Recentprogress:kinetics of thebainite transformation in steels
Manabu Takahashi *
Stel Rsrh Larrres Ni St/ Corton St 20-1 Fats Ch 293-8511. Jap
Received 15 August 2004; received in revised form 19 August 2004; accepted 23 August 2004
Abstraet
s fully accepted experimentally. The nucleation-controlled model based on the displacive and diffusionless nature of the bainite trans adopt the solute drag efeet and a quantitative approach is proposed although the difusion of substitutional alloying elements is notformation bined with the kinetics of cementite formation in austenite and in ferrite as the simultaneous reactions. 2004 Elsevier Ltd. All rights reserved.
Keywordx: Bainite; Bainite transformation; Kinetics; T curve; Transformation stasis; Solute drag
1. Introduction
ment conditions for a given production line. Even when steel contains substitutional alloying elements such as Siand Al which are known to retard the formation of cementite the untransformed austenite could de-pose to cementite and ferrite during extended holdingat bainite transformation temperatures. It is therefore essential to model the bainite reaction and the cementiteformation individually to permit the prediction of a ‘processing window' where it is possible to obtain themixed microstructure of ferrite bainite and retainedmation are still not well established as Bhadeshia austenite. However the kinetics of the bainite transfor-nite transformation [5]. In this review the author will try pointed out in his article on unresolved issues in the bai-to assess the recent work that is considered to contributetransformation in steels. the quantitative expression of the kinetics of the bainite
of the intensely discussed phenomena in steels. The The bainite transformation is as well recognized onemajor interest of the work is focused not only on themechanism of the bainite transformation itself but also on a quantitative expression of the kinetics of the bainitetransformation. As it is well known the bainite transfor- mation has been widely exploited in industry to improvethe properties of steels. One of the successful applica-tions of the bainite transformation is so called low-al- loyed TRIP steels [14]. An enrichment of carbon intomass% is maintained in the absence of cementite precip- untransformed austenite up to more than about 1itation during the bainite transformation. The untrans-formed austenite then remains stable even at ambient temperature. Since the rate of the bainite reaction de-pends strongly on the chemical position of the austenite at the beginning of the reaction the fulldescription of the kinetics of phase transformations islect the right alloy chemistry and appropriate heat treat- extremely useful from an engineering point of view to se-
2. Maximum degree of the reaction
It is always very important to know the maximumpossible degree of the transformation as a function ofthe reaction temperature. When bainitic ferrite forma- tion is acpanied by the cementite precipitation the
of bainitic ferrite can be casily calculated from the car- transformation is known to plete. And the amountbon content of the steel. It is not however the casewhere cementite precipitation is retarded effectively. When the formation of cementite is separated from thatof bainitic ferrite by the addition of Si or Al a two-stepreaction can be observed. The first step corresponds to the formation of bainitic ferrite and the onset of the sec-ond step is recognized as due to the start of cementite precipitation from carbon-supersaturated austenite.
The maximum degree of the first step is extremelyimportant since it determines the amount of retained austenite in low-alloyed TRIP steels. Two deferentmechanisms which govern the maximum degree of the reaction have been discussed. One is the effect of partic-ular alloying element on the movement of the transfor- mation interface and the other is the diffusionlesstransformation nature of bainite. In principle the recon-structive formation of ferrte procceds until the equilib- rium or para-equilibrium condition is maintained whenthe formation of cementite is absent. In order to explain the existence of the transformation stasis in the recon-structive reaction an additional mechanism such as sol-ute drag effect is always needed to introduce. Wu and his co-workers [*6] tried to explain the transformation stasisquantitatively by introducing the drag effect of Mo on the interface movement. Although they adopted 20 or30kJmol′ of the binding energy of Mo at the centerof the boundary the result showed only a qualitative agreement of the tendency of the degrees of the reactionas a function of the reaction temperature. Further improvement of the model may be nceded to deal quan-titatively with the effect of the various kinds of substitu-reaction. Furthermore there is a discrepancy among tional alloying element on the maximum degree of thethe measurements of substitutional alloying element segregations to the transformation interfaces. Althoughan extended segregation of Mo at the vicinity of aboundary is reported as a result of a STEM-EDX analysis [7] vast numbers of finest scale analyses usingAP-FIM preclude the existence of any segregation of substitutional alloying element [812].
On the other hand the calculation of the maximumpossible degree of the diffusionless ferrite formation is straightforward. Observed carbon concentrations in re-tained austenite are reported to agre well with the cal- culated T curve in different steels [13 14]. The negativeslope of the To curve predicts the carbon concentrationin residual austenite to be greater at lower bainite trans- formation temperatures as is observed experimentally.The effect of alloying elements on the carbon concentra- tion in retained austenite can also be explained as theireffect on the To curves [15]. Most of the alloying ele-ments decrease the carbon concentration at the T curve except the case of Si Al and Co. A1 [16 17] and Co [17]increase the carbon concentration at the T curve and
Si hardly changes it. This tendency quantitativelycorresponds to the effect of those alloying element on the maximum possible carbon concentration in re-tained austenite namely the maximum degree of the reaction.
3. Overall reaction kineties
nucleation and growth mechanism the overall reaction Since the bainite transformation proceeds by therate depends on the driving force and the mobilityAlthough the transformation pletes when the resulting in a C-curve nature of the TTT diagram.cementite formation is allowed to occur it is essential to treat the formations of bainitic ferrite and cementitecope with a large variety of alloy chemistries that are individually and allow them to pete cach other toadopted in industries. There are two major approachesto express the overall kinetics of the bainite transforma- tion one of which is the diffusion controlled growthmodel of the bainitic laths and the other is the nuclea-tion of bainitic ferrite plates controlled model.
3.1. Diffusion controlled growth model
The growth rate measurements of bainite bainiticferrite and sub-unit have been reported for various - binations of alloy chemistry. There seem to be big dis-Trivedi's model is reported to be applicable to predict crepancies among these measured growth rates.the growth rate of the bainite laths [18] although theculated one. It is however worth noting that bainite measured growth rate is somewhat smaller than the cal-consists of fine sub-unit of ferrite whose thickness is measured to be around 0.1 or 0.2 μm at 400°C and de-creases with the transformation temperature and anincreasing the alloy carbon content [19]. This result is consistent with the analysis made by Singh and Bhades-hia [20]. This fine structure of bainite was also repro- duced in a thre-dimensional way [*6 *21] supportingthe autocatalytic nucleation of ferrite. From these inves-tigations the growth rate discussed by Quidort and Bre- chet [18] could be that of sheaves of bainite rather thansub-units. Observed larger growth rates in the low Si steel can be understood in a different manner from theirdiscussion. Because of the existence of cementite precip-itation carbon supersaturation in untransformed auste- nite is reduced and then the total amount of bainitewas analyzed to be larger growth rates. In the case of increases. This results in larger sheaves of bainite thatthe diffusion controlled growth model of bainite laths ost9] it is always necessary to adopt the solute drag effectelement such as Mo is reported vast numbers of finest
scale analyses using AP-FIM support no segregation ofinterfaces [812] substitutional alloying element at the transformation
and the carbon concentration in the untransformedis inevitable within a realistic period of time. In this case austenite. In general however the carbide formationfrom the untransformed austenite and the sub-units of the model should include the formation of carbide bothbainite which petes with the formation of the sub- units. The overall process is illustrated in Fig. 1 [31].Bainite nucleates with the partitioning of carbon butgrows without any partitioning. The nucleation theory [32] of bainite was confirmed to be applicable to highcarbon steels [33] that is important when the carbonformed austenite at ambient temperature is required. enrichment up to an amount to stabilize the untrans-So the sub-unit of bainite is fully carbon supersaturated at the growth stage. Carbon atoms are however evacu-ated to the surrounding austenite after the formation be- cause the reaction temperature is high enough forcarbon to diffuse along a long distance. This carbonredistribution can cause a carbon enrichment of the untransformed austenite. When the carbide forms insidepleted the lower bainite is obtained whereas car- the sub-unit of bainite before the carbon evacuation isbide precipitates from the carbon-supersaturated auste-Azuma et al. recently modeled whole of this process nite resulting in the upper bainite microstructure.the formation of bainitic ferrite was adopted. And successfully [**34]. Rees and Bhadeshia's model forfor the cementite precipitation both in austenite and in ferrite. Avrami's extended volume concept was modifiedThis model can be used to predict the overall transfor- [35 36] to deal with the simultaneous transformations.mation kinetics with and without the cementite precipi-tation. As a result the model can reproduce the two-step reaction at 450°C and appearance of the lower bainite at
formers such as Mo and W are included alloy carbides In particular alloy systems in which strong carbideformation reported to occur at the transformation inter- faces which suggest a long range diffusion of substitu-tional alloying element during or after the formationquence of the reactions will be needed. of bainite [25 26]. Detailed investigations of the se-
3.2. Nucleation controlled model
sionless nature of the formation of sub-units of bainitic Another model is based on the displacive and diffu-ered to be large enough to simplify the overall transfor- ferrite. Since the growth rate of the sub-unit is consid-mation rate being controlled by the successivenucleation of bainitic laths or plates. Quantitative expression of this model is first reported by Bhadeshia[28 29] This approach is applicable to steels where no [27] and latterly modified by Rees and Bhadeshiacarbide precipitation is expected to occur until the car-bon concentration in untransformed austenite reaches the To curve above which the diffusionless transforma-tion is thermodynamically impossible. In this model it is essential to determine the size of sub-unit of bainite[19 20]. A theoretical approach was conducted recentlysub-units is determined as a point where the chemical by Van Dooren et al. [**30] in which the thickness offree energy released by the formation of a sub-unit is equal to the mechanical energy dissipated in the auste-nite due to the growth of a sub-unit. When the forma-tion of carbide is inhibited these models can be applied to predict the degree of the transformation
Fig. 1. Schematic illustration of the overall process of tbe bainite transformation in steels.
Fig. 2. Calculated change in volume fractions of bainitic ferrite cementite (q) and austenite in a Fe0.6% C1.5% Si1.5%Mn alloy transformed at 300 and 450°C. Plots are experimentally observedvolume fraction munication). of bainitic ferrite (M Azuma.personal
350°C in Fe0.6% C1.5% Si1.5%Mn alloy as shown inFig. 2 (M. Azuma personal munication). Effect of Si on the retardation of the cementite precipitation canbe understood as the result of a diminished driving force for the formation of para-equilibrium cementite. In thecase of the bainite transformation in low-alloyed TRIPsteels the sub-unit of bainite nucleates side by side due to a geometrical limitation of small austenite grains[*37 *38]. It is therefore important to take into account the nucleation of sub-units both at the austenite grainboundaries and at the sub-units. Matsuda and Bhades-hia [*39] reported a model that deals with these two dif- ferent nucleation sites in carbide free bainiticapplied to the microstructural development in low-al- microstructures. This approach could be extended andloyed TRIP steels. Further investigations are needed.
4. Effect of pre-deformation and external stress
There is a continued interest in the effect of externalstresses and pre-deformations on the kinetics of the bai- nite transformation [40] since the grain refinement of ashot rolled multiphase steels including low-alloyed TRIPsteels is one of the successful approaches to improve mechanical properties. It was confirmed experimentallythat the plastic deformation of austenite reduces the
5. Conclusions
Acknowledgement
References
total amount of the bainite transformation due tomechanical stabilization of austenite [4144] although external stresses accelerate the bainite transformation[40 41]. The effect of pre-deformation on the kinetics of the bainite transformation is rather plicated.retarded [40]. The retardation of the overall reaction The transformation rate is first accelerated and finallywas again confirmed experimentally [4244]. Quantita-external stresses is needed to develop. tive treatment of the effect of pre-deformations and
Modeling of the kinetics of the bainite transforma-petition of bainitic ferrite and cementite can be pre- tion has e to the stage where the formation anddicted quantitatively. The displacive and difusionlessnature of the bainite transformation is now widely ac- cepted since the theory can predict quantitatively theproducts obtained in actual production lines. However amount and the chemical positions of the reactionfurther developments are required for the model to dealIt is also worth noting that the kinetics of the lower bai- with the effect of pre-deformations and external stresses.nite transformation just above the Ms temperature is an area where more detailed investigations are nceded. Aninteraction between the formations of bainite and mar-of the overall reaction rate at the temperatures [45]. tensite could play an important role in determination
The author is grateful to Prof. H.K.D.H. Bhadeshiasions on the manuscript. and Dr. N. Fujita for their useful ments and discus-
The papers of particular interest have been high-lighted as:** of very special interest. of special interest;
[1] Matsumura O Sakuma Y Takechi H Enhancement ofelongation by retained austenite in intercritical annealed 0.4C1.5Si0.8Mn steel. Trans ISIJ 1987;27:5709.[2] Matsumura O Sakuma Y Takechi H. TRIP and its kinetic aspects in austempered 0.4C-1.5Si-0.8Mn steel. Scripta Metall[3] Sakuma Y Matsumura O Akisue O. Influence of C content 1987;21:13016.and annealing temperature 0 on microstructure and mechanical propertiesof 400 C transformed steel containing retained[4] Kawano O Takahashi M Wakita J Esaka K. Abe H. austenite. ISIJ Int 1991;31:134853.Improvement of ductility by retained austenite in bot rolled
[23] Quidort D Brechet Y. The role of carbon on the kinetics of o ss o q[24] Enomoto M. Partition of carbon and alloying elements during the growth of ferrous bainite. Scripta Mater 2002;47:1459.[25] Shiflet GJ Hackenberg RE. Partitioning and the growth of bainite. Seripta Mater 2002;47:1637.[26] Hackenberg RE Shiflet GJ. A microanalysis study of the bainite reaction at the bay in FeCMo. Acta Mater 2003;51:[27] Bhadeshia HKDH. Overall transformation kinetic. Journal 213147.de physique colloque C4 Supplement au No.12 1982;43:437 41.[28] Rees GI Bhadeshia HKDH. Bainite transformation kinetics Part I modified model. Mater Sci Technol 1992;8:98593.[29] Rees GI Bhadeshia HKDH. Bainite transformation kinetics Part 1 modifed model. Mater Sci Technol 1992;8:9946.[**30] Van Dooren D. De Cooman B.C. Thibaux P. Theoreticalof Int. Conf. on Advanced High Strength Steels for Automotive Applications AIST 2004;24757.[31] Takahashi M Bhadeshia HKDH. Model for transition from upper to lower bainite. Mater Sci Technol 1990;6:592603.[32] Bhadeshia HKDH A rationalization of shear transformation in steels. Acta Metall 1981;29:111730.[33] Garcia-Mateo C Bhadeshia HKDH. Nucleation theory for high-carbon bainite. Mater Sci Eng A 2004;378:28992.xddn u1 un se N mn nzy [] and lower bainite transformation in steels. Mater Sci Forum[35] Robson JD Bhadeshia HKDH. Kinetics of precipitation in 2003;426432:140512.[36] Jones SJ. Bhadeshia HKDH. Kinetis of the simultaneous power plant steels. Calphad 1996;20:44760.deposition of austenite Into several transformation prod- ucts. Acta Mater 1997:.[°37] Girault E Jacques P Ratchev P Van Humbeeck J Verlinden B Aermoudt E. Studly of the temperature dependence of thebainitic transformation rate in a multiphase TRIP-assisted steel. Mater Sci Eng A 1999;273275:4714.[*38] Godet S Harlet Ph Delannay F Jacques PJ. Critical assess- ment of the bainite transformation of finely grained and[*39] Matsuda H Bhadeshia HKDH. Kineties of the bainite trans- deformed austenite. Mater Sei Forum 2003;426432:14338.[40] For example see Bhadeshia HKDH. Bainite in Steels. Lon- formation. Proc R Soc London A 2004;460:170722.[41] Yao K-F Liu C Shi W Liu Z. The infuence of stress and don: Institute of Materials; 1992. 20124.strain on kinetis of bainite transformation. J Shanghai Jiaotong Univ 2000;E-5-1:2259.[42] Larn RH Yang J-R. The effect of pressive deformation of austenite on the bainitic ferrte transformation in FeMnSC[43] Godet S Harlet Ph Delannay F Jacques PJ. Critical asess- steels. Mater Sci Eng A 2000;278:27891.ment of the bainite transformation of finely grained and deformed austenite. Mater Sci Forum 2003;426432:14338.[44] Godet S Jacques PJ Harlet Ph Delannay F. Simulation of the hot rolling of TRIP-assisted multiphase steels. Mater Sci Forum[45] Chang LC. Microstructures and reaction kinetics of bainite 2003;426432:12738.transformation in Si-rich steels. Mater Sci Eng A 2004;368: 17582.
high strength steels. In: Proc. of Int. Conf. On Phys. Metall. on Thermomech. Proc. of Stels and Other Metals ISU[5] Bhadeshia HKDH. The bainite transformation: unresolved 1988;2:6928.[*6] Wu KM Kagayama M. Enomoto M. Kineties of ferite issues. Mater Sci Eng A 1999;273275:5866.transformation in an Fe0.28mass% C3mass%Mo alloy. Mater Sei Eng A 2003;343:14350.[7] Enomoto M Maruyama N Wu KM Tarui T. Alloying element accumulation at ferritelaustenite boundaties belowMo alloy. Mater Sei Eng A 2003;343:1517. the timetemperature-transformation diagram bay in an FeC[8] Peet M. Babu SS Miller MK. Bhadeshia HKDH. Three- u uop qe o sseu aqod e [9] Bhadeshia HKDH Waugh AR. Bainite: An atom-probe study temperature bainite. Scripta Mater 2004;50:127781.of the inplete reaction phenomenonActa Metall 1982:30:77584.[10] Stark I Smith GDW Bhadeshia HKDH. The distribution of substitutional alloying element during the bainite transforma-[11] Bhadeshia HKDH Waugh AR. An atom-probe study of tion. Metall Trans A 1990;21A:83744.bainite. In: Proc. of Int. Conf. on Solid-Solid Phase Transfor- mations ASM 1981:9938.[12] Josefsson B Andren HO. Microstructure and thermodynamic behavior of a chromium molybdenum submerged arc weldmetal in the as welded state Advances in the science and technology of welding. ASM Int 1989:2437.[13] Jacques P Harlet P Delannay F. Critical assessment of the phase transformations occurring during the heat treatment ofAided High Strength Ferrous Alloys GRIPS' Sparkling World TRIP-assisted multiphase steels. In: Proc. Int. Conf. on TRIP-[14] Matsuda H. Kitano F. Hasegawa K. Urabe T Hosoya Y. of Steel 2002;1:12934.sheet. In: Proc. Int. Conf. on TRIP-Aided High Strength Ferrous Metallurgy of continuously annealed high strength TRIP steel[15] Takahashi M Bhadeshia HKDH. A model for the microstruc- Alloys GRIPS' Sparkling World of Steel 2002;1:11319.ture of some advanced bainitic steels. Mater Trans JIM 1991;32:68996.[16] Traint S Pichler A Hauzenberger K Stiaszny P Wermer E. Influence of siicon aluminum phosphorus and copper on hephase transformation of low alloyed TRIP-steels. In: Proc. Int. Conf. on TRIP-Aided High Strength Ferrous Alloys GRIPS[17] Garci-Mateo C Caballoro FG Bhadeshia HKDH. Acceler. Sparkling World of Steel 2002;1:1218.[18] Quidort D Brechet Y. The role of carbon on the kinetics of ation of low-temperature bainite ISIJ Int 2003;43:18215.[19] Chang LC. Microstractures and reaction kinetics of bainite bainite transformation in steels. Scripta Mater 2002;47:1516.transformation in Si-rich steels. Mater Sci Eng A 2004;368: 17582.[20] Singh SB Bhadeshia HKDH Estimation of bainite plate- thickness in low-alloyed steels. Mater Sci Eng A 1998;245:729.[*21] Wu KM Enomoto M. Three-dimensional morphology of degenerate ferrite in an FeCMo alloy. Scripta Mater[22] Aaronson HI Reymolds WT Purdy GR. Coupled-solute drag 2002;46:56974.effeets on Ferrite formation in FeCX systems. Metall Matet Trans A 2004;35A:.
