ErRS Journal of the European Ceramic Society 21 (2001) 649-657 www.elsevier.com/locate/jeurceramsoc SiC-fibre reinforced glasses-electrical properties and their application Beate Fankhanel a, Eberhard Muller a,*, Ulrike Mosler b, Winfried Siegel b Freiberg University of Mining and Technology, Institute of Ceramic Materials, Gustay-Zeumner-Str. 3. 09596 Freiberg. Germany Received 30 March 2000; accepted 8 August 2000 Abstract Silicon carbide fibre reinforced glasses were investigated to create smart materials. The fibres were used not only as reinforcement ctrica I properties o the composite are not t only influe interface layer. Because of the electrical conductivity of the inlaid fibres and the fibre/matrix interface, the electrical properties of the composites have been used to detect the development of damage during mechanical testing. The development of damage due to fibre fracture and delamination under applied mechanical load in a unidirectionally reinforced specimen was monitored and micro deformation analysis was performed simultaneously Furthermore, local changes in temperature could be detected. Several mea- surements were taken to localize mechanical damage and temperature. C 2001 Elsevier Science Ltd. All rights reserved. Keywords: Composites; Electrical conductivity; Glass-matrix; SiC-fibres; Smart materials 1.Introduction introduction are associated with these methods. On the other hand, in some uses, already existing reinforcing In recent years, the role of online monitoring in the elements could operate as sensors. Among these meth- field of materials science has become more and more ods, there are already some by which electrical quan- important, especially in safety-sensitive sectors like tities were measured to obtain information about the aerospace, where composites play an important role. It "health"of a composite. 3-5 A requirement for this kind ld be useful to have methods fo met or accurate and rell- of smart composite is a large difference in electrical e self diagnosis in the materials to facilitate fracture properties between matrix and reinforcing fibres. Pre- prediction. ferably, the matrix is insulating and the fibres are elec- Often composites consist of materials with large dif- trically conductive. Often these so-called "smart"" ference in properties, such as electrical composites consist of carbon fibre reinforced polymers elastic modulus. Examples are composites containing and are used for real-time non-destructive evaluation of fibres as a reinforcement. Some histicated delamination and damage control during tensile and structures of these comp e to pro-fatigue testing. 6-14 But also fibre reinforced glasses and duce novel functions an glass-ceramics could be used as a self detecting compo- self diagnosis in the m site to check the degree of damage due to mechanical measure steadily and n e strength of a stress component could be realised by y known system was investil- special optical sensor gated regarding its electrical properties. The reinforcing whole smart compos fibres were used as sensors, detecting electric However, problems of mismatching and stress resistance as an indicator of mechanical strength, to create a fracture prediction technique. Furthermore, the Corresponding author. appearance of a warm-up due to external heating during E-mail address: mueller@ anw.ikw.tu-freiberg.de (E. Muiller). the application of a component should be detected, too. 0955-2219/01/- see front matter 2001 Elsevier Science Ltd. All rights reserved pii:s0955-2219(00)00240-5
!"#"$ %&' ( ) !"#"$ !" #$ %$$$& ' ( %$$$ ) " * *+ , ) - * * * * + . " ) * - / - - * -+ "- 0*1 * " "* * * + , "* * * * - * ) * * * - ) * * -+ * * + " * * ) 2 * * *+ %$$3 " 4 + ( " + *+ , *& "-& 5*1& & * * . - * * * * * * - - " ) * - * + . ) " * * + 6 * * ) "- * + 1* * * + * - * * " 77 88 * * *+ * ) * - "- * - 3 * ) * * *+% 9)" * ** ) * + 6 * - 1 * + (* * - * - ) : ) * * 7788 *+#; ( :* * * < ) *1 + = - *1 - "+ 6 77*88 * -* * " " * * +>3? * * * * +3; . - "- ) -* ) " + , ) * :+ * )* 1 * + $@;;%%3@0$30A * %$$3 " 4 + ( " + =..B $@;;%%3@C$$D$$%?$; E * - %3 C%$$3D >?@F>;G )))+"+*00H* + % , *I )+)+ + C+ D+
B. Fankhanel et al. Journal of the European Ceramic Society 21(2001)649-657 That means silicon carbide fibre reinforced glasses The preparation of all composites was made by a were investigated regarding their ability to inform about known slurry infiltration process described in detail he degree of impairment after and during mechanical elsewhere. 16, 17 The fibre bundles were infiltrated with a tress or about the component's present temperature slurry consisting of the glass powder and an organic binder. The prepregs were dried before densification The densification was performed by a hot-pressing pro- 2. Materials and experimental procedure cess under an inert atmosphere. In case of the specimens produced at Freiberg University, the hot-pressing 2. 1. Sample preparation temperature was 1100C SiC-fibre reinforced glasses were used for our investi- gations. Two types of specimens were chosen. At first, 2. 2. Measurements of electrical resistance vs mechanical simplified models of Sic-fibre reinforced glasses load as well as heat treatment produced at Freiberg University (for fibre content see Table 2). The glasses DURAN(borosilicate glass)and Afterwards the reinforced glasses were cut into rec- SUPREMAX(alumosilicate glass) from Schott Glas angularly shaped pieces. One part of the specimens was Germany) were selected and Tyranno TM-DIE08PX used to apply a three-point bending load or a pressure Sic-fibres from UBE Industries Ltd (Japan) were cho- load on the samples while simultaneously the electrical sen as reinforcing material. Later commercial specimens resistance was measured. The other part of the speci- known as FORTADUR from Schott Glas( Germany) mens was heat treated while the electrical resistance was were used. The commercial specimens were made of the measured glasses DURAN and 8252(alkali-earth alumosilicate Mechanical testing took place at room temperature glass) from Schott Glas(Germany) and of Tyranno using a universal testing machine(INSTRON Int. Ltd TM-SIC16PX SiC-fibres from UBE Industries Ltd. USA)while thermal treatments were performed in an (Japan).Table 1, Fig. 1) alumina tube furnace in air or later for local resolution Properties and composition of glass matrices and Sic-fibres Glass matrix SiC-fibre 8252 TM-DIEO8PX TM-SICl6PX 63 2.45 2900(10-6/K) 3.1 3.1 Tg(c) Youngs modulus(GPa) 187 Electrical resistivity(2 cm) Composition(wt o% SiO, 52 siO279.7 SiO, 59.9 Al,O3 22 B2O310.3 Al2O313.8 P2O38 Na,O 5.2 Cao 10.2 O2~12 MgO 7.5 Al2O33.1 Tie ao 7 B2O34.5 B2O32 Cao 0.8 Mgo 2.4 Bao 1.5 other 0.4 Table 2 Specimens and their application Matrix SiC-fibre Fibre volume Produced at Reported in SUPREMAX E08PX4.5±1% Freiberg University 3-point bending test 27, 28 DURANE E08PX4.5±1% Freiberg University Thermal treatment 31. this report DURANE EO8PX d Freiberg University hermal 29. this report 90° ocal resolution 8252 TM-SICl6PX40±5% 3-point bending test DURANE TM-SlCI6PX40±5% Schott glas 3-point bending test, 30. this report ocal resolution DURAN TM-SICl6PX40±5% this report TM-SICl6PX Schott glas Thermal treatment this
, * ) " - * ** * * 8 *+ -) ' '' ) " + ,) - * ) + ( * * ) "- C , %D+ , J!(K C D =!(L C* D * 5 C5* -D ) ,- ,J3$'=L * . 4 + CE D ) *+ 4 ** * ) 6!,(J! * 5 C5* -D ) + , ** * ) * J!(K '%;% C * D * 5 C5* -D ,- ,33>=L * . 4 + CE D+ C, 3 + 3D+ , * ) * - ) - )+3>3G , ) ) - ) + , ) + , ) * - *+ . * "- * ) 33$$+ -- + ( ) ) - + 6 * ) - * ) * - ) * + , * ) ) ) * + * * " * C.K,!6K . + 4 (D ) * * ) * * , % * 1 "* = ! ( ! =!(L ,J3$'=L ?+;3M "- # %G %' J!(K ,J3$'=L ?+;3M "- ,* * #3 J!(K ,J3$'=L + % "- ,) * C $F@$D ,* * %@ '%;% ,33>=L ?$;M 5 # %@ J!(K ,33>=L ?$;M 5 # #$ J!(K ,33>=L ?$;M 5 = '%;% ,33>=L ?$;M 5 ,* * , 3 = * * 5 *1 *1 J '%;% ,J3$'=L ,33>=L C0*# D %+;G %+%% %+># %+#G %+?; %$0#$$ C3$> 0ND ?+3 #+# ?+> #+3 #+3 C D G#$ ;%; G%; F F O 8 * C5=D 'G ># '3 3'% 3'G "- C *D F F F 3+@3$# 3G * C)+MD 6% ;% 6% G@+G 6% ;@+@ ;% (%6# %% %6# 3$+# (%6# 3#+' #% =%6# ' K%6 ;+% 6 3$+% 6%3% 6 G+; (%6# #+3 6 '+' ,? 6 G 6 $+@ %6# ?+; %6# % 6 $+' 6 %+? 6 3+; $+? >;$ . /0 1 2 % ' -) 3-!!)4 $5"6$#7
B. Fankhanel et al. Journal of the European Ceramic Society 21(2001)649-657 y using a flame and heating the specimens directly at fibres and matrix. Because the glass is an insulating pecific positions matrix, an assumption was made at first that the semi To detect the electrical resistance of the composites conducting SiC-fibres could affect the electrical proper the specimens were connected to copper electrodes. Sil- ties of the composite. However, the appearance of a ver paint was used for electrical contacts. The type of carbon layer at the fibre-matrix interface during hot connection depended on the kind of investigation, as pressing was found to have a much larger influence on shown in Fig. 2. A Keithley 2000 digital multimeter was the composites'electrical resistivity used for the electrical measurements The in-situ formation of such a carbon interface has been widely observed in glass and glass-ceramic matrix 2.3. Optical measurements vs mechanical deformation composites reinforced with SiC-fibres and has been reported by several authors 20-24 Usually the carbon Furthermore, a new method of a micro deformation interface between the fibre and the matrix provides a analysis, the so-called"grey scale correlation analysis, weak line and is responsible for the high toughness of was performed with help of the Chemnitzer Werk these composites because it allows crack deflection stoffmechanik GmbH. This method is based on imaging interfacial debonding and fibre pull-out. Such phenomena of specimens during their deformation. A series of pic- are of fundamental importance for the reinforcement of tures is recorded in the course of specimen deformation composites with brittle matrices and the positions of identical grey values in the images A scheme of such a carbon layer is given in Fig 3. It of different states of deformation - representing iden- shows the difference in structure between the existing tical points of the specimen- are correlated to give the carbon interface(turbostratic carbon) and a carbon vectors of movement. 8. 9 In our case, photos during the layer perfectly structured (graphite). The two-dimen mechanical testing of the specimens were taken. These sional honeycomb structure of turbostratic carbon con- photos were digitized and the grey scale pattern of digi- sists of uncorrelated individual carbon layers. Weak tized photos of different stress situations were corre- disorder results in stacking faults giving rise to a small lated. As a result, an optical impression of the increase in the interlayer distance. Therefore, the stack deformation of the specimen as well as figures of the ing of the individual carbon layers becomes uncorre- displacement as vector diagrams were obtained. From lated and turbostratic carbon occurs. The electronic these diagrams, bending lines could be calculated structure of turbostratic graphite, a zero gap semi- A summary of the specimens used and their conductor, is qualitatively different from that of ideal experimental application is given in Table 2. graphite, a semimetal with a small band overlap This carbon interface affects the electrical properties of the Sic-fibre/glass-matrix composites. This assump- 3. Results and discussion tion was supported by the values determined for the electrical resistivity of the single fibres as compared to 3.. Structural studies the composites. The used Sic-fibres were stable in their mechanical and electrical properties up to about A fundamental feature of Sic-fibre reinforced glasses 1200oC. 27,28 Table 3 shows the measured values of elec is the large difference in electrical properties between trical resistivities of the SiC-fibres and of the composite 静 Fig. 1. Examples of cross sections of investigated specimens.() SiC-fibre reinforced DURAN.glass matrix produced by Schott Glas(Germany) (b) SiC-fibre reinforced DURAN.glass matrix produced at Freiberg University
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
652 B. Fankhanel et al. Journal of the European Ceramic Society 21(2001)649-657 electrode bidirectionally reinforced specimen rodes a)3 b)3-point-bending test-local resolution of the electrical resistance at the specimens' surface electrode monitoringsimplified construction of a two-dimensionally d) Printed circuit board with a fixed speci (* R2 analogous at the back of the specin Fig. 2. Models of the measuring principles. (a)3-point-bending test. (b)3-point-bending test- local resolution of the electrical resistance at the pecimens surface. (c) Uniaxial compression load testing(R2 analogous at the back of the specimen). (d) Printed circuit board with a fixed speci- men(the fibre bundels were used as scanning channels 1-10 for the electrical resistance) idealized C-layer (ideal crystal structure of graphite) bre axis turbostratic carbon as it exists C-layer X≈50nm ig. 3. Scheme of the carbon layer at the fibre-matrix interface after hot-pressing
+ %+ * + CD # + CD # * 8 + CD 1 * C9% * D+ C D = ) 1 * C ) 3F3$ D+ + #+ * - F*1 + >;% . /0 1 2 % ' -) 3-!!)4 $5"6$#7
B. Fankhanel et al. Journal of the European Ceramic Society 21(2001)649-657 The electrical resistivity of the composite is much lower should arise the occasion to give warning before cracking than that of the Sic-fibres itself. Consequently there starts must be something else in the composite which is elec Fig. 5 shows this first part, until a first matrix crack ductive occurred, of such a relationship between applied load, Accordingly we assumed a carbon layer of a thickness deflection and electrical resistance to unidirectionally of about 50 nm to explain the observed decrease of the reinforced commercial composites. The value for the electrical resistance of the composite after hot pressing. electrical resistance was normalized to show more This estimation is in agreement with data from the lit- clearly its variation during mechanical testing. There- erature 20-24 Comparing the measured results with the fore, the initial value of the electrical resistance was calculated value of the electrical resistivity of the carbon 100%. From Fig. 5 it can be seen that at the beginning layer at the fibre-matrix interface(Table 3), it is clearly of the mechanical loading, the electrical resistance seen that the electrical properties of the composite are increased with increasing load. Though the change in affected by the carbon interface electrical resistance seemed to be not very significant, it It should be noted that in our case, the type of was o served in all the cases. By using a new method of matrix used here has no influence on the electrical a micro deformation analysis described above we esistivity of the composites. However, an increase in established that this first increase in resistance is just the isotropy of the fibres architecture in the composite, influenced by destroying the fibre/matrix interface, con from unidirectional via perpendicular to random orien- sisting of the electrically conductive carbon layer, and tated. leads to an increase in ity of not by matrix cracking with fibre pull-out and fibre the whole composite as we have published elsewhere. 29 3.2. Mechanical testing 33 In case of specimens on which mechanical load pplied, it was possible to monitor the degree of nsisance impairment during mechanical stress based on changes 2 in electrical resistance, because a change of the load always leads to a change in electrical resistance. At first, the unidirectionally reinforced specimens were investi- gated measuring the electrical resistance parallel to the reinforcing direction by applying a three-point bending load(Fig 2a). a distinction could be made between fibre fractures and delamination connected with fibre pull-out before the final macroscopic failure of the component deflection(mm) occurred.27,28 Fig. 4 shows typical results for this kind of Fig. 4. Load-deflection curve and electrical resistance when applying measurement. It is clear to be seen that the bend ts 3-point bending load to a specimen from Schott Glas(Germany), cf. the resistance curve were connected to fibre fractures. In Fig 2a for the measuring principle this case the load appearing on the specimen drops dra matically. On the other hand a steady rise in resistance is connected with delamination and fibre pull-out. 060120180240300360420 However the behavior of the electrical resistance 1.004 fore a first matrix crack occurred is much more important with regard to fracture prediction, since it ole 3 Comparison of electrical resistivities Material Electrical direction 0.998 (9 h urve and electrical resistance when applying C-layer at the fibre- matrix interface 3-point bendin a specimen from Schott Glas(Germany) (assumptions: 50 nm thick, parallel connection of resistances resistance was d to be 100%. cf. Fig. 2a for the measuring princip
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