Interfacial Characteristics and Fracture Toughness of Electrolytically Ni-Plated Carbon Fiber-Reinforced Phenolic Resin Matrix Composites

Journal of Colloid and Interface Science 237, 91–97 (2001) doi:10.1006/jcis.2001.7441, available online at http://www.idealibrary.com on

Interfacial Characteristics and Fracture Toughness of Electrolytically Ni-Plated Carbon Fiber-Reinforced Phenolic Resin Matrix Composites Soo-Jin Park1 and Yu-Sin Jang Advanced Materials Division, Korea Research Institute of Chemical Technology, P.O. Box 107, Yusong, Taejon 305-600, Korea (South) Received September 18, 2000; accepted January 22, 2001

between the fibers and the matrices cannot be achieved without intimate contact—that is, unless the fiber surface contacts the resin on an intermolecular equilibrium distance level (13–16). A large interfacial area of intimate contact by the matrix resin is a prerequisite whenever the interfacial bond is due primarily to van der Waals physical adsorption force. In addition, complete wetting of the fibers by the liquid polymeric matrix is advantageous when the surface free energy of the fiber is well above that of the liquid polymer in view of the London dispersive-specific polar components of surface free energy. Untreated commercial carbon fibers have a surface tension close to 40 mJ m−2 and an extremely hydrophobic nature due to the extremely high temperature of the manufacturing process leading to carbonization (700–1500◦ C) or to graphitization (above 2500◦ C). On the other hand, most polymeric matrix resins, such as phenolic resin, are slightly hydrophilic with a surface tension in the range of 35–45 mJ m−2 (13–16). Electroplating has been used to produce metal matrix composites reinforced with carbon fibers (17, 18). Carbon fiber surfaces are metallized by electrolysis in molten salt solutions (19). The adhesion of electroplated metal/carbon fiber composites has been shown to be excellent where the metal coating is grown from the carbon fiber surface (20). On the other hand, polymer/metal interaction is strong, mainly due to the higher metal surface energy that allows extensive wetting by the polymer (21). The objective of this study is to evaluate the influences of nickel-plating on the surface energy of PAN-based carbon fibers. The surfaces of uncoated and coated fibers are also analyzed by XPS to determine the elemental compositional changes of the fiber surface. The relationship between the fracture toughness of the material and the interfacial adhesion in the composites has been investigated.

The electrolytic plating of metallic nickel on a carbon fiber surface has been carried out in order to improve the interfacial adhesion and the mechanical properties in carbon fiber/phenolic matrix composite systems. The surface and the mechanical interfacial properties of composites are characterized by X-ray photoelectron spectrometry (XPS), surface free energy, and the critical stress intensity factor (K IC ). From the experimental results, it is clearly revealed that the oxygen functional groups and the metallic nickel on fibers largely affect the mechanical interfacial behavior of the composites, resulting in increased surface polarity, whereas the nitrogen functional groups have no effect. Also, a good correlation between surface oxygen functional groups and mechanical interfacial properties and between wettability and K IC is established and it is found that a 10 A m−2 current density is the optimum condition for this system. ° C 2001

Academic Press

Key Words: carbon fiber; Ni-plating; surface free energy; XPS; critical stress intensity factor (KIC ).

INTRODUCTION

In recent years, carbon fibers are widely used as reinforcing materials in high-performance composite materials. Carbon fibers present several advantages, such as high modulus and strength, good stiffness, and creep resistance. In addition, the fiber shows good compatibility with a phenolic matrix. The improvement of interfacial adhesion between these components is attributed to the presence of polar groups (hydroxyl and carboxyl) on the fiber surface, which are able to interact with the active groups present in the phenolic matrix (1, 2). The physical and chemical properties of carbon fibers are usually modified to achieve good adhesion between the reinforcement and the matrix materials. To enlarge the surface polarity of nonpolar carbon fibers, various surface treatment techniques are applied, such as oxidation in acid solutions (3, 4), dry oxidation in oxygen, anodic oxidation, plasma treatments, mild fluorination (5–9), and metallic coating (10–12). Interfacial adhesion

EXPERIMENTAL

Materials For the present investigation, the reinforcement materials were continuous polyacrylonitrile (PAN)-based carbon fibers (12 K, TZ-307) manufactured by Taekwang of Korea. Prior to use, the fiber surfaces were cleaned by placing them in a

1 To whom all correspondence should be addressed. E-mail: [email protected]. re.kr.

91

0021-9797/01 $35.00

C 2001 by Academic Press Copyright ° All rights of reproduction in any form reserved.

92

PARK AND JANG

cated in a hot-press at 150◦ C and 7.4 MPa for 150 min using a vacuum bagging method (22). The composites made were cut into test specimens for mechanical testing.

TABLE 1 The Physical Properties of Carbon Fibers Physical properties Fiber type Filament diameter (µm) Tensile strength (MPa) Tensile modulus (GPa) Ultimate elongation (%)

TZ–307 6.85 3528 245 1.5

XPS Studies of Fiber Surfaces

Soxhlet extractor with acetone. Details of these fibers are given in Table 1. Phenolic resin (CB-8057, supplied from Kangnam Chemical Co. of Korea) was used as the polymeric matrix. Nickel sulfate (NiSO4 ), nickel chloride (NiCl2 ), and boric acid (H3 BO3 ) were used as the electrolyte systems for the carbon fiber surface treatment. Electrolytic Ni-Plating An electroplating device was constructed that can continuously plate nickel onto the fiber surface. The speed of carbon fibers was controlled by a gearbox and the speed was normally about 0.72 m min−1 . Nickel sulfate was the main salt used in the electroplating solution, and electrolytic nickel plate was used as the anode. Carbon fibers are conductive, so they were used directly as the cathode to be plated. The compositions and operating conditions of the plating bath are given in Table 2. Before being plated, the carbon fibers were activated in nitric acid for 30 min in order to enhance the interfacial adhesion between the nickel coating and the carbon fibers. Prior to use following fiber surface analysis or preparation of the composites, the residual chemicals used were removed by Soxhlet extraction with acetone at 70◦ C for 2 h. Finally, the carbon fibers were washed several times with distilled water and dried in a vacuum oven at 60◦ C for 12 h. The amount of Ni-plating was determined by atomic absorption spectrophotometry (AAS). Preparation of the Carbon Fibers/Phenolic Resin Composites The composite preparations were done by filament winding, where the carbon fiber was continuously soaked in the phenolic matrix bath before being wound onto the mandrel. Specimens were prepared from laminates composed of 22 plies and fabri-

TABLE 2 Composition and Operating Conditions of Ni-Plating Bath (Watts Bath) Composition

NiSO4 · 6H2 O NiCl2 · 6H2 O H3 BO3

280 g/L 40 g/L 30 g/L

Conditions

pH Temperature (◦ C) Current density (A m−2 )

4.5–5.0 45±1 0–60

The surfaces of carbon fibers were analyzed using a VG Scientific LAB MK-II X-ray photoelectron spectrometer (XPS). The spectra were collected using a MgKα X-ray source (1253.6 eV). The pressure inside the chamber was held below 5 × 10−8 Torr during analysis. Both survey and high-resolution XPS spectra are recorded at a 45◦ take-off angle. A Gaussian–Lorentzian function was used for curve fitting O1s and C1s photopeaks. The C1s electron binding energy was referenced at 284.6 eV. Contact Angle Measurements Contact angle was used as a parameter to characterize the wetting performance and surface free energy of the surface-treated fibers. Contact angle measurements of carbon fibers were performed using a Kr¨uss Processor Tensiometer K 12 with fiber apparatus (23). Two grams of carbon fiber was packed into the apparatus and then mounted to the measuring arm of a microbalance. The packing factor of the fibers was measured for each continuous filament by measuring the increase in weight per unit time at zero depth of immersion of a completely wetting test liquid (in the current study, n-hexane). The test liquids used for contact angle measurements were n-hexane, deionized water, and diiodomethane. The surface tensions of these liquids are known (surface tension of water, γL = 72.8 mJ m−2 , polar component of the surface tension, γ LSP = 51.0 mJ m−2 , and London dispersive component of the surface tension, γ LL = 21.8 mJ m−2 . For diiodomethane γL = 50.8 mJ m−2 , γ LSP = 0.38 mJ m−2 , and γ LL = 50.42 mJ m−2 ) (14, 24). Mechanical Properties A three-point bending test was conducted using an Instron model Lloyd LR-5K mechanical tester according to the ASTM E399 (K IC , span-to-depth ratio, 4 : 1). The test speed for fracture toughness was 1 mm min−1 . The K IC was given by (25): K IC =

PS f (a/W ), BW 3/2

√ MPa m

[1]

and f (a/W ) =

3(a/W )1/2 [1.99 − (a/W )(1 − a/W )(2.15 − 3.93a/W + 2.7a 2 /W 2 )] , 2(1 + 2a/W )(1 − a/W )3/2

[2] where P is the rupture force, S the span between the supports, and W and B the specimen width and thickness, respectively.

93

PHENOLIC RESIN MATRIX COMPOSITES

RESULTS AND DISCUSSION

Nickel-Coated Carbon Fibers The SEM photographs are shown in Fig. 1 as a function of current density for the surfaces of nontreated (A) and nickel-plated (B–E) carbon fibers. As the electrodeposition current density was increased, the nickel coating thickness increased significantly. The metal coating appeared to cover the entire surface of the fiber (at >10 ∼ <30 A m−2 ). Also, above 30 A m−2 current density, the nickel coating formed rod-like microstructures. The amounts of nickel on the fibers, measured by AAS, are listed in Table 3.

TABLE 3 Ni Quantification of the Electrolytic Ni-Plated Carbon Fibers by AAS Current density (A m−2 )

Nickel/carbon fiber (mg g−1 )

0 5 10 30 60

0 3.1 5.7 10.6 12.3

FIG. 1. SEM micrographs of the nontreated and nickel-plated carbon fibers (×5000, original magnification). (A) 0, (B) 5, (C) 10, (D) 30, and (E) 60 A m−2 .

94

PARK AND JANG

TABLE 4 Elemental Composition and O1s /C1s Ratio of Nontreated and Nickel-Plated Carbon Fibers Elemental composition

FIG. 2. XPS spectra of the nontreated and nickel-plated carbon fibers. (a) 0, (b) 5, (c) 10, (d) 30, and (e) 60 A m−2 .

These SEM micrographs of the coated carbon fibers, as shown in Fig. 1, suggest that nickel first deposits on the fiber surface at energetically favored sites, such as kinks, edges, corners, and other structural irregularities and then covers the entire surface of the fiber. XPS Studies of Fiber Surfaces XPS spectra of nontreated and nickel-coated carbon fibers are shown in Fig. 2. As anticipated, the nontreated carbon fibers (Fig. 2a) show a C1s peak and a substantial O1s peak at 284.6 and 532.8 eV, respectively (26). The O1s peak is probably due to intrinsic surface carbonyl or carboxyl groups. Otherwise, for the nickel-plated carbon fibers (Figs. 2b–2e), carbon, oxygen, and nickel (BE = 857.6 eV) peaks are observable in XPS (26). The O1s peak of nickel-plated carbon fibers (Figs. 2b–2e) is probably due to NiO, C==O, –OH, and O–C–O groups.

Current density (A m−2 )

C1s ([AT]%)

O1s ([AT]%)

N1s ([AT]%)

0 5 10 30 60

68.8 64.4 62.8 63.3 63.0

25.8 25.4 28.0 26.8 23.9

0.8 0.8 0.8 0.8 0.8

Composition ratio O1s /C1s 0.375 0.394 0.446 0.423 0.379

Figure 3A shows expanded scale O1s XPS spectra for carbon fibers coated with nickel at 10 A m−2 current density. The O1s spectra reveal the presence of three peaks corresponding to NiO groups (peak 1, BE = 529.6 eV), C==O, or –OH groups (peak 2, BE = 531.6 eV), and O–C–O groups (peak 3, BE = 532.6 eV) groups (26–30). The Ni2p peak is shown on an expanded scale in Fig. 3B for nickel-coated carbon fibers. In the case of the nickelcoated carbon fibers in Fig. 3B, several subpeaks are seen in addition to the main peak (BE = 858.5 eV). These subpeaks which are separated computationally in Fig. 3B indicate that Ni metal (Peak 1, BE = 852.7 eV, and Peak 4, BE = 858.5 eV), NiO (Peak 2, BE = 854.6 eV), and Ni(OH)2 (Peak 3, BE = 856.4 eV) are present as a result of electrolytic nickel plating (31–33). However, these peaks are never seen for the nontreated carbon fibers. For the nontreated and nickel-plated carbon fibers, elemental composition and O1s /C1s ratio are listed in Table 4. The O1s /C1s composition ratios of nickel-coated carbon fibers are increased compared to nontreated due to the deposition of more active forms, such as, NiO, Ni(OH)2 , and Ni metal on the inactive carbon. However, nitrogen of carbon fiber surfaces have no significant changes as the concentration and distribution are varied.

FIG. 3. High-resolution O1s and Ni2p XPS spectra of nickel-plated carbon fibers (10 A m−2 current density). (A) 1, Ni-metal (BE = 852.7 eV); 2, NiO (BE = 854.6 eV); 3, Ni(OH)2 (BE = 856.4 eV); 4, Ni-metal (BE = 858.5 eV). (B) 1, NiO (BE = 529.6 eV); 2, C==O or –OH (BE = 531.6 eV); 3, O–C–O (BE = 532.6 eV).

PHENOLIC RESIN MATRIX COMPOSITES

95

From the XPS experimental results, it is found that the surface composition of carbon fibers changed substantially as a result of nickel plating. The carbon content of nickel-plated fibers decreased when the fibers were plated with metallic nickel, whereas the oxygen and nickel contents of coated fibers were higher than for nontreated fibers. The active groups on the carbon fiber surfaces after nickel plating may help to change the polarity and the functionality of fiber surfaces. Contact Angle Measurements and Surface Free Energy Analysis The wetting of a solid surface by a liquid and the concept of contact angle (θ ) was first formalized by Young (34), γS − γSL = γL cos θ,

[3]

where γL is the surface energy of the liquid, γSL is the interfacial energy of solid/liquid interface, and γS is the surface energy of solid. Fowkes (35) has suggested that the surface free energy of substances consists of two parts, the London dispersive and specific (or polar) components, γ = γ L + γ SP ,

[4]

where the superscript L refers to the contribution due to London dispersive forces which are common to all substances, and superscript SP relates to the specific polar contribution. In the context of surface energetic studies, Owens and Wendt (36) derived the following equation for the interfacial energy between liquid and solid assuming a geometric mean combination of the London dispersive and specific components: ¡ ¡ ¢1 / ¢1 / γSL = γS + γL − 2 γLL · γSL 2 − 2 γLSP · γSSP 2 .

[5]

Combining Eqs. [3] and [5] yields a linear equation, ¡ ¡ ¢1 / ¢1 / γL (1 + cos θ ) = 2 γLL · γSL 2 + 2 γLSP · γSSP 2 ,

[6]

where γL , γLL , and γLSP are known for the testing liquids (13, 21) and γS , γSL , and γSSP can be calculated by the contact angle measurements. Based on “harmonic” mean and force addition, Wu (37, 38) proposed the following equations: γSL = γS + γL −

4γSL γLL 4γ SP γ SP − SPS L SP · L L γS + γL γS + γL

[7]

Equation [7] can be written as follows with the aid of Eq. [3]: γL (1 + cos θ) =

4γSL γLL 4γSSP γLSP − · γSL + γLL γSSP + γLSP

[8]

FIG. 4. Surface free energy of the nontreated and nickel-plated carbon fibers (γS , surface free energy; γSSP , specific (polar) component; and γSL , London dispersive component).

Wu (37) claimed that this method applied accurately between polymers and between a polymer and an ordinary liquid. The contact angles of nontreated and nickel-coated carbon fibers were measured for two testing liquids, deionized water and diiodomethane. The γS , γSL , and γSSP of the carbon fibers studied are given in Fig. 4. As seen in Fig. 4, the treatments at the current density range of 5–60 A m−2 lead to a wetting level better than those for the sample which was not nickel-coated. The polar component, γSSP , of the nickel-plated carbon fibers was significantly increased, but the dispersive component γSL was barely changed. In this system, the surface polarity can also increase compared to the nontreated carbon fibers. These results reveal that the treatments lead to an improved interaction between the polar liquid (deionized water) and the carbon fiber surfaces. To improve wettability, the surface energy of the fibers should be made larger than or equal to the surface energy of the matrix. In the case of 10 A m−2 current density, the surface free energy, γs , reaches the highest value, 67 mJ m−2 . Thus, it is expected that the interface between metallized carbon fiber surface and phenolic resin should be good, since the metallized carbon fiber surface energy should allow extensive wetting by the phenolic resin (35–45 mJ m−2 ) (13, 14, 21). As mentioned above, these results are attributed to the introduction of van der Waals physical adsorption force and polar groups, such as –CO, –COO, NiO, Ni(OH)2 , and nickel metal of carbon fiber surfaces, resulting from increasing the specific component of the surface free energy, as shown in Fig. 3. Relationship between Surface Properties of Carbon Fibers and Fracture Toughness of the Composites The effect of surface treatment can be expressed in terms of the K IC values. The critical stress intensity factor (K IC , K C in mode I fracture), which is one of the fracture toughness parameters, is described by the state of stress in the vicinity of the tip of a crack as a function of the specimen geometry, the crack geometry and

96

FIG. 5. density.

PARK AND JANG

Evolution of critical stress intensity factor (K IC ) with the current

the applied load on the basis of linear elastic fracture mechanics (LEFM) (39). The K IC can be characterized by a single edge notched (SEN) beam fracture toughness test in the threepoint bending flexure (40). In Fig. 5, the results of the K IC tests of CFRP (carbon fibersreinforced plastic) are shown as a function of current density. The maximum strength is found at a current density of 10 A m−2 . Also, it is observed that the K IC of the CFRP for fibers treated at a relatively high current density (60 A m−2 ) is not significantly increased, compared with the composites made with untreated fibers. Nevertheless, it appears that in this system, the nickelplating of carbon fiber surfaces affects the K IC of the composites, resulting from both the presence of nickelized functional groups and increased oxide functional groups on the carbon fibers. It is interesting to note that the K IC results support the reliability of the data, since the trends in K IC values seem to be

FIG. 7. Dependence of the critical stress intensity factor (K IC ) on specific polar component, γSSP , of surface energy (R, coefficient of regression).

very similar to trends in O1s /C1s or surface free energy. In fact, the relationships between K IC and O1s /C1s and between K IC and γSSP are almost linear for all samples with different current densities, as shown in Fig. 6 and Fig. 7, respectively. Therefore, the O1s /C1s ratio and γSSP may prove to be the governing factors in the adhesion between the fibers and the phenolic matrices in this system. CONCLUSIONS

Electrolytic nickel-plating of carbon fiber surfaces which promote better fibers/matrix adhesion result primarily from an increase of C==O, O–C–O, NiO, Ni(OH)2 , and nickel metal concentration, as measured by XPS, and from an increase in the total surface free energy of the carbon fiber surface, as determined by contact angle measurement. The change in the total surface free energy comes about primarily through an increase in the specific polar component, γSSP , of the fiber surface free energy. Consequently nickel-coated fiber surfaces which have higher activity are effective in promoting increased adhesion due to the presence of oxide groups and nickel functional groups on the fiber surface, which produce changes in the O1s /C1s ratios as well as in the surface free energy of the carbon fibers. Also, the relationship between mechanical properties of the composites and fiber surface chemical properties of fiber are quite consistent. REFERENCES

FIG. 6. Dependence of the critical stress intensity factor (K IC ) on O1s /C1s ratio (R, coefficient of regression).

1. Wang, S., and Garton, A., J. Appl. Polym. Sci. 45, 1743 (1992). 2. Park, S. J., Cho, M. S., and Lee, J. R., J. Colloid Interface Sci. 226, 60 (1999). 3. Donnet, J. B., and Bansal, R. C., “Carbon Fibers.” 2nd ed. Dekker, New York, 1990. 4. Wu, Z., Pittman, C. U., Jr., and Gardner, S. D., Carbon 33, 597 (1995). 5. Fukunaga, A., and Ueda, S., Compos. Sci. Technol. 60, 249 (2000). 6. Yuan, L. T., Shyu, S. S., and Lai, J. Y., J. Appl. Polym. Sci. 42, 2525 (1991).

PHENOLIC RESIN MATRIX COMPOSITES 7. Park, S. J., and Kim, M. H., J. Mater. Sci. 35, 1901 (2000). 8. Oliver, S. R. J., Bowden, N., and Whitesides, G. M., J. Colloid Interface Sci. 224, 425 (2000). 9. Chang, Y., and Ohara, H. J., J. Fluorine Chem. 57, 169 (1992). 10. Beydon, R., Bernhart, G., and Segui, Y., Surf. Coat. Technol. 126, 39 (2000). 11. Hage, Jr, E., Costa, S. F., and Pessan, L. A., J. Adhes. Sci. Technol. 11, 1491 (1997). 12. Abraham, S., Pai, B. C., Satyanarayana, K. G., and Vaidyan, V. K., J. Mater. Sci. 25, 2839 (1990). 13. Hoecker, F., and Kocsis, J. K., J. Appl. Polym. Sci. 59, 139 (1996). 14. Park, S. J., in “Interfacial Forces and Fields: Theory and Applications” (J. P. Hsu, Ed.), pp. 385–442. Dekker, New York, 1999. 15. Park, S. J., Jin, J. S., and Lee, J. R., J. Adhes. Sci. Technol. 14, 1677 (2000). 16. Park, S. J., and Lee, J. R., J. Mater. Sci. 33, 647 (1998). 17. Lee, W. S., Sue, W. C., and Lin, C. F., Compos. Sci. Technol. 60, 1975 (2000). 18. Carotenuto, G., Gallo, A., and Nicolais, L., Adv. Compos. Lett. 3, 139 (1994). 19. Jahazi, M., and Jalilian, F., Compos. Sci. Technol. 59, 1969 (1999). 20. Brenner, A., “Electrodeposition of Alloys: Principle and Practice.” Academic Press, New York, 1963. 21. Park, J. M., and Bell, J. P., “Adhesion Aspects of Polymeric Coating.” Plenum, New York, 1983. 22. Park, S. J., and Cho, M. S., J. Mater. Sci. Lett. 18, 373 (1999). 23. Rho, S. B., and Lim, M. A., Polymer (Korea) 23, 662, (1999). 24. Wu, S., “Polymer Interface and Adhesion.” Dekker, New York, 1982.

97

25. Wang, T. W. H., Blum, F. D., and Dharani, L. R., J. Mater. Sci. 34, 4873 (1999). 26. Wagner, C. D., Riggs, W. M., Davis, L. E., and Moulder, J. F., “Handbook of X-ray Photoelectron Spectroscopy.” Perkin-Elmer Corp, Norwalk, 1979. 27. Gardner, S. D., Singamsetty, C. S. K., Booth, G. L., He, G. R., and Pittman, C. U., Jr., Carbon 33, 587 (1995). 28. Yang, W. P., Costa, D., and Marcus, P., J. Electrochem. Soc. 141, 2669 (1994). 29. Brown, N. M. D., Hewitt, J. A., and Meenan, B. J., Surf. Interface Anal. 18, 187 (1992). 30. Mcintyre, N. S., and Gook, M. G., Anal. Chem. 47, 2208 (1975). 31. Kim, K. S., and Davis, R. E., J. Electron Spectrosc. Relat. Phenom. 1, 251 (1973). 32. Barr, T. L., J. Phys. Chem. 82, 1801 (1978). 33. Barr, T. L., J. Vac. Sci. Technol. 9, 1793 (1991). 34. Kinloch, A. J., “Adhesion and Adhesives.” Chapman and Hall, London, 1987. 35. Fowkes, F. M., J. Phys. Chem. 67, 2538 (1963). 36. Owens, D. K., and Wendt, R. C., J. Appl. Polym. Sci. 13, 1741 (1996). 37. Wu, S., J. Polym. Sci., Part C 34, 19 (1971). 38. Wu, S., in “Adhesion and Adsorption of Polymers, Polymer Science and Technology” (L. H. Lee, Ed.), Vol. 12A, p. 53. Plenum, New York, 1980. 39. Griffith, A. A., Phil. Trans R. Soc. 221, 163 (1921). 40. Gopal, P., Dharani, L. R., and Blum, F. D., Polym. Polym. Compos. 5, 327 (1997).