Indian Journal of Engineering & Materials SciencesVol. 4, December 1997, pp.266-270

Carbon fibre/epoxy composites: Effect of epoxy network andsurface treatment of fibres on interfacial shear strength

D S Varma", Sneh Saxena", Neeraj Guptar& I K Varma'"

'Department of Textile Technology, "Centre for Polymer Science and EngineeringIndian Institute of Technology, Hauz Khas, New Delhi 110016, India

Received 18 September 1996; accepted 17 March 1997

The paper deals with studies on interfacial shear strength (T ) of carbon fibre composite using singlefibre fracture test. Surface modification of desized carbon fibres has been done by argon or oxygenplasma. The effect of network structure on T has been studied by using difunctional or multifunctionalepoxy resins. Interfacial shear strength depends on (i) chemical structure of the epoxy resin (ii) nature ofgas used in plasma generation and (iii) duration of plasma treatment. An increase in cross-link densityresulted in a decrease in T, thereby indicating that a more flexible interface is superior for betteradhesion with fibres. Oxygen plasma treated carbon fibres gave highest values of r.

Carbon fibre reinforced epoxy resins are beingextensively used for aerospace and defenceapplications primarily because of their highstrength-to-weight ratio, superior performance andexcellent mechanical properties. The majorrequirement in ensuring durability and stability ofsuch composites is the ability to bond together thetwo constituent materials. The transmission ofstress between fibre and matrix depends on astrong interfacial bond which resists failure. Forthis reason the degree of contact and the cohesiveforces at the interface are of considerableimportance.

In fibrous composites the interface or interfacialzone comprising of near surface layers of fibresand matrix and any other layers of materialexisting between these surfaces is a controllingfactor in obtaining optimum mechanical properties.The properties that are influenced by the interfaceinclude composite strength, Young's modulus,intcrlammar shear strength, bending stiffness andcompressive strength. The mode of compositefailure also depends upon interfacial bonding. Agood interfacial bond leads to sharp well definedbreak while poor interfacial bonding leads tomatrix failure followed by fibre failure. The failuremechanism is an important problem in applicationsas diverse as aerospace and defence vehicles,

• For correspondence

dental devices and microelectronic circuits.Good adhesion between fibre and matrix can be

achieved by (i) surface treatment of fibres and(ii) manipulation of adhesive composition. Surfacetreatment of carbon fibres for improvinginterlaminar shear strength of composites havebeen extensively reported in literature '.3.Controlled surface modification of the fibresurface by plasma treatment has also beenexamined", Generally, adhesion and weatherabilityis improved by plasma treatment':". Theeffectiveness of a given plasma treatment dependson a large number of process parameters such aspower, pressure, gas flow rate and treatmentgas/substrate material combination. Plasma ofacrylonitrile, styrene, ethylene and styrene/airmixed gases has been recently used formodification ·of graphite fibre surfaces". Oxygenplcsma can cause extensive surface oxidationleading to the formation of carboxyl, carbonyl andhydroxyl groups on the surface. In the inertatmosphere (Ar. He) the plasma treatment mayaffect the fibre surface by ablation of the lowmolecular weight compounds". Oxidation may alsotake place subsequently when these fibres areexposed to oxygen.

In the present studies the surface modificationof carbon fibre was done by plasma treatment inargon and oxygen atmosphere. The effect ofstructure of epoxy network on interfacial shear

VARMA et al.: CARBON FIBRE COMPOSITES

stress was evaluated using a difunctional and atetrafunctional epoxy resin.

In order to characterise the stress transfer atmatrix-fibre interface, single fibre was embeddedin the matrix and the specimen was stretched alongthe fibre axis until fibre breakage reaches asaturation level. The average fibre length (lt~) isthen related to interfacial shear strength ('t)according to relationship"

CYfdT=--

u;where d is the fibre diameter and af is the fibretensile strength. Since the fibre matrix interface isplaced under shear, therefore, the calculated valueof t is expected to be an excellent estimate of theshear strength between fibre and matrix"!',

Experimental ProcedureTenax. J carbon fibres (grade HM-35A-6000)

having a tensile strength of 3000 MN/m2 anddiameter 6.5 urn were obtained from Toho RayonCo. Ltd., Japan. Sizing of the fibres was removedby overnight soxhlet extraction with acetone. Thetensile strength of single fibres (de sized andplasma treated) (gauge length = 1 em) wasevaluated by Instron tensile testing machine modelNo. 1112. A cross-head speed of 1 cm/min and achart speed of 10 cm/min was used. More thantwenty fibres were tested and average tensilestrength was determined.

In order to study the effect of epoxy resin on theinterfacial shear strength, two epoxy resins wereused, i.e., diglycidylether of bisphenol-A(Araldite-Cibatul Ltd.) and a tetrafunctional epoxy,MY 720 (Ciba Geigy). Cured resins have beendesignated as A and MY respectively in thesubsequent discussion. Curing of epoxy resins wasdone by using a stoichiometric amount of 4,4'-diaminodiphenyl methane. The amine was purifiedby crystallisation from water.

Equipment for plasma treatment-Plasmatreatment of carbon fibres was carried out in a 25crrr diameter plasma chamber. A glow dischargeplasma was generated by applying potentialbetween the two disc shaped aluminium electrodesfacing each other. Argon or oxygen plasma wasobtained by continuously supplying either of thesegases. Working pressure of 10-1 torr wasmaintained throughout the experiment. A bundle of

267

unsized carbon fibres were placed axially betweenthe electrodes by means of glass support suspendedfrom the chamber side wall. In order to ensureclean working environment, the chamber wasflushed with the corresponding gases for 30 minprior to starting the fibre treatment. Typically 40-50 mA current was recorded in argon/oxygenplasma for a discharge voltage of 540 V. Carbonfibres were treated with argon plasma for 10 minand 30 min and with oxygen plasma for 30 min.

Fabrication of steel and rubber mould-Mildsteel mould having eight dogbone shapedspecimen was machined according to the ASTMstandards. RTV silicone rubber (Metro Arch) wasused for transfer moulding and to obtain 7 em longdogbone specimen cavities having a width, depthand gauge length of 0.3, 0.15 and 2.6 em .

Fabrication of composite specimens-Singlefilaments were picked from the fibre bundle andwere mounted with the help of a sticking tape on apaper cut tensile chamber and then placed onto thesilicone rubber mould. The mixture containing theepoxy and the amine was thoroughly compoundedat 80°C and degased prior to pouring in the mould.The assembly was then transferred to an oven andmaintained at 50°C overnight and 80°C for 1 h.After cool down in the oven, the mould was thencurled away from the specimen parallel to the fibreto prevent fibre damage.

Tensile and photo-elastic measurements-Theepoxy tensile coupons having a single carbon fibreat the centre were tested on tensile testing machine(Polymer Laboratories-Minimat) which wasmounted on a Leitz polarizing microscope having aWild MPS45 camera facility at the top. The carbonfibre encapsulated coupons were loaded in tensilem?de. The changes were monitored by polarisedmicroscope.

Results and DiscussionEffect of plasma treatment on fibres-The SEM

photographs of unsized carbon fibres and plasmatreated fibres are shown in Fig. 1. No significantdifference in the structure is observed on treatmentwith argon plasma. In oxygen plasma, on the otherhand, certain amount of longitudinal ridges andvalleys were present on the surface.

Critical length and interfacial shear strengthstudies-The stresses generated within the fibre-

268 INDIAN 1. ENG. MATER. SCI., DECEMBER 1997

,It

Fig.l-SEM photographs of (a) desized carbon fibrc (b)Argon-plasma treated fibre (J 0 min) (c) Argon-plasma treatedfibre (30 min) (d) Oxygen-plasma treated fibre (30 min)

Fig.2-Polarized transmitted light micrographs of carbon fibrein MY resin (x I00)

matriX interface observed in each specimen wererecolcied photographically with increasing strain.After a break an intense photo-elastic regionappears around the end of the fibre (Fig 2). Withincreasing strain, this region rapidly expands downthe fragment away from the break. This results in 8

stress pattern that has alternating light and darkareas in the micrograph, thereby leading to anincrease in the gap between fragments withincreasing strain. Coupons with plasma treatedfibres showed that fragment separation was

Fig.3-Polarized transmitted light micrographs of carbon fibretreated in argon plasma for 30 TllIn inA resin (x 100)

reduced by the presence of well bonded fibre'Lower values of critical lengths were observecwith treated fibres thereby showing an increase inadhesion of carbon fibre with epoxy matrix.

Carbon fibres treated with argon and oxygenplasma for 30 min were also used for fabrication ofsingle fibre composites. Such a treatment of carbon

fibre resulted in a decrease in critical length andincrease in interlaminar shear strength (Figs 3 and4).

In Table 1, the results of interfacial shear

VARMA et al.: CARBON FIBRE COMPOSITES

·----- -- - ---,----- .~-- ----.

269

Resin

A

MY

Table I-Results of interfacial shear strength of carbon fibre epoxy composites

Plasma treatment of C fibre

IjdO"ft

AtmosphereTime (min) (Mpa)(MPa)

41.7

188622.6

Ar

1034.5223332.3Ar

3029.9198533.2

O2

3022.7169737.4

79.7

188611.8

Ar

1067.1223316.6

Ar

3046.219852\.5

O2

3032.1169726.4

Fig.4-Po1arized transmitted light micrographs of carbon fibretreated in oxygen plasma for 30 min in A resin (x250)

strength in these single fibre epoxy composites aregiven. The critical length decreased with the argonor oxygen plasma treatment of the fibres. This isexpected because the surface impurities removedby plasma treatment and functional groups aregenerated in oxygen plasma.

It is also obvious from these results that the

chemical structure of the matrix resin (A or MY)influences the shear strength. The 'A' based epoxyresin composites had almost twice the shearstrength than the MY composites when unsized orargon plasma treated (10 min) carbon fibres wereused. The multifunctional epoxies yield a highlycross-linked matrix resin which may not bedeformable to that extent at interface as the less

cross-linked A resin is expected to be. In otherwords, a deformable interface may be a betterapproach in getting higher interfacial shearstrength.

Another explanation which can account forthese results is the shrinkage of matrix resins. A

highly cross-linked resin is expected to undergomore shrinkage on cooling. Residual stress is thusdeveloped at the interface due to differentialshrinkage of fibres and matrix 12.

Increase in duration of argon plasma treatmentof carbon fibres only marginally affects the

fibre/resin A composites by increasing the durationof plasma treatment from 10 to 30 min, in case ofMY resin composites, a 30% improvement in "'C

values was observed. In argon plasma only surfacecleaning (ablation of adsorbed species) is possible.The polar groups present on the surface of thefibres thus become more accessable to the polar

groups of the matrix and thus improvement ininterfacial shear strength was observed. Theconcentration of polar hyroxyl groups which aregenerated in the matrix resin due to opening ofoxirane rings of the epoxy resins is more intetrafunctional MY resin as compared todifunctional A resin and this may be responsiblefor the observed increase in "'C values. Oxygen

plasma on the other hand generates polar groupson the surface of carbon fibres. A better interaction

of epoxy and polar groups of fibre surface isresponsible for the observed high strength values.

AcknowledgementThe Department of Science & Technology,

Government of India and National Institute of

Standards & Technology, USA are gratefullyacknowledged for sponsoring the project andproviding fellowships. Thanks are due to Prof. R.P. Dahiya of Centre of Energy Studies, lIT Delhifor providing facility 'for plasma treatment.

ReferencesI Pinchin 0 J & Woodhams R T, J Mater Sci, 9 (1974) 300.

2 Kalnin I L, US Pat 3, 723, 607, March 1973 cf. Kalnin I L

& Jager H "Carbon fibre and their composites", Ed Fitzer,

E (Springer Verlag, New York) (1983) 63.3 Asai H, Nako F & Yoneyama H, Jpn Kokai Tokkyo Koho

jp, 61 194 271 (1986); cf Chern Abstr, 105 (1986)2100210v.

270 INDIAN J. ENG. MATER. SCI., DECEMBER 1997

4 Drazl L T, Rich M & Koenig M, J Adhesion, 18 (1985)49.

5 Wu S, Polymer Interface and Adhesion, (Marcel DekkerNew York), 1982,298.

6 Momose Y, Oaku T, Chuma H, Okazaki T, Masui M &Takeuchi M, Am Chem Soc Ploym Mater Sci Eng, 62(1990) 353.

7 Sohn D W & Sung N, Am Chem Soc Ploym Mater SciEng, 62 (1990) 43.

8 Kelly A & Tyson W R, J Mech Phys Solids, 13 (1965)

329.9 Drazl L T, Rich M J & Lloyd P J, Adhesion, 16 (1983) l.

10 Netravali A N, Henstenburg R B, Phoenix S L & SchwartzP, Composites, 10(4) (1989) 226.

II Wimolklatisak A S & Bell J P, Polym Comp, 10(3) (1989)162.

12 Aggarwal B D & Broutman L J, Analysis and performanceof fibre composites, (John Wiley & Sons New York),1980, 33-34.