Page 209 - Engineered Interfaces in Fiber Reinforced Composites
P. 209

Chapter 5. Surface  treatments of $hers  and e8ecr.s on composite properties   191

                additional bonding by developing weak adhesion mechanisms at the interface, such
                as dispersion force, dipolar interaction and hydrogen bonds.
                  In  the  light  of  the  foregoing  experimental  evidence,  the  following  can  be
                summarized  regarding the effects of carbon fiber surface treatments, depending on
                the methods and media employed:
                (1)  There are substantial changes in fiber surface area with associated variations in
                    rugosity depending on the oxidative treatment  medium.
                (2) A  weak  surface  layer  may  be  removed,  the  removal  being  more  serious  in
                    plasma etching than in wet oxidation.
                (3) There is an increase in the polar surface energy.
                (4)  Chemical modification takes place and carboxyl, hydroxyl and carbonyl groups
                    are produced on the fiber surface.
                  The mechanisms of chemical bonding due to the presence of functional groups
                have yet to be more thoroughly clarified, which change the energetics of the carbon
                fiber  surface  considerably.  Better  mechanical  anchoring  arising  from  the  surface
                rugosity and the increased physical surface area involved in adhesion as well as the
                beneficial effect of  removing the surface weak layers all contribute to an improved
                interfacial  bond.  However,  there  must  be  a  limit  to  the  improvement  in  bond
                strength  by  fiber  surface  treatment  only.  A  decrease in  bond  strength  and other
                deteriorating  effects  are expected  to  occur  if  the  surface  treatment  is  excessive,
                leading to severe damage of the fiber.

                5.3.2.3.  EJrects of surface  treatment on composite properties
                  The interlaminar shear, flexural and tensile strengths are increased as the principal
                effects of carbon fiber surface treatment on composite properties. The enhancement
                of  these  strength  properties  depend  on  the  fiber  elastic  modulus,  the  degree  of
                surface  treatment  and  the  type  of  resin  and  curing  agent  used.  The  largest
                improvement in ILSS is obtained for high modulus fibers. The compressive strength
                is also increased slightly (Norita et al., 1986), and the mode I interlaminar fracture
                toughness  GI,  for  crack  initiation  is  almost  doubled  (Ivens  et  al.,  1991) with
                increasing  degree  of  treatment.  In  general,  an  increase  in  the  interfacial  bond
                strength, q,, enhances the composite compressive strength by augmenting the load
                required to cause the interface to fail in transverse tension due to the fiber Poisson
                effect. Delamination is reduced in favor of microbuckling of surface treated fibers
                (Drzal and Madhukar, 1993). The improvement in interface  bond  strength, zb,  or
                fracture  toughness,  GI,,  due to fiber surface treatments  has been confirmed using
                microcomposite  tests, e.g. the fragmentation  tests (Drzal et  al.,  1983a, b) and the
                fiber pull-out tests (Baillie and Bader,  1991). However, all these beneficial effects of
                improved  strength properties  are inevitably  accompanied  by  a  loss in  the impact
                fracture toughness of unidirectional laminates or notched tensile strength of angle-
                ply laminates, as shown in Fig. 5.15 (Goan et al., 1973; Dauksys, 1973). Therefore, a
                careful balance  has to be sought to ensure both  adequate strength and toughness
                properties.
                  More recently Drzal and coworkers (Madhukar  and Drzal,  1991a, b,  1992a, b;
                Drzal and Madhukar, 1993) have spent significant research efforts to establish the
   204   205   206   207   208   209   210   211   212   213   214