Effects of focused ion beam milling on austenite stability in ferrous alloys

Transcription

1 MATERIALS CHARACTERIZATION 61 (2010) 1 6 available at Effects of focused ion beam milling on austenite stability in ferrous alloys K.E. Knipling, D.J. Rowenhorst, R.W. Fonda, G. Spanos Naval Research Laboratory, 4555 Overlook Ave, SW Washington, DC 20375, USA ARTICLE DATA Article history: Received 20 March 2009 Received in revised form 16 September 2009 Accepted 21 September 2009 Keywords: Martensitic steels Serial sectioning Focused ion beam milling Electron backscatter diffraction (EBSD) ABSTRACT The susceptibility of fcc austenite to transform to bcc during focused ion beam milling was studied in three commercial stainless steels. The alloys investigated, in order of increasing austenite stability, were: (i) a model maraging steel, Sandvik 1RK91; (ii) an AISI 304 austenitic stainless steel; and (iii) AL-6XN, a super-austenitic stainless steel. Small trenches were milled across multiple austenite grains in each alloy using a 30 kv Ga + ion beam at normal incidence to the specimen surface. The ion beam dose was controlled by varying the trench depth and the beam current. The factors influencing the transformation of fcc austenite to bcc (listed in order of decreasing influence) were found to be: (i) alloy composition (i.e., austenite stability), (ii) ion beam dose (or trench depth), and (iii) crystallographic orientation of the austenite grains. The ion beam current had a negligible influence on the FIB-induced transformation of austenite in these alloys. Published by Elsevier Inc. 1. Introduction Focused ion beam (FIB) milling has emerged as an effective specimen preparation tool in materials science. In a FIB instrument, a finely focused ion beam from a liquid metal ion source (usually Ga + ) impacts the specimen, leading to localized sputtering that can be controlled to mill nanometer-scale features. A dual beam FIB instrument, which also contains an electron source, can additionally be equipped with an electron backscatter diffraction (EBSD) detector for crystallographic orientation analysis. This combination enables the dual beam FIB to acquire crystallographic information in three dimensions by sequentially milling the sample with the FIB and analyzing each freshly produced surface by EBSD [1 6]. Irradiating a solid with a beam of energetic ions, however, invariably results in some degree of microscopic disruption and heat generation. The various interactions that can occur between the ion beam and the specimen have been reviewed by several authors (e.g., see [7 13]). In crystalline materials, ion beam irradiation can generate extensive lattice defects which, for a sufficient implantation dose, can lead to phase transformations and/or complete amorphization in the material. The present study investigates the stability of austenite in three steels as a function of several FIB milling conditions. The three steels examined represent various levels of inherent austenite stability, based on their alloy content. The austenite stability was evaluated in each alloy as a function of ion dose and beam current to determine the relative response to those milling parameters. This study was undertaken in large part to provide a critical knowledge base for potential FIB-based three-dimensional analytical experiments in ferrous alloys containing significant volume fractions of austenite, but it is also relevant to the FIB preparation of transmission electron microscopy specimens. Use of such FIB milling techniques for 3D microstructural studies has been reported previously in many non-ferrous alloy systems (e.g., see [1,4 6,14]) but less frequently in ferrous alloys (e.g., [1,2,15]). Corresponding author. Tel.: address: knipling@anvil.nrl.navy.mil (K.E. Knipling) /$ see front matter. Published by Elsevier Inc. doi: /j.matchar

2 2 MATERIALS CHARACTERIZATION 61 (2010) Experimental Procedures Three commercial stainless steels were investigated in this study: (i) a model maraging steel, Sandvik 1RK91, which forms isothermal martensite starting at temperatures around 40 C and achieves a maximum transformation rate at 40 C [16]; (ii) an AISI 304 metastable austenitic stainless steel; and (iii) an AL-6XN super-austenitic stainless steel. These three alloys were chosen to represent varying degrees of inherent austenite stability based on their increasing levels of alloying elements. Published alloy compositions are provided in Table 1. Metallographic preparation of the specimens was carried out using conventional techniques. Final polishing of the surface was accomplished with a solution of 20% hydrogen peroxide (30%) and 80% colloidal silica solution. The specimens were examined in an FEI Nova 600 NanoLab DualBeam TM SEM/FIB equipped with an HKL Channel 5 EBSD system. Some EBSD analyses were also performed with a JEOL JSM-7001F SEM, equipped with an EDAX-TSL Hikari EBSD detector. Rectangular trenches (usually ~5 5 µm 2 ) were milled using a 30 kv Ga + beam at normal incidence to the specimen surface to assess the effect of ion beam irradiation on the austenite stability. While most practical milling situations (e.g., in serial sectioning or in transmission electron microscopy specimen preparation) utilize high glancing angles of incident radiation, normal incidence was used in this study to provide a more reproducible ion dose and to confine the resulting damage to a well-defined region suitable for subsequent EBSD analysis. The sputtering yield is significantly less for incident radiation as compared with high glancing angles, but the relative damage induced by the beam is comparable [7,9]. The trenches were based on preprogrammed raster patterns in the FEI control software, which are defined by a number of parameters. The pattern depth and beam current were varied to adjust the delivered dose of the ion beam. The depth of the milled pattern (a nominal value based on the sputtering yield of silicon) is directly proportional to the delivered dose. The beam current (the rate of delivery of ions to the specimen) was varied from 0.1 to 5.0 na. The dwell time (the duration that the beam spends at each step) and the overlap (the distance between adjacent steps) were held constant at 1 μs and 50% overlap, respectively. The resulting damage and austenite transformation were examined using electron forescatter detector orientation contrast imaging and EBSD. 3. Results and Discussion A typical forescatter electron image of the polished surface of the Sandvik 1RK91 alloy is shown in Fig. 1(a). EBSD analyses of this and other regions indicated that the initial microstructure was primarily martensitic, with a minor constituent of small (~3 μm) equiaxed austenite grains. The surface morphology of the martensite and austenite constituents appeared similar in secondary electron images, although forescatter electron images typically displayed more visible striations in the martensitic material in the Sandvik 1RK91 alloy. Three austenitic grains are contained within the indicated boxed area, as revealed (in red) in the superimposed EBSD phase map in Fig. 1(b). Because the volume fraction of austenite was relatively small, a series of narrow trenches, 0.5 μm wide and 10 μm long, was milled across the austenite grains. Milling was performed with a 1.0 na beam current of 30 kv Ga + ions at normal incidence to the specimen surface with an ion dose of Ga + ions/cm 2. Fig. 1(a) shows a trench milled through one of the austenite grains. An EBSD phase map of this region, Fig. 1(b), indicates that the fcc (red) austenite in that milled region had transformed to bcc (blue). Fig. 1(c) shows the result of milling a second trench through a different austenite grain. A comparison of the superimposed phase identification maps in Fig. 1(b) and (c) shows unambiguously that the FIB beam has effected a transformation from fcc austenite to bcc in the milled region. The milling-induced decomposition of fcc austenite to a bcc phase in the Sandvik 1RK91 alloy led to a parametric study of the factors influencing such austenite decomposition. Two additional austenitic alloys with increasing levels of austenite stability (as indicated by the alloying constituents, Table 1) were selected for this study, in large part because of the marginal austenite stability of the Sandvik steel, which partially transforms to isothermal martensite at room temperature [16]. The current study thus focused predominantly on AISI 304 stainless steel due to its enhanced austenite stability as well as its widespread use and availability. An AL-6XN super-austenitic steel was also examined to evaluate the susceptibility to milling-induced transformation in an alloy with an even higher inherent austenite stability. To evaluate the effects of ion beam dose on the stability of austenite, a series of 5 5 μm 2 trenches was milled in the AISI 304 stainless steel using a beam current of 0.1 na. The Ga + ion dose, calculated from the beam current and required milling time, was varied from 4.2 to Ga + ions/cm 2, by varying the pattern depth. The forescatter electron micrograph in Fig. 2(a) shows the expected increase in trench depth with increasing ion dose. After milling, the microstructure was analyzed by EBSD, as shown in Fig. 2(b) and (c). The original microstructure was entirely austenitic and remained untransformed for doses less than Ga + ions/cm 2. For moderate doses (13 to Ga + ions/cm 2 ), the transformation was nonuniform across the milled area. For the most severe milling conditions (doses greater than Ga + ions/cm 2 ), all of the milled area was transformed to bcc. Table 1 Compositions (wt.%) of alloys investigated [16 18]. Ni Cr Mo Mn Si Ti Al N Cu C P S Sandvik 1RK <0.01 AISI AL-6XN

3 MATERIALS CHARACTERIZATION 61 (2010) Fig. 1 Localized FIB-induced damage in Sandvik 1RK91. (a) Forescatter detector electron micrograph showing a milled trench through one of the austenite grains. (b) Phase identification map where red is fcc and blue is bcc. The milled part of the austenite (fcc) grain has transformed to bcc. (c) A second milled trench through a different austenite grain, again showing the transformation in the irradiated area. The delivered dose of the ion beam was Ga + ions/cm 2. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Because the area investigated in Fig. 2 contained only a single predominant austenite orientation, a series of larger trenches, spanning multiple grain orientations, was milled in the AISI 304 stainless steel using a beam current of 1.0 na see Fig. 3. The delivered ion dose was Ga + ions/cm 2, which is near the threshold of austenite to bcc transformation observed in Fig. 2. Three of the austenite grains in Fig. 3 (labeled A, B, and C) were aligned near the three low-index directions (<100>, <110>, and <111>, respectively), as indicated by their red, blue and green colors in the inverse pole figure map of Fig. 3(b). Qualitatively, grain B (<111> surface normal, blue) transforms much more readily than either grain A (<100> surface normal, red) or grain C (<110> surface normal, green), while grain A (<100> surface normal, red) appears to transform slightly more readily than grain C (<110> surface normal, green). This demonstrates that crystallographic orientation also plays a strong role in determining the degree of transformation. It is well known that during FIB milling, the sputtering yield can vary as a function of crystallographic orientation within the specimen [19 21]. This so-called channeling effect is related to the ease with which ions are able to penetrate the crystal lattice before being scattered and is thus related to both the crystallographic orientation of the specimen and the crystal structure. Channeling appears to be the basis for the strong crystallographic effect observed in Fig. 3. Channeling is highest along the austenite <110> directions, where the Ga + ions can penetrate deeper between the closely packed {111} planes before being scattered. Similarly, channeling is lowest along the austenite <111> directions and the Ga + ions scatter Fig. 2 Series of 5 5 μm 2 trenches milled in AISI 304 stainless steel using a beam current of 0.1 na and beam doses of Ga + ions/cm 2. (a) Forescatter detector electron micrograph. (b) Phase identification map where red is fcc and blue is bcc. (c) Inverse pole figure map showing the initial orientation of austenite. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

4 4 MATERIALS CHARACTERIZATION 61 (2010) 1 6 Fig. 3 Series of μm 2 milled trenches spanning multiple grain orientations in AISI 304 stainless steel. FIB milling was performed using a beam current of 1.0 na and a beam dose of Ga + ions/cm 2. (a) EBSD image quality map showing the regions exposed to ion beam damage. (b) Inverse pole figure map showing the orientation of the fcc austenite. (c) Inverse pole figure map, showing the orientation of the transformed bcc phase. closer to the surface. Thus, austenite grain orientations aligned favorably for channeling are more resistant to transformation, whereas those oriented for less channeling, which also correspond to a higher sputtering yield (and damage), are more susceptible to transformation to bcc. The transformation often produces a consistent orientation in the transformed bcc phase, as seen in grains B and C. By examining the orientations of several regions, it is found that the close packed {111} planes in the original fcc grain are parallel with the close packed {110} planes in the transformed bcc. Similarly, the close packed <1 10> fcc directions are parallel with the close packed <1 11> bcc directions, suggesting a near Kurdjumov Sachs orientation relationship. In addition, some areas produce inconsistent transformation orientations, which may indicate a higher degree of damage, or may be a reflection of the deformation already present in some of the parent crystals (such as in crystal A in Fig. 3). To determine the effects of milling rate on the threshold for austenite decomposition, specimens of AISI 304 were milled at a constant ion dose of Ga + ions/cm 2 with beam currents (milling rates) ranging from 0.1 to 5.0 na. Ion milling under these conditions failed to transform the fcc austenite throughout this range of beam currents. However, partial transformation of the austenite was accomplished at total doses of Ga + ions/ cm 2 and was nearly complete at total doses greater than Ga + ions/cm 2, irrespective of the beam current (within the range of 0.1 to 5.0 na). Thus, variations in the milling rate (by changing the beam current) have a negligible effect on the transformation probability, particularly when compared with the effect of total delivered ion dose. A super-austenitic stainless steel, AL-6XN, was also milled under a variety of beam conditions to evaluate the austenite transformation susceptibility in an alloy with even greater austenite stability. Fig. 4 shows a series of six trenches in AL- 6XN, which were milled at beam currents ranging from 0.1 to 5.0 na and a total ion dose of Ga + ions/cm 2. The milled areas are readily apparent in the forescatter electron image, Fig. 4(a). However, the corresponding EBSD phase identification map, Fig. 4(b), and the austenite EBSD orientation map, Fig. 4(c), indicate that there was no transformation of the austenite observed at any of the beam currents studied. Additional EBSD analyses performed at 5.0 na indicated that this alloy did not transform even after exposure to ion doses as high as Ga + ions/cm 2. (Beyond this depth, EBSD indexing becomes problematic, as the trench walls block the electrons emitted from the specimen surface). Thus, an ionbeam-induced decomposition of fcc austenite to bcc was never observed in this alloy. This study provides some insights into the mechanism by which the ion beam transforms the fcc austenite to bcc. The

5 MATERIALS CHARACTERIZATION 61 (2010) Fig. 4 Series of 5 5 μm 2 trenches milled in alloy Al-6XN at various beam currents (as indicated in na) at a delivered ion dose of Ga + ions/cm 2. (a) Forescatter detector electron micrograph. (b) Phase identification map where red is fcc and blue is bcc. No transformation of austenite was observed. (c) Inverse pole figure map, showing the orientation of austenite. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) dependence on crystallographic orientation suggests that the damage introduced into the material, as modified by channeling effects, is a significant contributor to this transformation. Because Ga is a ferrite stabilizer, it likely also contributes to the driving force for conversion to bcc. Using the method of Ishitani and Kaga [8], we have estimated that the temperature of the specimen during FIB milling would rise less than 15 C for the conditions used here, which has a negligible influence. Finally, the observance of a Kurdjumov Sachs (K S) orientation relationship between the bcc phase and the fcc austenite from which it forms indicates that the transformation occurs directly from the fcc austenite. 4. Summary and Conclusions The effect of focused ion beam milling on the stability of the fcc austenite phase and its transformation to the bcc structure was studied in three commercial stainless steels with increasing austenite stability: (i) a model maraging steel, Sandvik 1RK91; (ii) AISI 304 austenitic stainless steel; and (iii) AL-6XN, a super-austenitic stainless steel. Using a 30 kv Ga + ion beam at normal incidence to the specimen surface, the milling conditions were varied to evaluate their effects on the transformation of austenite in these alloys. The factors influencing the transformation of the austenite (listed in order of decreasing importance) were found to be: (i) alloy composition (i.e., austenite stability), (ii) ion beam dose or trench depth, and (iii) crystallographic orientation of the austenite grains. The FIB-induced transformation of austenite in these alloys was not observed to be influenced by the ion beam's current. REFERENCES [1] Zaefferer S, Wright SI, Raabe D. Metall Mater Trans, A 2008;39A: [2] Inkson BJ, Mulvihill M, Mobus G. Scripta Mater 2001;45: [3] Konrad J, Zaefferer S, Raabe D. Acta Mater 2006;54: [4] Zaafarani N, Raabe D, Singh RN, Roters F, Zaefferer S. Acta Mater 2006: [5] Groeber MA, Haley BK, Uchic MD, Dimiduk DM, Ghosh S. Mater Charact 2006;57: [6] Uchic MD, Holzer L, Inkson BJ, Principe EL, Munroe P. MRS Bull 2007;32: [7] Gianuzzia Lucille A. Introduction to focused ion beams: instrumentation, theory, techniques and practice. Springer; [8] Ishitani T, Kaga H. J Electron Microsc 1995;44: [9] Ishitani T, Yaguchi T. Microsc Res Tech 1996;35: [10] Giannuzzi LA, Stevie FA. Micron 1999;30: [11] Matteson TL, Schwarz SW, Houge EC, Kempshall BW, Giannuzzi LA. J Electron Mater 2002;31:33 9.

6 6 MATERIALS CHARACTERIZATION 61 (2010) 1 6 [12] Prenitzer BI, Urbanik-Shannon CA, Giannuzzi LA, Brown SR, et al. Microsc Microanal 2003;9: [13] Mayer J, Giannuzzi LA, Kamino T, Michael J. MRS Bull 2007;32: [14] Mulders JJL, Day AP. Mater Sci Forum 2005; : [15] Keehan E, Karlsson L, Bhadeshia HKDH, Thuvander M. Mater Charact 2008;59: [16] Holmquist M, Nilsson JO, Stigenberg AH. Scripta Metall Mater 1995;33: [17] Metals handbook. American Society for Metals; p. 5. [18] AL-6XN sourcebook. Allegheny Technologies; [19] Phillips JR, Griffis DP, Russell PE. J Vac Sci Technol, A 2000;18: [20] Kempshall BW, Schwarz SM, Prenitzer BI, Giannuzzi LA, et al. J Vac Sci Technol, B 2001;19: [21] Yahiro Y, Kaneko K, Fujita T, Moon WJ, Horita Z. J Electron Microsc 2004;53:571 6.