Elevated temperature effects on R-curve behavior in alumina ceramics

• Main Author: Webb, James Ernest
• Language: ENG
• Published: ScholarWorks@UMass Amherst 1995
• Subjects: Materials science
• Online Access: https://scholarworks.umass.edu/dissertations/AAI9541165

I. Crack wake bridges were studied in glassy alumina s-DCB specimens at elevated temperatures. Direct observation of fracture surfaces in unmodified s-DCB specimens revealed the sizes and shapes of the high temperature bridges in the crack wake relative to the position of the crack front. Similar fracture surface observations made on s-DCB specimens modified to include a rear notch wedge, allowing the measurement of the high temperature bridging forces, were used to analyze bridge behavior in terms of viscous and power law creep models. The sizes and positions of the bridges were found to be highly stochastic but there was a clear trend toward decreased area bridged with increasing COD. Simple uniaxial viscous or creeping columns were inadequate to explain the load supported by the bridges. II. R-curve behavior and slow crack growth were studied in 99.5% alumina chevron notched short bar specimens at temperatures up to $1200\sp\circ\rm C.$ Constant loading rate tests measured toughness as a function of crack length at various loading rates. Constant load tests measured subcritical crack growth as a function of time. The intrinsic toughness decreased with increasing temperature while the bridging contribution to the R-curve remained independent of temperature. The effect of loading rate on R-curve tests was masked by experimental scatter at all temperatures except $1200\sp\circ\rm C.$ Subcritical crack growth tests showed high n-values at 700 and $1000\sp\circ\rm C,$ indicative of region III type behavior and an n-value indicative of region I at $1200\sp\circ\rm C.$ Predictions using the obtained subcritical crack growth parameters showed a small decrease in the level of the R-curve with decreasing loading rate at 700 and $1000\sp\circ\rm C$ that was within the range of the experimental scatter. A discrepancy between the R-curve predictions and the data at $1200\sp\circ\rm C$ was attributed to a rate dependence of the bridging behavior that was not considered in the analysis.