G E. Youngblood et al./Composites Science and Technology 62(2002)1127-1139 lars m about -0.7 0×104s/m2.The S-term in Eq.(10) is related to the strength of the Umklapp or thermal phonon-phonon scattering resis- h=5000w(cm2K) tance. For high-purity CVD-SiC, several observations in this laboratory have indicated that Sa48+5 s/(m2K) approximately independent of irradiation defect con- W/(cm?K centrations [36]. Although these values may not be quite correct for the Tyranno SA fiber or the CvI-SiC matrix, they are representative enough to predict the correct temperature effects on Keff In Fig. 10(a), the analytic h a 50 W/(cm, K) solutions predicted for the hypothetical Tyranno SA, CVI-SiC composite are replotted as a function of tem- perature forf=0.4 and selected values h=5000, 500 and Temperature (C A well-adhered PyC fiber coating with thickness 0.2 um and thermal conductivity 10 w/m K) would have an equivalent haKa/t= 5000 W/(cm- K), a value that 空(b) represents good f/m thermal coupling. For such a case, Kef(T) ranges from a value of about 29 W/(m K)at 27C up to an apparent maximum of 34 w/(m K)at about 200C, and back down to 26 w/(m K)at 1000C The maximum in Kef at about 200C, observed for all h-values, is related to the rather steep positive tempera ture dependence of the SiC heat capacity for T<600C The two lower h-values (500 and 50 w/(cm-K) repre sent moderate and poor fim thermal coupling, respec tively.Forh≈500W/cm2K),R≈ I and Keff≈Km≈20 W/(m K)for all temperatures. Finally, for h=50 W/ h=50 (cm K), Kefr<12 W/(m K) in the 27<T<1000 oC range. This hypothetical case again emphasizes the 1000 00 mportant role that the interfacial conductance plays in determining Kef(h) when h-values fall in the transition ature resisted b y dm al for ac iai y irma ased land tio af To account fo o account for degradation of Kef(t)due to irra irradiated hypothetical 2D-Tyranno SA / CVI-SiC composite with diation damage, the K(T)and Km(T)-values were h=50.,500or5000W/cm21 lected data values: a=5 um,f=0. 4, Kr(27C)=65w/(m K)and Km(27C)=20 w/(m K). separately adjusted for expected temperature and dose effects prior to substituting them into Eq(1). The ther- mal conductivity degradation for irradiated SiC depends primarily on the radiation temperature when Thermal diffusivity data for unirradiated Tyranno Sa doses are in excess of saturation, i.e. >1 dpa for Sic Sic fiber and for a CVI-SiC matrix were generated and [32]. To estimate the degradation effect of irradiation fit to an expression with linear temperature dependence: quadratic curves were separately fit to the temperature- dependent thermal conductivity data given in reference 32] for high-purity CVD-SiC and SiCe/CVI-Sic irra l/a=P+S(1) (10) diated above saturation. For 27<T< 1000 oC. the equations K()rad=[0. 10+3.5x10-T2K,(T)and for the temperature range 300-1300 K starting with Km(T)rad=0.20+3.6x10-TKm(T were used to known values of a at 300 K. Then KiT)-and Km(T)- represent the temperature dependence of Tyranno SA values were calculated from the generated a-data using fiber and CVI-SiC matrix irradiated above saturation Eq( 8)for both cases. In Eq. (10), the P-term depends doses, respectively upon the initial purity of the SiC as well as the radiation In Fig. 10(b), the analytic solutions predicted for the concentrations. For instance, P=-(0.72+0.08) hypothetical irradiated Tyranno SA/CVI-SiC composite s/m2 for high-purity CVD-SiC as determined by are replotted as a function of temperature after the ke Senor et al. [36], while CVd-Sic irradiated above and Km values were adjusted for the effects of irradia saturation doses exhibited a P-value of about 10x10+s/ tion. In contrast to the unirradiated condition, for each m[36]. Thus, the P-term for SiC may cover a relatively selected h-value the predicted Keft-values for irradiatedThermal diffusivity data for unirradiated Tyranno SA SiC fiber and for a CVI-SiC matrix were generated and fit to an expression with linear temperature dependence: 1= ¼ P þ S TðÞ ð10Þ for the temperature range 300–1300 K starting with known values of at 300 K. Then Kf(T)- and Km(T)- values were calculated from the generated a-data using Eq. (8) for both cases. In Eq. (10),the P-term depends upon the initial purity of the SiC as well as the radiation defect concentrations. For instance, P=-(0.720.08) 104 s/m2 for high-purity CVD-SiC as determined by Senor et al. [36],while CVD-SiC irradiated above saturation doses exhibited a P-value of about 10 104 s/ m2 [36]. Thus,the P-term for SiC may cover a relatively large range from about 0.7 to +10 104 s/m2 . The S-term in Eq. (10) is related to the strength of the Umklapp or thermal phonon-phonon scattering resistance. For high-purity CVD-SiC,several observations in this laboratory have indicated that S 485 s/(m2 K) approximately independent of irradiation defect concentrations [36]. Although these values may not be quite correct for the Tyranno SA fiber or the CVI-SiC matrix, they are representative enough to predict the correct temperature effects on Keff. In Fig. 10(a),the analytic solutions predicted for the hypothetical Tyranno SA/ CVI-SiC composite are replotted as a function of temperature for f=0.4 and selected values h=5000,500 and 50 W/(cm2 K). A well-adhered PyC fiber coating with thickness 0.2 mm and thermal conductivity 10 W/(m K) would have an equivalent h Kd/t=5000 W/(cm2 K),a value that represents good f/m thermal coupling. For such a case, Keff (T ) ranges from a value of about 29 W/(m K) at 27 C up to an apparent maximum of 34 W/(m K) at about 200 C,and back down to 26 W/(m K) at 1000 C. The maximum in Keff at about 200 C,observed for all h-values,is related to the rather steep positive temperature dependence of the SiC heat capacity for T<600 C. The two lower h-values (500 and 50 W/(cm2 K) represent moderate and poor f/m thermal coupling,respectively. For h 500 W/(cm2 K), R 1 and Keff Km 20 W/(m K) for all temperatures. Finally,for h=50 W/ (cm2 K), Keff<12 W/(m K) in the 27<T<1000 C range. This hypothetical case again emphasizes the important role that the interfacial conductance plays in determining Keff(h) when h-values fall in the transition range (see Fig. 9). To account for degradation of Keff (T) due to irradiation damage,the Kf (T )- and Km(T )-values were separately adjusted for expected temperature and dose effects prior to substituting them into Eq. (1). The thermal conductivity degradation for irradiated SiC depends primarily on the radiation temperature when doses are in excess of saturation,i.e. 51 dpa for SiC [32]. To estimate the degradation effect of irradiation, quadratic curves were separately fit to the temperaturedependent thermal conductivity data given in reference [32] for high-purity CVD-SiC and SiCf/CVI-SiC irradiated above saturation. For 27<T<1000 C,the equations Kf (T)rad=[0.10+3.5 107 T2 ]Kf (T ) and Km(T )rad=[0.20+3.6 107 T2 ]Km(T) were used to represent the temperature dependence of Tyranno SA fiber and CVI-SiC matrix irradiated above saturation doses,respectively. In Fig. 10(b),the analytic solutions predicted for the hypothetical irradiated Tyranno SA/CVI-SiC composite are replotted as a function of temperature after the Kfand Km-values were adjusted for the effects of irradiation. In contrast to the unirradiated condition,for each selected h-value the predicted Keff-values for irradiated Fig. 10. The transverse thermal conductivity, Keff,as a function of temperature predicted by Eq. (1) for (a) an unirradiated and (b) an irradiated hypothetical 2D-Tyranno SATM/CVI-SiC composite with h=50,500 or 5000 W/(cm2 K). Selected data values: a=5 mm, f=0.4, Kf (27 C)=65 W/(m K) and Km (27 C)=20 W/(m K). G.E. Youngbloodet al. / Composites Science andTechnology 62 (2002) 1127–1139 1137