Reactive infiltration processing and compression creep of NiAl and NiAl composites

Abstract

Reactive infiltration processing of bulk and composite NiAl was investigated with powder and wire preforms of nickel. Inhomogeneous microstructures were often obtained with powder preforms because their high surface-to-volume ratio, low permeability, and irregular infiltration paths lead to simultaneous infiltration and reaction. Homogenous NiAl could be obtained with nickel-wire preforms which had a lower surface-to-volume ratio, higher permeability, and regular infiltration paths, because infiltration was completed before the onset of reaction. Composites with continuous tungsten (W) and sapphire fibers were also successfully fabricated by reactive infiltration, while composites with molybdenum particulates and short-fibers showed significant dissolution in NiAl. The high-temperature uni-axial compression creep behavior of uni-directionally reinforced continuous fiber composite materials was investigated using NiAl-W as a model system for the case where both the NiAl matrix and the W fiber underwent plastic deformation by creep. The creep behavior of the constituents NiAl and W and NiAl composites reinforced with 5-20 volume % W was characterized at 1025 °C and 715 °C. At 1025°C, the NiAl-W composites exhibited three stage creep behavior with distinct primary, secondary, and tertiary creep, wherein the composite creep-rate decreased monotonically, remained constant, and increased rapidly, respectively. At 715C, the NiAl-W composites exhibited insignificant primary and tertiary creep but significant secondary creep. Microstructurally, primary and secondary creep were characterized by pure uni-axial compression of W fibers while brooming, bulging, buckling, and kinking were four fiber deformation modes that contributed to tertiary creep. The composite primary creep was modeled by solving for transient stress-states while loads transferred from the weaker phase (matrix) to the stronger phase (fiber) as the composite transitioned from the elastic state present at loading to steady-states attained at later times. The effects of primary creep of the constituents on the primary creep of the composite were also captured. Composite primary creep strains were predicted to be significant at high applied composite stresses and for high fiber volume fraction composites, while the composite primary time was uniquely related to the composite steady-state creep-rate by a power-law at a given temperature and for the stress range investigated. Good correlation between the primary creep model predictions and experiments was obtained when the observed composite steady-state creep behavior converged to the McLean steady-state. The composite secondary creep was observed to correlate reasonably well with the rule-of-mixtures model developed by McLean. The composite tertiary creep was modeled by solving for global or local kink-band evolution with composite deformation respectively contributing to fiber buckling or kinking. The model predicted the critical threshold strain for the onset of tertiary stage to be most sensitive to the initial kink angles while being relatively insensitive to the initial kink-band heights and varied inversely with the volume fraction of fiber in the composite. Reasonable correlation between the model and experiments was obtained when the observed composite steady-state correlated well with the McLean steady-state.