The evaluation of a numerical criterion to provide quantitative insight on freckling conditions is critical to the successful manufacture of large superalloy castings. Of the criteria reported in the literature, those based on the Rayleigh number seem best suited to predict the onset of freckle formation. However, in their current form, these criteria cannot explain why freckles develop predominantly at the surface of single crystal (SX) castings and at midradius in VAR/ESR ingots. An experimental Bridgman-type furnace has been built to directionally solidify freckle-prone superalloys, CMSX-11B, RENÉ 88, NIM80A, WASPALOY, MAR-M247, and a variation of IN718 with high silicon content, at various angles to the vertical. Under typical industrial solidification conditions (thermal gradient between 500 and 4000 K m Ϫ1 (5 Ͻ G Ͻ 40 ЊC cm Ϫ1 ) and solidification rate between 1.67 ϫ 10 Ϫ5 and 1.0 ϫ 10 Ϫ4 m s Ϫ1 (1 Ͻ R Ͻ 6mm min Ϫ1 )), the results indicate a dependency of freckling on growth front angle likely related to the anisotropy in permeability. A modified Rayleigh criterion has been developed which accounts for directional permeability and orientation of the growth front relative to the gravity vector. Application to the experimental data shows good correlation with the onset of freckling for the range of solidification conditions examined in the study. The approximate threshold value for the modified Rayleigh number was estimated to be for CMSX-11B, 0.88, for RENÉ 88, 0.90, for NIM80A, 0.85, for WASPALOY, 0.95, for MAR-M247, 0.86, and for IN718-Si, 0.65.

No abstract

The analysis of the formation mechanisms for freckling and for random gram formation in single crystal castings has resulted in a general understanding of the principles involved, but has not been tested extensively against results in actual superalloy castings. In this work, we present the results of experimental work on the production of freckling in superalloy castings, made directionally in a Bridgemau furnace capable of producing DS castings in which the directional axis can be rotated with respect to gravity.The results indicate that we may account quantitatively for freckle formation through an analysis based on fluid-flow as described by a modified Rayleigh criterion. The theoretical critical condition for freckle initiation is found to be a Rayleigh Number of unity; the experimental determination of this critical value is found to be in the region of 0.7 -0.9 depending on the alloy examined. We ascribe the difference in these values to the secondary features of the dendrite morphology. The relation to random grain nucleation is found to lie in the role of isotherm curvature in the solidifying region. This aspect is examined by the use of ProCAST computations in both castings and remelt ingots. morphology may have a large influence on the numerical value of Ra, and therefore on the development of freckling in a casting. Thirdly, the common industrial solution to freckling in directional castings, that of increasing the solidification temperature gradient, is shown to be reasonable in that its primary influence is on the dendrite spacing, but it is also seen to have clear limitations in the case of large thermal sections as would, for example, exist in large castings. Finally, the role of casting geometry and imperfect isotherm control during solidification is not at present part of the analysis and should be included for a full description of the phenomenon.

Macrosegregation defects, such as freckles and center-segregates, are currently one of the main problems encountered by attempts at scaling-up ingot diameters in order to supply industrial gas turbine manufacturers. Focusing on niobium-containing superalloys (IN7 18, IN706 and IN625), interdendritic liquid segregation and density profiles are presented, and the occurrence of fluid flow leading to macrosegregation is linked to the solidification sequence of these alloys. This paper also mentions the possible use of the Rayleigh number as a criterion to predict detrimental fluid flow patterns. Finally, it is suggested that slight modifications of alloy chemistry may be the only way to eliminate fluid flow related macrosegregation in conventional remelting processes.

No abstract