Micron 35 (2004) 273 279 www.elsevier.com/locate/micron Comparison of different sample preparation techniques in TEM observation of microstructure of INCONEL alloy 783 subjected to prolonged isothermal exposure Longzhou Ma* Harry Reid Center for Environmental Studies, University of Nevada, Las Vegas, 4505 Maryland Parkway, Las Vegas, NV 89154-4009, USA Received 14 February 2003; revised 16 September 2003; accepted 20 October 2003 Abstract INCONEL alloy 783 was annealed and aged following the standard heat treatment procedure. One set of specimens was then isothermally exposed at 500 8C for 3000 h. Mechanical properties were measured at room temperature and 650 8C, and the results showed the prolonged exposure increased the strength and decreased elongation of alloy 783. The microstructures of as-produced and exposed material were examined using optical microscope, SEM and TEM, respectively. Three techniques, jet electro-polishing, ion milling, and focused ion beam, were employed to prepare the TEM samples to observe the variation of microstructure of alloy 783 due to isothermal exposure. TEM images of samples prepared by different methods were analyzed and compared. The results indicate that the jet electro-polishing technique allows the detail microstructure of alloy 783 subjected to different treatments to be well revealed, and thereby the TEM images can be used to explain the enhancement of strength of alloy 783 caused by isothermal exposure. q 2004 Elsevier Ltd. All rights reserved. Keywords: Alloy 783; Jet electro-polishing; Ion milling; Focused ion beam (FIB) 1. Introduction Recently developed INCONEL alloy 783 (nominal composition of Ni 34Co 25Fe 5.4Al 3Nb 3Cr) is a low coefficient of thermal expansion (CTE) superalloy. Due to high aluminum content (5.4 wt%), in addition to formation of coherent g 0 (Ni 3 Al-type) phase to provide alloy strength, alloy 783 precipitates incoherent b (NiAl-type) phase in an austenite matrix. The beneficial property of alloy 783 is achieved by the introduction of b-nial precipitates (Smith and Heck, 1996; Ma et al., 2000a). The preliminary work showed that b phase could be processed to improve the resistance to stress accelerated grain boundary oxidation (SAGBO), while providing low thermal expansion and useful mechanical properties at up to 600 8C (Mannan and debarbadillo, 1998; Ma et al., 2000b). Due to its low CTE, high strength, and good SAGBO resistance, alloy 783 has been specified for use in aircraft gas turbine components such as rings, casings, shrouds and seals. Gas turbine components usually experience thermal * Tel.: þ1-702-8925-024; fax: þ1-702-895-3094. E-mail address: lma@unlv.nevada.edu (L. Ma). exposure of up to 600 8C up to 30,000 h or more at condition of service. Therefore, the engine designers have concerned the influences of prolonged thermal exposure on the properties of superalloys for the over past decade. Early work indicated that the prolonged isothermal exposure could promote the SAGBO-induced time-dependent fatigue crack growth rate of alloy 783 at elevated temperature due to precipitation of g 0 phase within b phase (Ma et al., 2002). The transmission electron microscopy (TEM) has been an indispensable tool in materials research for obtaining valuable information about material structure and properties, since it was developed in the 1930s. To gather true information about material structure and exclude artifact effect, the selection of sample preparation technique is very important for the TEM analysis. The advancements in modern material science such as traditional structure material, semiconductor, ceramics, and metal-composite require the TEM sample preparation technique with sub-micron-level precision, high sample yield, and minimum of artifact. Since sample preparation techniques are very material dependent, there are currently three major methods, jet electro-polishing, ion milling technique, and focused ion beam (FIB) technique (Williams and Barry Carter, 1996; Phaneuf, 1999), to be 0968-4328/$ - see front matter q 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.micron.2003.10.048
274 L. Ma / Micron 35 (2004) 273 279 applied for different materials. Jet electro-polishing has been a successful technology for the preparation of most metallic materials and superconductors. Jet electro-polishing is accomplished by pumping a stream of negatively charged electrolyte against the surface of a positively charged sample. This process is utilized to generate a dished or dimpled area and is continued until perforation of the sample occurs. Usually, electro-polishing produces a large electron transparent area without mechanical-induced damage, but this technique is very electrolyte dependent and limited to some alloys susceptible to chemical etching. Ion milling technique actually consists of the ultrasonic cutting, initial grinding, dimpling and ion milling. This technology is usually applicable for cross-sectional and plan-view TEM sample, most electronic materials and composites. The basic principal of ion milling involves bombarding a specimen with energetic ions or neutral atoms accelerated and formed into a tightly ion beam. Material is sputtered from the sample resulting in an electron transparent area around the area of interest. The high energy ion bombardment sometimes generates the artifacts including preferential sputtering, sample heating and radiation damage. FIB technology uses the gallium ion beam to image the sample and to remove material by ion bombarding from both sides of sample, and consequently to make the sample transparent to electron in a TEM. Since material was thinned from two sides of sample, FIB is more effective in sample preparation than ion milling technique. The state-ofthe-art FIB technology with in-situ imaging capabilities can control a finely focused gallium ion beam to perform submicron high-precision milling, while the secondary electrons and ions released in the process are used to produce highresolution and high-contrast images. Therefore, FIB technique for TEM sample preparation has been increasingly applied for semiconductor integrated circuits (ICs) to prepare the specific site cross-sectional sample. Due to less timeconsuming and labor involvement, FIB techniques are extending its application in metallic materials, electrolytesensitive alloys and multi-layer composites to produce in situ stress-free TEM sample. The purpose of this study is to measure the mechanical properties of alloy 783, subjected to the standard heat treatment and prolonged isothermal exposure, as well as examine in-depth their microstructures with TEM. The TEM samples are to be prepared using three techniques, jet electro-polishing, ion milling and FIB, and the effects of different sample preparation techniques on the TEM imaging are to be compared and evaluated. 2. Materials and experiments 2.1. Materials, heat treatments and material property measurements A hot rolled flat of INCONEL alloy 783 with a cross section of 120.00 mm 9.50 mm was supplied by Special Metals Corporation, Huntington, WV. Analyzed chemical composition of alloy 783 with weight percentage in parenthesis was: Co (34.39), Ni (28.21), Fe (24.88), Al (5.32), Nb (3.11), Cr (3.24), Ti (0.32), C (0.008), and Si (0.06). The blank samples were cut from as-rolled flats and were subjected to a standard heat treatment as follows: Material was annealed at 1120 8C for 1 h and quenched in water. Then, the annealed material was heated to 843 8C for 4 h and cooled in air. This heat treatment, referred as b aging in the text, precipitated b phase. After b aging, the material was heated to 720 8C for 8 h and cooled in furnace at a rate of 55 8C/h to 620 8C, held for 8 h and then cooled in air. This procedure is called g 0 aging, which results in the formation of g 0 in the matrix. After the standard heat treatment, one set of specimens was subjected to isothermal exposure at 500 8C for 3000 h. All samples were machined into the dog-bone tensile specimens with a cross-section of 3.18 3.18 mm 2, and the effective length of 44.45 mm. The orientation of specimen is longitudinal. The tensile test and hardness measurement were conducted at room temperature and 650 8C in a MTS servo-hydraulic system and AGFA hardness tester, respectively. The tensile rate was kept at 0.02 mm/s for all specimens. The grain size and intergranular precipitate developed during hot rolling and heat treatments were examined with an optical microscope and scanning electron microscope (SEM). The internal structure of precipitates was observed using TEM. Samples for optical microscope were sectioned and mounted for polishing following the standard metallographic preparation procedures. Kalling s reagent containing 6 gm CuCl 2, 100 ml HCl, 100 ml H 2 O, and 100 ml CH 3 OH was used to reveal the microstructure. Optical microscopy was done with a Leitz Laborux 12 ME optical microscope equipped with a CCD camera and a PC computer for digital recording. A Hitachi S-750 SEM was used to examine b precipitates morphology, and a Tecnai F-20 TEM was utilized to observe the internal microstructure. 2.2. TEM Sample preparation TEM samples cut from as-produced and isothermally exposed materials were prepared using three sample preparation techniques, jet electro-polishing, ion milling and FIB. To prepare TEM sample by the jet electro-polishing technology, 400 500 mm thick foils were cut with a precision diamond saw. The 3 mm discs were mechanically punched from the foils and ground to 150 200 mm thickness. The discs were then electrolytically polished in a twin jet electro-polishing apparatus using a solution of 10% perchloric acid in methanol at 28 V and 250 8C. To prepare sample by the ion milling technique, the 3 mm discs punched from 400 to 500 mm thick foils were attached to a mount using a low melting point thermoplastic polymer. These mounted discs were ground to below 80 mm thickness, and then dimpled to below 10 mm. The dimpling procedure could produce a thin central region in the disc while leaving a thick supporting rim to protect the sample from damage.
L. Ma / Micron 35 (2004) 273 279 275 beam was operated at 30 kev. After the sample was thinned below 10 mm, a TEM-wizard program would be activated and the current of Ga ion beam would be reduced to 300 PA and then 100 PA to further polish the sample down to 100 nm. As the thickness of sample was below 100 nm, the current of Ga ion beam was switched to 50 PA to manually clean the sample. Finally, a thin membrane was plucked onto a holey carbon support film grid via a static glass needle. 3. Results and discussions 3.1. Microstructures and mechanical properties Fig. 1. FIB secondary electron image showing the cross section of a lift-out sample ready to be cut free. After dimpling, the samples were ion beam polished to generate a large electron transparent area using a Gatan Model 691 Precision Ion Polishing System (PIPS) instrument. Directional milling was used to eliminate milling the aluminum stubs. This was accomplished by turning the ion guns on and off during the sample rotation. The double-sided milling was used, while the left gun was tilted 58 from the top of sample and right gun was tilted 48 from the bottom. The voltage of ion beam was kept at 5.0 kv until sample was perforated. As the perforation was detected, the voltage of ion beam was reduced to 2.5 kv and held 20 min. After this, the voltage was reduced to 1.5 kv again and held another 20 min. A FEI Strata DB 235 M instrument, which combines the FIB technology with SEM in a single tool, was also used to prepare TEM samples. This instrument can provide SEM imaging in order to monitor the whole process during FIB milling. Fig. 1 shows SEM photograph of the cross section of alloy 783 sample during milling. Firstly, the samples were thinned to below 10 mm using the high current Ga single ion beam with a current of 5000 PA (Pico Ampere) for initial milling, while the ion The optical microstructures of isothermally exposed and as-produced material are shown in Fig. 2. Both treatments produced an isotropic microstructure with a similar grain size of ASTM 5-7. The residual strain structure introduced by the hot rolling has been removed after annealing. SEM photographs of as-polished materials show numerous plate-like b precipitate networks formed along grain boundaries (Fig. 3). These intergranular precipitates are likely to have formed during b aging. Since the solvus temperature of b phase is approximately 1175 8C and the annealing temperature is 1121 8C, globular b particles formed during processing, which were not completely dissolved during annealing, are shown in the intra- and inter-grain boundaries. Asproduced and isothermally exposed samples show the identical morphology of b precipitates, indicating the prolonged isothermal exposure could not modify the morphology of b precipitates. Table 1 shows the effect of isothermal exposure on tensile properties and hardness of alloy 783 at room temperature and 650 8C. At room temperature, prolonged isothermal exposure marginally increased the yield strength and lowered the elongation, while producing the comparable ultimate strength and hardness. However with increasing of temperature, isothermal exposure caused a remarkable enhancement of ultimate strength and reduction of elongation together with the comparable yield strength. Fig. 2. Optical microstructure of alloy 783. (a) Isothermally exposed; (b) as-produced.
276 L. Ma / Micron 35 (2004) 273 279 Fig. 3. SEM micrograph of b precipitate. (a) Isothermally exposed; (b) as-produced. 3.2. TEM characterization 3.2.1. TEM imaging of samples prepared by jet electro-polishing All samples, prepared by Jet Electro-polishing, Ion Milling, and FIB technique were examined with a Tecnai F-20 transmission electron microscope operating at 200 kev. Fig. 4 shows the TEM photographs of as-produced sample prepared by jet electro-polishing technique. Bright field photo shows that lots of plate-like NiAl-type b precipitates distribute along grain boundary (Fig. 4(a)). Dark field photo shows that numerous ordered cuboidal Ni 3 Al-type g 0 particles with an average diameter of 30 40 nm precipitate in the austenitic matrix nearby grain boundaries (Fig. 4(b)). The existence of grain boundary g 0 Precipitate Free Zone (PFZ) is evident. Fig. 5 compares the TEM photographs of NiAl-type b phase in as-produced and isothermally exposed specimens. It was observed that the well-defined g 0 particles precipitated around b phase, and b phase in the as-produced material is clean and essentially devoid of any internal precipitation (Fig. 5(a)). In contrast, on 3000 h exposure at 593 8C b phase was found to contain numerous spherical g 0 precipitates with an average diameter of 5 10 nm (Fig. 5(b)). Alloy 783 primarily consists of b, g 0, and g phase. It has been reported that b phase contained approximately 30% Ni, 24% Co, 15% Fe, and 31% Al by atom percentage (Mannan and debarbadillo, 1998). Since Co and Fe would occupy Ni lattice sites, the (NiCoFe)Al-type b phase will be Al-rich. This phase has demonstrated the good resistance to SAGBO-induced degradation, such as stress rupture and intergranular cracking. Like other Ni-based superalloys, the ordered Ni 3 Al-type g 0 phase provides strength of alloy 783. Studies have shown that Al-rich b phase in Ni Al binary system could transform to g 0,Ni 5 Al 3, and NiAl-type b phase on exposure in the temperature ranging from 500 to 700 8C (Mannan and debarbadillo, 1998). Therefore, it is not surprising that the ordered spherical g 0 phase precipitated within b phase of alloy 783 on exposure at 593 8C for 3000 h. For as-produced alloy 783, b phase is hardened by substitution solid strengthening with Co and Fe and also by point defects owing to non-equiatomic composition, and alloy strength is primarily provided by g 0 phase. For isothermally exposed material, the whole volume fraction of g 0 phase was increased due to internal precipitation of g 0 phase within b phase, and b phase was dispersion strengthened by g 0 phase compared to the as-produced material. As indicated in Figs. 2 and 3, and Table 1, although the grain size and the morphology of b phase are identical in as-produced and isothermally exposed specimens, their mechanical properties are different. Therefore, the structure modification of b phase due to internal precipitation of g 0 phase is responsible for increase of room and high temperature strength of isothermal exposed specimens, as well as reduction of the high temperature elongation, which is related to the SAGBO-induced degradation. 3.3. TEM Imaging of samples prepared by ion milling Fig. 6 shows the TEM photographs of as-produced and isothermally exposed sample prepared by ion milling technique. As shown in Fig. 6(a), b phase in as-produced Table 1 Comparison of mechanical properties at room temperature (RT) and 650 8C Properties As-produced Isothermal exposed RT 650 8C RT 650 8C s 0:2 (MPa) 852 716 953 741 s b (MPa) 1225 985 1236 1018 d (%) 25.9 36.3 20.2 17.6 Hardness (HRC) 31 34 Fig. 4. TEM photograph of as-produced alloy 783 sample prepared by Jet Electro-polishing. (a) Intergranular NiAl-type b phase; (b) NiAl-type b phase, PFZ, and Ni 3 Al-type g 0 phase.
L. Ma / Micron 35 (2004) 273 279 277 Fig. 5. TEM photograph of b phase is as-produced and isothermally exposed sample prepared by jet electro-polishing. (a) Internal structure b phase in as-produced sample; (b) internal structure of b phase in isothermally exposed sample showing the internal precipitation of g 0 within b phase. material is clean, smooth and lack of any internal precipitates. Compared to the samples prepared by jet electro-polishing as shown in Figs. 4(b) and 5(a), the morphology of ordered g 0 precipitates is somewhat blurred. Numerous tiny particles and holes appear on the matrix, indicating that lots of spherical precipitates were sputtered out from the austenitic matrix during ion milling. These tiny particles and same size holes are also found within b phase in the isothermally exposed sample, suggesting that the internal precipitation occurred within b phase. Therefore, the TEM image analysis for the samples prepared by ion milling technique comes to the same conclusion as illustrated in Fig. 5 that there is internal precipitation of g 0 phase within b phase in alloy 783 on isothermally exposure. Both ion milling and jet electro-polishing techniques allow the TEM observation to clearly show the morphology of b phase, but the TEM images of samples prepared by ion milling cannot reveal the clear morphology of g 0 phase. Since the ion milling technique uses the ion beam with energy of 5.0 kev to bombard the sample to remove the material from sample surface, the bombardment procedure with high energy ion would produce artifacts including preferential sputtering, specimen heating, and radiation on the electron-transparent region of the sample. The ordered spherical Ni 3 Al-type g 0 phase is a primary hardening phase in alloy 783, and its hardness and strength are much higher than the austenitic matrix and b phase. During ion milling process, ion beam may sputter the g 0 phase, b phase and austenitic matrix at different rates causing the preferential sputtering effect. Therefore, the surface of austenitic matrix and b phase with internal precipitation in isothermally exposed material is relatively rough, and b phase without internal precipitation in as-produced sample is clean and smooth. Also, the high energy ion beam is likely to generate the amorphous damage to the g 0 particle resulting in blurred morphology of g 0 phase in TEM image. 3.4. TEM imaging of samples prepared by FIB Of interest to note is that TEM examination of all asproduced and isothermally exposed samples prepared by FIB did not show the reasonable morphology of b and g 0 phase, and the grain boundary profile did not appear. The fact that whole sample surface appeared to be damaged and full of artifacts suggests that the amorphous effects to samples occurred during FIB milling. As seen in Fig. 7(a), TEM imaging of as-produced sample prepared by FIB shows the interface between large plate-like precipitates, which are believed to be b phase, and matrix as seen in Fig. 7(a). However Fig. 7(a) also shows that there are lots of unknown particles, which are thought to be artifacts, attached on the b phase, and the ordered g 0 particles could not be observed on the matrix. Unlike the isothermal exposed sample prepared by jet electro-polishing and ion Fig. 6. TEM photograph of b phase in as-produced and isothermally exposed sample prepared by ion milling technique. (a) Internal structure b phase in asproduced sample; (b) internal structure of b phase in isothermally exposed sample showing the internal precipitation within b phase.
278 L. Ma / Micron 35 (2004) 273 279 Fig. 7. TEM photograph of b phase in as-produced and isothermally exposed sample prepared by FIB. (a) b Phase in as-produced sample; (b) b phase in isothermally exposed sample; (c) high magnification view of (b). milling, the ordered g 0 precipitates did not appear within b phase in the isothermally exposed sample prepared by FIB as seen in Fig. 7(b). Many unknown particles are attached on the surface of b phase. High magnification view of Fig. 7(b) as seen in Fig. 7(c) shows the amorphous structure of b phase, and the internal precipitation is not observed. Study has indicated that for sample prepared with the FIB technique, the beam of Ga þ with an accelerated voltage of 30 50 kev can induce the damage of sample surface, and then produce artifacts including implantation, amorphization, and mixing (Young et al., 1998). In this study, the surface damage caused by the action of the high energy Ga þ beam was likely to occur during milling. Although the low energy cleaning was used in the final polishing, the artifacts seemed not to be removed completely. As a result, the TEM images show a shallow artifact layer, which may suppress the details of b and g 0 phase. 4. Summary Jet electro-polishing, ion milling and FIB technique was utilized to prepare TEM sample of Ni-based INCONEL alloy 783, subjected to the standard heat treatment and prolonged isothermal exposure. TEM examination indicates that the jet electro-polishing can produce a large artifactfree electron transparent region, wherein the detailed microstructure of alloy 783 can be well observed. The results of TEM analysis for samples prepared by jet electropolishing demonstrate the internal precipitation of g 0 within b phase is consistent with the early studies as well as is able to be employed to explain the change of mechanical properties of alloy 783 due to isothermal exposure. Ion milling technique induces artifacts on the austenitic matrix of both as-produced and exposed samples so that the morphology of g 0 phase cannot be well revealed, but the structure of large plate-like b phase is observable. This study shows FIB causes the extensive artifacts including surface damage and amorphous structure for the asproduced and exposed samples. Therefore, the TEM images of these samples cannot be used to illustrate the microstructure and property variation of alloy 783 due to long term exposure. All above mentioned illustrations suggest that the jet electro-polishing technique is the best method to prepare the TEM sample of superalloy 783, as well as FIB and ion milling techniques should be very carefully used to prepare the samples of dispersion-strengthening metallic materials, ensuring that the artifacts can be minimized. Acknowledgements The author would like to thank Dr Y.-C. Wang and J. Ringnalda in FEI Company, Hillsboro, Oregon, and Mrs R. Alani and M.A. Izquierdo in Gatan Inc., Pleasanton California, for assistance in sample preparation and TEM observation. References Ma, L.Z., Chang, K.-M., Mannan, S.K., Patel, S.J., 2000a. Effect of thermomechanical processing on fatigue crack propagation in INCONEL alloy 783. In: Pollock, T.M., Kissinger, R.D., Bowman, P.R., Green, K.A., McLean, M., Olson, S., Schira, J.J. (Eds.),
L. Ma / Micron 35 (2004) 273 279 279 Proceedings of the Ninth International Symposium on Superalloys, TMS, Seven Springs, PA, pp. 601 609. Ma, L.Z., Chang, K.-M., Mannan, S.K., 2000b. Effect of heat treatment on fatigue crack propagation in INCONEL 783. In: Chang, K.-M., Srivastava, S.K., Furrer, D.U., Bain, K.R. (Eds.), Advanced Technologies for Superalloy Affordability, TMS Annual Meeting, Nashville, TN, pp. 131 140. Ma, L.Z., Chang, K.-M., Mannan, S.K., Patel, S.J., 2002. Effect of prolonged isothermal exposure on elevated-temperature, time-dependent fatigue-crack propagation in INCONEL alloy 783. Metall. Trans. A 33A, 3465 3478. Mannan, S., debarbadillo, J., 1998. Long term thermal stability of INCONEL alloy 783. International Gas Turbine and Aeroengine Congress and Exhibition, Stockholm, Sweden, June 2 5, pp. 1 10. Phaneuf, M.W., 1999. Application of focused ion beam microscopy to materials science specimens. Micron 30, 277 288. Williams, D.B., Barry Carter, C., 1996. Transmission Electron Microscopy I. Plenum Press, New York. Smith, J.S., Heck, K.A., 1996. Development of low thermal expansion, crack growth resistant superalloy. In: Kissinger, R.D., Deye, D.J., Anton, D.L., Cetel, A.D., Nathal, M.V., Pollock, T.M., Woodford, D.A. (Eds.), Proceedings of the Eighth International Symposium on Superalloys, TMS, Seven Springs, PA, pp. 91 100. Young, R.J., Carleson, P.D., Da, X., Hunt, T., Walker, J.F., 1998. Highyield and high-throughput TEM sample preparation using focused ion beam automation, In: Proceedings of the 24th International Symposium for Testing and Failure Analysis, ASM International, Dallas, Texas, November, pp. 329 336.