Professional Documents
Culture Documents
The Royal Society
The Royal Society
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
The Royal Society is collaborating with JSTOR to digitize, preserve and extend access to Philosophical
Transactions of the Royal Society of London. Series A, Mathematical and Physical Sciences.
http://www.jstor.org
Interfacein compositematerials
[Plate 1]
INTRODUCTION
It is a well known fact that unidirectional composites reinforcedwith high modulus fibres
exhibitexcellentcharacteristicsin thedirectionof the reinforcement.However, theirweak point
lies in theirmoderate shear strength.This may be aggravated by the decrease in strengthof the
composite during weathering,which renders the problem of the coupling between fibre and
matrix even more intricate.
Interfacial stressesin compositesare due to the shrinkingof the resin and the differencein
elasticitymodulus between matrix and reinforcingagent. The stresstransferat the interface
thereforerequires an efficientcoupling between fibreand matrix. It is, however,importantto
optimize the interfacialbonding, since a direct linkage between fibreand matrixgives rise to
rigid, low impact resistancematerial.
Among the many factorsthat govern the characteristicsof composites involving a fibrous
material and a macromolecular matrix,it appears certain that the adhesion between fibreand
matrixplays a predominantpart. Although the differentauthorsdo not always agree as to the
nature, the intensityand the properties of the interfacial bonding (physical or chemical,
reversibleor irreversible)it seems clear that one of the main criteria is the absence of inter-
facial defects,i.e. the existenceof a smooth interfacialadhesion.
Many models have been suggested,but only thosewith a directbearing on compositeswill be
dealt with in the firstpart. As to the influenceof surfacechemical functionsin carbon fibres-
epoxy resin composites, attention will be paid to them in the second part.
[ 87 ]
Mechanicalbonding
This model consistsof a linkage phenomenon, a mechanical anchoringof the polymerinto
the pores and the unevennessof the fibre,afterhardeningof the polymer.The interpenetration
of two solids is, in fact,a mechanical factorwhich promotesthe adherence between two bodies.
Wake (I959) has shown that the adherence to fabricsbased on continuousbut variouslyshaped
yarns is proportional to the surface area of the fabrics. However, the rugosityis a positive
factoronly in so far as the substrateis perfectlywetted by the liquid. As a matterof fact,if the
liquid cannot penetrateinto the asperitiesof the substrate,the hardeningof the resinis accom-
panied by the formationof interfacialcavities which are liable to initiate the failure of the
interfacialbond.
In this model, the adherence and wettingcriteriaare thereforecloselyrelated. Thus, in the
case of a good wetting,an increase in the specificsurfacearea of graphite fibres,achieved by
nitric oxidation improves the interlaminarstrength(i.l.s.s.) of the correspondingcomposites
(table 1) (Donnet et al. 1974) .
TABLE 1
fibre(typeI) SB.E T. (fibre)
/(mlg-1) composite)
i.l.s.s.(carbonfibre-epoxy /MPa
non-treated 0.4 22
nitricacidtreated 1.4 43
It is also worthnoting that pittingthe surfaceof carbon fibresby oxidation in air at 870 K
significantlyincreases the shear strengthof the resultingthermosettingcomposites (Clark et al.
I 974) -
Physicalcoupling
In this model, the forcesof interfacialbonding are so-called secondary or van der Waals
forces. They include mostly London dispersion forces, dipolar interactions and hydrogen
bonding. Their energyis usually comprised between 8 and 16 kJ/mol.
Generally speaking, these forces are involved in the wetting and adhesion phenomena
(Zisman I963). In the case of composite formation,this model stipulatesthat strongadhesion
may only be achieved ifthe materialsfacingeach other are in close contact. In thisconnection,
it is worth noting that the shear strengthhas been related to the wettingangle between resin
and carbon fibre(Yamamoto et al. 1971).
Chemicalcoupling
In the chemical bonding model, the adsorption of chemical groups and the formationof
true covalent bonds between fibreand matrix are assumed. The forcesof interfacialbonding
are, in this case 'primary forces' of high energy (40-400 kJ/mol)i.e. covalent, ionic or metal
bonds. Chemical groups that give rise to such bonding are seldom available on a surface,and
the aim of surfacetreatmentsis to insertthem at the interface.
For glass fibresthe surfaceis treatedwithcoupling agents,mainlyof the silane type,described
among others by Plueddemann (I 970). The bonding between fibre and resin involves the
[ 88 ]
formationof hydrogenbonds between the silanol groups of the fibreand the partlyhydrolysed
silanes according to the scheme shown in figure 1, where R is a group compatible with the
resin.
An overall interpretationof the characteristicsof the coupling agents according to the theory
of the chemical bond is however not completely satisfactory.Other theorieshave been pro-
posed, more particularlyan enhanced ageing resistanceof the compositein a humid atmosphere
(Outwater & Murphy I970).
R resin
- Si-
0-
-o
Si
glass fibre
FIGURE 1. Schematicdiagramofsilane fixationon a glassfibre.
C-OH + CH2-CH-CHj-C1
_'.~0'
C-O-CH2-CH-CHj-C1
OH
The model of the chemical bonding, however,is only an extensionof the physical coupling
as suggestedby Kaelble (I97I). Both are dependent on a good contact and hence upon a good
wettingbetween fibreand matrix. Only the bonding energymay vary fromone model to the
other even thoughthe broad scatteredspectrumof the various possible bonds (hydrogen,ionic,
covalent bonds) renderstheiridentificationdiicult.
[ 89 ]
Flexible
interlayer
The insertionof a non-rigidinterlayerbetween the fibreand the matrixconstitutesanother
practical means of ensuringa good stresstransfer.Broutman& Agarwal (I 973) calculated that
although the strengthof the composite varies little with the modulus of the interlayer,the
ruptureenergyofthe compositepasses througha maximumfora modulus lower than thatof the
resin used. It appears, therefore,that coating the fibreswith a polymer provides a means of
increasing the impact strengthof the composite (Fila et al. I972; Hancox & Wells I977).
A simple process to establish such a non-rigidlayer and ensure a good adherence between
fibre and matrix has been employed by Riess et al. (1974). They modifiedcarbon fibresby
graftinga copolymer bearing flexiblesegments (polyisoprene,PI) and segmentscompatible
with the matrix(styrenemaleic anhydridecopolymer,SMA). The bonding between the carbon
fibreand the polyisoprenesequence consistsof dipole-dipole interactionsresultingfromsalified
carboxylic groups, as shown in figure3.
COONa
NaOOC
A - B C :| D
TABLE 2t
carbonfibre-epoxy
composites
The 'Kevlar type' organic fibreshave likewise been polymer treated with positive results.
According to Wake (1972), the fibresurfacemay react with an isocyanate linked polymeras
indicated in figure4.
Using polyurethane treatment,Riess & Reeb (1977) significantlyimproved the torsion
of Kevlar-epoxy composites.
stiffness
Beforeconcluding thissection,it is worthnotingthat the model of the flexibleinterlayerhas
also been suggestedto account forthe behaviour of silane coupling agents in glass-fibrebased
composites.In fact,Bascom (1972) reportsthat afterhydrolysis,the silanes may have an effect
on resin polymerizationand hence on its physico-chemicalpropertiesin the vicinityof the
[ 90 ]
\\ \-C-NH\ \ \ + NCO
11
C=O
I
NH
Anodic
oxidation
ofthefibres
In order to investigatethe effectof surfacegroups on the fibre-resinbonding, we developed
surfacetreatmentsthat alter the nature and the number of surfacegroups withoutaffectingthe
surface area. Previous studies showed that shorttime anodic oxidations in various electrolytes
(diluted nitricacid, alkaline medium) permit the achievement of such conditions (Ehrburger
et al. I 974, I 975) .
Experimentshave been carried out with carbon fibre,ex-PAN fibre (AC fibre,type II). The
acidic groups were of two types: the strongerones (carboxylic and some phenolic groups) were
estimated with sodium hydroxide,whereas sodium ethanolate was used for the weaker ones
(hydroxylicgroups). In the latter case, all the acidic functionson the carbon surface were
neutralized (Scholz & Boehm I970). The data in table 3 reveal that the number of functions
TABLE 3
acidic surfacegroupsneutralizedwith
that may be measured with sodium ethanolate is similar whateverthe type of electrolyteused
forthe anodic etching (HNO3 or NaOH). Yet, the treatmentof the fibresin alkaline medium,
AC-NaOH, yields comparativelystrongeracidic groups since theymay all be neutralizedwith
sodium hydroxide.The uncertaintyof the measurementsis of + 2 jiequiv/g forAC-HNO3 and
AC-NaOH fibres.
of modified
Bondingproperties carbon
fibres
Unidirectional composites have been formedby using treated carbon fibresand an epoxy
resin (Ehrburger et al. 1974). The interlaminar shear strengthof the composites has been
measured with the cantilever test as previouslydescribed (Donnet et al. I974). The mode of
failure of the composite in this test has been determined by analysing the load-deflexion
diagram. In fact, the shape of this diagram reveals the nature of the bonding between the
componentsof the material (Hanna & SteinsigerI969). Two extremecases may occur.
The interfacialbonding is weak; in that case 'pull out' is responsibleforthe rupture.As the
fibresare not matrixbound, theyslide in theirbeddings and load transfercan no longer take
place (shear failure).
The interfacialbondingis strong;thenthe matrixtransfers the load to the fibresuntilreaching
of the breakingpoint. In thiscase, the failureis sudden and catastrophic (tensile failure).
Most cases of failurelie between theseextremes,the relativeimportanceofwhich depends on
the interfacialbonding.
The load-deflexiondiagram of compositescomprisingAC fibresare shownin figure5.
1.5
1.0
0 o.4 0.$
I/mm
deflexion,
FIGURE 5. Load-deflexiondiagramsforcarbonfibres-epoxy
composites:curve 1, non-treatedAC fibres;
2, AC-HNO3; 3, AC-NaOH.
tD- s~~~~~~~~~~~~~~~~~~~~~.A
(;Facing p. 500)
TABLE 4
carbonfibre-epoxycomposites
In order to relate the propertiesof the compositesto thoseof the carbon fibres,the number of
acidic groups was varied at the carbon surfacethroughpyrolysisat 1273 K for 3 hours, in an
inert atmosphere.The data are included in table 5.
TABLE 5
acidic surfacegroupsneutralizedwith carbonfibre-epoxy
composite
It is seen that, afterpyrolysis,the strongacidic groups that may be neutralized with sodium
hydroxide are destroyed,whereas the weak acidic groups are only partlyremoved. Moreover,
the nature of the acidic functionsdetermineslargely the failuremode of the composite in the
Rsheartest. The presence of strongacidic groups resultsin a tensilefailure mode, i.e. a strong
interfacialbonding. On the other hand, the failurein shear would rathercorrespondto weak
acidic groups.
In figure8, the i.l.s.s. of the material has been plotted against the total amount of acidic
groups. A clear correlation indeed appears between the content of acidic functionsand the
ultimate value of the shear strength.
10
C1 FW I
/
_6 - /
.{/
4/
10 20
acidic groups/(jiequivg-1)
FIGURE 8. Interlaminarshear strengthof the compositeplottedagainstthe amountof acidic group
neutralizedby sodiumethanolate.
[ 93 ]
DISCUSSION
Discussion
A. R. UBBELOHDE, F.R.S. (Department of ChemicalEngineeringand ChemicalTechnology, Imperial
College,London,SW7, U.K.). When certainsurfacetreatmentsare used to 'improve' the contact
between carbon fibresand various matrices,it must be rememberedthat such fibresare often
quite well graphitized. Various formsof anodic treatmentin aqueous electrolytescontaining
(say) OH' or NO' may actually promote formationof intercalation compounds of graphite.
Even if mostof the intrudinganions are subsequentlyremoved,e.g. by hydrolysis,intercalation
would probably involve penetrationof layersin some depth, and leave a fluffier interfaceafter
subsequent treatmentbeforecontact with the resin,than if surfaceimpuritieshad been merely
removed by high temperatureoxidation, or ion bombardment. Titration of the foreignanions
with Boehm's methods may not reveal all thosepresentin the formofintercalationcompounds,
near the surfaceof the carbon fibre,because 'residue compounds' are formed.
Fibres treated anodically should thus be considered rathercritically,since some anions may
actually produce a fluffy interfaceextendinginwards to some depth. Since a great variety of
anions can be used foranodic intercalation,thisprocedure can be varied to check how far the
interfacehas been modifiedin depth, by varyingthe anion.
J. W. JOHNSON(Rolls-Royce
Limited,AeroDivision,Non-MetallicsLaboratories,
PlasticsandComposite
Materials Department, AlfretonRoad, Derby, U.K.). ProfessorUbbelohde's observations on the
electrolyticformationof intercalationcompounds in carbon fibresare very relevant and have
great significanceforcarbon fibrecompositetechnology.Mild anodic electrolysisof carbon fibre
in a varietyof aqueous electrolytesoftenproduces thin film interferencecolours on the fibre
surface; extended electrolysiswill ultimatelyrender the fibre completely transparentso that
it can be examined in the light microscope. The electrolysisis most easily accomplished with
type A or type III fibres,although the swelling accompanying the intercalation bursts the
fibresalong theirlengthsand the propertiesare irretrievablyruined.
Types I and II fibres,ratheroddly, are more resistantto the process but are still susceptible
to a large extent.Orientationis mantained in the swollen state as shown by birefringence,and
details of fibremorphologyare preserved (see figure9).
[ 95 ]
32 Vol. 294. A
! FM
[ 97 ] ~~~~~~~~~32-2