dynamic e ect on strength in sic /sic · pdf file ·...

International Journal of Impact Engineering 25 (2001) 29}40

Dynamic e!ect on strength in SiC&/SiC

.composite

M. Futakawa!,*, Y Tanabe", T. Wakui", H. Kogawa!, R. Hino!, M. Eto!

!Japan Atomic Energy Research Institute, Tokai-mura, Naka-gun, Ibaraki-ken 319-11, Japan"Niigata University, Ikarashi Ni-nocho, Niigata-shi, Niigata-ken 950-2195, Japan

Received 24 February 1999; received in revised form 20 April 2000

Abstract

Loading rate dependence of mechanical properties of SiC "bre-reinforced SiC composites (SiC&/SiC

.) has

been experimentally investigated as to the "bre volume fraction and coating materials for SiC "bre. Thecomposites consisting of monolithic SiC and SiC "bre (Hi-Nicalon) coated with Boron-Nitride (BN) orCarbon (C) with "bre volume fractions of 20, 30 and 40% were fabricated by polymer in"ltration}pyrolysis(PIP) process. The stress}strain response and strength were measured in tension over a wide range of strainrate,10~4&200 s~1. It was shown that the higher volume fraction, the larger tensile strength regardless ofthe kind of coating and strain rate. The interface friction stress evaluated by the "bre pullout lengththat is measured through microscopic observations of fractured specimens is larger in dynamic loading thanin static loading. The BN-coated "bre gave the composite superior tensile strength to the C-coated"bre. This trend results from the variety of the interface friction stress associated with the coatingthickness. ( 2000 Elsevier Science Ltd. All rights reserved.

Keywords: Ceramic matrix composite; Fibre coating material; Mechanical properties; Strain rate; Fibre volume fraction

1. Introduction

Because of its advantages such as high-temperature strength, high corrosion resistance and lowradioactivity, SiC monolithic ceramics are prospective structural materials for nuclear reactorcomponents used under severe environments. However, there are several objections to theirutilization: its inherent brittleness with a wide variation of strength; engineering issues such asfabricability of large components and joining technology for the parts of components. Recently, an

*Corresponding author. Tel.: #81-292-82-5363; fax: #81-292-82-6489.E-mail address: [email protected] (M. Futakawa).

0734-743X/01/$ - see front matter ( 2000 Elsevier Science Ltd. All rights reserved.PII: S 0 7 3 4 - 7 4 3 X ( 0 0 ) 0 0 0 3 6 - 1

e!ort has been made to overcome the brittleness by its reinforcement with continuous ceramic"bres, referred to as continuous "bre-reinforced ceramic matrix composites (CMCs), and SiC"bre-reinforced SiC composites (SiC

&/SiC

.) have been developed. The CMCs have generally

superior mechanical properties to monolithic ceramics primarily due to the toughening mecha-nisms involving matrix microcracking and frictional, energy-absorbing "bre pull-out [1]. Theonset of these energy-dissipating mechanisms requires an adequately low interfacial strengthbetween "bres and matrix [2,3], and this requirement is usually achieved by the use of an interfacialcompliant material. Carbon (C) has been widely used as an interfacial material so far [4}6].Recently, Boron-Nitride (BN) has also been suggested as an alternative interfacial material invarious CMCs to carbon [7}10] since the e$ciency of C-coated interface is limited by its lowresistance to oxidation at high temperatures [11,12]. Assessment of mechanical properties of theseCMCs is the important key to development work of materials with more favourable properties aswell as con"rmation of their applicability to structural components. Most of the previous studies inthis "eld have been carried out on the failure and deformation behaviour under quasi-static loadingwithout concern about the strain rate e!ect although many structural applications requireresistance against dynamic loads inducing complicated deformation with various strain rates [13].Hence, in order to investigate the strain rate e!ect on the dynamic failure and deformationbehaviour of SiC

&/SiC

., which could be a!ected by the interface between the "bre and matrix and

"bre volume fraction, the tensile tests were carried out under very wide loading rate, includingimpact loading, using the specimens of SiC

&/SiC

.with di!erent coatings on the "bres.

2. Materials and methods

2.1. Materials

The SiC&/SiC

.composites used in the test were produced by polymer in"ltration}

pyrolysis (PIP) process [14]. The "bres to reinforce them bidirectionally (0/90) are the SiC "bre(Hi-Nicalon) coated with C (0.06lm thickness) or BN (0.4lm thickness) [15]. Their "brevolume fractions,<

&, were chosen to be 20, 30 and 40%. The average diameter of the "bres is about

14lm.Small sheet tension specimens were prepared for both low and high strain rate tests. The

specimens have the gauge section con"guration of 6 mm long, 3 mm wide and 2 mm thick, and the"bre bundles with the stacking sequence of (0/90) were observed at the mid-cross-section ofa specimen, as shown in Fig. 1. The grip regions of a specimen were "xed into parallel-sided slots ofthe threaded aluminium grips using epoxy adhesive.

2.2. Methods

2.2.1. Tensile deformation and strengthThe tensile tests were performed at room temperature using an Instron-type testing machine

for quasi-static load at strain rate of 10~4 s~1 and a split-Hopkinson pressure bar (SHPB)apparatus, originally developed by Nicholas [16], for dynamic loading at strain rate of approxim-ately 200 s~1.

30 M. Futakawa et al. / International Journal of Impact Engineering 25 (2001) 29}40

Fig. 1. Geometry of sheet tension specimen.

Fig. 2. Schematic illustration of tensile split-Hopkinson pressure bar apparatus.

For quasi-static tests, the strain measurement was made by an extensometer attached directly tothe specimen gauge section. For dynamic tests, the tensile SHPB apparatus was used as illustratedin Fig. 2. The apparatus consists basically of an air gun, a striker bar, two Hopkinson (or one inputand one output) pressure bars, and recording equipments. The sheet tension specimen withthreaded grip ends described above is attached to the two Hopkinson bars as depicted in Fig. 1.The principle of the SHPB technique and the detail of specimen attachment to the apparatus canbe found in our previous study [17]. Note that the collar in which the entire compressive pulsepassed through is unable to support any tensile loads because it is not fastened to the bars. Thisprior compressive loading may not introduce serious damages in the specimen because theabsolute value of the compressive stress passing through the specimen, which can be estimatedfrom the ratio of the cross-sectional area of the specimen to that of the collar, is less than 10% valueof the re#ected tensile stress.

M. Futakawa et al. / International Journal of Impact Engineering 25 (2001) 29}40 31

Based on the elementary theory of one-dimensional elastic wave propagation, the average straine, strain rate e5 and stress p in the specimen can be evaluated from the SHPB test records as

e"2

o0c0lP

t

0

(pI!p

T) dt, (1)

e5"2

o0c0l(p

I!p

T), (2)

p"S0S

pT, (3)

where pI

and pT

are the incident and transmitted stress pulses, respectively, c0"JE/o

0the

longitudinal elastic wave velocity in the Hopkinson bars, E Young's modulus of the Hopkinsonbars, o

0the density of the Hopkinson bars, l the gauge length of the specimen, t the time from the

start of the pulse, S0

the cross-sectional area of the Hopkinson bars, S the cross-sectional area of thespecimen. Eqs. (1)}(3) can be applied only when the relation p

I#p

R"p

Tis established, where

pR

is the re#ected stress pulse. Eliminating time t yields the nominal stress}strain curve for thespecimen at the strain rate given through Eq. (2).

Fracture surfaces of all broken specimens were observed using a scanning electron microscopeSEM.

2.2.2. Interface characterisationThe energy consumed during the failure of a composite material is due to various competing

failure mechanisms that operate during a fracture. These can be classi"ed into matrix fracture,"bre}matrix interface debonding, "bre pullout, and "bre fracture. In particular, in brittle matrixcomposites, "bre pullout resistance, i.e. debonding stress, post-debonding friction and "bre pulloutmay be important mechanisms of energy consumption during failure, which is strongly dependenton the interfacial condition, i.e. the types of coatings on "bres. The debonding interfacial strengthwas evaluated directly from debonding force in push-in test using microindentation technique [18].Indentation test was carried out by a Shimadzu DUH-200, using a diamond indentor which hasa cone tip with a top angle of 140.63 to establish the same depth to displaced volume ratio asVickers and Berkovich. The push load is quasi-statically imposed on the end of the "bre. When theload reaches to the critical load to cause the debonding at the interface, the step appears on theload}depth curve. The debonding force F

$was determined readily as the load casing the step on

the load}depth curve. The quasi-static debonding strength p$

is de"ned [19] as

p$"F

$/pR2, (4)

where R is an average radius of "bres.The sliding friction stress q, which can be directly measured from the microindentation tech-

nique, is limited in a quasi-static loading condition. There are the other methods to evaluate q,which are based on the average "bre pullout lengths on the fracture surface, the saturation matrixcrack spacing and the width of the hysteresis loop during loading and unloading [21]. Here, inorder to discuss on the di!erence of q between dynamic and static loading conditions, s was


Fig. 3. Typical stress}strain curve (BN coating with V&"20, 30, and 40%).

evaluated from the average "bre pullout length L after failure by using the following equation[21,22]:

q"j(m)p#R/¸, (5)

where j(m) is a pullout factor, m is a shape parameter for "bre strength distribution and q#

isa characteristic stress of a "bre.

3. Results and discussion

3.1. Tensile deformation and ultimate strength

Typical tensile stress}strain curves of the SiC&/BN/SiC

.under dynamic loading are shown in

Fig. 3. All stress}strain curves show non-linear characteristics, or convex upwards, independentlyof the volume fraction. The ultimate strength increases with the volume fraction. This tendency isalso observed in the static loading as well as for the SiC

&/C/SiC

.in the dynamic and static loading.

The ultimate strengths in the static and dynamic stress}strain curves for each composite are plottedas a function of strain rate in Fig. 4. It seems that the dynamic strengths at e5'10 are higher thanthe static strengths at e5"4]10~4 regardless of the coating and the "bre volume fraction and thatthe strength of the SiC

&/BN/SiC

.is higher than that of the SiC

&/C/SiC

.throughout experimental

conditions. In order to make clear the strain rate dependency of the strength, the results shownin Fig. 4 were replotted by normalizing the dynamic strength p

$with static strength p

4, p

$/p

4,

for each composite in Fig. 5. Fig. 5 also shows the bending strengths which are obtained onthe same composites [22]. The large scatter on the dynamic strength is observed in the C coatings.The scatter is much less for the BN coatings, which exhibit the strain rate dependency on the


Fig. 4. Strain rate dependency of tensile strength.

Fig. 5. Strain rate dependency of strength.


Fig. 6. Relationship between static strength and "bre volume fraction.

strength: the strength increases with the strain rate slightly. The tensile and bending static strengthsp4used for the normalized stresses are plotted as a function of volume fraction in Fig. 6. The larger

the "bre volume fraction, the higher the strength becomes regardless of the coating materials. Thistrend was theoretically explained as p

4J<

&p#

[20]. The strength is higher for the BN coatingindependently of the loading type, tensile or bending, and the "bre volume fraction.

3.2. Interfacial properties

Mechanical properties of a "bre-reinforced composite are largely governed by the stress-transfercapability of the "bre/matrix interface. The degree of capability is related to an interfacial shearstress which characterizes the combination of the stress necessary to debond the interface and thefrictional force developed at this interface. Fig. 7 shows the relationship, obtained by the indenta-tion technique, between debonding force F

$and "bre radius R for the SiC

&/BN/SiC

.and

SiC&/C/SiC

.composites. The measured "bre radius ranges from 4 to 11lm approximately. It is

clear that the debonding forces for each "bre radius in the C-coating is higher than those in theBN-coating. The debonding strengths estimated using Eq. (4) were about 3 GPa for the C-coatingand about 1 GPa for the BN-coating. These are higher than the value measured elsewhere,500MPa to 1 GPa [19]. This discrepancy seems to be induced by using a corn shaped indenter tip,which generates a radial stress on the interface, instead of a #at-bottomed diamond indenter.Nevertheless, the di!erence on the debonding strength due to the coating is clearly recognised andthe debonding strength of the BN-coating is about one-third of the C-coating.

Fig. 8 shows typical microscopic views of fracture surfaces of the SiC&/BN/SiC

.and the

SiC&/C/SiC

.with 30% "bre volume fraction. There is more appreciable "bre pullout in the

BN-coating than in the C-coating despite loading condition, dynamic or static. The "bre pulloutlength appears to be longer in static loading than in dynamic loading. Average "bre pullout lengths


Fig. 7. Relation between "bre radius and debonding force measured by indentation technique.

Fig. 8. Microscopic views of fracture surface at 30% "bre volume fraction.


Fig. 9. Fibre pullout length measured from SEM observation.

for each material and loading condition are plotted together in Fig. 9. It is con"rmed that thepullout length is strongly associated with the loading condition and the type of coatings asmentioned above.

Because the friction coe$cients are indicated to be about 0.1 for both the C and BN coatings atroom temperature [23], the internal friction stress can be given by Eq. (5) without considering thedi!erence of coating material. On the other hand, the interface friction stress is a!ected by theinterfacial stress "eld due to mainly thermomechanical mis"t between the matrix and "bre, whichis very dependent on the coating thickness [24]. So that, the interface friction stress could beobtained from Eq. (5) using the observed "bre pullout length, as assuming that j(m)+0.3 fora shape parameter with multiple matrix cracks in the SiC/SiC, m"5 [23], and a characteristicstress p

#"1.6GPa for Hi-Nicalon "bre [20]. Fig. 10 shows the relationship between the

interface friction stress and the interface thickness (0.06lm for C-coating and 0.4lm for BN-coating [15]). The interface friction stress is larger in dynamic loading than in static loadingand increases as decreasing the interface thickness. The latter trend coincides with the resultsobtained by Lara-Curzio [19] using composites with various values of C-coating thickness, asshown in Fig. 10.

The "bre sliding velocity should be higher at higher strain rate. Such condition would increasethe interface friction stress because the friction stress is considered to be proportional to a frictioncoe$cient, which increases with the "bre sliding velocity [25,26]. The toughening mechanism dueto the "bres could not be expected when the internal friction becomes too large for the "bres to pullout. The increase of tensile strength with the strain rate has been observed for monolithic SiC andAl

2O

3in the previous studies [27,28]. The same dynamic e!ect could contribute dominantly to an

increase in the strength of the matrix and "bre in composite at high strain rate region, although


Fig. 10. Relationship between interface coating-layer thickness and friction stress evaluated using Eq. (5).

toughening mechanism due to the "bre pullout cannot be expected under such condition becauseof the increase in the interface friction stress. This could be the reason why larger ultimate strengthbut shorter "bre pullout length were observed in both composites under dynamic loading.

4. Conclusions

The static and dynamic tests on the SiC&/SiC

.composite with various "bre volume fractions

and coatings on the "bre are carried out to investigate the dynamic e!ect on its strength anddeformation. The specimen with higher "bre volume fraction exhibited larger tensile strengthindependently of coating and strain rate. The BN-coating composite exhibits superior tensilestrength to the C-coating. The "bre pullout length was longer in the BN-coating composite than inthe C-coating one in the present experimental range. In each composite it was longer for higher"bre volume fraction and lower strain rate. The trend results from the variety of the interfacefriction stress that is associated with the coating thickness and the strain rate.

Acknowledgements

The authors would like to thank Messrs. T. Gohno and K. Kobayashi, Niigata University, fortheir assistance in these experiments.


References

[1] Evans AG, Marshall DB. The mechanical behavior of ceramic matrix composites. Acta Metall 1989;37:2567}83.[2] Cao HC, Bischo! E, Sbaizero O, Ruhle M, Evans AG, Marshall DB, Brennan JJ. E!ect of interfaces on the

properties of "ber-reinforced ceramics. J Am Ceram Soc 1990;73(6):1691}9.[3] Peres P, Anquez L, Jamet J. Micromechanical analysis of fracture in ceramic composites. Rev Phys Appl

1988;23:213}228 [in French].[4] Caputo AJ, Stinton DP, Lowden RA, Besmann TM. Fiber-reinforced SiC composites with improved mechanical

properties. Am Ceram Soc Bull 1987;66(2):268}72.[5] Thebault JG. Fabrication process of a composite material with a refractory reinforcement and a ceramic matrix and

structure elaborated with such a process. French Patent No. 2567874A1, July 20, 1984.[6] Brennan JJ. Interfacial characteristics of glass}ceramic matrix/SiC "ber composites. J Phys Colloq C5 1988;10(49)

Suppl:S791}809.[7] Rice RW. BN coating of ceramic "bers for ceramic "ber composites. US Patent No. 4642271, February 10, 1987.[8] Singh RN, Brun MK. E!ect of boron nitride coating on "ber/matrix interactions. Adv. Ceram. Mater.

1988;3(3):235}7.[9] Lowden RA, More KL. The e!ect of "ber coating on interfacial shear strength and the mechanical behavior of

ceramic composites. Mater Res Soc Symp Proc 1990;170:205}14.[10] Naslain R, Dugne O, Guette A, Sevely J, Robin-Brosse C, Rocher JP, Cotteret J. Boron nitride interphase in

ceramic matrix composites. J Am Ceram Soc 1991;74(10):2482}8.[11] Filipuzzi L, Camus G, Thebault J, Naslain R. E!ect of high temperature ageing treatments on the mechanical

behavior of unidirectional SiC/SiC "brous composites. In: Bentzen JJ et al., editors. Proceedings of the 11thInternational Symposium on Metallurgy and Materials Science, National Laboratory, Roskilde, Denmark, 1990.p. 283}89.

[12] Coyle TW, Chan HM, Deshmukh UV. E!ects of heat treatments in air on the microstructure and mechanicalbehavior of the "ber/matrix interface in an LAS glass}ceramic matrix-SiC "ber composite. In: Ishida H, editor.Proceedings of the Second International Conference on Composite Interfaces. New York: Elsevier Science, 1988.p. 489}501.

[13] Lankford J. The failure of "ber-reinforced ceramic}matrix composites under dynamic loading. JOM (1995)64}68.

[14] Mitsuno S, Imai Y, Ichikawa, H. Kagawa Y. Fabrication and thermal stability of SiC "ber reinforced SiCcomposites. Proceedings of the Fifth Japan International SAMPE, 1997. p. 1507}12.

[15] Takeda M, Imai Y, Ichikawa H, Kagawa Y, Iba H, Kakisawa H. Some mechanical properties of SiC(Hi-NicalonTM) "ber-reinforced SiC matrix NicaloceramTM composites. Ceram Engng Sci Proc1997;18(3):779}86.

[16] Nicholas T. Tensile testing of materials at high rates of strain. Exp Mech 1985;21:177}85.[17] Tanabe Y, Komiya S, Wakui T, Futakawa M, Kogawa H, Stockmann Y, Hino R. Dynamic mechanical behaviour

of SiC/SiC composites. ICONE-6, No. 6219, 1998.[18] Marshall DB. An indentation method for measuring matrix-"bre frictional stresses in ceramic composites.

J Am Ceram Soc 1984;67(12):C259.[19] Lara-Curzio E, Ferber MK, Lowden RA. The e!ect of "ber coating thickness on the interfacial properties of

a continuous "ber ceramic matrix composite. Ceram Eng Sci Proc 1994;15(5):989}95.[20] Evans AG, Domergue JM, Vagaggini E. Methodology for relating the tensile constitutive behavior of ce-

ramic}matrix composites to constituent properties. J Am Ceram Soc 1994;77(6):1425}35.[21] Curtin WA. Theory of mechanical properties of ceramic}matrix composites. J Am Ceram Soc 1991;74(11):

2837}45.[22] Stockmann YJ, Futakawa M, Komiyama S, Tanabe Y, Kogawa H, Hino R, Mitsuno S. E!ect of loading rate on the

strength of SiC&/SiC composites. Ceramic material systems with composite structures, Ceramic transaction vol. 99.

Westerville, OH: The American Ceramic Society, 1998. p. 357}69.[23] Evans AG, Zok FW. The physics and mechanics of "ber-reinforced brittle matrix composites. J Mater Sci

1994;29:3857}96.


[24] Mikata Y, Taya M. Stress "eld in a coated continuous "ber composite subjected to thermo-mechanical loadings.J Comp Mater 1985;19:554}78.

[25] Lankford J, Couque H, Nicholls A. E!ect of dynamic loading on tensile strength and failure mechanisms in a SiC"ber reinforced ceramic matrix composite. J Mater Sci 1992;27:930}6.

[26] Cranmer DC. Tribological properties of glass}ceramics. Com Am Ceram Soc 1984;C180}2.[27] Awaji H, Yamada T. Impact strength of ceramics. J Soc Mater Sci 1990;39(445):1457}61 ([in Japanese]).[28] Futakawa M, Kikuchi K, Tanabe Y, Muto Y. Dynamic e!ect on fatigue strength of brittle materials. J Euro Ceram

Soc 1997;17:1573}8.


dynamic e ect on strength in sic /sic · pdf file ·...

Documents