Fatigue-crack-propagation (FCP) tests were conducted on the powder metallurgy nickel-base superalloy KM4 at temperatures of 20 °C, 550 °C, and 650 °C. Two different heat treatments were investigated, one yielding a relatively coarse grain size of 55 µm and another yielding a fine grain size of 6 µm. Tests were conducted at 100 Hz and 1000 Hz and at load ratios between 0.3 and 0.7. In the Paris regime, trends observed at high frequencies for KM4 were identical to those observed by earlier investigators at lower frequencies: coarse grains, low load ratios, low temperatures, and higher frequencies generally resulted in lower crack-propagation rates. However, in contrast to the Paris-regime behavior, thresholds were a complicated function of microstructure, load ratio, temperature, and frequency, and the only variable that resulted in a consistent trend in threshold was the load ratio. For example, thresholds increased from 100 to 1000 Hz for the fine-grained material at 550 °C, but decreased with the same frequency variation at 650 °C. One reason for this complexity was a change to intergranular fracture in the fine-grained microstructure at 650 °C, which was beneficial for high-frequency thresholds. Higher load ratios and lower frequencies promoted intergranular fracture. However, not all of the complexity could be explained by changing fracture mechanisms. Scanning electron microscope (SEM) stereofractography was utilized to determine quantitative measures of fracture-surface roughness. The most useful quantitative measure was found to be the standard deviation of the fracture-surface height, which is a physically meaningful length parameter and which corresponded to about half the grain size during room-temperature fatigue at near-threshold ΔK levels. The roughness of the fracture surface was found to increase as the load ratio was increased for both microstructures. For the coarse-grained microstructure, there was a direct correlation between fracture-surface roughness and FCP threshold over the entire range of temperatures, frequencies, and load ratios. However, measurements of closure loads indicated that roughness-induced closure was not the sole reason for the varying FCP thresholds.
Read full abstract