1200 3.0 乏1000 2.8 FPHPS 日800 C-SNC 巴2.6 60 SNF/SNC SNF/PHPS 200 22 0040.8121.62.0 Displacement (mm) 2.0 Fig 3. Stress-deflection curves of UD composites 100011001200130014001500 Firing Temperature('C PHPS Si c SNC tter Times(min) Fig. 4. AES depth profiles of coated SNF. 5. The variation of polysilazanes in pyrolysis. (a) truc density as a function of firing temperature.(b) XRD patterns of polysilazanes pyrolysed at 1350C surface of the composite, the fibre pull-out has would inherently w are similar in Si based com- an increase in density. provided similar profiles as above. These results Since both matrices are would suggest that the pulled out face is the inter- ponent and structure, densification(an increase in face between the coated surface and the matrix. It atomic density) would provide an increase in elas- is reasonable to speculate that coating the fibre tic modulus. The true density of fired SNC and would give adequate interface shear strength by PHPS are 2.5 and 2. 8 g/cm, respectively. there controlling the interactie fore, it is presumed that the elastic modulus(em) The properties of composites are different of pyrolysed PHPS was much larger than that of according to the kinds of matrix precursor. The pyrolysed SNC. If Er> Em, the stress in the fibre is flexural strength at room temperature of SNF/ greater than in the matrix, because the fibre bears PHPS and SNF/SNC are 649 MPa and the major part of the applied load. This effect 1049 MPa, respectively. This difference is due to results in an increase in the effectiveness of fibre the crystallization behaviour of the precursors. strcngth to compositc strength. The flexural Though PHPS is selected as a high impregnation strength of samples using SNC matrix has become precursor with low viscosity, this PHPS contains high as noted above. To neglect the influence of excess Si compared with the stoichiometric com- the amount of scatter in the Vr on the composite position of silicon nitride. The deviation from the strength, the apparent effectiveness of fibre heoretical Si/N ratio results in low crystallization strength to composite strength(R)is defined as temperature ( 1200C). On the other hand, follows SNCs crystallization temperature is above R=Oe/(orVrA) 1.c, because snc does not contained excess Si, and its pyrolysis product has a stoichiometric where oc and or are the strength of composites and composition of a complex silicon nitride and sili- fibres, and a is the ratio of the fibre which is con carbide. The true density and X-ray diffraction aligned to the orientation of the tensile stress. For atterns of pyrolysed polymers are presented in UD and 2D composites, A is assumed to be 1.0 and Fig. 5. In preceramic poly le cry 0.5, ively182 H. Morozumi et al. 0 0.4 0.8 1.2 1.6 2.0 Displacement (mm) Fig. 3. Stresswleflection curves of UD composites Sputter Times (min) Fig. 4. AES depth profiles of coated SNF. surface of the composite, the fibre pull-out has provided similar profiles as above. These results would suggest that the pulled out face is the inter￾face between the coated surface and the matrix. It is reasonable to speculate that coating the fibre would give adequate interface shear strength by controlling the interaction. The properties of composites are different according to the kinds of matrix precursor. The flexural strength at room temperature of SNF/ PHPS and SNF/SNC are 649 MPa and 1049 MPa, respectively. This difference is due to the crystallization behaviour of the precursors. Though PHPS is selected as a high impregnation precursor with low viscosity, this PHPS contains excess Si compared with the stoichiometric com￾position of silicon nitride. The deviation from the theoretical Si/N ratio results in low crystallization temperature (M 1200°C). On the other hand, SNC’s crystallization temperature is above 15OO”C, because SNC does not contained excess Si, and its pyrolysis product has a stoichiometric composition of a complex silicon nitride and sili￾con carbide. The true density and X-ray diffraction patterns of pyrolysed polymers are presented in Fig. 5. In preceramic polymer, the crystallization 3.0 2.01 * I , , , I ! , , 1 1000 1100 1200 1300 1400 1500 Firing Temperature (“C) 10 20 30 40 50 60 70 26 (deg.) Fig. 5. The variation of polysilazanes in pyrolysis. (a) Plot of true density as a function of firing temperature. (b) XRD patterns of polysilazanes pyrolysed at 1350°C. would inherently cause an increase in density. Since both matrices are similar in Si based com￾ponent and structure, densification (an increase in atomic density) would provide an increase in elas￾tic modulus. The true density of fired SNC and PHPS are 2.5 and 2.8 g/cm3, respectively. There￾fore, it is presumed that the elastic modulus (E,) of pyrolysed PHPS was much larger than that of pyrolysed SNC. If Er> E,,.,, the stress in the fibre is greater than in the matrix, because the fibre bears the major part of the applied load. This effect results in an increase in the effectiveness of fibre strength to composite strength. The flexural strength of samples using SNC matrix has become high as noted above. To neglect the influence of the amount of scatter in the Vf on the composite strength, the apparent effectiveness of fibre strength to composite strength (R) is defined as follows: R = a,l(arvfA) (3) where o, and of are the strength of composites and fibres, and A is the ratio of the fibre which is aligned to the orientation of the tensile stress. For UD and 2D composites, A is assumed to be 1 .O and 0.5, respectively