Reinforcing Porous Silica with Carbon Nanotubes to Enhance Mechanical Performance

http://jcm.sagepub.com

Journal of Composite Materials

DOI: 10.1177/0021998306067262

2007; 41; 979

Journal of Composite Materials

Jen-Tsung Luo, Hua-Chiang Wen, Chang-Pin Chou, Wen-Fa Wu and Ben-Zu Wan

Performance

Reinforcing Porous Silica with Carbon Nanotubes to Enhance Mechanical

http://jcm.sagepub.com/cgi/content/abstract/41/8/979

The online version of this article can be found at:

Published by:

http://www.sagepublications.com

On behalf of:

American Society for Composites

can be found at: Journal of Composite Materials

Additional services and information for

http://jcm.sagepub.com/cgi/alerts Email Alerts: http://jcm.sagepub.com/subscriptions Subscriptions: http://www.sagepub.com/journalsReprints.nav Reprints: http://www.sagepub.co.uk/journalsPermissions.nav Permissions: http://jcm.sagepub.com/cgi/content/refs/41/8/979 Citations

JEN-TSUNGLUO,* HUA-CHIANG WEN ANDCHANG-PINCHOU Department of Mechanical Engineering, National Chiao Tung University

Hsinchu, Taiwan

WEN-FA WU

National Nano Device Laboratories, Hsinchu, Taiwan

BEN-ZUWAN

Department of Chemical Engineering, National Taiwan University Taipei, Taiwan

ABSTRACT: In this study, porous silica (SiO2) film is reinforced with carbon

nanotubes (CNTs), which gets homogeneously dispersed in porous SiO2 gel by

ultrasonic stirring. The size of SiO2 molecules is the same as that of nanotubes,

so a fine CNTs/SiO2network is formed. Nanoindenters used to test the mechanical

properties of the resulting specimens show a marked enhancement in the properties as a function of filler content. Interfacial bonding and the distribution of CNTs also strongly affected the mechanical performance. Although a high concentration of CNTs improves the mechanical characteristics of the matrix, it also causes agglomeration such that the mechanical performance at the surface was dispropor-tionate. Fourier transform infrared spectroscopy (FTIR) analysis indicates that there is little chemical reaction occurring between CNTs and SiO2. The results suggest that

CNTs are ineffective reinforcements.

KEY WORDS: porous SiO2, carbon nanotubes (CNTs), nanocomposites,

mechanical properties.

INTRODUCTION

P

integration circuit processes, microelectronic technology, environmental technology,

*Author to whom correspondence should be addressed. E-mail address: oam.me90g@nctu.edu.tw Figures 2–6 and 9–11 appear in color online: http://jcm.sagepub.com

Journal ofCOMPOSITEMATERIALS, Vol. 41, No. 8/2007 979

sensors, membranes, and insulators [1–4]. The conventional porous SiO2was mesoporous

in nature when it was first discovered in 1992 [5,6]. It is formed using surfactant micelles and liquid crystal as a template. Mesoporous SiO2 is easily fabricated, and

the pore diameters are in the range of 10–50 nm. However, the use of a high-porosity and low-polarity material and formation of inorganic material with uniform pores of arbitrary size, shape, and orientation, to improve the electrical and dielectrical properties, remains a challenge. Porous SiO2 is of loose structure, with low mechanical strength, making

further process fabrication difficult. Hence, significant interest exists in the development of methods using reinforced filler, to obtain favorable mechanical characteristics of porous SiO2material.

Since the first discovery of carbon nanotubes (CNTs) in 1991 [7], a new and powerful method to overcome the mechanical disadvantages of traditional materials has been developed. The CNTs are regarded as suitable composite fillers because they have excellent mechanical properties, such as high Young’s modulus, excellent flexibility, low density, and high thermal conductivity.

Two main issues must be addressed to improve the mechanical properties of the composite material. They are the interfacial bonding and proper dispersion of fillers in the matrix [8,9]. According to the report, if only 1% of carbon atoms are combined with the matrix material, then the interfacial shear stress will be effectively transferred to the CNTs [10,11]. The dispersion of CNTs in the matrix is another problem to be solved. The CNT is a nanoscale material that has a giant surface area (more than 1000 m2/g). The surface area is so enormous that van der Waal’s force between any two carbon nanotubes is very strong. This force causes them to aggregate easily, which causes difficulty in dispersing CNTs in the matrix material. The homogeneous dispersion of CNTs or an exfoliation of the agglomerates and good wetting of the matrix material are considered to solve the problem of dispersion [12].

In this study, CNTs are employed as reinforcements that increase the mechanical strength of porous SiO2film. The solgel process is used to disperse CNTs homogeneously

in a porous SiO2 matrix. The mechanical properties, such as Young’s modulus and

hardness were evaluated.

EXPERIMENTAL

Multiwalled carbon nanotubes (MWCNTs), which grew from the cobalt catalyst, were synthesized from natural gas by catalytic chemical vapor deposition (CVD), with an average length of several microns, and a diameter of 10–50 nm (measured by transmission electron microscope, TEM), shown in Figure 1. The MWCNTs samples were treated by immersing them in nitric acid solution for 6 h to reduce the van der Waal’s force, remove the impurities, and eliminate the metallic characteristics of the catalyst. Subsequently, samples were washed with deionized water on a filtration membrane until the acidic solution became neutral. Finally, the MWCNTs were dried in a vacuum oven at 80
_{C for 36 h.}

The porous SiO2solution was prepared using Tween80 (polyoxyethylene (20) sorbitan

monooleate), deionized water, ethanol (95%), hydrochloric acid (70%), and TEOS (tetraethylorthosilicate). Hydrochloric acid was used as an activating agent. Deionized water and ethanol were used as solvents. The TEOS is the precursor of SiO2and Tween80

is a surface-activating agent. The mixing ratio (by weight) of Tween80: deionized water: ethanol: hydrochloric acid: TEOS was 2.7 : 1 : 6 : 0.3 : 3.4. The CNTs, dispersed

in ethanol as a suspension, were added to the porous SiO2 solution with weight

fractions from 0 to 5%. The mixed CNTs/porous SiO2 solution was stirred for 3 h at

room temperature to make the solution more homogeneous. Then, the solution was spin-coated on glasses and silicon wafers. The spin-coated samples were then soft-baked at 106
_C for 1 h to solidify the mixture. Then, the film was calcined at 400
_{C in a furnace for 30 min} to burn out the template molecules and solidify the pores. Finally, the surface was hydrophobically treated by immersing the samples in hexamethyldisilazane (HMDS) solution.

Samples from extruded composite strips were characterized using a scanning electron microscope (SEM), a transmission electron microscope (TEM), and an atomic force microscope (AFM), which yielded micrographs and surface roughness after dispersing CNTs. The nanotubes, embedded in the matrix, are invisible from the surface of the film, so the film is fractured by mechanical force or chemical etching. The fractured surface is then scanned to elucidate the microstructure. Chemical effects, such as functional groups and specific bonding characteristics were examined by FTIR analyses. The hardness and Young’s modulus were measured using a nanoindenter system from nano-instruments, operated at a constant displacement rate. The stiffness was continuously measured at a small penetration depth.

RESULTS AND DISCUSSION

Figure 2 shows the AFM image of the films. It shows that the average surface roughness of the films increase from 0.34 to 1.8 nm as the weight fraction of CNT content increases from 0 to 5%. Figure 3 is the 3D magnified image of the film, with 1% CNTs, where CNTs lie on the surface. A bed-like structure of CNTs had the expected height and increased at the end point. Many researchers have found a relation between film surface morphology and electrical conductivity [13]. It is said that the nodular structure has been associated

20 nm

with high-doped conducting film in which the nodules represent doping rich and highly conductive areas.

Figure 4(a) displays the SEM micrographs of cross sections of randomly oriented CNTs embedded in the porous silica matrix. Figure 4(b) presents the etching fracture surface on which some CNTs are observed. The diameter of the tubes is approximately 25 nm. Some CNTs were vertically drawn out by an extensive polishing force and integrated, as shown in Figure 4(c). Interfacial bonding between CNTs and porous SiO2matrix was

important for the loading test. A load transfer, from the matrix to CNTs, will be effective, if the interfacial bonding is strong. The effective reinforcing modulus also strongly depends on the geometry of the filler and the combination ratio of CNTs to porous SiO2.

A bent nanotube reduces the stiffness of the composite. The constraint of the surrounding matrix on the straightening of the wavy nanotube may be significant. The geometry of wavy CNTs reduces the effective reinforcing modulus and unbent nanotubes correspond to stronger reinforcement. (c) 5%, RMS = 1.80 nm (a) 0%, RMS = 0.34 nm (b) 1%, RMS = 0.45 nm 20.00 nm 20.00 nm 20.00 nm 0.2 0.4 0.6 0.8 0.2 0.4 0.6 0.8 0.2 0.4 0.6 0.8 _µm

20.000 nm 0.2 0.4 0.6 0.8 µm

Figure 3. Three-dimensional magnified AFM image of CNTs in porous silica matrix.

CNTs (a) (b) CNTs (c) CNT (d)

Figure 4. SEM micrographs of CNTs/SiO2composites: (a) cross section image of the CNTs/SiO2composite; (b) etching surface of CNTs/SiO2composite; (c) integrated image of CNTs; and (d) aggregate image of CNTs.

The importance of a homogeneous dispersion was determined from the image of the fractured structure. The nanotubes were dispersed by sonicating them in ethanol; by mixing the suspension with porous silica resin, and evaporating the solvent. Figure 4(b) and (c) shows that some CNTs in the porous silica matrix were randomly distributed but some were agglomerated together, as presented in Figure 4(d). The agglomerate degrades the mechanical properties of the film. As reported in literatures, the agglomerate of CNTs depends on the uniformity of dispersion and size of the CNTs [14].

Figure 5 displays the results of the FTIR analysis. The FTIR spectroscopic analysis is a typical method for characterizing functional groups of CNTs/porous SiO2 composite.

The most intense absorption peak is that of SiO (1060 cm1) in the main skeleton of SiO2.

This peak, centered at around 1100 cm1, was a shoulder at the high-frequency end of the peak at 1060 cm1, because the transverse optical vibration mode corresponded to the asymmetric optical stretching of the oxygen atoms in the Si-O-Si linkage [15–17]. The broad peak between 3300 and 3600 cm1 is related to the stretching of –OH. The hydrophobic trimethylsilyl peaks were present at 1277 and 2975 cm1, and are considered as SiC and CH3. These alkyl-like groups can protect the film from attracted moisture.

The mechanical performance of nanocomposite depends on the interfacial reaction between CNTs and porous silica. The atoms of Si and O are present in a strong network that is connected at short distances [18]; a loose pore structure would have a poor mechanical strength. The interfacial strength between the porous silica and the CNTs was very strong. Some C¼O groups (17001750 cm1), which may be carboxyl, ester, ketone, and aldehyde were found on the surface of the material. These groups take part in the polymeric reaction and enable the linkage of CNTs to the matrix, perhaps because of the mixed motion of CH2 and COC, with deformation in inorganic silicide solution,

where CH2is the
carbon of the matrix material and COC belongs to the CNTs. Jia et al.

[19] also suggested that the  bond of CNTs may be opened and interact with the matrix. The oxidation of CNTs are responsible for reducing the amount of impurities,

0 500 1000 1500 2000 2500 3000 3500 4000 4500 Wave numbers (cm−1)

Absorbance (arbitrary units)

Si-C Si-O OH C=O CH3 Without CNTs 1 % 2 % 3 % 4 % 5 %

the factor which may also contribute to the interaction between the CNTs and the silica matrix. However, C¼O groups may also be regarded as an oxidation product of ethanol which is used as a solvent. The presence of C¼O resulted in a polar surface and caused a chemical change.

The mechanical properties, such as hardness and Young’s modulus, were tested using a nanoindenter system. They were obtained from load–displacement data, with penetration depths of 1/10 of the thickness of the film, to avoid surface and substrate damage [20–23]. Figure 6 plots the typical load–displacement curve obtained from the nanoindenter system. The indenter is a triangular pyramid-shaped diamond with a shape angle of 115
_{. Every sample was tested 10 times at different locations of the surface.} The Young’s modulus and hardness were average values. Equation (1) yields the indentation depth, with the given pyramidal geometry, from the diagonal length of the indentation (lp). hp¼ lp ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 3  tan2_ð=2Þ p 3 tan =2ð Þ ð1Þ

where  is the edge angle of the indenter, lpis the diagonal length of the indentation, and hp

is the penetration depth. The surface area must be determined to measure the surface hardness, which equals the mean pressure under the indenter. The length of the diagonal of the indentation must be measured to determine the surface area.

Sp¼

9h2tan =2ð Þ

3  tan2_ð=2Þ: ð2Þ

Equation (2) yields the surface area (Sp) for a given pyramidal geometry. The plastic

depth can be determined from the penetration depth using heffas the plastic depth, which is

0 200 400 600 800 1000 −20 0 20 40 60 80 100 120 h_eff h_fin h_max Loading Unloading Displacement (nm) Loading force (mN)

associated with the point of intersection between the sloping line of the maximum load point and the x-axis, and is measured by the method of least squares [24]. The hfinis the

position where the indenter was detached from the surface of the sample. The hmax is

the position at which the indenter touches the sample’s maximum depth. Figure 6 presents the values of the aforementioned parameters.

Figure 7 plots the maximum loading force versus CNT content of the samples. The maximum loading force increased with CNT content, reaching 168 mN at a CNT weight fraction of up to 5%. The load force, without the CNT sample, was only about 116 mN. It cannot be regarded as an elastic or plastic modulus when the part of the curve from which the average gradient was determined involves both elastic and plastic deformation. However, the maximum load corresponded to a combination of the resistance of the material to the deformation before and after yielding. This value is also related to the yield strength of the material, because this value marks the transition between the maximum values of the load during the penetration depth experiment. This is a measure of the capacity of the material to absorb energy before breaking [25,26]. The result also indicated that this property is related to the wear resistance of a class of porous SiO2, including fillers. Figure 7 clearly shows that the addition of CNTs changed the

mechanical characteristics of porous SiO2.

Figure 8 presents the relationship between load and hardness. All of the samples considered herein were prepared under the same conditions, except for the CNT content. Hence, the hardness curve is assumed to be caused by the change in CNT content. The hardness increased with the penetration load. When the nanoindenter test is performed with a small load, the area of the sphere at the tip of the indenter is not negligible. The hardness of the composite was also shown to be a function of the filler content. The hardness curves at CNT contents of 1–3% are similar and the curves at CNT contents between 4 and 5% are also similar. Accordingly, a step-region threshold concentration of the filler may be considered to exist. When the filler content exceeded this threshold concentration, the mechanical performance of the matrix was substantially higher. The higher CNT content in porous silica film corresponds to

0 1 2 3 4 5 110 120 130 140 150 160 170 CNTs content (wt%)

Maximum loading force (mN)

greater hardness, because more CNT interacts with SiO2 molecules, strengthening the

transfer of loading.

The contact stiffness must be obtained from the gradient of the unload curve to find the effective Young’s modulus [27]. The effective Young’s modulus is related to the stiffness and given by Equation (3).

S ¼dP dh¼CAE  ffiffiffiffiffi A_r p ð3Þ where 1 E¼ 1  2 1 E1 þ1   2 s Es where E* is the reduced modulus; CAequals 2= ffiffiffi

p

, and Ar is the projected real contact

area between indenter and surface. E1and 1are the Young’s modulus and the Poisson’s

ratio of the indenter. Es and s are the corresponding parameters of the test sample. S

represents the contact stiffness. When the nanoindentation test is undertaken with the triangular pyramidal indenter with a shape angle of  ¼ 115
_{, the constant (C}

A) must

be 2=pffiffiffi.

When the indenter tip penetrated the composites, the loading force was significantly dispersed by the large quantity of CNTs. Figure 9 plots Young’s modulus versus displacement. The modulus increases with the displacement. The gradient of Young’s modulus declines at shallow depths, until a contact depth of 320 nm is reached. The curve suggests that the sample with CNT fillers is stiffer than that without fillers. If the surface layer, without CNTs fillers, undergoes elastic deformation, then plastic deformation will occur. The similar plot of hardness versus displacement (Figure 10) shows analogous behavior as in Figure 9. 0 20 40 60 80 100 120 140 160 180 0 1 2 3 4 5 6 7 0% 1% 2% 3% 4% 5% Hardness (GPa) Loading force (mN)

The effective modulus of the CNTs/porous SiO2 composite is related to the weight

fraction of CNTs. Figure 11 plots the hardness and Young’s modulus of the composites as functions of the weight fraction of CNTs from 1 to 5%. At a CNT content of 1%, the Young’s modulus was about four times larger than that of the unreinforced sample, and increased with CNT content. The dependence of Young’s modulus on the nanotube weight fraction is associated with the effect of load transfer from the matrix to CNTs, which have a large aspect ratio and excellent mechanical properties. The effective distribution of CNTs in the surrounding matrix at low content also contributes to this dependence. If CNTs are dispersed uniformly in the matrix, the effective load transfer

0 200 400 600 800 1000 −20 0 20 40 60 80 100 120 140 160 180 0% 1% 2% 3% 4% 5% Modulus (GPa) Penetration depth (nm) CNT contents (%)

Figure 9. Modulus vs displacement curve of CNTs with different weight contents on the CNTs/SiO2 composite. 0 200 400 600 800 1000 0 5 10 15 20 0% 1% 2% 3% 4% 5% Penetration depth (nm) Hardness (GPa) CNT contents (%)

from the matrix to CNTs is very enormous. Load sharing between the CNTs and matrix enhances the mechanical properties of the composite. The results also demonstrate that CNT-reinforced polymer or alumina composite are stronger than other fillers [28,29]. However, higher CNT contents of over 10% resulted in the agglomeration of CNTs in the surrounding matrix, and suppressed densification. A high fraction of nanotubes also caused the deflection of cracks and continuous crack propagation [30]. It also degraded the mechanical properties. Additionally, sintering impurities may promote the crystallization of SiO2by heterogeneous nucleation [31].

CONCLUSION

Different weight fractions of CNTs and porous SiO2 films were produced. The

CNTs were homogeneously dispersed into porous silica solgel using a sonication method. A functional nanotube which does not interact chemically with porous SiO2expresses itself

in the presence of agglomerates. The hardness and Young’s modulus of CNTs/porous SiO2

composites were enhanced with the increase of CNT content. It was due to the load transfer between the CNTs and porous silica from interfacial bonding. However, higher CNT content made CNTs agglomerate together and degraded the mechanical performance. The method presented in this study provides a means for parametrically exploring the reinforcement–matrix relationships. Further improvement should optimize the process of procedure to enhance the reinforcing role of CNTs.

ACKNOWLEDGMENT

The authors would like to thank the National Science Council of the Republic of China, Taiwan, for financially supporting this research under Contract No. NSC 94-2216-E-009-027. −1 0 1 2 3 4 5 6 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 10 20 30 40 50 60 Hardness (GPa) CNTs content (wt%)

REFERENCES

1. Wang, Y.H., Moitreyee, M.R., Kumar, R., Wu, S.Y., Xie, J.L., Yew, P., Subramanian, B., Shen, L. and Zeng, K.Y. (2004). The Mechanical Properties of Ultra-low-dielectric-constant Films, Thin Solid Films, 462–463: 227.

2. Sugahara, G., Aoi, N., Kubo, M., Arai, K. and Sawada, K. (1997). Nano-scale Structure Control of Mesoporous Silica, In: Proceedings of the Third International Dielectrics for ULSI Multilevel InterconnectionConference, Santa Clara, CA, USA February (1997), p. 19.

3. Martin, S.J., Goldschalx, J.P., Mills, M.E., Shaffer II, E.O. and Townsend, P.H. (2000). Development of a Low-dielectric-constant Polymer for the Fabrication of Integrated Circuit Interconnect, Adv. Mater., 12: 1769.

4. Morgen, M., Ryan, E.T., Zhao, J.H., Hu, C., Cho, T.H. and Ho, P.S. (2000). Pore Size and Pore-size Distribution Control of Porous Silica, Annu. Rev. Mater. Sci., 30: 645.

5. Kresge, C.T., Leonowicz, M.E. and Roth, W.J. (1992). Ordered Mesoporous Molecular Sieves Synthesized by a Liquid-crystal Template Mechanism, Nature, 359: 710.

6. Beck, J.S. and Vartuli, J.C. (1992). A New Family of Mesoporous Molecular Sieves Prepared with Liquid Crystal Templates, J. Am. Chem. Soc., 114: 10834.

7. Iijima, S. (1991). Pentagons, Heptagons and Negative Curvature in Graphite Microtubule Growth, Nature, 354: 56.

8. Cooper, C.A., Cohen, S.R., Barber, A.H. and Wagner, H.D. (2002). Detachment of Nanotubes from a Polymer Matrix, Appl. Phys. Lett., 81: 3873.

9. Lu, K.L., Lago, M., Chen, Y.K., Green, M.L.H., Harris, P.J.F. and Tsang, S.C. (1996). Mechanical Damage of Carbon Nanotubes by Ultrasound, Carbon, 34: 814.

10. Shelimov, K.B., Esenaliev, R.O., Rinzler, A.G., Huffman, C.B. and Smalley, R.E. (2004). Carbon Nanotube-reinforced Epoxy-composites: Enhanced Stiffness and Fracture Toughness at Low Nanotube Content, Compos. Sci. Technol., 64: 2363.

11. Frankland, S.J.V., Caglar, A., Brenner, D.W. and Griebel, M. (2002). Molecular Simulation of the Influence of Chemical Cross-Links on the Shear Strength of Carbon Nanotube-Polymer Interfaces, J. Phys. Chem. B, 106: 3046.

12. Zhang, M., Yudasaka, M., Koshio, A. and Iijima, S. (2001). Effect of Polymer and Solvent on Purification and Cutting of Single-wall Carbon Nanotubes, Chem. Phys. Lett., 349: 25. 13. Barishi, J.N., Stella, R., Spinks, G.M. and Wallace, G.G. (2000). Carbon Nanotubes with

Single-layer Walls, Electrochim. Acta., 46(4): 519.

14. Liu, W., Hoa, S.V. and Pugh, M. (July 2003). Effect of Carbon Nanotube Addition on the Tribological Behavior of Carbon/Carbon Composites, In: Proceedings of 14th International Conference on Composite Materials, ICCM-14, San Diego, CA, July (2003), Paper ID 1407. 15. Yu, S., Wong, T.K.S., Pita, K., Hu, X. and Ligatchev, V. (2002). Skeletal Silica Characterization

in Porous-silica Low-dielectric-constant Films by Infrared Spectroscopic Ellipsometry, J. Appl. Phys., 92: 333.

16. Gorman, B.P., Rosa, A. and Mueller, D.W. (2001). High Strength, Low Dielectric Constant Fluorinated Silica Xerogel Films, Appl. Phys. Lett., 79: 4010.

17. Lim, S.W., Shimogaki, Y., Nakano, Y., Tada, K. and Komiyama, H. (1996). Preparation of Low-Dielectric-Constant F-Doped SiO2 Films by Plasma-Enhanced Chemical Vapor

Deposition, Jpn. J. Appl. Phys., 35: 1468.

18. Han, Y.S., Li, J.B., Wei, Q.M. and Wen, Z.H. (2003). The Effect of SiO2Addition on Porous

Silica Composite Strength, Mater. Lett., 57: 3847.

19. Jia, Z., Wang, Z., Xu, C., Liang, J., Wei, B., Wu, D. and Zhu, S. (1999). Study on Poly(methyl methacrylate)/carbon Nanotube Composites, Mater. Sci. Eng., A271: 395.

20. Shen, L. and Zeng, K.Y. (2004). Comparison of Mechanical Properties of Porous and Non-porous Low-k Dielectric Films, Microelectron. Eng., 71: 221.

21. Shen, L., Zeng, K.Y., Wang, Y.H., Narayanan, B. and Kumar, R. (2003). Determination of the Hardness and Elastic Modulus of Low-k Thin Films and their Barrier Layer for Microelectronic Applications, Microelectron. Eng., 70: 115.

Nanoindentation Techniques, Ultramicroscopy, 97: 481.

24. Doener, M.F. and Nix, W.D. (1986). On the Axial and Interfacial Shear Stresses due to Thermal Mismatch in Hybrid Composite Sheets, J. Mater. Res., 1: 601.

25. Edidin, A.A. and Kurtz, S.M. (2001). Model Experiment on a Protrusion Method for Measurement of Interface Shear Sliding Stress in Fiber-reinforced Composite, Funct. Biomater., 198: 1.

26. Mao, X., Saito, M. and Takahahi, H. (1987). A New Look at Shear Correction Factors and Warping Functions of Anisotropic Laminates, Scripta Metal Mater., 25: 2481.

27. Fukuhara, M. and Yamauchi, I. (1993). Effect of Indenter Tip Radius on the Load Deflection Behavior of Thin Plates, J. Mater. Sci., 28: 4681.

28. Odegard, G.M., Gates, T.S., Wise, K.E., Park, C. and Siochi, E.J. (2003). Constitutive Modeling of Nanotube-reinforced Polymer Composites, Compos. Sci. Technol., 63: 1671.

29. Mo, C.B., Cha, S.I., Kim, K.T., Lee, K.H. and Hong, S.H. (2005). Fabrication of Carbon Nanotube Reinforced Alumina Matrix Nanocomposite by Sol–Gel Process, Mater. Sci. Eng. A, 395: 124.

30. Liao, Y.H., Olivier, M.T., Liang, Z., Zhang, C. and Wang, B. (2004). Investigation of the Dispersion Process of SWNTs/SC-15 Epoxy Resin Nanocomposites, Mater. Sci. Eng. A, 385: 175.

31. Peter, P.B. (1983). Stress Distribution along Broken Fibres in Polymer-matrix Composites, J. Am. Ceram. Soc., 66: C188.