cm assigned to the stretching modes of the C=C bond of olefinic hydrocarbons are not observed. The bands assigned to the in-plane and the out-of-plane bending modes for olefinic CH groups are also not observed distinctly in the 14501300/cm and 1000-800/cm regions of the infrared spectra of the copolymers. The symmetrical and the asymmetrical stretching bands of the epoxy ring seem to occur near 1250/cm and 910/cm, respectively. The intensity of the band at 910/cm decreased and disappeared in the spectra of the heated samples which became insoluble and infusible. The homo- and copolymers of these epoxides with the vinylpyridines were highly amorphous as judged by the X-ray diffractograms Under the experimental conditions, the resulting copolymers were soluble in organic solvents

such

as

tetrahydrofuran,

chloroform,

and

N,N-

dimethylformamide, while in a humid atmosphere the copolymers, especially those with 4-vinylpyridine, were apt to be converted to the insoluble materials during reprecipitation or drying at a room temperature. Differential scanning calorimetry and thermogravimetry studies show these copolymers do not melt but react autocatalytically at 90-200°C and degrade extensively above 300°C. The polymeric materials were crushed, mixed with KBr powder and pressed into the form of a KBr disk. The samples were heated in a forced draft oven, taken out at convenient

intervals,

cooled

in

a

desiccator,

and

the

intensity of the band at 910/cm was measured. The intensity of IR absorption spectrum of the CpOXyYeqnoupmat 910/cm was found to decrease in proportion to the decrease of the epoxy compound in the reaction SYS tem (@2))re

ee

UT

AOF posn SenTeRA peumssy ta FO UOTALCTNOTeS “SonieA “2 pue SOnNTeA peRzeTNoTeDd 9yq WOAF *“SONTeA 3-0 pounsse AOF pesn ejep soueAsToyY *SsenTeA 3-0 FO UOTReTNOTeO JO y-STOW G°O YIM UeAN;F "9009 3e& eTtTazazTUOATAANGOSTOze-,7‘z UOT eZTASWATOG :SUOTFTPUOD ‘szewouOU FO STOW 7ZO°Q uoTRZezTAewWATOd

UOIIPy Ajsiwayy JewAjog ‘aua!as JaWAJOg JO |eUINOS

Stent

-orpAyeizeq

p od q

e

SsuTptaAdTAuta-Z

eutptzAdyTAuta-z-TAyWa-S

oTES 0

PEST

oo lene

el a 2SL9°0 6£9°0 290€9°0 909S°0 9SS°0

Boel

a li ll o8e°T 8E°T 82 °T D6E aL Cou

0S*0-

ee aie ales Ai po TT°1= Docit= TL 0= eS Tes

pect

lie es Oe aL orl ps*T pork Tmt

Of; ft

cw

ouTpTrAdTAutA-F

pue

sotzey

ARTATZOVSY

TewoUuoW

-7% suezueq (Axodoid-q{-Axods-¢‘Z)

Les

O7TS"0 L9v°0

8-H

Coon

oSTL*O 8€9°0

sentTeA

yAt™M

Ls

pe La7 es q’?

O25n07

TO

ut

seutptTzAdTAuTA uotTZeztzrewATodoD

o(?W)

('W)

prt 977

FO

Z8°0 q?0

-TAueuzg-[T

9eTqdeL IIT

203 Self-Cross-Linkable Polyepoxides TANAKA

14,

ODS

O00E

00Sz

000Z

ur ay}

ee

AWM =

OOLT

:tauhjod ([) oo

OST

‘\

Ger

OO9T OOST USaWAN)

ODST

OOTT OODT 006

008

‘sawhjod (TJ)owoy

002 059

siawhjo pupdowoy auripithd

Aujs|way) UO!}IPZ

uoyovif fo tauhjodoo

JawAjOg ‘eouals JawAjog

OOZT

fixodosd ( -{-hxoda auazuaq pup -¢ sp sasayjuaSnoYs ind ay} ajou

jeuinor Jo

7-WO (

OOHT

RESIN

‘sau0Uuc (J) u

)-p-)huayja-T ‘aurpruhdjhua-g saqunn uy

OST

GMO,

ieee

EPOXY

aundig YF vuyoads fo ssauhjod Yynnoo

OOOH

wy

204 CHEMISTRY

14. The

TANAKA

Self-Cross-Linkable Polyepoxides

characteristic

the phenyl dards for

bands

of

groups at 1440 copolymers of

the

205

methyl

or

and 1505/cm were these epoxides,

because their intensities reactions. The results of

methylene,

and

used as stanrespectively,

did not change during the typical experiments with the copolymers of 1-ethenyl-4-(2,3-epoxy-l-pr)benzene opoxy are plotted in Figure 5, in which the relative intensity of 910 to 1505/cm are plotted changes against the reactime and follow, tion cases, in some sigmoidal curves MThe induction period. period was an induction showing to be dependent on the reaction found and on temperature unit vinylpyridine -in copolymers. the content the of The relationship of the rate of consumption of the

epoxy

group,

-d(Mj,¢/Mj,0)/dt,

with

the

epoxy

groups

remaining Mi t7 io tn the =copolymers = of l-etheny!— 4-(2,3-epoxy-l-propoxy)benzene is shown in Figure 6. Pie etner content’ of the vepoxy, group) at reaction time i. The maximum rate of consumption of the epoxy group increased as the mole fraction of the pyridine to the epoxy group increased. At the maximum rate of reaction the rate of consumption of the epoxy group reached the stationary stage, independent of the content of the epoxy group, the region of which became longer as the fraction of the pyridine increased. The Slope of the linear part of the curve obtained for the relationship decreased with increasing the content of the pyridine unit except in the initial stage of the reaction. Then, the rate of disappearance of the epoxy group varies with the mole fraction of the epoxy group and the order of reaction with respect to the concentration of the epoxy group cannot be determined by this experiment. The time-conversion curves, however, are Similar to those observed in the tertiary amine-catalyzed oligomerization of epoxy compounds (2), in which the rate is first-order with respect to the epoxy com-

pound the the

and

except in Jf'the rate concentrations rate equation = OM)

may

be

the initial stage of the reaction. Of reaction iS given aS a function of of the epoxy and the pyridyl groups, is shown by: /dt = kf (Mz) g(M>)

rewritten

eee

oN

as:

72M) ) = keg (Mo)

cy

the concentraOt EUnCELONS are g(Mz) £(M]) and where rate k is the pyridine, the and epoxide the of tions the converwhen time reaction t; is the and Constant, that g(M2) Assuming i. sion of the epoxy group reaches we can obtain the condition, under is constant

GUM2) ta SKY where

Kj

is

a constant

shown

by

-[JfdM]/f£(M))]/k.

EPOXY

206

RESIN

CHEMISTRY

0.0

0.2

oO e=)

Doig /Dy505

0,8

REACTION

60

80

TIME

( MIN)

100

120

Journal of Polymer Science, Polymer Chemistry Edition

Figure 5, Time-conversion curves for the epoxy group of poly[1-ethenyl-4-(2,3epoxy-I-propoxy)benzene] in KBr pellet at 105°C for copolymers with various mole fractions of pyridine: (I) 0.274; (II) 0.399; (III) 0.584

14.

TANAKA

Self-Cross-Linkable Polyepoxides

207

400

NOOooOo

100 co(=,

qo-* 2) isa

—=DD =) (=

ine)Oo

,

9) 4(My //My at-¢

4

6 810

20

40

EPOXY GROUPS REMAINING

60 80 100 ( % )

Journal of Polymer Science, Polymer Letters Edition

Figure 6. Dependence of the rate of consumption of the epoxy group at 130°C on the remaining epoxy group in copolymers obtained at reduced pressure with various mole fractions of pyridine: (I) 0.258; (IL) 0.502; (IIT) 0.597

EPOXY

208

Thus,

various

reciprocal

of

values

RESIN

the

CHEMISTRY

reaction

of the epoxy group the conversion ty, at which time i, are plotted against the mole fraction of the reaches rate The reaction in Figure 7. Mj, as shown pyridine, pyridyl the of content the increasing with increases a maximum value about at 0.5 or 0.6, and reaches group, The of the pyridine. the fraction with then decreases cothese in group epoxy the of reaction ring-opening the of presence in the out carried be may polymers action of the hydroxyl pyridyl group by the cocatalytic The mechanism of this selfgroup in the polymer chain. crosslinking reaction of the copolymers seems to be similar to those (2) proposed for the oligomerization of the epoxy compound and the curing of the epoxy resin by a tertiary amine. The reactivity difference in these self-crosslinking reactions between these copolymers seems to be described mainly by the polar effect of any of the possible organic moieties of the copolymer except the epoxy group. The ring-opening reaction of the epoxy compound can be generally explained by the modified Taft equation (6) As a linear relationship has been observed between the polar substituent constant 9* and the basicity of ethers (2), the reactivity towards nucleophilic) species will decrease and that towards electrophilic reagents may increase aS the basicity of the epoxy compound increases, i.e.,

as

the

substituents

o*

value

decreases.

R[R=CH2CHCg6H4OCH2

The

and

for

the

CH 2C(CH3)CO2CH2]

o*

values

of

the epoxy compounds, OCH2CHR, can not be found in the literature, but their relative order may be estimated from those (o*=0.600 and 1.9) (6) for CgEéHs and CH3C0, o* for 1,2-epoxypropane |(R-CH3) ebeing 0. Therefore, the reactivity of the epoxy group in the copolymers of letheny1-4-(2,3-epoxy-l-propoxy)benzene can be considered to be higher towards electrophilic species and to be lower towards nucleophilic compounds than that in the copolymers of 2,3-epoxy-l-propyl methacrylate. In the self-crosslinking reaction, actually, the former was converted) \to .the) insoluble ‘materials» fastergsand: at ea lower temperature than the latter. The pKa values of unsubstituted, 2-vinyl and 4vinylpridines “are 5-17, .4:.92—4.98andil5 62 mnie A methyl or an ethyl group in 2— or 6-position of the pyridine nucleus causes an increase in basicity as well as that in) 93—" or@4—-postion. Consequently, a similar increase in rate would be anticipated if the rate was proportional to the basicity of the pyridyl group in copolymers. Nevertheless, the rate of reaction catalyzed by 2,5-disubstituted pyridyl group are not larger but

smaller

than

2-

or

4-substituted

pyridyl

group

as

shown

14.

TANAKA

Self-Cross-Linkable Polyepoxides

209

12

10

(MIN )

ats 40

VINYLPYRIDINE ( MOLE

IN

COPOLYMER

FRACTION )

Journal of Polymer Science, Polymer Chemistry Edition

Figure 7. Effect of the mole fraction of pyridine on the reaction at 130°C of the epoxy group in copolymers of 1-ethenyl-4-(2,3-epoxy-1-propoxy) benzene (O, ©, @) or 2,3-epoxy-1-propoxy methacrylate (O) with various vinylpyridines. Where 10/t, at 30% conversion of the epoxy group is plotted: (©) copolymers with 2vinylpyridine; (@) copolymers with 4-vinylpyridine; (O) and O copolymers with 5-ethyl-2-vinylpyridine.

EPOXY

210

RESIN

CHEMISTRY

to the steric be attributed This might ceo. in hLOUL of the moieties organic of any of the possible effect idenThe similar group. the pyridyl except copolymer in the reaction results has been retained tity of these acid benzoic with either phenyl 2,3-epoxy-l-propyl of other many in and pyridines substituted by catalyzed reactions®

(2)%

Synopsis

Soluble and self-crosslinkable linear copolymers with pendant epoxy and pyridyl groups were obtained from l-ethenyl-4-(2,3 epoxy-l-propoxy) benzene or 2,3-epoxy-1propyl methacrylate and vinylpyridines by the action of 2,2'-azoisobutyronitrile. The monomer reactivity ratios were

determined

in

tetrahydrofuran

at

60°C,

and

the

Q

and e values for these epoxides were calculated with the reported Q-e values for these pyridines. The intrinsic viscosities of the copolymers were found to be 0.15-0.38 in tetrahydrofuran at 30°C and to be dependent on the copolymer composition. The copolymers were amorphous, had no clear melting points, and became insoluble crosslinked polymers under heating without further addition of any curing agents. The reactivity difference in these self-crosslinking reactions among these copolymers was described by the steric as well as the polar effect of any of the possible organic moieties of the copolymers. Literature

1.

2.

Sem Aan ». 6-9

vs

Cited

Tanka, Y; Okada, A; Tomizuka, I.,in May, C. A., Tanaka, Y., Ed.; Dekker:

"Epoxy Resins"; New York, 1973;

Chaprmr eTanaka, -Y.;-Mika, T. fF. inv sEpOxy Resins.) seeMaye C. A., Tanaka, Y.;1 Ed.; )Dekker> New York 7o1973; Ciiapraese EMayO gal aa Rc ©LEWLS mee. Mc, Js AMEGL. Chem OCey UO OG, ISOs, PY) Ceitm Cam Cre adic Polym. Sci.s,/1948, 3, 772. slamikado; Tf. #J s:Polym-e SCi.,01960,8 4576460" Taft, R-9 W.i ree, sine Stericebf fects sin Organic Chemistry"; Newman, M. S., Ed.; Wiley: New York, LI56:eChapel3. Perrin, °D.. 9D.” Eds, "Dissociation Constanta or Organic Bases in Aqueous Solution"; Butterworths: London,

1965.

RECEIVED May 21, 1979.

i) Some Studies on the Preparation of Glycidyl 2-Ethylhexanoate D. A. CORNFORTH, B. G. COWBURN, C. W. STEPHENS, and C. G. TILLEY

K. M. SMITH,

Anchor Chemical Company Ltd., Clayton, Manchester M11 4SR, England The preparation of glycidyl 2-ethylhexanoate has been studied in order to gain an insight into the general mechanism of the reaction of epichlorohydrin with carboxylic acids. The investigation was carried out in two stages. The first stage of the reaction involves the catalysed addition of 2-ethylhexanoic acid to epichlorohydrin to yield 2-hydroxy-3-chloropropyl 2-ethylhexanoate. The reaction is then completed by dehydrochlorination of 2-hydroxy-3-chloropropyl 2-ethyl-hexanoate

to yield glycidyl 2-ethylhexanoate (Fig. 1). This compound was selected in order to give intermediates and products of acceptable volatility to allow gas chromatography to be used the principal analytical tool. Results

and

as

Discussion

Reaction Acid

of Epichlorohydrin

with

2-Ethylhexanoic

The cetyl trimethylammonium bromide catalysed addition of epichlorohydrin to 2-ethylhexanoic acid has been studied at levels of 1, 5 and 7 molar ratios of epichlorohydrin to 2ethylhexanoic acid. The reactions were carried out in toluene in order to facilitate removal of water from the final product (glycidyl 2-ethylhexanoate) by azeotropic distillation. A further advantage in using toluene as solvent is that where glycidyl esters of high molecular weight are being considered the solution viscosities may be kept sufficiently low to effect filtration, if required, before final distillation. The reaction products were determined by gas chromatography. The rate of consumption of the starting materials was also determined by gas chromatography and other standard analytical techniques.

0-8412-0525-6/79/47-114-211$05.00/0

© 1979 American

Chemical

Society

RESIN

EPOXY

DAD

0. L CH2) = CHe

R= COOH

:

Rie CORIO

eee

CHEMISTRY

=) Cha.

Cl

Cage!

OH

R = COO

Figure 1.

Chae Chives

SZ

Overall reaction scheme

O

40

3-0 F vp)

i, AO

2-EHA

1.0 L

O

ECH

1

2

3

4

REACTION

Figure 2.

S

6

TIME

7

8

3)

10

- HRS

Reaction of epichlorohydrin with 2-ethylhexanoic acid (1:1 mol ratio)

15.

CORNFORTH ET AL. 1:1 Molar

Ratio

Glycidyl 2-Ethylhexanoate of Epichlorohydrin

to

213 2-Ethylhexanoic

Acid

The concentration-time curve for the reactants is shown in Fig. 2, and the concentration-time curve for the products is Examination of Fig. 3 shows that as well depicted in Fig. 3. as the desired reaction product (the chlorohydrin ester), there are two other major product components. These were identified as 1,3-dichloropropan-2-o0l (@-dichlorohydrin) and glycerol (hydroxy diester). The structure of 1,3-di(2-ethylhexanoate) the hydroxy diester was confirmed by its independent synthesis The from glycidyl 2-ethyl hexanoate and 2-ethyl hexanoic acid. yield of chlorohydrin ester was only 50.4%, the remainder of the 2-ethylhexanoic acid being converted to the undesirable hydroxy Fig. 4 shows the structure of the products obtained diester. Fig. 5 depicts the formation of the in the 1:1 reaction. hydroxy diester from glycidyl 2-ethylhexanoate and 2-ethylhexanoic acid. A mechanism which is consistent with the experimental observations for the production of the chlorohydrin ester is given in Fig. 6. The first stage of the reaction involves ionisation of 2-ethylhexanoic acid by the catalyst cetyltrimethylammonium bromide. Subsequent attack of the carboxyanion on epichlorohydrin leads ultimately to the desired reaction product of the first stage of the reaction. Examination of Fig. 3 shows that the rate of formation of &-dichlorohydrin and the hydroxy diester closely follow each other indicating that their formation is interdependent. The by-product formation seems best accommodated by the mechanism shown in Fig. 7 which involves breakdown of the intermediate chloro-oxy anion by internal nucleophilic displacement, to yield glycidyl 2-ethylhexanoate.

The

glycidyl

2-ethylhexanoate

subsequently

reacts

with the carboxyanion to yield ultimately the hydroxy diester. The X-dichlorohydrin may be formed by chloride ion attack on epichlorohydrin with subsequent protonation of the dichloro-oxy anion. !Further evidence for this mechanism is also provided by the direct synthesis of the hydroxy diester from glycidyl 2ethylhexanoate and 2-ethylhexanoic acid (Fig. 5). LEesmalso possible that the chlorohydrin may break down thermally to give the undesirable by-products (Fig. 8).

Summary

of

1:1

Reaction

The reaction of epichlorohydrin with 2-ethylhexanoic acid in a 1:1 molar ratio gives rise to only 50% of the desired The mechanism depicted above product, the chlorohydrin ester. indicates that the undesirable hydroxy diester is formed by reaction of 2-ethylhexanoic acid with glycidyl 2-ethylhexanoate. It appears therefore that formation of the hydroxy diester competes directly with the chlorohydrin ester formation (Fig. 9).

214

EPOXY

2:5

RESIN

CHEMISTRY

CHLOROHYDRIN

2.0 eh lJ ‘ail

g wa

HYDROXYDIESTER

oc-DCH

05 r 0

>

fe

1

(a

eh

A

1

eG

REACTION

Figure 3.

1

1

i

A

EE

4

I

ah

TIME - HRS

Reaction of epichlorohydrin with 2-ethylhexanoic acid (1:1 mol ratio; product composition)

@)

R-CO-O-CH2- CH |

OH 2-HYDROXY - 3 CHLOROPROPYL

CHe C2

-‘!2-ETHYLHEXANOATE!

(CHLOROHYDRIN)

@)

Cl-CHa-CH- CH - Ci

OH 1,3 - DICHLOROPROPAN -2- OL (& - DICHLOROHYDRIN) a-DCH 8)

R -CO-O-CH2-CH-CH,-0-CO-R

OH GLYCEROL - 1,3- DI -(2 - ETHYLHEXANOATE) ‘4YDROXYDIESTER HDE’ Figure 4.

Product structures

15.

CORNFORTH ET AL.

Glycidyl 2-Ethylhexanoate

R CO2 H +R CO2 CHz CH- CH, Weg 0

215

—®R CO, CH, CH CH, CO, R | OH

R=C4Hg ° CH — |

C2 Hs

Figure 5.

Glycerol-1,3-di-2-ethylhexanoate synthesis

/

(3)

RaNBr + CHa-CHCH, Cl——=—® BrCH.-CH-CH.Cl + RaN Se | Q

Br CH2-CH - CHaCl + R COzH —— Br CH; - CH - CH) CI +RCO> ge CHz-

OH

CH - CH, Cl

+

C) RCO,

———}®

RCO?

- CH2 -

Sd

CH-CH,2Cl

le

(e)

O RCO2H @

RCO,

tc

R COz - CH2 - Oe = CHCl

OH

Figure 6.

RCO. -

Proposed reaction mechanism Part A

CH2.- CH-CH2

et Oo

ro™ e Cl ———®RCOz - CH2 - CH- CH2 +Cl phenols > water > alcohols nitriles > aromatic > hydrocarbons > dioxane > diisopropylether (6,10). The alcohols are included in this list due to their weakly acidic nature. The accelerating effect of alcohols is dependent

on

butanol

> n-butanol

structure:

It was virtually dent

on

methanol

suggested

by Harrod

independent the

amine

>

> isobutanol

and

of

ethanol

(5) that

temperature

epoxide

> n-propanol

> cyclohexanol

used.

and

the

ratio

k2/k]

composition

Dusek

et

al

>

tert-

(6,10). but

Ci)

is depen-

state

that

ky and kz depend on the autocatalytic effect of the hydroxyls and retardation caused by hydroxyl complexing with amine bydrogens but present no evidence to support this statement. They also

mention

data

to

support

the

constancy

of k2/kj.

The presence of hydroxyl groups during the reaction should be included within the rate expressions. Schecter et al (1) first proposed a termolecular mechanism to describe the effect of hydroxyl groups on the reaction, reactions (5) and (6).

R,NH + ore’ + HO-X —

RyNH---(-C-R?

(5)

HO*X R 5 NH---G-GR’ ars Ts) NC-CR' + HO*x 2

HO*X

In this system, hydroxyl groups in the epoxy ring opening by hydrogen bonding transition. Smith

(10)

(6)

OH

proposed

a variation

of

solvent accelerate to the epoxy group this

mechanism.

the during He

suggests that a hydrogen bond forms between the acid and the oxygen on the epoxide ring followed by a three molecule transition state which is rate determining, reactions (la), (1b) and (lc), Figures Lva

'

a laDG See

tes

AN CF O

(la)

HX

HX

0 R,NH + R'CE

R,NH

—>

2|

a

(1b)

16.

KING AND BELL

Ne,

-OR

Epoxy-Aliphatic Diamine System

RN

SiO

wane

a

ie

RX

fast

eee

227

+ HX

(1c)

H

—-HX

Figure 1. More

recently

Mika

and

Tanaka (12) suggested a mechanism of amine to a hydrogen donor, re-

based on the hydrogen bonding actions (2a), (2b) and (2c).

RONE

EX

[R,NH-- -HX] t,CCR 0

[R NH: - -HX]

(2a)

(eee) H-+-0

(2b)

Pee fa a R,NH*

**C-CR'

| ——

R_N-C-CR'

+

Ge)

HX

we

H

Figure 2. This

system

is

supported

by the work

of

Iwakura

et

al

(8) who

studied the reaction of amines and epoxide in acrylonitrileglycidyl methacrylate copolymers. An accelerating effect by formed hydroxyl groups was noted but the cyano group, which should retard activity, by the hypothesis of Smith, acted as an accelerator by its interaction with amine. Harrod (13) observed hydrogen bonding in epoxy-amine systems with both nitrogen and oxygen. Equilibrium hydrogen bonding would be more favorable between hydroxyl and amine groups than epoxy and hydroxyls due to the difference in basicity. It was also found that epoxy reacted more readily when the added water was equal to amine.

Mika

and

Tanaka

also

suggested

that if reaction (10) occurs, maleic and phthalic acids, to formation of imides, caused by dioxane, tetrabe attributed to the fact

then the retardation effect caused by ketones and esters may be attributed ketimines and amides. The inhibition hydrofuran and diisopropyl ethers may that

the

(R2NH---HX)

reacts

with

ethers

to

yield

a

complex,

thereby preventing the epoxides from reacting. This would be expected because the four, five and six member rings are better hydrogen donors than three component epoxy rings. This report compares the three mechanisms in light of data collected for an Epon 828/diaminobutane epoxy system. It also studies the effect of low molecular weight hydroxyl groups on the mechanisms and rate of reaction.

Experimental Materials.

(Epon the

828)

structure

tests.

The

diglycidyl

(Registered shown

in

ether

trademark

of

Structure

1

bisphenol

of

Shell (n *

Chemical

0.2)

was

A,

DGEBA,

Co.),

used

in

having all

OH

3 cH

0-¢-

=

-tH-CH

{3

-9-0-CH,-

H

oe fi-cu,

-$-0-CH,—

-0--

H-CH.,

H3 Structure

The resin was placed sealed in a nitrogen content was prepared

the

bottle,

isopropyl

added.

Characterization of

Knoll

1

in a vacuum of 30 in Hg for two hours, then atmosphere. Resin with a higher hydroxyl in a similar manner except before sealing

was

procedure

CHEMISTRY

RESIN

EPOXY

228

et

alcohol al

(Reagent, of

(14)

these

gave

Hydroxyl content was calculated These epoxy values were used to ratios of curing agent.

Lehigh

Valley

materials

the

epoxide

by

Chemical)

a modified

content,

Table

lI.

from the structural formula. determine the stoichiometric

1-4, Diaminobutane, DAB (Aldrich), Di-n-Butylamine, DBA (Aldrich), and Tri-n-Butylamine, TBA (Eastman Organic Chemicals), were used as curing agents. Each was distilled under Helium at one atmosphere and sealed. The amine analysis of Critchfield and Johnson (15)was conducted to determine purity, Table I. The purity

the

and

weight

the

of

theoretical

amine

needed

equivalence

were

for

weight

a given

used

of

to

determine

resin.

Curing Process. Two to five grams of resin were weighed into a 10 milliliter beaker. The appropriate amount of amine was added and mixed thoroughly by stirring with a spatula for 30 seconds. This mixture was placed in several 1 cc syringes which served as batch reactors during the cure. Every half hour, 0.5 to 1.0 milliliter of sample was added to a reagent solution for analysis. All reagent solutions used stopped the cure almost immediately. The syringe was weighed before and after the addition to obtain the exact sample weight. For these studies it was desirable to maintain a constant temperature of about 22° C (room temperature). Since the amineepoxy polymerization is an exothermic process, the syringes must have sufficient surface area to dissipate the evolved heat rapidly. Tests with Leeds and Northrup Potentiometer and a small alumel-chromel thermocouple, inserted into the center of the syringe, showed a maximum temperature rise of less than 2° C during the entire reaction time. This was sufficiently small to assume that isothermal conditions were maintained.

this

Analytical paper were

Bell (16) and and secondary

a spectrophotometric amines based on work

Pyridinium

analysis

was

Methods. The analytical methods utilized for a combined epoxy-amine analysis as described by

Chloride

developed

Analysis.

by Bell

(16)

determination for primary by Toome and Manhart (17). The

pyridinium

based

method

on methods

of

described

16.

Epoxy-Aliphatic Diamine System

KING AND BELL

Table

I —

Resin

Starting

Epon

Materials

Epoxy

meq/g 822

229

Hydroxyl

- sample

meq/g

-

sample

adhe

ONS2

4.65

2d AL?

AO)

Bho ey

Pay

6272

Resin

Prepared Resin

[

Prepared Resin

II

Prepared Resin

IIT

Curing

Agent

Purity

1,4-Diamino

butane H,N- (CH, ) 4-NH.,

Di-n-butylamine

(DAB)

SUH

(DBA)

(CH.,CH, CH, CH, ) .NH

Tri-n-butylamine (CH

CH CHAGH)

Date

ae 2S

NG

(TBA) Ni

98.62

EPOXY

230

RESIN

CHEMISTRY

by Knoll et al (14) and Critchfield and Johnson (15). Epoxy first reacts with excess pyridinium chloride in the reagent solution, the remaining pyridinium chloride is titrated with 0.5N NaOH to a pH endpoint of about 11.0. The difference between a blank run and this first endpoint is the epoxy equivalent present. Carbon disulfide is added; this reacts with primary and secondary amine hydrogens to form their corresponding thiocarbamic acids. The acid is then titrated with 0.5N NaOH to a pH endpoint of around 10.0. The difference between the first and second endpoint is the amount of primary and secondary amine. Spectrophotometric Analysis. Florescamine (Fluram*, Roche Diagnostics) is a nonfluorescent compound that reacts with primary and secondary amines to form fluorescent compounds. It has been used chiefly in the spectrophotometric investigation and assays of biological compounds. This investigation utilized the compound to determine concentrations of primary and secondary amines. Excess Fluram reacts with primary and secondary amine forming structures that have absorptions of 375 - 410 and 310 320 nm, respectively. The remaining Fluram is hydrolyzed by water. The procedure developed was based on work done by Toome

and Manhart

(17).

The change in primary and secondary amine concentrations versus time is shown in Figure 3. The system used is Epon 828 and DAB. Both curves are accurate when compared with primary and secondary amine points obtained by another method (16). The secondary amine peak was found to disappear at 2.0 hours. This is explainable by the large size of the Fluram molecule. At 90 minutes tertiary amine production is initiated; this is where branching of the molecule most likely starts. When 2 hours is reached, a highly branched molecule is present. At this point the secondary amines are probably too sterically hindered to allow the large planar Fluram molecule to react. The primary amine is still available because it is a dangling end, exposed to

the

reagent

solution.

Results

A set of material balances were developed to determine the concentrations of various reactive groups based on data collected by the method of Bell (16). These material balances assume that any side reactions are negligible. Aj - initial primary amine concentration -primary amine concentration epoxy concentration ° - initial - epoxy concentration

-

initial

secondary

amine

concéntration

- secondary amine concentration tertiary amine concentration fe) ° - initial WP nNnNWWDW - tertiary amine concentration

16.

KING AND BELL

Epoxy-Aliphatic Diamine System

231

5.0

4.0

3.0

2.0

Primary amine Secondary

EPOXY MEQ/G-SAMPLE AMINES, AND

TIME

Figure 3.

amine

, HRS

Change in epoxy and primary, secondary, and tertiary amines vs. time:

(#) method of Bell (16), calculated with 95% confidence limits, compared to (A) spectrophotometric method (1.0:1.0 epoxy to amine stoichiometry)

EPOXY

DEVs

RESIN

CHEMISTRY

Given: Aj, Bg and Rg = So = 0.0 Data collected: B, (A + R) It

follows

that:

=

A, -

0.0

=

(A +R)

(B -

A=R-

BA) +

(7) (R -

Ry) + 2.065

=

Ss)

(8)

(A+R)

(9)

This data alone is sufficient to determine all reactive groups present, that is, epoxy, hydroxyl and primary, secondary and tertiary amines, provided no significant reactions take place other than between the amino hydrogens and epoxy resin, reactions (Gayman 2) To date, investigations have supported this absence of side

reactions (1,6,18,19) with some special exceptions; low amine concentrations (20,21) less than a 1.0:1.0 amine:epoxy stoichiometry, yield etherification, reaction (3). High curing temperature has had the same effect (6). Since the stoichiometry of epoxy to amine is not less than 1.0:1.0 and the cure proceeds at a low temperature, one would predict that only reactions (1) and (2) would take place. To test this assumption for the DAB- DGEBA system, several identical runs were analyzed by the method of Bell (16). Utilizing the material balances found above (which assume no side reactions), the various reactive groups were determined and averaged with 95% confidence limits. A similar run using the spectro-photometric method gave the amounts of primary and secondary amines separa-— tely. This yields a calculated and actual concentration of primary and secondary amines. Provided there are no side reactions, the two values should be identical. The results were plotted in Figure 3 and found to fall on the same curve, within experimental error. This verifies the assumption that only reactions (1) and (2) need be considered. DBA and TBA, a secondary and tertiary amine respectively, when used as curing agents at 1.0:1.0 stoichiometries, showed no noticeable reaction with Epon 828 over a period of five hours. After two days, however, the DBA-resin mixture turned into a clear glassy solid. Hydroxyl groups were calculated from the structure of epoxy given in Structure 1 with n = 0.2. For each epoxy consumed one hydroxyl was produced.

OH

—

OH, -

hydroxyl added

OH = Original data and this paper appear

concentration

hydroxyl

0.1

concentration

(B) + OH, ar (B

(isopropyl

alcohol)

- B)

calculated values used in in Tables II through VII.

the

(10) development

of

233

Epoxy-Aliphatic D iamine System BELL AND KING

16.

etTdues -3/bow

0°0

0°0

TO*O

60°0O

E710

6¢c°0

TGa0

La)

72) 0

Min)

Cand)

08°0

60°C

coal

LYE

ey AL

Ocaas

LGr i

EG

€6°0

Gilat

oe Il

ee)

oma

ORG

EasG

60°C

Ic ¢

60°C

TO°¢

Bey

Sie

0 Gms

ae

MRS

WES

0

O€

09

06

(et

OST

O8T

eTIeL

ouTJ,

GOYY

(3 / baw) kxody

e3eq

-

(3 / bow) ATepuodseas Azewtig

OSG

sutTue pue

Ny 0)

(3 /bew) TAxorpAy

O€ “¢

(3 / Dew) ouTue

10F

TeJuowtiedxy

QO°T:Q'T

eqeq

Axods

II

0°0

Areutig

‘“AaqQewoTyoToIs

outwe

0°0

(3 /bew) outTue

TetTjTuy

Te) peqweTno

Ale puoseas

vAxoipdhy

eqeq

(3 / bow) ouTue

= UOTe1}ZUBIUOD

AreTAIIL

97°Q

GOmG

OST

/aaal ey

CSG

OcT

O8T

G6°C

06

“ES SE

TSS,

O€

09

ey)

kxody /bew) (3

eIeq AoF

0

J, OuT

OferTit-

cane

88°E

60°7

617

Gy

GG)

UC

Axeutig pue ATepuodss eutTue Dau) / (3

Tequollt eqeq isedxy

Q'z:Q°T sutue-Axodsa

TL

O

Teme

EEGs

LOG

PO

OiGaals

8Z°0

Gy

pA TAx02 baw) / (3

G

SECG

OGmG

I”

Ome

LES

Ws)

UG

TE

T

0)

Shy

67°

ey

Sill

om

8e°0

0°0

yO

ear)

GiO

c0°0

0°*0

0°0O

0°0O

AIeTALe ouTUe (3/bew)

IeAQUeDUO WoT= D Z74°Q

Azepuoosas |, suTUe (3/bew)

TAxoIpdy

Aweutig ouTwe bow) / (3

peqeptno eReq te9

*Azjowotyotojs Terazuy

oTdues

3/bow

234 EPOXY RESIN CHEMISTRY

LG

oy WE

One

IOs

OS

06°0

7S°0

OST

O8T

OCT

=

6S°T

OoE

LDS

=

LEO" G

09

06

fe O Sa

OG

Glue

= O™

ios

OT*O

WLS

O€

80°C

HEE

Gr

Wee's

Ey

S8°0

iO ie,

1G S

SIT LO

O Aart

Os

Q*E:Q°T

-

kxody

SIeC

OTIPL

out]

0

eutTue ATE puodaS pue AleuTig

ouTWe-Axoda

(3 / baw)

Come

pnopeopee TeTITUL

(3 / baw)

WL G

ejeq

(3 /bow)

8e°O (3/bew)

WAS,

(3 /bew)

0°O

0°0

AI

0°0

AOF

Ore (3 /bew)

TAxXOI pAY

TAxOIpsy

*ALQoOWOTYOTOJS

ouTwe ATewTIg

= uoT,eAqueduU0D

ouTUe Ale puodsas

ge*Q

ouTue Ale TIA,

3/bew

UeWTIsedxg Te eqeq

aTdues

235

Epoxy-Aliphatic Diamine System BELL AND KING

16.

Faeq@

— IN Spar,

CIS

WES

78°C

UD

E

S

O€

09

06

OcT

OST

IL

Th

Glee

ee

cme:

L(y

00°72

Gila

G

*Az}QOWOTYDTOIS

Ne

OCW

0

Oeb0eT

SILAS,

OS°€

LGN

LS”

Gi7zaaG

tO

Se

Goum)

SOsT

Te

oa

GUE

AZewtig outue (3/beaw)

OLE

TL OP

8L°0

DOO

(0) us

OMG

puoovss Are ouTUe Dow) / (3

SBE ouTue /bew) (3

lie

Cian)

62°0

Siw)

GW

OmO

Sey Ea

eTdues [T6*T

OUT],

TeTATuT

TAxOIpAH /bou) (3

peqepno eqeq

TAxoapdy

Azewtig pue ATepuodes euTue /bew) (3

Te)

3/ball

kxody bow)(3/

[ejUSuTIe eq e dx|

236 EPOXY

= uotzerAUSs0U0D

|suTWe-fxods

20

RESIN CHEMISTRY

237

System Epoxy-Aliphatic Diamine BELL AND KING

16.

ere @

=

€S°0

om

cO'L

=

66°0

60°T

8L°0

7¢°0

=

com)

Gomez

BGS

sty S

a

coms

rie

OStall

Sa

=

GE I

Ws) 2G

Sy G

OSs

=

O€

09

06

O¢cL

OST

O8T

Je 0 L£S°0

Siac

WLS

we

6T°O

O8'T

0

ecm,

-

Axody

BIeq

STWeL

out,

Se 1

ATepuodasS Arewtig

TOF

[Te ueWTiedxy

Q*T:Q°T

eqeq

suTWe-—Axode

(3/bew)

GS°0

*AtZOWOTYOTOAS

(3 / baw)

¢0°0

(3 / Daw)

come

(3 /bew)

OSE

(3 /bew)

TetqyTuy

pewepnoTe)

{Ax 01 pAY

TAxoipAy

eqeq

ouTue ATeuTIg

UOT eAQUSeDUOD

ouTue Aie puosas

c3°T

euTue pue

IA

0°0

=

0°0

(3 / bew)

Qg*¢g

ouTUe ACTIV

eTdues 3/beu

12

09

O8T

=

Lalla

ise

O7T

OST

671

06

IE

=

Oks

OG WL

O€

O°ws@O°w

98°0

Oma

LG

ie

~0) el

TL

=

ama

ATepuosss euTue (3/bew)

Azewtig pue

uewtiedxg| Te e7eq

BaeG sor

Axodq (3/bew)

0

UTI,

Srey WN= suTWe-—kxoda TAxorpéy

=

=

€6°8

Tan

09°8

62°8

80°L

3

eo

°O 6€

cea

SO LG

=

SEL

TAxOoIpAH ouTwue baw) / (3 bow)(3/

ATewtIig

pezepTnope) ejeq

*ArQOWOTYD TeTITUT TOAS

7

°C AS

Lae

SLO

OAO

0°0

ouTue bow) / (3

puoves Are

=

a

3/bou atTdues

09°0

m0

IEC

GiaaO

0°0

surwe baw) / (3

ELo ATE

eAqUadUO0D uot= go°/

238 EPOXY RESIN CHEMISTRY

16.

KING AND BELL

Epoxy-Aliphatic Diamine System

239

Figures 4, 5, and 6 show the concentration of various reactants as a function of time for varying stoichiometries. In all three cases, the epoxy drop seems relatively linear throughout the test period and shows an increasing rate of consumption as the stoichiometric amount of amine increases. Tertiary amine formation is not apparent until 1.5 hours of reaction for all amine concentrations. Figures 7, 8, and 9 demonstrate the effect of increased hydroxyl concentration (by adding isopropyl alcohol) on the rate of production or consumption of the various reactive groups. The rate of epoxy drop is no longer constant in contrast with the earlier situation when no alcohol was added, but varies with time.

The

production

of

hydroxyl groups earlier case.

tertiary

were

added;

amine

is

immediate

this

is

also

in

in all

contrast

cases to

where

the

Discussion

The Effect of Hydroxyl Concentration on Epoxy-Amine Reactions. Prior work completed with polyamine-epoxy systems has indicated a second order rate law, supporting the S,2 mechanism for nucleophilie attack on the epoxy ring. ($,10,22,23) .Figure 10 is the result of an initial rate analysis, described by Levinspiel, (24) of the present epoxy resins containing low molecular weight hydroxyls, Resins I, II, III, Table I. The graph demonstrates a possible second order reaction. This the apparent zero order dependence of Figures 4, 5 and 6. To be valid, any

is in contrast, however, to the epoxy reaction shown in model must be able to re-

solve

this difference in order. The mechanisms of Smith (10) and Mika and Tanaka (12), although differing, both contain an intermediate complex, Structures 2 and 3:

aR

[Ce

Wes ]

p

; HX Structure

2

Structure

Smith

Mika

It can be seen by examination of the two and 2, that if the hydroxyl groups, when

intermediates, are sterically slow

is

rate

formed.

of

and

3

Tanaka

mechanisms, Figures 1 complexing to form their

Structures 4 and 5, in equations (la) and (2a), hindered (thereby being less reactive) causing a

epoxy

consumption,

a small

amount

of

intermediate

EPOXY

240

RESIN

CHEMISTRY

5.0 aaa

4.0

3.0

SAMPLE MEQ/GAMINES, AND HYDROXYL EPOXY,

TIME,

HRS

Figure 4. Change in primary, secondary, and tertiary amines, epoxy, and hydroxyl groups vs. time for 1.0:1.0 stoichiometry (initial hydroxyl concentration = 0.46 meq/g-sample) (O) epoxy; (CT) primary amine; (A) hydroxyl groups; (@) secondary amine; ([§)) tertiary amine

e@

MEQ/GAMINES, SAMPLE AND HYDROXYL EPOXY, 0.0 uz

ee

ee

a ge 10

raat ee

ee

2.0

ae

a

eee

eee 3.0

TIME , HRS

Figure 5. Change in primary, secondary, and tertiary amines, epoxy, and hydroxyl groups vs. time for 1:2 stoichiometry (initial hydroxyl concentration = 0.42 meq/g-sample) (O) epoxy; (L]) primary amine; (A) hydroxyl groups; (@) secondary amine; (§)) tertiary amine

16.

KING AND BELL

Epoxy-Aliphatic Diamine System

241

uJ aa a

s

B=q w

i

6.0

'

S

6.0 0

=

L|

a

x 4.0

f

:

:

=

“eps:

S fag

:?

20

= ae

ads

40 = ae

;

”=

i

©

30s

° Loge

z

=

:

0)

>

H

=

to 10

f

=

Sd

20

re)

EA

:

uJ

SLO

Zz >

ea ee

O

w

~)

< =

2

Ws

{e)

(a)

==

UV

5.0

o2—__—__2—____-8)

1.0

2.0

3.0

4 1.0

TIME , HRS

[eg

a oO z

[o) oO

J

wn

Figure 6. Change in primary, secondary, and tertiary amines, epoxy, and hydroxyl groups vs. time for 1.0:3.0 stoichiometry (initial hydroxyl concentration = 0.38 meq/g-sample) (OQ) epoxy; (CT) primary amine; (A) hydroxyl groups; (@) secondary amine; (§) tertiary amine

fh

EPOXY

242,

RESIN

CHEMISTRY

MEQ/G-SAMPLE AMINES, AND HYDROXYL EPOXY,

TIME , HRS

Figure 7. Change in primary, secondary, and tertiary amines, epoxy, and hydroxyl groups vs. times for 1.0:1.0 stoichiometry (initial hydroxyl concentration = 1.91 meq/g-sample) () tertiary amine; (O) epoxy; (1)) primary amine; (A) hydroxyl groups; (®@)secondary amine

16.

KING AND BELL

Epoxy-Aliphatic Diamine System

243

5.0

4.0

3.0

SAMPLE MEQ/GAMINES, AND HYDROXYL EPOXY,

TIME

, HRS

Figure 8. Change in primary, secondary, and tertiary amines, epoxy, and hydroxyl groups vs. time for 1.0:1.0 stoichiometry (initial hydroxyl concentration = 3.50 meq/g-sample) (QO) epoxy; (CT) primary amine; (A) hydroxyl groups; (@) secondary amine; ([§)) tertiary amine

EPOXY, HYDROXYL AMINES, AND SAMPLE MEQ/G-

TIME

, HRS

Figure 9. Change in primary, secondary, and tertiary amines, epoxy, and hydroxyl groups vs. time for 1.0:1.0 stoichiometry (initial hydroxyl concentration = 7.08 meq/g-sample) (O) epoxy; (I) primary amine; (A) hydroxyl groups; (®@) secondary amine; (§§) tertiary amine

EPOXY

244

RESIN

CHEMISTRY

0.8

MEQ 0.7

*SAMPLE-HR OF GM 0.6

O5-

—

\-==-slope

=2.0

0.4

0.3

0.2

0.1

0.0

0.1

0.2

0.3

0.4

0.5

LOG OF INITIAL PRIMARY AMINE, MEQ/G- SAMPLE

DISAPPEARANCE AMINE PRIMARY OF RATE THEINITIAL LOG VALUE ABSOLUTE THE

Figure 10. Log of the initial rate of consumption of primary amine vs. the log of the initial primary amine concentration to determine the reaction order

16.

KING AND BELL

= Epoxy—Aliphatic Diamine System

R'G-£

245

[&NH- “BK )

i Structure

4

Structure

Smith

An equilibrium order behavior ted

to

the

Mika

is set up in observed for

slow

breakdown

equations the epoxy

and

low

and

5

Tanaka

(1b) and (2b) and the zero consumption may be attribu-

concentration

of

the

ternary

Structures 2 and 3. intermediate, On the other hand, if the hydroxyls are more reactive, equilibrium formation of the intermediates, Structures 4 and 5 will be far to the right and equations (la) and (2a) may be neglected. It may also be assumed that when the hydroxyls are much more tightly bound to the amine or epoxy, shifting equations (1b) and (2b) to the left, then the association of hydroxyl complexes with epoxy or amine, equations (1b) and (2b), tend to control the rate of reaction and not the

of

breakdown would

cause Figure

complex,

ternary

equations

(lc)

and

(2c).

This

apparent second order reaction. represents the fractional epoxy

an 11

drop for the data shown in Figures 4, 5 and 6. It may be noted that as the stoichiometric amounts of amine increase by factors of two and three, the slopes of the corresponding lines increase by the same fac-— tor. Figures 12 and 13 are plots of the fractional drops of primary amine and production of tertiary amine, respectively. A single curve may be drawn through the data for the different stoichiometries. Taking the value of epoxy present as one, it can be observed in the model of Mika and Tanaka, (12) that as the amine concentration is increased by factors of two and three, the intermediate

present,

creasing amine is half and

5 is

Structure

5,

increases

the consumption of taken as one, then a third. Assuming

low,

due

to

the

low

by

that

factor

thereby

in-

epoxy by the same factor. When the the epoxy is reduced by factors of a that the concentration of Structure

reactivity

of

the

hydroxyls,

then

the

consumption of amine would remain constant due to the large excess of epoxy ready to react. The same line of reasoning may be used to explain the mechanism of Smith (10). When the epoxy is one, then the equilibrium established in equation (la) is constant and as the amine is increased by factors of two and three, the If the disappearance of epoxy is accelerated by the same factor. amine is unity and the equilibrium in equation (la) produces a Structure 4, to react with the sufficient amount of intermediate,

amine be

the

then same

the for

rates lower

of amine epoxy

consumption

concentrations.

ing also supports the model of summarize the possible effects mechanisms.

Smith (10). of hydroxyl

in equation This

line

(1b)

could

of

reason-

Figures 14 and 15 groups on the two

EPOXY

246

RESIN

CHEMISTRY

0.8

0.6

0.4

0.2 REMAINING EPOXY FRACTIONAL

Cee

1.0

2.0 TIME

3.0

, HRS

Figure 11. Effect of initial epoxy: amine stoichiometries on the fractional consumption of epoxy (©) 1.0:1.0, epoxy:amine; (1) 1.0:2.0, epoxy:amine; (A) 1.0:3.0, epoxy:amine

1.0

0.8

0.6

0.4

0.2 REMAINING AMINE PRIMARY FRACTIONAL

0.0

eee 1.0

=——t 2.0 TIME, HRS

cB inst ee 3.0

Figure 12. Effect of initial epoxy:amine stoichiometries on the fractional consumption of primary amines (O) 1.0:1.0, epoxy:amine; (1) 1.0:2,0, epoxy:amine; (A) 1.0:3.0, epoxy:amine

16.

KING AND BELL

z

Epoxy-Aliphatic Diamine System

247

0.15

=

a

< |ol

a

ae

0.10

w

(e) WJ

n

jeg ea |

3°

0.05

|

oO

¢ joa Ww

0.00 1.0

2.0

3.0

TIME , HRS

Figure 13. Effect of initial epoxy:amine stoichiometries on the fractional rise of tertiary amine. Total conversion to tertiary amine is 1.0. (O) 1.0:1.0, epoxy:amine; (() 1.0:2.0, epoxy:amine; (/\) 1.0:3.0, epoxy:amine

Slightly

reactive

hydroxyl

groups:

R2NH + HX (RoNHeoeHX]

9wee

+ a

—[R2NHoo oHX]

fast

RgNH *°C~CR}

‘S

XHoeoe

RaNHe + eC=CR}

slow

©o00

Highly

reactive R2NHee

eHX

+

C-CR]

R2NHe ee

+ HX

H

groups: slow

@)

RzNHe oe C=C]

—

(RNH, * * + HX)

(11)

R,NH

(R,NH* * HX)

C12)

+ HX =>

Such an association would depend on many factors, including the relative ease of removal of the amine hydrogen, i.e., basicity. This ability may be related to tte pKg values of the amine. DAB consists of two primary amine groups separated by four carbons. It has been found that these two amines have differing pK, values (25) of 11.15 and 9.71. Such values would be interpreted to mean that the first primary amine could react more quickly

than

the

remaining

one.

Damusis

(26)

found

that

the

reaction

of

two to six carbon, aliphatic diamines is rapid for the first amino group, but the reaction of the second primary group is not detectable, being achieved only by catalytic action. It is also indicated (25) that aliphatic secondary amines have a pK, value of around 10.7, falling between the values of the two primary

groups. If the hydroxyl groups present are not very reactive, the high molecular weight resins, then it would be assumed the first group on the diamine would react first, followed the

secondary amine and the second primary Figure 12 shows a leveling out of the

tration at 50%, suggesting that later mary does not react as quickly as the a

as in that by

plot

of

the

percent

of

tertiary

amine group. primary amine concenin the reaction the prisecondary. Figure 16 is

amine

‘produced

versus

primary

amine consumed. As can be seen, only primary amine reacts until about 30% consumption then tertiary amine is produced. At about 50% primary amine remaining, the rate of secondary amine reacting

is twice that of primary. This would tend to indicate a threestep reaction which could be described by a two-step primary and a secondary reaction. The secondary-alcohol amine has been found to be more reactive due to the hydroxyl group (8). Laird and Parker (7) have suggested and presented evidence that the hydroxyl group

has an intermolecular "self-solvating" effect; a similar statement was made by Allen and Hunter (28). A hypothetical intermediate being formed between the secondary-alcohol amine and epoxy

is

presented

in

Structure

6.

EPOXY

250

RESIN

CHEMISTRY

AMINE TERTIARY OF RISE FRACTIONAL 0.6

07

0.8

0.9

FRACTIONAL PRIMARY AMINE

1.0

REMAINING

Figure 16. Primary vs. tertiary amine, fractional changes for varying stoichiometries (OQ) 1.0:1.0, epoxy:amine; ()) 1.0:2.0, epoxy:amine; (A) 1.0:3.0, epoxy:amine

0.8

06

0.4 REMAINING EPOXY FRACTIONAL

0.2

0.0

= 2.0

TIME , HRS. Figure 17. Effect of initial hydroxyl concentration on the fractional drop of epoxy for a 1.0:1.0 stoichiometry (O) intial hydroxyl = 0.46 meq/g-sample; (()) initial hydroxyl = 1.91 meq/g-sample; (A) initial hydroxyl = 3.50 meq/g-sample; (@) initial hydroxyl = 7.08 meq/g-sample

16.

KING AND BELL

Epoxy-Aliphatic Diamine System

251

CH,mas——CHR RNH7*

\H

te

Structure

Further

evidence

of

the

6

probability

of

the

Mika

and

Tanaka

model (of hydroxyls complexing with amine) is an observation during experimentation. In Resin III, the isopropyl alcohol was not totally soluble; when the mix was homogenized it formed an As amine was added, the mixture became miscible as emulsion. evidenced by the almost immediate disappearance of turbidity.

Added

low

molecular

weight

hydroxyl

groups

accelerate

the

reaction between amines and epoxy as noted in Figure 17. This representation shows the fractional loss of epoxy for Epon 828

and

Resin

systems

I,

II

and

III,

each

of which

has

increasing

hydroxyl concentrations, Table I. There is an obvious disappearance of linearity when one compares Epon 828 to systems I, II and III; this change is probably due to the more active added hydroxyls as opposed to the formed hydroxyls.

Figure

18

demonstrates

the

increased

reactivity

of

the

amine

groups when in the presence of hydroxyls. It plots the change of primary to tertiary amines. Resins with high hydroxyl group concentrations are more reactive and show an immediate production of tertiary amine.

A relation was developed in Figure 19 for hydroxyl to epoxy ratio on the initial rate of hydroxyl creases;

concentration rises, the this would probably hold

effect of reaction.

the As

the

initial rate of reaction intrue until dilution effects

dominate. Similar relations have been developed by Bowen and Whiteside (29). It should be further noted that this curve represents isopropyl alcohol as the hydroxyl source. The slope would shift for other alcohols depending on their reactivity. Kinetic

Modeling

Approaches.

This

section

contains

some

possible approaches to modeling of the data and mechanisms proposed in this paper. It also includes recommendations for future work to analyze the validity of the suggested models. The cases presented take into account the zero order behavior at low hydroxyl concentrations (Case I) and the second order reaction observed at high hydroxyl concentrations in conjunction with the three step mechanism proposed previously (Case II). Case

IL:

tions

Case I examines the runs made with low initial concentraThis data is represenof high molecular weight hydroxyls.

ted

in Figures

4, 5 and

6.

EPOXY

252,

RESIN

CHEMISTRY

SAMPLE MEQ/GAMINE, PRIMARY

0.0

0.1

0.2

0.3

TERTIARY

0.4

Os

0.6

0.7

AMINE, MEQ/ G-SAMPLE

Figure 18. Effect of initial hydroxyl concentration on primary vs. tertiary amine for a 1.0:1.0 stoichiometry (O) hydroxyl = 0.46 meq/g-sample; (T)) hydroxyl = 1.91 meq/g-sample; (A) hydroxyl = 3.50 meq/g-sample; (@) hydroxyl = 7.08 meq/g-sample

40 Po

fe) |

Qa

=

a

3.0

ae on

Cw

ra

23 aa aes

2.0

fone) lu a

Ew ce

1.0

=

ct

=

=

0.0

1.0

2.0

3.0

INITIAL HYDROXYL TO EPOXY

Figure 19.

4.0 RATIO

Initial hydroxyl to epoxy ratio vs. initial rate of epoxy consumption for a 1.0:1.0 stoichiometry

16.

KING AND BELL Figure

zero

11

order.

Epoxy-Aliphatic Diamine System shows

It

the rate of in contrast

that

also

the

shows

consumption

that

as

the

of

amount

253

epoxy of

is

amine

apparently increases,

epoxy consumption increases proportionally. This to the normal zero order equation that states: r, B =

dB=k dt

is

Gls)

=a

This equation has no provision for changes in stoichiometry. To explain this, one must take into account the complexing of hydroxyls with amine as shown in Figure 15. Assume: 1) hydroxyls only complex with amines 2) all amine-hydroxyl complexes are equally reactive with

epoxies.

Initially only primary amines are present. An increase in the amount of amine will cause a proportional increase in the amount of hydroxyl intermediate formed:

k ela

=

k

A-H

(14)

2,

el eae Se

"5

7@ @

ise!

Increasing the amine, then, does not change the value of the equilibrium constant but does allow for a proportional increase in the amount of hydroxyl intermediate present, which in turn, is available to react with the epoxy. This would account for the paradoxial effect of the varying rates in Figure 11. Assume: 3) rate of reaction is dependent only on the breakdown of the ternary intermediate (zero order); all ternary complexes breakdown at the same rate 4) consumption of various reactive groups is dependent only upon their relative composition in the hydroxyl complex state. = lst primary amine Given: = 2nd primary amine = secondary amine lI hydroxyl = ternary complex = epoxy = tertiary RAWPp HUA

The

first

is

reaction

At+H

B+

amine

the

hydroxyl

ies

H-——>

6 +H >

complexing

with

amine:

(16)

B’H

(17)

CH

(18)

EPOXY

254 Assume: 5)

of

the complexing rapid.

epoxy

CHEMISTRY

complex

hydroxyl

the

with

RESIN

JNA

ae 1B)

Ey

(19)

OVE

a= 1b) =~

ER

(20)

Gok

se Dae

Eo

2D)

EL

———

products

2H

is

(22)

x

From

the

above

the

assumptions

following

k

k

A+H=>E,—>C+B

model

is

proposed:

+ 2H

ey

(23)

k

1 ap ial =

Ky

EB — > C + 2H

(24)

1

i k, Wi Coemeas Wee een

an

(25)

iro, Ou

2

The the

disappearance of the various reactants following differential equations:

may

be

described

sa =r, =-k, (4) @ +k, (E,)

(26)

dB _ = = ky ’ (B) foes

(H) + ky ' (E,) + k, (E,)

27 C27)

“ a

(HB), +

fee Sige

a

(C)

k,"

(EQ) aS k3 (E, +

The consumption of epoxy would equal described by the zero order reaction

= Therefore

the

the

of k,.

value

slope

It

of

is

= the

Ep)

1.0:1.0

recommended

stants be determined by equilibrium with the use of IR to determine the

(28)

the breakdown of E, and observed in Figure ll

Ss

by

be

(29) stoichiometric

that

the

case

kj and

would

kg rate

yield

con-

experiments done separately extent of hydrogen bonding. The use of IR for determination of Ea> Ep and Ec would be impossible in the reaction mixture since the absorption bands for these three components would fall within the same area; an overall amount of hydrogen bonding may be calculated by IR. It

16.

KING AND BELL

should amines

Epoxy-Aliphatic Diamine System

255

be noted that hydroxyls need not only associate but may also bond with epoxies simultaneously.

with

Case

II. Case II examines the runs made with added low molecular weight hydroxyls. The data is shown in Figures 7, 8 and 9. Figure 15 shows the proposed sequence of reactions in the presence of these low molecular weight hydroxyl groups. Based on this and the following assumptions, a model may be proposed. Assume: 1) the formation of hydroxyl intermediate is rapid and

A= A-H B = B-H Ci=eCeH

(30) (31) (32)

2) From

the breakdown of the ternary intermediate compared to the formation. these assumptions the following mechanism may

ic [Pay Sees he eee =) C + 2H B + D ——~— GP)

The the

an eee

is

rapid

be

proposed:

(33) (34)

a od

(35)

disappearance of the various reactants following differential equations:

may

be

described

a = 1, = -k, (4) (@)

(36)

dBye r= _ ik, (A) @) e -k, (B) ©) Bee

(37)

“ =r, =

(38)

dividing

dB

the

k, (A) @) +k, () @) - k, () (@)

first

two

equations

we

when

by

obtain:

(B)

(39)

When this is solved, using the initial and final concentrations of A and B as limits, the following equation is obtained:

DS

hy

ba

= § rf

aes

«attest a

serie eet

fyae

Atty

ye

"ie

are

ey

_ : SoH

74

1 =.cts

+f Sort? ced

oe

;

vtbbyD ee: $ ie

iene

id

ba Wht i,

é eer

4 pnp omitl pot

2 Sled

oy

I esnelf

hn tay

“= i

emia Se Yous

)

re

id oul

4

ae

ADE

Ey, bishye

‘

aie

a

£1)

eae

ins’

Th ;

if:

at

Qe

q

i m

phim ani

a

ae

“vet 0 wl oma, etn?

he

all

259

17 Epoxy Resins Containing a Specific Vulnerability JAMES

R. GRIFFITH

Chemistry Division, Naval Research Laboratory, Washington, DC 20875

The concept of a coating with a built in vulnerability to destruction by specific reagents has been used successfully in some applications. For example, floor finishes which can be removed readily when renewal is required by the application of dilute ammonia, or another mild reagent, are available and The extension of this concept to rugged, durable convenient. entails the requirement that a means such as epoxies, coatings,

be

to

found

degrade

an

network"

"infinite

molecular

structure

A need exists for such which is inherently resistant to attack. materials, however, because the factors of toughness and durability which are desirable when epoxy coatings are applied become expensive, troublesome obstacles to removal when the time arrives

for

renewal.

The

present

work

concerns

some

efforts

to

synthesize epoxy resins which contain a cleavable weak link that allows the cured network to be chemically degraded into smaller The consequence of such degradation should be that fragments. such coatings would peel from surfaces or become soluble in

common

solvents

after

treatment

with

For convenience, such a selectively called a command-—destruct resin. Properties

of

An

Ideal

a "triggering" destructible

Command—Destruct

reagent.

polymer

is

Resin

An ideal command-destruct resin should retain the desirable properties of conventional epoxies; including, reasonable cost, cure by available curing agents, ability to be compounded with etc. to form high-quality paints, and solvents, pigments, In addition, the cured coatings provide long service life. should retain the ability to be destroyed quickly and effectively by selected reagents not likely to be encountered in the service A high degree of destruct agent specificity insures environment. Also, the destruct against accidental loss of the coating.

This chapter not subject to U.S. copyright. Published 1979 American Chemical Society

EPOXY

260 agents

should

be

mild

chemicals

without

harmful

RESIN

CHEMISTRY

environmental

Cm neCinsr

All of the ideal properties may never be found in real systems because of a multiplicity of problems concerning economics, synthesis possibilities, destruct effectiveness limitations and environmental stability. However, some of the necessary attributes of effective materials are being incorporated into the following resin types. Cleavage

of

Olefinic

Bonds

One approach to an effective command-destruct epoxy is that of incorporating one or more olefinic bonds into each molecule such that the final network can be degraded by the well-known reaction of olefins with powerful oxidizing agents, such as acidic

permanganate:

[oJ R-CH=CH-R' One

of the

simpler

or be cease point

(284°C)

2

RCOOH

compounds

ener

of this

CH? CECH M/

————3

+ HOOCR'

of this

ee although diphenol

00) cucu

type

the

make

(O) oc

is

the

expense

the

diglycidyl

ether

and high-melting

substance

impractical:

CHCH

e\

Wi 2

O

The product obtained in a standard diglycidyl ether reaction with epichlorohydrin and. alkali was an off-white crystalline powder melting at 191 21028 Cs A lower melting diphenol which contains several oxidizable sites is curcumin. This is an intensely colored pH indicator which undergoes various vivid color changes during the synthesis of the diglycidyl ether. The following formula shows the simplest form, although the actual products obtained to date have apparently been polymeric: HCO

0

(al

0

CH, 27 CHCH,0 5 (O) CH=CHCCH, 5 CCH= CH cx 0)

OCH, (O)0c, ce jie

0

A third compound of the olefinic type, and the most promising obtained to date, is the diglycidyl ether of 2,6divanillylidene cyclohexanone, which is produced by the conventional reaction of the diphenol with epichlorohydrin:

17.

GRIFFITH

Epoxy Resins with a Specific Vulnerability Cr B Ome

O

OCH

i

5

weoren,0 (CO) CO ca) Om 0

CH, CHCH,0

This

range

bright

CH=

yellow,

of 143°-148°C,

and

261

=

crystalline

compound

it is virtually

0 has

insoluble

a melting

in common

epoxy solvents. However, it will react with polyamide curing agents, and films of good quality have been produced by use of dimethyl sulfoxide as solvent. The films so produced are readily destroyed by a solution of potassium permanganate in 50% acetic acid - water. Permanganate in water alone is ineffective, and the solvency of the acetic acid is necessary in order to swell and soften the cured films enough for the permanganate to penetrate to the olefinic sites. After degradation, the film residue may be easily brushed away from the surface to which the film was formerly tightly bonded. Another approach to a diglycidyl ether of this type which results in a product of improved solubility characteristics is the reaction of two moles of a pure diglycidyl ether of Bisphenol-A with one mole of 2,6-divanillylidene cyclohexanone. This results in a resin with glycidyl ether terminals which is partially soluble in ketone solvents, particularly methyl ethyl ketone. Further work will be necessary to produce practical paints from the 2,6-divanillylidene cyclohexanone system, but as an example of the possibilities of the olefinic approach to cleavable epoxy resins, it has been reasonably successful. Cleavable

Tertiary

Carbamate

e)

Resins

Tertiary carbamates, or urethanes, are not readily as the primary and secondary compounds,

synthesis

has been

described

(1).

This molecular

readily cleaved by warm solutions of mineral ease of cleavage may be controlled by proper the components of the tertiary radical: 0

R'R"R™

\]

produced as but an effective

unit

can be

acids (2), and selection of

the

H+

COCNHR ——> A

R'R"R™

COH + CO, + HNR

Epoxy resins may be produced which contain tertiary carbamates by the use of glycidol as the chain-terminating For example, a alcohol in a diisocyanate - diol reaction. equivalent of one which in. used been has stepwise reaction glycidol has been reacted with 1,6-hexamethylene diisocyanate to produce a mixture represented ideally by the reaction:

CHEMISTRY

RESIN

EPOXY

262

CH,CHCH,OH + OF CSIhCH 5) = C=O:———

NG, 0 0

i CH, CHCH OCNH( CH, ) ae

ane

0

CH bes CH,

3°

system by the

0 i

0 H

'

CH, CH, OCNH

is a lowformula:

CH,

Gree

CH,

ons

Jee

0

This resin may be melted and blended with amine curing agents to produce tough polymers which are vulnerable to cleavage at the tertiary linkage. As with the olefin - containing epoxies, much is yet to be

done with tertiary carbamate epoxies in order to produce practical systems, but the synthesis possibility has been established. A variety of diisocyanates and tertiary alcohols are available. A general synthesis effort is needed to screen the most effective compositions with regard to solvency, compounding, cure, environmental stability, and the effectiveness of various destruct agents and conditions. Literature

1. é.

Bailey,

pie)

Bailey,

5(1),

Cited

W.

J. and Griffith,

J.

R.

Polymer

Preprints,

196},

W.

J. and Griffith,

J.

R.

Polymer

Preprints,

1964,

meebo =2 7c.

2279-290.

RECEIVED May 21, 1979.

INDEX

INDEX A

B

Acid—amine equilibrium .......0..0........... 249 Acrylonitrile-glycidyl methacrylate COPOLVINICKS vere een ie eee 227 Adhesion promoters, alkali-resistant .. 59 Alcohol(s) ....... 6 SN Bat Cmca os teeta, 226 amines SCCONGALy= = eye enn: 249 Alkali-resistant adhesion promoters .. 59 Amidoamines: 0ins.. 5 attire 101 Amine(s) -catalyzed oligomerization of epoxy compounds, tertiary ............ 205, 208 -cured epoxy networks, creep behavior Ob dow att oe peed RY 183-195 hydantoin epoxy resin chemical resistance of aro-

Benzenediazonium hexafluorophosDIACSSMinAurore tit moc citieter nett 26t Binder(s) ANION Clee eee a eee Semen eh oe 57-69 LE BIR 25 te ech cece a eee ee ee ted 59 CallOnlG™ -pyeeieraee ie Gre kee eee 65-69 for electrodeposition coatings with high corrosion resistance, CDPOxyaLesin= bascdie ese eee 57-69 with high corrosion resistance, AnOMICy Li) eee eae ee Ree 61-65 hydrolytic stability of ....0.000000...... 64 E:MPB: based emma ner sae tere, 57-69 O Herich garde. tire se tah a ei 65 OUEITEERCALLOTICH ee sean eee 67 performance of anionic ED ............ 58t Bisphenol A -based epoxy prepoly-

THATIC Me ee ey Bee.

129¢, 131¢

solvent resistance of aroIMAG’

ve .c te eee eee ee128t, 130¢

solvent/chemical resistance of BTOMaAL Ce eee ee eee 132¢ resins, hydrophobic—hydrophilic balance ob = eee,

118, 121

curing of hydantoin epoxy resins, room

temperature

...123¢, 124t, 127¢

Gy cloalipia trees meee ee eee 101 equuliiriumdisea Cid —sees ere eee 249 with epoxy, effect of pK, values on themeacktonyotipees wt mee 249 —hydroxyl complexes ..............2..:00. 253 reactions, effect of hydroxyl concentration on epoxy ................ 239 Secondaty=alCODO) ema aes neeee 249 tertiary ......... Se rE dete eee tien, 225 Ammonium halides, quaternary ........ 48 Anhydride-cured epoxy resins, quaternary phosphonium compound latent accelerators for .................. 47-56 Anhydride curing agents ...................... 48 Aqueous epoxy resins for electricalreinforced plastics industry ........ 77-82 PIRATE ASS a core eee eee aac 71-75 Arrhenius-type relationship ................ 187 Arsonium compounds, quaternary ...... 59 Ary Idiazonium Salts faqctem settee. 1-16 IG LONT SIS OL Mer Meee asad sca siai: 1-16 Ary odontum, Salts? noe ee... eco hvinstlc 19 Automotive primer coatings ................ 73 formulation and performance properties LOme en ea ee Me hs TAt DO AAN7OISO Duty TOUT Cmene eee mee 198

265

LENGIEY.

Sa lnesetsanasael 49, 108, 137-156, 158

diglycidy] ether of

(DORE

eee eee 83-90, 227, 261

Blushing of can coatings, sterilizaLIOMETESIStANCC OLNe ee teeta 91-98 Blushmeterpee oe ee 93-98, 96f BY GnSteclhacicl Serene eee es eee 2 7 Bucherer TeachOngeeset eee een 116 Cc

Calorimeter, differential scanning ...... Can

8

coatings, sterilization resistance

(oliishing)) oleae ener 91-98 end Coating ieee eee eee ween 72 formulation and performance JOIKO) YESH ALES GY?’ os seseecr cao san aconce: 73t WEYeCoRUNES TPE corcceonnencunnreceeests hosoner 91-98 Carbamate epoxy resins, cleavable bertlar ye carrer ct ectcomer eae:261-262 Gar bona oeNIV) ae eee een 84, 87

Carboxylic acids, reaction of epichlorohydrin with «0.0.0.0... 211-224 @ardolitey arate eee es99-114 GARD URAC E LO Renin as ats eet 62 Gatalystsmlatenty meme sen ereser ae 47-56 Cationic polymerization Oley clickethersmmemmart treter ace 19 of epoxides, photoinitiators for ........ 7 of epoxy resins, photoinitiated ........ 1-15 Cetyltrimethylammonium bromide ...._ 213 Chemical resistance immersion tests .. 100

266

EPOXY

p-Chlorobenzenediazonium hexaHUGKTOPMOSp ate area ameet 19 Ghiorobydrinvéster mean re 213 Gecomposition mes te ene 218f Circuit boards, epoxy printed. ............ 84 Coating (s) Call CG ierrcen eee ee ere ener oe ae 72 EDOXY. ESteL menu ne eee eee 59 formulations with phenalkamines ...........0..0...... 112¢ and performance properties for AULOMOLIVED PLIMCI Ieee

TAt

and performance properties for Can CNC age ee eee 73t solvent-free liquid epoxy .............. 108 with high corrosion resistance, epoxy resin-based binders for electrodeposition ..................-. 57-69 light-curable epoxide ..............0...... 17-42 prepared from high molecular weight epoxy resins, waterIbOrne bee torte ig rcens ene Be71-75 salt-spray resistance of cationic ED 69 sterilization resistance (blushing) of Cant ee OR ea 91-98 SUittaComme eee ak. ee 1-16 Command-destruct resin, properties Oban 1deal'* iene reece: 259 (Cormosioniresistances ee eee 59 epoxy resin-based binders for electrodeposition coatings with niche eae ee 57-69 GreEp ee ieee nen 183-195 behavior of amine-cured epoxy netWOLK Sete e ot en Sate oreo 183-195 OlfepOxyaTesiNs emer neers 145, 187t ETOG UNSC Dav Ole an eee 184 Cross-link density (M._) .......... 137-156, 161 distribution on the engineering behavior of epoxies, effect OLA eee 137-154, 183-185 properties affected by the distribuTION OL Na ee, HS A Cue eee 154 properties not affected by the disEDO ULE ONO Lae een 154 Cross-linking of epoxides ............0.0..... 42 G@urcuminwee, enn eee, oe Cure Of DEE BATTEsin sraterOLee ee ee effect of epoxide structure on rate of...

260 88t

35 35

of humidity on rate of .................. of photoinitiator concentrator on rate: ob eeee ee 22¢ of Gpoxy resins, UV. 2 eet 1-16 Tate studies =e see eek ee ene near infrared spectroscopic

CNER YS cee as hr aes

103

100

RESIN

CHEMISTRY

Curing agents

anhydride?—san.. anne 48 comparison of properties of ........ 102

-diluent-accelerator mixtures, epOxy Tesi) pee eee 105 CDOKY Sp ereec eee ere ae ee eee 99-114 PYOPErtles; cond eae ee ee 100

Dhena kane Process

ween

°c. eee

Be

ee

ee

102¢ 33

Of thermoscts = eens 157-182 Cycloaliphatic amines'# Aiea See ey See 101 anhydride curing of 1,3-diglycidylhy dantoinsaes eee eee 119¢, 120¢ epoxide i Se ee 35 Cyclohexane oxide, photopolymeriZAtIOUPOL ere sends ce eee 8, Of Cyclopentamethylenehydantoin .......... 118 D

DAB (1-4, diaminobutane) ................ DBA (Di-n-butylamine) .................. 228, Dehydrochlorination reaction ............ DGEBA (see Diglycidy] ether of bisphenol A) .............. 24, 48, 83-90, 5,0-Dialkylhydantoins .............0..00 Diaminobutane epoxy system, Epon $28 se ahah, MEE eee 1,4-Diaminobutane (DAB) ................ Diary hodoniumipsalts =, eee er arylationVols Sie. Meee eee

228 232 219 116 116

227

228 2 3 DHOOM ALOLS Ne eee 7, 24, 28 Diazontumiysalt((S) eee ee ee 19-24 reactivity of epoxides toward .......... 25¢ Di-n-butylamine (DBA) ...................... 228

a-Dichlorohydrin (1,3-dichloropropan-2-ol) Wa ee ee Dielectric breakdown voltage ............ Differential scanning calorimeter ...... Diglycidyis 1)ME see eee Diglycidyl ether(s) of bisphenol A (DGEBA)

213 79 8 126

24, 83-90, 227, 261

epoxy resins based on .................. 83-90 resin, rate Of Cure Of eee

eee

of 4,4-dihydroxystilbene ................. of 2,6-divanillylidene cyclohexanone 1,3-Diglycidyl-5,5-cyclopentamethy1enehydantoi), 4:4... see 1,3-Diglycidyl-5,5-dimethylhydantoin 1,3-Diglycidylhydantoins ................ 116, cycloaliphatic anhydride curing Of

88t

260 260 118 116 117¢

ce nrtateuie ae eee119€, 120¢

1,3-Diglycidyl-5-methy1-5-ethylhydantoin S290) Poe ar ayer 4,4-Dihydroxystilbene, diglycidyl ether oftcri) i. tines 2a tee See

118

260

INDEX

267

Diisocyanate—diol reaction .................. Dimethylhydantoin (DMH) .............. -based hydrophilic resin .................. ie chy Pe ic ered thederesees cx! Dimethylolpropionic acid (DMPA)... Diphenol ir". Webern scinicigttckissatn 2,6-Divanillylidene cyclohexanone, dighycidyl ether Of ..........icisscwe DMPA (Dimethylolpropionic acid) .. Dynamic mechanical properties of EDOXVeTCSINS, Fee, tues seats peer eens

261 12 126 126 61 260 61 141

E

ECC (3,4-epoxycyclohexylmethyl-3,4epoxy cyclohexane carboxylate). 32 ED (see Electrodeposition ) EEW (Epoxide equivalent weight) . 83-90 Electrical properties of cured epoxyanhydride resins containing quaternary phosphonium compounds 54¢ Electrodeposition (ED) ..............0..... 57-69 EABCSTS Ie AL we easiicscs 59 CAEL OMI CRearns oteretdeer arth 65-69 with high corrosion resistance, AUMLOWIC ached. Pree ee .... 61-65 performance of anionic ................ 58t coatings

with high corrosion resistance, epoxy resin-based binders LOT Mee teat ae eeeea 57-69 salt-spray resistance of cationic... 69 DAUMESem Io meMECC secures ere 65 Electron microscope, scanning

IML

sae,babeorrertart texted Yo 160

Electron microscope, transmission

(TEM) See een eae te. 160 Emulsions, oil-in-water ...... fee eee 71-75 E/PF (see Epoxy resin/phenolictormaldehyde))/Meu.ckeee te 01-08 iEpichlorohy. cna te eee eke 116, 260 with carboxylic acids, reaction Ogee EE cre eee OL Looe with 2-ethylhexanoic acid, reaction (6) ee en meer tartare 211-224 iy COLYSISHOLtnenee pete en eee oe 219-224 EPIROUE, LOOT (CE-LOOT) ....05..040.5. 59 Epon 828/diaminobutane epoxy system .. 227 828-methylene dianiline system ...... 184 DUSDOLVIDENS narasscsMerae ce brtttee 138 Epoxide(s) coatings, light-curable ..................... 17-42 Gross-linkingy Ober eee ee: 42 cycloaliphatic). SESW ince 35 equivalent weight (EEW) ............. 83-90 determination by carbon-13 nuclear magnetic resonance

ENN Ieee ober oe

83-89

Epoxide(s) (continued ) monomers, low-viscosity .............0.... 40 photoinitiators for cationic polymerization of ............ ie CYOsS- LINN OL heey eee 18¢ Photosensitized (cto esradans 17-42 SLOLACGRstaDilit val maaan 22724732, structure on rate of cure, effect of .. 35 tack-tree. timeyol at.inet eee 36t

toward diazonium salt, reactivity of 25t HES Dorey asks ee mee ere nredommeacearnie 22) Epoxy (see also Resin(s) ) —aliphatic diamine system, reactions LIN ALL eter At eewdttrn) tal toetet225-256 —amine reactions, effect of hydroxyl CONCENtTrAlON OD te 23) —anhydride resins containing quaternary phosphonium compounds, electrical properties of cured ...... 54t gel time data for quaternary phosphonium latent accelerBtOrS ie ten peter 5lt viscosity (storage) data for quaternary phosphonium Acceleratorswiniete aero 52t bisphemoliga yarncaeeraah. teeerste 108 coating formulations, solvent-free IWejbutsle cca amecee rede posh wor eetee eee 108 coatings, volatility of photosensiLIZ e aged: Reenter cane aaeee 40-41 compounds, tertiary amine-catalyzed oligomerization of ......205, 208 CULING kaCCNtSiee en eee ae99-114 effect of cross-link density distribution on the engineer behayLON OL erent ate reenter te137-154 effect of pK, values on the reaction

Ohi ANTM ARAL yoy cess een 249 ESLET COAUIIGS maeanmme tin meet ceed 59

Iho ri[gies as Gaal Pek Ry Ea? networks creep behavior of amine-

108

CULE Jrateoh cee sh cube et 183-195 micro- and macro-structures of MD A-curedie ss areaee 160 morphology effect of molecular weight of the prepolymer on ............ 166 effect of stoichiometry on ........ 161 and the mechanical behavior Olen ie reo cet 157-180 LWO-DMASC a mente een tenes 157-182 PrepOL Mele streak hae 177 bisphenol-A-based. .................... 137-154 printed-circuit boards ...........0..0....... 84 rate determining factors in the photoinitiated polymerization (O30 cco sachin scteaa de a mcr eae 10

268

EPOXY

Epoxy (continued ) resin(s) -based binders for electrodeposition coatings with high corrosion resistance ....... Mie ere 57-69 pasedson) CH:B Agent ere 83-90 chemical resistance of aromatic amine-cured hydantoin ..129t, 131¢ cleavable tertiary carbamate ....261—262 creep behavior of ....................145, 187¢ -curing agent-diluent-accelerator mixtures

dynamic mechanical properties of 141 effect of alkyl substituents on the properties of cured ihnydantoin yes ere 115-135 for electrical-reinforced plastics industry, aqueous ................. 77-82 emulsion, thermoplastic modified 75t esters, amine-modified .................. 67 glass transition temperature

(15) Tobe

ee em.

142

hydantoin- Ses cat ARO Fh ea De 116 intermolecular forces of hydantoin-based latent catalysts for . /phenolicformaldehyde (E/PF) 91-98 Canylacguens maar ante te eee 91-98 photoinitiated cationic polymerization of .. ... 1-15 properties of cured hydantoin: Jeo alas} quaternary phosphonium compound latent accelerators for anhydride-cured .................... —56 room temperature amine curing

of hydantoin ..122¢, 123¢, 1248, 127¢ rubbery modulus of .....................- 142 solvent resistance of aromatic

amine-cured hydantoin ..128¢, 130¢ solvent/chemical resistance of aromatic amine-cured

hy dantornae er systems, bisphenol A ....................

132¢ 49

tensile behavior of ................00.0145 UNVecurelOl ea er er eee 1-16 Walter DaSCClmm et een eee ee eS?

water-borne coatings prepared from high molecular weight Hele Young’s modulus of Solid Ges hie Sei arene eee one ee een ae 3,4-Epoxycyclohexylmethy]-3,4-epoxycyclohexane carboxylate (ECC) 82 2,3-Epoxy-1-propyl methacrylate ....197—210 COPOLVIMETS TO Li =e eennee 208 with vinylpyridines, monomer reactivity ratios in copolymeriABT ON AOE seAerie Beret aie ner geleys 201t

RESIN

CHEMISTRY

Ester(s) amine-modified epoxy resin ............ 67 chlorohydrin’ (23. 23s 213 GeCOMPpOsitionieetet eee ats 218f glycidyl Wi... a eae 211-224 2-hydroxy-3-chloropheny] ......... ae 221 1-Etheny]-4-(2,3-epoxy-1-propoxy fae oe : a EES, SOE 197-210 COPOlyanersixO Lees ye eee 205, 208 with vinylpyridines, monomer reacLEAL? DEMOS) sax doacnpennaeenaeosar scones 203¢ Ethers, cationic polymerization of cyclic ped «SR ee ean ee 19 9 -Bihythexanoics acid: eae en a 213 reaction of epichlorohydrin with 211—224 5-Ethyl-2-vinylpyridine .................... 197-210 F

Florescamine (Fluram)) ..............0....:. 230 Foaming, of themesinge ee eee 78-79 Fourier transform (FT) NMR............ 87 G

Gehman Torsional Stiffness Tester ... 138 Glass transition temperature 7 eee ee 118, 179, ie OEPOKY, NCSI Saye eeae Glycerol-1,3-di-2- ethylhexanoate cts 313 synthesis eT See EN: 218, 215f Glycidoly: Baeeaas Aaah ee ee 261 Glycidylated heterocycles ................ 115-135 Glycidyl STO

Seth

eae camara ee

211-294

2-ethylhexanoate, preparation of 211-224 methacrylate copolymers, acryloTHUY G5 cht eee ee PRU H

Haltester linkage) 2 eee 64 Halfesters of trimellitic anhydride (CENA Se ered oh Ores 64 Halides, quaternary ammonium .......... 48 Heterocycles, glycidylated .............. 115-135 Hexafluorophosphates benzenediazonin 9.0.4 een 8t p-chlorobenzenediazonium .............. 19 preparation of p-methoxybenzenediazonium te... as eee 41 Hexahydrophthalic anhydride

(HHP A),) raves 6. gp eee

18

Hexamethoxymethylmelamine CEM MM) testers ata tegs Rees

59

INDEX

269

Hydantoin “based epoxy resins et ccnscenc, intermolecular forces of ................

116 133

epoxy resin

chemical resistance of aromatic amime-cured

........cass....... 129¢, 131¢

DLOPELMesHOLCUTeC ae eae meen. 118 effect of alkyl substituents OM sie ed lean 115-135 room temperature amine curing

OR

etetn

tinct 1226, 123%, W24e LOTt

solvent/chemical resistance of aromatic amine-cured 128¢, 130t, 132¢

TL ag eee eR Biot nc bast ne 115-135 Hydrophilic resin, DMH-based .........._ 126 Hydrophobic—hydrophilic balance of amine-cured

resins ................... TOS) wea

nivdroxym diecsterare sere eee 2-Hydroxy-3-chloropheny] esters ........ Hydroxycitronellol ....... Rrra te. ecere Hydroxyl Complexes amine— eee wie

218 221 262 253

concentration on epoxy—amine reactions,

PROUD S me ern

effect of -...........00......

rh Sinead

239

225-246

p-Methoxybenzenediazonium hexafluorophosphate (PM) ..........00.0.... 22 PNOLGIMIEIACONS serrate yet) cern 81 LC DAUA UCL i 4] Methylene dianiline (MDA) ..137-154, 158 -cured epoxy networks, micro- and TMA CKOSS EU CEC SmOlena Sypoilenatn, LEGON CNS) en cuessnonssnee eons 1-Methyltetrahydrophthalic anhydride 49 Micra geloy, mntiee, cee ees 157-180 CORA ONC arcra scr cet er eee 172 Obs Dial y erie re ate mT et 175 GiESCCODCATY smear, oe enter tae igs ISO At OLNO Lae eer te tence me eee etter ee 160 Modulus of epoxy resins, rubbery ...... 142 Modulus of epoxy resins, Yonug’s ...... 142 5-Monoalkylhydantoins .....00.000.000...... 116 N

Near infrared spectroscopy (NIR) ..._ 103 UIE: TENS? GUBIOHIES, sccateseeaccoccnaseoe seane 100 Network(s) creep behavior of amine-cured DOK Vaart raat ren ee 183-195 LONMa LOMSETMNOG CEO Leeann

I

Initial deformation temperature (GUID Te ees ee eens. toes. ... 118, 134 limmgnllenatosn mab gernEMS. ones 47-56 Insulation properties of laminate, Glisaisnrom bk ve,aetoo ete: pe oma 19 K

ISLETS: TNOLONNIE

oe

ec ier

251-256

L

Laminate(s) electrical insulation properties of .. [DEOCESS ite trees ak Lat art. oh Pesdiient yee’ jOVROJOVSES. OH coc trcosctderecenventear ina _80f, T,, and pressure cooker test results 79, Wasenw EGSN Cae

ee eee meee

Laser light, back-scattering of ............ Low-molecular-weight polybutadienes

CDA

ok Se

ee

79 T1f 82f 81f 93

i morphology effect of molecular weight of the prepolymer on epoxy ............ 166 effect of stoichiometry on epoxy. 161 and the mechanical behavior of CDOKICS Beer: teenie 157-180 surface properties of the .................. 158 WOO) CITE OXON Srsecocroossoeksocerecoansat157-182 NIR (Near infrared spectroscopy) ...._ 103 CUTE sTALENSEUGICS arte tee nen cree 100 NMR (see Nuclear Magnetic Resonance) NOE (Nuclear Overhauser effect) ... 88 Nuclear magnetic resonance (NMR) CAT DODO! ay, tata meena tes are sea e 84 epoxide equivalent weight determination by carbon-13 .............. 83-90 IMouaatere ies IAM. (LEW) caccecornnoccer 87 OROUOM ESdecreas eerece Mente tare ore 84

Nuclear Overhauser effect (NOE)...

88

93

ae 57-69

BASeC DUNC CLS mene eee eee ee 57-69 TevOphODICILYy Maren edn tek mieenans. 126 M

Macrogel, formation of ................... Ts, a

MDA (Methylene dianiline) . 137-154, 158 Mechanical shift factors, dynamic ...... 190

O

Olefinic bonds, cleavage of .............. 260-261 Olefins ep oxidizeds ae ner et 35 Oligomerization of epoxy compounds, tertiary amine-catalyzed .......... 205, 208 Oligomierstr snes cee ee ee. oe 83-90 (Oya

GAMES su. cconaeecunnne osorastareunccer etooce 17, 19, 28

Onium salt photoinitiators 0.0.0.0... UV absorbance spectra of ................

20t 30f

EPOXY

270 P

IRamntse plemented i)1) meena ares Rermanganate, acidic ee

65

260, 261

Birenalicaminesi ae eae een ee 99-114 coating formulations with ................ 112¢ Guriheva cents PLOpeltlesme eee eens 102¢ Phenolicformaldehyde, epoxy resin/

EY) pin ehieels Is teil ae

nee 1-98

Phosphonium compounds electrical properties of cured epoxy-anhydride resins containing quaternary ................ latent accelerators for anhydride-

54t

cured epoxy resins, quater-

TAT Vor ee een eae are ee 47-56 NONNAlG ewer eee ee ee eee 53 qQuatemary esti centre ees 49 SEMICCUTESSOL ee eeenene eee 50t latent accelerators in epoxy-anhydride resins, gel time data for GuateI ahyaaa eeeeres Slt latent accelerators in epoxy-anhydride resins, viscosity (storage) Gata tonquatemary ee ee 52t Phosphorous pentafluoride .................. 21 Photoinitiated cationic polymerizaTLOMMOLe POXVALCSTIAS etter natate 1-15 Bhotomn ciaton (6) meee eee eee 2 for cationic polymerization of CDOXICES me meer rer teee tere rs 17 concentration on rate of cure, effect Ola et cn eee ere ee 225 for cross-linking of epoxides ............ 18¢ Ciany 0G ONTOS.

tsi

BS

ONLI ASal Ge ee eee eee eee 20t UV absorbance spectra of ............ 30f P.M eee ane eee ee re 31 EDS Ce eee ee eee ete ee 81 Prianyisninonnunitsa| ta ers 2-15 Photolysis of an aryldiazonium salt .... 1-16 Photopolymerizations .......:....:cs..:se0-0: 15 Photosensitized epoxides, storage Sta bilityaO Guetta eee 92, 24, 82 Photosensitized epoxy coatings, volaTU Ob geese ch tae sac adn eee 40-41 pK, values on the reaction of amines with epoxy, effect of ............c0000 249 Plastics industry, aqueous epoxy resins for electrical-reinforced ................ 77-82 PM (p-methoxybenzenediazonium hexafluorophosphate) .................. 22 Polvamides {ota nec ates an. era aes 101 BRolyeniines a5. e tous er tee 101 Polybutadienes, low-molecular WWIEAINE (TEIN MEVBYS)) ooo cect ceevne57-69 Polyepoxides, self-cross-linkable ....197-210

RESIN

CHEMISTRY

Polymerization of cyclic ethers, cationic .................. 19 of epoxides, photoinitiators for CallOnic® 2c.) aete-atse ee yy of epoxides, rate determining factors in the photoinitiated .......... 10 of epoxy resins, photoinitiated CatlOnie’s hee eee wee 1-15 Livi moss merge eae eee 33 Prepolymersy Epouye..2.s eee 138 BrepolymenswepOxyar eet lias Pressure: cookeritestq, ene 79 results, comparison of laminate Ae aXe ene debepreteen taminen 79, 81f Primer coatings, automotive ................ 73 formulation and performance properties for 2a.2 eee eee 74t Printed circuit board industry ............ 77-82 Pyridine. 232. are eee ee 205 Pyridinium chloride analysis .............. 228

Q QO-e schéme2 eee ee 198 Quaternary ATION |Val C CS 48 arsonium compounds .................0.+. 55 phosphonium compound(s) ............ 49 electrical properties of cured epoxy-anhydride resins COMtALIINSee eens eee ee 54t latent accelerators for anhydridecured epoxy resins ................ 47-56 SEV CCUTESTOLee sees eee eer 50t phosphonium latent accelerators in epoxy-anhydride resins, gel time datas {Oras eee Slt phosphonium latent accelerators in epoxy-anhydride resins, viscosity (storage) data for .......... 52t stibonium compounds ...................... 55 R

Relaxation-time spectrum .................... Resin(s) (see also Epoxy )

192

DMH-based hydrophilic .................. 126 Foaming. oben ee: 78-79 hydrophobic—hydrophilic balance of amine-cured)

5.

118, 121

hydrophobic TETA-cured ................ 121 latent catalysts for epoxy .................. 48 PLeparation see eee eee nee 159-160 properties of an ideal commanddestruct |,2:11 te. Seve eerah tae 259 rate of cure of DGEBA .................... 38t solvent-borne thermoset

INDEX

271

Rheovibron viscoelastometer .............. Ring closure reaction ..0.....0...0.0.0000. 219-224 Rubbery modulus of epoxy resins ...... 142 S

Salt-spray resistance ...........0.0.ccccsee 59, 74 of cationic ED coatings 00.00.0000... 69 DaAlb-SDERY: LEstie@ Mana haik tixiicauenen 59 SA (Succinic anhydride) .................... 61 Scanning electron microscope (SEM) 160 Scattering of laser light, back- ............ 93 Schieman reaction (0. c.s.dccie csc 21 SEM (Scanning electron microscope) 160 Sodiumrcarbouatemse a ec 219 SOdimyhyaroxic omy wi can eee ees 219 Solvent-bormestesiny w... aces ee 72 Spectroscopy, near infrared (NIR) ... 103 Sterilization resistance (blushing ) of CaMICOAUIMES ae eaee ee eee 91-98 Stibonium compounds, WE. wo StresserclaxatiOn. me mans hah ees 183-195 Styrene oxide, photopolymerization of 8, 4 Succinic anhydride (SA) Giullicoromiireny GAMES os Acncosaccusenceie coco 98-39 ap

Tack-free time ......... set pene nt aH 81 device for measuring ..........0......0.... 30t PETC ORIC Cate en ee Se eo ahoacls 36t PPACKSL GSU amen, eet een Creare nee 31 TBA (Tri-N-butylamine) ............... 228, 232 TEM (Transmission electron TNNIGLOSCOD GC)mare tesserae sees 160 Tensile behavior of epoxy resins ........ 145 TETA (Triethylenetetramine) .......... 118 T, and pressure cooker test results, comparison of laminate .............. 79, 81f Thermoplastic-modified epoxy resin EI SLOT per ee ee ee Hot phen OSCteLeSUUs mrt tee eee i 157-182 SLO NIGTICIMEM fee fine eee cece ase 211-224 Torsional Stiffness Tester, Gehman

....

138

TPR (Triphenylsulfonium hexaHuorophosphate)) eevee eee 28, 29 flimamise DOxiC atl OT Mewes ete arnt te neo 221) TOACLOM erie eee oes ossstces Transmission electron microscope

(ES CAN

ster

ee ee

eee

eee

222F

160

Triarysulfonium halides ......0.0...00........ 3 Tmarysultomiunm salt(s) %4..4..c00., 2,18 photomittiators yayoeai-a.ceee 2-15 DEE DATEL OURO Meets ae eee 4t—5t Triethylenetetramine (TETA)) .......... 118 -cured resins, hydrophobic .............. PAL Trimellitic anhydride (TMA) halfCSHOSS COL eae | alee teers 64 Tri-N-butylamine (TBA) .......0.......... 228 Triphenylsulfonium, hexafluoroATSEMALC gece eerie ean en eee cee 3 Triphenylsulfonium hexafluorophosphate CUPS nr rern, e 28, 29 HOtOM IEA LONS eee 31 U

Uirany lenitiatc iene heen eee 55 Wirethanes mee pe eats ce ere aes 261 UV-cure of epoxy resins ............c0000 1-16 V

Varnish formula, organic solvent ........ 78f Varnish formula, water-based ............ 78f Varnishes, problems with waterDaSCC gierey wet te te ayre eee cere 78-79 WAL AyeNEKENTNE(S)) —csocoscicon ccovcrononnacheosbaste 198 Q-e values in copolymerization of 1-etheny]-4-( 2,3-epoxy-1propoxy )benzene with .............. 203t Q-e values in copolymerization of 2,3-epoxy-1-propyl methEXC ENS) AMES, fe ecnssosanatencooarete 201t 4-Vinylcyclohexene dioxide ................ 3 GEN tay lonMaGWhnES) .o-sccssosccnoonsmascseeonsebsxe197-210 ANany pyridine sierwm senses eee 197-210 Viscoelastic behavior ...............0-+ 137-156 Viscoelastic parameters ...................: 184-192 Viscoelastometer, Rheovibron ............ 138 W

Water-based epoxy resins ...............00 77-82 WLFE ( Williams-Landel-Ferry ) (ELC[LENATO}OV:for soccctsonsndcoumeherenbase Geceaen 183, 187 Ys

Young’s modulus of epoxy resins ........

142

“cor

(i

—

es

er 7

Facet (245 77

LOGS] ue

6

Jeary) peas qultenalay Sm smehy

er

ae

le

7

‘

iNe) hed olen cet 7 ake

‘ht wigs

=

ef ey] A

i

eel

Bsj

ei

;

MT)