Elsevier

Acta Materialia

Volume 73, July 2014, Pages 287-297
Acta Materialia

Localized recrystallization during creep in nickel-based superalloys GTD444 and René N5

https://doi.org/10.1016/j.actamat.2014.03.052Get rights and content

Abstract

Creep damage mechanisms in two nickel-based superalloys, directionally solidified (DS) GTD444 and single crystal (SX) René N5, crept at high temperature (982 °C) and low stress (179–206 MPa), have been studied. Electron backscatter diffraction analyses showed localized dynamic recrystallization in specimens crept to high strain levels and to final fracture. Recrystallization was observed around clusters of carbides and creep cavities, and was less common in areas away from the fracture surface. The average recrystallized grain diameter was 0.71 μm, which, after 50 h at 982 °C, had grown to an average of 2.31 μm. Growth of the recrystallized grains occurred by the dissolution of the γ precipitates at the interface followed by discontinuous precipitation to relieve supersaturation. To consider the influence of localized recrystallization on macroscopic creep rates, a model for recrystallization-accelerated tertiary creep was developed. The model predicts that the axial strain rate increases by approximately one order of magnitude from the onset of recrystallization to rupture, comparing favorably to the experimentally measured accelerations in strain rate in the tertiary creep regime from 180 h to rupture. The observation of localized dynamic recrystallization provides new insight into the damage processes that occur in the tertiary creep regime.

Introduction

Nickel-based superalloys are essential components of gas turbine engines due to their excellent combination of high-temperature strength, toughness, and corrosion and oxidation resistance [1]. The typical superalloy microstructure contains a face-centered cubic (fcc), nickel-rich γ matrix phase and γ precipitates. Superalloys derive their strength primarily from the γ precipitate phase; this is especially true at high fractions of their melting point [2]. Minor alloying elements result in the formation of carbides or borides with cubic crystal structures [3]. Of particular importance is the creep resistance of directionally solidified and single crystal nickel-based superalloys, due to the tight tolerance between turbine airfoils and the surrounding engine housing. In pure or single phase simple materials, creep is well described by three distinct regimes: primary, secondary and tertiary [4]. However, complex engineering alloys, such as nickel-based superalloys, typically do not exhibit an extended secondary or steady stage regime with a balance between hardening and recovery processes (constant creep rate); instead, the creep rate continuously increases from a minimum rate established after a small primary creep strain [5], [6]. Thus, a majority of creep life is spent in the tertiary creep regime, as seen in Fig. 1(a). The mechanisms that result in the continuous acceleration of the creep rate in this class of materials are not well understood.

The present research focuses on creep damage mechanisms at high temperature (982 °C) and low stress (179–206 MPa). Once the load is applied at temperature, there is an initial incubation period in which dislocations propagate through the matrix γ phase [7], [8], [9]. This process is impeded by the presence of the γ precipitates, forcing the dislocations to cross-slip and bow through the narrow matrix channels. Dislocation glide occurs preferentially along the horizontal matrix channels perpendicular to the applied tensile load due to an estimated 2.4 times higher resolved shear stress on 110{111} slip systems in the horizontal channels [7], [10], [11]. This is a result of the superposition of the applied load and the misfit stresses between the γ matrix and γ precipitates. Once dislocations have permeated the material to a point at which macroscopic strains can be measured, primary creep begins. During primary creep, dislocations continue to move through the γ channels and accumulate on γ/γ interfaces, gradually relaxing the misfit and consequently the misfit stresses, leading to a decrease in the strain rate over time. In this temperature and stress regime the γ precipitates are not sheared until after a minimum creep rate has been reached.

Directional coarsening (rafting) of the precipitates also occurs during creep at high temperatures (>900°C) and is influenced by the misfit stresses. Rafting occurs by the coalescence of the γ precipitates along the less stressed matrix channels, which, for negative misfit alloys experiencing a tensile stress, are the vertical channels [12], [13], [14], [15]. At the end of the primary creep regime, a minimum creep rate is reached, and after several hours the strain rate begins to increase with time, signifying the beginning of the tertiary creep regime. It has been observed that the strain rate becomes roughly proportional to the accumulated strain in the tertiary creep regime. This behavior has been rationalized by assuming that the mobile dislocation density increases as a function of the macroscopic strain and a constant describing the dislocation multiplication rate [6], [16], [17]. The increase in mobile dislocation density results in an increase in the number of jogged segments, increasing the climb velocity and consequently the creep rate [18]. Later in the tertiary creep regime, local stresses are high enough for dislocations to shear the γ precipitates [7], [1]. Shearing has also been observed immediately after the primary creep regime in some cases, and may be a contributing factor to the acceleration of the creep rate in the tertiary regime [19].

In the tertiary creep regime, damage develops in the form of cavities nucleating and growing from casting porosity and carbides [20], [21]. The carbides are non-deformable particles, creating local stress concentrations in the surrounding material that can be relieved by decohesion of the matrix/particle interface or by particle cracking [22], [23]. In polycrystalline alloys it is well known that cavities preferentially grow on grain boundaries transverse to the loading axis, limiting creep ductility. However, in directionally solidified and single crystal alloys, very little damage is observed early in the tertiary creep regime due to the elimination of transverse boundaries [6], [24]. As creep progresses, nearby cavities coalesce and rupture occurs when cavities reach a critical fraction of the load-bearing cross-sectional area [25], [26]. The large strains associated with tertiary creep are not expected globally in turbine components, but in isolated cases could occur locally due to stress concentrations.

To date, there has been limited research on the damage mechanisms at high creep strains that lead to the rapid increase in creep strain to rupture, or on the process by which creep cavities coalesce. The objective of this research is to examine the damage processes that occur as a function of strain in the tertiary creep regime of a columnar grained directional solidified alloy and a single crystal alloy. Detailed electron backscatter diffraction (EBSD) analysis provides new insights that are incorporated into a creep damage model.

Section snippets

Experimental procedure

Directionally solidified (DS) GTD444 and single crystal (SX) René N5 Bridgman cast plates were machined into [001] oriented creep specimens. The plates were solution treated and aged with standard commercial cycles prior to machining. The surface damage in the creep specimens was minimized by using low stress grinding. Nominal alloy compositions are listed in Table 1. Creep tests were performed under constant load in air under two conditions, depending on the alloy: 982°C/179 MPa for GTD444(DS)

Results

Creep tests of both GTD444(DS) and René N5(SX) to rupture exhibited an extended tertiary creep regime, as expected. Representative strain rate and creep strain vs. time plots of a René N5(SX) specimen are shown in Fig. 1. A photograph of a specimen crept to rupture (Fig. 2) shows significant necking that extends several millimeters away from the fracture surface. A majority of the creep cavities are located near the fracture surface (Fig. 3(a)). Several typical creep cavities located in the

Recrystallization

The driving force for recrystallization is the stored energy in the material, as reflected in the dislocation density. The boundary energy of the newly formed grain and the coherent γ precipitates serve as the barriers to recrystallization. During recrystallization in nickel-based alloys there are two types of boundary–particle interactions of relevance [27]:

  • 1.

    Dissolution of γ precipitates at the moving interface and reprecipitation in the new grain, either discontinuously or continuously.

  • 2.

A model for recrystallization-accelerated tertiary creep

In general, creep rupture occurs by the growth and coalescence of voids on grain boundaries. However, in the directionally solidified and single crystal alloys used in this study, transverse boundaries have been removed to improve material properties, causing the void growth to be controlled by power-law creep of the surrounding material rather than a boundary or surface diffusion process. In the Cocks and Ashby model, a bound theorem is used to produce approximate analytical expressions for

Conclusions

Based on the results of detailed EBSD and analytical modeling, the following conclusions are made;

  • 1.

    Localized dynamic recrystallization has been observed in directionally solidified and single crystal nickel-based superalloys with as little as 20% creep strain. Dynamic recrystallization during creep testing of GTD444(DS) and René N5(SX) is expected to occur first at a critical defect in the specimen, such as a large void or a cluster of carbides, or a combination of both. The initial

Acknowledgments

The authors would like to thank GE Power & Water, specifically Steve Balsone and Jon Schaeffer, for providing financial support and technical guidance. A special thanks to Kate Gallup, Michael Titus, and Chris Torbet for additional technical support.

References (57)

  • T. Grosdidier et al.

    Mater Sci Eng A

    (1998)
  • B.F. Dyson et al.

    Acta Metall

    (1983)
  • T.M. Pollock et al.

    Acta Metall Mater

    (1992)
  • P. Caron et al.

    Mater Sci Eng

    (1983)
  • T.M. Pollock et al.

    Acta Metall Mater

    (1994)
  • B. Dyson et al.

    Acta Metall

    (1987)
  • A.S. Argon et al.

    Acta Metall

    (1981)
  • R. Srinivasan et al.

    Acta Mater

    (2000)
  • S.H. Ai et al.

    Scr Metall Mater

    (1992)
  • M. Loveday et al.

    Acta Metall

    (1983)
  • J. Cormier et al.

    Mater

    Sci Eng A

    (2009)
  • A. Cocks et al.

    Prog Mater Sci

    (1982)
  • M. Ashby et al.

    Acta Metall

    (1979)
  • A. Porter et al.

    Mater Sci Eng

    (1983)
  • S. Chen et al.

    Acta Metall

    (1988)
  • R. Doherty et al.

    Mater Sci Eng A

    (1997)
  • M.D. Sangid et al.

    Mater Sci Eng A

    (2010)
  • J.O. Andersson et al.

    Calphad

    (2002)
  • P. Karduck et al.

    Acta Metall

    (1983)
  • G. Gottstein et al.

    Acta Metall

    (1983)
  • L. Wang et al.

    Acta Mater

    (2006)
  • C. Zambaldi et al.

    Mater Sci Eng A

    (2007)
  • J.J. Moverare et al.

    Acta Mater

    (2009)
  • R.C. Reed

    The superalloys fundamentals and applications

    (2006)
  • T.M. Pollock et al.

    J Propul Power

    (2006)
  • O. Sherby et al.

    Prog Mater Sci

    (1967)
  • D. McLean

    Rep Prog Phys

    (1966)
  • G.R. Leverant et al.

    Metall Mater Trans B

    (1970)
  • Cited by (64)

    View all citing articles on Scopus
    View full text