Influence of Mn Content on the Microstructure and Mechanical Properties of Ultrafine Grained C Mn Steels

, pp. 1721 1726 Influence of Mn Content on the Microstructure and Mechanical Properties of Ultrafine Grained C Mn Steels Rongjie SONG, Dirk PONGE and Dierk RAABE Max-Planck-Institut für Eisenforschung, Max-Planck-Str. 1, 40237 Dusseldorf, Germany. E-mail: ponge@mpie.de; raabe@mpie.de (Received on February 17, 2005; accepted on July 6, 2005) The effect of Mn content on the microstructure and mechanical properties of two ultrafine grained 0.2%C Mn steels has been investigated. The ultrafine grained microstructure was produced by use of large strain warm deformation and subsequent annealing. The final microstructure consists of fine cementite particles within an ultrafine grained ferrite matrix. The increase in the Mn content leads to a decrease in the average ferrite grain size (from 1.3 to 0.8 mm for an increase in the Mn content from 0.74 to 1.52 mass%). This can be attributed to the enrichment of Mn in the cementite particles, which becomes finer in the steel with a higher Mn content. The increase in the Mn content results in an increase in strength at equal ductility and toughness. KEY WORDS: ultrafine grained steel; Mn; EDS; cohenite; ductility; toughness. 1. Introduction In recent years, ultrafine grained steels have attracted much attention due to their excellent strength and toughness. 1,2) However, due to a low work hardening rate of single phase ultrafine grained steels the tensile ductility at room temperature (such as the uniform elongation) is rather low. 3 5) While the static mechanical properties (e.g. tensile properties) of ultrafine grained steels have been discussed in the literature, 3,6,7) their dynamic mechanical properties 8) (e.g. Charpy impact properties) are seldom studied due to a limitation in the sample sizes produced. Majka et al. 9) have shown that Mn is responsible for the development of microstructural banding which is due to the segregation of substitutional alloying elements during dendritic solidification in low-alloy steels. The effects of the Mn contents on the constitution and morphology of the phases, which affect the mechanical properties, have also been studied. 10) However, no investigation on the effect of the Mn contents on grain refinement was carried out in the field of ultrafine grained steels. This study presents an approach to enhance the uniform elongation of ultrafine grained steels by the introduction of a second phase in the form of a homogeneous dispersion of fine particles in the ultrafine grained matrix. The microstructure and the mechanical properties of two ultrafine grained plain 0.2%C Mn steels (with Mn contents of 0.74 and 1.52 mass%, respectively) with a dispersion of globular cementite, produced by large strain warm deformation and annealing, 11 13) are investigated. The effects of the Mn contents on grain refinement and on the resulting mechanical properties are discussed in detail. 2. Experimental Procedures Table 1 shows the compositions of the two plain C Mn steels (hereafter referred to as steel 2C and steel 2CMn, respectively) studied in this work. The A e3 temperatures (equilibrium austenite to ferrite transformation temperature) of the steels shown in Table 1 were calculated by using Thermocalc. 14) The laboratory samples were machined directly from the cast ingot into rectangular parallelepiped samples of 50 40 60 mm 3 (width length height). The plane strain compression tests were conducted by use of a large scale 2.5 MN hot press, 15) where the compression direction was parallel to the sample height. After reheating the specimens at a heating rate of 10 K/s, the samples were austenitized at 100 K above A e3 for 3 min. After air cooling to 1 143 K, a one-step deformation treatment was imposed with a logarithmic strain of e 0.3 at a strain rate of 10 s 1. This was followed by a controlled cooling down to pearlite finish temperature in order to obtain a bainite free ferrite pearlite microstructure. The cooling rate was 6.5 and 5.0 K/s for the steel 2C and 2CMn, respectively. After holding at 823 K for 2 min, the large strain warm deformation was performed by exerting a four-step plane strain compression series with an inter-step time of 0.5 s. Each of the four subsequent deformation steps imposed a logarithmic strain of e 0.4 at a strain rate of 10 s 1. For both steels the temperature increased during the Table 1. Chemical compositions (mass%) and calculated A e3 temperatures (K). 1721 2005 ISIJ

deformation, and reached to approximately 923 K by the end of the last deformation step. Subsequently an annealing treatment of 2 h at 823 K was conducted. The distribution of C and Mn in the microstructure was qualitatively investigated by use of energy-dispersive spectrometry (EDS) operated in a scanning electron microscope (SEM), and was quantitatively measured using EDS operated in a transmission electron microscope (TEM), (see below). Small cementite particles with the diameter of particles less than 500 nm can not be well analyzed by the EDS technique due to its limited resolution of 500 nm in the case of cementite. Therefore, a high temperature annealing was applied to obtain a larger size of the cementite particles. For this purpose the steel 2C was annealed at an elevated temperature of 973 K for 2 h after the large strain warm deformation at 823 K. The samples for TEM observation were produced by mechanical polishing to a thickness of about 50 nm, followed by electropolishing using a Tenupol double jet instrument in 95%C 2 H 4 O 2 5%HClO 4 at 288 K and 40 V. The TEM observations were carried out using a Philips CM20 equipped with an EDS system and a scanning device. The chemical composition of the particles was determined by EDS measurements which were performed using the nanobeam mode of the electron beam position in the scanning mode (STEM) of the microscope. The sizes of the cementite particles were determined by mean linear intercepts from the scanning electron microscope (SEM) and bright field TEM images. The ferrite grain size (counting only crystals with grain boundary misorientations above 15 ) was determined by use of experimentally obtained electron backscatter diffraction (EBSD) maps in conjunction with the mean linear intercept method. 11) The grain size was defined as the average diameter of the equivalent area circles which match the area of the elliptically shaped grains. In the present study, highangle grain boundaries (HAGB) were defined as homophase interfaces with a misorientation angle of q 15. Lower values of the local misorientation (2 q 15 ) represent low-angle grain boundaries (LAGB). Misorientations below 2 were neglected in the study. Tensile properties were measured by using tensile test specimens with a cylindrical cross section (f 5 mm) and a gauge length of 25 mm, which were machined in accordance with the corresponding ASTM standard. Tensile tests were conducted at room temperature with a constant crosshead speed of 0.5 mm/min. Impact tests were conducted in a temperature range from 103 to 423 K. Subsize Charpy V-notched specimens with a ligament size of 3 4mm 2 were machined along the rolling direction (RD) according to the German Industry Norm DIN 50 115. The V-notch was along the normal direction (ND) (standard longitudinal sample). 3. Results 3.1. Microstructure Figure 1 shows the TEM micrographs of the two steels after the large strain warm deformation (e 1.6) and subsequent annealing treatment at 823 K for 2 h. The microstructure of the steels consists of ultrafine ferrite grains and Fig. 1. TEM micrographs of the ultrafine grained steels after the large strain warm deformation (e 1.6) and subsequent 2h annealing at 823 K. (a) Steel 2C (0.74 mass% Mn). Arrows 1 point out the fine cementite particles inside the ferrite grains; arrows 2 point out the coarse cementite particles at the ferrite grain boundaries. (b) Steel 2CMn (1.52 mass% Mn), see Table 1. globular cementite particles. Two different size groups of cementite particles are observed in the steel 2C (Fig. 1(a)). The finer cementite particles (5 90 nm) are distributed inside the ferrite grains (see arrows 1). Planar arrays of larger cementite particles (90 350 nm) are located at the ferrite grain boundaries (see arrows 2), acting as obstacles for the migration of ferrite grain boundaries. In contrast to this, in the steel 2CMn, homogeneously dispersed fine cementite particles (5 120 nm) (not clearly dividable in two size groups) were observed (Fig. 1(b)). The volume fraction of the cementite particles is dependent on the carbon content, and it was evaluated to be approximately 3.2% for the steel with 0.2% C by thermodynamic equilibrium calculations. With an increase in the Mn content from 0.74 to 1.52 mass% the average ferrite grain size becomes smaller (from 1.3 to 0.8 mm) and the grain shape becomes more equiaxed. The increase in the Mn content leads also to an increase in the fraction of high-angle grain boundaries from 64 to 73%. Figure 2 shows the dislocation structures and the individual cementite particles inside the ferrite matrix in the two steels. The black arrows point out the cementite particles pinning the dislocations in the steels. 2005 ISIJ 1722

Fig. 3. Mechanical properties of the ultrafine grained steels 2C and 2CMn processed by the large strain warm deformation (e 1.6) and subsequent 2 h annealing at 823 K. (a) Strength; (b) ductility. Error bars fall within the symbols. Fig. 2. Dislocation structures and individual cementite particles inside the ferrite matrix of the steels processed by the large strain warm deformation (e 1.6) and subsequent 2 h annealing at 823 K. (a) Steel 2C (0.74 mass% Mn); (b) steel 2CMn (1.52 mass% Mn), see Table 1. 3.2. Mechanical Properties Figure 3 shows the tensile properties of the ultrafine grained steels tested at room temperature. Each dot represents an average value obtained from three separate tensile tests. The increase in the Mn contents from 0.74 to 1.52 mass% leads to an increase in the lower yield stress as well as in the ultimate tensile stress by about 80 MPa, Fig. 3(a). However, the difference in the Mn contents does not affect the ductility, as is documented by the total elongation of about 22%, the uniform elongation of about 10% and the reduction in area of about 65% in both steels, Fig. 3(b). Figure 4 shows the Charpy impact transition curves for the two types of ultrafine grained steels. Both steels have quite similar transition curves. 3.3. Distribution of Mn and C 3.3.1. Energy-dispersive Spectrometry Analysis Figure 5(a) shows the image quality map of the steel 2C (0.74 mass% Mn) after the large strain warm deformation (e 1.6) at 823 K and subsequent annealing at 973 K for 2 h. The microstructure in Fig. 5(a) consists of ferrite and globular cementite. The arrows in Fig. 5(a) point out two individual cementite particles. Figures 5(b) and 5(c) show the lateral distributions of C and Mn, respectively. The gray band from white to black displays an increase of the local C or Mn concentration. The arrows in Figs. 5(b) and 5(c) Fig. 4. Charpy impact transition curves for subsize specimens of the steels 2C and 2CMn processed by large strain warm deformation (e 1.6) and subsequent 2 h annealing at 823 K. point out the locations of the two cementite particles highlighted in Fig. 5(a), respectively. It shows that besides C, Mn is also enriched inside the cementite particles. 3.3.2. TEM Analysis By use of EDS measurement which was performed using the nano-beam mode of the electron beam position in the scanning mode of the microscope, the amount of Mn in the ferrite matrix and cementite particles can be quantitatively determined. Figure 6 shows the occurrence of Mn in both matrix and carbides for the steels processed by large strain warm deformation and subsequent annealing. Each symbol represents an average value from five to ten measurements. For both steels, the Mn content in the ferrite matrix is close to the respective nominal composition, i.e. 0.74 and 1.52 mass% for the steel 2C and 2CMn, respectively. However, it 1723 2005 ISIJ

Fig. 5. EBSD (a) and EDS (b, c) maps of the steel 2C (0.74 mass% Mn) processed by large strain warm deformation (e 1.6) at 823 K and subsequent annealing at 973 K for 2 h. (a) Kikuchi image quality map; (b) C distribution; (c) Mn distribution. Fig. 6. Manganese contents in the matrix and in the cementite particles, which are determined by the STEM-EDS technique, for the steels 2C and 2CMn processed by large strain warm deformation (e 1.6) and subsequent 2 h annealing at 823 K. is important to note that the Mn content in the cementite particles in the steel 2CMn (about 7.5 mass%, Table 1) is significantly higher than that in the steel 2C (about 3.0 mass%, Table 1). 4. Discussion 4.1. Effect of the Mn Content on the Microstructure After the large strain warm deformation and subsequent annealing, the steel with a higher Mn content shows improved microstructural characteristics when compared to the steel with a lower Mn content. In Fig. 1(b), the steel 2CMn showed an ultrafine ferrite microstructure with an average grain size of 800 nm, a nearly equiaxed grain shape, and a homogeneous distribution of the cementite particles. This clear effect shows that the influence of Mn on grain refinement deserves a more detailed discussion. Mn affects the microstructure in different ways during the processing. On the one hand, the addition of Mn significantly retards the formation of ferrite and pearlite by shifting the transformation to a lower temperature range. Thus, a higher Mn content leads to a finer transformed ferrite pearlite microstructure. On the other hand, Mn will partition differently between cementite and ferrite at thermodynamic equilibrium. As a carbide-forming element, it will be more concentrated in the cementite. 16) According to the work by Zhang and Schleich, 17) the complex of Mn Fe carbide (Fe 3 X Mn X C) is termed cohenite. During the formation of pearlite, Mn will not partition between ferrite and cementite since the transformation time is too short for a significant substitutional diffusion of Mn. After applying the large strain warm deformation procedure to the transformed ferrite pearlite microstructure, the diffusion of Mn is enhanced. This is due to the presence of a high dislocation density and an excess vacancy concentration. During and after the recovery process, which may partially proceed dynamically, the introduction of high-angle grain boundaries into the microstructure additionally facilitates the diffusion of Mn. Therefore, it is conceivable that a cohenite Fe 3 C type phase, Fig. 5, can be stabilized by further substituting Fe by Mn. According to Fig. 6, there is a larger enrichment of Mn in the cementite particles in the steel with a higher Mn content. The results correspond approximately to a molecular formula of Fe 29 MnC 10 and Fe 11 MnC 4 for the steel with the lower and the higher Mn content, respectively. During the large strain warm deformation and in the course of the subsequent annealing, smaller particles of Fe 3 X Mn X C may go into solution and the larger particles grow at the expense of the smaller ones. The Gibbs Thomson equation 18) shows that smaller particles have a higher concentration of solute atoms at the matrix/particle interface than larger ones, so that a redistribution of C and Mn occurs between smaller and larger particles. The process of Ostwald ripening is, therefore, more likely controlled by the slow substitutional diffusion of Mn instead of the fast interstitial diffusion of C. 19) As a result, a slow Ostwald ripening could be expected in the steel with a higher Mn content. 20) This leads to a smaller size of particles in this steel, Fig. 2(b). However, even in the steel with a lower Mn content the particles are still fine enough to pin the ferrite grain boundaries during the 2 h annealing at 823 K. Therefore, ultrafine grained microstructures can be obtained in both steels, Fig. 1. 2005 ISIJ 1724

According to the study by Tsuji et al., 21) in a single phase IF steel the increase in strain, assisted by recovery, leads to a refined substructure and to an increase in the misorientations between the gradually evolving neighboring subgrains. In his work a large fraction of high-angle grain boundaries (70 80%) was obtained after a severe plastic deformation of e 5.6 at 773 K. 21) In our present investigation a similar microstructural development was observed, however, at a significantly lower strain of e 1.6 and a higher deformation temperature of 823 K. This can be attributed to the presence of fine particles in the present case. Very fine particles dispersed inside the ferrite matrix may exert a strong pinning effect even on individual dislocations at certain temperatures, 13) Fig. 2. Geometrically necessary dislocations will accumulate around such particles. It is likely, that this facilitates the formation of new subgrain boundaries inside the ferrite grains. At a later stage of the procedure new high-angle grain boundaries can form preferably at such particle-stabilized subgrain boundaries. 22) As a result, this mechanism might promote the formation of fine continuously recrystallized ferrite grains. On the other hand, the presence of these fine particles drastically retards the softening of the ferrite. Both the formation and growth of the ferrite subgrains are controlled by the fine particles. Consequently, the spacing of the particles determines the size of the subgrains. 23,24) In accordance with this assumption, in the present study a smaller strain applied at a higher deformation temperature is sufficient to produce a similar ultrafine ferrite microstructure as in the previous work by Tsuji et al. 21) This is important in terms of a possible large scale industrial production of ultrafine grained steels. 4.2. Tensile Properties of the Ultrafine Grained Steels Compared with the tensile properties of the steel with a lower Mn content, there is a large enhancement of the strength in the steel with a higher Mn content and finer microstructure, Fig. 3(a). Numerous investigations have shown that the yield stress of low-carbon steel increases inversely proportional to the square root of the ferrite grain size in terms of the Hall Petch relationship. Figure 3(a) shows that the ultimate tensile stress reveals the same increase as the yield stress (e.g. 80 MPa) with increasing Mn contents from 0.74 to 1.52 mass% under the same uniform elongation, Fig. 3(b). This means that both steels have a comparable work hardening rate. As mentioned in the introduction, grain refinement typically leads to a pronounced decrease in the work hardening rate in single phase ultrafine grained steels. However, in the present case a decrease in the average ferrite grain size from 1.3 mm (for the steel with 0.74 mass% Mn) to 0.8 mm (for the steel with 1.52 mass% Mn) does not lead to a drop in the uniform elongation, Fig. 3(b). This can be attributed to the finer particles in the steel with a higher Mn content which effectively increases the work hardening rate by the accumulation of geometrically necessary dislocations in the vicinity of the fine particles. Geometrically-necessary dislocations control the work hardening of the specimen when their density exceeds that of the statistically-stored ones. They contribute to work hardening in two ways 25) : They act as individual obstacles to slip, and (collectively) by creating a long-range back-stress, with a wavelength equal to the particle spacing. 4.3. Charpy Impact Properties of the Ultrafine Grained Steels The microstructural evaluation and the tensile tests show that the increase in the Mn contents and/or the accompanying decrease in grain size improve the strength of the steels. Different effects on toughness can be expected from the grain size and the Mn contents. A reduction in average grain size commonly leads to a lower ductile-to-brittle transition temperature. For example, the transition temperature was lowered by about 40 K when the grain size was reduced from 6.8 to 1.3 mm for the lower Mn steel. 20) This can be understood in terms of cleavage crack initiation and propagation. It is known that the grain size is one of the major factors determining the cleavage fracture unit. 26,27) According to the work of Kim and Brozzo, 26,27) the cleavage fracture unit is the length of a cleavage fracture between two neighboring break-through points on its propagation direction, which is reduced as the grain size decreases. When the cleavage crack propagates through several grains, both the emission of crack-tip dislocations and the formation of cleavage facets are interrupted by the grain boundaries. If the cleavage crack moves across a high-angle grain boundary the crack front must be branched, which, together with the separation of the grain boundary between the break-through points, result in an additional fracture work. This toughening effect can lower the ductile-to-brittle transition temperature considerably. 28 30) For example, when the cleavage crack paths are carefully observed at high magnifications, it can be seen that the paths were changed by high-angle grain boundaries. Furthermore, it was observed that cleavage cracks can also be arrested by high-angle grain boundaries. 31) Thus, the decrease in grain size can limit the propagation of the initiated cleavage crack and raise the fracture toughness in the transition region. The effects of Mn on toughness are manifold. First of all, the increase in the Mn content may contribute to grain refinement, which is beneficial to improve the toughness. Secondly, the strengthening by Mn in solid solution depends on temperature. The change in the yield stress of a- iron by alloying with Mn is the highest near room temperature but it is severely reduced as temperature decreases. Below 200 K a softening effect (alloy softening) is found. 32,33) That means, the yield stress of the alloyed a- iron is below that of pure iron. In other words, when the temperature is below 200 K, alloy softening may also lead to a decrease in the ductile-to-transition temperature due to a decrease in the yield stress. In this context, the higher Mn content would contribute to an improved toughness because of alloy softening at lower temperatures. Thirdly, the large volume fraction of finer particles may result in an unfavorable effect on toughness. This is attributed to precipitation hardening which effectively increases the yield stress of the steel with a higher Mn content. Therefore, we assume that the similar Charpy impact transition curves observed for the different Mn contents (Fig. 6) may be due to the joint effect of grain refinement, alloy softening at low temperatures (positive) and precipitation hardening (negative). 1725 2005 ISIJ

5. Conclusions Two plain C Mn steels with ultrafine ferrite grains (average grain size of 1.3 and 0.8 mm for the steel with 0.74 and 1.52 mass% Mn, respectively) and homogeneously distributed carbides were produced by large strain warm deformation (e 1.6) and subsequent annealing. The ultrafine microstructures were stable against grain coarsening even during a 2 h annealing treatment at 823 K. The increase in Mn content results in a finer ferrite grain size and a more equiaxed grain shape. This is, firstly, due to the initial finer transformed ferrite pearlite microstructure in the steel with higher Mn content. Secondly, it can be attributed to the fact that Mn substitutes Fe in cementite and forms a more stable particle (Fe 3 X Mn X C with X 0.1 for the steel with 0.74 mass% Mn and X 0.25 for the steel with 1.52 mass% Mn). The fine carbides are also beneficial for the accumulation of geometrically necessary dislocations and substructures inside the ferrite grains, which facilitate grain subdivision. Therefore, a finer microstructure and a more equiaxed grain shape were observed in the steel with a higher Mn content after the deformation and annealing process. The increase in Mn contents is beneficial to improve the strength of steel. Good ductility was observed in the ultrafine grained steels, as is documented by a uniform elongation of 10% and a total elongation of 22%. This can be attributed to the additional carbides improving the work hardening capacity. Acknowledgements The authors would like to express their gratitude to the financial support of the European Coal and Steel Community (ECSC). Project title: ultrafine grained steel by innovative deformation cycles; number: 7210-PR/288. REFERENCES 1) R. Z. Valiev, R. K. Islamgaliev and I. V. Alexandrov: Prog. Mater. Sci., 45 (2000), 103. 2) R. Song, R. Kaspar, D. Ponge and D. Raabe: Ultrafine Grained Materials III, ed. by Y. T. Zhu, T. G. Langdon, R. Z. Valiev, S. L. Semiatin, D. H. Shin and T. C. Lowe, TMS, Warrendale, PA, (2004), 445. 3) P. D. Hodgson, M. R. Hickson and R. K. Gibbs: Scr. Mater., 40 (1999), 1179. 4) D. H. Shin, J. J. Paka, Y. K. Kima, K. T. Parkb and Y. S. Kimc: Mater. Sci. Eng. A, A325 (2002), 31. 5) Z. C. Wang and P. B. Prangnell: Mater. Sci. Eng. A, A328 (2002), 87. 6) K. T. Park, Y. S. Kim, J. G. Lee and D. H. Shin: Mater. Sci. Eng. A, A293 (2000), 165. 7) N. Tsuji, Y. Ito, Y. Saito and Y. Minamino: Scr. Mater., 47 (2002), 893. 8) N. Tsuji, S. Okuno, Y. Koizumi and Y. Minamino: Mater. Trans., 45 (2004), 2272. 9) T. F. Majka, D. K. Matlock and G. Krauss: Metall. Mater. Trans. A, A33 (2002), 1627. 10) J. S. Byun, J. H. Shim and Y. W. Cho: Scr. Mater., 48 (2003), 449. 11) R. Song, D. Ponge, R. Kaspar and D. Raabe: Z. Metallkd., 95 (2004), 513. 12) R. Song, D. Ponge and R. Kaspar: Steel Res., 75 (2004), 33. 13) R. Song, D. Ponge, D. Raabe and R. Kaspar: Acta Mater., 53 (2005), 845. 14) B. Jansson, M. Schalin and B. Sundman: Phase Equilibria, 14 (1993), 557. 15) R. Kaspar and O. Pawelski: Materialprüfung, 31 (1989), 14. 16) D. A. Portern and K. E. Easterling: Phase Transformations in Metals and Alloys, UK. Van Nostrand Reinhold, Workingham, England, (1981), 339. 17) Y. Zhang and D. M. Schleich: Solid State Chem., 110 (1994), 270. 18) G. Gottstein: Physical Foundations of Materials Science, Springer- Verlag, Berlin, Heidelberg, (2004), 410. 19) S. Björklund, L. F. Donaghey and M. Hillert: Acta Metall., 20 (1972), 867. 20) R. Song: Microstructure and Mechanical Properties of Ultrafine Grained C Mn Steels, Shaker Verlag GmbH, Aachen, (2005), 115. 21) N. Tsuji, R. Ueji and Y. Minamino: Proc. IF Steels 2003, ed. by H. Takechi, ISIJ, Tokyo, (2003), 347. 22) F. J. Humphreys and M. Hatherly: Recrystallization and Related Annealing Phenomena, Elsevier Science Ltd, Oxford, (1995), 164. 23) A. Gholinia, F. J. Humphreys and P. B. Prangnell: Acta Mater., 50 (2002), 4461. 24) L. Storojeva, D. Ponge, R. Kaspar and D. Raabe: Acta Mater., 52 (2004), 2209. 25) M. F. Ashby: Philos. Mag., 21 (1970), 399. 26) S. Kim, Y. R. Im, S. Lee, H. C. Lee, Y. J. Oh and J. H. Hong: Metall. Mater. Trans. A, A32 (2001), 903. 27) P. Brozzo, G. Buzzichelli, A. Mascanzoni and M. Mirabile: Met. Sci., 11 (1977), 123. 28) D. Hulla: Acta Metall., 8 (1960), 11. 29) R. O. Ritchie, J. F. Knott and J. R. Rice: Mech. Phys. Solids, 21 (1973), 395. 30) J. F. Knott: Pressure Vessels Piping, 64 (1995), 225. 31) S. Kim, S. Lee and B. S. Lee: Mater. Sci. Eng. A, A359 (2003), 198. 32) W. C. Leslie: Metall. Trans., 3 (1972), 5. 33) E. Pink: Z. Metallkd., 64 (1973), 871. 2005 ISIJ 1726