66 Sangnamdong, Changwon, Gyeongnam , Rep. of Korea. Keywords: Alloy718, VAR, Ingot, Billet Cogging. Abstract

Transcription

1 Superalloys 718, 625, 76 and Derivatives 25 Edited by E.A. Loria TMS (The Merals, Metals & Materials Society), 25 CHARACTERISTICS OF VIM/VAR-PROCESSED ALLOY 718 INGOT AND THE EVOLUTION OF MICROSTRUCTURE DURING COGGING - N.-K. Park 1, J.-T. Yeom 1, J. -H. Kim 1, X.-X. Cui 1 1 Materials Processg Research Center, Korea Institute of Machery & Materials 66 Sangnamdong, Changwon, Gyeongnam 641-1, Rep. of Korea Keywords: Alloy718, VAR, Ingot, Billet Coggg Abstract Alloy 718 got with a diameter of 3mm was made by the vacuum meltg process; VIM (vacuum duction meltg) followed by VAR (vacuum arc re-meltg). The VIM/VAR got was heat-treated for homogenization, and castg structure of the got was broken down for uniform microstructures and mechanical properties by controlled coggg processes usg a hydraulic press. The VIM/VAR-processed got contas three different microstructure zones along radial direction, i.e. surface chill zone, termediate columnar ga zone, and central equiaxed zone, because the local solidification procedure varies dependg on locations with the got. To understand the local deformation behavior and microstructure evolution, compression tests were conducted on samples collected from different zones of the got wide temperature and stra rate ranges, i.e. 9~115 o C and.1 s -1. The existence of different microstructures with the got resulted different compression behaviors, which was attributed to the preferred orientation the columnar gra zone, comparison with the equiaxed gras the central region. At large stras, the itial difference microstructure eventually disappeared due to dynamic, meta-dynamic, and static recrystallizations. Constitutive relations were established for the simulation of microstructure evolution, which was applied to the billet coggg process. Introduction Large diameter gots of a nickel base Alloy 718 are normally made by vacuum meltg processes, i.e., VIM (vacuum duction meltg) followed by VAR (vacuum arc re-meltg) or ESR (electro-slag re-meltg) double meltg processes, or by the combation of ESR and VAR triple meltg processes. Even though VIM/VAR- or VIM/ESR-processed gots conta relatively uniform microstructures and alloy chemistry when compared to the as-vim processed got, there rema some undesired castg structures the order of tens of microns, cludg Laves phase, carbides, etc. The castg structures can be broken down by appropriate coggg processes for homogeneous chemistry and microstructures, and sound mechanical properties [1]. Gra size is a typical way of controllg mechanical properties of Alloy 718, i.e., coarser gra size is favored for creep strength and crack-growth resistance, and fer-gra structure for lowcycle fatigue life and tensile yield strength. The gra size can be controlled by dynamic recrystallization, meta-dynamic recrystallization, static recrystallization, and gra growth. The gra size of the billet varies dependg on locations with the billet, which is often observed actual billet makg processes. The difference gra size greater than ASTM #2 is frequent large diameter billets. The difference the gra size with the billet may be duced by the differences the got quality, i.e. chemistry and gra structure, and by the differences billet conditions, i.e. temperature, stra rate, stra, etc.[2] The process parameters for billet coggg 253

2 can be defed by the state variables, cludg stra, stra rate, and temperature. The feed rate and upset ratio billet coggg are important to ensure the soundness of the billet. The aim of this research is to vestigate the local homogeneities the microstructure of Alloy 718 got, and their potential effects on coggg. The microstructure of VIM/VAR processed got was carefully characterized, and the coggg process was evaluated an effort to achieve a uniform microstructure the fal billet. The effects of homogeneous microstructures VIM/VAR-processed billet on the flow behavior and microstructure evolution are discussed for the simulation of press forgg. The template of Deform 3D was used for coggg analysis. Decoupled FE simulation method, usg numerous steps for coggg process to calculate accumulated state variables, was employed for the study of recrystallizations and gra growth with the billet. Characteristics of the Ingot Structure A VIM/VAR-processed Alloy 718 got with a diameter of 3mm was vestigated this study. The chemical compositions of the alloy are presented Table 1. The VIM/VARprocessed got contas three microstructural features on its cross-section. The got is composed of outermost chill zone, termediate columnar gra zone, and central equiaxed zone. And, it is found that the thickness of each zone varies with the local solidification procedure. Fig. 1 shows that the difference microstructure can be attributed to the difference overall solidification process that varies with locations with the got. Due to the heat loss from the got to the lower bottom part and also from the got to the outer copper crucible, the gras grew a concave way from the bottom to the top and from the surface to the center of the got. The details of dendritic structure also vary from the center to the surface, and from the bottom to the top of the got. For the study of microstructures of the got and billet, the outermost surface, top and bottom parts of the got were discarded before coggg. The center part of the got consists of coarser microstructure than those of the columnar region, and the dendrite arm spacg of the center part is thicker than that of the columnar region as shown Fig. 2. It is here to be noted that the macrosegregation was negligible regardless of the size of the local microstructure. The amount of Nb, which is most prone to segregation, may vary dependg on got size from 2 wt.% the dendrite arms to 1 wt.% the terdendritic regions for Alloy 718 gots [3]. Accordgly, an extensive Laves phase as well as carbides is precipitated the terdendritic regions. The Laves phase is represented as A 2 B, where A atoms are primarily Ni, Fe, and Cr, and the B atoms are Nb, Mo, and Ti. [4]. Table 1. Chemical compositions of Alloy 718 used this study. (wt.%) C P S Mn Si Cr Mo Co Ti Al B Fe Zr Ni comp Bal. (a) Cross section (b) Vertical section Fig. 1 Macrostructure of VIM/VAR-processed got 254

3 Fig. 2. Center-to-surface distribution of dendritic structure of the VIM/VAR-processed got. A homogenization treatment at 115 o C was employed to remove the micro-segregation of the VIM/VAR-processed got, and to dissolve the Laves phase formed VAR process. Dendritic microstructure disappeared and alloy chemistry became uniform followg the homogenization treatment. In addition, gra size was creased somewhat, due to the dissolution of the Laves phase and carbides. X-RD analysis was made order to understand the microstructure of columnar gra zone and central equiaxed gra zone. The central equiaxed gra zone showed a random orientation relationship, but columnar gra zone showed a strong preferred dex, which will be discussed later. The different local solidification process resulted the change microstructure and mechanical properties of the billet. The variation of preferred orientations columnar gra zone was assessed durg high temperature compression tests. Flow Behavior of the Ingot under Compression It is expected that the locally different microstructures of the got show different mechanical behavior the coggg process. It is conjectured from Fig. 1 that the flow stress levels of columnar gra zone would be different from those of the central equiaxed gra zone. To vestigate the local mechanical behavior, compression tests were conducted on the samples that were mached from the center region and also from the columnar region of the got. All the compression test samples were made parallel to the length direction of the got. A Thermecmaster testg mache was used for compression tests, and the specimens were heated to the test temperature at the heatg rate of 5 o C/s, and held at each temperature for 5 m. to ensure temperature uniformity. Compression tests were conducted wide temperature and stra rate ranges, i.e. 9~115 o C and.1~1 s -1, respectively, and all the samples were gasquenched after each testg. The present presumption that a preferred orientation would result a different deformation behavior can be confirmed by the shape change compression-tested samples, as shown Fig. 3. Compression-tested specimens obtaed from the columnar gra zone, showed a square crosssection with measurable anisotropy. Fig. 4 also shows stress-stra curves of various samples obtaed from different sections with the got. It is noted that the flow stress of columnar microstructure is generally higher than that of equiaxed one at relatively low stra level (ε ~.3). This fact dicates that the gras the columnar gra zone have hard orientations compared to those of the equiaxed gra zone. (a) (b) Fig. 3. Deformation anisotropy formed compression specimens obtaed from (a) equiaxed gra zone, (b) columnar gra zone. 255

4 ε=.1 s -1 9 o C out 9 o C 95 o C out 95 o C 1 o C 1 o C out ε=.1s 115 o C out 15 o C out 15 o C 11 o C out 11 o C 115 o C ε=.1 s o C out 9 o C out 1 o C out 9 o C 95 o C 1 o C ε=1. s o C out 9 o C out 95 o C 1 o C 1 o C out 9 o C ε=1.s -1 1 o C out 9 o C out 1 o C 95 o C out 95 o C 9 o C Fig. 4. Effects of specimen locations on the flow behavior compression tests ε=.1. s -1 ε=1. s o C out 15 o C 11 o C out 11 o C 115 o C out 115 o C 15 o out 11 o C out 115 o C out 15 o C 11 o C 115 o C ε=1. s o C 15 o C out 15 o C 11 o C out 11 o C 115 o C out

5 Evolution of Microstructure durg Deformation Recrystallized gras start to form preferentially at gra boundaries, as well as at some carbides or carbonitride particles. Near the surface, where the temperature is low due to die chillg and radiation coolg, only a small fraction of recrystallization takes place. When the microstructure is reheated at sufficiently high temperatures, i.e. above 15 o C, recrystallization and gra growth take place. As such, a fully recrystallized structure can be produced by repeated deformation and reheatg. It is worth notg that the preferred orientation observed the columnar gra zone decreases with the deformation at high temperatures, due to the formation of randomly oriented gras. It is expected that the heavy deformation eventually reduces the itial orientation difference the got microstructure and forms a homogeneous microstructure after the coggg process. As for validation tests, compression specimens were cut and tested parallel to the three different axes of the got. Flow stress curves are presented Fig. 5. It is clear that the difference itial flow stress level decreases with creasg stra when the compression tests are conducted at 11 o C, dicatg dynamic recystallization occurs actively durg deformation. The amount of deformation needs to be large enough to produce a uniform microstructure, without formg a duplex structure consistg of coarse and unrecrystallized gras together with fe recrystallized gras formed near itial gra boundaries A B. T=11 o C, ε=.1s -1 1 stress,mpa C 2 A:cylder axis//r.d B:cylder axis//z.d C:cylder axis//t.d stra,ε Fig. 5. Variation of flow stresses with compression axis and stra. The existence of preferred orientations the got can be seen Fig. 6. The X-ray diffraction patterns are obtaed from three different axes of the got; i.e. (a) dicates the X-ray pattern when the X-ray cidence plane is perpendicular to the radial direction of the got, (b) the X-ray pattern when the X-ray cident plane is perpendicular to the tangential direction of the got, and (c) the X-ray pattern when the X-ray cident plane is perpendicular to the longitudal direction of the got. It is clear that the itial difference the flow stress shown Fig. 5 is attributed to the difference preferred orientation of the samples. Fig. 7 shows that the relative tensity of {2} plane the columnar gras becomes stronger with stra, and X-ray patterns eventually show the randomly oriented structure pattern. Both the gradual decrease flow stress and the loss of the preferred orientation with stra dicate that dynamic recrystallization is proment when the compression tests are conducted at 11 o C and.1s

6 (a) (a) Relative Intensity(cps) (b) (C) (d) (111) (2) (22) (311) (222) Relative Intensity(cps) (b) (c) (d) (111) (2) (22) (311)(222) θ(degrees) θ(degrees) Fig. 6. XRD patterns obtaed before compression tests. (a) X-ray cidence plane R.D, (b) X-ray cidence plane T.D, (c) X-ray cidence plane Z.D (d) Standard sample Fig. 7. Change X-ray diffraction pattern with stras when testes at 11 o C and.1 s -1 : (a).13, (b).39, (c).55, (d).73. Simulation of Gra Structure under Side Pressg For the precise control of gra size, constitutive equations for the evolution of gra structure are established under normal coggg conditions. The constitutive model for gra structure covers dynamic recrystallization, meta-dynamic recrystallization, static recrystallization, and gra growth.[5,6] The decoupled FE simulation, which considers accumulated state parameters durg deformation, is previously shown to be appropriate for the prediction of microstructure of Alloy 718 [7]. The process parameters were carefully selected to represent the actual coggg process. With constitutive equations for the gra structure evolution, the evolution of microstructure of the billet can be simulated precisely. Fig. 8 shows a conceptual procedure for the microstructure simulation of the billet. Deform-3D with a user-subroute was employed for the analysis of gra structure. Local gra size was evaluated usg constitutive equations based on experimental compression tests. The microstructure evolution was simulated for the side pressg of columnar gra sections of the got. Sections with a diameter of 5mm were mached from the columnar gra zone, and were hot forged at the simulation condition. Fig. 9 shows the comparison of simulated results and experimental results, which has been carried out at the temperature of 11 C with the height reduction of 5%. Predicted gra size the center region was below 6µm and volume fraction of recrystallized gra was very high (above.7). Meanwhile, the outer side has larger gra size because recrystallization is not significant. The actual gra size is also compared with that of the simulated result. It can be seen that the predicted gra size matches well with experimental results made on the billet, dicatg that the constitutive modelg successfully predicts the evolution of microstructure for coggg simulation. 258

7 Fig. 8. Constitutive model-base simulation of gra size with the billet. Fig. 9. Simulation results of the gra size match well with gra size distribution the billet. 259

8 Conclusion The different gra zones of VIM/VAR-processed got with a diameter of 3mm show a different flow behavior due to different solidification procedure durg the got makg process. Columnar gras show high flow stresses at normal coggg conditions, dicatg that the crystal orientation the columnar zone has hard orientation compared to that of the central equiaxed gra zone. The difference gra structure shows different flow behaviors, but disappears followg the deformation at high temperatures, dicatg the different microstructures follow a similar recrystallization route regardless of preferred orientations. Even with the orientation difference castg structure, a large deformation the billet coggg results a similar recrystallization behavior and uniform microstructure the end. The evolution of microstructure durg coggg process was confirmed by decoupled FE simulation method, and the gra structure was well presented by the constitutive modelg based on the recrystallization and gra growth. The present methodology can be successfully used for the prediction of microstructure of Alloy 718 billet coggg process. Acknowledgement Present research is sponsored by MOCIE. Discussions on the metallographic work with Prof. H.K. Park of Korea Polytechnic University are appreciated. References [1] C.A. Dandre, S.M. Roberts, R.W. Evans, and R.C. Reed, Mat. Sci. Tech. 16 (2), p.14. [2] D. Zhao, S. Guillard, and A.T. Male, High Temperature Deformation Behavior of Cast Alloy 718, Superalloy 718, 625, 76 and Various Derivatives, ed. E.A. Loria (Warrendale, PA: TMS, 1997), pp [3] J. F. Radavich, "The Physical Metallurgy of Cast and Wrought Alloy 718," Superalloy 718 Metallurgy and Applications, ed. E.A. Loria (Warrendale, PA: TMS, 1989), pp [4] R. M. F. Jones and L. A. Jackman, JOM 51(1) (1999), p.27. [5] Y.S. Na, J.T. Yeom, N.K. Park and J.Y. Lee, Met. and Mater. Int 9(1) (23), p15. [6].Y.S. Na, J.T. Yeom, N.K. Park and J.Y. Lee, J. Mater. Proc. Technology, 141 (23), p.337. [7] N. K. Park, I. S. Kim, Y. S. Na, and J. T. Yeom, J. Mater. Proc. Technology, 111 (21), p