W.Li, Z.H. Chen/Ceramics Intemational 35 (2009)747-753 the inter-shielding status of the large and small pores, reflecting loaded into the penetrometer, undergoing evacuation and the medias topological characters [12], while the latter is pressurization at the low pressure port and the high one referred to the ones caused by the variation in contact angles, respectively, the pressures range from 1 to 55,000 psia surface tensions during mercury advancing or retreating from (6.9 x 103 N 380 MPa). Usually, the mercury extrusion the samples, and can be eliminated by adjusting the extrusion begins as the pressure reaches the highest value and continues contact angles carefully [13]. The permanent mercury till the atmospheric pressure level (36 psia)reached. However, entrapment is believed to occur in the large pores shielded to make Hg withdraw from the samples as much as possible, a by the small ones, caused by the snap-off of mercury flow at the novel procedure was adopted. Firstly, the penetrometer wa pore throats when retreating from the chambers [14], which is extracted from the high-pressure port when the Ist extrusion more likely to happen if the chamber/throat size ratio reaches 6 ended, then the sample segments were taken out carefully, or higher [15. loaded into another new penetrometer, evacuated to a very low MIP is an indirect characterization and the interpretation of pressure (5.3 Pa), and were subjected to the next intrusion- the experimental data needs some theories and well-established extrusion cycle. The raw data of 2nd Mip were adjusted by the models. Moreover, to extract more useful porosity information, original sample mass. The contact angle of Hg on the specific operations are often applied during usual MIP specimens during intrusion was assumed to be 130, and its procedure, among which applying secondary intrusion-extru- surface tension was 0.485 N/m The equilibration time was 15 sion after the primary cycle is the most familiar. This both at intrusion and extrusion. The morphology of specimens application can help to estimate the contribution of the cross-section was characterized by scanning electron micro- shielded pores to mercury entrapment, evaluate the structures scope (JSM-5600LV, JEOL). The pore size and distribution was damage of the samples induced by high pressures during also characterized by the bubble point method according to the intrusion, and distinguish the continuous, accessible parts of all GB/T 5249-1985, GB/T 5250-1993 standards(PR China) According to the previous literature, during MIP procedure, the pp ying re-intrusion started immediately as soon as the pressures 3. Results and discussion reduced to a low level where the primary extrusion came to an end. However, to avoid dangers and damages upon apparatus, 3. 1. Capillary pressure curve the primary/lst extrusion can only be carried out in the high pressure ports, which determines the end pressure is higher than The primary and secondary MIP capillary pressure curves e atmospheric pressure, e.g,29 psia [18], or 250 psia [16]. 3D-CpSiC are plotted in Fig. 1. On the priNs gL slopes.Atthe trusion curve Obviously, these pressures are not low enough to let the there are several phases distinguished by different slopes. At the mercury retreat thoroughly, and mercury droplets will still beginning of 1st extrusion, Hg does not retreat from the sample retain in the large pores, which is immingled with the"real" until the pressure reduces to the value of 0.2 um-sized capillary, entrapment. thus the extrusion branch lies above the intrusion one. with ur preliminary work has shown that MIP is preferable to depressurization continuing, the extrusions deviation from INS for characterization the porosity of 3D-CrSic composites, intrusion becomes wider and wider, till the pressure reaches the because of the limited probing ranges and disability to macro- ambient level and retreating finishes. At the low capillary pores(>l um) of INS [19]. In this work, a novel secondary pressures below 20 um, the 2nd intrusion curve is well mercury intrusion-extrusion procedure was conducted to 3D- consistent with the former one, but drops behind at higher CSiC besides the primary, to investigate the actual porosity of the composites. By interpreting the differences between these two MIP cycles, supported by SEM and bubble point method ■一 First Cycle results, the pore geometry was analyzed and described 0.091-+Second Cycle 2. Materials and methods 0.07 The 3D-CfSiC specimens were produced by subjecting the 50.06 braided 3-dimensional carbon fiber fabrics(T300, Toray Inc. Japan)to some infiltration-pyrolysis cycles, using polycarbo- 20.05 silane (PCS) as the polymer precursor. To track the 3 0.04 microstructural evolution of Cr/SiC, specimens underwent various fabricating cycles before finish were also chosen for 0.03 0.02 All MIP measurements were carried out with micromeritics Autoporelll 9420. Prior to characterization, the 3D-CiSic specimens were sliced into segments about 4 cm x 4 cm to Fig. 1. Capillary pressure curves of 3D-C/SiC after primary and secondary expose the inner pores. After repeated impregnation with MIP cycles, respectively. The dash line shows the possible trend of the Ist ethanol, cleanout, and drying, some of these segments were extrusion plot if the pressures keep reducingthe inter-shielding status of the large and small pores, reflecting the media’s topological characters [12], while the latter is referred to the ones caused by the variation in contact angles, surface tensions during mercury advancing or retreating from the samples, and can be eliminated by adjusting the extrusion contact angles carefully [13]. The permanent mercury entrapment is believed to occur in the large pores shielded by the small ones, caused by the snap-off of mercury flow at the pore throats when retreating from the chambers [14], which is more likely to happen if the chamber/throat size ratio reaches 6 or higher [15]. MIP is an indirect characterization and the interpretation of the experimental data needs some theories and well-established models. Moreover, to extract more useful porosity information, specific operations are often applied during usual MIP procedure, among which applying secondary intrusion–extru￾sion after the primary cycle is the most familiar. This application can help to estimate the contribution of the shielded pores to mercury entrapment, evaluate the structures’ damage of the samples induced by high pressures during intrusion, and distinguish the continuous, accessible parts of all the pore network that is crucial to permeability [16–18]. According to the previous literature, during MIP procedure, the re-intrusion started immediately as soon as the pressures reduced to a low level where the primary extrusion came to an end. However, to avoid dangers and damages upon apparatus, the primary/1st extrusion can only be carried out in the high￾pressure ports, which determines the end pressure is higher than atmospheric pressure, e.g., 29 psia [18], or 250 psia [16]. Obviously, these pressures are not low enough to let the mercury retreat thoroughly, and mercury droplets will still retain in the large pores, which is immingled with the ‘‘real’’ entrapment. Our preliminary work has shown that MIP is preferable to INS for characterization the porosity of 3D-Cf/SiC composites, because of the limited probing ranges and disability to macro￾pores (>1 mm) of INS [19]. In this work, a novel secondary mercury intrusion–extrusion procedure was conducted to 3D￾Cf/SiC besides the primary, to investigate the actual porosity of the composites. By interpreting the differences between these two MIP cycles, supported by SEM and bubble point method results, the pore geometry was analyzed and described. 2. Materials and methods The 3D-Cf/SiC specimens were produced by subjecting the braided 3-dimensional carbon fiber fabrics (T300, Toray Inc., Japan) to some infiltration–pyrolysis cycles, using polycarbo￾silane (PCS) as the polymer precursor. To track the microstructural evolution of Cf/SiC, specimens underwent various fabricating cycles before finish were also chosen for characterization. All MIP measurements were carried out with Micromeritics AutoporeIII 9420. Prior to characterization, the 3D-Cf/SiC specimens were sliced into segments about 4 cm 4 cm to expose the inner pores. After repeated impregnation with ethanol, cleanout, and drying, some of these segments were loaded into the penetrometer, undergoing evacuation and pressurization at the low pressure port and the high one respectively, the pressures range from 1 to 55,000 psia (6.9 103  380 MPa). Usually, the mercury extrusion begins as the pressure reaches the highest value and continues till the atmospheric pressure level (36 psia) reached. However, to make Hg withdraw from the samples as much as possible, a novel procedure 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 to a very low pressure (5.3 Pa), and were subjected to the next intrusion– extrusion cycle. The raw data of 2nd MIP were adjusted by the original sample mass. The contact angle of Hg on the specimens during intrusion was assumed to be 1308, and its surface tension was 0.485 N/m. The equilibration time was 15 s both at intrusion and extrusion. The morphology of specimens cross-section was characterized by scanning electron micro￾scope (JSM-5600LV, JEOL). The pore size and distribution was also characterized by the bubble point method according to the GB/T 5249-1985, GB/T 5250-1993 standards (PR China), applying ethanol as the saturation liquid. 3. Results and discussion 3.1. Capillary pressure curve The primary and secondary MIP capillary pressure curves of 3D-Cf/SiC are plotted in Fig. 1. On the primary intrusion curve, there are several phases distinguished by different slopes. At the beginning of 1st extrusion, Hg does not retreat from the sample until the pressure reduces to the value of 0.2 mm-sized capillary, thus the extrusion branch lies above the intrusion one. With depressurization continuing, the extrusion’s deviation from intrusion becomes wider and wider, till the pressure reaches the ambient level and retreating finishes. At the low capillary pressures below 20 mm, the 2nd intrusion curve is well consistent with the former one, but drops behind at higher Fig. 1. Capillary pressure curves of 3D-Cf/SiC after primary and secondary MIP cycles, respectively. The dash line shows the possible trend of the 1st extrusion plot if the pressures keep reducing. 748 W. Li, Z.H. Chen / Ceramics International 35 (2009) 747–753