W.Li, Z.H. Chen/Ceramics Intemational 35 (2009)747-753 adopted. Firstly, the penetrometer was extracted from the high- CWVADXY pressure port when the Ist extrusion ended, then the sample B segments were taken out carefully, loaded into another new penetrometer, evacuated the penetrometer to a very low pressure(40 um Hg, 5.3 Pa)and were subjected to the next intrusion-extrusion cycle. Through this operation, Hg is able to withdraw from the samples as much as possible, eliminating the B temporarily retained Hg at atmospheric pressure from the real entrapment due to topological reasons. From the interpretation of the MiP results according to the models of hysteresis and Hg entrapment, accompanied by SEM analysis and bubble point Fig9. Physical model 3D-CASiC for fluids invasion. Symbol A represents the measurement, a model to describe the porosity of 3D-C sicis ter-bundle chamber exposed by slicing, Symbol B refers to that shielded by proposed. In 3D-C/ SiC, a 3-dimensional network of pores channel, as C pointed out. D is the channel that leads to the surface directly. E exists, which originates mainly from the architectures of the represents a group of micro-cracks which connects with channels and will be braided carbon fabrics. This network includes several main shielded by them during extrusion of MIP. components. The most important are the channels at the bundles' borders with Back to the PSD curve of 3D-CiSiC after 3 PIP cycles in skeleton of porosity, communicating with the bigger or smaller Fig. 6, the 3rd peak locating at 1.5-0. 1 um, which is not pores, and offering passages for fuids'flowing Another are the observed since the 7th PIP cycle, is supposed to be covered into inter-bundle chambers, with size of hundreds of microns, which the plot among 40.1 um, due to the pores becoming smaller. are the reservoirs of fluids. The last ones are the pores below 0 I um, which are attributed to micro-cracks, e.g., inside the 3. 4. Pore geometry of 3D-C /Sic bulk matrix or near the fiber-matrix interface. This model 3D-CUSiC Based on the interpretation of MiP data and analysis of SEM porosity. A more accurate model by means of network ages above, a rough depiction about the pore geometry of modeling or fractal geometry theory is under research at 3D-CfSiC is deduced, as the Fig 9 shown In 3D-CSiC, there present for simulating the fluids flow in the 3D-CSiC is a complicated 3-dimensional pore network, which originates mainly from the architectures of the braided carbon fabrics. Acknowledgement This network includes several main components. The important are the channels at the bundles' borders with sizes We would like to thank Mr. H. T. Zhang( China Building about 4-20 um(symbols C and D), making up the skeleton of Materials Academy, CBMA)and MrDWChang(Tsinghua porosity, connecting the bigger and smaller pores, and offer University) for their great help on MIP measurements and also passages for fluids flowing directly or via percolation valuable advices mechanism. Another are the inter-bundle chambers(symbols A and B), with size of hundreds of microns, and they are the reservoirs of fluids. The last one, from the MIP results, are pores References below 0.1 um, which are attributed to micro-cracks, e. g, inside the bulk matrix or near the fiber-matrix interface based on our [1] S.R. Qiao, S.M. Du, G.C. Ji, et al. Damage mechanism of 3D-C/Sic preliminary results, and they also connect with the channels mposite,J.Mechan. Strength 6(2004)307-312 [2] D.P.H. Hasselman. Effect of cracks on thermal conductivity, J. Compos. Certainly, this model is only a simulation of the real status, and need further improvement and correction, for it neglects the [3] S.M. Dong,YKatoh, AKohyama, et al.Microstructural evolution and regularity of pores'arrangement in the fabrics, and also omits mechanical performances of SiC/SiC composites by polymer impreg the intra-bundle pores which are not distinguishable by MIP, ation/microwave pyrolysis(PIMP) process, Ceram. Int. 28(2002) nd so on. However, our research provides some hints on the 99-905 [4] w. Krenkel, Microstructure tailoring of C/C-SiC composite, Ceram. Eng insights into the microstructure of 3D-CfSiC composites. Sci.Proc.24(2003)471-476 5] C.A.L. Leon, New perspectives in mercury porosimetry, Adv. Colloid 4. Summary Intert.Sci.76-77(1998)341-372 [6] L B. Liu, X.H. Wang, Fractal analysis of bentonite porosity using 3D-C/SiC fabricated by PIP is a kind of porous material, 7 S. Suyama, T.Kameda, Y Itoh, Evaluation of microstructure for SiC/SiC because of its specific processing. MIP is an effective composites, Int J Mater. Prod. Tech. 16(2001)232-238 characterization method for porous media, nevertheless the [8] S Suyama, Y Itoh, Evaluation of microstructure for SiC/SiC composites porosity information from the raw data of MiP deviates from g mercury intrusion method, Ceram. Eng. Sci. Proc. 20(1999)1 the actual cases, for the over-simplified model applied. To I3 terpret MIP data accurately and gain more reliable [9] E.w. Washburn, The dynamics of capillary flow, Phys. Rev. 17(1921) 273-283. description, additional theories and operations are needed. In [10) E. Moro, H. Bohni, Ink-bottle effect in mercury intrusion porosimetry of his work, a novel secondary intrusion-extrusion cycle was ement-based materials, J Colloid Interf. Sci 246(2002)135-149Back to the PSD curve of 3D-Cf/SiC after 3 PIP cycles in Fig. 6, the 3rd peak locating at 1.5–0.1 mm, which is not observed since the 7th PIP cycle, is supposed to be covered into the plot among 4–0.1 mm, due to the pores becoming smaller. 3.4. Pore geometry of 3D-Cf /SiC Based on the interpretation of MIP data and analysis of SEM images above, a rough depiction about the pore geometry of 3D-Cf/SiC is deduced, as the Fig. 9 shown. In 3D-Cf/SiC, there is a complicated 3-dimensional pore network, which originates mainly from the architectures of the braided carbon fabrics. This network includes several main components. The most important are the channels at the bundles’ borders with sizes about 4–20 mm (symbols C and D), making up the skeleton of porosity, connecting the bigger and smaller pores, and offer passages for fluids’ flowing directly or via percolation mechanism. Another are the inter-bundle chambers (symbols A and B), with size of hundreds of microns, and they are the reservoirs of fluids. The last one, from the MIP results, are pores below 0.1 mm, which are attributed to micro-cracks, e.g., inside the bulk matrix or near the fiber–matrix interface based on our preliminary results, and they also connect with the channels. Certainly, this model is only a simulation of the real status, and need further improvement and correction, for it neglects the regularity of pores’ arrangement in the fabrics, and also omits the intra-bundle pores which are not distinguishable by MIP, and so on. However, our research provides some hints on the insights into the microstructure of 3D-Cf/SiC composites. 4. Summary 3D-Cf/SiC fabricated by PIP is a kind of porous material, because of its specific processing. MIP is an effective characterization method for porous media, nevertheless the porosity information from the raw data of MIP deviates from the actual cases, for the over-simplified model applied. To interpret MIP data accurately and gain more reliable description, additional theories and operations are needed. In this work, a novel secondary intrusion–extrusion cycle was adopted. Firstly, the penetrometer was extracted from the high￾pressure port when the 1st extrusion ended, then the sample segments were taken out carefully, loaded into another new penetrometer, evacuated the penetrometer to a very low pressure (40 mm Hg, 5.3 Pa) and were subjected to the next intrusion–extrusion cycle. Through this operation, Hg is able to withdraw from the samples as much as possible, eliminating the temporarily retained Hg at atmospheric pressure from the real entrapment due to topological reasons. From the interpretation of the MIP results according to the models of hysteresis and Hg entrapment, accompanied by SEM analysis and bubble point measurement, a model to describe the porosity of 3D-Cf/SiC is proposed. In 3D-Cf/SiC, a 3-dimensional network of pores exists, which originates mainly from the architectures of the braided carbon fabrics. This network includes several main components. The most important are the channels at the bundles’ borders with sizes about 4–20 mm, making up the skeleton of porosity, communicating with the bigger or smaller pores, and offering passages for fluids’ flowing. Another are the inter-bundle chambers, with size of hundreds of microns, which are the reservoirs of fluids. The last ones are the pores below 0.1 mm, which are attributed to micro-cracks, e.g., inside the bulk matrix or near the fiber–matrix interface. This model provides preliminary understanding of the 3D-Cf/SiC’s porosity. A more accurate model by means of network modeling or fractal geometry theory is under research at present for simulating the fluids flow in the 3D-Cf/SiC. Acknowledgements We would like to thank Mr. H. T. Zhang (China Building Materials Academy, CBMA) and Mr. D. W. Chang (Tsinghua University) for their great help on MIP measurements and also valuable advices. References [1] S.R. Qiao, S.M. Du, G.C. Ji, et al., Damage mechanism of 3D-C/SiC composite, J. Mechan. Strength 6 (2004) 307–312. [2] D.P.H. Hasselman, Effect of cracks on thermal conductivity, J. Compos. Mater. 12 (1978) 403–407. [3] S.M. Dong, Y. Katoh, A. Kohyama, et al., Microstructural evolution and mechanical performances of SiC/SiC composites by polymer impreg￾nation/microwave pyrolysis (PIMP) process, Ceram. Int. 28 (2002) 899–905. [4] W. Krenkel, Microstructure tailoring of C/C-SiC composite, Ceram. Eng. Sci. Proc. 24 (2003) 471–476. [5] C.A.L. Leo´n, New perspectives in mercury porosimetry, Adv. Colloid Interf. Sci. 76–77 (1998) 341–372. [6] L.B. Liu, X.H. Wang, Fractal analysis of bentonite porosity using nitrogen adsorption isotherms, J. Chem. Eng. Chin. Univ. 17 (2003) 591–595. [7] S. Suyama, T. Kameda, Y. Itoh, Evaluation of microstructure for SiC/SiC composites, Int. J. Mater. Prod. Tech. 16 (2001) 232–238. [8] S. Suyama, Y. Itoh, Evaluation of microstructure for SiC/SiC composites using mercury intrusion method, Ceram. Eng. Sci. Proc. 20 (1999) 181– 189. [9] E.W. Washburn, The dynamics of capillary flow, Phys. Rev. 17 (1921) 273–283. [10] F. Moro, H. Bo¨hni, Ink-bottle effect in mercury intrusion porosimetry of cement-based materials, J. Colloid Interf. Sci 246 (2002) 135–149. Fig. 9. Physical model 3D-Cf/SiC for fluids invasion. Symbol A represents the inter-bundle chamber exposed by slicing, Symbol B refers to that shielded by channel, as C pointed out. D is the channel that leads to the surface directly. E represents a group of micro-cracks which connects with channels and will be shielded by them during extrusion of MIP. 752 W. Li, Z.H. Chen / Ceramics International 35 (2009) 747–753