MICROSTRAIN ACCUMULATION IN MULTIPHASE SUPERALLOYS

Transcription

1 732 MICROSTRAIN ACCUMULATION IN MULTIPHASE SUPERALLOYS J. Repper 1, M. Hofmann 1, C. Krempaszky 2, R.C. Wimpory 3, W. Petry 1, E. Werner 4 1 Forschungsneutronenquelle Heinz Maier-Leibnitz (FRM II), TU München, D Garching 2 Christian-Doppler-Labor für Werkstoffmechanik von Hochleistungslegierungen, TU München, D Garching, Germany 3 Helmholtz-Zentrum Berlin für Materialien und Energie, D Berlin Wannsee, Germany 4 Lehrstuhl für Werkstoffkunde und Werkstoffmechanik, TU München, D Garching, Germany ABSTRACT Four matrix-phase crystallographic directions of IN718 are investigated by in-situ tensile tests using neutron diffraction. The elastic diffraction constants for all directions measured are compared to theoretical values calculated by the Kröner model. The differences between the microscopic and the macroscopic material response are given. The accumulation of microstrains in the different crystallographic directions is discussed. A comparison between the results of a single phase material (ingot IN718) and two differently thermal treated multiphase materials is presented. INTRODUCTION Thermal and mechanical properties of high performance alloys are strongly affected by changes in the microstructure due to thermo-mechanical treatments during the production processes. In addition to changes in the microstructure these processes often go along with the accumulation of residual stresses on different length scales [1]. The macroscopic stresses (type I stresses) occur over the whole component. Superposed to these, intergranular and interphase residual stresses (type II stresses) on the microscopic scale may be generated. These microscopic stresses are strongly dependent on the microstructure of the sample. They result from the elastic and plastic anisotropy of both, grains of the same phase (intergranular stresses) and grains of different phases (interphase stresses) [2]. This coexistence of macrostress and microstress is also present, at least in a simple way, during external loading of a uniaxial tensile specimen. A powerful tool to determine the occurring stresses are diffraction methods [3]. In this paper we present the results

2 This document was presented at the Denver X-ray Conference (DXC) on Applications of X-ray Analysis. Sponsored by the International Centre for Diffraction Data (ICDD). This document is provided by ICDD in cooperation with the authors and presenters of the DXC for the express purpose of educating the scientific community. All copyrights for the document are retained by ICDD. Usage is restricted for the purposes of education and scientific research. DXC Website ICDD Website -

3 733 of in-situ load experiments using neutron diffraction on Inconel 718 samples possessing three different microstructures differing in grain size, volume fraction and types of precipitates. MATERIAL Inconel 718 (IN718) is a nickel based high temperature superalloy. Its chemical composition can be seen in table 1. The alloy is strengthened by sequences of thermal processes resulting in the precipitation of additional phases in the fcc Nickel matrix. Precipitates which can be found in the Ni matrix are: the stable δ-phase, the stable γ -phase, and the metastable γ -phase [4, 5]. The hard orthorhombic δ-phase, which is formed inevitably during the service of high temperature components, leads to an undesirable embrittlement of the material. The face centered cubic γ - and the body centered tetragonal γ -phases are used to strengthen the superalloy by an increase in yield strength and ductility. Therefore, different thermal standard treatments resulting in the precipitation of γ - and γ - phase are widely used in industry. Table 1: Chemical composition of IN718 in weight %. Element Ni Fe Cr Nb Mo Ti Al C B Weight % ,3 3,1 0,95 0,55 0,03 0,005 SAMPLES All measurements were made on round tensile test specimens of IN718. The total length of the specimens was 60 mm with a gauge length diameter of 6 mm. We explored three sample sets differing in microstructure. The first sample set is made from ingot material of IN718 used in industry (Sample A). The other two sample sets consist of thermally treated ingot material. The first thermal treatment favours the precipitation of the brittle δ-phase (Sample B). The second treatment (Sample C) is equivalent to the standard heat treatment of IN718 used in industry and to facilitate the precipitation of the γ - and γ - phases. Because all specimens were machined from the same cylindrical ingot material and subjected to similar heat treatment procedures texture can be assumed as equivalent in all sample sets.

4 734 EXPERIMENT The measurements on all three sample states were carried out at the neutron residual stress diffractometer E3 at BENSC at Helmholtz-Zentrum Berlin, Germany [6]. The wavelength was λ = Å with a gauge size of 4 x 4 x 6 mm³. Due to the restricted 2θ range of the instrument set-up only four lattice planes of the fcc Ni-matrix ({200}, {220}, {311}, {222}) were observable. The measurement direction given by the normal vector of the measured lattice planes was parallel to the load direction for all measurements. The scattering angle coverage of the detector allowed the observation of one Bragg reflection per measurement only. Therefore, four identical test specimens for each sample state were used to investigate the behaviour of four lattice planes. During the in-situ measurements the applied load was increased continuously with a velocity of v F = 5 N/s. The count time for each measurement point was 90 s. RESULTS AND DISCUSSION The stress- lattice strain diagrams for all observed lattice planes are shown in figures 2a) to 2c) with the macroscopic stress bulk strain curves for all three sample states given in figure 2d). A macroscopic Young s modulus of E = 194 ± 5 GPa was determined for all sample sets. The elastic diffraction constant was determined for each lattice plane and the values are shown in table 2. A comparison of the results for one single direction gives only small variations between the different sample states. The diffraction elastic constants for a pure Nickel matrix calculated from the Kröner model [7] using the XEC [8] program is also given in table 3. Regarding the error bars a good agreement of theoretically calculated and experimentally determined elastic constants can be seen for all sample states. In terms of the determined macroscopic Young s modulus of E = 194 GPa, the best representation of the macroscopic material response is given by the {311} and the {200} planes for all three sample states (see table 2). The {311} lattice planes slightly overestimate the macroscopic Young s modulus whereas, for the {200} planes it is somewhat too low. The worst agreement between the macroscopic material response and the stress lattice strain behaviour is found for the {222} planes. The accumulated residual lattice strains (microstrains) after a macroscopic unloading of the samples are shown in table 3. As seen in figure 2d) the plastic bulk strain levels of samples A (approx. 7%) and C (approx. 6%) are similar, whereas sample B was strained to approx. 11% up to the onset of necking. Correction of the accumulated microstrains to the same plastic bulk

5 735 Figure 2: Stress-strain diagrams for the four investigated lattice planes of the Ni matrix-phase of three differently treated IN718 samples. a) ingot material (A), b) δ-phase enriched IN718 (B), c) IN718 prepared with the industrially used standard heat treatment (C). Differences in the mechanical behaviour are evident. For clarity the symbols only show the average values over three (elastic regime) and two (plastic regime) measurement points, respectively. In addition the macroscopic stress strain diagrams for all three sample states are given in d) in which the remaining plastic bulk strains after unloading can also be seen. strains was omitted for sample B because the resulting changes in the residual lattice strains doesn t affect the tendencies discussed in the following. For all three sample states the lattice planes with the highest accumulated microstrains are the {200} (tensile strains) and the {220} (compressive strains) planes. Therefore, the Bragg reflections resulting from scattering at these lattice planes are less favourable choices to derive residual stresses because the results could be highly influenced by microstrains. This conclusion is in accordance with earlier results [2, 9]. The residual lattice strains of the {311} planes are almost zero for the A and C sample states, whereas the {311} planes of sample state B show a clear tendency to accumulate compressive microstrains after the macroscopic unloading of the sample. A similar effect can be seen for the

6 736 {222} planes of the C sample state. Here, the microstrain state is in tension, whereas the {222} planes of the sample states A and B clearly accumulate compressive microstrains. The different behaviour of the same lattice planes in different sample states can be explained with the mechanisms of residual lattice strain evolution in the samples. The accumulated microstrains after macroscopic unloading of the samples are a mixture of intergranular and interphase microstrains. Therefore, they result from the plastic and elastic anisotropies of the different crystallographic directions of both, the Ni-matrix and the additional precipitates in the matrix. Table 2: Young s modulus for different sample states and lattice planes determined by neutron diffraction compared with diffraction elastic constants for pure Nickel calculated with a computer software based on the Kröner model [7,8]. Sample Diffraction elastic constant [GPa] of lattice planes {200} {220} {311} {222} A 187 ± ± ± ± 9 B 194 ± ± ± ± 24 C 170 ± ± ± ± 24 Kröner Table 3: Accumulated residual lattice strains of different sample states and lattice planes determined by neutron diffraction. Sample Accumulated microstrains [MPa] of lattice planes {200} {220} {311} {222} A 1541 ± ± ± ± 91 B 442 ± ± ± ± 85 C 3872 ± ± ± ± 145

7 737 Here, the microstrains were determined only for the Ni matrix, not for precipitates. The ingot material of IN718 is almost free of additional phases. It can be treated as a single phase material. In the B and C sample states different phases are precipitated. The material of sample state C shows mainly the γ - and γ -phases, whereas the material of sample state B exhibits a high amount of the brittle and hard δ-phase. The different precipitates have a significant effect on the heterogeneous stress distribution amongst the different grain orientations. In case of the {311} and {222} lattice planes even the sign of the residual lattice strain development is affected. For a detailed explanation of this effect, the determination of the elasto-plastic material parameters of the different phases and the role of coherency stresses have to be investigated. CONCLUSION The data show that the microstructure of the sample has to be taken into account for the choice of the best Bragg reflection for residual stress analyses. For the investigated IN718 sample states at hand one can draw the following conclusions: Due to its capability to render the macroscopic Young s modulus and its very low accumulation of residual lattice strains, the (311) matrix Bragg reflection will be the best choice for residual stress determination in the matrix-phase of the ingot material and the γ - and γ -strengthened IN718 sample. For the δ- phase enriched material, the (200) reflection appears to be the most reliable reflection for the determination of macroscopic residual stresses. More investigation is necessary to see if this conclusion can be more generally applied. ACKNOWLEDGMENTS The authors acknowledge the BENSC in the HZB for providing beam time and the DFG for funding this research within projects WE 2351/11-1 and PE 580/7-1. REFERENCES [1] H. Behnken: Mikrospannungen in vielkristallinen und heterogenen Werkstoffen, Habilitationsschrift, Shaker Verlag, Aachen, 2003.

8 738 [2] D. Dye, H.J. Stone, R.C. Reed: Intergranular and interphase microstresses, Current Opinion in Solid State and Materials Science, 5(2001), [3] M.T. Hutchings, P.J. Withers, T.M. Holden, T. Lorentzen: Introduction to the Characterization of Residual Stress by Neutron Diffraction, Taylor & Francis, Boca Raton, [4] R. Bürgel: Handbuch Hochtemperatur- Werkstofftechnik, Vieweg, Wiesbaden, [5] M. Durand-Charre: The Microstructure of Superalloys, Gordon and Breach Science Publishers, Amsterdam, [6] R.C. Wimpory, P. Mikula, J. Saroun, T. Poeste, J. Li, M. Hofmann, R. Schneider: Efficiency boost of the materials science diffractometer E3 at BENSC: One order of magnitude due to a horizontally and vertically focusing monochromator, Neutron News, 19 (2008), [7] E. Kröner: Berechnung der elastischen Konstanten des Vielkristalls aus den Konstanten des Einkristalls, Zeitschrift für Physik, Bd. 151 (1958), [8] H. Wern, R. Johannes, H. Walz: Dependence of the X-ray Elastic Constants on the Diffraction Plane; Phys. Stat. Sol. (b), 206 (1998), [9] T.M. Holden, R.A. Holt, A.P. Clarke: Intergranular stains in Inconel-600 and the impact on interpreting stress fields in bent steam-generator tubing, Materials Science and Engineering A, 246 (1998),