Preparation of Unidirectional Carbon Fiber Preform

for Aluminium Matrix Composites*1

Moonhee Lee1;*2, Yongbum Choi2, Kenjiro Sugio2, Kazuhiro Matsugi2 and Gen Sasaki2

1Department of Machanical Science and Engineering, Graduate School of Engineering, Hiroshima University,Higashi-Hiroshima 739-8527, Japan2Department of Materials and Production Engineering, Faculty of Engineering, Hiroshima University,Higashi-Hiroshima 739-8527, Japan

The unidirectional carbon fiber (CF) preform for carbon fiber reinforced aluminium (CF/Al) composites has been investigated in terms ofthe fabrication condition and the strength property. The CF preform which consists of CFs and Cu particles was fabricated by spark plasmasintering (SPS) process with different fabrication temperatures. In order to determine the infiltration pressure of molten Al to the CF preform, thecompression test was performed to the CF preform. The unidirectional CF preform by SPS process was well formed with the fabricationtemperature condition at 1123 K in accordance with the formation of Cu particle bridging between fibers and Cu deposition on fibers. Thecompression strength of the CF preform increased with increasing fabrication temperature. Besides, the CF preform was deformed such as fibermicro-buckling or fiber kinking phenomena over the maximum compression strength. The infiltration pressure of the molten Al can be decidedunder the maximum compression strength. [doi:10.2320/matertrans.L-MZ201110]

(Received October 1, 2010; Accepted January 11, 2011; Published April 13, 2011)

Keywords: carbon fiber preform, spark plasma sintering, compression test, microstructure, carbon fiber (CF)/aluminium (Al) composites

1. Introduction

The various composite materials have been developed fora field of thermal dissipation components providing excellentadvantages such like the improvement of the thermalconductivity, tailorable coefficient of thermal expansion,weight saving and net shape fabrication process to replace theconventional materials.1,2) The thermal packaging of compo-site materials which possessed the excellent thermal con-ductivity can be used as the heat sink applications for the highheat generation electronic components. These componentshave been exploited the increase of power levels such assemiconductors, light emitting diode (LED) and converter.3,4)

Especially, the thermal dissipation of high power capacityconverter modules including insulated gate bipolar transistor(IGBT) or metal oxide semiconductor field effect transistor(MOSFET) can be one of the issues in accordance with thedevelopment of the hybrid electric vehicles (HEV) andelectric vehicles (EV) in an automobile field.5–7)

Recently, SiCp/Al and Diamond/Cu composites with highthermal conductivity have been investigated at the thermalmanagement industry.8,9) Although Diamond/Cu compositeswere expected to represent the high thermal conductivity, theheavy weight of Cu and high cost of diamond are able to limittheir application. In addition, both of SiC and diamond can bea defect for machining with their high hardness property.However, pitch-based CF/Al composites are able to accom-plish the high thermal conductivity and solve the problemsmentioned above.

The squeeze casting process has been well adopted as thefabrication process of composite materials by means ofmolten metal infiltration to porous fiber preform. However,

squeeze casting process for fabrication of fiber reinforcedcomposite materials is able to cause fiber fracture and/orinhomogeneous fiber distribution with by using high pres-sure.10,11) On the other hand, the low pressure infiltration(LPI) process is one of the promising infiltration methods byusing low applied pressure. It was also known that the LPIprocess for the composite materials enabled relatively simplefacilities, cost-effective and complex shape fabrication usinglow applied pressure.12,13) Prior to the fabrication of CF/Alcomposites by the LPI process, the optimization of fabrica-tion process of CF preform is one of the important parameterto obtain high performace composite materials.

In the present study, the unidirectional CF preform withaddition of Cu particles has been prepared by SPS methodwith different sintering temperatures. The compression teston CF preform was carried out in order to determine therange of the infiltration pressure of molten Al to the fiber axisdirection. In addition, the effect of the fabrication temper-ature on the deformation profiles of CF preform aftermaximum compression strength has been also discussed inconjunction with the observation of microstructures.

2. Experimental Procedure

The unidirectional CF preform contained of coal tar pitchbased K13D2U CFs (Mitsubishi Plastics, Inc.) and atomizedCu powders (Fukuda metal, foil & powder Co, LTD.) withan average particle size of about 2.55 mm. The inherent sizetreatment on the commercial fiber surface was removed inacetone for 40 min with ultrasonic cleaning. Subsequently,Cu powders with polyethylene glycol (PEG) were dispersedinto the fiber bundles for Cu spacing and bridging betweenthe CFs. The fiber mixtures were sintered to be CF preformby SPS method. The fiber mixtures put into the graphite moldand voltage and current were applied from the upper punch tolower with 4 V and the range of about 380–410 A for 30 min.

*1The Paper Contains Partial Overlap with the ICAA12 Proceedings by

USB under the Permission of the Editorial Committee.*2Corresponding author, E-mail: leemoonhee@hiroshima-u.ac.jp

Materials Transactions, Vol. 52, No. 5 (2011) pp. 939 to 942Special Issue on Aluminium Alloys 2010#2011 The Japan Institute of Light Metals

The CF preform was prepared with the different fabricationtemperature of 1073 K, 1123 K and 1173 K with about 380 A,390 A and 410 A, respectively. The dimension of the CFpreform was �10� 10ðtÞmm3. In addition, the compressiontest by direct loading to the fiber axis for unidirectional CFpreform was carried out to identify the compressive behaviorand determine the infiltration pressure of the molten Al. Thecrosshead speed of the compression test was 0.5 mm/min atroom temperature. The microstructure of as-received andafter compression test of CF preform was observed by SEM.The image analysis on microstructure of both CFs and Cubridging particles from CF preform has been carried out tocalculate the occupied area of Cu deposition particles and thecontact area of Cu bridging particles on CFs.

3. Results and Discussion

Figure 1 represents the microstructure of CF preformfabricated at the different sintering temperatures of 1073 K,1123 K and 1173 K, respectively. The addition of Cu powderswas determined to disperse into the CFs as spacer and fiberbridging material. In the studies of SPS process,14,15) thepowder compact with extremely small contact area betweenpowders proceeds mass transport by high local currentdensities and occurs melt and vaporization of adjacentparticles under SPS conditions. In this study, the addition ofCu powders is able to carry out not only the Cu bridgingbetween fibers, but also deposition on fibers by melt andvaporization under SPS process. It was also reported that Cucoatings by deposition on CFs facilitated the wettability withthe molten Al.16) As shown in Fig. 1(a), the CF preformconsists of CFs and well dispersed Cu particles betweenfibers. However there are little amount of Cu deposition onfibers and Cu bridging between fibers. In addition, littleamount of Cu bridging can indicate lower compressionstrength level of CF preform which debond easily betweenCFs and Cu particles by compression loading than wellbridged preform. The fiber bridging is inevitable to producethe preform shape formation with the expected dimension bybonding fibers each other and possess the strength of CFpreform against the infiltration pressure. The CF preformfabricated at 1123 K (Fig. 1(b)) shows significant Cu bridg-ing between CFs by coalescence of contact particles. Besides,the Cu deposition on CFs remarkably increased in accord-ance with the increase of current density conditions.

Figure 1(c) also shows the significant Cu deposition onfibers and bridging between fibers into the CF preform atthe sintering temperature of 1173 K. However, the Cubridging particles formed enlarged shape by bonding theadjacent particles continuously. It is possible that theinfiltration of molten Al can be disturbed by enlarged Cuparticles when the molten Al flows into the inter fiber regionby LPI process.

Figure 2 indicates the occupied area of Cu deposition onCF surface depending on the fabrication temperature. The CFpreform fabricated at 1073 K shows low deposition ratio ofabout 2.87%, whereas the CF preform fabricated at 1123 Kshows high percent of deposition of about 32.19% withincrease of current density. However, the Cu depositionratio of the preform fabricated at 1173 K decreased to about

16.49%, because of bonding and coalescence of adjacentdeposition particles.

Figure 3 represents the contact area between Cu bridgingparticles and fibers depending on the fabrication temperature.The contact area increased such as 2.95, 6.32 and 10.81 mm2

with the increase of the fabrication temperature of 1073 K,1123 K and 1173 K, respectively. It means that the Cuparticles grow up their size in accordance with the elevationof the fabrication temperature and bridge the fibers bycontinuous bonding and coalescence of adjacent Cu particles

Fig. 1 Microstructure of unidirectional CF preform depending on the

sintering temperature of (a) 1073 K, (b) 1123 K and (c) 1173 K,

respectively.

940 M. Lee, Y. Choi, K. Sugio, K. Matsugi and G. Sasaki

each other. In addition, the increase of contact area wasexpected to improve the compression strength properties ofCF preform with a part of resistance of fiber deformationby direct compression loading. As the results from Fig. 1 toFig. 3, the CF preform fabricated at the temperature of1123 K will be eligible for the low pressure infiltration ofmolten Al.

Figure 4 shows the compressive stress-strain curves ofthe unidirectional CF preform depending on the fabricationtemperature. The initial maximum strength of CF preformfabricated at 1073 K, 1123 K and 1173 K was evaluated as

1.02 MPa, 1.67 MPa and 2.15 MPa, respectively. The initialmaximun stress of preform increased with the increasingfabrication temperature. On the other hand, the plasticdeformation with energy absorption against the applied loadis started to occur after the initial maximum stress, so called‘plateau stage’ in case of the metal foam structure.17)

However, unidirectional CF preform is able to appear thefiber micro-buckling18) and fiber kinking19) phenomenoninstead of the plastic deformation. The amount of energyabsorption of the CF preform fabricated at 1173 K is morethan that of the preform fabricated at 1123 K. The increase

1050 1100 1150 12000

10

20

30

40

50A

rea

ratio

of

Cu

depo

sitio

n (%

)

Sintering temperature, T/K

Fig. 2 Magnitude of Cu deposition area on CF surface depending on the

fiber temperature.

1050 1100 1150 12000

5

10

15

Con

tact

are

a, A

/μm

2

Sintering temperature, T/K

Fig. 3 Size of contact area between Cu bridging particles and CFs

depending on the fabrication temperature.

Stre

ss,σ

/MPa

Strain (%)

0 5 10 15 20 25 300.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

1123K

1173K

1073K

Fig. 4 S-S curves of unidirectional CF preform by compression test.Fig. 5 Microstructure of CF preform fabricated at (a) 1123 K, (b) 1173 K

and (c) magnification of (b) after compression strain of 10%.

Preparation of Unidirectional Carbon Fiber Preform for Aluminium Matrix Composites 941

of both the initial maximun stress and energy absorptioncan be considered that the relatively wide bonding area ofCu particles between CFs, as shown in Fig. 3, delays thedeformation such like micro-buckling and kinking of CF bypreventing form fiber bending.

Figure 5 shows the cross section of CF preform aftercompression strain of 10%. In the microstructure of CFpreform fabricated at 1123 K (Fig. 3(a)), most of fiber did notfailed by compression above the initial maximum strengtheven if there were observed some of local fiber fractures. Inother words, the deformation behavior after initial maximumstrength can be mainly governed by fiber micro-bucklingmodel without fiber kinking model mostly. In the contrary,the CFs of the preform fabricated at 1173 K (Fig. 3(b))showed the catastrophic fracture profile after the compressiontest, even if the compression strength was higher than thatof the preform fabricated at 1123 K. The CF failed by theformation of fiber kinking mode which might cause by stressconcentration between enlarged Cu particles and fibers, asshown in Fig. 3(c). Furthermore, the enlarged Cu particlesare able to be obstacles to flow of molten Al when the moltenAl penetrates into the CF preform by LPI process. On theother hand, since both of fiber micro-buckling and kinkingby deformation of CF preform are possible to bring out theimperfect infiltration of molten Al, the infiltration pressurehave to be determined below the maximum compressionstrength without fiber deformation and fracture.

4. Conclusions

(1) The Cu particles between the CFs well transformed tofiber bridging and deposition on CFs at the fabricationtemperature above 1123 K.

(2) The maximum compression strength of the CF preformwas 1.02, 1.67 and 2.15 MPa depending on thefabrication temperature of 1073 K, 1123 K and 1173 K.

(3) The CF preform showed the fiber deformation such asfiber micro-buckling and fiber kinking after maximumcompression strength depending on the fabricationtemperature of 1123 K and 1173 K.

(4) The infiltration pressure for LPI process has to bedetermined under maximum compression strength.

Acknowledgements

This study was supported by the light metal educationalfoundation in Japan and Grant-in-Aid for Scientific Research(c) (21560771) from the Ministry of Education, Culture,Sports, Science and Technology, Japan and the hightechnological research project on ‘‘Research and Devel-opment Center for Advanced Composite Materials’’ ofDoshisha University and the Ministry of Education, Culture,Sports, Science and Technology, Japan.

REFERENCES

1) C. Zweben: JOM 50 (1998) 47–51.

2) S. Jin: JOM 50 (1998) 46.

3) C. Zweben: Proc. 4th Int. Conf. on Solid State Lighting, ed. by I. T.

Ferguson, N. Narendran, S. P. DenBaars and J. C. Carrano 5530, (SPIE,

Bellingham, WA, 2004) pp. 194–206.

4) S. Azzopardi, A. Benmansour, M. Ishiko and E. Woirgard: Micro-

electron Reliab. 45 (2005) 1700–1705.

5) S. Januszewski, M. Kocizewska-Szczerbik and H. Swiatek: Micro-

electron Reliab. 38 (1998) 1325–1330.

6) M. Ciappa, W. Fichtner, T. Kojima, Y. Yamada and Y. Nishibe:

Microelectron Reliab. 45 (2005) 1694–1699.

7) L. Dupont, S. Lefebvre, M. Bouaroudj, Z. Khatir and J. C. Faugiers:

Microelectron Reliab. 47 (2007) 1767–1772.

8) J. M. Molina, R. Prieto, J. Narciso and E. Louis: Scr. Mater. 60 (2009)

582–585.

9) K. Hanada, K. Matsuzaki and T. Sano: J. Mater. Proc. Tech. 153–154

(2004) 514–518.

10) T. W. Clyne and J. F. Mason: Metall. Trans. A 18 (1987) 1519–1530.

11) E. Carreno-Morelli, T. Cutard, R. Schaller and C. Bonjour: Mater. Sci.

Eng. A 251 (1998) 48–57.

12) Y. B. Choi, G. Sasaki, K. Matsugi and O. Yanagisawa: Mater. Trans. 46

(2005) 2156–2158.

13) Y. B. Choi, K. Matsugi, G. Sasaki, K. Arita and O. Yanagisawa: Mater.

Trans. 47 (2006) 1227–1231.

14) U. Anselmi-Tamburini, S. Gennari, J. E. Garay and Z. A. Munir: Mater.

Sci. Eng. A 394 (2005) 139–148.

15) K. Matsugi, H. Kuramoto, O. Yanagisawa and M. Kiritani: Mater. Sci.

Eng. A 354 (2003) 234–242.

16) S. Abraham, B. C. Pai and K. G. Satyanarayama: J. Mater. Sci. 27

(1992) 3479–3486.

17) I. Jeon and T. Asahina: Acta Mater. 53 (2005) 3415–3423.

18) P. Berbinau, C. Soutis and I. A. Guz: Compos. Sci. 59 (1999) 1451–

1455.

19) A. Jumahat, C. Soutis, F. R. Jones and A. Hodzic: Compos. Struct. 92

(2010) 295–305.

942 M. Lee, Y. Choi, K. Sugio, K. Matsugi and G. Sasaki