August 2, 2005
Effective Grain Size and Charpy Impact Properties of High-Toughness X70 Pipeline Steels
The correlation of microstructure and Charpy V-notch (CVN) impact properties of a high-toughness API X70 pipeline steel was investigated in this study. Six kinds of steel were fabricated by varying the hot-rolling conditions, and their microstructures, effective grain sizes, and CVN impact properties were analyzed. The CVN impact test results indicated that the steels rolled in the single-phase region had higher upper-shelf energies (USEs) and lower energy-transition temperatures (ETTs) than the steels rolled in the two-phase region because their microstructures were composed of acicular ferrite (AF) and fine polygonal ferrite (PF). The decreased ETT in the steels rolled in the single-phase region could be explained by the decrease in the overall effective grain size due to the presence of AF having a smaller effective grain size. On the other hand, the absorbed energy of the steels rolled in the two- phase region was considerably lower because a large amount of dislocations were generated inside PFs during rolling. It was further decreased when coarse martensite or cementite was formed during the cooling process.
I. INTRODUCTIONMOST pipeline steels recently produced have been increasingly getting stronger, tougher, thicker, and larger in order to increase the transportation efficiency and to reduce the cost by long- distance transportation of a large amount of natural gas or crude oil under high pressure. In addition, as the gas composition has become richer and many drilling activities have been undertaken in severe conditions such as extremely cold or deep regions, the demands for much stronger and tougher pipeline steels have been rising.[1-5] Since the increase in strength is generally accompanied with a considerable deterioration in toughness, it is highly important to scrutinize the structural integrity related with microstructure so as to produce new high-strength high-toughness pipeline steels. Charpy V-notch (CVN) impact test and drop-weight tear test (DWTT) have been used as important testing methods to guarantee the required fracture resistance.[1,6,8] Presently, the 85 pet shear-appearance transition temperature by DWTT and the CVN upper-shelf energy (USE) are most widely used as standards to evaluate the fracture propagation transition temperature (FPTT) and the resistance to ductile fracture of pipeline steels, respectively. These testing methods have correlated well with the actual fracture propagation behavior of conventional pipeline steels having CVN USE below 100 J. However, this correlation has become less obvious due to phenomena such as the rising upper shelf,[1,8,9] separation,[7,10] and abnormal fracture appearance[7,11,12] occurring during DWTT, because the toughness of pipeline steels has greatly improved by manufacturing advancements through controlled rolling and accelerated cooling over the last decades. As a way to solve this problem, Chevron notch or static precracked DWTT specimens, whose notch is adjusted so that the fracture initiation energy is lower than that of the standard pressed notch DWTT specimen, are used for testing high-toughness pipeline steels.[9,12- 14] Various other criteria such as CVN energy for 100 pet shear area, DWTT energy, and crack-tip opening angle have been presented.[8,9,15] Nevertheless, the CVN impact test is still most widely used as a simple way for measuring the toughness and transition temperature of pipeline steels.
In the present study, based on the research results of previous studies, high-toughness pipeline steels were fabricated by optimal alloy designing and by varying the rolling conditions. Effects of microstructure on CVN energy and transition temperature were investigated. By comparing the "crystallographic unit" of microstructures determined from electron backscatter diffraction (EBSD) analysis with "morphological unit" obtained from unit crack path and cleavage facet unit of CVN specimens fractured at low temperatures, effects of effective grain size on transition temperature were characterized.
The steel used in this study was an API X70 grade steel having a nominal yield strength level of 483 MPa (70 ksi), and its chemical composition was Fe-0.05C-0.27Si-l.24Mn-0.5(Cu + Ni + Mo)-0.1(Nb + V + Ti) (wt pet). Six kinds of steels were fabricated by varying the rolling conditions, as shown in Table I and Figure 1. An overall grain refinement effect was expected by rolling with a high rolling reduction ratio in the nonrecrystallized region of austenite after austenitization at 1200 C. Rolling was finished at one of two temperatures, which were the temperature of the austenite single- phase region above Ar^sub 3^ or the temperature of the (austenite + ferrite) two-phase region below Ar^sub 3^, respectively. After the finish rolling, the steels were rapidly cooled to a finish cooling temperature (FCT) of about 400 C, 500 C, or 600 C. For convenience, the steels, which were rolled in the single-phase region and cooled to different finish cooling temperatures of 400 C, 500 C, and 600 C, are referred to as "S4,""S5," and "S6," respectively, while those rolled in the two-phase region and cooled to different finish cooling temperatures of 400 C, 500 C, and 600 C are referred to as "T4,""T5," and "T6" steels, respectively (Table I).
Table I. Rolling Conditions of the API X70 Steels
Fig. 1-Schematic illustration of the rolling conditions.
In consideration of the direction of tensile and CVN impact test specimens, the longitudinal-short transverse plane of the rolled steels was polished and etched by a 2 pet nital solution, and microstructures were observed by optical microscopy. In addition, secondary phases such as martensite were identified by a two-step etching method, and their volume fraction was measured by an image analyzer. In this two-step etching, specimens were etched first in a solution of 96 pet ethanol, 4 pet picric acid, and a few drops of hydrochloric acid, and were then etched in a solution of sodium hydroxide and metabisulfate. This way, secondary phases can be discerned. Tensile bars with a gage diameter of 6 mm and a gage length of 30 mm were prepared in the transverse direction, and were tested at room temperature at a crosshead speed of 5 mm/min.
Charpy impact tests were performed in the temperature range from - 196 C to 20 C by a Tinius Olsen impact tester (JT Toshi, Japan) of 500 J capacity on subsize CVN bars with a 7.5 10 55-mm size, which were machined in the transverse-longitudinal orientation. In order to reduce errors in data interpretation, a regression analysis for absorbed impact energy vs test temperature was done by a hyperbolic tangent curve fitting method. Based on the data from the regression analysis, the energy-transition temperature (ETT), which corresponds to the average value of USE and lower-shelf energy, was determined. The fracture-appearance transition temperature (FATT) at which the area fractions of the cleavage and ductile shear fracture modes were 50 pet was also determined from the observation of fractured CVN bars. In order to examine the cleavage fracture unit and crack propagation path, the fracture surface and the cross-sectional area beneath the fracture surface of the CVN specimens fractured at - 196 C were observed by a scanning electron microscope (SEM) after the fracture surface was coated by nickel.
The EBSD analysis (resolution 0.2 m) was conducted on the cross- sectional area beneath the fracture surface of CVN specimens fractured at -196 C by a field emission -scanning electron microscope (FE-SEM, model S-4300SE, Hitachi, Japan). The data were then interpreted by orientation imaging microscopy analysis software provided by TexSEM Laboratories, Inc.
A. Microstructure and Cristallographie Orientation
Figures 2(a) through (f) are optical micrographs of the steels rolled in the single- and two-phase regions, and the basic microstructures and volume fraction of each phase are summarized in Table II. The steels rolled in the single-phase region are mostly composed of acicular ferrite (AF) and polygonal ferrite (PF), and about 5 vol pet of martensiteaustenite (MA) constituent is present (Figures 2(a) through (c)). They show almost no microstructural variations with FCT. The S4 steel contains somewhat more AF than the S5 steel because of the rather faster cooling rate during accelerated cooling (Figures 2(a) and (b) and Table I). The S6 steel shows a slightly coarsened ferrite microstructure as some ferrites were grown (Figure 2(c)). The volume fraction of PF increases with increasing FCT, whereas the fraction of AF decreases (Table II). In the steels rolled in the two-phase region, the volume fraction of PF transformed before or during finish rolling exceeds 85 pet (Figures 2(d) through (f)), and the kind of the other phases transformed from retained austenites during or after cooling varies with the FCT. Five volume percent of martensite is present in the T4 steel (Figure 2(d)). In the T5 and T6 steels, on the other hand, pearlite is found, together with a small amount of martensite (Figures 2(e) and (f)). The volume fraction of s\econd phases tends to decrease with increasing FCT (Table II).
The SEM micrographs and the results of the EBSD analysis of the S5 and T5 steels are provided in Figures 3(a) through (d). Grain boundaries are displayed in different colors according to the grain boundary misorientation, as shown in Figures 3(b) and (d). There are few grain boundaries having misorientations of 1 to 5 deg in the two steels. In the S5 steel, the boundaries between AFs having different orientations are high-angle boundaries of 15 deg or higher, whereas boundaries between those having similar orientations are mostly low- angle ones of 5 to 15 deg (Figures 3(a) and (b)).[20-23] The boundaries between PFs or between AF and PF are mostly high-angle ones of 15 deg or higher. In the T5 steel, the boundaries between PFs are high-angle ones, whereas boundaries between sub-boundaries formed inside PFs are low-angle ones (Figures 3(c) and (d)). Generally, high-angle grain boundaries of 15 deg or higher obtained from EBSD can be used as a crystallographic domain parameter representing the effective grain size.[20,21] The effective grain sizes of the steels rolled in the single-and two-phase regions are measured to be 3.5 and 5.1 m, respectively. This indicates that the steels rolled in the singlephase region expect to have better low- temperature toughness than those rolled in the two-phase region because of the smaller effective grain size.
Fig. 2-Optical micrographs of the (a) S4, (b) S5, (c) S6, (d) T4, (e) T5, and (J) T6 steels. Nital etched.
Table II. Basic Microstructures and Volume Fractions of Various Phases of the API X70 Steels
Distributions of grain boundary misorientation and  pole figure of the S5 and T5 steels are shown in Figures 4(a) and (b). In the S5 steel rolled in the single-phase region, the relative ratio of high-angle vs low-angle boundaries is lower than in the T5 steel rolled in the two-phase region. This is because the S5 steel containing a number of AFs (Figure 2(b)) shows a higher fraction of low-angle misorientation found between parallel AFs, whereas the T5 steel contains a number of PFs (over 85 vol pet) having highangle boundaries. This can also be confirmed in terms of the  pole figure. In the S5 steel, some of the points are aggregated in certain areas, whereas they are more or less evenly dispersed in the T5 steel. Thus, the relative fraction of high-angle boundaries of the S5 steel is lower than that of the T5 steel, but the effective grain size of the S5 steel is comparatively smaller than that of the T5 steel because of the presence of more AF and PF, which are finer than PF formed in the T5 steel.
B. Tensile Properties
Room-temperature tensile test results of the steels rolled in the single- and two-phase regions are listed in Table III. All the steels show yield strengths over 483 MPa (70 ksi), satisfying the strength requirement of APi X70 grade pipeline steels. The S4 through S6 steels show lower yield and tensile strengths and higher elongation by about 3 pet than the T4 through T6 steels, and their yield ratios (yield ratio = σ,^sub y^/σ^sub uts^) are low (below 85 pet). Since the S4 through S6 steels have similar microstructures, they show little change in tensile properties according to the FCT. In the steels rolled in the two-phase region, the T4 steel containing some martensite shows a continuous yielding behavior, but the T5 and T6 steels show a discontinuous yielding behavior. Thus, the T4 steel has higher tensile strength but lower yield strength than the T5 and T6 steels, and shows lower yield ratio. The T5 and T6 steels show high yield ratios of about 90 pet, which are considerably higher than that of the T4 steel or the S4 through S6 steels.
C. CVN impact Properties
Figure 5 presents the CVN energy data, and the results of USE, FATT, and ETT obtained from them are summarized in Table IV. The S4 through S6 steels show excellent CVN . properties because they have lower ETT (below -130 C) and higher USE than the T4 through T6 steels, irrespective of the FCT, and the dependence of USE, energy tested at - 20 C, ETT, and FATT are the same in the three steels. The CVN properties of the S4 through S6 steels do not show much difference because of little microstructural variation according to the FCT. The T4 steel containing a considerable amount of martensite shows much lower USE than the T5 and T6 steels, but does not differ much in the transition temperatures of ETT or FATT. The S5 or T5 steels show the best CVN properties in the steels rolled in the single- or two-phase regions, respectively.
Fig. 3-(a) and (c) SEM micrographs and (b) and (d) misorientation maps of the (a) and (b) S5 and (c) and (d) T5 steel.
Fig. 4-Distribution of grain boundary misorientation and  pole figure of the (a) S5 and (b) T5 steels.
Table III. Room-Temperature Tensile Properties
Figures 6(a) through (f) are SEM fractographs of the CVN specimens fractured at -196 C. The cleavage fracture units of the S4 through S6 steels are smaller (below 10 m) than those of the T4 through T6 steels (10 to 15 m). These units are well related with the effective grain sizes having misorientations of 15 deg or higher, although they are somewhat larger than the effective grain sizes. The difference between the fracture unit size and effective grain size might be associated with the tilting angle of the fracture surface inside the SEM.
The SEM observations of the crack propagation path of the S4 through S6 steels and the T4 through T6 steels are shown in Figures 7(a) through (f). Here, secondary phases such as martensite were identified by a two-step etching method,'171 and are marked in Figures 7(a) through (f). The unit crack path (UCP) is generally defined as the length of a crack in which the crack propagates almost in a straight line. According to Pickering, this is known to be roughly equal to the distance between two neighboring high- angle boundaries. Many researchers have found that the UCP is almost the same as the ferrite grain size in ferrite steels, while it is 1.3 to 1.5 times as long as the packet size in bainite steels.[25,26,27] According to the observation of the crack propagation path of this study, the steels rolled in the twophase region were found to show the larger UCP than those rolled in the single-phase region. This is because the S4 through S6 steels containing fine AF have changes in crack path between AF and at AF/ PF interfaces, while many cracks of the T4 through T6 steels are in relatively coarsened PF.
Fig. 5-CVN energy vs test temperature of the (a) S4, (b) S5, (c) S6, (d) T4, (e) T5, and (f) T6 steels.
Table IV. CVN Impact Properties
Fig. 6-SEM fractographs of CVN specimens fractured at -196 C for the (a) S4, (b) S5, (c) S6, (d) T4, (e) T5, and (f) T6 steels.
Fig. 7-SEM micrographs of the cross-sectional area beneath the cleavage fracture surface of the CVN specimens fractured at -196 0C for the (a) S4, (b) S5, (c) S6, (d) T4, (e) T5, and (f) T6 steels, showing the crack propagation path. Fractured surfaces were coated by Ni. Secondary phases such as martensite can be identified by a two-step etching method.
The high-toughness pipeline steels fabricated in this study have different microstructures, depending on the rolling conditions (Figures 2(a) through (f)). Though high-toughness X70 pipeline steels are mainly composed of AFs formed by controlled rolling and accelerated cooling, they are also known to contain many other phases such as PF, quasi-polygonal ferrite, Widmansttten ferrite, granular bainitic ferrite, bainitic ferrite, and martensite, which may form during hot rolling.[16,20,28-31] Based on me resuits of microstructural and EBSD analyses, the effects of microstructural and crystallographic properties on CVN absorbed energy and transition temperature are discussed here.
The steels rolled in the two-phase region show lower USE than those rolled in the single-phase region (Table IV). This is because a considerable amount of dislocations are formed inside PF when PF transformed before the finish rolling is deformed during the finish rolling, and further because hard phases such as martensite and cementite are formed. The hardness values of PFs in the steels rolled in the singleand two-phase regions were measured to be 170 to 200 VHN and 220 to 230 VHN, respectively. This indicates that PFs in the steels rolled in the two-phase region are harder than that those in the steels in the single-phase region because of the strain hardening during the finish rolling. This is also confirmed by SEM micrographs of Figures 3(a) and (c), which show that substructures are well developed inside PFs of the steels rolled in the two-phase region, unlikely in the case of the steels rolled in the single- phase region. Since the steels rolled in the single-phase region undergo little microstructural change with the FCT, their USE values are about the same, whereas the USE of the steels rolled in the two- phase region varies widely because there are other phases formed besides PF (Figures 2(a) through (f)). Second phases in the steels rolled in the two-phase region are observed by an SEM, as shown in Figures 8(a) through (c). Most secondary phases in the T4 steel are martensites, and a small amount of AFs formed during accelerated cooling (Figure 8(a)). In the T5 steel, martensite and cementite exist, together with a small amount of AF (Figure 8(b). In the T6 steel, cementites formed by the high FCT exist in a form of pearlite, and a small amount of martensite is also observed, as shown in Figure 8(c). The USE of the T4 steel is low as martensite has formed, and the USE of the T5 and T6 steels containing cementite is higher than that of the T4 steel.
The transition temperatures of steels are known to be affected by effective grain size, which is influenced by the kind, size, and volume fraction of phases.[10,30-32] In this study, \thus, the ETT was interpreted from the perspectives of grain boundary orientation and effective grain size based on the EBSD analysis. All the steels fabricated in this study have excellent low-temperature toughness in general as they were rolled under the high rolling reduction ratio in the nonrecrystallized region of austenite to achieve grain refinement (Table IV). In fact, the ETT of the steels rolled in the single-phase region is extremely low (-130 C or lower), while that of the steels rolled in the two-phase region is relatively low (- 105 C to ~-85 C). The difference in the transition temperature of the steels rolled in the single- and twophase regions largely depends on the grain size and grain boundary misorientation. The transition temperature decreases as the effective grain size decreases with an increase in the number of high-angle boundaries of 15 deg or higher.
According to the typical phase transformation process of AFs mainly formed in the steels rolled in the single-phase region, they are first nucleated at nonmetallic inclusions inside austenites, or are nucleated secondly and grown at AF/austenite interfaces or at interfaces between AFs.[20,33-36] The term AF originated in welding and has been extended to refer to the ferrite plates nucleated at secondary phases distributed inside grains, with a wide range of morphologies. As the AF plates typically exhibit a chaotic arrangement, they are known to have different spatial orientations. According to a recent study, however, in some AFs nucleated secondly, plate-type ferrite consists of packets of parallel plates, such as those of bainites, at certain transformation temperature, chemical composition, and particle density. Consequently, parallel AF plates, similar to bainitic laths, have mostly low-angle grain boundaries, whereas AF, which is arranged differently, has mostly high-angle grain boundaries. These findings can be confirmed from Figure 3(b), in which grain boundary characteristics of the S5 steel composed of AF and PF are analyzed.
Fig. 8-SEM micrographs of the (a) T4, (b) T5, and (c) T6 steels, showing secondary phases in detail.
Since the grain boundaries of AF act as obstacles of cleavage cracks propagating at low temperatures and can effectively deflect the propagation direction, they help the transition temperature to decrease.[16,20] To verify this effect, the cross-sectional area beneath the cleavage fracture surface of the CVN specimen of the S5 steel fractured at -196C was analyzed by EBSD, and the results are presented in Figure 9. Very fine AF grains of 1 to 3 m in size are observed in the image-quality map. In the  pole figure, points are evenly dispersed overall, although some aggregated regions with points are observed, indicating the presence of many high-angle grain boundaries. According to the analysis results of the misorientation between these individual grains, most of the boundaries between fine AFs and between AF and PF are found to be high-angle ones, which work more effectively as obstacles to the cleavage crack propagation than relatively coarse PFs, because of the frequent change of the crack propagation path. This can be verified from Figures 6(a) through (c) and 7(a) through (c) showing that the cleavage fracture unit and UCP of the steels rolled in the single-phase region are smaller than those of the steels rolled in the two-phase region. Therefore, the S4 through S6 steels containing AF have a lower transition temperature by over 40 C than the T4 through T6 steels mainly composed of PF (Table IV). The ETT and FATT of the S6 steel are somewhat higher than those of the S4 and S5 steels because the S6 steel has slightly coarsened AF and PF.
As explained earlier, the steels rolled in the two-phase region are mainly composed of PF formed before or during finish rolling, regardless of the FCT, and this PF determines the overall effective grain size. Thus, the ETT does not vary much with the FCT, like those rolled in the single-phase region, but it can be somewhat affected by the FCT as the FCT affects the formation of other phases. The T4 steel containing some martensite shows much lower USE than the T5 and T6 steels, although the ETT does not show much difference. This seems to be because the martensite transformed from the fine-grained austenite does not show much difference from the effective grain size of PF. Furthermore, since a part of martensite can effectively work to prevent the cleavage crack propagation at grain boundaries with PF, they can more or less reduce the overall effective grain size. In fact, the T4 through T6 steels show less variations in the ETT than in the USE.
Fig. 9-Image-quality map and  pole figure of the cross- sectional area beneath the cleavage fracture surface of the CVN specimen fractured at -196 C for the S5 steel, showing the crack propagation path. The misorientations between neighboring individual grains are given in the table.
As discussed previously, the S4 through S6 steels rolled in the single-phase region, in which AF and PF are properly formed, show better CVN properties of higher absorbed energy and lower transition temperatures than the T4 through T6 steels mainly composed of PF, and the effects of AF on transition temperatures can be explained in detail by the EBSD analysis. In order to further clarify the effects of effective grain size on transition temperature in the future, more in-depth studies are required for determining crystallographic microstructural factors and for introducing quantitative correlations between them and the transition temperature.
In the present study, high-toughness pipeline steels having different microstructures were fabricated by varying the hot- rolling conditions, and the effects of microstructure on absorbed energy and transition temperature were investigated by the CVN impact test.
1. The steels rolled in the single-phase region composed of AF and fine PF sized smaller than 5 m showed higher absorbed energies and lower transition temperatures than the steels rolled in the two- phase region mostly composed of relatively coarser PF.
2. According to the EBSD analysis results, grain boundaries of fine AF present in the steels rolled in the single-phase region were high-angle ones, which effectively worked as obstacles to the cleavage crack propagation. This eventually led to lower transition temperatures than in the steels rolled in the two-phase region predominated with coarse-grained PF.
3. The absorbed energy of the steels rolled in the two-phase region was considerably low because a large amount of dislocations were generated inside PFs during rolling. It was further decreased when some coarse martensite or cementite was formed during the cooling process.
This work was financially supported by the 2003 National Research Laboratory Program of the Ministry of Science and Technology of Korea and by POSCO under Contract No. PL-03909. The authors thank Drs. Ki Bong Kang, Dong Han Suh, and Seong Soo Ahn, POSCO, and Mr. Sang Yong Shin, POSTECH, for their help with the CVN testing and data analyses.
1. R. Denys: Pipeline Technology, Elsevier, Amsterdam, 2000, vols. I-II.
2. D.P. Fairchild, M.L. Macia, S.D. Papka, C.W. Petersen, J.H. Stevens, S.T. Barbas, N.V. Bangaru, J.Y. Koo, and M.J. Luton: Proc. Int. Pipe Dreamer's Conf., M. Toyoda and R. Denys, eds., Yokohama, Japan, 2002, pp. 307-21.
3. K.T. Corbett, R.R. Bowen, and C.W. Petersen: Int. J. Offshore Polar Eng., 2004, vol. 14, pp. 75-80.
4. M. Matsuda and H. Miura: Met. Mater. Int., 2003, vol. 9, pp. 537-42.
5. C.-S. Oh, H.N. Han, C.G. Lee, T.-H. Lee, and S.-J. Kim: Met. Mater. Int., 2004, vol. 10, pp. 399-406.
6. Recommended Practice for Conducting Drop-Weight Tear Tests on Line Pipe, API Recommended Practice 5L3, American Petroleum Institute, 1996.
7. G. Mannucci and D. Harris: "Fracture Properties of API X100 Gas Pipeline Steels," Final Report, European Commission, Brussels, Belgium, 2002.
8. D.J. Horsley: Eng. Fract. Mech., 2003, vol. 70, pp. 547-52.
9. G.M. Wilkowski, W.A. Maxey, and R.J. Eiber: Can. Metall. Q., 1980, vol. 19, pp. 59-77.
10. I. Tamura, H. Sekine, T. Tanaka, and C. Ouchi: Thermomechanical Processing of High-Strength Low-Alloy Steels, Butterworth & Co., Ltd., London, 1988.
11. E. Heier: "Drop Weight Tear Testing of High Toughness Pipeline Steel," Technical Report, DET NORSKE VERITAS, Norway, 2003.
12. B. Hwang, S. Lee, Y.M. Kim, N.J. Kim, J.Y. Yoo, and C.S. Woo: Mater. Sci. Eng. A, 2004, vol. A368, pp. 18-27.
13. H. Kashimura, M. Ogasawara, and H. Mimura: Met. Progr., 1976, Nov., pp. 58-62.
14. T. Ishihara, J. Kondo, T. Kitada, and T. Akiyama: Trans. Iron Steel Inst. Jpn., 1987, vol. 27, pp. 219-21.
15. J.C. Newman, Jr., M.A. James, and U. Zerbst: Eng. Fract. Mech., 2003, vol. 70, pp. 371-85.
16. B. Hwang, S. Lee, Y.M. Kim, N.J. Kim, and S.S. Ahn: Metall. Mater. Trans. A, 2005, vol. 36A, pp. 725-39.
17. A.K. De, J.G. Speer, and D.K. Matlock: Adv. Steels Processing, 2003, vol. 161, pp. 27-30.
18. ASTM Standard E23-02, ASTM, Philadelphia, PA, 2002.
19. W. Oldfield: ASTM Standardizations News, 1975, pp. 24-29.
20. M. Diaz-Fuentes, A. Iza-Mendia. and I. Gutierrez: Metall. Mater. Trans. A, 2003, vol. 34A, pp. 2505-16.
21. J.Y. Koo, M.J. Luton, N.V. Bangaru, R.A. Petkovic, D.P. Fairchild, C.W. Petersen, H. Asahi, T. Hara, Y. Terada, M. Sugiyama, H. Tamehiro, Y. Komizo, S. Okaguchi, M. Hamada, A. Yamamoto, and I. Takeuchi: Int. J. Offshore Polar Eng., 2004, vol. 14, pp. 2-10.
22. F.B. Pickering: Mater. Sci. Technol., 1993, vol. 7, pp. 45- 94.
23. A.F. Gourgues, H.M. Flower, and T.C. Lindley: Mater. Sci. Technol., 2000, vol. 16, pp. 26-40.
24. F.B. Pickering: Proc. Symp. on Transformation and Hardenability in Steels, Climax Molybdenum Co. and The University of Michigan, Ann Arbor, MI, 1967, pp. 109-29.
25. P. Brozzo, G. Buzzichelli, A. Mascanzoni, and M. Mirabile: Met.Sci., 1977, pp. 123-29.
26. Y. Ohomori, H. Ohtani, and T. Kunitake: Met. Sci., 1974, vol. 8, pp. 357-66.
27. J.P. Naylor and P.R. Krahe: Metall. Trans., 1974, vol. 5, pp. 1699-1701.
28. H.N. Han, C.-S. Oh, D.W. Sun, C.G. Lee, T.-H. Lee, and S.-J. Kim: Met. Mater. Int., 2004, vol. 10, pp. 221-29.
29. T.-H. Lee, C.-S. Oh, C.G. Lee, S.-J. Kim, and S. Takaki: Met. Mater. Int., 2004, vol. 10, pp. 231-36.
30. N.J. Kim: J. Met., 1983, vol. 35, pp. 21-27.
31. J.Y. Yoo and J.S. Woo: Proc. Int. Pipe Dreamer's Conf., M. Toyota and R. Denys, eds., Scientific Surveys, Ltd., Yokohama, Japan, 2002, pp. 441-56.
32. D. Bhattacharjee, J.F. Knott, and C.L. Davis: Metall. Mater. Trans. A, 2004, vol. 35A, pp. 121-30.
33. J.Y. Yang and H.K.D.H. Bhadeshia: Mater. Sci. Technol., 1989, vol. 5, pp. 93-97.
34. F.J. Barbara, P. Kraulis, and K.E. Easterling: Mater. Sci. Technol., 1989, vol. 5, pp. 1057-68.
35. I. Madariaga, I. Gutierrez, and H.K.D.H. Bhadeshia: Metall. Mater. Trans. A, 2001, vol. 32A, pp. 2187-97.
36. E. Sarath Kumar-Menon and H.I. Aaronson: Acta Metall., 1987, vol. 35, pp. 549-63.
BYOUNGCHUL HWANG, Post-doctoral Research Associate, YANG GON KIM, Research Assistant, SUNGHAK LEE, Professor, YOUNG MIN KIM, Research Assistant, and NACK J. KIM, Professor, are with the Center for Advanced Aerospace Steels, Pohang University of Science and Technology, Pohang 790-784, Korea. JANG YONG YOO is with the Plate Research Group, Technical Research Laboratories, POSCO, Pohang 790- 785, Korea. Contact e-mail: [email protected]
Manuscript submitted August 6, 2004.
Copyright Minerals, Metals & Materials Society Aug 2005