Fabrication of parts by metal injection molding (MIM) results in very fine grain sizes. In the present investigation, heat treatments were employed to gain a larger grain size that is more creep resistant. The alloy under investigation was CM 247 LC. The composition was slightly modified to facilitate grain growth. Secondary recrystallization was observed to occur during post-sintering annealing treatments approximately 50 °C below the sintering temperature. The grain size increased from 25 µm to > 2 mm. The increase of grain size was found to improve creep strength significantly. The samples in the as-sintered condition exhibited a bimodal grain size distribution. The grain size in a thin surface zone after sintering was slightly increased because of a lower C and O content in this zone that promoted normal grain growth. Secondary recrystallization did not occur in the surface zone. This is attributed to a lack of driving force in this area.
Superalloys are key materials in gas turbines and petrochemical applications due to their superior mechanical properties and corrosion resistance at elevated temperatures.[1,2] The use of superalloys, however, leads to high component cost as casting, forming and machining processes are all difficult and expensive. In case of more complex geometries, net shape components cannot be achieved and machining is particularly expensive. A potential solution for cost reduction and net shape processing that has been pursued for a long time is powder metallurgy (PM),[3,4] in particular hot isostatic pressing of canned powder (HIP).[5,6,7] Newer PM processes with a high potential include metal injection molding (MIM)[8,9,10,11] and additive manufacturing (AM).[12,13,14]
MIM, which has been investigated in the present study, typically includes the following four steps: (1) Feedstock preparation. The metal powder and thermoplastic polymer binder are mixed together in an optimized ratio to prepare the pelletized feedstock. (2) Injection molding. The feedstock is brought to a temperature above the polymer melting point and injected into the mold with high pressure to obtain the green part. (3) Debinding. The green part is first solvent debound and then thermally debound before sintering. After solvent and thermal debinding, most of the binder is already removed. There are continuous channels inside the parts after thermal debinding so that the residual binder can be evaporated during the sintering process. (4) Sintering. The parts after debinding are sintered in high-temperature vacuum furnaces, close to the solidus temperature of the alloys, to obtain the final dense parts. Although MIM is a new technology in the superalloy community, it has already received much attention.[16,17,18]
Some of the first investigations on MIM superalloys focused on alloy compositions such as IN 718 or IN 625.[16,17,19] These alloys contain very small amounts of aluminum and are not suitable for applications where very high strength and hot oxidation resistance are required. Recently attention has turned to high aluminum chemistries such as IN 713 LC and CM 247 LC.[18,21] Creep properties of IN 713 LC after MIM processing are comparable to the cast and HIP IN 713 LC alloy at intermediate temperatures such as 600 °C and intermediate rupture times. However, if higher temperatures and longer times are considered, the creep properties of MIM material are inferior. This is due to the very fine grain size of MIM alloys (10 to 25 μm). The fine grain size is detrimental to the deformation resistance of superalloys at high temperatures because of severe grain boundary sliding.[22,23]
The present work focuses on grain size modification to improve the high temperature creep properties of MIM CM 247 LC. When it comes to normal grain growth, the challenge is the sluggish grain growth kinetics of MIM superalloys. A high volume fraction of carbide particles is present after sintering, which will retard the migration of the grain boundaries.[18,24,25] To improve the grain growth kinetics of MIM superalloys, the composition of the nickel-base superalloy CM 247 LC is slightly modified in the present study. C and Hf contents in CM 247 LC were reduced in the pre-alloyed powder as they are the main elements that will form MC carbide precipitates.[18,26] A further consideration in this context is the high carbon uptake in MIM processing of superalloys, i.e., there is always enough C present in the material. In the same vein, a high Hf content is not really required in MIM alloys. Hf is known to improve the hot tearing behavior during directional solidification, a phenomenon that is of no concern in MIM processing.[27,28,29] High Hf concentrations are otherwise undesirable as they lead to a decrease of the melting point and formation of γ/γ′-eutectic, which makes heat treatment difficult.
The alloy powder chemistry used in the present investigation is a derivate of superalloy CM 247 LC. The powder was prepared through Ar inert gas atomization by a commercial vendor, H.C. Starck (Goslar, Germany). The compositions of the pre-alloyed MIM powder and CM 247 LC cast alloy are listed in Table I. Particle size distribution analysis indicated that the D90 vol pct is < 32 µm, which indicates the powder is suitable for MIM. The density of the powder was 8.57 g/cm3 as analyzed by gas pycnometry.
MIM Specimen Preparation
MIM feedstock and injection molded samples (green samples) were prepared by Schunk Sintermetalltechnik GmbH (Thale, Germany). Two standard geometries were used for our investigations, including tensile bars and creep specimens. The specimens were received after solvent debinding at Schunk Sintermetalltechnik GmbH. They were then thermally debound in our laboratory with a Gero GLO furnace (Carbolite-Gero GmbH, Germany) in H2 flow. After thermal debinding, they were sintered from 1275 °C to 1325 °C for 3 hours under high vacuum with a Gero HTK furnace (Carbolite-Gero GmbH, Germany). The sintering temperature was optimized and then was used for other specimens. The as-sintered specimens were used for subsequent heat treatments and mechanical tests. The densities of as-sintered samples were measured using the Archimedes method according to DIN EN ISO 3369.
Other Experimental Details
C and O contents were analyzed at Neue Materialien Fürth GmbH (Fürth, Germany) using Horiba EMIA-320 V2 and EMGA-620W/C devices (Horiba, Japan). Grain growth heat treatment was performed from 1240 °C to 1300 °C for different holding times with a Gero LHTM furnace (Carbolite-Gero GmbH, Germany). The specimens were protected with Ar flow (1 L/hour) during isothermal holding. Samples cut from MIM tensile bars were used for the grain growth heat treatment investigation.
Metallography specimens were all prepared with polishing machines (Struers, Germany). The specimens were cut with Accutom (Struers, Germany), cleaned in an ultrasonic bath, cold mounted with Technovit® 4071 resin (Kulzer GmbH, Germany), ground with SiC paper (# 200 to # 4000, Struers, Germany) and finally polished using standard colloidal suspension OP-U (0.04 µm, Struers, Germany). Morphology of pores was observed using polished specimens with the Imager.M1m optical microscope (Carl Zeiss Microscopy, Germany). The area fraction of porosity was measured using ImageJ software. The prepared specimens were etched with Spüli etchant [85 ml distilled water, 60 ml hydrochloric acid HCl (32 pct), 1 g molybdic acid MoO3 (85 pct), 15 ml nitric acid HNO3 and 5 drops “Spuelmittel” (dish detergent/dish soap)] at room temperature for 2 seconds. Micrographs were obtained with the Imager.M1m optical microscope (Carl Zeiss Microscopy, Germany) and Quanta 450 scanning electron microscope (SEM) (FEI). Electron backscattered diffraction (EBSD) analysis was performed on a FEI Helios NanoLab 600i FIB Workstation equipped with an Oxford Instruments NordlysNano EBSD detector. Transmission electron microscopy (TEM) analysis was conducted on Philips CM30 (FEI) with an accelerating voltage of 300 kV. TEM samples were prepared with a twin jet polishing machine (Struers, Germany) at Institute 1 (WW1), University of Erlangen-Nuremberg. Electron probe microanalysis (EPMA) was performed using JXA-8100 (JEOL, Japan). Differential scanning calorimetry (DSC) analysis with a STA 409 C/CD thermal analyzer (Netzsch, Germany) was used to obtain the transformation temperatures, such as the γ′ solvus temperature, solidus and liquidus temperature of MIM powder and as-sintered specimen.
Creep tests were performed at the Joint Institute of Advanced Materials and Processes (ZMP), Friedrich-Alexander University of Erlangen-Nuremberg. Three groups of specimens were tested: as-sintered specimens (MIM), specimens after solution and aging heat treatment (MIM + HT1) and specimens after recrystallization and aging (MIM + HT2). The treatment details are listed in Table II.
Microstructure in Starting Powder and After Sintering
The morphology of the powder used for the present study is presented in Figure 1. Obviously, most of the powder particles are spherical, and there are only a few irregular particles (Figure 1(a)). Cross-sectional check of individual powder particles reveals a cellular solidification structure due to the fast cooling during atomization (Figure 1(b)). Bright eutectic precipitates formed in between the cellular structures, indicating that some segregation still occurs. The fine grain structure inside the individual powder particles can be best observed through EBSD measurements as in Figure 1(c). The average grain size is typically < 5 µm.
Figure 2 presents the density and porosity after sintering at different temperatures in MIM tensile bar specimens. With the increase of the sintering temperature, the density increased while the corresponding porosity decreased; 1305 °C is chosen as the optimized sintering temperature in the following. Unnecessarily high sintering temperatures are to be avoided because of the evaporation of Cr at higher sintering temperatures.
Typical macro- and microstructures of the as-sintered MIM tensile bars are shown in Figure 3. One representative MIM tensile bar after sintering is presented in Figure 3(a). The head of the MIM tensile bar is cut along the yellow line (Figure 3(a)) to obtain a cross-sectional view. As shown in Figures 3(b) and (c), a bimodal grain size distribution in the cross section is observed. There are coarse grains (mean grain size > 100 µm) in the surface region of the tensile bar, while in the center there are still fine grains (mean size ≈ 25 µm). The fine-grain zone is enveloped by a coarse grain zone.
During sintering, the shrinkage of the brown part is about 14 pct. Some residual pores can be observed after sintering (Figure 3(d)). The pores are distributed homogeneously in the whole cross section, and the pore size is < 20 µm. Further check of the as-sintered microstructure proved that large and irregular γ′ particles precipitate along the grain boundaries (Figure 3(e)), while small γ′ particles are distributed homogeneously inside the grains. Ta- and Hf-rich MC carbides are also observed.
The atmosphere in sintering is such that C and O are removed from the sample. On the other hand, as the polymer binder system contains C and O and some binder residue will be left after solvent debinding, there is also a chance for C and O uptake during thermal debinding. C and O content was analyzed in the powder and after various stages of heat treatment. The results are shown in Table III. Obviously, C and O increased after thermal debinding, but both C and O decreased at the specimen surface after sintering. Interestingly, the C content in the surface region of the as-sintered specimen is even lower than the C content in the powder, while the center has absorbed about 200 ppm C and 400 ppm O. Obviously the samples pick up C and O during thermal debinding. During sintering C and O are removed, but the times are not long enough for the diffusional flow to reach the center of the sample.
Grain Growth During Additional Post-sintering Heat Treatment
Secondary recrystallization in the fine-grain center, normal grain growth in the outer rim
A series of post-sintering heat treatments was performed above the γ′ solvus temperature to introduce grain growth as shown in Figure 4. As stated above, before the additional heat treatment, the sample exhibited a bimodal grain size distribution, with a fine-grain zone in the center enveloped by a coarse grain zone at the surface. As can be seen in Figure 4, the additional heat treatment caused little change in the surface coarse grain zone. The average grain size is only about 200 µm even after holding at 1280 °C for 10 hours indicating the expected rather sluggish growth kinetics for normal grain growth. More surprisingly, the central fine-grain region exhibits very different grain growth behavior that will be interpreted as secondary recrystallization later; see the discussion section. Very large grains, several mm in size, are formed at heat treatment temperatures ranging from 1240 °C to 1280 °C.
In Figure 5(a), the secondary recrystallized area fraction is plotted as a function of the heat treatment temperature. Since secondary recrystallization only occurred in the central fine-grain region, the quantitative evaluation is limited to this area. The measurement is performed by drawing four parallel test lines of predefined length (= 65 mm) in the fine-grain region and by determining the sum of intercepts of recrystallized grains (Σ) with these four lines. The four lines are drawn with the same intervals (0.6 mm) from the center of the cross section. The secondary recrystallization fraction is then calculated as
where SRX is secondary recrystallization. Figure 5(b) also illustrates the secondary recrystallization temperature window together with the DSC curves of powder and an as-sintered specimen. Obviously, the secondary recrystallization window is above the γ′ solvus temperature and below the sintering temperature. There is no noticeable event that can be noted in DSC that might trigger the secondary recrystallization process. Figure 6 shows the development of recrystallization with time. A Johnson–Mehl–Avrami type of behavior is observed with more rapid growth at the higher temperature, as expected.
Grain size as a function of temperature
In Figure 7, the average size of secondary recrystallization grains (secondary recrystallization of fine-grain center region of the sample) is plotted as a function of temperature. The largest recrystallized grains are observed right above the recrystallization temperature, while the size of secondary recrystallization grains will decrease with the increase of the holding temperature. This is in contrast to the normal grain growth results in Figure 4 (normal grain growth in the coarse grain rim of the sample) where the average grain size increases with temperature.
Grain boundary morphologies
For secondary recrystallization to occur, an inhibition mechanism is required, which prevents most grain boundaries from moving. The inhibition typically occurs by the presence of dispersed second phase particles or by grain boundary impurities. Figure 8 shows the results of a cursory TEM investigation of the grain boundaries before and after recrystallization. Before recrystallization, the grain boundaries are curved and twisted, with MC particles randomly precipitating along the GB (Figure 8(a)). After recrystallization, the grain boundaries are rather straight (Figure 8(b)), suggesting the disappearance of an inhibition mechanism. The exact nature of the inhibition is still unclear at this stage.
Three groups of specimens with different grain sizes were creep tested at different temperatures. The sample geometry is according to the Rolls-Royce standard. The processing conditions for the different specimens are listed in Table II. The grain sizes for MIM (as-sintered fine grain), MIM + HT1 (fine grain) and MIM + HT2 (SRX grain) specimens are about 25 µm, 43 µm and > 2 mm, respectively. Strain vs. time curves of different specimens tested at 700 °C, 800 °C and 900 °C are shown in Figure 9. Obviously, creep resistance of secondary recrystallized grain specimens increases compared with the MIM and MIM + HT1 specimens. At 700 °C, the modification of the γ′ morphology and distribution through heat treatment is more effective in improving the creep properties. The increase of creep life is more obvious at 800 °C and 900 °C than at lower temperatures. Stress vs. time curves at both temperatures are presented in Figures 9(b) and (c). The loss of ductility with secondary recrystallization is notable.
Grain Growth Mechanism
Grain growth can occur by primary recrystallization, secondary recrystallization or normal grain growth. Secondary recrystallization is also referred to as discontinuous grain coarsening or abnormal grain growth. Primary recrystallization is of course not under discussion in the present context, as the material has not been plastically deformed and contains only a few dislocations.
Normal grain growth is unavoidable in a material that contains grain boundaries because they have interfacial energy. Normal grain growth in its simplest form can be described by the following equations. Grain boundary velocity v is proportional to grain boundary mobility M and driving force \( 2\gamma /L \),
where L is the mean grain intercept, t is the time, and γ is the interfacial energy. Integrating with respect to time, we obtain
Grain boundary mobility depends on the diffusion coefficient, i.e., normal grain growth is suppressed at low temperatures. It can also be impeded by dispersoids and impurities.
Secondary recrystallization describes the sudden and relatively rapid growth of a very small number of grains, consuming all the other existing grains. Often it is only one out of a million existing grains or more that actually grows, and the resulting final grain size is on the order of mm or cm. The considerations of driving force and mobility of grain boundaries are in principal the same as in normal grain growth. The reason for secondary recrystallization to take place is then the presence of dispersions or impurities that prevent normal grain growth from occurring. A slight degree of coarsening of the dispersoids or a slight degree of desegregation of impurities can unpin the boundaries of the very few largest grains only. They will take off and consume all the other smaller ones.
As stated above, in the present investigation the as-sintered material, which is then subjected to another heat treatment, is characterized by a bimodal grain structure, a coarser grained outer rim and a finer grained center. We interpret the grain growth that takes place during sintering and during subsequent heat treatment in the outer rim as normal grain growth. The grain growth in the center of the sample that leads to millimeter-sized grains is attributed to secondary recrystallization.
The following observations lead to the conclusion that secondary recrystallization takes place:
Normal grain growth is effectively suppressed A thorough investigation of the kinetics of normal grain growth will follow later. The results on normal grain growth in Figure 7(b) point already to the fact that at temperatures ranging from 1250 °C to 1270 °C, where secondary recrystallization is meant to occur, normal grain growth is suppressed because of low grain boundary mobility.
There is a well-defined grain-coarsening temperature As Figures 4 and 5(a) show, grain growth occurs all of a sudden once a temperature of 1250 °C is exceeded. This is characteristic for secondary recrystallization where at a given temperature for a small portion of grains the blocking mechanisms disappears.
Bimodal Grain Structure
In the as-sintered condition the fine-grained core region of the samples is surrounded by a coarser grained rim (Figure 3(b)). The fine-grained core region displays a higher C and O content than the coarse-grained outer rim (Table III). The difference in grain size is attributed to a difference in normal grain growth. The lower C and O content in the surface implies lower retarding forces and faster grain growth. After post-sintering heat treatment, the situation is reversed; see Figure 4. Now the center is coarser grained because of secondary recrystallization. Such a grain structure is highly desirable from an application point of view. Large grains are necessary for high creep strength. Secondary recrystallization does not occur in the outer coarse grain rim of the samples. This is attributed to the lower driving force in this area.
Grain size Dependence of Creep Properties
Under creep conditions that are most relevant to practical application, superalloys deform by dislocation mechanisms. Typically, the creep rate is found to be independent of grain size. This picture will change if alloys with very small grain sizes are considered (grain size < 100 μm). Grain boundary sliding will then yield contributions to plastic deformation that grow with increasing temperature and decreasing strain rate. There is ample experimental evidence that for alloys with a higher γ′-volume fraction the weakening influence of grain boundaries becomes noticeable earlier, i.e., a grain size of 25 μm might not lead to noticeable weakening in a solid solution strengthened NiCr alloy, but it definitely will in an alloy with high γ′-volume fraction such as CM 247 LC.
In the light of these considerations, the experimental finding is not surprising that an increase of grain size from 25 μm to > 2 mm results in superior creep strength (Figure 9), in particular at the high temperatures. The increase would probably be larger if longer times and higher temperatures were tested.
Quite surprising is the drop in strain to failure in our experiments. Previously, Robert also reported that the superalloy IN 713 C after secondary recrystallization exhibits low ductility. One explanation is suggested by the results in Table III. The C and O concentration in the material is much higher than what is aimed at in a superalloy. Changing the grain size from 25 μm to > 2 mm means the grain boundary area per volume will decrease dramatically by a factor of > 200. This implies the concentration of undesirable elements on the grain boundary increases sharply. In the future the processing route should be optimized to avoid such an increase.
The secondary recrystallization behavior of MIM CM 247 superalloy was investigated in the present work. The following conclusions can be drawn.
C and O are reduced in the surface rim of the as-sintered specimens during sintering in vacuum. Bimodal grain sizes (surface coarse grain and central fine grains) are generated as a result.
Secondary recrystallization occurred in the central fine-grain regions during post-sintering heat treatments, which will produce large SRX grains. This secondary recrystallization occurred above the γ′ solvus temperature and below the solidus temperature.
Creep properties of large grain specimens after secondary recrystallization exhibit much better creep life compared with the as-sintered specimens at high temperatures. However, the ductility is still quite low, which might be due to the high O and N content.
 R.C. Reed: The superalloys: fundamentals and applications, Cambridge University Press, London, 2009, pp. 1-28.
 C.T. Sims, N.S. Stoloff, W.C. Hagel: Superalloys II, John Wiley & Sons, New York, 1987, pp. 3-25.
 G.H. Gessinger, M.J. Bomford: Metallurgical Reviews, 1984, vol. 19(1), pp. 51-76.
H.J. Brown: Powder metallurgy, 1964, US Patent, US3,155,502 A.
 K. Essa, R. Khan, H. Hassanin, M.M. Attallah, R. Reed: The International Journal of Advanced Manufacturing Technology, 2016, vol. 83(9), pp. 1835-45.
 L. Chang, W. Sun, Y. Cui, R. Yang: Metallurgical and Materials Transactions A, 2017, vol. 48(3), pp. 1273-87.
 C. Qiu, X. Wu: Philosophical Magazine, 2014, vol. 94(3), pp. 242-64.
R.M. German: Injection Molding of Metals and Ceramics. Metal Powder Industries Federation, 1997.
 H.Ö. Gülsoy, Ö. Özgün, S. Bilketay: Materials Science and Engineering: A, 2016, vol. 651, pp. 914-24.
 O. Dugauguez, J.M. Torralba, T. Barriere, J.-C. Gelin: Powder Metallurgy, 2017, vol. 60(2), 131-38.
Ö. Özgün, R. Yılmaz, H. ÖzkanGülsoy, F. Fındık: Materials Characterization, 2015, vol. 108, pp. 8-15.
 W.E. Frazier: Journal of Materials Engineering & Performance, 2014, vol. 23(6), pp. 1917-28.
 H.E. Helmer, C. Körner, R.F. Singer: Journal of Materials Research, 2014, vol. 29(17), pp. 1987-96.
 M. Ramsperger, R.F. Singer, C. Körner: Metallurgical & Materials Transactions A, 2016, vol. 47(3), pp. 1469-80.
R.D. Rivers: Method of injection molding powder metal parts, 1978, US Patent, US4,113,480 A.
Ö. Özgün, H. ÖzkanGülsoy, R. Yilmaz, F. Findik: Journal of Alloys and Compounds, 2013, vol. 546, pp. 192-207.
 B. Albert, R. Völkl, U. Glatzel: Metallurgical and Materials Transactions A, 2014, vol. 45(10), pp. 4561-71.
 A. Meyer, E. Daenicke, K. Horke, M. Moor, S. Müller, I. Langer, R.F. Singer: Powder Metallurgy, 2016, vol. 59(1), pp. 51-56.
 E. Hnatkova, B. Hausnerova, A. Hales, L. Jiranek, F. Derguti, I. Todd: Powder Technology, 2017, vol. 322, pp. 439-46.
K. Horke, A. Meyer, E. Daenicke, R. Singer: Proceedings of World PM 2016, Hamburg, 2016.
A. Meyer, K. Horke, E. Daenicke, S. Müller, I. Langer, R.F. Singer: Proceedings of Powdermet 2017, Las Vegas, 2017.
 T.G. Langdon: Acta Metallurgica et Materialia, 1994, vol. 42(7), pp. 2437-43.
T. Langdon: Philosophical Magazine, 1970, vol. 22(178), pp. 689-700.
 I. Andersen, Ø. Grong, N. Ryum: Acta Metallurgica Et Materialia, 1995, vol. 43(7), pp. 2689-2700.
 V.Y. Novikov: Acta Metallurgica Et Materialia, 1994, vol. 42(5), pp. 1639-42.
K. Harris, G.L. Erickson, R.E. Schwer: in Superalloys 1984: Proceedings of the 5th International Symposium of Superalloys. Wiley Online Library, 1984, pp. 221–30.
 K. Heck, R. Blackford, R.F. Singer: Materials Science and Technology, 1999, vol. 15(2), pp. 213-20.
 J.E. Doherty, B.H. Kea, A.F. Giamei: JOM, 1971, vol. 23(11), 59-62.
 J. Zhang, R.F. Singer: Acta Materialia, 2002, vol. 50(7), pp. 1869-79.
 D. Raabe: Phys. Metall., 2014, vol. 3, pp. 2291–397.
R. W. Cahn: Phys. Metall., 1996, vol. 3, pp. 2400–500.
 A. Lasalmonie, J. L. Strudel: J. Mater. Sci., 1986, vol. 21, pp. 1837-52.
R.V. Miner, Jr.: NASA Technical Memorandum, 1972, pp. 1–17.
The authors thank Rolls-Royce Deutschland and Bundesministerium für Wirtschaft und Energie BMWi (Federal Ministry for Economic Affairs and Energy) for financial support (Funding Code 20T1312A). Some of the work was done within the framework of the LuFo project “AdCoTurb: Advanced Components for Turbines – Fortschrittliche Turbinenkomponenten.” We also acknowledge the support from Sieglinde Müller at Schunk Sintermetalltechnik GmbH in preparation of the MIM preforms. Naicheng Sheng thanks Jan P. Liebig from WW1, Friedrich-Alexander University of Erlangen-Nuremberg, for his help in grain boundary structure analysis. We thank the technical staff at WTM and ZMP, Friedrich-Alexander University of Erlangen-Nuremberg, for the help during metallography preparation, chemical analysis and machining.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Manuscript submitted June 17, 2020; accepted October 27, 2020.
About this article
Cite this article
Sheng, N., Meyer, A., Horke, K. et al. Secondary Recrystallization of Nickel-Base Superalloy CM 247 LC After Processing by Metal Injection Molding. Metall Mater Trans A 52, 512–519 (2021). https://doi.org/10.1007/s11661-020-06087-3